WHY PEACE FAILS

This page intentionally left blank

WHY PEACE FAILS
The Causes and Prevention of
Civil War Recurrence

C H A R L E S T. C A L L

GEORGETOWN UNIVERSITY PRESS

Washington, DC
Georgetown University Press, Washington, DC
www.press.georgetown.edu

© 2012 by Georgetown University Press. All rights reserved. No

part of this book may be reproduced or utilized in any form or
by any means, electronic or mechanical, including photocopying
and recording, or by any information storage and retrieval system,
without permission in writing from the publisher.

Library of Congress Cataloging-in-Publication Data

Call, Charles.
Why peace fails : the causes and prevention of civil war
recurrence / Charles T. Call.
p. cm.
Includes bibliographical references and index.
ISBN 978-1-58901-894-5 (pbk. : alk. paper)
1. Peace—Political aspects. 2. War—Causes. 3. Peace-building.
4. War—Causes—Case studies. 5. Peace-building—Case studies.
I. Title.
JZ5538.C34 2012
303.6'4—dc23
2011036106

o This book is printed on acid-free paper meeting the requirements

of the American National Standard for Permanence in Paper for
Printed Library Materials.

15 14 13 12 9 8 7 6 5 4 3 2
First printing

Printed in the United States of America

 To Tracy
and
Shayla, Dash, and Jag
This page intentionally left blank
Contents

List of Tables x
Acknowledgments xi

Introduction: The Tragedy of Civil War Recurrence 1

The Importance of This Book 2
The Central Argument 3
Contributions to Theory 6
Research Design and Methodology 7
Organization of the Book 17
Notes 20

Part I. Why Peace Fails: Theory

1 What Do We Know about Why Peace Fails? 25

What We Know about Civil Wars and Ethnic Conflict 26
Four Approaches to Peacebuilding 30
Clarifying Concepts: Exclusion, Inclusion, and Legitimacy 36
Conclusion 47
Notes 48

2 Is Civil War Recurrence Distinct from Its Onset? A Quantitative

Analysis and the Limits Thereof 50
A Regression Analysis of Civil War Recurrence 51
The Contributions and Limitations of Quantitative Methods for
Studying Civil Wars 59
Conclusion 65
Notes 66
viii Contents

Part II. Examining the Cases

3 Liberia: Exclusion and Civil War Recurrence 71

The First Civil War 72
The Onset of Peace 74
The Second Civil War: A Brief Summary 76
Charles Taylor’s Exclusionary Behavior 78
Alternative Explanations 81
Insights from Liberia’s Second Postwar Peace Process 88
Conclusion 91
Notes 94

4 Separatist Recurrences of Civil War 96

Sudan: The Marginalization of the South 100
Chechnya: Reneging and Resistance 106
Georgia and South Ossetia: Integration Backfires 110
China and Tibet: Compelled from Autonomy 115
Analyzing Cases of Reneging on Territorial Autonomy 118
Notes 120

5 Nonseparatist Recurrences of Civil War 122

Precipitating Exclusionary Behavior 123
The Central African Republic: Exclusion and State Weakness 124
Haiti: Political Exclusion and Recurrence 129
East Timor: Liberation, Statehood, and Exclusion 136
Zimbabwe: Liberation, Statehood, and Exclusion 146
Burundi and Rwanda: Chronic Exclusionary Behavior 150
Alternative Explanations and Conclusions 158
Notes 160

6 Recurrences That Defy the Argument 162

Lebanon: Failed Powersharing 163
Mali: Failed Powersharing 167
Nicaragua: Externally Driven Recurrence 173
Peru: Exclusion, Coca, and Rebel Resurgence 177
Conclusion 179
Notes 181

7 Making Peace Stick: Inclusionary Politics and Twenty-Seven

Nonrecurrent Civil Wars 183
Inclusion, Powersharing, and Peacebuilding Success 186
Powersharing and Peace Consolidation: Examining the Pool of Cases 192
Contents ix

Beyond Powersharing: Inclusionary Behavior and Peace 195

Peace and Exclusionary Behavior? 196
International Troops and “Frozen” Conflicts 202
Notes 209

Part III. Implications for Theory and Practice

8 Conclusions for Theory: Legitimacy-Focused Peacebuilding 213

The Main Findings of the Book 213
Rethinking the Aims and Approaches of Peacebuilding 218
Addressing Limitations 230
Notes 235

9 Conclusions for Policy and Practice: Can External Actors Build

Legitimacy after War? 236
Why Legitimacy Building Is Exceptionally Difficult 237
Beyond Blanket Inclusionary Formulas: Four “Moments” for Key
Choices and External Strategy 245
Conclusion 273
Notes 275

References 277
Index 303
Tables

I.1 All Civil Wars and Recurrences, 1946–99 14

I.2 Fifteen Core Cases of Recurrence 15

1.1 Risk Factors for Civil War Onset and Recurrence:

Quantitative Findings 29

2.1 Comparing Civil War Onset and Recurrence: A First Cut 53

2.2 Civil War Recurrence, Adding Ethnic Fractionalization

Squared 55

2.3 How Robust Is Ethnic Fractionalization? 56

4.1 Fifteen Core Cases of Civil War Recurrence 98

7.1 Mode of Civil War Termination by Historical Period,

All Nonrecurrent Civil Wars, 1944–2007 185

7.2 Powersharing by Type, All Cases of Recurrence and

Nonrecurrence, 1944–2007 187

7.3 Incidence of Powersharing among All Recurrences and

Nonrecurrences, 1944–2007 193

7.4 Inclusion and Peace after the Cold War: All Cases of Recurrence
and Nonrecurrence, by Political Behavior, 1988–2007 199

8.1 Key Contributions and Shortcomings of the Main Theoretical

Approaches to Peacebuilding 225
Acknowledgments

THIS BOOK WAS long in the making, and I am grateful for the support of many
institutions and individuals in that process. I was honored to receive a senior
fellowship at the US Institute of Peace in 2008–9 for this project, then titled
“Making Peace Stick.” My fellowship at the Institute was rewarding not just for
the stimulating intellectual environment but also for my supportive colleagues,
including Ginny Bouvier, Chantal de Jonge Oudraat, and Shira Lowinger.
I am also grateful for support I received as a visiting scholar at the Center on
International Cooperation of New York University. Thanks to the director, Bruce
Jones, and to the former deputy director, Rahul Chandran, for their support and
friendship.
This book benefited in many intangible ways from my close collaborations
with the International Peace Institute (IPI), especially with Elizabeth Cousens,
Jenna Slotin, and Vanessa Wyeth, as part of IPI’s peacebuilding project. Indeed
the idea for the book first emerged during a 2006 event organized by IPI for the
UN Peacebuilding Commission, its first informal meeting even before it was for-
mally convened. The policy implications of the book have been honed through
my ongoing engagement with the UN System, often through IPI and the Center
on International Cooperation.
I am grateful to my colleagues in the School of International Service at Amer-
ican University, including our encouraging deans, Lou Goodman and Maria
Green Cowles. I especially want to acknowledge my supportive colleagues in
the International Peace and Conflict Resolution Program, notably our dedicated
director, Ron Fisher, our inspirational founding director, Abdul Aziz Said, and
our talented coordinator, Becca Davis. Many thanks also to Georgetown Univer-
sity Press and especially to my editor, Don Jacobs, for his support and flexibility.
Numerous individuals contributed directly to the research and content of this
project. The economist John Schmitt worked with me to prepare a quantitative
analysis of civil war recurrence that was presented at the 2009 annual meeting
of the International Studies Association. His quantitative skills and thoughtful
analysis were invaluable in my drafting of that paper, which lay the groundwork
xii Acknowledgments

for the quantitative analysis and the analysis of quantitative methods in chap-
ter 2.
Several research assistants provided indispensable help. Yolande Bouka and
Heather Svanidze were especially helpful in researching the entire pool of civil
wars to determine which qualified as full-fledged “recurrences” and “nonrecur-
rences.” Leah Gates and Heather Svanidze provided research support with the
statistical tables and the fifteen core cases of recurrence and the twenty-seven
cases of nonrecurrence.
A number of people also provided feedback on portions of the draft, for which
I am very appreciative: Peter Andreas, Pierre Englebert, Tracy Fitzsimmons,
Leah Gates, Sanjeev Khagram, Madalene O’Donnell, Jim Ron, Susan Shepler,
and Bill Stanley. My thanks also to International Studies Association panelists
who commented on the quantitative portion: Pamela Aall, Michael Brown,
Chantal Jonge de Oudraat, and Roy Licklider. I also thank the anonymous re-
viewers whose feedback improved the manuscript, including the reviewers of an
early version of the Liberia case study for the journal Civil Wars. Of course, I am
responsible for any errors of fact or judgment.
Several friends have provided intangible but very welcome support as I
worked on the volume. I owe an enormous debt of thanks to Carol Bergman, Bob
Burke, Andy Fleischmann, David Holiday, Libby Ingham, Will Ingham, Sanjeev
Khagram, Robert Lynch, Caroline Moser, Madalene O’Donnell, Peter Sollis, Bill
Stanley, and Missy Wolff-Burke. My thanks also to my extended family, from
my parents, David and JoAnn Call, and my siblings, to my parents-in-law, Don
and Carolyn Fitzsimmons, whose support in different ways aided my ability to
complete the book.
Finally, my wonderful wife, Tracy, and fantastic children, Shayla, Jag, and
Dash, have been unfailingly patient and supportive, and it is to them I dedicate
this enterprise.
Introduction
The Tragedy of Civil War Recurrence

WHY DOES PEACE, once it is widely seen as consolidated, fail? Consider the
tragic experience of Liberia. In 1996 a cease-fire between rebel factions and an
embattled government ended a bloody war that had taken 200,000 lives and
displaced most of the country during the previous seven years. With the election
of a new president in 1997 and the cessation of combat, the war was deemed
over. Peacekeepers went home, postconflict reconstruction drew to an end, and
scholars analyzed the causes of the war’s beginning and end. However, by 2000,
a new civil war, involving many of the same fighters and same issues, broke out
in Liberia. The second war displaced hundreds of thousands anew, and another
150,000 were killed. Many of these new casualties were women and children
who had developed cautious optimism with the end of the prior war, only to see
their hopes dashed and their loved ones once again threatened, conscripted, or
brutally killed.
Liberia’s experience—a devastating relapse into armed conflict after hav-
ing achieved apparent peace—is shared by other countries: Lebanon in 1975,
Nicaragua in the early 1980s, Russia in Chechnya in 1999, Burundi in 1993, the
Central African Republic in 2001, Haiti in 2004, and East Timor in 2006. In
the latter two cases, as in Liberia, renewed hostilities followed lengthy expensive
international peacekeeping operations that concluded amid declarations of suc-
cess. Given the expanding investment of troops and international resources to
stabilize places like Afghanistan, Sierra Leone, Bosnia, Iraq, and Kosovo, ensur-
ing that peace “sticks” is more important than ever.
What explains why peace sticks in some cases but not in others? Other schol-
ars have only partially addressed this question. Many have studied peace agree-
ments and have postulated that successful implementation depends alternately
on the character of the agreement, on the degree of commitment among the par-
ties, on the commitment of outside third parties, or on implementation efforts
(Fortna 2004a, 2004b; Hampson 1996; Harbom, Högbladh, and Wallentsteen
2006; Walter 2004; Stedman, Rothchild, and Cousens 2002). Yet these studies
2 Introduction

omit cases of civil wars (often with salient international elements) that have not
ended in peace agreements, such as Kosovo and East Timor in 1999, Afghani-
stan in 2001, Haiti in 2004, and Sri Lanka in 2009. These cases of victory (and,
for opponents, defeat) constitute an important and growing component of civil
wars today.
Other researchers have examined the effectiveness of peacekeeping as a
tool. Yet their research does not address societies where peacekeeping has been
viewed as successful but armed conflicts still recur, as in Liberia, Haiti, and East
Timor. Much statistical work and survival analysis examine recurrences after a
cease-fire, with either no minimal threshold once a cease-fire is reached, or a
minimal threshold of only one or two months of a cease-fire holding to enter the
data set as a terminated war (e.g., Fortna 2004a). Such data sets lump together
very “young” cease-fires, where the population or parties may well believe that
the cease-fire is temporary or tactical, with stable cease-fires, where most citizens
genuinely believe the war to have ended. Here I am not interested in short-lived
cease-fires that unsurprisingly fail but rather cease-fires characterized by wide-
spread expectation that the prior war is over and where citizens act according
to that expectation—that is, where most former combatants demobilize, most
refugees return, most political parties compete for power, and civil society modi-
fies its behavior to reflect a stable cease-fire. Rather than examine cases where the
seed of peace has been planted but its emergence remains uncertain, this book
is concerned with countries where the sapling of peace has sprung forth and
survived its first year.

The Importance of This Book

This book’s inquiry—why does peace fail?—is important for four reasons. First,
civil wars are devastating. During the past fifteen years, civil wars have collec-
tively caused more deaths than any other form of organized mass violence—in-
ternational wars, genocide, or terrorism. In addition a large portion of civil wars
occur in countries that have experienced prior civil war. According to one data
set, 30 percent of all civil wars that started between 1988 and 1999 represented
recurrences of prior civil wars that had ended in the previous ten years.1 The
recent World Development Report (World Bank 2011, 1.10) claims an even higher
percentage of onsets in postwar countries for the 1990s—67 percent—and a
much higher rate, 90 percent, for the first decade of the 2000s. Thus the failure of
peace is a principal factor in the outbreak of the most common and deadly form
of warfare in the world today.
Second, once a process of peace consolidation is on the path to success, its
failure is doubly devastating. Renewed conflict is destructive for the populations
that are once again uprooted after beginning to rebuild, that again fear for their
Introduction 3

loved ones’ survival, and that see fragile economic regeneration evaporate. But it
is deeply demoralizing as well. Investing in a renewed peace in one’s country and
community is more difficult when prior hopes and investment have been shat-
tered. The disillusion inspired by bloodshed after its apparent end extends also
to outside peacemakers, peacekeepers, and peacebuilders. Failure after apparent
success calls into question the entire effort and industry of peace.
Third, these cases represent a new burden for international peacekeeping.
Many observers are familiar with cases where international troops have been re-
deployed after peacekeepers withdrew amid failure (e.g., in Angola in 1991, and
the former Yugoslavia in 1992). However, redeploying peacekeepers after a prior
peacekeeping operation seemed to succeed—and then peace failed—is a new
challenge for the international community. In only three cases has a UN peace-
keeping operation been deployed after another had been previously withdrawn
amid widespread perception of a successful peace process. These cases—Liberia
in 2003, Haiti in 2004, and East Timor 2006—all occurred in the past decade
and only three years before the formation of a new United Nations body de-
signed to tackle these challenges, the UN Peacebuilding Commission.
Fourth, these cases go to the heart of the contemporary peacebuilding chal-
lenge: How, in the wake of bloody war, can external actors help foster a society
that can resolve its conflicts without recourse to mass violence? With the ex-
ception of the United States–led wars in Afghanistan and Iraq, in recent years
the international community can claim some successes in stabilizing postwar
societies. From Bosnia and Kosovo to Sierra Leone and Liberia, from Georgia
and Moldova to Azerbaijan and Haiti, international troops have enjoyed some
success in quashing destabilizing threats to postwar societies and their regimes.
However, the threat of renewed warfare underlies the decision to retain external
troops in all these cases. What is the exit strategy for these international troops?
For many analysts, postwar state building is the answer to the challenge of
consolidating peace. Building self-sustaining legitimate and effective states obvi-
ates the need for foreign troops, reduces the international peacekeeping burden,
and advances numerous international security objectives. As Fukuyama (2004,
ix) says, “State building . . . should be at the top of our agenda.” These cases—
where peace has been secured but subsequently failed—offer a window into the
factors that determine state-building success and failure, and its relation to suc-
cessful peacebuilding.

The Central Argument

The first finding of this book, which precedes the main argument, is that no
single factor or variable accounts for success in consolidating peace and averting
war recurrence. Although this may seem unsurprising, some scholarly studies
4 Introduction

suggest that only one or two variables are responsible for the onset or recurrence
of war. This examination of the pool of cases of reversion to internal armed con-
flict after civil war decisively undermines such monocausal claims.
Nevertheless, this study finds that one factor seems to play a more common
causal role in civil war recurrence than others. This central finding is that po-
litical exclusion, rather than economic or social factors, plays the decisive role
in most cases of civil war recurrence: Political exclusion acts as a trigger for re-
newed armed conflict.2 Conversely, political inclusion, including but not limited
to powersharing arrangements, is highly correlated with consolidation of peace.
The contribution of this book lies not only in its identification of this factor in
almost all recent cases of nonrecurrence of war and in cases of renewed armed
conflict after apparent peace, but also in its critique and subsequent development
of current postwar peacebuilding theory in ways that respond to this finding.
Political exclusion here refers to the perceived or actual deprivation of an ex-
pected opportunity for former warring parties, or the social groups associated
with them, to participate in state administration, through either appointed posts
or elective office. In separatist struggles, those opportunities are concentrated in
authority over a substate territory. Inclusionary postwar policies respond both to
the need for security guarantees among former rebels and to the need for some
degree of voice in governance among formerly mobilized elites. Inclusionary be-
havior corresponds very closely (but not perfectly) to an ability to avoid civil
war recurrence.
Two further characteristics define exclusion. First, political exclusion here
refers to processes rather than substantive policy outcomes.3 It refers to partici-
pation in political or policy processes: the opportunity either to compete in the
electoral arena or to enjoy representation in appointed state offices. Political pro-
cesses signify both the potential and the actual exercise of power, and thus they
provide a means of attaining desired ends. Chief among these ends is security
for the group and its adherents or associated populations. As such, participa-
tion in the security forces (military or police) is an especially important arena
of state inclusionary or exclusionary behavior affecting the likelihood of subse-
quent violent conflict. It is especially perilous for postconflict regimes to violate
expectations about the ability of formerly armed and mobilized social groups to
participate in the policing and defense of their own country or territory.
Second, political exclusion is contextually defined rather than measured by a
single objective standard of participation. Exclusion triggers a return to armed
conflict if it violates the expectations of either a faction of the prior war or a
social group associated with that faction. That is, exclusion refers here not to
a globally applicable objective level of political participation but to subjective
perceptions among former warring factions about whether the state engages in
Introduction 5

exclusion based on prior understandings that emerged from prior war termina-
tion. As noted above, these expectations may concern either an opportunity for
electoral participation or for representation within the armed forces, the police,
legislative posts, or Cabinet positions. And these expectations are often heavily
shaped by peace agreements, which serve as an initial indicator of minimal ex-
pectations. However, expectations are also shaped by promises by one party to
others and by initial practices that are subsequently changed. Actions that may
not be formally proscribed in an agreement can constitute a violation of expected
inclusionary behavior, such as firing or persecuting a number of members of one
or more parties associated with the prior war. Such a definition complicates, but
also enriches, cross-national comparisons.
The book also reaches a number of other conclusions. First, the nature of
peace and the factors that shape peace depend in important ways on the global
normative context. The findings suggest that how wars ended during the Cold
War, and the role of exclusion and inclusionary behavior, were not the same as
after the Cold War. Having been shaped by changes in great power policies and
transnational expectations, the parties to warfare reach negotiated settlements
more often. More pertinent, exclusionary behavior, which correlated with nonre-
currence of armed conflict during the Cold War, has since then ceased to success-
fully suppress the recurrence of armed conflict. The implications of this finding
are explored in the concluding chapters.
The research also finds that inclusionary behavior closely corresponds to suc-
cessful peacebuilding in the post–Cold War era. In separatist struggles, inclu-
sionary behavior involves a satisfactory degree of representation in the governing
apparatus of the relevant territory, through either appointive or elective posi-
tions, including the forces providing security in the separatist territory. In this
sense inclusivity refers to perceived chances to participate in state institutions
more broadly than simply the integration of former enemies into state institu-
tions. This form of “territorial powersharing” is a form of powersharing broadly
defined. I find that the integration of former enemies into state political or se-
curity institutions is the norm among successful cases of peace consolidation
after civil wars ended through negotiated settlements. Similarly, powersharing
broadly defined (either through integration of given state institutions or through
divvying up different state agencies without internal integration of each) is the
norm among successfully consolidated peace agreements. However, neither inte-
gration nor powersharing is necessary for inclusionary peacebuilding. Inclusion-
ary peacebuilding also includes the opportunity to compete electorally, without
any given guaranteed state offices. Thus inclusionary peacebuilding generally
involves powersharing but may also involve “power dividing” (Roeder 2005), so
long as the expectations of the chances for participation in politics, including key
6 Introduction

state offices or posts, are not violated. Ultimately, the importance of inclusion
and exclusion leads this book to emphasize the role of legitimate authority and
to advocate an approach I call legitimacy-focused peacebuilding.

Contributions to Theory
This book’s findings contribute to theory by calling into question four domi-
nant understandings or approaches to civil wars and postwar peacebuilding. First
and foremost, it challenges prevailing theories that economic factors, especially
poverty and a decline in economic growth, are the first place to look for factors
in civil war recurrence. The influential work of Paul Collier and his colleagues
has especially given rise to approaches emphasizing poverty, declining economic
growth, and the role of natural resources and export dependency.4 Instead, this
book shows that political decisions that exclude certain groups—whether they
are in governance, security, or economic policy—are the first line of factors in
understanding why certain countries experience recurrence and others do not.
Drawing on processes within qualitative studies of numerous cases, rather than
risk factors in quantitative analysis, the book may offer the most compelling an-
tidote to the prevalent economy-centered approaches.
Second, the book’s findings point to a central role for national actors in the
success or failure of peace. Much of the recent literature on civil wars and peace-
building explores the role of peacekeeping troops, “liberal” or “republican” mod-
els, or mediators in the duration of peace settlements (e.g., Fortna 2008, 2004a;
Paris 2004; Barnett 2006; Ghani and Lockhart 2008; Walter 2004). The book’s
case studies confirm some of the important findings of this research. Most im-
portant, the presence of third-party military troops in maintaining stability once
a cease-fire is reached seems very important in some cases. Three recurrences
took place within two years of the departure of international troops, and most
analysts believe that several “frozen conflicts” would fall into recurrence if for-
eign peacekeeping troops were to depart. However, this book finds that national
actors’ calculations hold a privileged and underemphasized role in the determi-
nants of civil war recurrence. The decisions by national elites about how to treat
former enemies, both in the state and in society, prove crucial in peacebuilding’s
failure and success. The book suggests that international actors might have more
impact through engagement with strategies for more inclusionary behavior by
elected governments.
Third, the book points the recent state-building literature in a new direc-
tion: a focus on state legitimation rather than state capacity. Several scholars
and policymakers have embraced “state building” as the principal solution to
the problem not only of “failed states” but also of postconflict reconstruction
Introduction 7

and peacebuilding (Fukuyama 2004; Ghani and Lockhart 2008; Chesterman,

Ignatieff, and Thakur 2005; Call 2008a; Brinkerhoff 2007; Barnett 2006). In
addition, international organizations like the World Bank, the United Nations
Development Program, and donor governments have moved to institutionalize
state building as an element of development and postconflict recovery (Stew-
art and Brown 2006; Keating and Knight 2004; Whitman 1999). Although this
work conceives state building as encompassing both legitimacy and capacity,
scholarship and policy work have actually stressed capacity over legitimacy (see
Fukuyama 2004). Doyle and Sambanis (2000, 2006), for instance, emphasize
the role of “local capacity” and “international capacities.”5 Ghani and Lockhart
(2008) discuss legitimacy, but they treat it as one sector among ten whose ca-
pacity requires strengthening. In contrast, the approach here demonstrates the
salience of legitimacy of postwar regimes—especially that component of legiti-
macy that derives from inclusionary practices of salient social groups—rather
than capacity.
Fourth, the book finds that the emphasis on electoral democracy by interna-
tional actors as the framework for postwar governance (Ponzio 2010) is insuf-
ficient, and perhaps harmful. In particular recurrence in Liberia and three other
cases occurred once elected governments, absent the presence of international
peacekeeping troops, felt sufficiently secure in power to exclude certain social
actors or factions from the state, and thus sparked renewed armed conflict. These
findings contradict prevalent international practice more than scholarship,
which has long criticized placing elections at the center of democracy promo-
tion. Recent research has proposed expanded understandings of governance in
postconflict societies beyond “democracy” to encompass different forms of dis-
tributing power and animate inclusion and participation. This book reinforces
these trends.
In sum the research here contrasts with and challenges four dominant trends
in postconflict scholarship: the emphasis on economic factors in the literature
on civil wars, a separate emphasis in the peacebuilding literature on international
determinants of peace, the focus on state capacity in the state-building litera-
ture, and the promotion of electoral democracy as the unquestioned and optimal
framework for governance in postwar societies.

Research Design and Methodology

What is the best way to approach this research endeavor? This section estab-
lishes key definitions and thresholds of concepts and variables. It starts by clari-
fying the threshold for the “failure” of postwar peace, then sets forth the pool of
cases, and then explains my methodology.
8 Introduction

The Object of Study: When Has Peace “Failed”?

A number of standards for “failed peace” exist. Some of these are economic in
nature (poor countries are structurally violent), some are political (dictator-
ships cannot experience “true” peace), and some are institutional (weak states
are never stable).6 As its main object of study, this book adopts a circumscribed
definition of the failure of peace: the recurrence of internal armed conflict where a
prior civil war is widely perceived to have ended. This definition is for a minimalist
“negative peace,” or the absence of armed conflict (Galtung 1964, 1–4). Most
studies of “peacebuilding” adopt a broader concept of peace, something that is
more than the simple absence of war. Some scholars (Doyle and Sambanis 2006;
Call 2008c; Call and Cousens 2008) advocate a standard that includes some
minimal level of nonauthoritarian politics, to which I am sympathetic. Others
adopt a “root causes” discourse that addresses structural violence (Paris 2004;
Richmond and Franks 2009).
Why, then, have I adopted this minimalist standard of success / failure? For
two reasons. First, the purpose of this study is strictly to examine the recurrence
of warfare, and the recurrence of organized mass violence is the sole universally
accepted indicator of the failure of peace. For anyone who has lived in the midst
of war, its recurrence is a worthwhile phenomenon of study. No matter how
broad the concept of peace one embraces, the unique, commonly shared stan-
dard of failure in armed contests for state power is the return of organized armed
conflict.7 Even though statistical research indicates that peace is more likely to
be sustainable without poverty, inequality, or political unrest, peace is defined
here so it can coexist with poverty, economic inequalities, social inequalities, and
even nondemocratic or authoritarian regimes.
Second, the study grew out of a lacuna regarding a core need in the realm
of policy. In May 2006 I was an invited guest at the first informal meeting of a
new UN body, the Peacebuilding Commission. At this meeting, which had been
organized by two think tanks (an analysis of which I drafted; see CIC and IPI
2006), the just-named thirty-one members of the commission deliberated what
its purpose and methods might be. One striking thing about the meeting was the
assumption, articulated by several nations’ ambassadors to the United Nations,
that preventing war recurrence was the central aim of the new body.8 Despite the
vast and growing literature on peacekeeping, postconflict reconstruction, and
peacebuilding, few scholars had analyzed civil war recurrence, and not a single
book addressed the topic. That meeting led to a proposal (which received sup-
port from the US Institute of Peace, from New York University’s Center on In-
ternational Cooperation, and from American University) to research this ques-
tion of primary concern for this new UN body: how to prevent internal armed
conflicts from recurring, especially where UN efforts had seemed to succeed.
Introduction 9

One of the motives of this study was specifically to inform policymaking in

cases where postconflict peacebuilding operations involving significant interna-
tional efforts had failed. For the purposes of this exercise, I heuristically exam-
ined the universe of UN peacebuilding operations that had succeeded, failed,
or resulted in mixed outcomes. Aside from peacekeeping operations that never
reached any level of recognized stability (e.g., Angola and Somalia), there were
only three cases in which perceived peacekeeping “successes” ended in recurrent
armed conflict to such an extent that it was necessary for the UN to redeploy
a new peace operation. In each of these three cases—Liberia, Haiti, and East
Timor—despite an earlier perception that the mandate had been fulfilled and a
minimal level of stability had been achieved, subsequent armed conflict never-
theless erupted to such an extent that a new UN peacekeeping force was neces-
sary.9 These cases exclude UN peacekeeping failures where troops withdrew amid
widespread perception of failure, as from Somalia in 1995 and Angola in 1997.10

Measuring the Failure of Peace: The Pool of Cases

How can one measure the recurrence of organized armed conflict in a post–civil
war society? To preclude bias in selecting the pool of cases, I drew on a well-
regarded data set, specifically that of Fearon and Laitin (hereafter F&L), as the
best one for civil wars for quantitative analysis across a large number of years,
with a relatively large number of control variables available. Following this data
set, I use countries rather than civil wars themselves as units of analysis, and I de-
velop a pool of cases from a list of countries that have experienced recurring civil
wars. As with the data sets of Doyle and Sambanis (2000) and Sambanis (2004),
F&L draw on the industry standard first introduced by David Singer and Mel-
vin Small of using 1,000 minimum battle-related deaths (including civilians) for
inclusion in the pool of cases of “wars” in their Correlates of War Project.11 F&L
code a civil war where a state in a given year engages against an organized non-
state challenger for state power in an internal armed conflict resulting in at least
100 deaths in that year, with a cumulative total of 1,000 battle deaths (including
civilians). The resulting data set includes sixty-eight countries that experienced
civil war onsets between 1946 and 1999.12
However, many cases of civil wars revert to armed conflict sufficient to mark
a failure of peace but whose levels of violence do not reach 1,000 battle deaths.
These armed conflicts (of a lesser status than “civil wars”) are seen both inside
their societies and by outside actors as failures of consolidation of peace. These
cases should not be counted as successes but as failures. Consequently, I lower
the level of “recurrence” from the 1,000 battle deaths conventionally required
for a full “civil war” to include armed conflicts that appear in the data set of the
Peace Research Institute Oslo (PRIO) and the Uppsala Conflict Data Program
10 Introduction

(hereafter the PRIO data set). Their threshold to count as an “armed conflict”
requires only 25 battle-related deaths in a single year.
This choice has three flaws. First, the data for these lower-level armed con-
flicts are not available for all countries before 1988. Where they are available,
they are less reliable because news sources were themselves not as reliable, avail-
able, or accessible as in more recent years. Consequently, I have used internal
armed conflicts to supplement and bring up to date the F&L data set, covering
only the period 1999–2007.
Second, it also includes a number of cases that do not correspond to com-
mon concepts of “civil war” or even a smaller-scale armed conflict. Some of these
cases may be better characterized as coups that involved organized violence (as
in the Central African Republic), sporadic and poorly organized violence with
ill-defined nonstate armed factions (e.g., Haiti in 2004), or conflicts with very
few deaths (e.g., East Timor in 2006).
Nevertheless, these armed conflicts do represent a failed peace process, and
they should be included even though our language and concepts prove inad-
equate. Including these lower-level armed conflicts as recurrences offers the ad-
vantage of including a number of cases widely considered relevant for the failure
of peace. Internal armed conflicts often prove destabilizing for the state or re-
gime and hold tragic consequences for the population. Indeed, a civil war with
thousands of battle deaths is not necessarily more destabilizing than an armed
conflict of a few dozen. Consider East Timor, where riots and armed combat be-
tween different factions of the army and police in 2006 led to the fall of the gov-
ernment, the displacement of tens of thousands of people, and an Australian-led
peacekeeping intervention of 1,300 troops. Yet only thirty-four persons died
in the fighting. Conversely, few would maintain that the Aceh independence
struggle, which resulted in more than five thousand deaths in its 1998–2005
phase, threatened the stability of Indonesia. Or consider Haiti’s recurrence in
2004 that involved few armed confrontations and “only” dozens of casualties yet
led to political upheaval, an interruption of the regime, the ouster of the presi-
dent, and the deployment of foreign troops and a UN operation that remains
as of this writing. In truth this book is about civil wars that recur as “internal
armed conflict recurrences after 1999,” but this seems rather cumbersome as a
moniker. Thus I still use “civil war recurrence,” even though this label should be
understood in a broader sense.
Third and last, the lowering of the threshold of deaths to twenty-five for re-
currences also may draw in cases that are never referred to by international or
national analysts as even an armed conflict. After reviewing the secondary litera-
ture, I have therefore excluded a single case that would fall into this category—
Azerbaijan’s 2004 instances of violent attacks involving a few dozen deaths—
Introduction 11

from inclusion in the pool of cases of “recurrence.” I could find no record of any
secondary literature referring to this as a “new” or second armed conflict, but
only as isolated instances of violence.
Having offered the above explanation of what count as “civil war” and “recur-
rence” in the definition of “failed peace,” I offer the following list, which gives a
more detailed characterization of the definition used here of the failure of post-
war peace: the recurrence of internal armed conflict where a prior civil war is widely
perceived to have ended:

• “Armed conflicts” do not necessarily rise to the level of a “civil war,”

and thus generally involve a lower minimum threshold of deaths.
Thus recurrences include both recurrent civil wars and renewed vio-
lence that does not rise to the level of a civil war but qualifies as the
lesser “internal armed conflict.”
• “Civil wars” and “armed conflicts” involve contestation for political
power over a given territory, usually either the state or a region seek-
ing some characteristics of a state (autonomy). This definition thus
excludes social violence that involves no attempt to capture some
state power.
• An “internal” armed conflict does not preclude the involvement of
external armed actors. The definition holds so long as one nonstate
armed group credibly claims to represent an internal population.
• The requirement that the prior conflict is “widely perceived” to have
ended excludes short-term cease-fires, and it limits the cases to those
that have seen a longer period (defined here as at least a calendar
year) without armed combat.
• Although “new” armed conflicts in postwar societies can also be
devastating, a “recurrent” conflict conceptually requires some relation
to a prior armed conflict. Hence, the definition here requires that at
least one nonstate actor (either the organization or its senior leaders)
be active in both conflicts.
• The definition does not prejudge the juridical status of the territory
where the conflicts occurred. Thus recurrent conflicts are included
even if the state’s name or borders shift in the interim.

In practice, it is extremely difficult to measure these characteristics of a war or

a society sufficiently to create a pool of cases. The distinction between an “armed
conflict” and a “civil war” also runs into problems. Despite rapidly expanding
research on society-based internal armed conflicts that do not involve govern-
ment forces, the term here is restricted to those in which government forces are a
12 Introduction

faction and the acquisition of some state power is at stake.13 All major data sets of
internal armed conflict use a total of 1,000 battle-related deaths as the threshold
for when an “internal armed conflict” qualifies as a “civil war.” This distinction is
arbitrary but useful.14
Similarly, knowing when a war is “widely perceived” to have ended is diffi-
cult to judge and impossible to standardize across numerous cases over time.
Certainly such perceptions can be documented through an examination of
how observers describe wars, describe periods after a cease-fire, and describe
the events that may constitute recurrence. For instance, Liberia and Chechnya
are unusual cases insofar as scholars and average citizens speak of a “first” and a
“second” civil war—a clear indication of a widespread perception of war termi-
nation. However, such language is not employed in some other cases that consti-
tute recurrence, such as the counterrevolution in “postrevolutionary” Nicaragua
from 1981 to 1990. In cases like Cambodia (1975–78), quantitative indicators
misleadingly suggest that situations of one-sided violence, or even genocide, be
scored as periods of interwar “peace.” As a proxy for countries “at peace,” I start
with countries in which Fearon and Laitin (2003) show that a civil war ended
for at least a calendar year.
In determining a perception of peace, I supplement the F&L list with an anal-
ysis of discourse in public news accounts and policy analysis that recognizes an
end to the first war. Because survey data are rarely available, I rely on analysts’
claims and the secondary literature about the extent to which the population
understood the war to be over. I also draw secondarily on these analysts’ own
perceptions about whether the first civil war ever was concluded. I count cases
as having concluded a prior war if at any time in the intervening period the wide-
spread perception emerges that the first civil war had ended. Hence it does not
matter if the population distrusted a cease-fire or peace accord in the immediate
aftermath of its adoption or if the parties had little faith in a peace accord at the
moment of signing it. What matters is whether, at some point subsequent to a
cease-fire or peace agreement (but before any recurrence), the majority of citi-
zens perceived the prior war to have ended and acted accordingly. One indicator
of this perception, for instance, is the return of many refugees. Another is the
demobilization or incorporation into the armed forces of a significant majority
of former combatants. This standard is still compatible with some minority of
people believing that the “peace” was illusory and that postwar stability repre-
sented only a lull in the fighting and prelude to renewed combat. Ultimately,
where secondary works clearly belie the F&L categorization of a period of inter-
war peace, such cases are dropped. Omitted cases include Cambodia (1970–75,
1978–92), Senegal (1989–2001, 2003), and Somalia (1981–2002, 2006–).
Nevertheless, expectations and perceptions about warfare are fluid, contextual,
and hard to determine.15 They are also shaped by interaction among the parties
Introduction 13

(Hampson 1996) and by third parties (Walter 2004). In addition, the parties to
conflict are often divided and their interests change. This study tries to be faithful
to the available evidence about expectations and their evolution, but this remains
an imperfect process.
Finally, a change in the status of a state poses challenges for categorizing such
wars as internal versus external. Consider the difficulty of coding the wars of the
former Yugoslavia after the 1991 declarations of independence by Croatia and
Bosnia, or the 1999 NATO bombing of Kosovo in which the Kosovo Liberation
Army played an enormous role as the ground troops. Analyzing war recurrence
where the status of the territory changes across decades is especially difficult.16
Many wars for independence (excluded from most data sets) are shortly fol-
lowed by related internal armed conflicts.17 None of the options is appealing,
whether it means (1) counting the recurrence as a single new war within the new
state (ignoring the prior war), (2) counting the two wars as new wars in each of
two states (ignoring their relation to one another and dropping the recurrence),
or (3) counting them as two wars within the new state (which did not exist dur-
ing the first war). Although I adopt the third option here (namely, regarding East
Timor, whose war for independence within Indonesia was followed by an inter-
nal armed conflict in 2006), no choice here is optimal.
What, then, is the pool of cases of recurrence of armed conflict? Given the dif-
ficulties described above, I develop a list using these data sets that approximates
the definition above. I use F&L’s sixty-eight cases as the initial pool, and add any
cases of recurrence of these civil wars as armed conflicts in the PRIO data set.
Thus the pool of cases of “recurrence” consists of

• any country among F&L’s sixty-eight countries that experienced a

civil war after a prior civil war followed by at least one calendar year
of peace (thus Nicaragua in 1978–79 and 1981–88 counts as a recur-
rence, whereas Afghanistan’s entries of 1978–92 and 1992– do not);
• added to this list, any subsequent civil war or internal armed con-
flict in the PRIO data set in a country that appears on F&L’s list of
civil wars;
• recurrence requires at least one nonstate actor common to both
armed conflicts (possibly subsequently as the state, or vice versa), so
I thus exclude cases of “new wars” in the same country (or the same
territory under a different name) in which none of the prior nonstate
actors are involved, and I focus instead on only “repeat” or “recurrent”
conflicts.

These categories permit one to sort all countries into three categories: those
that did not experience any new wars after a given civil war (nonrecurrence),
 14 Introduction

Table I.1 All Civil Wars and Recurrences, 1946–99, plus Recurrences as Internal Armed Conflicts,
1999–2007
New Civil War, Same Country Civil War (and Internal Armed
Nonrecurrent Civil Wars or Territory Conflict) Recurrences
Angola, 1975–2002 Algeria, 1962–63, 1992–present Azerbaijan, 1992–94, 2005a
Bangladesh, 1976–97 Argentina, 1955, 1973–77 Burundi, 1972, 1988, 1993–2006
Bolivia, 1952 Chad, 1965–2002, 2005–present Cambodia, 1970–75, 1978–92a
Bosnia, 1992–95 China, 1946–50, 1991–present Central African Republic,
Costa Rica, 1948 Colombia, 1948–62, 1963– 1996–97, 2001–2
Croatia, 1992–95 present China, 1950–51, 1956–59
Cuba, 1958–59 Democratic Republic of Congo, East Timor (Indonesia), 1975–99,
Cyprus, 1974 1960–65, 1977–78, 1996– 2006
Djibouti, 1993–94 2001, 2006–present Georgia (South Ossetia),
Dominican Republic, 1965 Ethiopia, 1974–92, 1997–present 1992–94, 2004
El Salvador, 1979–92 Georgia (Abkhazia), 1992–94 Haiti, 1991–95, 2004
Greece, 1945–49 Indonesia, 1950, 1953, 1958–60, Lebanon, 1958, 1975–90
Guatemala, 1968–96 1965–present Liberia, 1989–96, 1999–2003
Jordan, 1970 Iran, 1978–79, 1979–93, 1984 Mali, 1989–94, 2007
South Korea, 1949–50 Iraq, 1959–59, 1961–74, 2004– Nicaragua, 1978–79, 1981–88
Laos, 1960–73 present Peru, 1981–95, 2007
Moldova, 1992 Kenya, 1963–67, 1991–93 Russia (Chechnya), 1994–96,
Morocco, 1975–88 Nigeria, 1967–70, 1980–85 1999–present
Mozambique, 1976–95 Pakistan, 1971, 1973–77, Rwanda, 1962–65, 1990–2002
Nepal, 1997–2006 1993–99, 2004–present Senegal, 1989–2001, 2003a
Papua New Guinea, 1988–98 Philippines, 1946–52, 1968– Sri Lanka, 1971, 1983–2002,
Paraguay, 1947 present 2003–presenta
Sierra Leone, 1991–2000 Turkey, 1977–80, 1984–present Somalia, 1981–2002, 2006–
South Africa, 1983–94 Yemen, 1962–69, 1986–87, 1994 presenta
Tajikistan, 1992–97 Sudan, 1963–72, 1983–present
United Kingdom, 1969–99 Zimbabwe, 1972–79, 1983–87
South Vietnam, 1960–75
a
Insufficient interwar peace.

those that experienced a “new” civil war or multiple civil wars simultaneously
(or, after 1999, internal armed conflict), and those that experienced a related
“recurrent” or “repeat” civil war (or, after 1999, internal armed conflict). Table I.1
presents all cases of civil wars that originated between 1946 and 1999, plus any
instances of renewed civil war during that period, plus any renewed internal
armed conflict from 1999 to 2007.18 As one can see, the total pool of what I shall
for simplicity call “war recurrence” (recognizing that this refers to civil war re-
currences from 1946 to 2007, plus internal armed conflicts from 1999 to 2007)
contains twenty countries.19
The cases of interest to this book are the twenty-seven countries of nonre-
currence in the first column and the twenty countries in the third column that
experienced “recurrent” or “repeat” civil wars (including, as described, countries
Introduction 15

Table I.2 Fifteen Core Cases of Recurrence

Burundi, 1872, 1988, 1993–2006
Central African Republic, 1996–97, 2001–2, 2006
China, 1950–51, 1956–59
East Timor (Indonesia), 1975–99, 2006
Georgia, 1992–94, 2004
Haiti, 1991–95, 2004
Lebanon, 1958, 1975–90
Liberia, 1989–96, 1999–2003
Mali, 1989–94, 2007
Nicaragua, 1978–79, 1981–88
Peru, 1981–1995, 2007
Russia, 1994–96, 1999–present
Rwanda, 1962–65, 1990–2002
Sudan, 1963–72, 1983–2005
Zimbabwe, 1972–79, 1983–87

that experienced civil war, then peace, then a recurrent internal armed conflict
after 1999). As indicated above, some “recurrences” are more clearly cases of an
internal war, followed by a widely shared perception of peace, followed by re-
emergence to the level of armed conflict. In seeking to advance qualitative stud-
ies, I sought to exclude cases where the perception of peace was questionable.
Thus, cases such as Cambodia’s interwar period between 1975 and 1978 (the
most intense period of genocide) were not widely perceived periods of “peace.”
Similarly, although the qualitative nature of the armed conflict in Somalia after
2006 differed from that of the prior few “interwar” years, the absence of a coher-
ent state calls into question the comparability of this case. In addition, no wide-
spread perception of “peace” prevailed.
If we remove these cases, we are left with what I call the “core cases” of oc-
currence of destabilizing internal armed conflict after a post–civil war period of
peace. (Note that no time limit for recurrence is used here.) Table I.2 contains
these fifteen clearer core cases. These core cases serve as the basis for qualitative
exploration (presented principally in chapters 3 through 7) of the book’s main
question.

Methodology
The main research objective is to discern causal patterns that explain the recur-
rence of organized political violence after a civil war has concluded. I use three
methods. First, drawing on qualitative and quantitative studies of all types of civil
wars, I briefly present a series of linear regression analyses conducted in collabo-
ration with the economist John Schmitt (Call and Schmitt 2009). The purpose
of this quantitative component of the study is to determine whether and how
16 Introduction

recurrence differs from the more widely studied onset. I also explore the limits
of quantitative methods and demonstrate the utility of qualitative methods for
understanding the causes of civil war recurrence and its absence, nonrecurrence.
Second, I present an extensive analysis of Liberia’s civil war recurrence as a
heuristic case study of an exemplary recurrence of civil war. This case was se-
lected initially as one of only three instances of reintroduction of UN peacekeep-
ing troops after the completion, amid perceived success, of a prior UN mission.
In addition, the Liberia case represents perhaps the clearest instance of a civil
war that was considered concluded amid great fanfare with heavy international
involvement, but then resumed with some of the same leading actors after a min-
imum period of stability.
The case study is theory oriented, and thus is organized around and explicitly
engages competing arguments about the causes of civil war onset and recurrence.
To the extent that the case advances the hypothesized argument over economic
or resource-based arguments, it constitutes what Eckstein (1975, 127) calls a
“least-likely” case—one that holds particular significance to call a theory into
question as a relationship is found where it is least likely (see also King, Keohane,
and Verba 1994, 209–12). This is because Liberia is generally considered one of
the principal examples of resource-based arguments. Although I conducted brief
field research in Liberia in 2004, the case study relies mainly on the extensive
primary and secondary materials available. Because the Liberia case was instru-
mental in generating the argument, it is solely illustrative and thus does not test
the argument about the role of exclusion in the context of understandings at the
close of a prior war. I also explore the role of hypothesized factors in the recur-
rence of internal armed conflict, and I examine interaction among these factors.
Third, I use structured, focused theoretically driven “mini–case studies” of
the remaining fourteen cases of recurrence of armed conflict. Here I employ
what Eckstein (1975, 108) calls a “plausibility probe” to explore the extent to
which “candidate-theories” seem to be causally operative in other cases of recur-
rence of internal armed conflicts. Alexander George and Andrew Bennett (2005,
67) set forth the case method of “structured, focused comparison,” describing a
structured case study as one where a common set of questions “are asked of each
case,” and a focused comparison as one that “deals with only certain aspects of
the historical cases examined.” I conduct structured, focused mini–case studies
of the other fourteen “core” cases of recurrence. In them I probe the extent to
which inclusion or exclusion underlies civil war recurrence, and thus examine
the role of factors such as poverty, economic downturn, ethnic and religious frac-
tionalization, dependence on natural resources exports, and geography.
These mini–case studies run the risk of failing to provide sufficient detail for a
rich and contextualized understanding of the case and processes, and they surely
do not capture all the causal dynamics or even all the variables that a lengthier
Introduction 17

treatment might identify. Country experts may be frustrated. Nevertheless, I be-

lieve that a brief, focused analysis of the most pertinent hypothesized variables
provides a persuasive and useful probe of the plausibility of the hypothesized
causes of recurrence of armed conflict. Furthermore, each mini–case study ex-
plicitly seeks out variables or factors that might be relevant in examining other
cases. In this sense the mini–case studies also perform a heuristic function of
generating new (or reviving neglected) causal relationships. Because of their
length, these fourteen case studies are broken into three separate discussions
according to specific criteria: cases of recurrences of separatist armed conflicts,
cases of nonseparatist armed conflicts that support the role of political exclusion,
and four cases that seem not to conform to the main argument of the book.
I then turn to the twenty-seven cases of civil wars that did not experience
civil war between 1944 and 2007. Here I use midlevel generalizations rooted in
mini–case studies that are not presented. However, correlations between non-
recurrence and hypothesized variables are shown for a handful of key variables.
This analysis points strongly to a positive correlation between political inclu-
sion of former enemies in political, security, or territorial mechanisms. Thus in-
clusionary behavior comes close to representing a sufficient, but not necessary,
condition for nonrecurrence.
Although each of these three methods is rather conventional, the method-
ological novelty of the book lies in showing the divergent results of each method
and explicitly displaying its strengths and weaknesses. This book seeks to con-
tribute to the study of civil wars and peacebuilding by clarifying, in a single book
with a single object of inquiry, the strengths and weaknesses these three different
methods offer.

Organization of the Book

The book is divided into three parts. Part I includes two chapters, one contain-
ing theory and the other a quantitative exploration of civil war recurrence that
establishes the need for the sort of qualitative research that follows. Chapter 1
examines what we know about the failure of peace, beginning with the extensive
literature, mainly quantitative, on civil war onset and recurrence. The chapter
then draws on the full range of literature on peacebuilding—distinguished by its
a priori focus on external and national efforts to consolidate peace—to review
the four main theoretical approaches to peacebuilding—liberal, republican, gov-
ernance / state building, and critical—that frame the book’s theoretical issues.
The final section of this chapter defines and expounds upon the key concepts of
exclusion, inclusion, and state legitimacy.
Chapter 2 draws on the earlier literature on civil war onset and recurrence to
apply methods and identify what can be learned from the broader universe of
18 Introduction

cases of civil war onset and recurrence. After specifying variables and their ratio-
nale, I conduct a handful of linear regression analyses that specify how civil war
onset differs from recurrence based on a well-known data set. The chapter then re-
flects on some of the systematic weaknesses of quantitative methods in the study
of civil wars that also apply to my own findings. This analysis cautions scholars
and policymakers against excessive reliance on quantitative research, including
the celebrated causal role of poverty in civil war recurrence. It points to the need
for extensive qualitative studies of civil war and its recurrence.
Part II contains the qualitative empirics of the book. Chapter 3 contains the
most extensive qualitative case study in the book, an examination of the causes
of civil war recurrence in Liberia. This and the other mini–case studies in chap-
ters 4 through 6 explicitly address competing or prominent arguments.
Chapters 4, 5, and 6 then extend the analysis to the fourteen other core cases
of internal armed conflicts, using the structured, focused mini–case studies.
Chapter 4 examines separatist recurrences in the North / South recurrence in
Sudan in 1983 after a decade of peace; the recurrence of Tibet’s independence
struggle in 1956 after five years of stability following the 1950–51 uprising under
Mao Zedong’s new Communist regime; South Ossetia’s recurrent armed conflict
in 2004 after ten years of stability (my analysis stops before the 2008 Russian-
backed conflict that led to de facto secession from Georgia of this province and
Abkhazia); and Russia’s recurrent civil war in Chechnya in 1999 after three years
of a negotiated settlement providing for provisional autonomy.
Chapter 5 turns to nonsecessionist recurrences that support the main argu-
ment. The chapter begins by examining two cases of weak states that experi-
enced recurrences: Haiti in 2004, after ten years of United Nations–supported
peace following the United States–led restoration of President Aristide in 1994
after three years of violence; and the Central African Republic’s recurrence of
armed conflict in 2001 after four years of stability. The chapter then takes up two
cases whose recurrences of armed conflict represent a distinctive pattern (noted
also by Atlas and Licklider 1999): victorious liberation movements that engage
in exclusionary behavior against former allies once independence is achieved.
These cases are East Timor’s destabilizing violence in 2006, and Zimbabwe’s
relapse into violence in 1983 four years after independence had been achieved
from Great Britain. Finally, the chapter analyzes Burundi and Rwanda in con-
junction with one another, arguing that exclusion played a particular, chronic
role in the recurrences in these two countries’ interrelated recurrent episodes of
violence and civil war.
Chapter 6 examines the cases of Lebanon, Mali, Peru, and Nicaragua, as well
as, where the argument about exclusion is insufficient, pointing to other causal
factors. In particular, regional dynamics, both neighboring states as well as the
cross-border flows of goods and services and matériel, prove decisive in these
Introduction 19

cases. The interests and strategies of great powers also drive war recurrence in a
certain small portion of the cases in the pool. Weak state institutions and poverty
certainly correlate with most cases, but these factors require further specification
and theoretical claims.
In chapter 7 I turn to cases of nonrecurrence of civil war—twenty-seven civil
wars that had not yet recurred as of 2007. Cases of nonrecurrence include Cuba
in 1959, Jordan in 1970, Vietnam in 1975, El Salvador in 1992, Mozambique in
1992, South Africa in 1994, Guatemala in 1996, Papua New Guinea in 1998,
and the United Kingdom / Northern Ireland in 1998. The chapter begins with
observations about the historicity of civil war termination since World War II,
and then examines different types of inclusionary behavior, especially different
types of powersharing. I classify these cases by whether or not postwar regimes
acted in ways that were inclusionary or exclusionary against the expectations
of the main opposition groups. The chapter ends by examining exclusion and
inclusion for the combined pool of the fifteen core cases of recurrence and the
twenty-seven cases of nonrecurrence. The analysis shows that, in contrast to the
Cold War period, exclusionary conduct in the contemporary era is likely to yield
armed recurrence and resistance. Conversely, different forms of inclusionary
conduct are a factor in virtually all cases of successful peace consolidation.
Part III represents a twofold conclusion to the book. It consists of a chapter
of conclusions related to theory, and another of conclusions related to policy
and practice. Chapter 8 recaps the central findings of the empirical analysis, and
it then reexamines theory in light of them. I first summarize how the empirics of
the book call attention to specific shortcomings of the four dominant approaches
to peacebuilding. I then propose a “legitimacy-focused peacebuilding” approach
that draws on elements of the main approaches but differs from each, offering
especially an alternative to the dominant “liberal peacebuilding” and political
economic approaches. The chapter also considers some of the limitations of
the research and of the concepts used in this field of inquiry, along with their
broader implications for the study of civil wars and peacebuilding. Even as I was
pleasantly surprised to identify certain causal patterns, the project simultane-
ously deepened questions I had about some of the concepts, terms, methods,
and approaches found in the study of war and peace. Many of the assumptions of
students, policymakers, and practitioners are based on rather rudimentary con-
cepts and causal claims that do not capture the complexity and variation of these
armed conflicts, the multiple paths to termination and recurrence, and the fluid
cross-national and subnational processes and dynamics that shape what we call
“civil wars” and “stability.”
Chapter 9 turns to policy and practice. Social scientists tend to conclude
their works by briefly explicating some policy implications of their research,
then leaving those implications unanalyzed and untested. Unfortunately, even
20 Introduction

where research findings are sound, choices to take action to transform social re-
ality in the theoretically desired direction can fail or even prove harmful. This
chapter explores the policy implications and challenges posed by the findings of
the previous chapters. I seek to anticipate negative consequences and to make
considered judgments about how policymakers might incorporate these findings
in ways that minimize harm.

Notes
1. In this interval, Fearon and Laitin (2003) record thirty-seven civil war onsets
worldwide. Of these, 11 (29.73 percent) were in states that had civil wars in the previ-
ous ten years, and twenty (54.05 percent) had experienced a prior civil war since the
start of the data set in 1945. NB: This does not account for civil wars fought with
different actors (i.e., not recurrence in the complete sense), nor does it include wars
involving nonstate actors within a territory before it became an independent state.
2. Recent quantitative work similarly finds that political exclusion is crucial in civil
war onset and secession movements. See Wimmer, Cederman, and Min (2009).
3. The relevant aspects of political exclusion evolve over time and geographical
space, shaped by international interests and norms. Although policy outcomes seem
to have been more relevant in earlier periods, the ability to participate in state institu-
tions of decision making and security are the most frequent triggers of recurrence in
the contemporary period.
4. See Collier (2002, 2007, 2008, 2009); Collier and Hoeffler (2002, 2004, 2006);
Collier, Hoeffler, and Söderbom (2006a, 2006b); and Collier et al. (2003). Also see
Ballentine and Sherman (2003).
5. For international capacities, Doyle and Sambanis (2006) refer to the presence of
international troops as well as the breadth of their mandate, inter alia. Sambanis’s work
(2007) identifies other risk factors for civil war onset and recurrence.
6. For a lengthier discussion of these concepts, and their advantages and disadvan-
tages, see Call (2008c).
7. This definition excludes organized criminal violence as a standard of failure, as
has occurred in postwar El Salvador, Guatemala, and South Africa, inter alia. For a
discussion of the breadth of the problem of postwar criminal violence, and case stud-
ies, see Call (2007).
8. For more on the UN Peacebuilding Commission, see Martinetti (2007); Murithi
(2008); Scott (2008); Street, Mollet, and Smith (2008); and CIC and IPI (2008).
9. Note that I am not equating peacekeeping and peacebuilding here. Instead, the
redeployment of a peacekeeping force is used as an indicator of peacebuilding fail-
ure. To see the breakdown of all UN peacekeeping operations, and whether they were
coded as successfully ending the prior war, see Call (2008c, table 1, 179).
10. For Eckstein (1975, 108), plausibility probes fall short of testing theories with
“crucial cases” but are more disciplined than heuristic case studies.
Introduction 21

11. David J. Singer and Melvin Small’s 1963 Correlates of War Project (most recent
version 1997, also adapted and updated by several others, including Paul Collier and
his colleagues in 2001) set the “industry standard” for quantitative cross-national anal-
ysis of warfare by social scientists. See the official Correlates of War Project homepage,
www.correlatesofwar.org.
12. This formulation draws on their replication tables (which exclude their list of
colonial wars) and condenses overlapping wars because for this study the country-year,
rather than the war itself, is the unit of analysis. See Fearon and Laitin (2003, “Ad-
ditional Tables for ‘Ethnicity, Insurgency and Civil War,’ February 6, 2003,” appendix
with list of wars, p. 7).
13. See the work of the Human Security Report Project of Simon Fraser University,
especially its MiniAtlas on Human Security (2010), and Uppsala Conflict Data Program.
14. Classification of particular conflicts by states themselves is often highly politi-
cal, as seen in the 2005–6 debate in the United States over whether Iraq was experi-
encing a civil war. States also resist the classification of a conflict as an internal armed
conflict (much less a civil war) in order to elude the application of the laws of war.
15. I am grateful to an anonymous reviewer for emphasizing this point.
16. Fearon and Laitin (2003, 76), for instance, include colonial wars for indepen-
dence in their data set of civil wars on the understanding that these should not a priori
be excluded, counting Algeria’s war for independence as internal to France, etc. Yet
when conflict recurs in Algeria, it appears as a new war altogether. Such problems are
inherent to cross-national data collection and comparison.
17. Examples include recurrent armed conflicts in newly independent Algeria,
Rwanda, Vietnam, and Burundi.
18. The table includes the initial civil war, which appears in the “recurrence” col-
umn if it corresponds to any repeat war, and in the “new war” column if subsequent
wars involve none of the same nonstate actors. Walter (2004) refers to these types of
civil wars as “repeat wars.”
19. Five of these countries—Burundi, Iraq, Lebanon, Rwanda, and Sri Lanka—
experienced more than one repeat civil war. In addition, some countries (viz., China,
Iraq, and Yemen) also experienced more than one “new” civil war as well as a “repeat”
or “recurrent” civil war.
This page intentionally left blank
Part I
Why Peace Fails: Theory
This page intentionally left blank
1
What Do We Know about Why
Peace Fails?

WHAT DO WE KNOW ABOUT why peace fails? Most relevant for answering
this question are two key literatures that overlap: that on civil wars, and that on
peacebuilding. The former addresses the causes and character of civil wars: why
they occur, why they end, what makes them persist, what consequences they
have, and so on. Part of this analysis involves the roles of external actors, but the
units of analysis are generally countries (that experience civil wars or do not) and
civil wars themselves.
The latter literature—on peacebuilding—proposes and tests claims about
the role of external, and sometimes internal, actions or interventions on the
process of peace consolidation, stabilization, or war recurrence. Although the
term “peacebuilding” has expanded to encompass preconflict prevention and
wartime peacemaking efforts at mediation or negotiation, the bulk of publishing
on peacebuilding refers to postconflict societies, and thus de facto exclusively
addresses those societies that have experienced conflict and a cease-fire, either
through victory or a negotiated settlement. Therefore, the bulk of peacebuilding
research directly addresses the recurrence of armed conflict; however, the unit
of analysis is not so much wars and countries at war as peace processes; interna-
tional organizations and their structures, policies, or programs; or the combi-
nation of programs and actions undertaken by external and / or internal actors.
Although some quantitative works can be considered to lie in this field (as well
as in the “civil wars” literature, in some cases), the majority of the research in this
peacebuilding literature is qualitative, either case studies, comparative work, pol-
icy analysis, or midlevel generalizations across several cases of peace processes or
institutional structures or programs.
In this chapter I first examine the literature on the onset and recurrence of
civil wars. Some of this research is qualitative, but most works that are both in-
fluential and recent have tended to rely upon quantitative methods. The chapter
then presents the dominant approaches found in the peacebuilding literature.
I focus on four broad categories of understandings: the “liberal,” “republican,”
26 Why Peace Fails: Theory

“state-building,” and “critical” approaches to peacebuilding. The concluding

chapters (8 and 9) revisit these theoretical approaches in light of the book’s em-
pirics. The third and final section of this chapter seeks to define and clarify the
key concepts used throughout the book: exclusion, inclusion, and legitimacy.

What We Know about Civil Wars and Ethnic Conflict

Civil war theorizing reflects the multiple disciplines of history, sociology, po-
litical science, social psychology, and economics. Influential understandings of
civil war from the 1950s to the 1970s focused on national or ethnic groups and
their desire for autonomy. Much of the early literature examined whether, and to
what extent, ethnicity played a role in civil wars (e.g., Deutsch 1953; Shils 1957).
Different perspectives prevail.1 Early analyses of the postcolonial struggles rec-
ognized the role of international power configurations, including the interaction
of colonial authority patterns and tribal elites (Young 1965; Geertz 1963). Other
approaches emphasize the role of social group (especially ethnic) composition,
group capabilities, and group opportunities for insurrectionary challenges to the
state. Gurr (1970) emphasized rational response to frustration, ethnic group ca-
pacities, and later the social opportunities for overcoming the collective action
problem (Gurr 2000; Hegre et al. 2001).
Following the Cold War, some researchers adapted the “security dilemma” to
internal armed conflict, arguing that ethnic group misperceptions and fears re-
garding other groups’ actions trigger offensive armed actions (Walter and Snyder
1999; Posen 1993). In such approaches, externally provided guarantors of cred-
ible commitments (Fearon 1998) or third-party guarantees (Walter 1997, 2002)
played a role in overcoming the security dilemma and ensuring peace. Others
(Gagnon 1994) argued that ethnicity is a tool of entrepreneurial political leaders,
or more broadly a basis for political or economic organization. Under this view,
ethnic composition is not so determinative but interacts either with individual
leadership or with forms of participation and representation that shape whether
mass violence occurs. These approaches all share an emphasis on political or
group grievance in civil wars’ onset.

Greed over Grievance?

In the 1900s scholars began to challenge the emphasis on grievance and identity.
Most prominently, the Development Research Group of the World Bank, un-
der the direction of Paul Collier, produced influential research that “spurred the
creation of new paradigms for the study of civil war and its associated resource
curse” (Ron 2005). This work was initially presented under the catchy moniker
“greed versus grievance,” and its earliest formulations emphasized the predatory
What Do We Know about Why Peace Fails? 27

acquisition of primary commodity exports, using the examples of diamonds in

Sierra Leone and Angola, timber in Cambodia, and drugs in Colombia (Collier
and Hoeffler 2004). Warfare grew out of economic competition among rebel
commanders over primary commodities, was fueled by external investors, and
enlisted the participation of mass-level rational economic actors who enlisted—
all of which led to a spiral of deepening armed conflict.2 Subsequent formula-
tions diminished the role of elites and of primary commodities. Instead Collier
and his associates argued that valuable primary commodities offered economic
opportunities to which rational, utility-maximizing male individuals responded.
They explicitly compared the business of war to common organized crime (Col-
lier and Hoeffler 2004, 3; Kaldor 1999).
Other factors, such as poverty and a recent decline in growth, also assumed
greater importance (Collier et al. 2003). Fearon and Laitin (2003, 75) drew the
same conclusion, stating that “the factors that explain which countries are at risk
for civil war are not their ethnic or religious characteristics but rather the con-
ditions that favor insurgency. These include poverty, political instability, rough
terrain and large populations.”
Collier and his colleagues also found significance in these variables, and in
their subsequent work they saw these variables as proxies for the opportunity
for individuals to take up arms to rationally pursue economic gain. Fearon and
Laitin (2003), conversely, saw them as indicators of weak and underinstitution-
alized states that fostered armed resistance.
A second generation of scholars of political-economy explanations of civil war
and peace has emerged, adding nuance to the original research. Such research
draws on earlier research on rentier states and the resource curse of abundant
natural resources.3 Researchers have also sought to combine these explanations
with more purely political factors, such as forms and patterns of governance, state
institutional capacity, and power-driven irrational leaders (Humphreys 2005;
Fearon 2005; Dunning 2005).

Recurrence
I now turn to the quantitative research on civil war recurrence. Although re-
search on civil war’s recurrence is less developed than that on its onset, scholars
have offered analogous explanations. The dominant views of the causes of civil
war’s onset emphasize the political economy of war, including risk factors re-
lated to poverty. Walter (2004), for instance, uses data from the Correlates of
War Project and finds that indicators of “quality of life”—infant mortality, life
expectancy, and adult literacy rates—are the best predictors of war recurrence.4
She argues that low quality of life fosters war recurrence, although it is difficult
to see why such factors operate differently from war onset. She finds that factors
28 Why Peace Fails: Theory

alleged to explain war onset—including ethnic difference, the combatants’ goals,

and the extent to which those grievances were settled—make no difference in
war recurrence. She also distinguishes “repeat wars” from “new [recurrent] wars”
on the basis of whether the actors and goals of the recurrence match those of
the prior war (Walter 2004, 382–84). Her conclusions rest on a small number
of observations—only fourteen repeat wars between 1946 and 1996, and only
eight “new wars.”
Some scholars emphasize the role of coercion, especially in the likelihood of
averting recurrence via negotiated settlements. Walter and Snyder (1999) em-
phasize the role of third-party guarantors of cease-fires, especially peacekeeping
deployments. Fortna (2004a) and Hegre and Sambanis (2004) point to the size
of a state’s armed forces as crucial, though with opposing views. Collier and oth-
ers (2003) also argued that the existence of a demobilization, disarmament, and
reintegration program would help prevent recurrence, although Humphreys and
Weinstein (2007) found such programs of limited utility.
Apart from these views emphasizing coercion or military force, others stress
various political tools or conditions. Some scholars focus on grievances or politi-
cal opportunities (Gurr 2000; Hegre et al. 2001), whereas others focus on the
presence, degree, or character of democratic regimes (Esty et al. 1995; Reynal-
Querol 2002; Mansfield and Snyder 2005; Hegre and Sambanis 2004). Al-
though many researchers have argued in favor of powersharing arrangements,
Roeder and Rothchild (2005) argue that “power-dividing” arrangements, with
checks and balances over branches of government, result in longer-lasting peace.
Hoddie and Hartzell (2005), by contrast, find that certain forms of powersharing
yield sustained peace. Table 1.1, though not exhaustive, depicts these various
broad approaches to explaining war recurrence, along with the corresponding
authors (who are illustrative, but should not be exclusively associated with the
variable mentioned).
Other variables pertain to the dynamics of the prior armed conflict. On
whether the duration of the prior war makes a difference, some authors found
that longer wars make for longer peace (Fortna 2004b; Walter 2004), but others
found no statistical relationship (Doyle and Sambanis 2000). On the severity of
the war (e.g., the number of battle deaths), Doyle and Sambanis (2000) find a
significant and negative correlation with the recurrence of violence, but Walter
(2004) finds no such relationship.
In terms of identity or ethnic composition, scholarship is ambiguous. Some
researchers argue that greater ethnic diversity of a society makes civil war more
likely (Gellner 1991). However, more discriminating theories have gained cur-
rency, some of which draw on linguistic fractionalization indices that measure
the chance that two randomly selected individuals will speak the same language.
Ethnic fractionalization, which applies the same test to ethnic identity, has
What Do We Know about Why Peace Fails? 29

Table 1.1 Risk Factors for Civil War Onset and Recurrence: Quantitative Findings
Political factors
Powersharing agreements Hartzell and Hoddie 2003, 2007
Nonmajoritarian democratic system Mukherjee 2006
State institutions
Weak state institutions Fearon and Laitin 2003
Local capacity Doyle and Sambanis 2006
Coercion / military factors
Third-party guarantees Walter and Snyder 1999; Walter 1997, 2004;
Hartzell and Hoddie 2003
International military capacity (multi- Doyle and Sambanis 2006; Collier et al. 2003;
dimensional) to support negotiated Fortna 2004a
settlements
State with large military Fortna 2004a
Existence of DDR program Collier et al. 2003
Military spending Collier and Hoeffler 2006
Political economy indicators
Low infant mortality rate Walter 2004
Low GDP per capita Collier et al. 2003; Fearon and Laitin 2003
Economic growth Collier et al. 2003; Fearon and Laitin 2003
Social factors / conflict dynamics
Ethnolinguistic fractionalization Collier and Hoeffler 2000
Identity wars Doyle and Sambanis 2000
Duration of war Walter 2004; Fortna 2004a, 2004b
Partition Walter 2004; Sambanis 2000
Lack of large diaspora Collier et al. 2003; Lyons 2006
War ending in victory Collier et al. 2003; Suhrke and Samset 2007;
Quinn, Mason, and Gurses 2007; Licklider
1995; Hartzell and Hoddie 2003

yielded mixed results. Walter finds no significance in her study specifically of war
recurrence. However, Sambanis (2000) finds that Fearon and Laitin’s (2003)
variable ethnic fractionalization holds negative statistical significance for war
recurrence. He also finds that ethnic heterogeneity and ethnolinguistic fraction-
alization have no statistical significance for peacebuilding (a dependent variable
that includes a measure of recurrence).
Related to the questions of ethnic and religious identity is that of whether
a state breaks apart or achieves independence. Walter (2004) and Sambanis
(2000) found a high correlation between partition and war recurrence. Because
secession is generally tied to group identity and the nationalistic claims arising
therefrom, partition may signify a preemptive or failed attempt to manage eth-
nic conflict. It is apparent, however, that explanations emphasizing ethnicity or
30 Why Peace Fails: Theory

religious identity have fallen out of favor in the relatively limited scholarship on
civil war recurrence.
In sum I find tremendous disparity among scholars about whether certain
factors are important or not, and about the degree to which they are important.
Many variables are relatively static and do not lend themselves to easy manipu-
lation. Chapter 2 returns to this topic and draws on the variables mentioned in
this section, for there I present a quantitative analysis of the question of civil
war recurrence. However, it bears mention here that such methods require us-
ing indicators for variables, and finding accurate proxies for the variable is often
difficult, especially in poor or war-torn societies where data are often unavailable
or have been destroyed. These methods have particular difficulty getting at in-
ternational strategies. An early creative effort to address the role of international
actors and their different efforts was the work of Doyle and Sambanis (2000,
2006), who measured the presence of different types of UN peace efforts. Fortna
(2008) uses quantitative methods and case studies to show that peacekeeping
contributes to the duration of peace. She utilizes hazard or survival models to
address the problem of “censored” data whose projected future values cannot be
known (Fortna 2008, 10–11).5

Four Approaches to Peacebuilding

The contemporary scholarly debates on peacebuilding represent the culmina-
tion of two decades of rapid evolution of thinking among scholars and policy-
makers. Only twenty years ago, the study of peacemaking efforts to end civil wars
was a neglected field. That neglect partly reflected the marginal status of civil
wars more generally within scholarly inquiry and policy analysis, especially in
the United States. Wars between states—especially major powers—dominated
US security studies and policy debates, and peacemaking in civil wars was lim-
ited to those who chose to examine regional dynamics (e.g., Central America or
Southern Africa). Since then, however, internal wars have assumed prominence
in both policy and scholarship on international security. Postconflict reconstruc-
tion, the related concept “peacebuilding,” and the problems of fragile states rep-
resent new widely studied subfields within international relations.
A number of competing scholarly approaches, often conflated, have emerged.
These have emerged in historical and conceptual relation to one another. This
evolution was heavily shaped by the perceived “lessons” of successive experi-
ences of peace operations among international actors. Especially influential here
are the policy documents and thinking of actors engaged in militarized peace-
keeping or humanitarian operations, such as Central America, Cambodia, South-
ern Africa, Bosnia, Kosovo, East Timor, West Africa, Afghanistan, and Iraq.6 The
evolution of thinking about peacebuilding failure (and success) shows increasing
What Do We Know about Why Peace Fails? 31

attention to the multiple relationships among elements of peace operations; the

aggregation of new tasks and priorities with each set of “lessons learned” (rather
than discarding prior ideas that have more properly been superseded); greater
technification of core tasks, such as constitution writing and the rule of law; em-
phasis on coordination achieved in form more than fact; and a greater consensus
(with lagging practice) on the role of national and local actors.
This section presents what I consider the four dominant, coherent conceptual
approaches to peacebuilding in postconflict societies. A large literature on peace-
building that emphasizes social-level dialogue and reconciliation is not explicitly
included here, although it has much to offer that complements, and critiques,
some of the state-centered peacebuilding approaches considered below. In ad-
dition, the literature on mediation and conflict resolution—or “peacemaking”
in Boutros Ghali’s (1992) terms—serves as a backdrop to my understanding of
peacebuilding. However, that literature largely focuses at arriving at a cease-fire
among warring parties rather than how to consolidate peace once a cease-fire has
been established. The four approaches presented here are (1) liberal peacebuild-
ing, (2) republican peacebuilding, (3) peacebuilding as governance and / or state
building, and (4) critical approaches to peacebuilding.

Liberal Peacebuilding
The most prevalent alternative to an economic opportunity model of civil war
and peacebuilding is a political approach that has come to be called “liberal
peacebuilding.” Although no theorist holds himself or herself up as the avatar of
“liberal peacebuilding,” the concept is useful because it depicts a collection of ele-
ments that seem to underlie and infuse international policies and programs, seen
as a whole. In essence, liberal peacebuilding consists of the promotion of democ-
racy and of market-based economic reforms in postwar societies. Roland Paris
(2004, 5) defines it thus: “In the political realm, liberalization means democrati-
zation, or the promotion of periodic and genuine elections, constitutional limita-
tions on the exercise of governmental power, and respect for basic civil liberties,
including freedom of speech, assembly and conscience. In the economic realm,
liberalization means marketization, or movement toward a market-oriented eco-
nomic model, including measures aimed at minimizing government intrusion in
the economy, and maximizing the freedom for private investors, producers, and
consumers to pursue their respective economic interests.”
In the early years of the twenty-first century, liberal peacebuilding became the
label capturing the overall thrust and assumptions of Western donors and inter-
national organizations in their policies toward postconflict countries. Liberal-
ism inconvertibly captured common elements of international strategies, which
included the explicit expectation that relatively free elections would mark the
32 Why Peace Fails: Theory

transition of conflict resolution from the battlefield to the political realm. These
strategies also included liberal notions about providing the necessary supportive
social environment for electoral competition through civil liberties enshrined in
constitutions as well as space for civil society to represent interest groups and
check government behavior.
A sophisticated example of liberal peacebuilding is the work of Michael Doyle
and Nicholas Sambanis, whose 2000 article was the most serious quantitative
attempt to identify factors that account for peacebuilding success. They found
that international capacities deployed (including multidimensional mandates in
peacekeeping missions), local capacities (measured with indicators like electrical
capacity), and the ease of the conflict context (e.g., fewer and more coherent in-
surgent groups) proved decisive in increasing the chances for peacebuilding suc-
cess. They found that UN mediation efforts alone and traditional deployments,
without expansive efforts to build state institutions and hold elections, had no
statistical influence on peacebuilding success. However, more multidimensional
efforts—those “with extensive civilian functions, including economic recon-
struction, economic reform, and election oversight”—were statistically sig-
nificant for success. One of the most significant implications of their research,
enhanced in their book Making War and Building Peace (Doyle and Sambanis
2006), was to ratify the importance of expansive UN peace operations that as-
pired to the liberal goals of transforming societies. Thus, although Doyle and
Sambanis did not label their work “liberal,” it set a foundation for liberal peace-
building approaches by emphasizing elections and economic reform.
The liberal character of international policies has been most explicit in the
economic domain. International actors, especially international financial insti-
tutions like the World Bank and the International Monetary Fund, used their
influence to advance minimal states in postconflict countries. These programs
assumed that minimizing the state would animate free enterprise and interna-
tional investment and provide the space for the private sector to create growth
and jobs, and thus help to ease poverty and integrate the local economy into the
globalized economy.
One aspect of the liberal approach merits reconsideration: its combination of
economic and political liberalism. Paris (2004), for instance, defines the “liberal
international” approach to peacebuilding as including both neoliberal economic
policies and liberal democratic models centered on elections. However, “liberal”
political actors and policies frequently diverge from neoliberal economic pre-
scriptions, as shown in an early critique by the UN diplomats who mediated
the peace agreement in El Salvador. In a seminal article, de Soto and del Cas-
tillo (1994) bemoaned the neoliberal policies of the economic wing of the in-
ternational community (mainly the international financial institutions), which
worked at odds to their efforts at political (liberal) peacebuilding. Conversely,
What Do We Know about Why Peace Fails? 33

advocates of Asian models of economic development have argued that neolib-

eral policies are best pursued by nonliberal, nondemocratic political regimes. It
seems prudent to eschew the conflation of economic and political liberalism in
conceptualizing peacebuilding.

Republican Peacebuilding
Illiberal peacebuilding, however, is not necessarily authoritarian. Michael Bar-
nett (2006) proposes a “republican” alternative to liberal peacebuilding. His ap-
proach comes closer to providing the scaffolding for an inclusionary, legitimacy-
focused approach to peacebuilding by offering valuable remedies for liberalism’s
neglect of issues of exclusion and legitimacy. For instance, Barnett argues (2006,
94) that modern republicanism (represented by Niccolò Machiavelli and James
Madison, rather than their republican predecessor Aristotle) is concerned largely
with stability in the face of contentious factions. Drawing on republican theory,
Barnett suggests three core elements of republican postconflict peacebuilding:
deliberation, representation, and constitutionalism.
On deliberation, Barnett (2006, 94) praises Machiavelli’s view of “the domes-
ticating influence of public deliberation; by compelling individuals to speak in
the language of community they might develop a greater sense of patriotism, that
is, a love of country.” In line with the need for the multivalent forms of participa-
tion I advanced above, he suggests that representation go beyond democratic
elections to include alternative forms of participation (p. 102). Such participa-
tion may include consultations but rests principally with representatives that es-
chew direct participation by the masses. Finally, Barnett calls for constitutional
constraints that divide power and regulate the behavior of representatives. He
contends that “all constitutional systems have (1) an agreement over the rules of
the game and the underlying principles that are to maintain the political order;
(2) rules and institutions that limit the exercise of power; and (3) rules that are
relatively difficult to amend” (p. 105). In this Barnett falls not far from liberal
theory and its concern with the social contract.

Peacebuilding as State Building

Although political approaches to negotiation and postconflict reconstruction
have long existed, they came to the fore in a more coherent manner by the early
2000s. The failed cases of Angola in 1992 and Rwanda in 1994 showed that suc-
cessfully reaching a peace agreement was not sufficient to secure peace. Research
on peace implementation (Stedman, Rothchild, and Cousens 2002) emerged
that explored the various challenges of preventing war recurrence. By 2000 the
experiences of Bosnia, Haiti, Sierra Leone, Kosovo, and East Timor had shown
34 Why Peace Fails: Theory

that international peacekeeping troops could maintain stability, but it was ap-
parent that the withdrawal of peacekeepers meant a renewed risk of conflict. By
2001, for instance, the Dayton-negotiated cease-fire had held for five years in
Bosnia, with no end in sight for the need for NATO troops.
The main instrument of self-sustained peace, it was increasingly believed by
policymakers, was a self-sustaining polity or state with the ability to resolve con-
flict without reliance on outside troops. In essence the success of these peace
operations depended on the emergence of a successful state, which made gover-
nance the key to peace (Caplan 2004a, 2005; Chesterman 2001, 2004; Beauvais
2000). Some of the works in this field focus on governance broadly, especially
whether democracy emerges or not. Thus a burst of scholarship unfolded on
postwar democracy (Call and Cook 2003; Bermeo 2003; Ottaway 2000), on
powersharing (Sisk 1996; Hartzell and Hoddie 2003, 2007; Hoddie and Hartz-
ell 2005), on partition (Kaufmann 1996, 1998), and on governance (Sriram
2008). These approaches argued that the formula or the institutions of gover-
nance constituted key factors in the sustainability of postwar peace. They are
considered in more detail below.
In addition, scholars took up the question of the state as it relates to consoli-
dating peace. Some scholars believed that the liberal peacebuilding model was at
fault, and that its central weakness was insufficient attention to institutions and
the state. As exemplified here by the work of Roland Paris (2004), policymakers
and analysts concluded that spoilers were taking advantage of liberal practices
to block peace consolidation. Consider, for example, how former armed parties
in West Africa and the Great Lakes region used multiparty elections and liberal
freedoms to mobilize voters via hate, to engage in exclusionary behavior once
elected, and to mobilize opponents of peace through liberal democratic mecha-
nisms. Paris’s answer was “institutionalization before liberalization.” Stronger
state institutions must first be in place in order to moderate politics, to check ex-
treme behavior by elected officials, and to provide controls on hate-mongering in
the media and in public discourse (see also del Castillo 2008). In Paris’s (2004,
187) words, “What is needed in the immediate postconflict period is not demo-
cratic ferment and economic upheaval, but political stability and the establish-
ment of effective administration over the territory.”
The terrorist attacks of September 11, 2001, attributed to a “failed state” in
Afghanistan, elevated the emphasis on “state building” in both fragile states
and postwar countries. Thus a spate of new research focused on state building
emerged (Fukuyama 2004; Ghani and Lockhart 2008; Chesterman, Ignatieff,
and Thakur 2005; Call 2008a; Ghani, Lockhart, and Carnahan 2005), which
contributed to a new understanding of peacebuilding as consisting largely of
state building. The state-building approach focused international efforts on
strengthening the capacity and legitimacy of state institutions. It struck a con-
trast to prior emphases on replacing the state (as in East Timor; see Krasner
What Do We Know about Why Peace Fails? 35

2004), to liberal peacebuilding and to the development community’s emphasis

on civil society programming. To be clear, state building was developed prin-
cipally as a response to “failed,” “failing,” or “weak” states—not specifically to
postconflict countries. However, given that the multiple definitions of “state fail-
ure” tend to include at-war and postwar societies (Call 2011), the concepts have
become increasingly conflated.
Scholarship elaborated further on the challenges of state building. Fukuyama
(2004) drew on new institutional economics and public administration theory
to chart recommendations for international state builders. Ghani and Lockhart
(2008) set forth ten state functional areas of primary importance on which na-
tional and external actors should concentrate in the creation of a viable state.
Newer works, including case-specific accounts of efforts in Afghanistan, Bosnia,
and elsewhere (Rubin 2006; Berdal and Caplan 2004; Caplan 2004a, 2004b;
Fayemi 2004) pointed to the complications and tensions in state building in
postconflict environments (Call 2008a; Paris and Sisk 2009; Brahimi 2007; del
Castillo 2008).

Critical Theories of Peacebuilding

A group of scholars has recently launched a concentrated critique of state build-
ing, folding it into what it depicts as the liberal peacebuilding model. This ap-
proach has come to be broadly referred to as critical peacebuilding. It is critical
of what it sees as the status quo policy frameworks and of the hegemonic (Tay-
lor 2010) or neo-imperial character of the peacebuilding activities of the world’s
dominant bilateral and multilateral actors.
Critical peacebuilding theorists reject liberal peacebuilding as a universaliz-
ing, disciplining exercise of power by Western and other powerful countries on
poor societies. Richmond and Franks (2009, 6) claim that, through liberal peace-
building, “international actors draw on the epistemic knowledge that liberalism
supplies and represents in order to reorder the distribution of power, prestige,
rules and rights.” He continues, “The balance of power, hegemony, institutional-
ism and constitutionalism, and civil society converge in this version of peace in
an era of governmentality.” In this view (Richmond and Franks 2009, 6), the
state’s institution-building agenda is part and parcel of liberal constitutionalism,
because these two “have recently converged on a contemporary notion of ‘peace-
as-governance,’ whereby the liberal state provides the framework for the creation
of peace at the local, state and international levels.” This notion “in state-building
terms focuses on the institutions of the state as the basis for the construction of
the liberal peace.” Along these lines, Chandler (2006, 1) criticizes peacebuilding
efforts as “highly intrusive.”
This emergent literature is the most sustained, cogent scholarly alternative
to the liberal peacebuilding approach. However, its conflation of state building
36 Why Peace Fails: Theory

with the liberal approach seems mistaken. The liberal vision is not philosophi-
cally or historically rooted in an emphasis on state building. As Paris emphasizes
(2004), classic liberal philosophers like Locke and Kant assumed the presence
and prevalence of institutions of governance with some degree of state capac-
ity. Political liberalism emphasizes liberal civil rights, the individual as the key
unit of analysis, and elections as an elevated expression of individual preferences.
However, states and state institutions were not central to their understanding of
political development, of freedom, or of rights. These institutions were ancillary.
Liberal theory includes institutions but does not emphasize them.
In contrast state builders emphasize limitations on rights, the collective as the
unit of analysis, and institutional strength (rather than elections) as the keys to
effective governance. State institutionalization may reflect Western assumptions
and values, but it is not liberal. Instead, the emphasis on state institutionalization
reflects the conservative, republican value of order more than it does the em-
phasis on rights and individual freedom inherent in liberalism (Barnett 2006).
To paint state building and its concerns with the same brush as liberalism and
its concerns is to conflate important and distinct approaches to peacebuilding.
Furthermore, neither political liberalism nor state-building theorists necessarily
embrace conventional neoliberal economic policies, as Richmond and Franks
(2009) and Paris (2004) suggest.
At the same time, critical theorists’ contributions merit attention. They have
called for greater attention to the particularities of cultural and social context,
and have decried cookie-cutter bureaucratic approaches. Richmond and Franks
(2009, 182) argue that “the local, the everyday, the customary, human needs,
human security, and the social contract need to be positioned far higher up the
practical agenda than they currently are.” Richmond and Franks (2009, 33) have
called for “a detailed understanding (rather than co-optation or ‘tolerance’) of
local culture, traditions, and ontology.” They have also called for unmasking the
interests of powerful states and economic elites that underlie peacebuilding pro-
grams. Richmond and Franks (2009, 34) insist that “international support for
these processes” not introduce “hegemony, inequality, conditionality, or depen-
dency.” Critical peacebuilding theorists reflect frustration with the absence of
local voices in deciding what governance models will be adopted in their society,
and the injustices of external and self-serving imposition of certain economic
and political institutions.

Clarifying Concepts: Exclusion, Inclusion, and Legitimacy

These multiple approaches to peacebuilding help one understand the process
and explain why some countries have such difficulty attaining a self-sustained
peace, even once stability seems at hand. The recognition of the importance of
institutionalization of authority, and the important challenge to state legitima-
What Do We Know about Why Peace Fails? 37

tion, remain underemphasized and undertheorized. Because the concept of le-

gitimacy lies at the center of this book, I spend some pages here mapping the
concept and how it operates in postwar societies, recognizing the variety of states
that have experienced warfare.
This section seeks to clarify the main concepts used in the empirics and in
the theoretical discussion of the book. It begins by elaborating on the discussion
of the concept of exclusion in the introduction. I then explain how exclusion
and inclusion relate to state legitimacy, a difficult but useful concept that is espe-
cially important in improving theoretical approaches to postwar peacebuilding.
Finally, powersharing figures prominently in inclusionary arrangements and also
merits some discussion. In addressing these important concepts, where appro-
priate I relate them back to the four main theoretical frameworks for peacebuild-
ing. In seeking conceptual clarity, this section does not summarize the extensive
relevant literature on these terms.

Exclusion and Civil Wars

As indicated above, scholars have long looked to grievance as a cause for col-
lective violence, including internal armed conflicts. Among these grievances
is the exclusion of important social groups, often those rooted in ethnic and
religious identities. Exclusion is generally associated with “social exclusion,” a
term advanced by social scientists in the 1970s as a more discerning concept
than “poverty” per se, which did not distinguish between those whose poverty
also involved a lack of access to public or social goods (Levitas 2005; Hunter
and Jordan 2010). Scholars have written less on cross-national political exclu-
sion, often in connection with citizenship (Wimmer 2002). One influential
research project is based on the collective action of minorities and associated
with Ted Robert Gurr (1993) and his colleagues, who created the Minorities
at Risk data set. Exclusionary behavior is mediated by factors such as political
opportunity structures and electoral systems (inter alia, Horowitz 1985), demo-
graphics (Saideman and Ayres 2000), and the strength of loyalties and attach-
ments to group and state (Cohen 1997; Dowley and Silver 2000; Elkins and
Sides 2007).
As defined in the book’s introduction, “exclusion” here refers to the perceived
or actual deprivation of an expected opportunity for former warring parties, or
the social groups associated with them, to participate in state administration,
through either appointed posts or elective office. In separatist struggles those
opportunities are concentrated in authority over a substate territory. As specified
above, exclusion is conceptualized here vis-à-vis state processes rather than sub-
stantive outcomes, and its specific manifestation (not the definition) will vary
across cases depending on the expectations generated among the parties. The
fulfillment of expectations regarding access to state positions—or “inclusionary”
38 Why Peace Fails: Theory

postwar behavior—responds both to the need for security guarantees among

former rebels and the need for some degree of voice in governance among the
formerly mobilized elites associated with them. Inclusionary behavior closely
corresponds to an ability to avoid civil war recurrence.
Exclusion takes different forms. Exclusionary behavior varies along three im-
portant lines: by who does the excluding (the agent), by who is excluded (the
target), and by its severity. In terms of the agent, the excluding agent can be either
the state or a nonstate actor—the latter either an unarmed / social or armed /
military-political insurgent force. However, for a nonstate actor to engage in ex-
clusionary conduct to an extent that provokes armed conflict, that nonstate actor
would need to exercise quasi-state functions. The most likely scenario here is an
autonomous substate authority that exercises authority over a territory, and then
engages in exclusionary behavior in political, economic, security, or economic
arenas. Because the cases here yield no such cases of nonstate exclusionary be-
havior sparking armed conflict, exclusionary behavior here refers exclusively to
states as agents.7
In postwar settings, exclusionary behavior varies also by targets, principally
the treatment of former combatant groups and the populations associated with
them.8 Having fought and sacrificed loved ones in a bloody struggle against an
enemy renders it difficult to coexist peaceably, much less endorse placing those
former enemies into positions of authority. Where civil wars end in victory by
one side, it is unsurprising that the incorporation of former enemies is difficult.
As such (see the next chapter), exclusionary behavior after victories has in earlier
decades not always led to war recurrence. Three types of targets of exclusion-
ary behavior are relevant for war recurrence after a prior war has ended: (1) the
former insurgent forces that have relinquished their arms and either been rein-
corporated into civilian life or the state’s armed forces;9 (2) the social or eth-
nic groups associated with prior insurrection; and (3) former allies in the prior
armed struggle. Although other social groups and elites may spark new internal
armed conflicts, these three types of targets are those that prove relevant for re-
lated, rather than new and unrelated, recurrences.
The severity of exclusion varies in self-evident ways. At the worst, states can
completely subordinate a given ethnic or other social groups in ways that deprive
them of access to education, cultural rights and patrimony, jobs, social services,
positions of authority, and / or a host of human rights including physical integrity
in the face of repression.10 However, the severity of exclusionary behavior does
not necessarily serve as the trigger or a precipitating factor for armed conflict.
Severe repression against a people or a political group can persist for decades or
centuries without resulting in armed conflict.11
To be sure, exclusionary behavior is rooted in broader social configurations
shaped principally by struggles among elites within a society. Elites mobilize na-
tional, ethnic, or religious identities in pursuit of interests, and these processes
What Do We Know about Why Peace Fails? 39

are embedded in long historical processes of industrialization and colonization,

and in periods of history with certain ideas of which political forms and relations
are legitimate and which are not (Greenfield 1992; Tilly 1975; Mann 1993). In
recent decades the interaction of elite struggles with international incentives,
ideas, and coercion constrain or even reformulate the lines along which exclu-
sionary behavior by elites takes places (Chatterjee 1993). These complex and
interactive historical processes are not the focus of the theoretical distinctions
offered here. Nor are they taken up explicitly in the empirics of this study. In-
stead their eventual culmination in exclusionary behavior, like the visible tip of
an iceberg, is what tears the ship of postwar stability. The submerged historical
and complex social relations underlying the recent expression of exclusionary
behavior are not analyzed.

Inclusion and Powersharing

Given that inclusionary behavior is here defined as the opposite of exclusionary
conduct, it consists of behavior that does not exclude former parties to warfare or
the groups associated with them from access to state positions in ways contrary
to their expectations. This broad definition encompasses multiple behaviors
and thus extends well beyond an agreed-upon fixed share of state offices. Most
particularly, inclusion here includes electoral regimes where a former party to
warfare fails to win many parliamentary seats but the process and outcome com-
port with prior expectations about governance. Thus democratic governance or
electoral regimes in concept (and based on expectations) offer just as acceptable
forms of inclusion as powersharing. “Inclusion” here is compatible with control-
ling a tiny portion of a country’s local districts or mayorships, with controlling
the production and profits of a valuable natural resource even in the absence
of parliamentary representation, and even with a degree of exclusion from state
offices—so long as these circumstances did not violate the expectations of the
relevant parties.
As seen below, the empirics of this book demonstrate that inclusionary
conduct is associated with the nonrecurrence of civil wars. Moreover, the case
studies permit more specific claims about one form of inclusionary behavior:
powersharing. Consequently, I discuss some of the main conceptual debates and
empirical findings about powersharing here. This focused analysis of and discus-
sion of powersharing, here and later in chapters 7 and 8, should not be taken to
equate inclusionary behavior with powersharing. On the contrary, inclusionary
behavior is (forgive me) an inclusive concept, one that allows for a plurality of
manifestations that do not constitute exclusion.
I define powersharing as a system of governance in which social, political,
or ethnic groups are guaranteed a role in their own governance either through
allocated positions in the state or through institutionalized group participation
40 Why Peace Fails: Theory

in administrative, electoral, or appointive procedures.12 Powersharing is often

associated with Arendt Lijphart’s (1977) concept of consociational democ-
racy, but in fact it is a broader concept that encompasses other mechanisms of
guaranteed access to state authority. Two decades ago the main lines of debate
over powersharing contrasted consociational approaches with “integrative” ap-
proaches, emblemized by Horowitz’s (1985) sweeping comparative work on
ethnic formulas for governance. Integrative approaches offered incentives for
groups to work across ethnic or group lines in electing or selecting state lead-
ers, whereas consociational approaches emphasized guarantees to those groups
that reinforced their identities. More recent theoretical and empirical work has
honed these categories considerably (McGarry and O’Leary 2003, 2004; Sisk
1996; Sriram 2008; Wolff 2009; Mitchell, Evans, and O’Leary 2009; Milante
and Skaperdas 2009).
The research on powersharing is divided about its effectiveness in prevent-
ing recurrence to warfare. Some studies find that powersharing has little impact
on reversion to civil war, yet a preponderance of recent studies find that power-
sharing has a statistically significant impact in reducing the chances for civil war
occurrence and reversion. Jarstad (2008), for instance, finds that when power-
sharing precedes parliamentary elections in postwar societies, peace generally
lasts longer. Norris (2008) also finds that political forms of powersharing help
peace persist.
More recently, Caroline Hartzell and Matthew Hoddie have added both con-
cept and empirics to the analysis of powersharing. Historically, powersharing
has referred mainly to “political powersharing” (referring to government posts
or processes in the legislature or the executive branch). However, Hartzell and
Hoddie (2007) and Hoddie and Hartzell (2005) identify four types of power-
sharing: political, security (referring to the military, police, or security forces),
economic (referring to access to resources or processes of economic decision
making), and territorial (referring to forms of territorial autonomy). I adopt this
concept of powersharing in the case studies that follow. Peace processes have ex-
perienced increasing incorporation of multiple forms of powersharing in the past
two decades, and Hartzell and Hoddie (2003) argue that the success of peace
agreements is enhanced where multiple forms of powersharing are adopted. Us-
ing survival analysis (which acknowledges that the chances of failure are higher
in earlier years after war ends), Hoddie and Hartzell (2005) found that in 38 ne-
gotiated settlements, peace endured longer the more types of powersharing that
were adopted. They also (Hartzell and Hoddie 2003, 2007) found that power-
sharing in military and police institutions has a statistically significant favorable
correlation with longer periods of postwar peace. Binningsbø (2006) also found
that two forms of powersharing—territorial autonomy and proportional repre-
sentation—led to longer periods of peace in 118 postconflict societies.
What Do We Know about Why Peace Fails? 41

Powersharing has enjoyed wide acclaim in policy approaches to settling armed

conflicts in recent years. For example, Ted Robert Gurr (quoted by Sriram 2008,
17) states that powersharing and resource sharing have become part of emerging
doctrine for conflict management. However, critics of powersharing have also
grown in number and scholarly influence. The main criticisms are that power-
sharing exhibits (1) inflexibility; (2) a tendency to marginalize broader partici-
pation, at the expense of “vertical legitimacy”; and (3) exclusion of the most mar-
ginal or tertiary ethnic, social, or political groups.13 Powersharing arrangements,
in their quest to reassure insecure parties, often make no provision for their own
alteration, which creates incentives for those who see themselves losing under
these arrangements over time to resort to violence. As Rothchild and Roeder
(2005, 38) state, powersharing institutions tend to be “unable to adapt to rapidly
changing social conditions during a transition from intense conflict.”
Powersharing also exhibits inflexibility in its reification of the very identi-
ties that underlie division and potential violence. Where power is apportioned
according to identity, it reinforces those identities relative to others, undermin-
ing the potential to erode those distinctions in society. It “privilege[s] issues of
interethnic allocation of power and resources,” at the expense of other issues
(Rothchild and Roeder 2005, 37). According to Sriram (2008, 12), power-
sharing “may be counterproductive, channeling existing conflicts and mistrust
into institutions of governance.” These deficiencies are explored in the case stud-
ies that follow, as well as in the book’s concluding theoretical reflections.
The empirics that follow suggest that powersharing and other forms of in-
clusion, as well as exclusion, are important empirical elements of civil war re-
currence. However, their theoretical significance is less apparent. To relate these
concepts to competing theoretical approaches to peacebuilding, it is necessary to
explicate their relationship to the broader character of the state and its relation-
ship to society. I now turn to state legitimacy.

How Do Exclusion and Inclusion Relate to State Legitimacy?

Fueled by increased policy and scholarly attention to failed states and state build-
ing, the state has enjoyed a revival in research on war and peace.

Defining Legitimacy
I adopt here a definition from Mark Suchman (1995, 574) of legitimacy drawn
from organization theory that encompasses both the relationship between an
institution and its external environment as well as its relations with its internal
components or constituents: “Legitimacy is a generalized perception or assump-
tion that the actions of an entity are desirable, proper, or appropriate within some
socially constructed system of norms, values, beliefs, and definitions.”
42 Why Peace Fails: Theory

Three points are worth noting about this conceptualization of legitimacy.

First, legitimacy is an objective quality, but one that rests on subjective evalua-
tions. The perceptions or assumptions of people are inherent in whether legiti-
macy exists or does not. At the same time it rests not with specific individuals
but a collective. Second, legitimacy is socially constructed, and thus it cannot be
understood outside a temporal and spatial context. Consequently, understand-
ings and assessments of legitimacy can and do change over time and space.
Third, the definition explicitly encompasses the three main forms or perspec-
tives on the nature of legitimacy—what Suchman (1995, 572) calls “pragmatic”
legitimacy, based on the self-interest of the constituents; “moral” legitimacy,
based on oughtness or normative appeal; and “cognitive” legitimacy, based on
appropriateness, taken-for-grantedness, or embeddedness. Pragmatic legitimacy
accrues when constituents perceive a particular organization to serve their inter-
est and lend it their support, loyalty, and eventually perhaps a moral or cognitive
legitimacy.14
Legitimacy is a notoriously difficult concept to measure without making as-
sumptions about the freedom of the public or constituents to express their opin-
ions, which in the extreme can, unfortunately, directly reflect legitimacy. One
measures or “knows” the legitimacy of a regime or state given the extent to which
its authority is accepted or unquestioned. Yet the degree to which conformity
to rules represents consent, as opposed to obeisance in the face of coercion, is
difficult to determine. The absence of opposition or of vocal expressions of sup-
port for a particular regime may reflect the effective application of coercion. It
is impossible in an authoritarian or totalitarian regime to have confidence in the
results of even the best designed methods of gauging public opinion (polling, fo-
cus groups, written surveys, etc.). The possibility of psychological and social ha-
bituation to coercion among regime opponents or victims renders it impossible
to discount (or even adjust for) respondents’ fear and suspicion. No concept of
legitimacy embraces “support” deriving from coercion or its threat.
And yet authority involves the potential or actual application of force. More-
over, history is replete with examples of states or regimes that emerged through
coercion and repression yet eventually came to be seen as “legitimate” among
their populations. Legitimacy based on performance or pragmatism can be
transformed into a legitimacy that is more cognitively sensible and even mor-
ally “right,” which undermines the notion of individual choice and moral clarity.
Does this defect undermine the concept of legitimacy? Not as a rule. How-
ever, it imposes a certain minimal threshold of freedom of expression in order for
the possibility of perceptions to be empirically captured with some confidence.
In most societies perceptive analysts know when pervasive fear exists and under-
mines freedom of expression. Here serious problems of legitimacy exist, and nu-
What Do We Know about Why Peace Fails? 43

anced measurements are impossible. Where fear and its habituation are not ap-
parent, however, it should be possible to measure degrees of relative legitimacy.

State Legitimacy: Vertical and Horizontal

State legitimacy here is defined as a generalized perception of a political com-
munity that the claim to authority by a collection of institutions over its terri-
tory is proper or appropriate. State legitimacy is thus defined by the acceptance
by a political community of the “oughtness” of the state’s exercise of authority
and coercion over the territory. It signifies the perception by the population of
the territory that the underlying political order is appropriate, moral, or proper,
including as it relates to the state’s borders.
Because “legitimacy” is often used to refer to a polity’s regime, it is useful
to distinguish state legitimacy and regime legitimacy. Regime legitimacy refers
to the perception by the population of a political community that the norms,
principles, rules, and procedures governing the forms and access to government
positions, and the resources and strategies available to office holders, are proper
or appropriate. Regime legitimacy is thus more narrowly conceived than state
legitimacy. The legitimacy of a regime refers more particularly to who governs
and how their power relates to the rights and obligations of citizens. In addition
the relationship of authority to the territorial boundaries is an element of state
legitimacy in a way that is absent from regime legitimacy.
Kalevi Holsti (1996, chap. 5) usefully distinguishes two dimensions of state
legitimacy, which pose distinct challenges for would-be legitimacy builders.
Vertical legitimacy refers to the broad sense of appropriateness of the state and
its functioning, including the rules by which leaders are selected. A state enjoys
vertical legitimacy based on an undifferentiated sense of the degree to which
the state is appropriately intervening, regulating, representing, or advancing the
interests of society.15 The extent to which the national state matters vis-à-vis dif-
ferent levels of authority also forms part of vertical legitimacy. Thus states where
the idea, weight, and activities of the national state are relatively weak—for ex-
ample, Afghanistan, Malawi, and East Timor—exhibit low vertical legitimacy.
Horizontal legitimacy refers to the extent to which various social groups and
communities within a territory “accept and tolerate each other” (Holsti 1996,
87). It refers to the socially differentiated perceptions of social groups about one
another and their relation to the state. Horizontal state legitimacy is low if it is
divided; that is, if one ethnic or social community considers the state to be ex-
tremely legitimate but another community rejects its legitimacy.
A lack of horizontal legitimacy means that legitimacy is bifurcated among at
least two social or ethnic communities within a country. As Holsti points out,
“ethnicity” connotes a narrow dimension of social division, because religion,
44 Why Peace Fails: Theory

class, caste, tribe, and language can serve to distinguish social groups (Holsti
1996, 106). Thus, the state’s legitimacy within Israel is extremely high among
Israeli Jews but is low among those living in the West Bank and Gaza. This trans-
lates into a diminished level of horizontal legitimacy for that state. A regime can
also suffer divided legitimacy if the norms, rules, and procedures governing the
inclusion or selection of rulers are not perceived as proper or appropriate by a
salient social group. The depth and breadth of that sense of impropriety—and
thus of state or regime legitimacy—can vary.
These concepts of legitimacy are linked in important ways to the theoreti-
cal debates over peacebuilding. Liberal peacebuilding, for instance, is founded
on the individual unit of analysis, one that sees the state’s relations to society
mediated by individuals who hold the state accountable. This perspective em-
phasizes vertical legitimacy rather than horizontal legitimacy. Liberal theory sees
individuals as the undifferentiated central agent in society, rather than as social
groups. Republican peacebuilding, with its recognition of minority communi-
ties and dialogue among salient social groups, has a place for both vertical and
horizontal legitimacy. The state-building perspective explicitly embraces legiti-
macy as a necessary component for consolidation of peace, but theorists em-
phasize legitimacy versus capacity to different degrees. Moreover, state-building
approaches to peacebuilding generally fail to distinguish between vertical and
horizontal legitimacy.
How, then, does exclusion relate to state legitimacy? Exclusion is not an ex-
plicit component of the four principal approaches to peacebuilding. Certainly re-
publican theory’s notion of different factions involves social groups as important
units of analysis. Exclusion would contradict this vision. Critical peacebuilding
theorists eschew the collusion of international and national elites in the man-
ufacture of a modus vivendi that leaves the population out of procedural and
substantive agreements about postwar authority. However, these approaches
emphasize distinct roles for international actors rather than particular concepts
of intrastate and state–society relations as the foundation for peace.
Horizontal legitimacy provides the conceptual entry point of exclusionary
conduct into peacebuilding theory. Horizontal legitimacy requires some degree
of inclusion of the main social groups into the polity. Inclusionary practices or
institutions are not sufficient for horizontal legitimacy, but they are necessary.
Thus, the concept of inclusionary peacebuilding, with its focus on the inclusion
of important social groups, leads to a theoretical concern with horizontal legiti-
macy. Legitimacy is a more difficult, historically rooted, and broader concept
than inclusion.
If inclusion offers a causal mechanism tied to sustained peace, then legiti-
macy, especially horizontal legitimacy, offers the “why” that such inclusion is
necessary and a frame for understanding and advancing a holistic concept for
What Do We Know about Why Peace Fails? 45

inclusionary behavior. Chapters 8 and 9, with their conclusions for peacebuild-

ing theory and policy, return to these questions and address them in more depth
in light of the empirics of the book.

International Sources of Legitimacy and Their Impact on Exclusion

and Inclusion
The main sources of state legitimacy are internal. Jones and others (2008, 17)
identify types of state legitimacy reflecting three distinct sources: (1) embedded
legitimacy, which is derived from “prior state formation or other historical dy-
namics”; (2) performance legitimacy, which reflects the “effective and equitable”
delivery of expected services; and (3) process legitimacy, which reflects how ac-
cepted and proper the rules, procedures, and institutions of policymaking and
governance are perceived to be.
Yet state legitimacy derives not only from internal sources but also from inter-
national sources. This conceptual understanding of legitimacy provides a frame-
work for understanding the role of international actors in inhibiting or fostering
exclusionary behavior in postwar states. In any given collectivity of social ac-
tors or agents, membership derives from not just the decision or behavior of the
agent but also from the acknowledgment of the other members of the collectiv-
ity that a specific agent or individual unit is “one of us” (Bull 1977). For all states,
internal legitimacy, whether vertical or horizontal, can be enhanced, generated,
or undermined by external actors.
International actors shape legitimacy in two salient ways. First, they shape
where a state’s borders lie—and even whether it exists. The underlying causes
of many present-day armed conflicts in the postcolonial Middle East, Africa,
and Asia can be traced to great power decisions about how to demarcate the
borders of their colonies (Mamdani 1996; Herbst 2000; Englebert 2000). Co-
lonialism and its aftermath did much to destroy or distort structures of authority
in the colonies, and constructed ephemeral institutions or rules that were not
consistently implemented and systematically undermined the ability of national
or local elites to build relationships of representation and accountability with
their populations.
As we jump to the twenty-first century, the international community faces
similar challenges in confronting the legacy of its earlier conduct, which is now
aggravated by additional causes of wars. In separatist struggles like those of East
Timor, Kosovo, Palestine, Moldova, South Ossetia, Abkhazia, and Nogorno-
Karabakh, the decision by several or many states to recognize the status of an
independent state is definitive in its coming into existence. Of course, this in-
ternational decision is heavily shaped by factors internal to the territory, includ-
ing whether secessionist armed forces have ousted the army of the recognized
state authority. In this sense, the military and political realities on the ground are
46 Why Peace Fails: Theory

necessary conditions for status as an independent state to receive consideration

by most states. And where disputes over borders occur, international recogni-
tion, often through the International Court of Justice, is often determinant.
Second, external legitimacy or recognition shapes who speaks for or repre-
sents the population of a state. External actors can strengthen or weaken a par-
ticular group’s claim to hold sovereign authority within a state. International
recognition holds the key to status, a seat at the United Nations, and the appoint-
ments of diplomats to other capitals. It confers whatever power derives from
the political and social connections with other states, given that they and the
intergovernmental bodies they create deal almost exclusively with the externally
recognized government as the expression of popular will and consent in a state.
The group recognized externally as holding state power is the counterpart for
international treaties, trade negotiations, and agreements on regulations for ex-
ports or for the production and removal of natural resources. Equally important,
that group controls access to the development aid and other revenues provided
by other states and international organizations.
External and internal sources of legitimacy interact in complex ways that
make it inadvisable to try to examine solely one absent reference to the other,
as if in a vacuum. For instance, external recognition of a particular government
may provoke a backlash, which undermines claims to internal legitimacy. Inter-
nal support for a government, where based on ethnic or religious loyalties, may
make external powers reluctant to extend their support. It is hard to exaggerate
the importance of the international dimension in most postconflict countries,
where questionable authority presents opportunities for external actors to influ-
ence initial processes of state legitimation or relegitimation. In addition, many
states, especially in Africa, have polities whose principal sources of legitimacy lie
outside their borders. As Jackson and Rosberg (1986, 2) said in a seminal article,
“most Tropical [sic] African states exist primarily by means of international le-
gitimacy.” The legitimacy of war-torn states also reflects the interaction of force
and persuasion, and international organizations and outside states are often ac-
tive both in deploying troops and in exercising leverage.
What are the implications of the external sources of legitimacy for interna-
tional engagement with exclusionary behavior? Theorists of legitimacy have de-
veloped nuanced understandings of the interaction of external and internal ac-
tors in the complex production of state legitimacy. Their analysis offers insights
into the range of ways that external actors shape the exclusionary policies and
practices of postwar states.
Concepts of exclusion and legitimacy form the foundation for the empirical
work and theoretical arguments that follow in this book. State legitimacy is, the
cases will reveal, in many situations foundational to the persistence of peace in
What Do We Know about Why Peace Fails? 47

postwar societies. Capacity is important, as recent work has underscored. How-

ever, inclusionary behavior occurred in the vast majority of cases of nonrecur-
rence. State legitimacy—especially horizontal legitimacy, including its interac-
tive external and internal sources—offers concepts whereby one can understand
some of the persistent failures of peace.

Conclusion
Understanding the question of why peace fails or succeeds requires examining
two overlapping and related clusters of scholarship. The first is highly quantita-
tive in nature and focuses on risk factors for civil wars’ onset and recurrence. A
good deal of research on civil wars’ onset emerged in the wake of the Cold War.
The most influential body of this research is quantitative, and it points to widely
agreed-on variables as risk factors for onset. These variables include low gross
domestic product per capita, falling gross domestic product per capita, infant
mortality rates, mountainous terrain, and the prior experience of civil war. This
literature largely downplays the importance of ethnic or religious identity, and
other factors that may correspond to grievance-based approaches to the question
of civil war.
The second cluster of research focuses on external, and to a lesser extent in-
ternal, efforts to maintain and consolidate peace, which I bring under the widely
used term “peacebuilding.” Although many distinct approaches to peacebuild-
ing might be identified, I have focused here on the four coherent approaches
linked to theoretical traditions and roots—liberal, republican, state building /
governance, and critical. In the past few years liberalism has become the prevail-
ing framework for peacebuilding. As with realism in the theory of international
relations, liberal peacebuilding in its diverse guises occupies a central reference
point for most theoretical work. Those alternatives have offered criticism aimed
at either replacing or improving the dominant liberal peacebuilding framework,
but they have not yet supplanted it with a coherent alternative.
As is seen in later chapters, this book follows in this vein; it chips away at the
dominant liberal model and ultimately offers a theoretical approach that is sym-
pathetic to state building but explicitly challenges some of the assumptions, em-
phases, and recommendations of liberalism. My theorizing relies heavily on the
concepts of exclusion, inclusion, and state legitimacy explicated above. But be-
fore taking up this endeavor, I present the book’s empirics. The next chapter thus
begins with the widest possible scope of analysis, by examining the entire pool of
civil wars and analyzing those that have experienced recurrence and those that
have not. It takes up the quantitative findings and knowledge presented in this
chapter and applies them to the phenomenon of civil war recurrence.
48 Why Peace Fails: Theory

Notes
1. For discussions of multiple perspectives, see Brubaker and Laitin (1998) and
Horowitz (1985).
2. Note that Collier and Hoeffler (2002) found that countries with an ethnic group
composing between 45 to 90 percent of the population were more likely to experience
civil war onset, although they dismissed these findings later. Also see Collier et al.
(2003) and Collier and Hoeffler (2004).
3. This is from Ron (2005), giving these examples: Mahdavy (1970); Karl (1997);
Ross (1999).
4. See the website of the Correlates of War Project, www.correlatesofwar.org.
5. Fortna (2008, 11) uses Cox proportional hazard models and Weibull models,
which are robust.
6. Outstanding here are the knowledge created around all of the US military’s wars
since Vietnam, UN peacekeeping operations, and NATO’s interventions in Kosovo
and Afghanistan.
7. Discriminatory treatment of ethnic groups that belong to majorities nationally
but minorities within a region (e.g., the maltreatment of Azeris by authorities within
Nagorno-Karabakh within Azerbaijan) would count here, but such treatment was not
found to be a factor in armed conflict recurrences. One case that comes close is the
Marxist FARC’s abuses of Colombian government sympathizers within its autono-
mous zone in 1998–2002, largely leading to a renewal of offensive actions against the
zone; however, this civil war never ended during this period, and thus does not con-
stitute a recurrence. Armed attacks (as in Dagestan by Chechen warlords and in Mali)
would count as recurrences themselves.
8. Obviously, social exclusion occurs in virtually all societies. The threshold here
consists of an identifiable state policy or decision that abrogates the expectation of the
target population.
9. Note, as in the case of Afghanistan in 2001, that taking power by force in a suc-
cessful regime change does not necessarily mean battlefield “victory” and the end of
warfare but may transform insurgents into the state, and the state into insurgents in an
ongoing war.
10. Physical extermination, or politicide or genocide, is of course the most extreme
treatment of one group by another. However, this would come under a different label,
and is not commonly thought of as “exclusion.”
11. The exclusion of women in many societies exemplifies this absence of a correla-
tion or causal relationship between severe exclusion and an armed uprising.
12. With this definition I have tried to be broader than the conventional guaran-
teed posts within the state, yet simpler than Roeder and Rothchild (2005, 30), who
define powersharing institutions as rules that “seek to guarantee . . . inclusive decision-
making, partitioned decisionmaking, predetermined decisions, or some combination
of these.”
What Do We Know about Why Peace Fails? 49

13. Critics like Sriram 2008 and Roeder and Rothchild (2005) also criticize the in-
efficiency, costliness, and perverse policy incentives of powersharing. I consider these
to be less important, partly because alternatives also exhibit such deficiencies.
14. These concepts do not have to compete and may complement one another.
15. Holsti (1996, chap. 5, esp. 84–87) uses this term in a relatively narrow manner,
almost equivalent to regime legitimacy. However, this concept is too narrow to capture
the more embedded and performance elements of state legitimacy that go beyond the
selection of rulers.
2
Is Civil War Recurrence Distinct
from Its Onset?
A Quantitative Analysis and the Limits Thereof
ONE OF THE FIRST QUESTIONS to arise in this inquiry concerned the character
of civil war recurrence. Is the phenomenon of recurrence any different from that
of onset? If not, then it would make more sense not to pursue an inquiry into re-
currence at all but to draw on the more numerous and better-trodden ground of
civil war onset in thinking about recurrence. If recurrence does differ from onset,
then how does it differ? Although recurrence may differ from onset in its mani-
festation, its consequences, or its modes of termination, I am here concerned
with the causes or risk factors associated with recurrence. One can imagine that
societies that have experienced civil war once may be more prone to turn to vio-
lence again, because of the availability of arms, of former militarized networks,
of prior training and capacity among rebel groups, and / or of the socialization
toward war. One can also imagine that war weariness, scars of warfare, social and
individual traumas, the depletion of resources and social capital brought about
by prior warfare, and other aftereffects may make recurrence less likely. Micro-
level violence (Kalyvas 2006) may aggregate or articulate in different ways as
well, with unforeseen consequences.
This chapter seeks to do two things. First, I endeavor to answer this ques-
tion: Do the causes or risk factor of civil war recurrence differ from its onset? I
explore this question through a quantitative exploration of the question drawing
on well-known data sets. I draw on research conducted in collaboration with
the economist John Schmitt (Call and Schmitt 2009), using linear regressions
to determine the extent to which the recurrence of armed conflict differs in its
causes from that of initial onset. If recurrence were not to differ appreciably from
onset, then this inquiry would best be served by analyzing all civil war onsets.
After explaining the pool of cases, I proceed to specification of variables and then
present results.
Second, the chapter demonstrates why one should be deeply skeptical of the
findings of much of the quantitative literature in the absence of other methods
Is Civil War Recurrence Distinct from Its Onset? 51

that can confirm or provide greater causal sustenance to correlation-based the-

ses. The quantitative exercise also demonstrates the severe limits of quantitative
methods generally for explaining the phenomenon of civil war recurrence, not to
mention possible responses to it. This chapter details several of the reasons read-
ers should be extremely cautious, though not disregard entirely, findings about
armed conflict that rely exclusively on quantitative methods. The second section
of this chapter calls into question the weight one should give to poverty and to
primary export dependency as factors in analyzing and responding to civil war
recurrence.

A Regression Analysis of Civil War Recurrence

How does one determine whether civil war recurrence differs from its onset? I
draw on a strong data set to seek to replicate the findings of one serious quanti-
tative study and then to see how that analysis applies specifically to recurrence.
This section presents the pool of cases, data sets, and variables, as well as the
results of the analysis.

The Pool of Cases

Several databases of civil wars exist. One of the earliest and most frequently uti-
lized is the Small and Singer data set, used by Walter (2004), for instance. How-
ever, this data set was created initially for interstate wars, and its criteria for clas-
sifying the initiation and termination of civil wars were not as nuanced as other
data sets. The Small and Singer data also suffer from a large number of missing
values and imputations for important variables over periods of many years.
James Fearon, in his collaboration with David Laitin (Fearon and Laitin
2003), improved and updated the Small and Singer data set, drawing on others.1
In his collaborations with Michael Doyle, Nicholas Sambanis (Doyle and Sam-
banis 2000, 2006) prepared a strong data set with detailed analysis of specific
cases, but this set has the disadvantage for many purposes of not containing data
on countries (and years) without civil wars. The Fearon and Laitin (2003) data
set includes data on countries and years with no civil war, which permits the
introduction of more numerous variables.
Therefore, I here draw on Fearon and Laitin’s data set, which includes
country-years from 1945 to 1999, with a total of 6,610 observations, including
both interstate wars and colonial wars. Their definition of civil war includes
(1) an insurgency that either seeks to shape or take state power or to take control
of a territory; (2) a minimum of 1,000 battle deaths total, with (3) an average of
100 deaths / year; and (4) 100 total deaths on each side of the conflict.2
52 Why Peace Fails: Theory

Variables and Regressions

This section examines all the variables contained in the Fearon and Laitin data
set for their significance in a series of regressions. My colleague and I started
by replicating their own findings. We replicated their findings on civil war on-
set very closely, using their publicly posted data and estimation code (Call and
Schmitt 2009).3
Without elaborating, for the moment, on the substantive significance of the
variables, three variables—per capita gross domestic product (GDP), the size
of population, and whether or not the country became independent in the prior
two years—all carry high statistical significance (p < 0.001).4 Other variables—
the portion of mountainous terrain, the degree of dependence on oil exports,
and whether a country had experienced a prior civil war or prior instability—are
also statistically significant (p < 0.01). As per the central message of Fearon and
Laitin (2003), ethnic fractionalization and religious fractionalization are not sta-
tistically significant determinants of civil war onset.5 This finding echoes that of
Collier and his colleagues (see Collier, Hoeffler, and Söderbom 2006a, 2006b;
Collier and Hoeffler 2004; Collier et al. 2003). We include the side-by-side com-
parisons in columns 1 and 2 of table 2.1.
We then ran the same regressions using a new dependent variable, recur1,
which identified countries experiencing civil war during a given year, after that
country had already experienced a civil war and then experienced at least one
year of peace.6 This and all our regressions drew solely on the publicly avail-
able Fearon and Laitin data set, which shows sixty-nine countries experiencing
ninety-seven instances of civil war onset or recurrence between 1945 and 1999.7
Twenty-one countries experienced civil war recurrence, including six countries
that experienced more than one recurrence (of these, Pakistan had two recur-
rences). Thus a total of twenty-eight civil war recurrences were recorded be-
tween 1945 and 1999.
We found that several of the variables that were statistically significant for
civil war onset were also significant for civil war recurrence. Table 2.1 compares
the effects of the Fearon and Laitin variables on civil war onset and recurrence.
Column 3 presents the marginal effects implied by the logit coefficients for the
variables found in column 2.8 The marginal effects give a more intuitive idea of
the political impact of the variables than the logit coefficients. Being an oil ex-
porter (oil makes up more than one-third of exports), for instance, increases the
chances of civil war onset by almost exactly 1 percentage point, and being a new
state increases it by 3.5 percentage points.
The four variables (or risk factors) shared by civil war onset and recur-
rence include being an oil exporter, having a large population, experiencing re-
cent instability, and the percentage of the territory deemed mountainous. All
Is Civil War Recurrence Distinct from Its Onset? 53

Table 2.1 Comparing Civil War Onset and Recurrence: A First Cut
Variable or Parameter (1) (2) (3) (4)b
Data sourcea F&L 2003 F&L 2003 by C&S 2009 C&S 2009
C&S 2009
Estimation logit logit logit, mfx logit, mfx
Dependent variable onset onset onset recur1
Mean of dependent 0.017 0.017 0.017 0.040
variable
Prior war (0 / 1) –0.954** –0.954** –0.00587** 0.14375**
GDP per capita (lag) –0.377** –0.344** –0.00288** –0.00028
Log of population (lag) 0.263** 0.263** 0.00220** 0.00254**
% mountainous 0.219** 0.219** 0.00183* 0.00168**
Noncontiguous (0 / 1) 0.443 0.443 0.00430 –0.00125
Oil exporter (0 / 1) 0.858** 0.858** 0.01009* 0.01281**
New state (0 / 1) 1.709** 1.709** 0.03498* —
Instability (0 / 1) 0.618** 0.618** 0.00650* 0.00624**
Democracy (Polity IV) 0.021 0.021 0.00017 –0.00008
Ethnic frac. (ELF) 0.166 0.166 0.00139 –0.00635**
Relig. frac. (RLF) 0.285 0.285 0.00238 0.01225**
Constant –6.731** –6.731** — —
Sample size 6,327 6,327 6,327 6,165
Statistical significance: # at 10% level; * at 5% level; ** at 1% level.
a
F&L = Fearon and Laitin (2003); C&S = Call and Schmitt (2009).
b
New state predicts failure perfectly.

these are correlated with war recurrence in the same direction as onset, and
all have a roughly comparable impact.9 Thus oil exporters have on average a
1-percentage-point greater chance of experiencing civil war than non–oil ex-
porters, but a 1.2-percentage-points greater chance of experiencing recurrence.
Similarly, a doubling of the population (a 1-unit increase in the log of the popula-
tion) increases the chances for civil war onset by 0.22 percentage point and for
recurrence by 0.25 percent.10
Three other variables point to important differences between recurrence and
civil war onset. At the same time, ethnic fractionalization and religious fraction-
alization are statistically significant for civil war recurrence but not for onset. The
absence of significance of these variables was a key factor in the central conclu-
sion of Fearon and Laitin (and Collier and colleagues) that ethnic and religious
grievance did not underlie civil war onset. Instead, they posited that civil wars
reflected opportunity for insurgency provided by poor or weak states, as signified
by the confluence of poverty (GDP per capita) and primary commodity depen-
dence (proxied by the oil exporter variable).
One of the most important substantive findings of this chapter is that per cap-
ita GDP (lagged) is not a significant predictor of civil war recurrence. Although
54 Why Peace Fails: Theory

we confirmed Collier and others’ (2003) and Fearon and Laitin’s findings that
GDP per capita (lagged) is negatively correlated with onset, the coefficient loses
its statistical significance for recurrence. This finding—that poverty makes no
difference in a country’s risk of civil war recurrence—cuts against much of
the current thinking and policy emphasis on the economic conditions of poor
people in the wake of war.11
Ethnic fractionalization also seems to cut against long-standing theoretical
claims about civil war risks. The coefficient is negative, which suggests that an in-
crease in ethnic fractionalization (the chance that two random persons are from
different ethnic groups) by half a unit (say, from 0.25 to 0.75, for a variable rang-
ing from 0.001 to 0.925) lowers the chance of civil war recurrence by 0.3 percent-
age point. Although 0.3 percentage point may sound small, this effect is sizable
compared with other effects considered to be large in the literature on civil war
onset, such as being an oil exporter, which raises the chance of civil war onset by
about 1 percentage point. The average probability of civil war recurrence in the
sample is only 4 percentage points.
Religious fractionalization (RLF) is also highly statistically significant for re-
currence (p < 0.001). RLF (which runs from 0 to 0.783 in the sample) behaves
as theoretically expected, damaging chances for postwar peace. An identical half-
unit increase in the degree of religious fractionalization raises the chances of civil
war recurrence by about 0.6 percentage point. Its impact on recurrence is thus
more than half that of oil dependence and twice that of a doubling of GDP on
Fearon and Laitin’s onset variable. Recall that these last two marginal effects have
sent the field into a tizzy in the past several years.
What import do these findings carry at first glance? Most important, they
suggest that civil war recurrence is worth studying. Its causes seem to be different
from those of onset. These differences are not simply a matter of degree but also
reintroduce into one’s understanding of civil wars variables that have been dis-
missed in recent research on civil war onset. The factors that pointed to poverty
as a central risk factor of civil wars are not, in fact, crucial for whether countries
emerging from war risk relapse into armed conflict. Specifically, our research sug-
gests that the social composition or character of society is an important factor in
whether peace can be consolidated. Our initial findings confirm that some cases
are “hard” for consolidating postwar peace, specifically where religious fraction-
alization is high and where ethnic homogeneity prevails.

Refining the Analysis of War Recurrence

Having discovered that the portions of ethnic and religious fractionalization in
a society have unexpected impact on the chances of war recurrence, we exam-
ined whether the linear assumption of the relationship was sound. By adding
Is Civil War Recurrence Distinct from Its Onset? 55

Table 2.2 Civil War Recurrence, Adding Ethnic Fractionalization Squared

Recur1 dy / dx Std. Err. z P > |z| [ 95% C.I. ] X
warl* .1200274 .0158 7.60 0.000 .089065 .150989 .136577
gdpenl –.0006387 .00023 –2.84 0.005 –.00108 –.000198 3.66926
lpopl1 .0025729 .00047 5.48 0.000 .001653 .003493 9.08865
lmtnest .0009761 .00039 2.48 0.013 .000204 .001748 2.18254
ncontig* .0000593 .00106 0.06 0.955 –.002017 .002135 .179562
Oil* .0109419 .00347 3.15 0.002 .004134 .01775 .127494
instab* .0065471 .00191 3.42 0.001 .0028 .010295 .149392
polity2l –.0000444 .00007 –0.62 0.537 –.000185 .000097 –.40438
ethfrac .0239523 .007 3.42 0.001 .010232 .037672 .385665
ethfrac2 –.0346829 .00854 –4.06 0.000 –.051422 –.017944 .230602
relfrac .0126473 .00278 4.55 0.000 .007204 .018091 .364796
Note: The variable nwstate was dropped because it predicts failure perfectly, and 162 observations were
not used.

the square of ethnic fractionalization, we concluded that the relationship indeed

seems nonlinear. Table 2.2 shows the marginal effects using war recurrence (re-
cur1 again) as the dependent variable. Ethnic fractionalization remains highly
significant (p < 0.001), but its overall impact (taking into account both its linear
and squared impact) is concave. For countries at the middle and lower ends of
the scale of ethnic fractionalization, an increase in fractionalization augments the
chance of war recurrence. But beyond a certain degree of heterogeneity (0.69
ethnic fractz), greater fractionalization bodes well for peace, with a reduced
chance of war recurrence.
Even including the square of ethnic fractionalization, religious fractionaliza-
tion remains highly significant, and the same half-unit increase in fractionaliza-
tion increases the chance of war recurrence by roughly 0.6 percentage point. Re-
call that being an oil exporter only increases the chances of onset or recurrence
by about 1 percentage point, and a doubling of GDP per capita only has about
a 0.3-percentage-point impact on the onset or recurrence. The distinction be-
tween the factors accounting for civil war onset versus recurrence remains; even
when the squared ethnic fractionalization variable is added to Fearon and Lai-
tin’s Model 1, neither ethnic nor religious fractionalization is statistically signifi-
cant for onset. Note, however, that when the square of ethnic fractionalization is
introduced, GDP per capita becomes statistically significant, albeit with a very
small substantive impact (a doubling of national GDP decreases the likelihood
of recurrence by only 0.06 percent).
We also tested Fearon and Laitin’s Model 1, substituting various other mea-
sures of ethnic and religious fractionalization. These included using Fearon
and Laitin’s plural religion, the percentage of the largest ethnic group, as well as
56 Why Peace Fails: Theory

Table 2.3 How Robust Is Ethnic Fractionalization?

Variable or
Parameter (A) (B) (C) (D)
Estimation logit, mfx logit, mfx logit, mfx logit, mfx
Dependent variable recur1 recur2 recur3 recur4
Mean of dependent 0.04 0.018 0.036 0.014
variable
ethfrac 0.02395** 0.01434** 0.01617** 0.00679*
ethfrac2 –0.03468** –0.01978** –0.02309** –0.00914*
relfrac 0.01265** –0.00013 0.01007** –0.00127#
Statistical significance: # at 10% level; * at 5% level; ** at 1% level.
Sources: Call and Schmitt (2009), using Fearon and Laitin (2003).

combinations of squared ethnic variables with unsquared religious variables, and

vice versa. We found that religious fractionalization was statistically significant in
some models and not in others, but always in the same direction.
To explore the robustness of these correlations, we also ran the regressions
using three alternative dependent variables, as shown in table 2.3. Our second
dependent variable (recur2) identified cases where civil war recurred after one
year of peace but within five years of the end of a prior war. A third dependent
variable (recur3) scored whether war recurred at any point before 2000 after two
consecutive years of peace immediately following a civil war. And a final depen-
dent variable (recur4) scored cases where war recurred within five years after two
consecutive years of peace after a prior war.
Table 2.3 demonstrates that ethnic fractionalization and its square remain sta-
tistically significant with the same sign (though not identical form or comparable
impact) for all four definitions of civil war recurrence. Religious fractionalization
loses much of its statistical significance when civil war recurrence is limited to
within the first five years of the end of armed conflict (recur2 and recur4). Note
that the differences across the four dependent variables reflect not just different
definitions but also a different sample of countries and years that meet these
definitions.

Reflections on Ethnicity and War Recurrence

All these tests suggest that ethnic fractionalization has a complex but statistically
significant correlation with civil war recurrence, and that religious fractionaliza-
tion also works against consolidating peace. In particular, the better fit for the
nonlinear specification of ethnic fractionalization (and considering the specific
coefficients on the two variables) suggests that ethnic fractionalization raises the
chances of recurrence over a substantial range but that the effect diminished at
high levels of ethnic fractionalization. Thus, in keeping with Fearon and Laitin’s
Is Civil War Recurrence Distinct from Its Onset? 57

findings for onset, societies with many small ethnic groups are not at the apex of
risk for civil war recurrence either. Of course these theories of ethnic diversity are
replete with dated assumptions that ignore transnational influences on identity,
even if empiricists test for a regional risk of conflict generally. Moreover, they are
based on time-invariant and often decades-old measures of ethnic composition
in a given country.
This revised analysis also suggests that large population, oil exporters, and
instability all retain some influence on war recurrence, with comparable influ-
ence as with onset. One important exception is the impact of per capita GDP,
which has only one-fourth the impact on recurrence as it has on onset. Terrain
also loses half its impact relative to onset. In general we find that the same eco-
nomic and geographic variables remain significant in explaining war recurrence,
although some have less impact.
To repeat, the most important finding of the quantitative exercises presented
in this chapter is that ethnic and religious fractionalization have an important
and complex impact on civil war recurrence that they do not have for onset. Thus
it appears worthwhile to study civil war recurrence as a phenomenon in its own
right, especially examining the role of ethnic and religious fractionalization that
should have a lesser impact at the extremes of its values (i.e., very much or very
little fractionalized).
Three considerations—conceptual, theoretical, and empirical—call for
caution in relying too heavily on the findings given here concerning the role of
ethnicity in civil war recurrence. First, at the conceptual level, as with other vari-
ables identified by quantitative research, ethnic fractionalization and religious
fractionalization remain relatively static variables. Fractionalization and diversity
measures of ethnicity fail to identify which ethnic groups matter (Chandra and
Wilkinson 2008) and do not capture the relationship between ethnicity and the
state. The generally static and thus unstrategic conception of these variables rep-
resents one of the biggest obstacles to reliance on this application of Fearon and
Laitin’s data set to the study of war recurrence.
Second, there are important theoretical gaps in relying on quantitative correla-
tions from the quantitative analysis given above. Even if the ethnic structure of a
society constitutes a risk factor, this relationship tells us little about how and why
the ethnic composition of society influences the chances of converting social
intercourse into armed conflict. Statistical analyses in and of themselves offer no
theoretical rationale for the correlations identified. One theoretical rationale for
the causal role of ethnic fractionalization (up to a certain point, but thereafter)
and religious fractionalization playing a role is one of opportunity for political
manipulation. That is, this structure indicates something of the opportunities
that political leaders have to mobilize social divisions and conflict toward politi-
cal ends—possibly toward violent means. This account comports with political
entrepreneurial accounts of ethnic conflict (Gagnon 1994).
58 Why Peace Fails: Theory

A second potential theoretical rationale is that ethnic fractionalization signi-

fies the degree of insecurity perceived by particular ethnic groups. This account
comports with the internal security dilemma (Walter and Snyder 1999) and with
commitment problems (Fearon 1998) that different groups experience once ei-
ther the state or other security guarantees are lifted. Yet the quantitative data
do not come close to being sufficient to confirm or disconfirm these theoretical
rationales, nor would it be easy to devise statistical analyses that would discern
among these competing theoretical hypotheses about the role of ethnicity and a
fractionalized ethnic or religious structure in a society.
Third, and finally, recent empirical work calls into question the reliance on
ethnic or religious fractionalization in analyzing armed conflicts. Wimmer, Ce-
derman, and Min (2009) offer a quantitative test of the strategic role of ethnic
identity that merits special attention here. They create a data set that measures
the extent to which a population is excluded from senior executive state offices
(e.g., cabinet-level posts and senior army and other top posts) for countries from
1946 to 2005. They claim (Wimmer, Cederman, and Min 2009, 329) that their
quantitative study “is the first time that the ethnic exclusion argument has been
statistically confirmed based on a global data set that measures exclusion directly
and at the polity level.” They find that ethnic exclusion is statistically significant
in accounting for the onset of civil wars. Their argument finds that ethnic exclu-
sion is especially robust in explaining what they define as ethnic conflicts. In such
cases an increase of 1 standard deviation—from 6 to 32 percent, for instance—
of the excluded population increases the chances of ethnic conflict by 25 percent,
slightly higher than the 22 percent increase for a comparable increase in GDP
per capita and much greater than the impact of population size or mountainous
terrain (Wimmer, Cederman, and Min 2009, 331).
Moreover, once exclusion is tested, ethnic diversity and fractionalization dis-
appear as significant variables in explaining the onset of civil war (defined with
by 1,000 battle-related deaths). As Wimmer, Cederman, and Min report (2009,
329), “Rather than diversity as such, it is political exclusion along ethnic lines
that breeds ethnic conflict.” In essence, it is not the ethnic demographic structure
of society that matters but what political or individual actors do with this differ-
ence, how they emphasize and reify it, or how they downplay and diminish it.
Wimmer, Cederman, and Min (2009, 329) also find that “ethnic exclusion is
as consistently related to conflict as is GDP per capita,” which undercuts claims
that economic motives are preeminent in civil war onset. They point out that po-
litical exclusion cuts across the greed / grievance debate, because unequal access
to state office translates into unequal access to political, social, and economic
goods (p. 324). Their findings also reinforce the weak state approach, because
they find that conflict is more likely “in incoherent states” where central author-
ity has historically been absent or minimal (p. 334). Their study suggests dif-
Is Civil War Recurrence Distinct from Its Onset? 59

ferent causal pathways whereby ethnicity might affect rebellions and infighting
within powersharing arrangements, including those where external events (like
the US civil rights movement for Catholics in Northern Ireland, or the onset of
the North American Free Trade Agreement for the Zapatistas) trigger excluded
populations to mobilize or ethnic groups in power to seek to expand that power.
The study by Wimmer, Cederman, and Min (2009) provides a foundation
for understanding the causal relations examined in the case studies that follow.
For two reasons, however, I have not drawn on their data set or definition of
exclusion. First, their concept restricts exclusion to ethnic identity. Instead, dif-
ferent types of exclusion may be relevant in ways that are not captured in their
data set’s single set of indicators or by other universal measures of exclusion. The
proclaimed basis of exclusion (e.g., religious or ethnolinguistic) or the institu-
tional configuration may vary across societies. Second, the perception of exclu-
sion is pertinent and rests on prior expectations. Although quantitative meth-
ods do not permit such a contextualized understanding, qualitative methods do.
Perceptions provide a more detailed understanding of causal pathways. For this
reason, the concept of exclusion in this book is contextually specific, rooted in
expectations by elites of the main sociopolitical groups.
Before turning to the qualitative analysis, however, it is worth reflecting on
how much one can infer from statistical methods in general. There is good rea-
son to question the reliability and robustness of quantitative findings, both in the
analysis presented above and in research more broadly. Despite the many posi-
tive contributions of such methods, a number of systematic problems tend to be
overlooked. The next section examines both these positive and negative aspects.

The Contributions and Limitations of Quantitative

Methods for Studying Civil Wars
This section assesses the positive contributions and especially several often-
overlooked limitations of quantitative methods.12 Quantitative methods carry
some obvious advantages in the study of civil wars and peace. They permit the
testing of theoretical claims across the pool of cases of civil wars. Linear regres-
sion analysis offers the chance to control multiple variables, something that is
necessary where no single factor can plausibly account for the occurrence of civil
wars. Indeed, the acknowledged multiplicity of factors that shape civil war on-
set can best be sorted via statistical methods. As seen above, despite divergent
findings in much of the literature, some degree of consensus about variables like
poverty, terrain, and high oil exports has emerged.
At the same time, several shortcomings of quantitative methods are often
overlooked. Drawing policy conclusions from research that elides these con-
straints may produce misguided or harmful policies. Some of our findings call
60 Why Peace Fails: Theory

into question the reliability and perhaps utility of quantitative analysis based
on numerous observations. The questions raised here go beyond long-standing
general objections to the application of statistical inference in a scientific frame-
work to social phenomena.

Too Few Observations

First, the number of observations is small. The number of civil wars in the period
covered by this data set (1945–99) reaches ninety-seven, but one should proceed
with great caution in interpreting results inferred at standard levels of statistical
significance from a total number of recurrences, which is twenty-eight.13 Few
people would take an experimental drug that had been tested on only twenty-
eight people.

Unit Heterogeneity
The small sample size also effectively imposes the uncomfortable requirement
that researchers assume that the underlying data generation process is identical
across all cases of civil war recurrence, which is a strong assumption. To con-
tinue the medical drug analogy, we know that some drugs can affect men and
women, children and adults, or other groups differently, but the small size of our
experiment forces us to assume that the effects are the same across all countries,
religious groups, ethnic groups, and other groups.
How comparable are Russia’s two civil wars (in Chechnya) with Nicaragua’s?
Or Indonesia’s conflicts over Aceh with Rwanda’s society-sweeping war–cum–
genocide? Or Argentina’s insurgency against the military dictatorship with Chi-
na’s? The idea is that control variables are designed precisely to control for dif-
ferences, but these controls may not be effective in complex social environments.
Because regression analysis requires setting aside this limitation and trusting the
controls, I elect to elide this objection. However, it is useful to keep in mind the
potential limits of control variables in comparing such disparate units of analysis.

Arbitrary or Questionable Estimation

Slight changes in the data used to proxy for different political concepts or in the
particular specifications (including or excluding particular variables, linear vs.
nonlinear terms, linear vs. nonlinear estimation techniques) result in substan-
tially different outcomes (e.g., see the broadly differing results across research
produced by a range of leading researchers in the field, using different data sets,
variables, time periods, and estimation techniques). The alteration of either a
Is Civil War Recurrence Distinct from Its Onset? 61

single variable or the exclusion of a few country cases changes our result in im-
portant ways. A few examples suffice. First, in our regressions on war recurrence,
adding a single variable (viz., the square of ethnic fractionalization) restored the
statistical significance of per capita GDP for recurrence.
Second, perfectly reasonable alternative definitions of some variables lead
to divergent outcomes. Whereas ethnic fractionalization is negatively associ-
ated with war recurrence, introducing its square points to a largely positive, but
concave, relationship. Another example from modeling ethnic diversity under-
scores this point. When we used various definitions of ethnic fragmentation, the
variable proved significant in virtually all versions. However, the sign of the eth-
nic fractionalization variable was not consistent across all specifications! Thus
even with statistical significance, the fundamental causal direction was wholly
unreliable.
We have no a priori reason to presume that one model of ethnicity (e.g., eth-
nic fractionalization squared) or religious structure is better than another. Ana-
lysts have a tendency to conclude that a better fit means a better explanatory
variable. Consequently, concluding that ethnic fragmentation is linearly related
to war recurrence or concluding it is a concave function is somewhat arbitrary,
which relies heavily on the particularities of the variables used to proxy for ethnic
fractionalization, the countries and years in the sample, and the estimation tech-
niques used. Drawing either scholarly or policy implications from such choices
seems ill advised.
A third, simple example comes from the work of Paul Collier and colleagues.
Since 2002, he and various colleagues have produced several publications citing
the rate that armed conflicts recur within five years, statistics that have found
their way into pronouncements by the UN secretary-general. Yet between 2002
and 2006, they reported remarkably divergent rates of armed conflict recurrence,
including 50 percent, 44 percent, 23 percent, and 21 percent (the latter two refer
to recurrence within four years).14 That the same person reports such divergent
figures over a span of four years is another argument for caution against drawing
strong conclusions from statistically based exercises under these conditions.

Poor Proxies
Often, the connection between proxies and the things they purport to measure
is thin or questionable. As Gleditsch and Ruggeri (2010, 299) aver, “Many of the
indicators used in empirical studies of civil war are relatively crude indicators of
the underlying concepts and only loosely related to the theoretical rationale.”
Fearon and Laitin, for example, use per capita GDP and oil export dependence
as proxies for state weakness, whereas Collier and colleagues see these same
62 Why Peace Fails: Theory

variables as proxies for limited economic opportunities. Walter (2004) uses in-
fant mortality rate as a proxy for economic opportunities, and Doyle and Sam-
banis (2000) use electricity consumption as a proxy for local capacity.
Creative use of the available cross-national data is admirable and necessary.
However, scholars should be much more modest in declaring that these imper-
fect proxies confirm or reject the underlying theoretical relationships that the
proxies are meant to represent. For example, infant mortality responds much
more to trends in health provisions in a society, reflecting factors unrelated to
the possibilities of employment. States and international organizations run risks
when they base policies on research that rests upon poorly measured or only
loosely connected proxies.

Omitted Variable Bias

In addition, the things one can measure do not adequately get at the things that
theorists say are important. Put another way, the supply of variables does not
meet the demand of theoretical explanations for war onset or recurrence. As dis-
cussed above, theorists posit myriad explanatory variables for civil war onset.
We have no cross-national direct measures for several of these, including state
institutional capacity, powersharing arrangements, manipulative ethnic leaders,
limited employment opportunities for youth, exploitation of one social group by
another, social exclusion, personal insecurity, and group insecurity. Most impor-
tant for this study, “legitimacy” (which is notoriously hard to conceptualize) and
“inclusion” cannot be readily measured, and certainly not when actors’ expecta-
tions are taken into account. Proxies are necessarily used, but more often than
not researchers proceed without including many of these arguably important
variables in their statistical analyses. This omitted variable bias points to the need
to use qualitative methods—and thus the case studies in this book—to examine
the role of hypothesized variables (e.g., capacity building, relevance to Western
security interests, and legitimacy) in postconflict societies.

Policy Applicability
Related to the prior problem are challenges of policy relevance and utility. The
variables for which we have data and that have the most significant impact are
ones over which policymakers typically have little control, especially in the short
run. A handful of variables repeatedly appear in statistical analyses of civil war
onset and recurrence: the level of GDP, infant mortality rates, size of population,
primary commodity dependence, and ethnic or social fractionalization.
At best, these correlations can guide policymakers in identifying and ad-
dressing the long-term risk factors for armed conflict. The most direct logical
Is Civil War Recurrence Distinct from Its Onset? 63

implication of the research findings given above is that peace can be advanced
by reducing poverty, infant mortality rates, and diversified economies. And yet
does anyone really think that reducing infant mortality rates in a specific coun-
try is the ticket to peace, or that national or international actors can reduce in-
fant mortality quickly enough to shape the dynamics of a recent, impending,
or ongoing civil war? Such variables may provide guidance on detecting where
armed conflict may be a few percentage points more likely, but they are not help-
ful for developing policy responses to existing or recently ended conflicts. Some
of these variables—like population, terrain, and the national demographic pro-
file—cannot be effectively manipulated by authorities unless those authorities
engage in mass human rights violations.
In addition, specific steps to reduce these risk factors for civil war onset and
recurrence may themselves cause armed conflict more acutely than the marginal
effect of the risk factor. For example, Collier’s advocacy for adherence to the
World Bank’s neoliberal macroeconomic prescriptions as the best route to devel-
opment may in fact deepen other proximate causes of warfare, because the dislo-
cations of privatization or reducing the state payroll may foster instability. This
is precisely the finding of Hartzell and Hoddie (2010), who show that adoption
of the International Monetary Fund’s structural adjustment programs between
1970 and 1999 made countries more likely to experience civil war onset. As an-
other example, diversifying the economy quickly through expanded agroexports
may leave thousands of peasants landless and may prompt armed uprisings. Even
if one accepts the risk factors identified in statistical analyses of civil wars, these
tools provide little undisputed guidance for nuanced, midlevel policymaking,
much less for country-specific remedies.
Conversely, the variables that policymakers care most about are often diffi-
cult to put into quantitative models. Some scholars have in recent years made
headway in testing quantitative models of international peacebuilding efforts, or
actions designed to reduce war recurrence. Collier and colleagues, for example,
have drawn conclusions about the impact of international aid, the timing of this
aid, and the particular type of aid that helps prevent conflict recurrence.15
Michael Doyle and Nicholas Sambanis have performed the most significant
work in this area, based on Sambanis’s civil war data set. They code UN-authorized
international missions based on their degree of authority and mandates. Namely,
they distinguish among (1) monitoring, (2) consent-based traditional peace-
keeping, (3) multidimensional peacekeeping, and (4) peace enforcement (Doyle
and Sambanis 2000, 781). They examine the impact on peacebuilding success as
measured by both a return to violence and a minimal level of participation or de-
mocracy. Their salient findings regarding international agency are that (1) treaties
have a positive impact; (2) any sort of UN involvement has a positive but limited
impact compared with treaties, with which it is highly correlated; and (3) only
64 Why Peace Fails: Theory

multidimensional peacekeeping operations have a statistically significant impact

(positive) on consolidating peace.
More recently, Sambanis (2008) has built on this earlier work with more nu-
anced definitions of peacebuilding success that have involved sustainability after
the departure of a UN peacekeeping operation. He also devises ways to miti-
gate the impact of potential selection bias about where the UN intervenes. His
thoughtful treatment of variables and coding have led to interesting conclusions
that few troops are better for peacebuilding success and that UN missions have
a diminished impact after a few years.16 He has also drawn similar conclusions
about multidimensional peacekeeping. Most important here, he finds that “high
fractionalization has a significant (p = 0.007) negative effect on peace duration”
(Sambanis 2008, 19). Doyle and Sambanis (2006) had found this result, which
coincides with my own findings, earlier as well, but they did not emphasize its
significance. Although they reported testing the square of ethnic fractionaliza-
tion for onset (without finding significance), they did not indicate having tested
it on recurrence.17
Sambanis’s analysis devotes welcome attention to factors beyond war recur-
rence that are normatively attractive and more attuned to policy concepts and
interests. Even here, however, the policy utility remains limited, for two main
reasons. First, if scholars cannot find agreement on the factors that account for
simple war recurrence, then one should have less confidence regarding more
ambitious definitions of peace (including, e.g., participation, civil rights, and de-
mocracy) upon which there is less agreement on measurement.
Second, even these admirable efforts by Sambanis, Doyle, and Fortna (2004a,
2004b, 2008) offer only blunt answers to basic questions: Are peace treaties
good? Yes. Does peacekeeping help? Yes. Should states invest in peacekeeping?
Yes. Should peacekeeping entail multidimensional operations that involve insti-
tution building? Yes. Are demobilization, disarmament, and reintegration pro-
grams worthwhile? Maybe.
Statistical analysis of other policy instruments has added to our knowledge.
Despite the work questioning powersharing (Roeder and Rothchild 2005;
Sriram 2008), an increasing number of studies suggest that multiple forms of
powersharing, and powersharing by security institutions, increase the chances
for sustained peace (Hartzell and Hoddie 2003, 2007; Hoddie and Hartzell
2005). Gurses, Rost, and McLeod (2008) study the effect of several variables,
including types of mediation interventions, on the duration of peace after civil
wars. They find that “ethnic and bloody wars” have shorter periods of peace,
whereas democracies and large armies favor longer periods of peace (p. 150).
Somewhat confusingly, they find that mediation attempts favor more durable
peace but that mediation agreements themselves (along with superpower media-
tion) undermine stable peace. They admit that their study is unable to capture
Is Civil War Recurrence Distinct from Its Onset? 65

the quality of the peace agreement, which may be the key to success, and thus
they call for case studies (p. 151).
These research findings show the limits of statistical methods for studying
civil war recurrence. These methods have great difficulty providing clear answers
to questions such as these: What elements of a political settlement are most im-
portant? How can external actors foster elections that do not themselves spark
renewed violence? How can exclusionary power arrangements be transformed?
What state institutional reforms are most important? How are their capacity and
legitimacy enhanced? How can countries that are dependent on primary com-
modities reduce their risk of war and its recurrence without provoking conflict?
How important is justice for past atrocities? Although scholars can create prox-
ies for some of the variables relevant to these questions, they are the exception.
Where such variables do exist, the nexus is distant and the cross-national data
are inadequate and usually incomparable. Even Sambanis’s efforts, which go
well beyond most others’, do not go far enough to address these kinds of policy
concerns.

Conclusion
This chapter serves two purposes. First, it determines, based on one of the best
and most appropriate data sets available, whether civil war recurrence differs
from civil war onset. The answer is a muted yes. Onset and recurrence share a
number of risk factors—including having a large population, relying heavily on
oil exports, and political instability—and these factors have a comparable influ-
ence on recurrence as with onset.
However, in contrast to their role in civil war onset, ethnic fractionalization
and its square both appear to represent statistically significant risk factors for civil
war recurrence, as does religious fractionalization. This difference suggests that
these two identity issues are conceived, refracted, or manipulated (perhaps in
conjunction with fears, interests, or desires) so as to enhance the chances of re-
currence once a civil war has been dormant for a short time (one or two years). It
is also important for this book’s claims about grievance as the most salient causal
factor for civil war recurrence, rather than economic factors like poverty, that
GDP per capita is not statistically significant as a risk factor in the application of
linear regression analysis to the Fearon and Laitin data set. However, once ethnic
fractionalization is squared, GDP per capita reappears as a statistically significant
risk factor, but only with one-fourth its earlier influence.
Second, this chapter shows that scholars and policymakers alike should
regard the findings of quantitative research on civil wars with a degree of cau-
tion. Even when used with care and sophistication, they suffer from numerous
flaws, most of which cannot be overcome without an unrealistic investment of
66 Why Peace Fails: Theory

resources or time. Slight differences in which variable is selected to stand for a

concept, or in how such a variable is measured, can produce startlingly divergent
findings. This variation is true precisely for ethnic and religious fractionalization
or diversity. This disparate relationship also holds for the impact of per capita
GDP on the recurrence of civil wars, which points toward other variables to ex-
plain this phenomenon.
Although there is a greater consensus among researchers about risk factors
such as poverty, oil export reliance, and geography, the utility and implications of
these findings are multiple and contested. The implications for policy prescrip-
tions of these more widely agreed causal relationships are not clear. Adopting
what may seem the logical response to a given risk factor may produce the oppo-
site reaction in postwar societies. These methods certainly offer generalizations
that qualitative research cannot provide. However, their meaning in the complex
world of social phenomena requires both humility and a greater pairing with
other methods that may verify or alter the purported causal relations that have
been identified. I now turn to such qualitative methods, beginning with a case
study of Liberia’s civil war recurrence.

Notes
1. Their complete list is given by Fearon and Laitin (2003, n3).
2. Fearon and Laitin (2003, 5); for additional coding notes, see Fearon and Laitin
(2003, n4).
3. The findings differed only by a thousandth of a point and are available on re-
quest. The table replicated is labeled table 1 in Fearon and Laitin (2003, 84).
4. In replicating the work of Fearon and Laitin, we too use the log of lagged popula-
tion, the log of portion of mountainous terrain, and lagged per capita GDP.
5. They measure ethnic fractionalization (and separately and analogously, religious
fractionalization) using a composite of ethnolinguistic fractionalization (based on the
1964 Soviet Atlas’s measurement of the probability that any two people in a country
are from different ethnolinguistic groups), the share of the largest ethnic group, and the
number of distinct languages spoken by groups exceeding 1 percent of the population.
See Fearon and Laitin (2003, 78).
6. In this Call and Schmitt (2009) diverge slightly from Fearon and Laitin. Our def-
inition of recurrence codes every year of the recurred civil war as a recurrence; Fearon
and Laitin, by contrast, focus on onset and code only the first year of a new civil war as
“one.” We find their coding strange, because it codes identically both every country in
its second (or third) consecutive year of civil war (e.g., Guatemala in 1983) and every
country at peace (e.g., Sweden 1997). We believe that countries at war should be coded
as experiencing recurrence rather than no warfare. However, we also ran a regression
for onset (available from the authors), coding all country-years that civil war occurred
as “one.” We found that neither was statistically significant at the 5 percent level or
Is Civil War Recurrence Distinct from Its Onset? 67

better, though RLF was statistically significant in some cases at the 10 percent level,
reinforcing our earlier findings on the differences between onset and recurrence.
7. This tally diverges from the 73 countries and 127 civil wars cited by Fearon and
Laitin (2003, 75), who counted multiple wars. Because our unit of analysis is country
(not the war, as Sambanis), we did not distinguish among wars in a given country.
8. Marginal effects show what percentage change is associated with a 1 unit change
in continuous variables, and with a change from 0 / 1 in dichotomous variables.
9. Oil exporters and large countries are slightly more likely (by 20 percent and
14 percent, respectively) to experience recurrence than onset. Terrain and instability
are slightly (by 4–8 percent) less important for recurrence than onset.
10. “New states” drops from the regression because it perfectly predicts war recur-
rence. Following Fearon and Laitin, the regression also includes a lagged dependent
variable.
11. Of course the level of GDP per capita says nothing about the national income
distribution.
12. For a more extensive treatment of some of these issues, see Human Security
Report Project, Human Security Report 2009 / 2010, chapter 2, “Peace, War and Num-
bers” (Human Security Report Project 2009), which appeared after Call and Schmitt
(2009).
13. As mentioned above, Fearon and Laitin state their data show 127 civil war on-
sets, including recurrent ones. Our examination of their data yields 97.
14. See, in the order corresponding to the figures cited, Collier and Hoeffler (2002);
Collier et al. (2003); and Collier, Hoeffler, and Söderbom (2006a, 2006b). Also see
Suhrke and Samset (2007, 198).
15. See most recently, Collier (2007, 104–7) and Collier, Hoeffler, and Söderbom
(2006b).
16. Again, we did not use Sambanis’s data set because it captures only civil wars, ex-
cluding data on countries that did not experience civil war, which substantially reduces
the number of observations.
17. Doyle and Sambanis (2006, 102n78).
This page intentionally left blank
Part II
Examining the Cases
This page intentionally left blank
3
Liberia
Exclusion and Civil War Recurrence

IN AUGUST 1997, after winning by a landslide 75 percent in a relatively free and

fair election, Charles Taylor was inaugurated as Liberia’s president. His election
marked the culmination of a peace process ending a seven-year war that had
killed roughly 150,000, displaced 40 percent of the population, and sparked a
deadly war in Sierra Leone and instability in Côte d’Ivoire and Guinea.1 The
peace process, in keeping with the 1995 Abuja Accord, involved agreement
among the six factions and the government to disarm and demobilize the rebels,
to form a transitional government that included the main warlords, to accept ver-
ification by peacekeeping forces of the Economic Community of West African
States (ECOWAS) and the United Nations, and to hold internationally super-
vised national elections. Three months after Taylor’s election, the UN mission
declared its mandate fulfilled and departed. ECOWAS troops returned home a
year later (Adebajo 2002b). In his final report on the mission to Liberia, the UN
secretary-general celebrated “the successful peace process” and the ability of “the
international community to assist in bringing peace to Liberia” (United Nations
Secretary-General 1997a).
Yet in April 1999, rebels based in Guinea launched attacks on the border town
of Voinjama in northwest Liberia’s Lofa County (Ali-Dinar 1999).2 Further at-
tacks occurred later that year, marking the onset of what would become Liberia’s
second civil war (Pham 2004, 180). The main rebel groups in this war would be
Liberians United for Reconciliation and Democracy (LURD), consisting mainly
of former members of the rebel group ULIMO-K, and subsequently the Move-
ment for Democracy in Liberia (MODEL).3 After accelerating, the war would
eventually cost some 2,500 lives and displace tens of thousands.4 In 2003 inter-
national pressure on then–president Taylor forced his flight to Nigeria, and a
negotiated settlement between the rebel groups and the state ended the war. As
of this writing, the 2003 peace process has held without renewed warfare.
What explains this clear failure of a celebrated peace? The case of Liberia
offers a test for prevalent theories about civil war recurrence and the factors in
72 Examining the Cases

successful postwar peacebuilding. The case seems to fit a number of competing

hypotheses: poverty, inequality, natural resource dependency, bad governance,
bad neighborhoods, international malfeasance, and weak state institutions.
It is an especially likely case for theories associated with Paul Collier and his
colleagues that emphasizes the role of economic opportunities and natural re-
sources in explaining the onset, duration, and recurrence of civil wars (Collier
and Hoeffler 2004).
In this chapter I argue instead that Liberia’s war recurrence shows the rel-
evance of grievance-based arguments (Englebert 2000; Herbst 2000; Reno
1999). The Liberian case illustrates not so much the failure or success of an in-
stitutional arrangement like powersharing as the consequences of deliberate de-
fiance of sharing power once a government has been elected. Specifically, I argue
that exclusionary behavior by Taylor’s regime was the single most salient factor
in the failure of peace in 2000. Taylor’s actions abrogated expectations about
political participation and inclusion that prevailed in 1996–97 at the close of the
prior civil war. Although economic opportunities for private gain played a role
in Taylor’s governance, his exclusionary conduct, of which access to state rents
formed a part, led former allies to instigate a bloody reprise of the first civil war
(1990–96). In this sense Liberia represents a most difficult case study, because
it exhibits favorable characteristics for economic-based explanations. The case
study explores competing hypotheses about the role of economic motivations,
natural resource dependency, religious and ethnic fractionalization, weak institu-
tions, and poverty. In acknowledging the difficulty of distinguishing economic
opportunities from political motives, the chapter also analyzes factors that facili-
tated or impeded political exclusion by the Liberian government in the interwar
period from 1997 to 2000.
In addition, I explore the period of apparent postwar stability that Liberia
has enjoyed from 2003 to this writing. Thus the chapter’s final section examines
whether the factors that shaped the 1999 war recurrence seem to also account
for the more recent prolonged postwar peace. This chapter combines a heuristic
effort at generating hypotheses with a plausibility probe (Eckstein 1975) of these
existing claims in the literature. It seems to reinforce, rather than contradict, the
findings concerning the interwar period. The chapter concludes with a consider-
ation of the strengths and weaknesses of a case study approach to explaining the
failure of postwar peace and war recurrence.

The First Civil War

Liberia, Africa’s oldest independent country, is a small nation that had a popu-
lation of about 3 million in 1989. After its founding in 1847 by American ex-
slaves, Liberia was ruled by Americo-Liberians, who made up only 5 percent
Liberia 73

of the population. These elites repressed the roughly sixteen main ethnic “up-
country” groups, most of which had been divided by borders from their clans-
men in neighboring Guinea, Sierra Leone, and Côte d’Ivoire. Indigenous tribes,
for instance, only received the right to vote in 1964. In 1980 a coup led by Master
Sergeant Samuel Doe and other low-ranking officers, all from native tribes and
lacking a high school education, toppled American-Liberians from power for the
first time.
Doe’s regime continued the patronage-based, murderous predatory state of
his predecessors. His own ethnic group, the Krahns (who made up only 5 per-
cent of the populace), were overrepresented in the cabinet, and the Gios and
Manos (who constituted 15 percent of the population) suffered discrimination.
In the wake of a failed coup attempt in 1985, Doe’s Krahn-dominated army killed
a reported 3,000 Gios and Manos (Adebajo 2002a).5 As Adebajo (2002a, 46)
notes, “This single episode, more than any other, set the stage for the exploita-
tion of ethnic rivalries that would eventually culminate in Liberia’s civil war.”
On Christmas Eve 1989, Charles Taylor and some 167 other rebels invaded
from Côte d’Ivoire, launching a civil war that would persist through at least
eleven peace agreements and twenty cease-fires. Ethnic grievance played a role
in the conflict. Taylor’s National Patriotic Front of Liberia (NPFL), which was
composed mainly of Gios and Manos, gathered more of these ethnic nationals
as it swept across the country (Adebajo 2002a).6 NPFL rebels killed Krahns and
Mandingos, and Doe’s Armed Forces of Liberia also killed Gios and Mandingos
(Ellis 2001).
Six other rebel groups would emerge to play salient roles in the first civil war.7
Two splinter groups broke away from Taylor’s NPFL. Prince Johnson’s Indepen-
dent NPFL broke away in July 1990, and would later that year capture Presi-
dent Samuel Doe and videotape his grisly killing. Three senior NPFL officials
broke with Taylor in 1994 and formed the Central Revolutionary Council. In
contrast to these groups, other groups emerged that were composed mainly of
Krahns and Mandingos. Refugees from Sierra Leone formed the basis of United
Liberation Movement of Liberia for Democracy (ULIMO) in May 1991. Three
years later it split into two groups: ULIMO-J, led by Roosevelt Johnson, a Krahn,
with 3,800 members; and ULIMO-K, led by Alhaji G. V. Kromah, a Mandingo,
with 6,800 rebels (Adebajo 2002a). The 2,500-strong Liberia Peace Council
emerged in 1993 under the Krahn politician George Boley, who had served as a
minister under presidents William Tolbert and Doe. And the 400-member Lofa
Defense Force emerged in 1993.
At the same time, Liberia’s first civil war seems a good candidate for the eco-
nomic causes of armed conflict. The country is rich in natural resources, includ-
ing diamonds, timber, gold, rubber, and other minerals. In accord with some re-
search, its rebel commanders could enrich themselves and find the independent
74 Examining the Cases

means to sustain their fight. Taylor reportedly collected $75 million per year from
his territorial control of key resources, including $10 million per month from a
consortium of American, European, and Japanese mining interests. ULIMO-J
garnered income from diamond mining in the West, and the Liberia Peace
Council exported rubber. ULIMO-K sought to restore the wartime interruption
of Mandingos’ income from diamond exports. One study estimated that in 1995
alone, Liberia’s warlords exported more than $300 million worth of diamonds,
$53 million in timber, and $27 million worth of rubber (Atkinson 1997).
Fearon and Laitin (2003) argue that weak state structures typical of high
resource dependency provide the opportunity for armed rebellions to emerge.
Certainly the Liberian state was weak, and these huge incomes represented a
serious motive for rebel groups to continue their fight in a country whose gross
domestic product was only $135 million in 1995. Taylor’s use of the state to plun-
der the country’s resources after his 1997 inauguration supports this argument.
However, it is not clear that these resources constituted the main motives for the
start of the war.
Whether one sees ethnic or economic underpinnings of Liberia’s initial war,
ideology was not a driving factor. None of the insurgent groups articulated ideo-
logical motives for taking up arms. Although the anti-NPFL groups claimed to
fight for democracy, Adebajo (2002a, 47) states that these “were essentially ad
hoc ethnic armies led by individuals with dubious democratic credentials.” Com-
mand and control were extremely minimal in these organizations, with their
heavy reliance on spiritual and drug-induced states among a huge number of
child rebels. Horrible atrocities characterized the war.

The Onset of Peace

The peace process of the first civil war was long and tortuous—involving at least
eleven peace agreements between 1990 and 1996. Peace talks began in the early
months of the war. The Economic Community of West African States (ECOWAS)
provided the backdrop for negotiation efforts. Under Nigeria’s strong influence,
the regional body deployed 3,000 troops under the moniker ECOMOG (for “the
ECOWAS Cease-Fire Monitoring Group”). Their deployment reflected ambiva-
lence among the great powers and the United Nations, and greater African inter-
est in addressing its own problems. ECOMOG quickly became a party to the
conflict, fighting in support of the Armed Forces, ULIMO, and the Liberia Peace
Council, reflecting Nigeria’s antipathy toward Taylor’s NPFL. The regional force
grew to 12,500 (Howe 1996), even as the francophone states Côte d’Ivoire and
Burkina Faso initially refused to contribute troops, instead providing support to
Taylor (Adebajo 2002b; Ellis 2001). Liberians also grew to resent ECOMOG for
its joining in the plunder of Liberian resources, calling it “Every Car or Movable
Object Gone” (Shillington 2005, 102).
Liberia 75

By late 1990 Taylor’s NPFL had control of 80 percent of Liberia’s territory, all
outside Monrovia. The next five years were characterized by efforts by the anti-
NPFL forces to push back the NPFL’s quick territorial gains, and Taylor’s (un-
successful) efforts to take the capital (Alao, MacKinlay, and Olonisakin 2000).
ECOMOG, whose prompt initial intervention in 1990 had secured Monro-
via, would eventually also come under attack from anti-NPFL forces (Adebajo
2002b). As bloody battles affected tens of thousands of civilians, faction leaders
shuttled to West African capitals in a constant circuit of peace agreements, failure
to fulfill pledges, and renegotiations (Alao, MacKinlay, and Olonisakin 2000).
By 1995, in a crucial strategic shift, the Nigerian government had come to
accept the Ghanaian position of the need to accommodate Taylor and his forces
as part of a settlement of the war. The incorporation of troops from francophone
states into ECOMOG accompanied this shift, opening the way to a regional
consensus on a negotiated settlement that had the support of not only the anti-
NPFL forces but also Taylor’s NPFL. This newfound modus vivendi between
Taylor and Nigeria led to an agreement in the Nigerian capital of Abuja in August
1995. Although that accord foundered amid heavy violence in Monrovia by the
spring, in August 1996 exasperated West African heads of state foisted on the
parties a new, strengthened timetable for Abuja’s implementation, in what came
to be known as the “Abuja II” accord.
Abuja II would govern the end of the first civil war. It revised deadlines for a
cease-fire and concentration of forces; for verification of the same by ECOMOG,
the UN, and a transitional government composed of the factions (“Council of
State”); for disarming and demobilizing the factions; and for national elections.
It also empowered monitors to recommend travel restrictions and the seizure of
assets of recalcitrant warlords, and it authorized ECOWAS to remove any non-
compliant warlords from the Council of State and even from electoral candida-
cies (Pham 2004, 130).
Virtually all experts regard the civil war as concluded by 1996. The peace
agreement that ended the war was signed after Taylor reached a rapprochement
with the regional power, Nigeria, in June 1995, after which the August Abuja
accord was quickly signed (Adebajo 2002b). A cease-fire went into effect that
fall, and fewer than 2 percent of the war’s battle deaths occurred between that
time and early 1997, mainly during the 1996 battle over Monrovia.8 After some
delays, elections were held in August 1997. Taylor left little doubt that his loss
would lead to a return to the horrific warfare of the prior years. Partly due to the
pall of fear and intimidation, Taylor won a large majority. As Lyons (1998, 183)
notes, however, “Postelection powersharing or the need to negotiate the compo-
sition of the new Armed Forces of Liberia, for example, received little attention.
Political leaders preferred to allow the election to settle such contentious issues.”
What followed underscored the perils of using elections and democratization as
a conflict resolution mechanism.
76 Examining the Cases

The Second Civil War: A Brief Summary

About two years after Charles Taylor’s inauguration, small-scale attacks began
in Voinjama, along the Guinean border in Lofa County, a stronghold of the for-
mer ULIMO-K’s Alhaji Kromah. According to a statement issued (in the United
States) by the Liberia Coalition for Reconciliation and Democracy, referring to
itself simply as Resistance, the group was composed of former fighters from Tay-
lor’s NPFL, the disbanded Liberia Peace Council of George Boley, the breakaway
ULIMOs of Roosevelt Johnson and Alhaji Kromah, the Armed Forces of Libe-
ria, and the Lofa Defense Force of Taylor’s incumbent youth and sports minister,
François Massaquoi.9 The group, which would eventually dub itself “Liberians
United for Reconciliation and Democracy”—again, LURD—would follow a
trajectory similar to the NPFL, and thus it controlled most of the countryside
by 2003 but was unable to ever take Monrovia. Although the rebel organization
contained former rebels from numerous factions, LURD was seen as a reformu-
lation of the Mandingo-dominated ULIMO-K (IRIN 2004).10
The stated motive of the attacks that began in 1999 was to topple the govern-
ment of Charles Taylor: “The formation of LURD was motivated by its mem-
bers’ mutual opposition to what they viewed as a persistent pattern of ethnic bias,
political exclusion, human rights abuses, and corruption under Taylor. LURD’s
ethnically diverse makeup reflected the commonality of such views across ethnic
lines” (Cook 2004, 5). In other words, the renewed warfare did not simply reflect
ethnic aspirations of a single community or faction. Instead, the exclusionary
behavior was so egregious that it motivated a diverse collection of interests to
take up arms against the Taylor government.
Most of 1999 and 2000 saw sporadic attacks in the border area with Guinea, in-
cluding attacks inside Guinea by its own Taylor-backed dissidents. LURD would
escalate its attacks in 2000, forcing thousands to flee internally or across the Guin-
ean border. By the spring of 2001, LURD posed a serious threat to the govern-
ment’s survival. Its effort was boosted by the backlash from Taylor’s cross-border
meddling in both Guinea and Sierra Leone. At the same time Taylor was cam-
paigning for president in May 1997, a coup in Sierra Leone brought to power a
government allied with a Sierra Leonean rebel group that Taylor had practically
created, the Revolutionary United Front (RUF). As ECOMOG rushed to restore
the ousted elected president of that country, Ahmed Kabbah, 25,000 refugees
fled into Liberia. Eight months later, in February 1998, ECOMOG succeeded in
restoring the Kabbah government in Freetown. Rebels took to the countryside,
where they received increased support from Taylor’s government.
As a result of these regional dynamics, the governments of both Guinea and
Sierra Leone provided important support to the armed challengers to Taylor’s
government in Liberia during the second civil war. LURD’s military commander,
Liberia 77

Sekou Conneh (an ethnic Mandingo), was the son-in-law of Guinea’s presi-
dent. The ebb and flow of the regional conflagration saw LURD pushed back
into Guinea in September 2000 but then boosted, ironically, by Sierra Leone’s
peace process in 2001. Idle combatants from both sides crossed the border to
fight in Liberia. Some five hundred former army soldiers joined LURD, whereas
two thousand former RUF rebels fought alongside Taylor’s government forces
(Pham 2004, 182). Civilians continued to suffer the brunt of indiscriminate at-
tacks, as first one side and then the other would simply invade a town to loot it
and withdraw, and rarely come into direct combat with the opposing forces.
In 2002, LURD advanced to take important towns of Gbargna (home to Tay-
lor’s NPFL-controlled territory in the first war) and Tubmanburg, only to be
pushed back by Liberia’s security forces by October. Taylor’s forces benefited
from internal divisions within LURD, which had no clear political program be-
yond the ouster of Taylor (Human Rights Watch 2002a). Those divisions fell
along the divide between political leaders and military commanders, along with
tensions between Krahns and Mandingos. In early 2003 Liberian Krahns in Côte
d’Ivoire broke from LURD and formed (with Ivorian support) MODEL (Pham
2004, 185). LURD and MODEL constituted the two main rebel groups of the
second civil war.
Conversely, LURD benefited from growing international reprobation of
Taylor due to his continued support for the horrific fighting in Sierra Leone. In
1991, RUF, led by Taylor’s friend Foday Sankoh, launched its insurgency against
Sierra Leone’s government from Liberian soil. Taylor’s NPFL played a key role
in supplying arms. As president, Taylor continued his support of RUF and of the
Armed Forces Revolutionary Council (Olonisakin 2007). That destabilization
of the region undermined support for Taylor during the second war, both from
the big powers and from neighbors in the region.
In June 2003 sitting President Taylor was indicted for war crimes by Sierra
Leone’s Special Court, whose creation had been blessed by the UN Security
Council and the UN General Assembly. This indictment, together with deterio-
rating security and President George W. Bush’s ordering US warships to Liberia
in late July, contributed to Taylor’s resignation and flight to Nigeria. Two weeks
later, on August 18, 2003, a peace accord was signed between the two rebel fac-
tions and Taylor’s successor that led to a transitional government and ended the
second civil war.
Despite widespread corruption among former warlords in that interim gov-
ernment, a peace process enjoyed success seven years later. The warring factions
were demobilized relatively quickly, security sector reforms proceeded, donors
funded extensive reconstruction projects, and the election of President Ellen
Johnson- Sirleaf and a multiparty legislature in 2005 left no renewed warfare
on the horizon in 2010—despite persistent problems of corruption, a lack of
78 Examining the Cases

accountability, weak institutions, and poverty. What explains this second civil
war’s onset, given that end of the prior war and a serious international peace-
building effort? The following sections consider different explanatory factors.

Charles Taylor’s Exclusionary Behavior

Although the Abuja peace accord did not specify political powersharing, the op-
position’s expectations certainly included both political and security participa-
tion in the new government. In his 1997 inaugural speech, Taylor emphasized
reconciliation, human rights, and national unity. He promised that there would
be no witch hunts in his effort to rebuild the country from the ashes of war
(Adebajo 2002b). And he took several steps that appeared inclusive. He named
ULIMO-J’s leader, Roosevelt Johnson, to his cabinet, and ULIMO-K’s Alhaji
Kromah to head a National Reconciliation Commission (Cook 2003, 15). These
steps seem like inclusive measures.

Political Repression
By the end of 1998, however, these steps had proven themselves a chimera. The
most significant exclusionary behavior was Taylor’s repression of political oppo-
nents. By late 1998 Taylor had forced into exile his main political rivals, largely
through his partisan use of loyal security forces engaging in unaccountable ha-
rassment and repression. Sam Dokie, for instance, had become a prominent critic
of Taylor after having once been his ally. On November 28, 1997, Liberian police
detained Dokie and three family members. The next day members of Taylor’s
Special Security Service (SSS) forcibly took the Dokies away, and three days
afterward their mutilated bodies were found (US Department of State 1999, 1).
The SSS members were arrested and released a few months later, and the per-
petrators were never held accountable. The US State Department also reported
that, on several occasions in 1998, “security forces publicly disrobed, flogged,
and humiliated perceived opponents of the administration” (US Department of
State 1999, section 1c).
The most salient act of repression of political opponents targeted former
ULIMO-J leader Johnson and those ex-combatants close to him. On Septem-
ber 18, 1998, hundreds of Liberian security forces and irregulars launched a
planned operation called “Operation Camp Johnson Road” against the follow-
ers of Johnson, allegedly in response to an assassination plot (US Department
of State 1999, 1). Johnson survived the initial attack, and he sought refuge in the
US Embassy, at whose gates the government attacks continued, killing several,
including top deputy Madison Wion (Adebajo 2002a; IRBC 2001; US Depart-
ment of State 1999).11 After seventeen hours and several witnessed extrajudicial
Liberia 79

executions, the US State Department reported more than 300 dead, mostly eth-
nic Krahns (IRBC 2001).
Some observers saw the attack as an attempt to kill Johnson, who was sub-
sequently evacuated from the embassy. His departure ended powersharing in
the Taylor government. Following these seventeen days of unrest in the capi-
tal, some 9,000 persons, mostly ethnic Krahns, fled Liberia (US Department of
State 1999, 1). Two months later, Taylor accused thirty-two people—including
Johnson and his other main rivals, such as Ellen Johnson-Sirleaf, Alhaji Kromah,
and George Boley—of involvement in the plot (Adebajo 2002b). Several were
subsequently convicted in absentia or jailed, which seriously undermined the
possibility of unarmed political opposition.

Media Suppression
Taylor also used his presidential power to stifle media critics. Six months into
his term, his Ministry of Information ordered the country’s only private print-
ing press closed because of “inflammatory” editorials, although it reopened after
an apology. Police flogged the publisher of another twice-weekly newspaper in
January 1998, and Taylor’s security services periodically threatened journalists
and editors, leading to a degree of self-censorship among the most critical media
(US Department of State 1999, 3). Harassment of the media continued through
1999 and until the war ended (Williams 2006). Two radio stations were closed
in March 2000 (Adebajo 2002a).

Exclusion in the Security Forces

One of the most important manifestations of exclusionary behavior in Liberia
was in the composition and operations of the military, the police, and other
security forces. Reform of the military and other security forces constituted a
crucial security guarantee for the former insurgent groups in Liberia. Although
international actors sought to institute human rights changes and to purge some
of the most abusive elements, the expectations of Taylor’s former enemies in this
sector centered on incorporation of both senior officers and the rank-and-file
into the Armed Forces of Liberia and (to a lesser extent) the police and other
security forces.
Practically the first grievance articulated in LURD’s (2002) “Blue Print for
Stability in Liberia” is the Taylor regime’s malfeasance in the security sector. In
its fourth paragraph, this document states:

The [1997] elections however, were intended to have produced a government

that would guarantee the safety and security of the parties in particular and
80 Examining the Cases

the Liberian people in general. . . . This was in effect, a political settlement that
was intended to put an end to seven years of a bloody civil war. It was believed
then, that confidence, a necessary element of national reconciliation and lasting
peace, would have to be nurtured, beginning with the intents of the parties in
good faith to make peace. Evidence in this case may be found in the fact that the
first post-war government which is now headed by Mr. Taylor, was mandated
to restructure the Army, Police and the various security agencies with a view to
reflect neutrality. That is, none of the parties to the agreement was to maintain
dominance over the army or any of the security agencies. (LURD 2002)

The LURD blueprint goes on to quote the Akosombo Agreement, ratified in

Abuja I (1995) and Abuja II (1996), that “ECOWAS, the United Nations and
friendly governments will assist in planning the restructuring and training of the
[Armed Forces of Liberia].” LURD’s position was that four provisions of the
Abuja accord would be implemented, beginning with “1. The restructuring of
the military and security institutions of Liberia to reflect ethnic and geographi-
cal balance.”
Instead, the opposition perceived that the Taylor government violated these
expectations. After saying the above commitments were “dumped into the gar-
bage bin of history,” the LURD document states that “Mr. Taylor out turned [sic]
his NPFL into the national army and other security agencies. This worsened the
security situation in the country and broke down every little remaining trust the
citizens and residents had in national the security network” (LURD 2002).
Even if LURD is exaggerating the situation, analysts agreed that Taylor used
the security forces in a partisan manner, refusing to engage in the sort of security
reforms envisioned by either the opposition or the international community, and
ultimately resisting checks on the repressive behavior of the army and security
forces. First, Taylor took the position that the Abuja accords’ provisions about
restructuring the armed forces under ECOWAS and international guidance no
longer applied once his elected government took office (Malan 2008, 8).12 In-
stead, restructuring the army and police was his sole responsibility.
Second, Taylor’s notion of reforming the armed forces was to weaken them be-
cause he distrusted an institution still dominated by Samuel Doe’s ethnic Krahn
kinsmen. For instance, he demobilized more than 2,600 soldiers from the armed
forces, including most of the remaining Krahn officers, replacing some with his
former NPFL fighters (Adebajo 2002b; Pham 2004, 177). Instead Taylor sought
to strengthen other security forces like the National Police, the SSS, and the Anti-
Terrorism Unit (ATU). These units, used in operations against political oppo-
nents, became influential and feared. According to the US State Department, as
of the end of 1998, “The many, newly created security services absorbed Taylor’s
most experienced civil war fighters. Armed units within these services consisted
Liberia 81

almost exclusively of undisciplined Taylor loyalists” (US Department of State

1999, 1).13
Moreover, Taylor appointed relatives and other trusted aides to run these se-
curity agencies based on partisanship and loyalty rather than professional merit.
Thus Taylor’s son “Chucky” was appointed to lead the ATU, and Taylor’s cousin
Joe Tate headed the Liberian National Police (Malan 2008). Tate “was accused
of having led gangs of looters and a political death squad during the civil war”
(Itano 2003, 3). According to Itano, “Other [units], more informally organised,
had names like ‘Jungle Fire’ or no name at all, and were simply informally orga-
nised units of boys led by a slightly older boy who had been with Taylor during
his days in the bush” (Itano 2003, 3). Units like the SSS, the ATU, and Jungle Fire
roamed the streets with impunity, generating widespread fear.
These developments led analysts to conclude that Taylor’s policies consti-
tuted the primary factor in the resurgence of civil war. As Adebajo explains, “As
the Liberian president continues to use state power to silence his critics and aims
to remove all opposition to his rule prior to the next elections scheduled for
2003, he has created a situation similar to that produced by Doe’s fraudulent
election of 1985, where opponents were left with no legitimate way of challeng-
ing Doe except through violence” (Adebajo 2002a, 70). Itano says that, after
discussing Taylor’s persistent complaints that the international arms embargo
thwarted his defense of the country, “A weightier factor in his government’s de-
cline was Taylor’s own style of rule, which caused him to centralise power in
himself and keep the Liberian military fractured and divided” (Itano 2003, 3).

Alternative Explanations
Having set forth the case for a precipitating role by Charles Taylor’s exclusion-
ary behavior, what can one make of the several widely perceived causal factors
in Liberia’s war recurrence in 1999 and 2000? This section explores the main
alternative explanations for the renewed armed hostilities.

Regional Factors?
What role did regional factors play in Liberia’s civil war recurrence? Certainly
regional factors played a salient role in the onset, duration, character, and recur-
rence of internal armed conflicts throughout West Africa. In some ways Liberia’s
wars, along with those of Sierra Leone and Côte d’Ivoire, are best analyzed as
manifestations of a single regional dynamic from 1989 to 2007. Rebel armies were
formed with the diplomatic support and funding of neighboring states; cross-
border rearguards played important roles in all cases; smuggling networks cut
across borders; refugees became strategic resources in mobilizing combatants;
82 Examining the Cases

personal relationships with neighboring heads of states shaped warfare; and

the subregional and regional organizations (ECOWAS and the African Union)
played salient roles in making war and peace. A regional level of analysis, eschew-
ing the national level altogether, would shed light on the dynamics of conflict
both within and across each country.
Such regional dynamics certainly shaped Liberia’s second civil war. The peace
processes, held under the auspices of ECOWAS, came to fruition only after the
positions, and in some cases individual leaders, of neighboring states shifted.
Guinea, Côte d’Ivoire, and Sierra Leone all became staging grounds for the at-
tacks that heralded the second civil war. Most important, Taylor’s continued sup-
port for RUF and other rebels fighting against Sierra Leone’s President Kabbah
became the single most important strategic factor in big powers’ policies toward
postwar Liberia. And during the second war, RUF forces became a crucial ally
for the Liberian state, as perhaps five thousand RUF rebels “simply moved to a
different conflict, taking service with the RUF’s longtime ally, Liberian President
Charles Taylor, in his own ongoing civil conflict” (Itano 2003, 3).
Nevertheless, national-level factors are important in determining the specific
configuration and historical trajectory of armed conflict in each society. One
cannot understand or explain the causes, timing, and character of Liberia’s first
civil war, its peace, nor its resurgence solely with regard to the regional dynamics.
Nothing at the regional level made internal armed conflict in Liberia inevitable,
and understanding the processes and triggers of the second civil war requires first
understanding the actions taken by Taylor’s government vis-à-vis other social
sectors and, especially, his handling of the security sector and his treatment of
political opponents and the ethnic groups associated with them.

International Factors?
What role might international factors more broadly have played in civil war re-
currence? Some analysts, after all, believe that, with the end of the Cold War, the
United States’ neglect and the subsequent aid cutoff were the decisive factors
in the onset of war in Liberia a month after the Berlin Wall fell in 1989 (Pham
2004, 89–91). Did the West topple Taylor? Was the UN mission or its missteps
responsible for the recurrence of armed conflict?
International factors played a role in Liberia’s war recurrence, but they consti-
tuted neither the main cause of war recurrence nor a major factor in Taylor’s vio-
lation of expectations surrounding the peace process. Some initial steps showed
support for the peace process. International actors facilitated the end of a war-
time shipping embargo and the reopening of seaports and airports.
Yet Western governments approached Taylor’s presidency with hesitation.
Although they viewed Taylor as a necessary part of the peace process in the
Liberia 83

mid-1990s, Western diplomats were well aware of his reputation for avarice and
corruption, of his ethnicity-based alliances, and of horrific rights abuses. At the
first donor conference under Taylor’s presidency, donor pledges of $230 million
reached only half the $438 million requested by the government (Eziakonwa
1998, 26).14 In addition, diplomatic sources in Liberia insisted publicly in mid-
1998 “that the government must implement more macroeconomic adjustment,
control its security forces, and attend to human rights as the preconditions for
providing new foreign aid to Liberia.”15 These concerns with democracy, corrup-
tion, and human rights rendered disbursements very slow.
The most significant international actions toward the Taylor administration
stemmed from his role in fostering warfare in the region, especially Sierra Le-
one. As Liberia’s peace process came to fruition, Taylor-backed RUF forces and
disgruntled former soldiers (the Armed Forces Revolutionary Council, AFRC)
took the capital for the first time in May 1997. They ousted that country’s presi-
dent, Ahmed Kabbah, who enjoyed the backing of the West and ECOWAS.
ECOMOG deployed in force to Freetown, becoming the main force confront-
ing the RUF–AFRC alliance (International Crisis Group 2002, 2). The regional
force ousted the rebel regime from the capital in February 1998.
ECOWAS represented the single most important external relationship for
the Taylor administration, and tensions over Sierra Leone intensified. Although
ECOWAS leaders had decided to accommodate rather than continue fighting
Taylor in the Abuja agreements of 1995 and 1996, that decision was rooted in the
hope that peace in Liberia would curb Taylor’s support for RUF in Sierra Leone
(International Crisis Group 2002, 2). Yet as Liberia’s peace process progressed,
Taylor-supported RUF forces surged rather than ebbed, undermining the rap-
prochement with Taylor. RUF rebels had killed dozens of Nigerian ECOMOG
troops in May 1997, and as president, Taylor refused ECOMOG jets access to
Liberian airfields for their Sierra Leonean operations. Upon his election, Taylor
also successfully pressured Nigerian president Sani Abacha to replace ECOMOG
force commander Victor Malu, who had insisted on ECOMOG’s independence
from Taylor’s government and on ECOMOG’s role in restructuring Liberia’s
armed forces (Adebajo 2002b). Boosted by the legitimacy of his election and
by international ambivalence about paying for peacekeepers and reconstruction
activities in a corrupt environment, Taylor weakened ECOMOG’s influence, es-
pecially its ability to restructure the armed forces. Its troops departed Liberia in
late 1998.
Largely because of Taylor’s support for the insurgency in Sierra Leone, do-
nors failed to provide him with resources or diplomatic support. Rather than
attempting to use diplomatic leverage to shape his behavior, they withdrew their
support. In March 2001, the United States sponsored unilateral sanctions, and
the United States and United Kingdom backed UN Security Council resolutions
84 Examining the Cases

sanctioning trade in “conflict diamonds” (Cook 2003, i). These understandable

measures may have hastened war recurrence. They reinforced Taylor’s autonomy
to repress his opponents, removed the check provided by international troops,
and undermined the process of economic recovery.
Aside from Security Council resolutions passed in New York, the United Na-
tions had little military impact in Liberia. The UN Security Council, seeking to
shift the burden of peacekeeping to regional organizations in Africa, deferred to
ECOWAS, authorizing only 300 unarmed observers in Liberia. During the war,
the UN Observer Mission in Liberia (UNOMIL) was largely dependent for secu-
rity on ECOMOG. As the peace process progressed in 1996, fewer than 100 UN
military observers remained (United Nations Department of Public Informa-
tion, n.d.). At the urging of Taylor, those UN observers were withdrawn less than
two months after he took office. A smaller UN Peacebuilding Support Office was
erected with a handful of civilian staff. That office, which had few resources of its
own, sought to mobilize resources for the newly elected government from 1997
to 2000. Consequently, Liberian civil society and outside observers criticized
it for excessive cozying up to the Taylor regime (Adebajo 2002b).16 In 2002,
the US nongovernmental group Human Rights Watch (2002b, 65–69) would
conclude that this UN office had “played little or no active role in addressing the
growing repression and abuses in Liberia.”
Ultimately, neither the United Nations nor the major powers can be seen to
have brought about Liberia’s second civil war. Isolation by intergovernmental
actors of Taylor due to his regional warmongering weakened his government
and probably exacerbated his perverse, rent-seeking, abusive, and exclusionary
behavior. But that behavior was not solely or even mainly a response to inter-
national attitudes or actions. Instead Taylor used his office to extend his long-
standing exclusionary behavior. He also expanded his regional support for war-
fare, even while he sought to reassure the international donors and financial
institutions that he was satisfying their expectations. All the while he reneged on
the expectations and agreements found in the Abuja accords.

A Flawed Demobilization and Reintegration Process?

A number of theorists and policy analysts emphasize the process of disarma-
ment, demobilization, and reintegration (DDR) of combatants as a determinant
of success in peace processes. To what extent was the process of DDR flawed in
ways that led to war recurrence in Liberia? The implementation of disarmament
and demobilization under Abuja II offers divergent implications.
First, the process of demobilization was praised by many observers. Once the
factions became committed to elections, demobilization proceeded relatively
Liberia 85

smoothly. As the UN mission closed in September 1997, Secretary-General

Kofi Annan celebrated the impact of the DDR process: “The successful com-
pletion of the disarmament and demobilization exercise, which began in No-
vember 1996 and ended in February 1997, with the surrender to ECOMOG
under UNOMIL observation of about 10,000 weapons and 1.24 million pieces
of ammunition, marked a significant turn of events in Liberia” (United Nations
Secretary-General 1997a).
Indeed, the need to overcome delays in demobilization is one of the main jus-
tifications for postponing those elections twice. The two faction leaders commit-
ted to elections—Taylor and Kromah (ULIMO-K)—controlled an estimated
60 percent of fighters, and made up 84 percent of all demobilized combatants
(Alao, MacKinlay, and Olonisakin 2000, 100). The UN Secretariat’s final report
also indicated that reintegration programs had helped generate quick employ-
ment for thousands of ex-soldiers: “The initial reintegration programmes estab-
lished following the end of the disarmament exercise were crucial in providing
useful employment to thousands of former fighters and war-affected populations
as a means both of restoring social stability and of rehabilitating some of the
country’s basic infrastructure” (United Nations Secretary-General 1997a).
However, the process was certainly flawed. Of an estimated 33,000 combat-
ants to be demobilized, no more than 21,315 had demobilized by the time of the
July 1997 elections (United Nations Secretary-General 1997b, para. 14; Deng
Deng 2001). The factions retained sufficient arms to easily reinitiate warfare
if the peace process or elections did not go as planned. Following the path of
Angola’s Savimbi in 1991, Taylor’s NPFL depicted total demobilization before
the election as “unrealistic” (Deng Deng 2001, 98), which preserved its ability
to either dominate the security sector after electoral victory or become spoilers
after electoral loss.
And yet it would be misleading to attribute the recurrence of armed conflict to
decisions surrounding DDR. Even a full demobilization of Taylor’s NPFL would
not have stopped him, once elected, from remobilizing his forces and using the
army as much as feasible to advance his own agenda in the region and internally.
Moreover, the informal nature of the recruitment, organization, and command
structure of Taylor’s forces made it easy for him to reconstitute those same in-
formal forces in the 1997–2002 period, before and during the second civil war.
Further demobilization of the armed forces would, ironically, have weakened a
force that could act as a counterweight to the newly elected President Taylor and
his loyal forces.
Barbara Walter (2004) also emphasizes the individual-level decisions by po-
tential combatants, especially young unemployed men, to join rebel movements.
Looking more broadly at factors beyond the DDR process, Utas (2005) sees “the
86 Examining the Cases

failure of impoverished youth to find a place in Liberian society” as a key factor

both in the second civil war and in the chances for recurrence (Utas 2005, 151).
He persuasively shows a postwar social malaise and marginalization felt by youth
whose age progresses but whose social status, reflecting their jobs and income,
remains low and frustrating.
In counterpoint to this hypothetical relationship, the postwar period saw little
change in this structural situation (Utas 2005, 137–53). Utas also documents a
process of reintegration of ex-combatants in the 1996–98 period in rural areas,
but no such process in urban areas (Utas 2005, 150). Moreover, he depicts mar-
ginalized urban youth as less central to the recruitment effort of LURD than ex-
iles and rural populations far from the capital. Despite the continued difficulties
facing former combatants and youth, he acknowledges that these were not the
crucial factors in the onset or progress of the second war.

“Greed” and Natural Resources?

The current thinking about economic motivations in civil wars is perhaps the
most persuasive alternative account of civil war onset and its recurrence in Libe-
ria. The country is heavily dependent on natural resources, which played a role
in the financing and recruitment of armed groups. Although Liberia was a major
supplier of iron ore to the world before the war, timber and rubber were the main
exports in the late 1990s. According to one source, by 2000 Liberia’s timber ex-
ports had risen to $221 million in value, 230 percent of their 1987 levels, consti-
tuting 22 percent of gross domestic product (Outram 2004). The largest timber
operation in Liberia, called the Liberian Forest Development Company, secured
a remarkable concession from President Taylor in 1999–2000 for 1.6 million
hectares across four counties that made up 42 percent of the country’s produc-
tive forests.17 Two panels of experts appointed by the UN Security Council re-
ported in 2000–2001 that this timber company, which Taylor called “his pep-
perbush (private matter),” was trafficking arms through Liberia for Sierra Leone’s
war (Outram 2004, 613–14).18
Rebel groups exploited natural resources during and after the initial civil war.
One expert noted that “the resulting logique de guerre crippled the Liberian state
as an organ of administration and created a new set of economic opportunities
based on plunder.”19 During the very first year of the war, Taylor’s NPFL suc-
cessfully sought to negotiate with various international markets to supply iron
ore. Within months he was receiving $10 million per month by transporting
iron down NPFL rails to then NPFL-controlled port of Buchanan, and then on
mainly to France (Pham 2004, 122). When ULIMO split, Kromah’s ULIMO-K
assumed control of the Bomi mining operations, just as the Liberia Peace Coun-
cil assumed control of timber exports along the Ivorian coast (Pham 2004, 121).
Liberia 87

Will Reno observed that “the war has been as much a battle over commerce
inside and beyond Liberia’s borders as it has been a war for territory or control
of the government.”20 Once the first war ended, disarmament and demobiliza-
tion occurred to a lesser degree in areas of natural resource concentration. This
pattern suggests that control of those resources underlay that process (Alao,
MacKinlay, and Olonisakin 2000, 100).
Economic motivations must be included in Taylor’s motives for waging a
regional war even once it was apparent he would become president in Liberia.
His own survival and his political power were apparent motives for his regional
aggressiveness. He remained, for example, extremely frightened for his own
personal security. Similarly, ethnic alliances and the training he and ally Foday
Sankoh received in Libya fostered political affinities. Yet no analysts allege any
strong ideological character to Taylor’s platform or his behavior. Taylor’s in-
volvement in smuggling diamonds, timber, and other natural resources was an
important component of the regional dynamic. His involvement in supporting
RUF, despite his protestations, was documented by a panel of experts formed
by the UN Security Council in 2000. Although he tried to forestall sanctions
by grounding his aircraft and inviting international monitoring against diamond
smuggling, the international community was unconvinced (Adebajo 2002b). In
March 2001 the UN Security Council, at the behest of the United States and
the United Kingdom, approved a ban on the sale of Liberian diamonds, an arms
embargo, and a travel ban targeting Liberia’s political leaders and their families
(Pham 2004, 182).
Of course the state is a powerful source of legitimate access to money, as dem-
onstrated by the rapid increase in tax revenues once postwar elections were held.
According to the Taylor government, it increased state revenues from a little more
than $1 million in the first seven months of 1997, just before Taylor assumed of-
fice, to $125 million in the subsequent five months (Eziakonwa 1998, 26). This
view is supported also by Taylor’s efforts to expand presidential authority over
strategic natural resource contracts. In 2000 Taylor also got the Assembly to pass
a law granting him the sole power to conclude commercial contracts for strategic
commodities, and thus channel their revenues directly to his control (Thomson
Reuters Foundation 2001). One report asserts that the Taylor / Sankoh split that
led to the capture and holding of more than 500 ECOMOG soldiers in 2000
by RUF was sparked by a dispute over access to Sierra Leone’s diamond fields
(International Crisis Group 2002, 3).
These empirics suggest that natural resources and private gain were crucial in
accounting for the first war and its persistence. Because combatant factions do
not generally acknowledge their economic motives as ends rather than means,
it is impossible to conclude with certainty the role such factors play in warfare.21
Taylor’s exploitation of the state for personal gain is undeniable, and it is difficult
88 Examining the Cases

to determine whether his former allies and enemies took up arms because of
their political, social, or economic exclusion from the spoils of the state, and
whether their individual exclusion or that of their ethnic group was causal.
However, the role of natural resources and private gain seems less persuasive
in understanding why war resumed in 1999 and 2000. Taylor’s exclusionary be-
havior is the indispensable factor in explaining civil war’s recurrence in Liberia.
The factions opposed to Taylor did not immediately challenge his government,
indicating that their treatment by Taylor was relevant in the decision to defect.
At a minimum, their desire for access to the spoils of power was mediated by
the marginalization perceived to stem from Taylor’s repression and exclusion.
Although LURD’s financing proved murky, it was not involved in commercial
logging until very late in the war after it took control of most of the country
(Brabazon 2003, 5). Indeed, some LURD commanders feared that exploiting
diamond mines would lose them the moral high ground against Taylor, even
when they took control of important mining areas in early 2002 (Brabazon 2003,
6): “During LURD’s initial occupation of this area, individual commanders and
fighters reportedly sold stones on a personal basis, a practice which led to one
colonel being ambushed by his own troops. Indeed, it is this fear that diamond
wealth will split the organization with financial jealousy that has so far prevented
any concerted effort to exploit mineral resources.”
The empirics suggest that both “greed” and “grievance” played a role in civil
war recurrence but that the latter was the clearest and most important precipi-
tating factor in a complex interaction process. Natural resources were important
when Taylor excluded even his allies from the spoils of office, and they again
became a factor once LURD had launched the renewed war in funding opera-
tions and attracting combatants. Yet exclusionary behavior seemed to underlie
the motives as seen through both rebel actions and rhetoric during the second
civil war. Moreover, Taylor’s conduct also shows how weak institutions that are
created by exclusionary conduct not only reflected weak institutions but also
helped to reinforce, and even reconstitute anew, the country’s personalized and
weak state institutions (Sawyer 1990, 148–73).

Insights from Liberia’s Second Postwar

Peace Process, 2003–Present
Although it is too early to assess the durability of Liberia’s second postwar peace,
it is instructive to explore what factors have facilitated a different outcome: sta-
bility for seven years. The existence of two peace processes in the same country
within a few years permits an initial analysis of the pertinent choices and strategic
factors, given that static variables changed little in the period. This section probes
Liberia 89

the plausibility of the different hypotheses suggested by Liberia’s failure to con-

solidate peace in the 1990s.
After Taylor’s departure and the Accra Peace Agreement of 2003, a transi-
tional national government exercised control of all three branches of govern-
ment until elections were held in November 2005. More than 100,000 combat-
ants were demobilized with the support of a new peacekeeping operation, the
UN Mission in Liberia (UNMIL). The reintegration process was better con-
ceived in its extension to communities where ex-combatants would return, but
slow and imperfect in execution and in generating employment (United Nations
Secretary-General 2007, para. 32).22 Despite occasional political unrest, peace
has held from 2003 up to this writing.
The differing outcome of this second peace process vis-à-vis the first can-
not be due to largely static factors such as poverty rates, state weakness, high
dependence on natural resource exports, demographics, and geography. With
minor variations, these variables changed little between 1996 and 2003. Instead,
the stability that issued from the 2003 peace process lends support to a number
of theoretical claims about strategic action by national or international actors.
Two of these stand out against the backdrop of the failed peace process from
1996 to 2000: (1) the more seriously backed and better-resourced international
peacekeeping presence; and (2) the more inclusive approach to governance by
Liberian elites, both in the more encompassing powersharing arrangement dur-
ing the 2003–5 transitional government and in the nonexclusionary conduct
of President Ellen Johnson-Sirleaf ’s administration. Here I briefly explore these
two factors in the persistence of postwar peace.
First, international peacemaking and peacekeeping efforts involved more
serious diplomatic commitment and larger and more fully resourced military
deployments. Although the international community was pleased to see the first
civil war end, virtually all the major Western donors had long expressed their
opposition to Charles Taylor and were reluctant to commit many resources to
his government once elected. The UN Security Council’s decision to wrap up
its peacekeeping operation there and instead create a troopless, and thus tooth-
less, political mission known as the UN Peacebuilding Support Office in Liberia
emblemized this lack of diplomatic support. West Africa remained divided over
Taylor’s rule and preoccupied by his increasing destabilizing actions in neighbor-
ing countries, especially Sierra Leone.
The enhanced international commitment was visible in many ways. UNOMIL
cost $81 million over the almost four years from its inception to June 30, 1997,
and, on the eve of Taylor’s assuming the presidency, the UN authorized only
$20.4 million for the peace operation during the year immediately following
(United Nations Department of Public Information, n.d.).23 Conversely, UNMIL
90 Examining the Cases

had a budget of more than $781 million for the second year after a new presi-
dent was elected following the second civil war, from July 2007 to June 2008
(Dupuy and Detzel 2007, 17). The decision of the UN Security Council to itself
authorize 15,000 troops in 2003, rather than rely upon ECOWAS to provide
peacekeepers, also reflected greater external commitment. Although US troops
never landed on Liberia’s shores, President George W. Bush’s decision to station
marines just offshore in August 2003 signaled their potential to intervene and a
greater priority than during the 1990s.
More significant was the decision to maintain peacekeepers for a period of
years after the 2005 elections. This decision contrasted starkly with the with-
drawal of UN troops within two months of Taylor’s swearing-in in 1997. Dupuy
and Detzel (2007, 23) conclude that “the success of the CPA [Comprehensive
Peace Agreement] is likely due to the deployment of UNMIL, which has pro-
vided public security and maintained stability in the country.” The difference
between the swift departure of foreign troops after the 1997 election and the sus-
tained peacekeeping presence since 2003 lends support to the findings of Fortna
and of Doyle and Sambanis that peacekeeping troops’ prolonged presence, es-
pecially when part of a multidimensional peace operation, reduces the chance
of war recurrence (Fortna 2008; Doyle and Sambanis 2006; Sambanis 2008).
Another measure of the seriousness of the international commitment were
the sanctions adopted by the UN Security Council on the importing of any
Liberian-produced diamonds, on the exporting of arms into Liberia, and later
on the importing of Liberian timber.24 No such leverage was exercised during
or after the prior war. Crucially for international leverage, the UN considered
and rejected a 2004 Liberian request to list the restrictions, providing continued
leverage over the Liberian government until the revocation of all such targeted
sanctions in 2008. Linked to this effort was the broader, more comprehensive
approach to postwar peacebuilding by external actors. Although the approach
was similar, the breadth of civilian commitment was greater after the second war
ended in 2003, with longer-term and better funded reform projects for the secu-
rity and justice sectors, the financial sector, and the civil service.
Second, in terms of internal economic or political strategic choices, the most
salient factor in the preservation of peace since the 2003 cease-fire ended Li-
beria’s second civil war has been the inclusionary character of governance, eco-
nomic, and security arrangements. The August 2003 Accra Peace Agreement
brought an end to the war, and it enshrined an interim powersharing arrange-
ment. Like the peace accords signed during the first war (including Abuja), this
agreement has continued to reward faction leaders by granting them personal
positions in the Cabinet, and thus permitting them to plunder state resources for
personal gain in keeping with political economy theories.
Liberia 91

However, the postwar political arrangements in the Accra agreement dif-

fered in important ways from those that followed the first civil war. This accord
bolstered the role of civil society and sought to prevent a new strongman from
emerging. It stipulated that the chairman and vice chairmen come from civil so-
ciety. The accord also prevented the interim leaders—including the transitional
chairman, vice chairman, key cabinet members, all top transitional Supreme
Court magistrates, and the Assembly speaker and deputy speaker—from run-
ning for any elective office in the 2005 elections (Articles 24, 25, 27). And it
established a Governance Reform Commission, whose members are appointed
from a list provided by civil society organizations and include their representa-
tives, as well as a Contract and Monopolies Commission comprising civil society
representatives “independent of the commercial sector” (Articles 16 and 17).
The role of civil society and political parties was also strengthened in a novel
arrangement regarding powersharing in a transitional legislative assembly. Of
seventy-six members, twelve came from Taylor’s government and each of the
two armed movements, with fifteen representing each of the country’s counties,
plus eighteen from political parties and seven from civil society organizations
(Article 24). Furthermore, the army was restructured to include rebel forces (Ar-
ticle 7). The breaking of Taylor’s power was necessary to create the powersharing
arrangement contained in the 2003 accord, and the pressure of the war crimes
indictments from Sierra Leone, as well as diplomacy by the UN, regional organi-
zations, and the United States proved crucial for his departure.
These findings suggest that the sustained presence of international peace-
keeping troops, a more committed approach to resourcing and supporting
peacebuilding efforts, and the inclusionary behavior of elites have been crucial
to the stability enjoyed by Liberia since 2003. These tentative findings based on
the postwar peace reinforce the conclusions drawn here based on Liberia’s recur-
rence of civil war in the late 1990s. Even if one acknowledges the role that access
to the wealth of natural resources and their export have for the likelihood of
war recurrence, this recent experience points to the importance of how national
actors govern with regard to political access and economic resources.

Conclusion
What explanations of Liberia’s second civil war are most persuasive? How well
do the circumstances and factors surrounding Liberia’s renewed warfare con-
form to various theories?
Liberia’s experience from 1996 through 2000 provides support for several
theories of civil war and its recurrence. Liberia is a poor society with weak state
institutions, lying among war-torn neighbors, and highly dependent on natural
92 Examining the Cases

resources, where young men constitute a large proportion of the population and
where ethnic divisions are significant. Neighboring states and the international
response to Charles Taylor’s warmongering in Sierra Leone contributed a rear-
guard, resources, and even combatants to LURD and MODEL. However, these
new insurgent groups were not created by external actors nor principally by the
lure of economic rents but by Liberians reacting to Taylor’s policies.
As shown above, natural resources like timber, rubber, and minerals played an
important role as incentives to and resources for civil war. This political economy
explanation represents the most plausible alternative to the role of exclusionary
behavior during the interwar period. Yet Taylor’s exclusion of all his main op-
ponents from the political stage, not to mention his violence and discrimination
against them, was the primary driver of Liberia’s war recurrence. Unprovoked by
any coherent armed opposition, Taylor began to repress his political opponents
and regional rivals soon upon acceding to the presidency. He then reneged on the
international and national expectation surrounding the downsizing, reform, and
curricular revision of the armed forces and stalled on these stipulated changes.
Despite the absence of detailed provisions for inclusion in the peace process of
1996–97, he violated expectations by dismissing his former enemies from the
government after promising them a share of power, he increasingly pushed out
his former enemies from the army and security services, and he used violence
and threats of arrest against some of these dismissed officials.
Taylor’s exclusionary behavior, especially in the security services, involved
patron–client networks that marginalized ethnic Krahns and Mandingos. His
policies led former ULIMO combatants to reorganize and challenge the state.
The competing informal networks characteristic of a weak state fell along the
lines of ethnic difference and grievance, fueled by a lack of resources. Yet nei-
ther weak institutions nor ethnic diversity per se is a sufficient explanation. Tay-
lor’s transparent ethnic manipulation made the difference. His exclusion was so
egregious that many of his senior followers through the end of the first civil war
helped form LURD. In the end, grievance accounts for this “tough” case of civil
war recurrence at least as much as the economic factors prevalent in the contem-
porary literature.
Beyond the substance suggested by this case study, it offers a chance to reflect
on case studies as a method of researching civil wars and their recurrence. This
case study, though based principally on secondary sources, offers some clear
advantages over the quantitative methods examined in chapter 1. Quantitative
methods can identify and confirm the likely causal impact of variables that can be
measured comparably across many cases, including institutional arrangements
like powersharing and risk factors like poverty or dependence on oil exports.
Unless they find a 100 percent correlation, such methods are obviously unable to
specify whether such variables are operative in any given case. Moreover, quan-
Liberia 93

titative methods cannot capture the strategic choices made by elites, even when
these are conditioned by risk factors or institutional arrangements.
This Liberia case study shows the value of qualitative methods for under-
standing the role of elite choices in postwar societies. Taylor’s decision to renege
on the arrangements agreed upon is crucial for explaining Liberia’s war recur-
rence. Despite the deliberate self-interested misrepresentation of motives, how
rebels frame their motives is relevant for understanding causes of civil wars and
the lines along which mobilization does or does not occur.
Case studies also hold heuristic value. Even a relatively structured and fo-
cused case study such as this one reveals empirics that suggest new lines of in-
quiry and potential causal factors that may require greater attention. This case
study has revealed the salience of regional and international levels of analysis.
Quantitative analysis has addressed these factors by retaining country-level units
of analysis but also including variables for “regional” impact or neighbors at war;
however, these efforts do not entirely capture the insights provided by shifting
levels of analysis. This shift, though not executed in this study (e.g., I did not
examine West Africa’s conflict cluster, the Great Lakes conflict cluster, or the
Central American conflict cluster), can nevertheless be identified and explored
in a case study. The case would be strengthened by more detailed process tracing
of the various factors that shaped Taylor’s decision making, including access to
advice he received and the pressures of various domestic constituencies, as well
as the application of psychological methods.
Qualitative methods, and this case study specifically, also carry liabilities.
Aside from its execution, a case study of Liberia cannot wholly sift through the
economic versus the power-based political motives of Charles Taylor or his asso-
ciates once in office. That difficulty undermines the effectiveness of the method
and the persuasiveness of the argument. Even more empirical detail would prob-
ably not produce a definitive finding.
Moreover, the robustness of a particular variable is not verifiable based on a
single case study. Those who execute case studies perceive their findings to apply
more broadly, at least to similarly situated cases. However, a variable that seems
consequential may be irrelevant in all other cases. Only through an examination
of other cases (whether similar or not) can one determine whether a particular
factor is relevant for the entire pool of cases. It is to such a method that I subse-
quently turn by conducting briefer, more structured, and more focused “mini”
or abbreviated case studies.
94 Examining the Cases

Notes
1. Several sources cite the 150,000 estimate of total deaths; e.g., see Stedman (1996,
252 n36).
2. Although the government did not verify the identity of the rebels as Liberian,
Reuters reported that locals alleged them to be members of ULIMO-K and ULIMO-J.
3. LURD was formed in February 2000 by former Taylor allies, consisted mainly of
former ULIMO-K members, but included also former members of prior rebel groups
ULIMO-J, the Liberia Peace Council, and former government soldiers.
4. Peace Research Institute Oslo (PRIO) Conflict Summaries, 2003a. Available
from the author.
5. The coup attempt was led by General Thomas Quiwonkpa, the founder of the
National Patriotic Front of Liberia, which Taylor would eventually command.
6. The NPFL also included Burkinabe, Gambian, and Sierra Leonean mercenaries.
7. This paragraph draws on Adebajo (2002a, 46–47).
8. PRIO Conflict Summaries for Africa, 2003a.
9. Statement released April 25, 1999, from the Global Security Organization web-
site on military factions, www.globalsecurity.org / military / world / para / lurd.htm.
10. Peter Dennis, “A Brief History of Liberia,” paper for the International Center for
Transitional Justice, May 2006, author’s files, 14.
11. Wion had allegedly confessed to involvement in the assassination plot (IRBC
2001).
12. Taylor cited the wording of the Akosombo Agreement, which stipulated that
planning for the restructuring and training of the military would be the responsibil-
ity of the Liberian Transitional National Government, with support from interna-
tional actors. See Akosombo Agreement, September 12, 1994, Section H, Article 9,
Count 4.
13. Nicole Itano emphasizes the tactical consequences of this divisive style, as a
weakened military proved unable to defend Monrovia against the rebels’ advance. See
Itano (2003, 3).
14. The donor conference was held in Paris in April 1998.
15. “Liberian Daily News Bulletin,” Star Radio, July 24, 1998, as quoted by Geepu-
Nah Tiepoh, “The Hubbub over Foreign Aid: Facing the Sobering Reality,” www
.theperspective.org / hubbub.html.
16. Author’s personal interviews, Monrovia and Dakar, August 2004.
17. The company was earlier the Oriental Timber Company, and later the Liberian-
Malaysian Timber Company.
18. One was the December 2000 report of the Sierra Leone Expert Panel, and an-
other was the October 2001 report of the UN Panel of Experts on Liberia.
19. Stephen Ellis, quoted by Pham (2004, 121).
20. William Reno, quoted by Pham (2004, 121n1).
Liberia 95

21. As Collier et al. argue in Breaking the Conflict Trap (2003), rebel leaders have
many incentives to hide their true selfish motives behind rhetoric that instead empha-
sizes legitimate collective grievances.
22. Also from the author’s interviews with officials of UNMIL and Western embas-
sies, Monrovia, August 2004.
23. See UN General Assembly Resolution 51 / 3 C, June 13, 1997. Of course,
ECOWAS played a more prominent and expensive role than the UN in the war as
well as afterward. One scholar estimates that Nigeria spent more than $4 billion on
ECOMOG alone in the 1990s (Coleman 2007, 85); however, Western states’ defer-
ence to the African Union (rather than the UN) itself reflects the low priority given
Liberia’s peace process by those states.
24. See UN Security Council Resolution 1343, March 7, 2001, and UN Security
Council Resolution 1521, December 22, 2003. UNSC Resolution 1343 was revoked
on November 20, 2008, by UN Sanctions Regulations Amendment SR 2008 / 394.
4
Separatist Recurrences of
Civil War

THE PREVIOUS CHAPTER USED a full-fledged case study to examine Liberia’s

recurrence of civil war after an apparently successful UN peace operation. These
types of cases share other characteristics that do not necessarily hold for other
postconflict societies. Not only did they experience renewed UN peacekeeping
operations after apparent peacebuilding success but they also all experienced sus-
tained international military interventions and extensive civilian state-building
efforts after both their episodes of armed conflicts. They are all poor, weakly
institutionalized societies where local and / or tribal authorities wield important
power outside the reach or presence of the state. Despite the particularities of
these three cases, is it possible that the common elements of their shared experi-
ences of reversion to armed conflict might shed light on the broader phenom-
enon of the failure of peace? What do these cases suggest about the factors that
lead wars to reignite? And how well do these hypothesized factors hold up to
empirical scrutiny?
This chapter moves to the third research method utilized in this book: brief,
tightly focused, structured case studies. In chapter 1 the strengths, weaknesses,
and findings of quantitative methods are first examined. That chapter shows the
need for a more contextualized analysis of variables that are inaccessible to quan-
titative analysis, at least given constraints on the time and resources needed to
develop relevant data.
Chapter 2 sets forth crucial concepts for theorizing the recurrence of war.
And in chapter 3 I analyze the strengths, weaknesses, and findings of a focused
and somewhat in-depth case study of Liberia’s second civil war. That method
offers richer understandings of both the process by which certain variables act
causally and the interaction of these variables. However, a case study or two, even
including within-country comparisons with more recent postwar experiences,
such as that in Liberia after 2003, do not permit robust or reliable findings.
The next three chapters, therefore, extend the cases examined beyond Liberia
to the other fourteen cases among the entire pool of fifteen civil war recurrences
Separatist Recurrences of Civil War 97

defined in the introductory chapter. It does so by focusing on the hypothesized

factors generated by the existing literature and illustrated through the case study
of Liberia. Chief among these is the exclusionary behavior exhibited by post-
war elected civilian governments, in its various manifestations. In addition the
case studies examine regional-level factors, external operations, and their impact,
along with weak state institutions. What is the role of neighboring states and ter-
ritories, as either havens for insurgents or supporters of their suppression? Did
external actors play a decisive role in recurrence? Are institutions weak, and do
weak institutions seem to play a role?
These “mini–case studies” also explicitly take up variables identified in the
quantitative literature and segment of this book: ethnic and religious fractional-
ization, the presence and departure of external peacekeeping troops, and natural
resources and the economic opportunities they provide. The method applied to
these minicases therefore constitutes what Harry Eckstein (1975) called a “plau-
sibility probe,” designed to explore the extent to which hypothesized variables
usually drawn from one set of cases hold causal weight in another set of cases. In
addition these minicases serve a second, heuristic function, because the studies
will highlight factors that seem to have been important to case specialists in their
analysis of particular recurrences of internal armed conflict.
The cases explored in this chapter are drawn from the list of core recurrences
given in table 4.1. These cases’ selection criteria are explained in the introduc-
tion, but their eclectic character bears recapping. They represent an amalgam of
countries with disparate types of armed conflicts, but they reflect reliance on oth-
ers’ categorizations in an effort not to skew the findings. Thus countries with less
lethal recurrences of internal armed conflict (e.g., Haiti, East Timor, the Central
African Republic, and Peru) are included alongside countries with very deadly
repeat civil wars (e.g., Chechnya, Tibet, Liberia, Lebanon, and Sudan). Fearon
and Laitin’s list also includes countries with recurrent internal armed conflicts
where one-sided violence or genocide constituted the main portion of the initial
war (viz., Haiti, Burundi, and Rwanda).
My own judgments have entered through the exclusion of certain cases (e.g.,
Sri Lanka, Somalia, and Senegal) because of the absence of a broad perception
within those societies, and among analysts, that war ever ended.1 Beyond reli-
ance on secondary sources (especially Fearon and Laitin 2003; Sambanis 2006)
for identifying the parties for each conflict, I have also included civil wars in
three countries where separate conflicts existed—China in Tibet, South Os-
setia in Georgia, and the North / South conflicts within the Sudan. I excluded
other countries like Indonesia, Pakistan, and Iraq, where multiple internal wars
rendered it impossible to specify that there was a civil war recurrence, because
internal wars overlapped either temporally or spatially (i.e., through overlapping
territories or interlinked conflict recurrences) or involved different actors.
98 Examining the Cases

Table 4.1 Fifteen Core Cases of Civil War Recurrence

Burundi, 1972, 1988, 1993–2006
Central African Republic, 1996–97, 2001–2
China, 1950–51, 1956–59
East Timor (Indonesia), 1975–99, 2006
Georgia / South Ossetia, 1992–94, 2004
Haiti, 1991–95, 2004
Lebanon, 1958, 1975–90
Liberia, 1989–96, 1999–2003
Mali, 1989–94, 2007
Nicaragua, 1978–79, 1981–90
Peru, 1981–1995, 2007
Russia, 1994–96, 1999–present
Rwanda, 1962–65, 1990–2002
Sudan, 1955–72, 1985–2002
Zimbabwe, 1972–79, 1983–87

Before embarking on this and the next two chapters on the recurrent cases of
armed conflict, I summarize the six main findings presented in these three chap-
ters, as drawn from the fifteen total cases. First, precipitating exclusionary behavior
is the most evident and cited proximate cause for recurrence in most cases—in nine
of fifteen cases, or 60 percent. Recall that precipitating exclusionary behavior goes
beyond a static condition of exclusion. In these nine cases postwar states adopted
a policy that violated the expectations of the former rebels or their followers and
supporters, and this exclusionary conduct served as the principal precipitating
factor for the recurrence of civil war or internal armed conflict. These nine cases
are the Central African Republic, China / Tibet, East Timor, Georgia, Haiti, Li-
beria, Russia / Chechnya, Sudan, and Zimbabwe.
Four of these cases—those based on secessionist struggles—are the subject
of this chapter. In an additional four cases—Haiti, the Central African Republic,
East Timor, and Zimbabwe—political exclusion was a driving factor in the re-
currence of armed conflict. These cases, whose recurrences were not rooted in se-
cessionist aspirations (even where the original war may have been thus rooted, as
in East Timor), are addressed in chapter 5. Liberia also fits this pattern of exclu-
sion and thus serves as a key trigger for the recurrence of internal armed conflict.
Second, exclusion by the state against its former wartime allies constitutes a sepa-
rate subset of war recurrences. In some cases where a coalition of insurgent armies
emerges victorious, exclusionary behavior by one victorious actor against a war-
time ally constitutes the principal precipitating factor in a relapse into armed
conflict. East Timor and Zimbabwe present cases of triumphant liberation move-
ments whose leaders, once in power, acted in an exclusionary manner toward
their former allies, which sparked renewed armed conflict.
Separatist Recurrences of Civil War 99

Third, a violation of powersharing arrangements was the trigger for war recurrence
in 40 percent of the core cases of recurrence. Precipitating exclusionary conduct may
rest on formal agreements or on unwritten, informal arrangements that create
certain expectations (as in Haiti, Burundi, and East Timor). In six of the fifteen
cases of recurrence—Sudan, Russia (in Chechnya), Georgia (in South Ossetia),
Liberia, the Central African Republic, and Zimbabwe—the state reneged on an
explicit, signed powersharing agreement, either to cede central power over a ter-
ritory or to share institutional power in political offices or military / police posts.
In Chechnya the role of reneging on a powersharing arrangement was an impor-
tant factor, but not the sole trigger.
Fourth, severe, chronic exclusionary conduct, though insufficient to postulate as
a correlated or causal variable, played a role in two war recurrences. In these two
additional cases of Burundi and Rwanda, exclusionary behavior was not the pre-
cipitating factor but is cited by virtually all analysts as a crucial causal factor. In
these interactive recurrences in adjacent, demographically similar states, severe
exclusionary behavior served as the constant backdrop to the eruption of resis-
tance and its harsh repression, which cost hundreds of thousands of lives. These
state policies, following earlier mass violence that had been deemed (perhaps in-
correctly) a civil war, reflect the most severe exclusionary behavior for the cases
observed. I refer to this as severe, “chronic exclusion,” a term that identifies a
risk factor for conflict recurrence but inadequately specifies whether (or when or
how) armed conflict will recur in a given society. The concept of chronic, severe
exclusion is thus useful to understand a subset of recurrent internal armed con-
flicts but not as a trigger of war recurrence. The cases of severe, chronic exclusion
in Burundi and Rwanda are examined in chapter 5.
Fifth, exclusionary conduct is not the only salient factor in these recurrences; how-
ever, it is the most consistently important one. Natural resources and economic fac-
tors were not a primary causal factor in any of the fifteen core cases except Peru
(which is analyzed in chapter 6, along with other nonconforming cases), where
the reemergence of the Shining Path insurgency coincided with its new role not
just in taxing cultivators of coca but also in its trafficking. However, natural re-
sources acted as an interrelated, secondary factor in four other cases of recur-
rence. Thus recurrence was influenced by the interaction of exclusion with oil
production or transportation in the Sudan and Chechnya; and with minerals in
Liberia and the Central African Republic.
Except for four states—Russia, Peru, Lebanon, and Georgia—the core cases
were predominantly poor. These cases also score high on the various indicators
of state institutional weakness. Haiti, East Timor, the Central African Republic,
Sudan, Zimbabwe, and Mali rank among the world’s least institutionalized states.
The qualitative research here indicates that these two factors identified by the
economy-centered literature on civil wars are relevant risk factors. Poverty and
100 Examining the Cases

state weakness enhance the risk of recurrence, but recurrence is most often trig-
gered by state exclusionary behavior.
Sixth, and most important, the points noted above indicate that exclusionary be-
havior was the most important causal factor in eleven of the fifteen cases, or 73 percent,
of recurrence of armed conflicts. Postwar states engaged in exclusionary conduct in
thirteen of the fifteen cases, or 87 percent, although it was not the main cause in
two of these. Again, these six central findings are presented in this and the next
two chapters. I now turn to the first subset of cases whose recurrences reflected
separatist aspirations.
As in Liberia, exclusion by postwar governments served as a trigger for re-
newed conflict in eight other core cases. Exclusion took various forms. I begin by
analyzing the cases of powersharing over separatist territories. Rather than offer
key analytic points prematurely, here I present four mini–case studies: Sudan’s re-
current civil war (1955–72 and 1983–2005), Russia’s two civil wars in Chechnya
(1994–96 and 1999–), South Ossetia’s recurrent armed conflict against Georgia
(1992–94 and 2004; the 2008 war and the Russian invasion were too recent
for this study), and China’s two wars in and over Tibet’s autonomy (1950–51
and 1956–60). Subsequently, I analyze the findings from all four cases.

Sudan in 1955–72 and 1983–2005: The Marginalization of the South

Despite multiple conflicts within the territory, Sudan’s two North / South wars
are included here because commonly accepted analysis permits separate analy-
sis of the recurrent conflicts, despite the related but distinct conflicts elsewhere
(e.g., Darfur). In contrast to the civil wars in Uganda, Somalia, and Indonesia,
Sudan’s conflict from the 1980s onward is often referred to as the “second” civil
war. Major data sets score two distinct wars with a period of interwar peace from
1972 to 1983.2 Some data sets score the origins of the first war in 1963 because
the formation of coherent guerrilla forces occurred in that year. However, several
book-length treatments of Sudan’s wars (Poggo 2009; O’Ballance 2000; Assefa
1987) score the first war as commencing in 1955 or 1956. Although the start-
ing date of the initial war is contested, the ending date (1972) and the starting
and ending dates of the second civil war (1983–2005) reflect a relatively broad
consensus.
Analysts are virtually unanimous in emphasizing the exclusionary behavior
of the government in explaining why civil war recurred in the Sudan. The dis-
covery of oil was one factor in the recurrence, but not as important as exclusion
vis-à-vis land, Sharia law, and political representation in the early 1980s. The
first war also involved a falling out among the national factions that struggled for
independence from Britain, and it thus resembled (imperfectly) the dynamics of
the postwar recurrences in East Timor and Zimbabwe.
Separatist Recurrences of Civil War 101

Sudan is the largest country in Africa, and its northern part is occupied by
the Sahara Desert. A British census in the aftermath of independence in January
1956 showed that 55 percent of the population claimed African descent, and
39 percent claimed Arab ancestry. Most of those African Sudanese lived in the
southern region, but almost half (45 percent) of those African Sudanese lived
with their Arab compatriots in the eastern and other non-southern regions
(Poggo 2009, 15). Because Sudanese “Arabs” are generally descended from a
mixture of Arab, Nubian, and black African races, the term is more cultural than
racial (Suliman 1997, 100). In the nineteenth century, the slavery of some 2 mil-
lion black southerners mainly at the hands of Arabs and Western colonists led to
the assimilation of darker-skinned peoples into Arab ways, Islam, and intermar-
riage (Anderson 1999, 65). Thus the rebel leader John Garang was fond of say-
ing that more Africans lived in the supposedly Arab-dominated North of Sudan
than in the South (Poggo 2009, 19). Ethnic distinctions also characterized the
South, where the most numerous black African tribes were the Dinka and the
Nuer (Poggo 2009, 13).

The First Civil War, 1955–72

Sudanese history is replete with tensions between northern peoples more con-
nected to world trade through the Red Sea and Egypt and the less developed,
more isolated South. The British and Egyptians invaded and exercised control
over Sudan after 1898, and in 1902 they adopted a policy to treat the North and
South separately (Assefa 1987, 45). This meant banning Muslim traders and oth-
ers from entering the South, encouraging the adoption of the English language
rather than the Arabic language, revitalizing tribal law and customs in the South
while discouraging Arab names and customs, and even banning the sale of Arab
clothing (Assefa 1987, 45). To illustrate the differential approach to state institu-
tions, only thirty officials from Britain served in the entire South during its first
fifty years of occupation, and they made no effort to include southerners in train-
ing as administrators or into the police or military colleges (Poggo 2009, 23–24).
After World War II, the British reversed this policy of separation, began to de-
fer more to the North in governance questions about the South, and entertained
the notion of a unified Sudan that would receive independence (Assefa 1987,
45–48). As independence approached in the 1950s, Egypt in 1951 abrogated
the Condominium Agreement governing its and the British role in the South,
but it failed in its bid to establish sovereignty over the Sudan. In the process,
however, the South became organized and created the Southern Party in 1951. It
is an ahistoric oversimplification to reduce the identity lines of Sudan’s conflict
to Arab / nonblack / Muslim versus African / black / Christian. Nevertheless, the
mobilization of both civil wars has tended to occur mainly along these lines.
102 Examining the Cases

The roots of the first civil war center on the aspirations for self-government
of the South and perceived northern resistance to this end. This resistance pre-
dated and anticipated Sudan’s independence on January 1, 1956. Northern po-
litical parties, in an effort to preemptively occupy the state posts and political
space that independence would afford, became expansive and more absolute
in their expressions of Sudanese unity (Poggo 2009, 35). Because they were
wrapped up in their negotiations with the colonial occupiers of Egypt and Brit-
ain, the northern parties neglected southern actors. The southerners increas-
ingly saw that their marginal role in decision making during this interim period
foreshadowed their continued exclusion (Poggo 2009, 34). They increasingly
expressed their frustration and eventually withdrew from the transitional ar-
rangements. Thus the dynamics in the years 1954–56 in Sudan reflect a falling
out among the main domestic protagonists in a process of the emergence of a
new country.
On August 18, 1955, soldiers in the town of Torit in Sudan’s southernmost
province of Equatoria received long-rumored orders to report for transporta-
tion to northern postings, which allegedly were aimed at weakening and dividing
southern soldiers within the national army. They attacked their northern officers
and stole arms and ammunition, and the soldiers in other southern towns soon
followed suit, which led to roughly four hundred deaths, mostly of northerners
(Poggo 2009, 42). The mutiny marks the onset of the civil war, which continued
in the following months and accelerated partly in response to severe repression
carried out by the North (Poggo 2009, 52–53).
After the 1955 mutiny, a Commission of Inquiry found that “the Southern
Army revolt was the final outburst after the Southern political claims had been
frustrated” (quoted by Poggo 2009, 56). In a recent detailed study of the origins
of the first war, Poggo (2009) concludes that the 1955 mutiny of the Equatoria
Corps “was the culmination of frustration after many years of discrimination”
against southerners, grievances “compounded by the Sudanese government’s
refusal to grant federal status to the South” (p. 49). She adds that southern de-
mands for a federated system (not even independence) were “totally ignored” by
the northern political leadership (p. 50). Northerners’ exclusion of southerners
continued in the late 1950s: in discrimination in state posts, including their total
exclusion from the Foreign Ministry; in the 1956 formation of a forty-six-person
constitutional drafting commission that contained only three southerners; and
in the composition of the legislature, where southerners chafed at their degree
of underrepresentation (pp. 52–58). Moreover, the repression of any seeming
dissent further deepened southerners’ marginalization and the civil war.
The rebel bands remained inchoate and small between 1955 and 1963, which
partly accounts for the starting dates of the war in the 1960s. The Sudan African
National Union (SANU) was formed in 1962 in Kinshasa, but it was viewed by
Separatist Recurrences of Civil War 103

the rebels as a futile diplomatic forum (O’Ballance 2000, 17–19). But in 1963
the rebels organized, adopted the name the Anya-Nya, and launched more at-
tacks in the mid-1960s while using the Republic of the Congo and Uganda as
refuges (O’Ballance 2000, 21–23). By the mid-1960s, SANU had split to form
new political groups, some tied to the Anya-Nya, that by now demanded inde-
pendence rather than federation.
In contrast to other cases, Sudan’s religious and ethnic dimensions seem more
intertwined, because marginalization and mobilization have occurred along both
lines, and it is impossible to dismiss one or the other. Furthermore, rather than
an entirely Christian composition, the rebels of the South included numerous
Muslims (Iyob and Khadiagala 2006, 32). The parties to conflict utilized these
terms—“Arab,” “African”—explicitly to mobilize greater adherents. Anticolonial
nationalists reappropriated the term “Arab” in the 1920s in resistance to British
colonial discourse. Only later did some adopt “African” as a gesture against their
disenfranchisement under the “hegemonic rule of their Arabized compatriots”
(Iyob and Khadiagala 2006, 31).
External actors also played an important role in this first war, beyond the pre-
independence policies of Egypt and Britain. Between 1960 and 1972, the Soviet
Union, Great Britain, various Arab countries, East Germany, Yugoslavia, and
China all supplied the Sudanese government with sophisticated arms and aircraft
(Poggo 2009, 1). The southern rebels were greatly strengthened in resources
and arms by the Israeli government, which passed on arms captured in the Six-
Day War and sought to undermine the Arab-dominated Sudanese government
(Poggo 2009, 2).

The Peace
Colonel Gaafar Muhammad Nimeiri had come to power initially as a socialist
and secularist, but after a Communist coup was turned back in 1971, he moved
toward the West and away from radical change. The United States responded
warmly to this shift and gave $18 million for refugee resettlement in support of
the peace accord (Anderson 1999, 17). Partly out of a desire to bolster his sup-
port in the South, Nimeiri also sought to negotiate an end to the seventeen-year
civil war. In 1972, the Addis Ababa agreement was signed between the Nimeiri
government and the now more unified political-military insurgency, the South
Sudan Liberation Movement. A cease-fire went into effect with general success
that year. The negotiations were facilitated by the agreement of the Ethiopians
to cease supporting the rebels in exchange for Nimeiri’s commitment to cease
support for the Eritrean rebels. The negotiations process was formally chaired
by the emperor of Ethiopia and was mediated by Liberian canon Burgess Carr
of the All-African Council of Churches (Mitchell 1989, 8; Assefa 1987, 131–43).
104 Examining the Cases

The peace agreement provided for considerable autonomy for the South un-
der a regional government with its own high executive council and a regional as-
sembly—all enacted into law. The regional entity would have legislative author-
ity over education, health, public security, and natural resources (e.g., oil), but
not defense, currency, or foreign affairs (Daly 1993, 19). Mechanisms were also
created to help prevent and resolve tensions between the central government
(where southerners continued to have representation) and the new regional
bodies. Arabic continued as the official language of the whole country, but En-
glish became the “common language” of the South and was taught in schools
(Mitchell 1989, 8). After the most difficult negotiations (Assefa 1987, 131–34),
the parties agreed on the incorporation of an estimated 12,000 rebels into the
state’s armed forces and police. Mitchell (1989, 8) reports: “This force would,
for a transitional term, be under the command of a commission of northerners
and southerners, until the south had set up its own machinery for maintaining
law and order, which was to consist of an armed police force and between 2,000
and 3,000 frontier guards.”
The agreement retained the unity of the country while offering the prom-
ise of substantial autonomy and security guarantees. It is generally considered a
successful peace negotiation, and a crucial factor in both the eleven-year inter-
war peace and the recovery of some stability for the Nimeiri regime (Mitchell
1989, 9).

The Second Civil War, 1983–2005

Unfortunately, the agreement was only partially implemented, because Nimeiri
increasingly flouted the accord and marginalized southerners. Sudan’s second
civil war occurred principally because of a perceived reneging on the promise of
autonomy enjoyed as the status quo during the mid-1970s, which was embodied
mainly but not exclusively in the 1972 peace accord.
Central to the dynamics of politics within the North of the Sudan were ten-
sions between secularists and Islamists. In 1976 an Islamist movement led by
Sadiq al-Mahdi failed in a coup attempt. In contrast to the reaction away from the
left provoked by the 1971 failed coup, this experience demonstrated to Nimeiri
the power of the Islamists (Daly 1993, 20). He moved to accommodate them as
part of a six-year process begun in 1977 known as “National Reconciliation.” In
that year he created a commission to reform the Constitution to bring it and law
into line with Islam, and in 1983 he declared Sharia to be the law throughout the
country (Daly 1993, 20). As Daly (1993, 21) states, “While the degree of regional
autonomy achieved did not approximate the federalism many had long favored,
the powers devolved at Addis Ababa allowed protection of the South’s special
character—or would have if the agreement had been honored by Nimayri and
Separatist Recurrences of Civil War 105

the Southern politicians themselves.” Related to these actions to roll back the
autonomy that southerners expected in their public security and legal systems,
not to mention religious and cultural life, Nimeiri reneged on other elements of
the accord and thus redrew the South’s internal divisions.
The Anya-Nya II, led by some of the same commanders, joined with other
mutineers in 1983 to form the Sudan People’s Liberation Army, which was con-
nected to its political arm, the Sudan People’s Liberation Movement ( Johnson
and Prunier 1993, 125). Colonel John Garang, who had fought with the rebels
in the first civil war, was sent as an army officer to negotiate with rebels in 1983
when he defected to join them ( Johnson and Prunier 1993, 124).
Overall, academic treatments of the Sudanese recurrence of civil war empha-
size the exclusionary behavior of the Sudanese government. Mitchell (1989, 2),
for instance, says that the Addis Ababa agreement provided the basis for peace
between North and South “until the terms and the spirit of the Agreement were
unilaterally undermined by Nimeiry, one of its main architects, and war broke
out again.” Johnson and Prunier (1993, 124) say that “Nimayri prepared for the
abolition of the Southern region by arresting pro-unionist Southern politicians
and transferring Southern troops to the North.” And Wakoson (1993, 29) says
that “the Northern ruling elite have consistently ignored Southerners’ demand
for federalism. The result of this intransigence has been the continual constitu-
tional crisis and civil wars dominating Sudanese national politics since 1955.”
Although many analysts of the war recurrence stress the North’s efforts to gar-
ner resources from the South, none place economic factors or natural resources
at the center of their explanation. However, natural resources, especially the dis-
covery of larger oil reserves in the 1980s and 1990s mainly in the South, are a
relevant factor in renewed warfare. Suliman’s (1997) neo-Marxist analysis explic-
itly argues that the ethnoreligious divisions that are emphasized by virtually all
analysts of the second civil war are overstated and that economic factors play an
important role, especially as the war wore on. He notes that Chevron announced
the discovery of new oil reserves in the southwest in 1981, amounting to 12 per-
cent of the estimated national reserves of 2 billion barrels at the time. Nimeiri
angered southern elites in 1982 when he announced that the country’s first re-
finery would be constructed south of Port Sudan rather than closer to the South
(Suliman 1997, 115). Among the earliest attacks of the second war in 1984 were
on oil fields in the South, which disrupted production there for several years.
Even so, Suliman (1997, 110–15) and Daly (1993, 21–23) emphasize the ef-
fects of economic crisis and export-led economic and agricultural policies rather
than oil, which receives only a tertiary mention. Indeed, the case lends support
to the thesis, contrary to Collier and his colleagues, that negative World Bank
and International Monetary Fund policies proved more important than any par-
ticular natural resource in facilitating the second war. Those policies undercut
106 Examining the Cases

state subsidies, aggravated drought-related economic crisis in the 1970s and

1980s, and caused migration and resentment (Suliman 1997, 111–13; Daly 1993,
21–22). Thus oil receives scant mention from O’Ballance (whose 2000 book
mentions oil on only two pages), Daly (whose 1993 chapter on the political and
economic roots of the renewed war mentions oil as only a rising cost factor in the
economic crisis); and Alex de Waal (who does not mention it in his 1993 chapter
on the rise of militias in the 1980s).
Sudan seems to confirm some of the more nuanced and recent versions of the
economic opportunities literature, as access to economic livelihoods (broader
than simply natural resources like oil) interacts with identity claims to account for
the mobilization of the second war in 1983. As Iyob and Khadiagala (2006, 29)
say: “It is the conflicts over resources—agricultural land, water, pasturage, and,
recently, oil—that underpin the attenuated rivalries that followed, leaving behind
new hostilities that were . . . woven together to become the poles of identity /
ethnicity around which the contesting groups rallied.”
International factors were important, but again secondary. Ethiopia and Su-
dan’s other neighbors were used as refuges for the insurgents, and both sides
received arms from outside sources. Nevertheless, international actors played
a less prominent role in the 1983 recurrence than they played in the earlier civil
war and its termination.

Chechyna in 1994–96 and 1999–Present: Reneging and Resistance

The separatist war that emerged in Chechnya in 1994 is similar to the other
more successful separatist struggles that emerged in the wake of the breakup of
the Soviet Union. These conflicts all exhibited long-simmering resentment of
Russian-dominated rule from Moscow; resistance to decades of ethnic marginal-
ization, including the incursion of ethnic Russians who enjoyed more privileges;
and a desire for greater control over economic policy and natural resources.
Georgia, Azerbaijan, Armenia, Moldova, and Chechnya (and neighboring In-
gushetia and Dagestan) all lie in the mountains of the Caucasus, affirming one of
the risk factors for the onset of civil war. In an earlier act of Soviet repression that
instilled resentment, Chechens had been forcibly deported to Siberia amid an
earlier uprising during World War II. At the time of the first Chechen war, 73 per-
cent of Chechnya’s residents were ethnic Chechens, whereas about 26 percent
were ethnic Russians (Hill 1995).

The First Chechen War

In late 1994 Russian troops invaded Chechnya to put down the separatist move-
ment. Baev, Koehler, and Zurcher (2005) cite three reasons for Russia’s invasion
Separatist Recurrences of Civil War 107

at this time: Russian political elites hoped that a military victory would give them
political capital, they were concerned about organized crime, and they were wor-
ried about the precedent that Chechen secession might set.3 A Russian missile at-
tack killed Chechen president Dudayev in April 1996, bringing together the sep-
aratist forces. They recaptured the capital on August 6, 1996. That same month,
Russia’s popular general Alexsandr Lebed and then–general Aslan Maskhadov
signed the Khasavyurt Accords, providing for a cease-fire and the withdrawal of
Russian troops. A tenuous period of peace ensued.
Chechnya’s war recurrence in 1999 may seem like it does not meet the stan-
dard of reneging on the part of the state. Kremlinologists, however, viewed the
war in Chechnya as not just potentially destabilizing in Russia but also one
that had a major political impact, sealing the fate of one president and opening
the way for another. In the words of Lieven (1998, 102), “The first week of the
Chechen War marked one of the most critical moments in the history of the
Yeltsin administration and indeed of modern Russia. . . . This is because, for a
few days, there seemed to be a real possibility that the unity of the Russian army
would crack, and with it the obedience of the junior commanders to the Defence
Ministry and the military hierarchy.”
The first war “very nearly led to Boris Yeltsin’s impeachment” (Lieven 1998,
102). The second civil war was also intimately tied to regime stability in Moscow.
Once Vladimir Putin became president upon Yeltsin’s resignation three months
before the March 2000 elections, “renewal of the war against Chechnya . . . se-
cured Putin’s victory” (Evangelista 2002, 6). Putin’s career rose with the specter
of renewed warfare in Chechnya.

The Khasavyurt Accord, 1996

The expectations surrounding the end of the first Chechen war were set prin-
cipally in the Khasavyurt Accords, which were reaffirmed in a treaty-signing
ceremony by Boris Yeltsin and newly elected Chechen president Maskhadov in
May 1997. The agreement deferred a decision on the separatist republic’s status
until 2001, but it explicitly eschewed both the use and the threatened use of
armed forces in resolving the issues (Khasavyurt Accord, August 30, 1996). The
two parties created a joint commission to monitor Russia’s troop withdrawal, to
draw up proposals for financial and budgetary relations, to confront crime inside
Chechnya, and to make economic reconstruction plans. These plans implied an
acceptance of Chechnya’s autonomy in its budget, finances, and internal security.
Similar to the 1991–94 period, the republic retained de facto independence. The
Russian troops’ withdrawal was interpreted by Chechens as a “sign of capitula-
tion and endorsement of the republic’s de facto independence” (German 2003,
147–48).
108 Examining the Cases

A number of signs point to the Russian acceptance of Chechnya’s indepen-

dence. First, international law professor Francis Boyle sees the treaty itself as
recognizing the republic’s sovereignty. The title of the Khasavyurt agreement
itself is “Treaty on Peace and the Principles of Interrelations between the Rus-
sian Federation and the Chechen Republic of Ichkeria” (Cornell 2001, 244).
In addition, the parties adopted the term “treaty,” a term usually reserved for
sovereign states, rather than “accord,” “agreement,” “compact,” or another term
that would be more common for settlements of civil wars. During the signing
of the peace treaty, President Yeltsin even used the name “Ichkeria,” which had
been embraced as the homeland moniker by Chechens but previously rejected
by the Russian government (Cornell 2001, 243). Conversely, the resolution of
Chechnya’s formal status was clearly postponed. And given that no foreign coun-
try was likely to recognize Chechen independence without Moscow’s assent, the
Russian government knew that Chechnya’s ultimate status lay in its hands.

The Second Chechen War, 1999

In contrast to an argument emphasizing Russian exclusionary behavior, Chech-
ens took steps that also undermined the peace. Maskhadov’s government was
unable to extend his authority over the territory in the face of the power of key
warlords. In April 1998 the insurgent leader Shamil Basaev and Arab-born Ibn
Al-Khattab declared their aim of unifying the Russian republics of Chechnya
and adjacent Dagestan into an Islamist state. In August 1999 they led several at-
tacks into Dagestan that provoked an immediate military response from Russian
forces. After several bombings of civilian targets left about three hundred people
dead in Moscow in September, Russian air power began a bombing campaign
that was shortly followed by a land invasion that eventually took the Chechen
capital of Grozny.
Many observers saw these events as a response to Chechen initiatives, and
Chechen actors certainly exercised agency for the recurrence of war. In particular
Maskhadov was unable to exercise control over the state and territory. Radical ri-
vals began a new wave of kidnappings for ransom (German 2003, 148). The Rus-
sian general appointed to represent the Interior Ministry in Chechnya was kid-
napped in Grozny in March 1999 after Maskhadov had given guarantees for his
security. Competing warlords who emerged with resources, arms, followers, and
legitimacy from the first war persisted in their drive to advance autonomy and,
as seen above, Islamicism. Domestic pressure from Islamists forced Maskhadov
to adopt Sharia law, even to dissolve Parliament on the grounds that its presence
violated Islamic law (Evangelista 2002, 56).
However, most analysts see Russian policies and choices as underlying, or at
least facilitating, the weakness of Maskhadov’s rule and even the threat posed by
Separatist Recurrences of Civil War 109

spoiler commanders (Lieven 1998, 140–46; Cornell 2001, 240–50). Moscow

reneged on its commitments to provide reconstruction support after the 1996
peace accord (German 2003, 148; Evengelista 2002, 56). In late 1998 Yeltsin an-
nulled a 1997 directive to negotiate a treaty with Chechnya on the mutual dele-
gation of powers. As German (2003, 148) put it, Moscow engaged in “prevarica-
tion on Chechnya’s status,” which undermined Maskhadov’s ability to maintain
control of the territory in the face of rivals. In the face of rising terrorist attacks
in and from Chechnya in early 1999, Yeltsin failed to respond to warnings from
General Lebed about the worsening situation.
More pointedly, the Russian government developed and was advancing plans
to invade Chechnya in early 1999, long before the attacks of August and Septem-
ber. Six months into the second war, Sergei Stepashin, who was Russia’s interior
minister and then prime minister by July 1999, said, “We were planning to reach
the Terek River [dividing flat northern portion of Chechnya from the mountain-
ous south] in August or September. So this was going to happen, even if there
had been no explosions in Moscow. I was working actively on tightening borders
with Chechnya, preparing for an active offensive” (quoted by Hoffman 2000).
Putin, who had formerly been head of internal intelligence and was appointed
prime minister in August 1999, weeks before the Dagestan attacks and the recur-
rence of war, only accelerated plans to invade (Hoffman 2000). Baev (2005)
emphasizes the desire to redeem the Russian military after its defeat in 1996 and
to reinvigorate its role in society. Although Chechen division and dysfunction
had a hand, the Russian role is evident in Evangelista’s (2002, 2) question: “How
could Russia’s leaders have steered their country into such destructive and seem-
ingly self-defeating wars, at a cost of tens of thousands of lives, mainly Russian?”

Alternative Explanations
Other factors played at most a secondary role. Although state weakness as a factor
is generally assessed with regard to the central state (in which an internal conflict
occurs), the weakness of Chechen pseudo–state institutions by 1999 played a
role in conflict recurrence. Putin and Stepashin (Hoffman 2000) explicitly cited
the inability of Maskhadov’s government, which enjoyed de facto autonomy, to
control dissident factions of the prior insurgent movement by 1999. Although oil
has played a role in the duration of the second conflict and in the intensification
of warlord violence and profiteering, Chechnya’s oil only makes up about 1 per-
cent of the oil in Russia (Baev, Koehler, and Zurcher 2005; Shermatova 2003).
Therefore, it does not seem likely that oil alone has been enough of an incentive
for Russia to continue the conflict and oppose Chechen independence.
International factors played a less prominent role here than in other cases.
One unusual precedent is cited as contributing to Russia’s proclivity to invade
110 Examining the Cases

Chechnya after the Dagestan raids and the Moscow bombings (allegedly by
Chechen separatists, though the authorship remains murky): NATO’s bomb-
ing campaign in Kosovo from March to June 1999. That bombing occurred
despite Russian objections and without UN Security Council consent. More-
over, it showed the value of low-risk, high-altitude aerial bombardment, which
inflicted high enemy military and bystander casualties while subjecting airmen
and ground soldiers to virtually no risk. Consequently the NATO campaign em-
boldened the Russians to invade on political grounds and on tactical grounds,
which increased civilian casualties while reducing Russian troop exposure until
after the aerial bombardment had weakened Grozny tremendously.

Georgia and South Ossetia in 1991–92

and 2004: Integration Backfires
Georgia is one of three Soviet socialist republics that successfully sought inde-
pendence upon the breakup of the USSR, only to experience “secondary” inter-
nal secessionist armed movements. The other cases are Moldova (i.e., the Trans-
nistria secessionist struggle) and Azerbaijan (Nogorno-Karabakh). Secessionist
armed conflicts in Chechnya, North Ossetia, Dagestan, and Ingushetia remain
internal to Russia. Aside from their particular features, the original onset of these
secondary conflicts related to the uncertainty created by the independence pro-
cess and the breakup of the Soviet Union.4
The case of Georgia is complex. It involved three armed conflicts in the early
1990s. The first, a struggle for power in 1991 between the then–prime minister
and political opponents, resulted in a coup that installed Eduard Schevardnadze
as president and was not secessionist in nature. The other two armed conflicts
were secessionist struggles in Abkhazia and South Ossetia. Fearon and Laitin
code a single civil war, from 1992 to 1994. For clarity in this analysis, I use Sam-
banis’s (2004) coding of the South Ossetia conflict of 1991–92. The Abkhaz
conflict took an estimated 2,000 lives and drew more international attention to
the initial war in South Ossetia, which cost an estimated 1,100 lives and displaced
100,000 to 150,000 people. The Abkhaz war is not analyzed here because its peri-
ods of peace were not sufficiently long to score a full year of peace. The Georgia–
Russia war of 2008, which involved internal combat forces of both Abkhazia and
South Ossetia, falls outside the time line of analysis here, and in any case it would
be classified as primarily an interstate war.
South Ossetia’s recurrence occurred in the summer of 2004, and it appears
here because the Peace Research Institute Oslo (PRIO) codes it as an internal
armed conflict despite the fact that it resulted in only twenty-seven deaths.5 The
recurrence barely counts as an internal armed conflict, and there may have been
comparable recurrences in Moldova, Azerbaijan, or the Central African Republic
Separatist Recurrences of Civil War 111

that were not recorded as reaching twenty-five deaths in a single year. As with
China’s renewed warfare in Tibet, the Georgian case could have been counted
as a case of multiple and simultaneous internal armed conflicts and thus would
have been omitted from the analysis. Again, however, it is included here out of a
commitment to seek to include those cases from Fearon and Laitin where pos-
sible, and due to the ability to analyze these two wars in South Ossetia in relative
geographic isolation from Abkhazia.

The First South Ossetia War, 1991–93

In the midst of glasnost, the South Ossetian nationalists created the South Os-
setian Popular Front in 1988 to seek to withdraw from Georgian control. In Sep-
tember 1990, after the Georgian Supreme Soviet adopted an election law barring
regional parties, the South Ossetian Supreme Soviet responded by declaring its
sovereignty, still within the USSR.
The move was rejected by Georgia. After South Ossetia elected its own Par-
liament in December 1990, Georgia withdrew the region’s “autonomous” sta-
tus, and hostilities erupted in January 1991. König (2005, 241) and the Inter-
national Crisis Group (2004) estimate that a thousand died in this first South
Ossetian conflict, although Tishkov (1999) estimated 1,100 dead with 150,000
displaced. Georgian troops blockaded the South Ossetian capital, Tskhinvali, a
month later, and Soviet troops deployed in the region were openly on the side
of the South Ossetians. After Georgia declared its independence in April 1991
and the USSR was formally dissolved in December, Soviet troops withdrew in
the spring of 1992.

The 1992 Sochi Agreement

In June 1992 the Organization for Security and Cooperation in Europe (OSCE)
assisted in negotiating a cease-fire with South Ossetian forces, called the Sochi
Agreement. After Georgia’s first elected government fell in 1992, new president
Shevardnadze met with new Russian president Boris Yeltsin and paved the way
for the Sochi Agreement, which ended major hostilities, defined a security cor-
ridor along the border of South Ossetian territories, and enhanced the security
of the patchwork of Georgian villages interspersed with ethnic Ossetian villages.
It established a Joint Peacekeeping Force composed of Georgian, Russian, and
South Ossetian troops, under Russian command. Russia contributed 700 troops
to this force (Baev, Koehler, and Zurcher 2005). The Joint Peacekeeping Force
would perform the duties of a conventional peacekeeping force but was made
up of troops from the three main armed actors, rather than a third party that was
trusted by both sides.
112 Examining the Cases

Within six months the OSCE had deployed a civilian mission whose man-
date included promoting negotiations between the parties. The OSCE mission
launched programs of disarmament and weapons collection, as well as confidence
building and dialogue programs that brought together representatives of civil
society from both sides (König 2005, 242–45). But these programs proved futile
in the face of the shift in Georgian policy that triggered the 2004 armed conflict.
The United Nations also deployed an integrated mission in Tbilisi in 1993, but
its mandate was limited to the conflict over Abkhazia. The middle to late 1990s
were marked by increased attempts at cooperation between the South Ossetian
authorities and the Georgian government, including law enforcement coopera-
tion, road reconstruction, and improved electrical supply (Cvetkovski 2001).
Mobilization in both South Ossetian conflicts occurred along ethnic rather
than religious lines. In 1989 Georgia’s ethnic makeup was 70.1 percent Georgian,
1.8 percent Abkhaz, 3.0 percent Ossetian, and 6.3 percent Russian (US Depart-
ment of State 1998). The population of South Ossetia in 1989 was 66.0 percent
Ossetian, 29.3 percent Georgian, and 1.9 percent Russian (Baev, Koehler, and
Zurcher 2005, 265).6 At least an equal number of Ossetians lived within Georgia
outside South Ossetia, where they lived relatively peaceably with Georgians and
others (Dawisha and Parrott 1997). Mobilization for the armed conflict and its
recurrence occurred largely along ethnic lines, with many Georgians displaced
from South Ossetia in the first war.
Religious difference was not a salient point of mobilization in the conflict re-
currence in South Ossetia. In 1993, the religious makeup of Georgia was 65 per-
cent Georgian Orthodox, 11 percent Muslim, 10 percent Russian Orthodox, and
8 percent Armenian Apostolic (Curtis 1995). Both ethnic Ossetians and other
residents of South Ossetia are mostly Orthodox Christian, with a minority of
Sunni Muslims among ethnic Ossetians. Both the Abkhaz and South Ossetia
conflicts manifested ethnic more than religious identities in discourse and mo-
bilization, particularly in South Ossetia, where the majority share the Orthodox
Christian religion. South Ossetia is extremely mountainous, and thus the only
northern passage to Russia is through a single tunnel.

The 2004 Recurrence of Armed Conflict

In this case, again, exclusionary behavior plays an important role, because Geor-
gia’s newly elected president deliberately embarked on its top priority to bring
South Ossetia and the other two irredentist regions more firmly into the state’s
grasp (König 2005, 238). Unfortunately, integration backfired in South Osse-
tia. Georgian president Mikheil Saakashvili tightened controls on movement of
goods and persons in and out of the region, a step viewed as cutting off the re-
gion. As Welt (2010, 64) says, “The trigger to the 2004 conflict is not in dispute:
Separatist Recurrences of Civil War 113

it was new restrictions on trade and movement that Georgia imposed in and
around South Ossetia.” However, in contrast to Charles Taylor’s plunder of the
Liberian state for the gain of his in-group, Saakashvili was willing to take some
inclusionary measures that his government viewed as significant concessions.
Indeed, the 2004 armed conflict was not inevitable: Saakashvili made overtures
to what he considered the “easiest” of the three regions, Ajaria, and thus suc-
cessfully secured its greater integration into Georgia in ways that “unfroze” that
particular conflict (König 2005, 237–38; Welt 2010). Ajarians, unlike Ossetians
and Abkhazians, are ethnic Georgians.
South Ossetia presented more difficulty than expected. Saakashvili entered
office in January 2004 in the culmination of the “Rose Revolution.”7 Upon tak-
ing office, he had already signaled that the “fragmentation” of Georgia would
not be tolerated (König 2005, 238). Tbilisi—recognizing that de facto South
Ossetian president Eduard Kokoity had very little popular support or legiti-
macy—thought that it might shift South Ossetian loyalties to the Georgian
state. It launched a crackdown on smuggling and black markets, hoping to ex-
pose and undercut the corruption of Kokoity’s government before his people. It
also offered social assistance to the population. The Georgian “carrots” included
establishing a free medical service, paying pensions, accepting the term “South
Ossetia,” and reopening the main railway link from Georgia to South Ossetia
(König 2005, 238).
Georgia’s plan, which presumed that Ossetian authorities would lose support
as they were less able to deliver services, backfired (International Crisis Group
2004, 1). Ossetians who earned their living from the black market were angry,
and others saw assistance as an attempt to co-opt the population (PRIO 2003b).
The deployment of around-the-clock police and political volunteers to stop con-
traband was akin to a widening of “occupation” of civilians, which raised the
specter of more widespread incursion and loss of autonomy (Welt 2010, 65).
Georgia also deployed several hundred special forces in the “conflict zone” as
part of the Joint Peacekeeping Force, but in a perceived violation of the Sochi
Agreement. Welt argues that, in a classic security dilemma dynamic, this move
was seen as offensive by South Ossetians, whose counterdeployments with the
Russians were in turn viewed as offensive rather than defensive by the Georgian
government (Welt 2010, 65).
In the summer of 2004 Georgia accused Russia of providing arms to the South
Ossetians and seized two trucks full of missiles. South Ossetia responded by
briefly detaining 50 Georgian police officers, and exchanges of gunfire between
forces commenced for the first time since 1992. Both sides marshaled forces, and
for the first time since 1992 the defined conflict zone was remilitarized by up to
3,000 additional Georgian troops, 2,000 South Ossetian troops, and an influx of
some 1,000 Russia mercenaries (International Crisis Group 2004, 14).
114 Examining the Cases

The mediation and confidence-building bodies played a troubled but impor-

tant role. Negotiations between the Joint Control Commission and the Joint
Peacekeeping Force were initially hampered by the refusal of one or the other
parties to participate in meetings. One OSCE official said that “the [ Joint Peace-
keeping Force] was totally not functioning. . . . We had cases of one battalion fir-
ing at another” (quoted by International Crisis Group 2004, 15). An important
weakness of having a peacekeeping force and consultative mechanism made up
solely of the parties themselves is that these mechanisms can also be disabled
upon a breakdown of stability. Eventually, however, the Joint Control Commis-
sion facilitated a cease-fire and prevented the armed conflict from spiraling into
a more deadly war. Georgia was motivated partly by its recognition that it could
not prosecute a full-scale offensive that could reassert control over the terri-
tory without confronting lengthy armed resistance (International Crisis Group
2004, 14).
Economic factors entered into the recurrence through the tensions over bor-
ders and their recognition. Natural resources played no prominent role in these
conflicts. After the first war South Ossetians perceived an embargo by the Geor-
gians. Their affinity with the Russian Federation was enhanced, because the lat-
ter became a more prominent market for goods, provided humanitarian aid for
Ossetian refugees in North Ossetia, and covered the cost of road construction,
pensions, and health services (International Crisis Group 2004, 8–10). South
Ossetians complained that after the (first) war, they had received nothing from
Georgia. Georgia resented increased smuggling because the Ossetians refused to
permit their customs agents inside the territory along the northern border, and
Georgia resisted posting agents at the southern border of South Ossetia (with
Georgia) for fear of reinforcing the perception of South Ossetian sovereignty
(International Crisis Group 2004, 8–9).
In summary, the recurrences of internal armed conflict in Chechnya, South
Ossetia, and Tibet all reflect perceptions among separatist movements of be-
trayal by the sovereign state of the arrangement that emerged from the first war.
In these three cases separatist movements that did not succeed nevertheless
ended the prior war with some degree of de facto autonomy that was contained
in a written agreement signed by the central state authorities, but without a clear
resolution of permanent status. Greed per se did not enter into these war recur-
rences in any serious way. Although all the regions were poor, indeed poorer
than the core areas of the state to which they nominally belonged, that difference
was not as prominent as in other cases of deliberate ethnically aligned economic
disparity as in Burundi and Rwanda. In all three cases, the mountainous topog-
raphy played a historic role in defining the borders of nations, in inhibiting travel
into and out of these secessionist regions, and in complicating war fighting and
facilitating the carrying out of rebellion and its duration.
Separatist Recurrences of Civil War 115

China and Tibet in 1950–51 and 1956–60:

Compelled from Autonomy
Since the Chinese Communist Party’s 1949 victory, China has experienced
numerous internal armed conflicts, some ethnic separatist in nature. Although
other countries are excluded from this book’s analysis because they experienced
multiple internal wars, China’s civil wars with Tibet (1950–51 and 1956–60)
are included here because war termination and the second civil war can be ana-
lyzed in relative temporal and geographical isolation from the remainder of the
country’s wars. China’s war recurrence in Tibet grew out of a number of factors,
including opting for exclusionary policies rather than inclusionary policies by
the People’s Republic of China. At its core, the initial war and its recurrence rep-
resent resistance to an imposition of a social, political, and economic transforma-
tion as part of a military occupation.
Upon coming to power, the Chinese Communist government set about con-
solidating control over Tibet, a region that had long exercised de facto internal
autonomy but was also under Chinese sovereignty. It was immediately apparent
that the new regime would seek to reassert control over Tibet and other periph-
eral regions, and the Tibetan government undertook to remove all elements of
Chinese authority in Tibet and those associated with it (Shakya 2000, 5). After
a year of unsuccessful pressure on Tibetan officials to accept Chinese authority,
the People’s Liberation Army (PLA) invaded Tibet in October 1950.8
Several factors lay behind this invasion. Zhou Enlai stated that the advance
of the PLA was determined to “liberate the Tibetan people and defend the fron-
tiers of China” (quoted by Shakya 2000, 45). Concerns with outside countries
using Tibet as an ally and foothold onto China remained high. Goldstein (2007,
20–21) argues that national “honor” in the face of British imperial dominance
was a major factor in the determination to reassert control over Tibet.9 Com-
pounding these Chinese concerns were ongoing Tibetan relationships and even
entreaties abroad. At the same time, Tibetan reactions and policies were devel-
oped in iterated interaction with American, Indian, and other outside actors and
their (un)willingness to confront China over Tibet.
Although the Tibetan autonomous areas (as designated by the Chinese gov-
ernment) constitute one-fourth of the Chinese territory (Davis 2008, 230), in
1953 the Tibetan population of only 1.3 million represented less than 0.5 percent
of China’s population (Orleans 1966, 120). The new regime sought a security
buffer to its southwest and access to valuable minerals, although these resources
did not play a major role (Bowers 1994, 409). In addition, half of today’s na-
tional territory lies in five ethnic-minority regions (Tibet, Xinjiang, Ningxia,
Guanxi, and Inner Mongolia) that produce 80 percent of the country’s meat
and milk (Bowers 1994, 427). The precedent that Tibetan independence would
116 Examining the Cases

have set in the 1950s carried broad ramifications for both China’s economy and
its stability.
During the invasion of 1950–51, Chinese sources report an estimated 5,000
persons died, and thousands of Tibetans fled to India. The advance of the Chi-
nese army was also met with the surrender of many Tibetan soldiers. The PLA
forces halted about 200 kilometers from the capital, Lhasa, and sought to negoti-
ate the exercise of its control with the Tibetan authorities and the Dalai Lama,
who was urgently formally installed at the age of sixteen years in light of the inva-
sion. Eventually a negotiations team of Tibetan officials in Beijing signed what
became known as the Seventeen-Point Agreement in May 1951.10 The authority
of the team, and the conditions of freedom or coercion under which they signed
the document, remain in dispute (Shakya 2000, 45–91). To the Dalai Lama’s
surprise (Goldstein 2007, 20), the Seventeen-Point Agreement stated that Ti-
bet was part of China, pledged Tibetan assistance for the presence of Chinese
military troops, and provided for locating military and civilian authorities in
the region.
However, the accord otherwise reaffirmed the political and religious status
quo of Tibet. Points 4 and 5, for instance, stated that the “Central Authorities
will not alter the existing political system in Tibet,” and affirmed the “status, func-
tions and powers” of the Dalai Lama and the Panchen Lama. Point 11 said, “In
matters related to various reforms in Tibet, there will be no compulsion on the
part of the Central Authorities.”
What led to the renewed fighting in 1955 that culminated in 1959 uprisings
in Lhasa that were suppressed? Tibet’s renewed warfare can best be seen as a
reaction against a gradually more invasive and transformative occupation. The
Chinese Communists openly described their agenda of “liberating” Tibet from
imperialist influences. At Mao Zedong’s insistence, they quietly and patiently
began a long-term process of social transformation of the religious, economic,
and political structures of society (Goldstein 2007, 542). Thus diverse Tibetan
Buddhist sects were brought into the polity for the first time (Shakya 2000, 132).
Some lamas received salaries and honorific posts in the Preparatory Committee
appointed to oversee a transition of Tibet into an autonomous region, akin to
the other ethnic minorities in which autonomy was more reduced than in Tibet
(Shakya 2000, 132–33).
The bureaucratic advance of the social policies and tasks of ever more in-
trusive territorial administration provoked rebellion among Tibetans, initially
in the eastern provinces. In 1955 the Chinese government initiated its Major
Socialist Campaign aimed at accelerating the collectivization of the nation’s
agriculture, including the ethnic minority regions so that these would not “fall
behind” (Goldstein 2007, 549). Although Tibet was exempt from this policy,
ethnic Tibetans who lived in the adjacent jurisdictions of Kham and Amdo were
Separatist Recurrences of Civil War 117

not. In contrast to central Tibet, where a feudal system of aristocrats and serfs
and slaves prevailed, many Tibetans in these provinces were nomads or small
landowners. Here the campaign proceeded apace and gave rise to a bloody up-
rising that began in Kham by ethnic Tibetans (called “Khampas”) (Goldstein
2007, 549). Within months, the rebel activities had spread to western Kham and
to central Tibet.
Between 1956 and 1959 thousands, possibly tens of thousands, died.11 The
PLA made significant headway against the rebellion in 1957, leading thousands
to flee to central Tibet. The fighting culminated in uprisings in Lhasa, but the
PLA crushed these and wrought devastation on religious sites (destroying many
Buddhist monasteries and religious sites). In 1959 the gradualist approach to
Tibet ended, and Tibetan autonomy was curtailed. The Dalai Lama fled and es-
tablished a government in exile across the border in India.
Tibetans launched their renewed civil war in response not only to forced col-
lectivization but also to the broader influx of and exercise of control by Chinese
administrative officials, soldiers, and Han Chinese migrants (Grunfeld 1996,
131–35). Mao planned to stimulate massive Han migration to Tibet with the goal
of Hans outnumbering Tibetans five to one (Minahan 1996). That migration in-
creased in the mid-1950s. Another example of the bureaucratic ramping up of au-
thority is the effort to disarm the Tibetan population in accord with national pol-
icies. The Chinese pressed forward their plans, even though Tibetans objected
that weapons were a part of their culture, including inside Buddhist monasteries,
where they formed part of sacred ornamentation. Their initial hesitation in step-
ping up intrusion in social and political life derived partly from the sense of de-
pendence and isolation, because the thousands of occupying PLA soldiers relied
on Tibetans for food and the country was only accessible over land from India.
In addition the Chinese government not only paid their tens of thousands of
soldiers but also paid additional civilian administrators and offered interest-free
loans to Tibetan businessmen and purchased Tibetan wool and other goods at
inflated prices. As a result of the economic costs of these policies (and the Korean
War at the other end of China), the Chinese government “began to take greater
control of the Tibetan economy” (Shakya 2000, 135). Thus the very presence of
ethnic Han Chinese laid the groundwork for the perceived marginalization and
disempowerment of Tibetans under Chinese occupation. “By the middle of 1955,
every Tibetan began to feel the presence of the Chinese” (Shakya 2000, 133).
The exclusion perceived by the Tibetans in the mid-1950s may also derive
from the end of the period required for the Chinese government to complete the
logistical and communications conditions necessary to militarily support full ex-
ercise of its authority over Tibet. No year-round passable roads existed between
Tibet and China until construction projects, deemed a priority in 1951, com-
pleted two roads in 1955. Telecommunications were also difficult, and telegraph
118 Examining the Cases

towers were also constructed in this time frame. As Thomas (1959, 129) says,
“The roads now tied Tibet to China. Immediately the Chinese increased their
pressure on the Dalai Lama and high officials; they wanted no more delay about
the establishment of the administrative committee that would give them formal
control of the country.”
Did the Chinese policies constitute exclusionary behavior? Some argue that
the Chinese never violated the provisions of the Seventeen-Point Agreement.
The Chinese authorities permitted extensive self-rule, and their transformative
social reforms occurred through carrots much more than sticks. However, the
incursion and the prolonged presence of troops, civilian administrators, and
Han migrants led to small changes in governance. Thus the Chinese prohibited
the cross-border trading of the Tibetan silver ingots used as currency because it
was undermining the Chinese currency. Wages were cut after being raised. Small
steps like changing the uniforms of the Tibetan armed forces also signaled that
the Chinese government would, in fact, slowly transform the status quo, even
with the complicity of some Tibetan elites along the way.
Finally, international and regional actors influenced the war recurrence in
Tibet. In the first war, calculations about how much leverage Tibetans had in
negotiating hinged partly on whether the Americans or others would provide
overt aid and recognize them (Shakya 2000). Extensive diplomacy between
the Americans, the British, other powers, and the Tibetan government in exile
persisted (Grunfeld 1996, 151–65). After the Dalai Lama’s trip to Beijing and
embrace of initiating some elements of socialist transformation, his brother and
two other Tibetan government officials sought the support of India and the
United States for a secret resistance group. By early 1956 the US Central Intel-
ligence Agency was recruiting guerrillas, training them in camps in Nepal, and
air-dropping them into Kham and Tibet (Shakya 2000). Many of these (most are
presumed to have been captured and killed) were drawn from the aristocrats of
Tibet’s extremely archaic feudal system.
Even if one accepts that the letter of the Seventeen-Point Agreement was
never violated, the issues of ethnicity and exclusion lie at the core of Tibet’s war
recurrence. Natural resources are absent from the main historical works on Tibet
(Goldstein 2007; Shakya 2000; Richardson 1962; Patterson 1960). Although
outside forces propelled the conflict once it was under way, they did not insti-
gate it.

Analyzing Cases of Reneging on Territorial Autonomy

All four of these cases offer evidence of the importance of state exclusionary be-
havior as the most significant trigger for civil war recurrence. New wars reflected
objective policy changes in each case, but these changes represented significant
Separatist Recurrences of Civil War 119

departures from the expectations held by separatist populations and their lead-
ers, which undermined hopes for continued or expanded autonomy.
The case of Sudan’s North / South civil war recurrence offers perhaps the
starkest case. The decision by the North to impose Sharia law on the South, to
transfer southern soldiers to the North, to arrest southern politicians and to ig-
nore agreed-upon federalism all triggered renewed warfare. Analysts are unam-
biguous about the privileged role of these actions, seeing the resurgence of orga-
nized attacks in the South as a reaction to, rather than precipitator of, the 1983
war. The cases of Chechnya and South Ossetia are similar insofar as separatists
in an earlier contest had fought strongly enough to force a cease-fire agreement
involving de facto autonomy. But after some years, and for different political rea-
sons, central state leaders opted to change the arrangement, which prompted
renewed warfare.
National heads of government in all four cases took action that reflected po-
litical aspirations and / or perceived political constraints. Russian prime minis-
ter Putin and the Russian military saw potential gains in relaunching a second
Chechen war, for political and security reasons. Of all these separatist cases, only
Chechnya’s rebel forces provide serious evidence of proactive defection from
the Khasavyurt Agreement in 1999. However, even Russian officials see the at-
tacks of dissident separatist Chechen groups as serving mainly to accelerate and
complicate Russia’s planned defection from the accord. Georgia’s newly elected
president Saakashvili thought he might consolidate his power and that of Geor-
gia through new provocative policies against illicit trade in South Ossetia, as he
had done swiftly in Ajaria.
In contrast to these two new leaders acting out of perceived and anticipated
strength, Sudan’s president Nimeiri acted out of potential weakness. He was mo-
tivated partly by fear that Islamists were becoming stronger and needed to be
appeased, which he obliged by toppling the South’s autonomy. Because China’s
strategic superiority over Tibetan forces was evident in the early 1950s, the Ti-
bet case shows some differences. Even so, Mao Zedong’s government showed
some restraint until impatient advocates of speedier reform prevailed and im-
plemented land reforms that prompted renewed fighting there. The politics of
handling separatist territorial movements can have its origins in questions of
statehood or regime change in each of these four cases. In three cases initial wars
occurred in tandem with the arrival of statehood and a new government after
external occupation. In the fourth case, Tibet, the initial war took place directly
on the heels of the regime change wrought by revolution in 1949.
What of alternative arguments? Only in the case of Sudan did natural re-
sources play a salient role in war recurrence of these cases, and even so that role
was in tandem with exclusionary dynamics that received more analysis and at-
tention. It was also only one of a handful of other economic issues considered
120 Examining the Cases

important at the time for the South. Conventional measures of state strength for
Georgia and Russia—given the well-developed role of the state in society, the
economy, and security—place them on a considerably higher level of institu-
tionalization than countries such as Burundi, Rwanda, the Central African Re-
public, and Haiti. Except for the presence of a joint Russian-Ossetian-Georgian
peace force in South Ossetia, none of the other cases experienced the presence
of international troops. And in South Ossetia, the Russian presence was not seen
as “neutral.” This factor is revisited subsequently in later chapters, where we can
consider cases of both recurrence and nonrecurrence.
In both Tibet and South Ossetia, one sees examples of secessionist struggles
in the context of military occupations in regions that had enjoyed significant au-
tonomy in prior decades. The dynamics of these struggles contrast sharply with
that of Liberia’s civil war and its recurrence, and with the experiences described
in Haiti, the Central African Republic, and other countries with nonseparatist re-
currences. Although East Timor represented a secessionist struggle, it too differs
from the pattern of Tibet and South Ossetia given the falling out among victors
after a successful armed struggle. I have argued here that Tibet and South Ossetia
represent secessionists who revert to war in the face of initiatives by a perceived
occupier to transform the status quo in ways that would render permanent or
more extensive the exclusionary rule, and thus undermine the possibilities of
self-government. This research does not analyze the factors that lead national
states to initiate such transformational policies, nor does it consider why in some
cases separatist territories do not choose to respond with renewed armed con-
flict. Such questions are important for a fuller understanding of separatist wars
and their recurrence. I now turn to nonseparatist struggles and their recurrence.

Notes
1. In the case of Azerbaijan, the recurrence in Nagorno-Karabakh was excluded be-
cause I could find no analysts except PRIO, which considered the two dozen deaths
from border skirmishes to constitute an internal armed conflict, especially given that
similar such skirmishes with fewer than twenty-five deaths had taken place in several
years between 1994 and 2005.
2. Sambanis (2006), covering 1945–2002, Fearon and Laitin (2003), covering
1945–99, and the Political Instability Task Force (2008), covering 1955–2007, all
score the first civil war from 1963–72, and the onset of the second in 1983, with vary-
ing end dates (especially because the war continued past the dates of publication for
Sambanis and Fearon and Laitin).
3. For more on the origins of the Chechen wars, see Dunlop (1998), Lieven (1998),
Sakwa (2005), Evangelista (2002), Cornell (2001), Gall and de Waal (1998), Russell
(2005), Jean (2000), and Ashour (2004).
Separatist Recurrences of Civil War 121

4. See James Fearon’s (1998) emphasis on the lack of credible commitments created
by the disappearance of central authority upon the dissolution of the Soviet Union and
the Yugoslavian Federation.
5. In 1944–2002 data set, Sambanis codes two civil wars, in 1991–92 and 1992–94.
Walter’s 1946–96 data set does not code any wars in Georgia in the early 1990s, and the
Political Instability Task Force codes two wars, one in 1991–93 and one in 1992–93.
6. The most recent Georgian census in 2002 did not include the autonomous ter-
ritories of Abkhazia and South Ossetia.
7. Saakashvili’s populist, nationalist government replaced that of the more tradi-
tional Communist (and former Soviet foreign minister) Eduard Shevardnadze, who
was forced to resign in the face of mass demonstrations against electoral fraud.
8. The war could, therefore, also be considered interstate.
9. Tibetan army officers, for instance, continued to sport British uniforms (Gold-
stein 2007, 215).
10. This was officially the “Agreement of the Central People’s Government and the
Local Government of Tibet on Measures for the Peaceful Liberation of Tibet,” May 23,
1951.
11. Some scholars estimate that 45,000 Han died and 65,000 Tibetans, but these
are considered high (Grunfeld 1996, 133).
5
Nonseparatist Recurrences of
Civil War

THE PREVIOUS CHAPTER ADDRESSED armed conflicts of a separatist nature.

Such armed conflicts share certain characteristics and are generally considered
more difficult to resolve via negotiated settlement and in a sustainable fashion
(Wood 2003). Separatist conflicts mobilize support around identity and its rela-
tion to territory. Consequently, territorial powersharing, including substantial
autonomy in governance and security and economic matters, is among the most
common demands and features of a negotiated settlement. Unfortunately, terri-
torial powersharing tends not to prepare a population for more integrated social
and political practices but instead tends to deepen the expectation of autonomy
among separatist populations and those who fought for it. Territorial power-
sharing is thus difficult to reverse without violence emerging.
I now turn to nonseparatist internal armed conflicts, which exhibit more
diversity in the aims articulated by insurgent forces. These cases of recurrence
also exhibit a range of levels of violence, from the deadly recurrences of ethnic
conflict and violence in the Great Lakes cases of Rwanda and Burundi in the
1970s, 1980s, and 1990s to the relative skirmishes of East Timor and Haiti in the
2000s. The distinction used here between separatist and nonseparatist conflicts
is increasingly seen as worth exploring in the literature (Doyle and Sambanis
2006; Cederman and Gleditsch 2009). Nevertheless, it is not the sole relevant,
or even the most pertinent, distinguishing feature among armed conflicts or
their recurrence.
In this chapter I distinguish among three distinct patterns of postwar vio-
lence that reflect very different manifestations of exclusionary behavior. Those
patterns are inductively derived from my research, although the bases for them
can be found in existing scholarship. The first distinction is between the “pre-
cipitating” exclusionary behavior examined above that also applies to some
nonsecessionist cases like Haiti and East Timor, and what I call “chronic” ex-
clusionary behavior. The latter form of exclusion refers to a condition of severe
exclusion of a social or ethnic group from access to positions of political power
Nonseparatist Recurrences of Civil War 123

for a period of several years. Usually such exclusion extends beyond access to
state power to also include access to economic power and social opportunities.
Recall that precipitating exclusion can take the form of either formal exclusion
based on laws or written policies (e.g., apartheid) or informal treatment or be-
havior that systematically excludes a group from such positions. In addition, I
argue here that the important aspect of exclusion rests on the dashed expecta-
tions of the aggrieved ethnic or political group rather than an objective or stan-
dard schema.
This chapter argues that exclusionary behavior played a precipitating role
in recurrences not only in Liberia and the four separatist recurrences analyzed
above but also in five additional cases: Haiti, East Timor, Zimbabwe, Burundi,
and Rwanda. In the first three of these additional cases, this behavior was “pre-
cipitating” insofar as it was a trigger causally linked to the recurrence of armed
conflict. In the latter two cases, Burundi and Rwanda, I argue that the exclusion
was “chronic” and was generally considered the most important factor underly-
ing the repeated violent conflict in those two interrelated cases.
These nonseparatist cases also suggest the importance of a second pattern of
precipitating exclusion: the exclusion of former allies or coalition partners after
a successful rebellion. In most cases where newly elected or empowered postwar
governments engage in exclusionary behavior, it is against former enemies or
the social groups associated with them. However, in a small set of cases drawn
from successful insurrections, exclusion may precipitate renewed armed conflict
when postwar governments move to marginalize, exclude, or even persecute
their former rebel partners. The two central cases analyzed here are East Timor
and Zimbabwe, although such behavior contributed to Nicaragua’s postrevolu-
tionary recurrence as well, and has been identified by Atlas and Licklider (1999)
in Sudan and Chad.
As mentioned in the previous chapter, the state’s reneging on an explicit
powersharing deal in the territorial, political, or security realm constituted the
principal trigger of recurrence in six of fifteen core cases. Of these, Liberia, the
Central African Republic, and Zimbabwe represented nonseparatist recurrences.
In the other nonseparatist cases of Haiti and East Timor, recurrence was tied to
a failure to meet informal expectations for participation in postwar governance.

Precipitating Exclusionary Behavior

The dynamic of exclusion in the wake of an autonomy arrangement differs in
important ways from exclusion from struggles that seek integration. Inclusion in
various state offices can serve as both security guarantees for former rebels and
as guarantees that political, economic, and social priorities will be taken into ac-
count in state policies. The case of Liberia examined above reveals how armed
124 Examining the Cases

groups emerged after an elected government, which enjoyed a degree of popular

support and had come to power through multidimensional UN peacekeeping
operations, marginalized former enemies and motivated renewed armed con-
flict. As shown, societies can engage in inclusionary or exclusionary behavior in
multiple ways, beyond inclusion in the security forces or in political posts.
Subsequently, in chapter 7 I examine in more detail powersharing arrange-
ments and their relative success in consolidating stability in postwar societies.
This chapter, however, examines nonseparatist failures of peace. It begins with
cases of war recurrence that conform to the pattern of the case studies examined
above: renewed rebellion after exclusion. Each case reveals the crucial impor-
tance of discriminatory treatment within the armed forces and police forces, in-
cluding the first case, the Central African Republic.

The Central African Republic in 1996–97 and

2001–2: Exclusion and State Weakness
The Central African Republic is a country that conforms closely, perhaps most
closely of the recurrent cases, to arguments about state weakness and Collier and
his colleagues’ thesis of a “conflict trap.” The country experienced several peri-
ods of unrest after the adoption of a multiparty system in 1993 under President
Ange-Félix Patassé, until he was deposed in a coup in 2003. And it has since ex-
perienced low levels of armed conflict. In this sense, it resembles the experience
of Haiti, although with better-organized armed actors who are mobilized around
ethnic identity and grievance. The country’s recurrence of conflict in 2001–2
certainly reflects exclusionary politics and also underscores the importance of
the presence of third-party troops in weak states, but simultaneously supports
the role of poverty and natural resources, given that 80 percent of its export earn-
ings at the onset of recurrence derived from diamonds and timber. The country
ranked 166th out of 174 in the UN’s 2000 Human Development Index (United
Nations Development Program 2000, 157). Moreover, of all the cases of war
recurrence, this case most closely supports a causal role for the weak state, where
personalism and a lack of institutionalization prevailed. The Peace Research In-
stitute Oslo (PRIO) Conflict Data Set, for instance, includes an armed conflict
in 2006. Nevertheless, the only recurrent armed conflict to qualify as involving
the same actors with recognized interwar peace was in 2001–2, and exclusionary
behavior played a salient role.
The disagreement among sources on the dates of the country’s civil war em-
blemizes the gray line between political instability, coups, and warfare that char-
acterizes weakly institutionalized polities. Some cite 2002–3, but others point
to several years of civil war ending in 2007 with a peace agreement. Fearon and
Laitin score a civil war from 1996–97, and PRIO marks periods of civil conflict
Nonseparatist Recurrences of Civil War 125

in 2001–2 and again in 2006. The first two conflicts, in the periods 1996–97 and
2001–2, had similar motivations and actors. The Central African Republic has
no ethnic majority but instead conforms to a plurality of groups. The country’s
ethnic breakdown as of March 2009 was Baya, 33 percent; Banda, 27 percent;
Mandjia, 13 percent; Sara, 10 percent; Mboum, 7 percent; M’Baka, 4 percent;
Yakoma, 4 percent; and other, 2 percent. The religious breakdown was 35 per-
cent indigenous beliefs, 25 percent Protestant, 25 percent Roman Catholic, and
15 percent Muslim, although the Christian majority is strongly influenced by
animist beliefs and practices. Like many of the conflict recurrences after 1988,
these repeat armed conflicts reflected not religious division so much as mobiliza-
tion along ethnic lines.

The Initial Conflict, 1996–97

The 1996–97 conflict was actually a series of three mutinies led by soldiers un-
happy with the government of President Patassé. Patassé was a member of the
Kaba, a subgroup of the Sara ethnic group in the northern part of the country.
However, the armed forces were dominated by the Yakoma, mainly found in the
south, an ethnicity shared by former president André Kolingba. In the April 1996
mutiny, in which few persons died, complaints about the failure of the govern-
ment to pay salaries for several months lay at the center of grievances. However,
in May and again in November 1996, the second mutiny cost several hundred
lives—more than a thousand by some accounts (AlertNet, as quoted by Samba-
nis 2007). According to the United Nations, “The mutinies and subsequent loot-
ing sprees brought about the destruction of commerce and industry, devastating
the formal and informal sectors and triggering high unemployment.”1 Exclusion
played a central role in the grievances of the soldiers in the latter mutinies, pri-
marily because Yakoma soldiers accused Patassé’s government of ethnic favorit-
ism in the treatment of officers and complained of salary arrears.

The Bangui Agreements

Roughly 2,000 French troops were vital in preserving Patassé in power, and the
conflict ended with the January 1997 Bangui Agreements, which were intended
to settle social and economic grievances. The agreements provided for a return
to the powersharing “Government of National Unity” agreed to in 1996 and then
abandoned. It also provided for a restructuring of the armed forces to include
diverse ethnic groups, a task relegated to General Didace N’dayen, the minis-
ter delegate of defense for ex-combatants, war victims, and restructuring of the
army. N’dayen was named to his newly created post on behalf of the mutineers
on February 18, 1997, as part of a cabinet that included several leaders appointed
126 Examining the Cases

from the “Group of Eleven” opposition coalition. By September of that same

year, the opposition members of the powersharing cabinet had resigned.
In addition, the Bangui Accords provided for a new ad hoc African peace-
keeping force to deploy with French logistical support, the Inter-African Mission
to Monitor the Bangui Agreements, which was made up of 820 soldiers from six
West African countries and was initially deployed on February 8, 1997, to engage
in combat on occasion with mutinous troops. It served until April 1998, when
it was replaced by the UN Mission in the Central African Republic. The UN
Security Council terminated this mission’s mandate in February 2000, and virtu-
ally all French troops withdrew by July 2000 (Havermans 2000). Powersharing
and external troops combined to provide minimal stability for this interim four-
year period.

Recurrence in 2001
The presence of UN peacekeeping troops corresponded to a period of relative
stability, because there was little violence between the Bangui Agreements and
the end of the peacekeeping mission in the Central African Republic in April
2000. Violence emerged a year later, again in the form of a coup attempt involv-
ing dissatisfied elites who rallied troops rather than a mass-based civil war. In
May 2001 members of the country’s armed forces, led by former president André
Kolingba (from the Yakoma ethnic group), led a coup attempt against President
Patassé. They alleged that the military’s old grievances from the prior conflict
against Patassé had not been addressed.
The rebellion of mid-2001 was suppressed with the help of Libyan troops
and a neighboring rebel group, the Mouvement pour la Libération du Congo
(MLC). Because the MLC was intent on toppling the Kabila government in
the Democratic Republic of the Congo, it provided troops that supported the
Patassé government, largely because the Central African Republic had allowed
the MLC to operate within its southern territory. The UN denounced the coup
attempt, but it stated that the government should not purposefully target the
Yakoma tribe with violence. Former Malian president Amadou Toumani Touré
was sent as a UN peace envoy to speak with the government, opposition, and
civil society. Patassé accused his army chief of staff, General François Bozizé, of
helping to orchestrate the coup attempt and had him fired. The casualties had
been relatively few up to this point in 2001.
However, additional rebellion against the government occurred within a few
months in 2001. The ousted General Bozizé gathered troops to resist his firing,
dividing the armed forces between Patassé and Bozizé loyalists. In November
2001, as the conflict resurged, the UN stepped up its mediation efforts, sending
Lamine Cisse, a retired Senegalese army general, to lead the mediation effort.
Nonseparatist Recurrences of Civil War 127

The Chadian ambassador played a role, and an Organization of African Unity

team participated in talks as well. From late 2001 the Libyan troops in the Cen-
tral African Republic were joined by troops from Sudan and Djibouti to make
up a peacekeeping force of the Community of Sahel-Saharan States. The BBC
(2010) claims that 59 persons died in 2001, and PRIO estimates between 50 and
300 battle-related deaths that year.
After November 2001, Bozizé and some of his men fled into Chad, which
exacerbated tensions between the Central African Republic and Chad. During
2002, many clashes occurred along the border between the two countries, and
an emergency summit of the Economic Community of Central African States
(CEMAC) summit was held in August 2002. A CEMAC observer mission was
initially dispatched to examine the security situation along the border. After an-
other coup attempt by Bozizé in October 2002, Bozizé was granted asylum in
France, and a CEMAC mission was sent to protect the president in Bangui. As
these events indicate, the Central African Republic’s armed conflict reflected a
persistent recourse to coup attempts in ways that make evident the state’s weak
institutions. According to PRIO, the total battle-related deaths from the armed
conflict in 2001–2 ranged between 200 and 700.
The Central African Republic’s conflict recurrence of 2001–2 supports the
argument that security by third-party troops is especially important in weak
states, where informal institutions are the more prevalent norm (Walter 2002).
Powersharing seemed helpful, but it failed within a few months to take hold in
late 1997. However, three years passed before the conflict reignited. It was only
in 2001, several months after external peacekeeping troops withdrew, that the
renewed attempts to topple the government led the Central African Republic
into a second internal armed conflict. The timing suggests that, in this case, the
external forces provided a necessary condition for sustained stability.
In conformity with Collier and others’ (2003) arguments about the “con-
flict trap,” the country is also dependent on natural resources in its exports. In
2002, diamonds made up 50 percent of exports, and timber made up 30 percent
(US Department of State 2010a). The country has uranium deposits, but these,
along with gold, oil, and other potential mineral deposits, have not been mined.
The history of political instability makes such rents a less than sure source of
income (Ruane 2003, 29). The Chad–Cameroon oil pipeline only brushes up
against the Central African Republic. However, control over these minerals did
not figure prominently in the discourse of rebels, even though discourse may be
misleading (Collier et al. 2003).
Regional dimensions also help account for the conflict. Armed conflicts in
Sudan, Chad, and Uganda have helped fuel the conflicts in the Central African
Republic, particularly in the north. Rebel and military forces from the neighbor-
ing countries have used the northern Central African Republic for operations,
128 Examining the Cases

leading to small arms proliferation. In 2007, rebels from Sudan fled into the Cen-
tral African Republic, as did units of the Lord’s Resistance Army from Uganda
in 2008 (BBC 2010).

Subsequent Armed Conflict and Coup Attempts, 2002–Present

On March 15, 2003, Bozizé led a coup that successfully ousted sitting President
Patassé, and the CEMAC peacekeeping force did not resist. In 2002, France had
withdrawn its support for Patassé and its troops from the country (Ruane 2003,
29). Because only thirteen people died in the coup, 2003 is not coded as an active
conflict. After the coup, Gabon and the Republic of the Congo sent delegates to
lead peace talks. The CEMAC force was dedicated to improving the security sit-
uation around the capital, and some later moved to provide security in the north,
where violence had continued since 2002. The CEMAC force, called FOMUC,
also helped in the disarmament process. Between September 15 and October 15,
2003, the country engaged in a national dialogue at which delegates discussed a
range of issues aimed at reconciling the country and preventing further violence.
Attacks continued at lower levels in subsequent years. In 2003, some of the
forces that had helped bring Bozizé to power began attacking government troops
in the capital. They called themselves Ex-Liberators and expressed discontent over
the amount Bozizé paid for their services. In November 2004, forces allegedly
associated with a group calling themselves Patriots for Renewal / Ex-Liberators
attacked government troops in the northwest of the country. Again, casualties
did not rise to the level of armed conflict, but they also dictate that the period
between the coded conflicts of 2001–2 and 2006 not be viewed as a period of
widely perceived “peace.” There was no widespread perception that the armed
conflict of 2002 ended and was followed by a period of peace. Instead, the level
of casualties fell below a certain threshold as dissident groups continued to mo-
bilize and carry out sporadic attacks in 2003–4 and surrounding the elections
of March 2005. Hence I do not include the 2006 recurrence as a separate armed
conflict for the purposes of this study.
If the 2006 recurrence were included, however, it would reflect the factors
identified above, namely, weak state institutions and ethnic exclusion in appoint-
ments to state offices, including within the armed forces. After the May 2005
elections, in which Bozizé was reelected in a runoff, rebel groups took shape in
the country’s far northwestern and northeastern regions, where the central gov-
ernment has done little to address local grievances. The Union of Democratic
Forces for Unity (UFDR) took military control of parts of the northeast in late
2006, and France assisted the Central African Republic government in taking
back the area. The UFDR was made up principally of members of the Gula eth-
Nonseparatist Recurrences of Civil War 129

nic group, and the northeastern rebellion prompted anti-Gula sentiment and re-
taliation, leading many Gula to flee their homes. Some of the Ex-Liberators (who
helped Bozizé to power) were also active in the rebellion. The Armée Populaire
pour la Restauration de la République et la Démocratie (APRD) was made up
mainly of former presidential guards of Patassé, who was from the same region.
The APRD claimed that its goal was to engage in dialogue with the government
to address the exclusion of Patassé supporters from elections, as well as the secu-
rity situation in the northwest, where armed bandits attack villagers.

Haiti in 1990–95 and 2004: Political Exclusion and Recurrence

Haiti’s war recurrence demonstrates a number of important questions and
claims. First, it is a clear failure of peacebuilding, even after peacekeeping had
been declared successful and stability had been established. Second, it illustrates
the conceptual difficulties of measuring and comparing mass violence, because
neither of its episodes neatly meets prevailing definitions of civil war or interstate
war. The case also demonstrates how armed recurrence stemmed from perceived
marginalization of opposition parties and factions in the wake of a successful
postwar election. That 2003–4 armed uprising occurred only once international
troops had departed and donor engagement had declined, along with important
international influence. Haiti’s patterns of internal armed conflicts also reflect the
weak state and social institutions in one of the world’s poorest countries.
Haiti is celebrated as having mounted the first successful slave revolt of the
world, but the independent country that emerged in 1804 replicated many el-
ements of the colonial state. New elites reproduced the plantation economy,
substituting feudal, low-wage peasant relations for slavery and maintaining a
“predatory” state that extracted resources from the majority of the population
and provided negligible goods and services in return.2 Haiti’s politics reflect a
long history of small-scale political violence, feared rule by brutal authoritarians,
poverty and deep inequality, and susceptibility to external forces, especially US
policy. As of the 1990s, the country’s light-skinned elites, who made up only
5 percent of its population (CIA 2010), had dominated the dark-skinned masses
in society, the economy, and the state for decades (Trouillot 1989, 1994).3
Religious difference plays little part in political divides. Despite a small and
growing Pentecostal presence, most Haitians regard themselves as Catholic, even
though the great majority of Haitians also believe in elements of Voodoo, includ-
ing its spirits (which draw on Catholic saints) and spells (Desmangles 1992).
Political differences do not fall along these lines, nor are followers mobilized via
religion. Father Jean-Bertrand Aristide’s political beliefs and activism, rather than
his religion, inspired his followers and alarmed his critics.
130 Examining the Cases

The First Armed Conflict, 1990–95

The first armed conflict unfolded in the aftermath of the 1986 departure of Jean-
Claude “Baby Doc” Duvalier, an ironfisted dictator whose rule extended that
of his father’s (1957–71). A series of elected and often military leaders ensued,
emblemizing Haiti’s condition of weak state and social institutions and personal-
istic rule with few institutionalized political parties or civil society organizations.
Although coded as a civil war recurrence, one must stretch the concept for either
of Haiti’s episodes of organized armed violence to qualify.
The initial armed conflict commenced with election-related armed attacks
between leftist supporters of the populist priest Father Jean-Bertrand Aristide,
who spoke against the military’s continued rule and rallied the poor, and the
United States–backed government and thuggish security forces known as the
Tonton Macoutes (Farmer 2004). A 1991 coup d’état ousted the country’s first
democratically elected president, the leftist priest Father Aristide, and led to
one-sided state repression that cost more than 3,000 lives and sent 20,000 flee-
ing the country.4 This initial episode concluded with a near-war, where a UN-
authorized, United States–led invasion was launched from American air bases
on a September night in 1994 but was called off once Haiti’s de facto military
leaders learned of the attack and surrendered before the warplanes reached
Haitian airspace (Shacochis 1999, 67–77). This example of US “coercive diplo-
macy” converted what would have been a hostile military invasion into an armed
“support” or “assistance” mission at the invitation of Haiti’s rulers. The same US
forces that aborted their invasion arrived days later as a multinational force that
treated Haitian state forces as allies, even as the coup’s leaders were forced into
exile and supplanted on October 15, 1994, by a reinstated President Aristide.
Although the organized violence reached sufficient levels, without arms the Hai-
tian leftist opposition did not cohere as an insurgent army and relatively few state
agents were killed.
Scholars differ about the causes of this armed conflict. As mentioned above,
religious difference was not a factor. Racial tensions between light-skinned and
darker-skinned Haitians underlay political tensions and exclusion but have not
been used as mobilizing elements since the Duvalier period. Economic ten-
sions certainly underlay the political instability in Haiti, and thus poverty as a
risk factor seems relevant. However, it is hard to point to any given economic
trend, decision, or event as a salient trigger in the violence and coup that began
in 1990–91. Per capita gross domestic product (GDP) fell in 1987 but climbed
steadily afterward, defying Collier and colleagues’ causal model.5 It was in the
late 1980s and early 1990s that the country underwent its first significant neolib-
eral reforms of trade and macroeconomic policies (World Bank 2002, 15–16).
This correlation with the onset of armed conflict confirms Hartzell and Hoddie’s
Nonseparatist Recurrences of Civil War 131

(2010) work on the negative impact for civil wars of the International Monetary
Fund’s structural adjustment packages, but it flies in the face of Collier’s recom-
mendations to adopt such policies for stabilization.
Two pertinent factors seem to be the resurgence of repression by the 1988–90
coup government led by General Prosper Avril that preceded Aristide’s election,
and the diffusion of democratization in the region that raised uncertainty amid
more competitive electoral events. Avril engaged in brutal repression, especially
targeting leftist critics and followers of Aristide, who was then a parish priest and
vocal critic of the government. Violence against the Haitian police and military
also increased. The fall of the Berlin Wall and the spread of electoral regimes in
Latin America and Central and Eastern Europe, which also extended to much
of Africa in the early 1990s, increased pressure on Avril’s regime to permit elec-
tions. Much of the violence in Port-au-Prince centered on the period preceding
the country’s first truly open and competitive elections in December 1990.
To the extent that recurrent political violence among elites, especially coup
attempts and political assassinations, signifies a weak state, then state weakness
is perhaps the single most apparent factor. Yet the mechanisms for how state
weakness is triggered and when and why specific armed conflicts emerge, even if
violence is neither widespread nor in the form of armed combat between formed
armies, remain an open question.
In the case of the initial armed conflict—especially the coup d’état of Septem-
ber 1991—some acts of exclusion by Aristide are apparent; however, these fac-
tors were eclipsed by the broader anti-Aristide (and anti–popular democracy)
sentiments that underlay the support for the coup. Even before president-elect
Aristide was sworn into office the first time on February 7, 1991, he faced a coup
attempt by Duvalier’s former interior minister (and director of the infamous Ma-
coutes), Roger Lafontant. The failed coup attempt was followed by the lynching
of some seventy former Duvalierists by pro-Aristide mobs (Serrill 1991). Ongo-
ing organizing against and demonizing of Aristide persisted, even though the
reforms he launched at this point were of a “moderate social democratic” nature,
neither barring the elite from access to state offices nor jeopardizing the coun-
try’s free market system (Dupuy 2007, 112). Aristide did, however, seem to con-
done or encourage deadly mob violence against those wealthy elites who tried to
thwart the will of the people in seeking to transform the exploitative and unequal
society and economy (Dupuy 2007, 122; 2009, 166–68). As one scholar puts it
(Dupuy 2007, 168–69), “The military, backed by the dominant classes, would
have overthrown Aristide in any case for what he (then) stood for. But he gave
the justifications for their actions that they would not have had otherwise.”
The strength of the popular movement that helped bring Aristide to power also
ensured that the coup would lead to broader violence, both by the poor support-
ers of Aristide and, most especially, against his supporters by the coup-spawned
132 Examining the Cases

government led by General Raoul Cedras. Similar to what I show below in the
cases of Rwanda and Burundi, political exclusion, backed by deadly violence,
underlay the rise in organized violent conflict in the early 1990s in Haiti, but it
was not a precipitating factor in the same way that it would prove to be in the
2003–4 recurrence.

The Onset of Peace in Haiti

Upon taking office, the Cedras government immediately targeted hundreds of
Aristide supporters with political violence. The violence—which was carried out
by the army, the police, local sheriffs, and a paramilitary group called the Front
for the Advancement and Progress of Haiti that was supported by the US Central
Intelligence Agency—also targeted many human rights groups and civil society
organizations (Human Rights Watch 1993). The coup and its attendant political
violence prompted international denunciation and an immediate and ineffective
economic embargo by the Organization of American States (OAS) (Bromley
2007, 3). The OAS and the United Nations created a joint human rights mission
called the Civilian Mission in Haiti. Eventually, UN Security Council Resolution
841 demanded Aristide’s return and imposed economic sanctions and an arms
embargo (Bromley 2007, 2).
Despite a negotiated agreement on the return of the elected president that
was brokered in July 1993, the military reneged and initiated a new round of po-
litical assassinations, leading to expanded UN sanctions and UN Security Coun-
cil Resolution 941, which mandated the use of force to restore President Aristide.
The administration of US president Bill Clinton mustered a multinational coali-
tion and declared its intention to intervene militarily before a delegation led by
former president Jimmy Carter secured the capitulation of September 1994.
The role of international actors proved significant. The economic sanctions
proved sufficient to bring the Cedras regime to the negotiating table, but they
were insufficient to force its ouster (Bromley 2007; Malone 1998). Only with the
direct threat of UN-authorized military intervention did the regime cede power.
The departure of Cedras and the restoration of President Aristide led to a brief
US-led peacekeeping mission of 21,000 troops. This was followed by a new UN
peacekeeping force, which was replaced by a UN support mission, whose pres-
ence was extended, in a modified and reduced form, until 2001, when the last
UN peace mission departed.

Haiti’s Recurrence of Armed Conflict

Following the 1994 restoration of Aristide, he was persuaded to accede to the
prior constitutional electoral schedule and permit new elections for president
Nonseparatist Recurrences of Civil War 133

in 1995. His prime minister, René Preval, was elected president for the first time
in a flawed process whose result was widely accepted internally and externally.
Under Preval, relative stability prevailed, facilitated by a large international UN
and OAS presence and UN troops. However, Parliament was not functioning for
much of his tenure, and by 1997 the government had gained a reputation for cor-
ruption (Dailey 2003). In 1996, the Lavalas political movement split as Aristide
formed his Famille Lavalas (Lavalas Family) political party with mass support.
But he left behind many disenchanted members of the intelligentsia and artists
who resisted his personalistic style of leadership.
Political violence began rising in anticipation of the May 2000 parliamentary
and municipal elections and the November 2000 presidential ballot. A number
of factors rendered the elections questionable and fueled instability that year:
postponed balloting, violence against some candidates and campaigns, manipu-
lation of the Senate results, politicization of the Provisional Electoral Commis-
sion, and polarization over the imminent return of Aristide to a second term
(OAS 2000). Referring to the elections of 2000, one journalist (Dailey 2003)
concluded, “Gross electoral fraud by the ruling party has deprived the entire po-
litical apparatus of legitimacy.” Although the breakdown in government under
Preval had prompted the World Bank and others to suspend aid, the inaugura-
tion of Aristide’s second administration led to the suspension of all official inter-
national assistance except humanitarian aid (Dailey 2003).
In 2003, tensions between the Aristide government and leftist and rightist
opponents grew. A former pro-Aristide gang leader, Amiot Metayer, formed
what he called the Cannibal Army to oppose the government. Metayer’s struggle
gained prominence until he was killed, presumably by government forces, in Sep-
tember (Beidas, Granderson, and Neild 2007). In January 2004, as the country
celebrated two hundred years of independence, thousands of people marched
against the government. On February 5, the former Cannibal Army (now re-
named the Artibonite Resistance Front for the Overthrow of Jean-Bertrand
Aristide), stormed the police headquarters in Gonaives and released dozens of
prisoners. The group was led partly by the slain Metayer’s brother, but also by
many former soldiers from the Duvalier period who had fled the country to the
Dominican Republic. In the next few days the uprising spread to a dozen other
cities, often facing little government resistance. On February 22, after the group
took control of the country’s second-largest city, insurgent leader and ex-soldier
Guy Philippe announced that he would take the capital within two weeks unless
Aristide stepped down.
Aristide and the rebel forces both rejected any powersharing arrangement.
The president also rejected the French government’s suggestion that he abandon
power. As forces advanced on the capital, Aristide resigned his post on Leap Day,
and was escorted from his home to the airport and flew into exile (Robinson
134 Examining the Cases

2007; Farmer 2004). Evidence of US tacit, and perhaps financial, support for
the uprising exists, and many analysts have debated whether US officials coerced
Aristide into his resignation and flight. In any case, a new interim government led
by Gerard Latortue was installed with the United States’ blessing. That govern-
ment engaged in continued repression and killings of Aristide’s supporters for
several months. Immediately upon Aristide’s departure, the UN Security Coun-
cil authorized a new UN Stabilization Mission in Haiti, deploying 3,500 troops
initially. In 2006, a new election was held in which former President René Preval
was elected to a second full term.

Sorting through the Explanatory Factors in Haiti

Dupuy (2009, 169) views Aristide’s predatory behavior in his approach to the
state, which excluded both former enemies and friends from access to it in a man-
ner reminiscent of Charles Taylor, as the main factor in the decision to oust him
in the uprising of 2003: “That Aristide and the FL [Famille Lavalas, his party]
would contemplate prolonging their hold on power indefinitely was in keeping
with the historical tradition of using the state as a means of social promotion and
personal enrichment. . . . These facts more than anything else Aristide did during
his second term were the principal reasons the middle- and dominant-class op-
position coalitions backed by the imperial troika [the United States, France, and
Canada] wanted Aristide out of power.”
These descriptions make it very difficult to distinguish Haiti’s weak state in-
stitutions from the exclusionary behavior of Aristide. State weakness is unques-
tionably a significant factor in the unrest emblemized by both the violence of
1990–94 and the political violence of 2003–4.
However, weak state institutions do not require that rulers consistently en-
gage in exclusionary behavior. Indeed, many intellectuals and leaders within
Aristide’s own political movement sought to institutionalize the party and move
away from the personalistic leadership characteristic of weak states. Aristide de-
liberately rejected and broke with this faction when he formed a new political
party, the Lavalas Family, in 1996. Here weak state institutions constituted a risk
factor, as elsewhere, but they did not predetermine Aristide’s or other leaders’
decisions to exclude former enemies and friends from access to state power.
President Aristide increasingly ruled in a manner that directed state resources
to fewer of his political supporters, which permitted known business friends to
profit. According to Dupuy (2009, 170), his government became a “mafia-like
organization” that engaged in “rampant corruption, drug trafficking,” and reli-
ance upon gangs to engage in violence and the threat of violence against regime
opponents. In a nepotistic fashion, Aristide abandoned these allies once he was
again in office, and he relied instead on those with long-standing personal loy-
Nonseparatist Recurrences of Civil War 135

alty to him and his wife. Political killings escalated, and the police were seen as a
political tool and pervaded by corruption linked to an expanding drug trade. In
June 2003, the head of the Haitian National Police fled the country, fearing for
his life because he had resisted attempts by the president and members of Parlia-
ment from his party to force favoritism in promotions and the hiring of unquali-
fied members for the force (Lynch 2003). In August 2003, the United States–
based Committee to Protect Journalists announced that Haiti had rejoined Cuba
as one of the two most dangerous countries in the Americas for journalists, citing
alleged state complicity in the murder of two journalists and the exile of thirty
others (Regan 2003).
Middle-class Haitians, as well as many from the poorer classes, became disil-
lusioned by the predatory approach to the state and the more overt antidemo-
cratic practices carried out with the electoral machinery and the legislature. As
the journalist Peter Dailey wrote (2004), “Aristide’s growing authoritarianism
has been denounced by virtually every element of the coalition that supported
his rise to the presidency in 1990: the priests and laypersons of the liberation
theology wing of the Haitian church, the network of grassroots organisations,
peasant co-operatives and labour unions, and every single Haitian intellectual or
artist of note. Aristide tried to compensate for the defection of his former sup-
porters by recruiting criminal gangs.”
Ironically, elections became the means by which a popular, populist politician
built a semiauthoritarian regime, over the dissidence of those who represented
peasant organizations, artists, labor unions, and other civil society organizations
(Fatton 2002). In this sense, Aristide’s behavior was exclusionary not only to-
ward his former enemies but also toward his former allies, a pattern that is more
decisive in the recurrences of armed conflict in Zimbabwe and East Timor.
Supporting Collier’s claims, the four years preceding the armed conflict of
2003–4 showed an unusual economic decline. Although GDP per capita rose
from 1996 to 2000, it fell in 2001, 2002, and 2004, before rebounding in 2005
and climbing until 2009 (Fasano 2009, 5). That decline in the early 2000s, like
much in Haiti, was highly conditioned by external factors, primarily the reduc-
tion in international aid and loans from the United States, the international finan-
cial institutions, and eventually Canada and European bilateral donors.
Because of the proximity of Haiti to the United States, it is difficult to distin-
guish the role of the international factors, especially as they reflect great power
interests, from regional factors in Haiti. Although the United States deployed
force to restore Aristide to power, neither the Clinton nor the Bush administra-
tion was ever comfortable with his leftist political platform, especially his resis-
tance to neoliberal economic prescriptions. The Clinton administration initially
conditioned aid and loans, eventually suspending aid to the very police force
that it had created and underwritten (Beidas, Granderson, and Neild 2007). The
136 Examining the Cases

cutoff of all international assistance except emergency humanitarian relief once

Aristide was elected anew in 2000 constituted a key factor in the reduction of
private capital flows to the country and its economic decline.
The presence of UN-authorized international peace missions also corre-
sponds to periods of relative stability. The recurrence of armed conflict in late
2003 and early 2004 occurred within three years of the withdrawal of the last
UN peace presence, and five years after the last UN troops departed. In response
to political instability in 2001 and the departure of the UN mission, the OAS
created a special mission in the country in 2002. However, that mission, which
included twenty-five police advisers, proved unable to prevent instability and
eventual armed conflict. In sum, a number of regional and international factors,
which are hard to differentiate, played an important role in recurrence in Haiti.

East Timor in 1975–99 and 2006: Liberation,

Statehood, and Exclusion
East Timor, at only 15,000 square kilometers about half the size of Belgium, lies
on the eastern half of the island of Timor. The country is a highly rural agrar-
ian society and ranks among the dozen poorest countries in the world.6 Tribal
identities are important for most Timorese, and the state’s formal institutions
are extremely weak, with low levels of human resources for education, training,
and professional preparation. The country attained independence in 2002 af-
ter three years of UN administration following bloody violence that marked the
final spasms of an armed struggle for independence that commenced in 1975.
At first glance, East Timor appears to be a highly homogenous society. The
vast majority of its population are proclaimed Catholics (up to 98 percent), and
there are tiny Muslim and Protestant minorities (CIA 2010). Throughout both
conflicts, religious difference did not serve as a mobilizing issue. Most Timorese
also share certain racial or ethnic classifications (CIA 2010), but the population
speaks sixteen different native languages, including the official Tetum. However,
different regional, local, and kin-based relationships also produce social distinc-
tions in East Timor. Extended families coincide with different villages histori-
cally, called “houses,” that intermarry with other houses (Hohe 2002). Groups
of houses have historically formed the basis for the kingdoms that exercise sig-
nificant local autonomy.
Timorese themselves distinguish among peoples based on their area of ori-
gin or their native language (Kingsbury 2009, 8–14). These distinctions prove
relevant for analyzing East Timor’s recurrence of armed conflict. The Timorese
people also distinguish between peoples from the western portion of East Timor,
kaladi, and those from the eastern end of the territory, firaku (Harrington 2007).
The kaladi are considered to be more reserved and taciturn, in contrast to the
Nonseparatist Recurrences of Civil War 137

more talkative and outgoing firaku (Harrington 2007). These categories—each

of which encompasses multiple language, religious, and tribal groups—are truly
geographical rather than racial or ethnic in the conventional sense.

Civil War, Intervention, Occupation, and Resistance

East Timor was a Portuguese colony from the sixteenth century until the 1974
revolution in Lisbon led to a process of decolonization. Three political parties
emerged in Timor that engaged this process. Two—FRETILIN (the Revolu-
tionary Front of East Timor) and UDT (the Timorese Democratic Union)—
advocated independence, although the UDT favored a gradualist approach. A
third—APODETI (the Popular Democratic Association of Timor)—sought in-
tegration with Indonesia, as had occurred with West Timor in 1947 (Elliott 1978,
238–39). During the years 1974–75, negotiations between the FRETILIN–
UDT coalition and Indonesia occurred, even as the latter expressed disinterest
in annexation.
The initial war started as a coup and countercoup, followed by three months
of armed conflict among Timorese political parties before the occupation and its
severe repression amid armed resistance. In June 1975, the Portuguese adminis-
tration set a time line of October 1978 for full independence. However, the UDT
executed a coup on August 11, 1975, claiming that it was preempting a planned
coup by FRETILIN (Elliott 1978, 239). Armed clashes broke out between the
three factions within the territory. FRETILIN gained the advantage and drove
the UDT combatants into West Timor, where the latter allied with Indonesian
covert forces (Hoadley 1975, 126). FRETILIN and its armed wing, FALINTIL
(Armed Forces for the Liberation of East Timor), claimed control over the capi-
tal, Dili, and most of the territory by late August. In response to intelligence indi-
cating a planned Indonesian invasion ( Jardine and Pinto 1997, 15), FRETILIN
claimed independence on November 28, 1975, and set up a government. How-
ever, two days later, the other parties countered with a declaration of integration
into Indonesia. The initial phase of the internal armed conflict, among Timorese
factions before the invasion, left about two thousand persons dead, according to
FRETILIN’s leader, Jose Ramos-Horta (1987).
Indonesia, which FRETILIN claimed had infiltrated the country covertly in
September, overtly sent troops into East Timor on December 7 (Elliott 1978,
239). They quickly gained control of Dili, but they faced determined resistance
from FRETILIN forces in the countryside. Ten days later, the UDT-APODETI
coalition formed the Provisional Government of East Timor. In May 1976 this
government organized a People’s Assembly that formally requested integra-
tion with Indonesia, which the Indonesian government embraced in July 1976
(Elliott 1978, 239). FRETILIN denounced the assembly and spearheaded the
138 Examining the Cases

armed resistance for the next twenty-five years, despite that fact that it received
virtually no assistance from abroad (Steele 2002, 77). After the annihilation of
tens of thousands of civilians by Indonesian forces and their allies, it became
apparent by the late 1970s that ending the occupation would require a political
strategy that rested on a broad civilian resistance movement, although the insur-
gent military forces retained symbolic importance (Rees 2004, 2).
As part of the political effort, FALINTIL’s commander, Xanana Gusmao,
sought to diminish long-standing Timorese disunity that would become impor-
tant in the 2006 recurrence. In 1987 he created the National Council of Mau-
bere Resistance (Conselho Nacional de Resistência Maubere, CNRM) to create
a broad, inclusive, and nonpartisan political movement. FALINTIL separated
itself from the political party FRETILIN, Xanana resigned from the party, and
FRETILIN ceased to claim to be the sole legitimate representative of the Ti-
morese people (Simonsen 2006, 579). In 1998 the CNRM reconstituted itself
as the National Council of Timorese Resistance (CNRT), without changing its
nonpartisan, unifying aspirations.
The numerous casualties of the occupation and the resistance lead many to
overlook the internal character of the original armed conflict. The occupation
involved the killing of between 100,000 and 200,000 persons (Niner 2001, 19).
Although the violence, oppression, and social transformation of the occupation
defined the last quarter of the twentieth century in East Timor, the armed con-
flict’s origins lie in internal divisions over independence in the context of Portu-
gal’s decline and withdrawal as a colonial power. Indonesia was a powerful strate-
gic actor that allied with internal factions; however, it would be an exaggeration
to attribute the origins of the original armed conflict primarily to Indonesia.

The End of the First War

Just as the first war commenced in the wake of the democratization of one oc-
cupier, so too did democratization in the successor occupier spark the end of
that occupation and war. In 1998, Indonesian president Suharto fell from power
amid initial steps toward democratization. After coming to power, President B. J.
Habibie initially rejected the possibility of independence for East Timor, and he
instead proposed a new relationship of special autonomy within Indonesia. In
the face of insistence by Portugal and the United Nations for some expression of
self-determination by the Timorese, he shifted course, and in January 1999 he an-
nounced a referendum on East Timor’s status (Martin and Mayer-Reickh 2005,
126–27). The UN Security Council authorized the creation of the UN Mission
in East Timor in June 1999 to administer the vote. On August 30, 1999, with 98
percent of those registered voting, 78.5 percent of East Timorese opted for out-
right independence rather than a special autonomous status within Indonesia.
Nonseparatist Recurrences of Civil War 139

The war for independence ended with a horrible spike in violence in reaction
to this referendum. A massive and deliberate campaign of violence, carried out
almost entirely by Timorese militias opposed to independence, broke out in Dili
two days after the vote and spread to the countryside. Although serious violence
had occurred in 1999 before the referendum, the scale and scope of this violence
had not been seen in years—it involved widespread rape, looting, and the killing
of some 1,400 persons (Cohen 2006, ii). Some 250,000 refugees fled to West
Timor, many more were displaced, and the great majority of schools and health
facilities were destroyed (Steele 2002, 77). The militias had received training,
organization, and equipment from the Indonesian armed forces, whose troops
also participated in much of the killing and the destruction of many of the main
buildings of the capital and elsewhere (Cohen 2006, ii).
On September 12, 1999, the UN Security Council authorized a Multinational
Force with Indonesian acquiescence, which deployed a week later under Austra-
lian leadership with the name International Force for East Timor (INTERFET).
INTERFET was eventually able to reduce the violence, because many anti-
independence militia members fled among civilians to West Timor. In October
1999, the UN Security Council authorized the UN Transitional Administration
in East Timor (UNTAET), which effectively exercised full sovereignty over the
territory. A special representative of the UN secretary-general was both head of
the UN peacekeeping operation and the administrator over the territory, and his
decrees had the force of law (Chopra 2000). Although there was no peace agree-
ment, the 1999 postreferendum violence and the authorization of INTERFET
marked the demise of Indonesian rule and the inevitability of internationally em-
braced independence for East Timor. Although violent attacks persisted through
late 1999, the war was essentially over.

The UN’s Role and Independence

UNTAET exercised a greater degree of state sovereignty over East Timor than
any UN entity ever had done. Although the CNRT had long favored a period of
UN tutelage before full Timorese self-government (Suhrke 2001, 4), it resented
what it regarded as its wholesale marginalization in the peacekeeping operation’s
planning and initial deployment (Suhrke 2001, 9–12; Steele 2002, 78). Early
proposals of “dual desks” in planning were shelved, and a Timorese consultative
commission was only brought in after UNTAET had been criticized for its inad-
equate inclusion of Timorese representatives in decision making (Chopra 2000,
2002; Simonsen 2006; Suhrke 2001; Steele 2002). Before the autumn of 2000,
the UN transitional administrator consulted primarily with Xanana Gusmao,
but the UN mission resisted relying on the CNRT, which was widely perceived
as legitimate, as its national-level counterpart.7 Shepard Forman believes that
140 Examining the Cases

this shortsighted decision contributed to the CNRT’s breakup into five factions
in 2000. After that time an East Timor Transitional Authority was created that
oversaw some administrative capacities, although only two of thirteen districts
were under Timorese leadership in March 2001 (Suhrke 2001, 13). Some believe
that these tensions later undermined the UN’s standing and thus its ability to
influence the factional tensions leading up to the 2006 violence.
Upon independence in May 2002, UNTAET’s mandate was changed to sup-
port the new Timorese state through the renamed and downsized UN Support
Mission in East Timor (UNMISET).8 Between the withdrawal of Indonesia’s
forces in 1999 and independence in 2002, more than 80 percent of the 250,000
refugees had returned from West Timor (Martin and Mayer-Rieckh 2005, 135),
and thousands of poor and displaced Timorese had converged on Dili. Political
wrangling, including courting former prointegration political factions, prevailed,
but without extensive violence leading up to the election of a Constituent As-
sembly in August 2001 and presidential elections in April 2002. Partly because
of the UN’s eagerness to depart swiftly, insufficient time to hold elections for a
Parliament left the Constituent Assembly converting itself into the Parliament
(Goldstone 2004, 87–90). FRETILIN gained fifty-five of the eighty-eight seats
in the Constituent Assembly in 2001, and no other party received more than
seven seats. Its political dominance was cemented with the Assembly’s vote
to extend the life of the initial election into a five-year Parliament without a
new election.

The Internal Armed Conflict of 2006

In 2006, tensions between rival factions of the security services became heated.
Under UNTAET, a new army, the Defense Forces of Timor-Leste (FDTL) was
formed as the country’s army, and many combatants from the independence
struggle demobilized. Although FALINTIL’s commanders had been involved in
the selection of the new force of 1,600 persons, many former FALINTIL mem-
bers felt that the process was unfair and that their patriotic service (widely sup-
ported by the population) had gone unrecognized. Both FRETILIN and many
ex-combatants were dissatisfied at the prevalence of Xanana loyalists in the new
and smaller army. A disproportionate number of commanders also originated
from the Firaco district (Simonsen 2006, 589–90). Shoesmith (2003, 247)
states, “The core of the new defense force is identified, then, not only with the
president and commander-in-chief, rather than the government, but with one
ethnic collectivity rather than another.” Veterans organizations began to agitate
against the new army (renamed FALINTIL-FDTL, or F-FDTL, once FRETILIN
came into power) and for a return to FALINTIL (United Nations Special Com-
mission of Inquiry for Timor-Leste 2006, 19).
Nonseparatist Recurrences of Civil War 141

In January 2006, 159 members of the F-FDTL submitted a petition to the

president seeking redress for a number of grievances, including discrimination
against westerners by the eastern-dominated force.9 After receiving no response,
the petitioners abandoned their barracks in February, and in March the defense
minister dismissed almost 600 petitioners and absentee soldiers (of some 1,600
total members) from the F-FDTL. Because FRETILIN enjoys strength in the
eastern part of the country, it was seen as complicit in the discriminatory behav-
ior. A week after the mass dismissal, President Xanana Gusmao criticized the
decision and acknowledged east / west tensions, perhaps fueling them.
On April 28, 2006, violence erupted between petitioners on the fourth day
of a protest outside the Government Palace, killing two civilians and injuring
six others. The army and the new National Police of Timor-Leste (PNTL) were
called in to restore order, as violent attacks took isolated lives and caused injuries
over the subsequent days. On May 3, Major Alfredo Reinado deserted from the
Military Police, taking with him other officers and troops of that force as well as
civilian PNTL officers and weapons. In early May, Interior Minister Rogelio Lo-
bato armed two groups of civilians with weapons and ammunition of the PNTL
border unit. Although the F-FDTL was withdrawn by May 4, tensions and vio-
lent incidents involving youth gangs and ex-combatants continued throughout
May. In late May violence resurged in several incidents, including one where
active army officers, police, and civilians attacked the police headquarters, kill-
ing eight policemen and wounding twenty-five others who had surrendered and
were exiting unarmed with UN accompaniment (UN S / 2006 / 628, August 8,
2006, para. 9). In addition, dozens of houses were burned, including the home
of the in-laws of Lobato, where six civilians died.
The mass violence ended by early June, when the government experienced
the resignations of Defense Minister Lobato, Foreign Minister Ramos-Horta,
and finally, on June 25, Prime Minister Alkatiri. In these three months from April
to June 2006, at least thirty-four people died, dozens were injured, and 15 per-
cent of the country’s population was displaced (United Nations Special Com-
mission of Inquiry for Timor-Leste 2006). In this sense, the violence barely rates
as sufficiently deadly to qualify as an internal armed conflict. Yet the social and
political consequences were far-reaching. Most visibly, 150,000 Timorese (some
15 percent of the country’s population) fled their homes in fear of the persistent
battles (Goldsmith 2009, 122). The government fell and was reconstituted with
Ramos-Horta as prime minister, but confidence in the Timorese state and the
political parties had been deeply shaken. Australia deployed a new peacekeeping
force in 2006, and the UN authorized a newly mandated peacekeeping opera-
tion to foster security and assist further with state building. Despite unsuccessful
assassination attempts on Ramos-Horta and Xanana Gusmao in 2008, broader
violence has not resurfaced to the extent of 2006.
142 Examining the Cases

Exclusionary Behavior in Independent East Timor

East Timor’s recurrent violence in 2006 had more than a single cause. State
weakness, the withdrawal of international troops, and missteps by the United
Nations and the major powers contributed to the renewed conflict. However, ex-
clusionary behavior also played a significant role. That marginalization centered
on the divisions within the broad coalition embodied by the CNRT. Rather than
exclusion of former enemies, the divisions that became violent lay with former
allies in the liberation struggle who turned on one another. Problems began even
before the electoral process. Opposition parties realized in 2001 that FRETILIN
had been able to accumulate the power that was gradually transferred to the Ti-
morese by the UN mission. After independence, the FRETILIN-dominated leg-
islature governed with little regard for other parties. In the first sixteen months
after independence in May 2002, the Parliament had adopted only sixteen laws,
most of which had been initiated by the governing party and underwent very few
changes. Almost twice as many decrees—issued by the prime minister or the
president—had the force of law without any chance for revision by the opposi-
tion (Simonsen 2006, 582).
The marginalization by FRETILIN of its former coalition partners within the
victorious resistance movement continued after independence. The United Na-
tions Special Commission of Inquiry for Timor-Leste (2006, 19) found that “the
power imbalance between [FRETILIN] and its political opponents has been an
issue since 2002 and informed the crisis of April and May 2006.” It noted that
the last opposition figure remaining in the Parliament’s leadership resigned in
March 2005 (p. 19). Regan’s (2003) interviews with opposition members sug-
gested that “they were fairly bitter, an indication, in my view, of the opportunities
for a more inclusive legitimizing process which were perhaps lost.” Simonsen
(2006, 582) echoed these findings: “The opposition’s resentment of Fretilin’s
exclusivist way of doing politics is increasingly shared by other actors, in particu-
lar the young, educated intelligentsia.”
The country’s security forces—the army and the police—also reflected ex-
clusion of a different sort. The FALINTIL army and its leadership had from its
early days come predominantly from the eastern portion of the territory, given
its more favorable terrain and its distance from Indonesia. By 2004 the Timorese
authorities had largely remedied the regional imbalance among the rank and
file, but easterners continued to dominate the senior officer corps (Bowles and
Chopra 2008, 239). As Harrington (2007) notes, “The result was command
structure within the F-FDTL largely dominated by those from the East, with
those from the West generally feeling discriminated against without opportunity
for advancement within the forces. . . . A letter of complaint singled out a high-
ranking F-FDTL commander for his alleged public humiliation of kaladi sub-
Nonseparatist Recurrences of Civil War 143

ordinates, referring to them [as] ‘sons of pro-Indonesia militia,’ a huge insult for
those who sacrificed many years of their lives in service of independence.”
The F-FDTL leadership introduced and perpetuated the east / west discrimi-
nation, ignoring complaints from those inside and outside the army about its
significance. Although Xanana helped foster this distinction, FRETILIN’s own
history also reflected this east / west divide, visible in its demonstrably greater
support from the east in the 2001 elections. FRETILIN’s “top down, non-
inclusive way of doing politics” (Simonsen 2006, 584) coincided with the per-
ception of favoritism over easterners and unwillingness to respond to evident
discrimination. This exclusionary conduct constituted the single most salient
trigger for the renewed conflict in 2006, but others also played a role.

State Weakness in East Timor

It is difficult to imagine a society with less state capacity than East Timor on
the heels of the occupation and the 1999 referendum. During the occupation,
Indonesians came into East Timor and occupied virtually all meaningful admin-
istrative positions. The Indonesian army and its police force constituted the sole
recognized armed forces of the territory. In the aftermath of the violence sur-
rounding the 1999 referendum, most Indonesian bureaucrats fled the country,
leaving hardly anyone with experience to fill the administrative slots at local or
provincial or even national levels in the state ministries (Hood 2006, 62; Stroh-
meyer 2001). Postoccupation East Timor thus came to emblemize a territory
with extremely low human resources and state capacity.
The Defense Ministry also showed low capacity, with many posts unfilled and
personal trust governing relations among its leaders (Rees 2004, 11–12). Several
analysts blame the UN mission for ceding the management of the army to the
leaders of the resistance movement and for not doing more to institutionalize ci-
vilian control and also procedures and policies (Hood 2006; Rees 2004; United
Nations Special Commission of Inquiry for Timor-Leste 2006). While acknowl-
edging the role of exclusion, the United Nations Special Commission of Inquiry
for Timor-Leste (2006, 2) emphasized weak state institutions in its explanation
of the renewed conflict: “It is the view of the Commission that the crisis which
occurred in Timor-Leste can be explained largely by the frailty of State institu-
tions and the weakness of the rule of law. Governance structures and existing
chains of command broke down or were bypassed; roles and responsibilities
became blurred; solutions were sought outside the existing legal framework.”
International organizations have a tendency to overly attribute problems to
insufficient capacity and insufficient coordination, as this analysis leads to tech-
nical solutions (favoring their programs) rather than more difficult diplomatic or
structural measures. Few other analysts emphasized capacity as the main causal
144 Examining the Cases

factors of the 2006 violence. However, the low level of state capacity correlates
highly with the problems in East Timor, and this is difficult to discount. The low
level of institutionalization of the army, as well as of the hastily formed and highly
personalized and political police force, made it difficult to control the conduct
of these forces and constituted an absence of channels for dissent (Rees 2004).

International Troops and Stability in East Timor

East Timor’s 2006 recurrence also offers support for the importance of interna-
tional troops in dissuading renewed political violence in postconflict societies.
In 1999 Australia’s INTERFET troops and then UNTAET’s forces showed that
deploying armed peacekeepers contributed decisively to a reduction in violence.
UNTAET’s force levels reached more than 7,000, falling to 6,200 in 2002, and
then the troop level under UNMISET dropped from 4,776 to fewer than 500.10
Upon UNMISET’s disbanding three years after independence—and exactly
one year before the renewed hostilities of April 2006—it withdrew 469 military
troops, authorizing a new political mission nicknamed UNOTIL with only 15
unarmed military advisers and fewer than 100 civilians and international police
advisers (UN S / Res / 1599, April 28, 2005). Thus by early 2006, the presence of
international peacekeeping troops had fallen from 5,000 in 2002 to these three
dozen unarmed advisers. Despite the recommendation of UN secretary-general
Kofi Annan to extend the presence of over-the-horizon infantry in early 2005, the
UN Security Council never approved that extended capability (Gorjao 2008).
The deployments of foreign troops during the 2006 armed conflict support
the importance of their role in dissuading violence—at least in small, weak
states. In response to a request from Foreign Minister Ramos-Horta, Australia
and other countries deployed about 1,300 troops in late May and early June
2006, as well as more than 200 police (Goldsmith 2009, 121). Those troops
helped quell the violence. In August 2006, the Security Council authorized a
new UN Integrated Mission in East Timor, with significantly more international
police advisers (1,608) and 34 unarmed military liaison officers (UN S / Res /
1704). The UN, Australia, and other states agreed that the international forces—
numbering 475 Australian and New Zealand troops as of 2010—would remain
outside UN command.11

Economic Factors in East Timor

East Timor is extremely poor. In per capita income it ranks 164th out of the
world’s 183 countries (International Monetary Fund 2010), and about 20 per-
cent of the country lives on $1 or less per day.12 Despite Indonesian investment
in Timorese road systems and increased economic growth during the 1980s
Nonseparatist Recurrences of Civil War 145

and 1990s, the territory remained the poorest in Indonesia in 1996 (Moxham
2008, 9). However, this level of poverty is not considered by Timorese observers
or by international analysts to have been crucial in the outbreak of armed conflict
of 2006. Harrington (2007) emphasizes economic factors more than others, but
he attributes that recurrence to the cultural distinctions between firaku and ka-
ladi, exacerbated by multiple horizontal inequalities, including greater access by
easterners to land, property, and housing after the 1999 referendum. Outbreaks
of violence among youth gangs that mobilized along the east / west divide began
in late 1999 and were suppressed by UN troops and police (Harrington 2007).
In addition, easterners were seen as dominating and as enjoying privileged access
to jobs and control of the markets in Dili after independence. These horizontal
inequalities underlay the exclusion perceived by the western FALINTIL peti-
tioners in 2006. Harrington (2007) argues that UNTAET was ill equipped and
unable to manage the influx of people into the Dili area after 1999 or to manage
land disputes adequately, despite having some recognition of the problems and
having created an office early on to address disputes. In short, economic factors
were not precipitating causes, but they exacerbated the tensions over land and
access to jobs, especially for the demobilized combatants and those people serv-
ing in the army and the police.

International and Regional Factors in East Timor

International factors in the recurrence of conflict mainly refer to the critiques of
the role played by the UN missions and the UN Security Council, especially after
1999. As noted above, some analysts blame the UN missions for shortsighted
policies that inflamed divisions among the Timorese elites, for pulling out troops
precipitately, and for failing to do more to institutionalize a less personalistic and
politicized security sector (Rees 2004; United Nations Special Commission of
Inquiry for Timor-Leste 2006; Gorjao 2008; Bowles and Chopra 2008).
Regional factors are limited largely to the role of Indonesia itself on the small
island of Timor. Indonesia’s looming presence, and its reticence to cooperate
effectively with attempts to clarify and rectify the human rights atrocities of the
counterinsurgency struggle, contribute to a fear in East Timor of renewed in-
terference and intrusion by its most resented former colonial power (UN S /
2005 / 458; West 2007; International Crisis Group 2006a). The new state ad-
opted Portuguese and native Tetum as official languages in 2001, attempting to
sideline Indonesian, which led to the firing and retraining of virtually the en-
tire judicial sector (Bowles and Chopra 2008). Yet that looming presence had
little direct influence on the violent conflict of 2006 or on the management of
tensions within the Timorese political elite or the security sector. Australia has
provided troops that helped quell violence in 2006 and subsequently. Despite
146 Examining the Cases

the thorough internationalization of East Timor for decades, international and

regional factors in its recurrence of armed conflict were remote and indirect.
In sum East Timor’s recurrence derived from several hypothesized factors,
the most salient of which was the perceived exclusion by former allies in the
initial lengthy armed struggle for independence. The weak character of Timor’s
state institutions, in a very poor and agriculture-based economy, contributed to
the ease with which a handful of armed dissidents could challenge the stability
of the state and the security of its top officials. The reversion to armed conflict
occurred a year after the withdrawal of international peacekeepers, even amid
warnings that the absence of an over-the-horizon force might embolden those
seeking a violent division. Primary commodity exports and regional-level factors
played little role in the recurrence. However, the shortsightedness of the UN mis-
sion also fueled the postwar political fragmentation and ability of FRETILIN to
marginalize opposition parties, including former allies.

Zimbabwe in 1972–79 and 1983–87:

Liberation, Statehood, and Exclusion
British Rhodesia first experienced armed resistance in 1964. Two armies of na-
tional liberation carried out attacks that assumed more organization and lethality
upon a December 1972 attack that elicited a severe government response, includ-
ing operations against guerrilla camps located in neighboring colonies. These in-
surgent groups were the Zimbabwe African National Union (ZANU) party and
its armed wing, ZANLA, and the Zimbabwe African People’s Union (ZAPU)
and its armed wing, ZIPRA .13 The two rebel armies had their rearguard in differ-
ent countries, backed by different Communist powers, and operated in different
parts of Rhodesia where different ethnic groups predominated (mainly Shona-
speaking people in ZANLA areas and Ndebele-speaking people in ZIPRA areas)
(Alexander and McGregor 2004). The two armies often fought one another.

The Initial War for Independence, 1972–79

The initial war for independence, coded from 1972 to 1979, was an indepen-
dence struggle by black Africans against white rule by Britain and its claimed
white Rhodesian successor (in an unrecognized unilateral declaration of inde-
pendence in 1965), led by Ian Smith. Ethnic differences converged with deep
economic disparity between whites and blacks. Outside military action by South
Africa fueled both the initial war and, through covert operations, the recurrence
of the 1980s. Natural resources played little role in the origins of the war. The war
was settled by negotiations brokered by the British government, which led to the
recognition of Zimbabwe’s independence in 1979 (Stedman 1993). President
Nonseparatist Recurrences of Civil War 147

Robert Mugabe, the leader of ZANU, was elected prime minister in 1980, and he
subsequently became president and head of the government.
The establishment of a new government in an internationally recognized,
majority-ruled Zimbabwe saw the inclusion by Mugabe of several ZANU of-
ficials in the Cabinet in 1980 (Catholic Commission for Justice and Peace in
Zimbabwe 1997, 7). ZAPU leader Joshua Nkomo (the “Father of Zimbabwe”)
declined the ceremonial post of president but became minister of home affairs.
The latent tensions between the newly empowered ZANU government and the
ZAPU opposition saw isolated attacks in 1981 and 1982, which cost up to 300
lives in 1982. In 1982, after the discovery of an arms cache, which very possibly
had been orchestrated, the government arrested several high-ranking ZIPRA
commanders and ousted ZAPU members from the Cabinet.14 Nkomo later fled
into exile for his life.15
In early 1983, the violence intensified. It took different forms, including ini-
tially attacks by “dissidents” on buses, trains, and stores, mainly against civilians,
with dozens of deaths (Lawyers Committee for Human Rights 1986, 34). Those
active in attacks were mainly former ZAPU guerrillas, and ZAPU leaders helped
organize the rebellion; however, the degree of command and control exercised
by the former guerrilla party over all dissident action is unclear (Lawyers Com-
mittee for Human Rights 1986, 20–22). In a vicious spiral, state military, para-
trooper, and police units stepped up operations against the dissidents alleged to
be members of ZAPU and its ZIPRA armed forces (Catholic Commission for
Justice and Peace in Zimbabwe 1997). In addition, in 1983 government forces
(especially a North Korean–trained unit called the Fifth Brigade, which con-
sisted only of ex-ZANLA members once ZIPRA ex-combatants were marginal-
ized, left, or were ousted) launched widespread attacks against ZAPU civilian
supporters, mainly Ndebele tribe members.16 The government announced in
January 1984 that “dissidents” had in 1983 murdered 120 people, mutilated 25,
and raped 47, and had committed 284 robberies.17
A significant portion of the violence experienced as part of Zimbabwe’s recur-
rent conflict after 1983 consisted of mass atrocities and executions, rather than
combat between uniformed armies. A report by the country’s Catholic Com-
mission for Justice and Peace found that “the presence of the Fifth Brigade in an
area in 1983 meant an initial outburst of intense brutality, usually lasting a few
days, followed by random incidents of beatings, burnings and murders in the
ensuing weeks, months, and years” (Catholic Commission for Justice and Peace
in Zimbabwe 1997, 54).
One extensive study of human rights abuses alleged that about 1,500 Nde-
bele civilians were killed in 1983 alone (Lawyers Committee for Human Rights
1986, 2). Another study based on interviews with former guerrillas stated that
after the war of independence, ZIPRA guerrillas were seen as a “political danger”
148 Examining the Cases

and, “along with ZANU civilians and ‘the Ndebele’ as a whole with whom they
were associated, were hunted down, jailed, tortured and killed” (Alexander and
McGregor 2004, 81–82). Nkomo later claimed that 20,000 Ndebeles were
killed, mostly executed outside combat.
Given the extensive and important role played by other countries in the in-
dependence struggle in Zimbabwe, external actors played a surprisingly reduced
role in the 1983–87 war recurrence. Based on detainee accounts, South Africa’s
government was implicated in providing training and weapons for the rebellious
dissidents (Lawyers Committee for Human Rights 1986, 22–24), which intensi-
fied the strength of the Zimbabwean government’s repressive response (Ohlson
and Stedman 1994, 90). However, third countries played no role in guiding the
strategies of either side. Instead, the resurgent armed conflict was dampened
and then ended by a political deal struck between ZAPU’s Nkomo and Mugabe.
Nkomo was eager to return from exile, and ZAPU gained little through these
violent attacks, especially given the harsh response by the government forces
(Ohlson and Stedman 1994, 91).
Just as the conflict commenced with exclusionary conduct, it ended with in-
clusionary powersharing. In 1987, ZANU-PF and ZAPU concluded a power-
sharing agreement that reinstated posts in the cabinet and made Nkomo the
country’s vice president from 1987 until his death in 1999. Crucial for the power-
sharing deal was the absorption of ZAPU into a renamed ZANU-PF (Patriotic
Forces). Competitive democracy was sacrificed to peace, as Mugabe’s goal of a
one-party state was realized with ZAPU’s incorporation into the ruling party
(Ohlson and Stedman 1994, 91).
Again, exclusionary conduct is central to the recurrence of armed conflict in
Zimbabwe, but it stemmed from a falling out among the victorious allied forces
after one faction went ahead and assumed control of a new state. The dissident
violence, led by former ZIRPA members, stemmed from a perception that for-
mer combatants and Ndebele peoples were being excluded from power by the
new regime. The conflict’s origins, according to Ohlson and Stedman (1994,
90–91), lay in “claims by former ZAPU members that they had been reduced to
political impotence.” ZAPU had “suffered from its exclusion from power.” The
violence differed from that of the East Timor insofar as the targets of the latter
were mainly and initially members of the armed forces and the Timorese police.
However, the dynamics of these recurrences are remarkably similar.

Analyzing Nonseparatist Cases of Reneging on Political Inclusion

The cases of nonseparatist recurrences further illustrate the core findings of this
book. First, as in the five cases already examined, exclusionary behavior played
Nonseparatist Recurrences of Civil War 149

a salient role in recurrence in these four additional nonseparatist cases. In the

Central African Republic, the armed faction led by ex-president Kolingba placed
primacy on their exclusion from the armed forces as their motive for rebellion
in 2001. Like Liberia and Haiti, weak state institutions set the stage, but elected
government leaders exercised agency in excluding certain groups. Even in East
Timor in 2006 and Zimbabwe in 1983, elected, inaugural administrations of
weak new states engaged in exclusionary behavior. These cases all demonstrate
the book’s central thesis.
East Timor and Zimbabwe also suggest a particular pattern of postwar exclu-
sion: that of former allies in a liberation struggle that experience marginalization
by the dominant faction of the victorious armed resistance movement. Sociolo-
gists and political scientists have long noted the violence that is engendered his-
torically by internal revolutionary struggles (Tilly 1975; Skocpol 1979; Trim-
berger 1978). Krain (2000) argues that victorious revolutionary movements
combine repression and accommodation to advance internal power and to avoid
armed challenges, at times failing and experiencing challenges from former allies.
Atlas and Licklider (1999, 37) note this particular phenomenon in post–civil
war societies, as “post-settlement political tensions often arise, not from reopen-
ing fissures between former foes but from deepening divisions among former
allies.” Despite the focus on harmony among former enemies in peacebuilding
theory, the danger of violence among former allies is especially real after victori-
ous revolutionary struggles, especially those of national liberation that tend to
unite disparate political tendencies within a society. On the basis of the cases of
Sudan, Lebanon, Zimbabwe, and Chad, Atlas and Licklider (1999, 37) find that
“certain groups or factions feel that they have not received their just desserts
from the settlement or that the terms of the settlement threaten their interests
or security.” It is unclear what circumstances determine when civil wars end in a
violent falling out among former allies in an internal armed struggle, but given
the cases of Zimbabwe, with its 1983 recurrence, and East Timor, with its situa-
tion in 2006 when former comrades in arms attacked one another, this subset of
war recurrence requires acknowledgment and further attention.
Second, the presence of third-party peacekeeping troops in the weakest states
offers important security guarantees. The case of the Central African Republic il-
lustrates this point, as powersharing and inclusionary behavior experienced sev-
eral ruptures, including within months of the end of the first civil war in 1997.
Yet the presence of peacekeeping forces helped maintain stability through these
difficulties. Within a year of their withdrawal in 2000, armed conflict recurred.
Conversely, the deployment of Libyan, Sudanese, and other troops denominated
as a peace mission—the Community of Sahel-Saharan States—did not, how-
ever, staunch the recurrence of conflict once sparked in 2001. The recurrence
150 Examining the Cases

of armed conflict in East Timor, one year after the withdrawal of the peacekeep-
ing operation in favor of a political mission with a handful of unarmed military
observers, reveals the value of the international troops as a dissuasive presence.
Third, regional dynamics and the initiative taken by neighbors in advancing
or mitigating warfare were apparent. In Liberia both the witting and unwitting
involvement of Guinean and Sierra Leonean state armed forces, along with non-
state actors, abetted the recurrence of war, although this was more pronounced
in the organized violence in the territory of Liberia’s neighbors. In Haiti, the US
government played a crucial diplomatic role in signaling support for those who
would undermine Aristide’s government after 2000. Perhaps their support for
the rebels was material as well. In the Central African Republic, the troops of
France and neighboring states (especially Libya) provided a lifeline for Presi-
dent Patassé at crucial moments, and the downgrading of their commitment and
deployments facilitated the slide into recurrent armed conflict. In East Timor,
Indonesia was an important factor in the initial war, even though regional and
external factors played little direct role in the 2006 recurrence.
In the cases of East Timor and Zimbabwe, reneging on presumed commit-
ments by elected postwar governments also triggered renewed armed conflict.
Alongside these two cases, the cases analyzed above—the Central African Re-
public, Liberia, and Haiti—demonstrate the relevance of weak state institutions
as a backdrop for exclusionary behavior and add weight to that alternative expla-
nation. Other explanations, including that of reliance on natural resources, were
not nearly as salient as the decisions surrounding the exclusion of groups from
particular ethnicities or geographic regions.

Burundi in 1972, 1988, and 1993–2002 and Rwanda in 1962–

65 and 1990–Present: Chronic Exclusionary Behavior
Two cases—Burundi and Rwanda—are here presented jointly because of their
similarities and interrelated patterns of armed conflict and mass violence during
the past five decades. Their joint treatment embodies the importance of the re-
gional level of analysis for understanding and accounting for war recurrence. In
addition, they suggest that “chronic exclusion” is a form of exclusionary behavior
that plays an important role in the recurrence of armed conflicts. As with many
cases of genocide and other forms of one-sided violence, it is especially difficult
in these two cases to distinguish between one-sided violence and armed con-
flict as defined between two organized armed parties (Human Security Research
Group 2009). However, the characteristics of armed conflict and its recurrence
exist. Against the backdrop of chronic exclusion, these cases also show that other
factors such as democratization, state weakness, and regional dynamics played
a role in recurrence. The degree and form of “chronic exclusion” remain under-
Nonseparatist Recurrences of Civil War 151

specified here, given that numerous cases of oppression and discrimination can
coexist with the absence of armed conflict. Nevertheless, in seeking to explain
these two interrelated cases, scholars repeatedly cite exclusion as a central factor.
The succession of internal armed conflicts in Burundi and Rwanda since the
1960s demonstrates the importance of the regional level of analysis. Some of
the most prominent theorists of mass violence in the two countries—including
Lemarchand (2004), Mamdani (2001, 2002), Prunier (1995), and Uvin (1998,
1999)—concur on the important role of cross-border dynamics in understand-
ing the origins of warfare and genocide in the “Great Lakes” area encompassing
and extending beyond Rwanda and Burundi. Consequently, I present both cases
here in a parallel, simultaneous fashion. While acknowledging that the processes
of violence differ in the cases (Uvin 1999, 265), I examine these two case stud-
ies in a single narrative, reflecting a structured-focused analysis of each case and
the interaction between them, presented in a chronological fashion that alter-
nates between each case based on stages of conflict and relative stability. The
structured, focused approach for each case is retained, but I analyze one set of
hypothesized factors for each case before moving on to the next set of factors.
These sets of recurrences of warfare stretch the boundaries of our concepts
of civil war and stable peace. The character of violence often took the form of
mass atrocities, or “selective genocide” (Lemarchand 1975), more than civil war
or even internal armed conflict. In these episodes of mass violence, there have
been “rebels” organized under some command. However, mass civilian violence
has often unfolded alongside these internal armed actions. Moreover, it has of-
ten been difficult to document the coherence, command structure, or even indi-
vidual leadership of insurrectionary forces that have attacked state targets. That
is not to say that such command-and-control mechanisms did not exist, but they
were ephemeral and poorly documented, and they have often been so wholly
annihilated that it is impossible to exhume their specifics. This mass violence
makes it difficult to differentiate genocide from warfare (something genocide
perpetrators count on), and the levels of this mass violence make it difficult to
exclude from data sets of civil wars.
Fearon and Laitin (2003) score civil wars in Burundi in 1972, 1988, and
1993–2002; and in Rwanda in 1962–65 and since 1990. Scholars disagree about
the specific timing of these conflicts. Thus Uvin (1999) and Sambanis (2006)
cite mass atrocities in the 1960s in Burundi, next recurring in 1972, for instance.
Sambanis scores civil wars also in 1972 and 1988, but commences the third civil
war in 1991. As for Rwanda, data sets (Fearon and Laitin 2003; Doyle and Sam-
banis 2006; Political Instability Task Force 2008) concur on a civil war episode
in the 1960s and another in the 1990s, but they disagree on precise starting dates
(1962 or 1963) and ending dates (1964, 1965, or 1966). Because of the con-
tinuity across the 1994 genocide and throughout the 1990s, I score the 1990
152 Examining the Cases

internal war as continuing throughout the 1990s, even though the majority of
war-related deaths occurred outside the borders at times, and even though the
dynamics shifted in the 1990s. It is relatively clear that from war’s onset in 1990,
right through the July 1994 fall of Kigali to the Rwandan Patriotic Front, there
was no period of widely perceived peace followed by a renewed warfare. There-
fore, for our purposes, the 1990 period was the most recent, and only, episode of
civil war recurrence in Rwanda.
Before turning to these wars, what are the relevant demographic and other
characteristics of these societies? The two countries formed a single Belgian
colony from World War I until July 1962, when they split and each received in-
dependence. Belgium had fostered a Tutsi-dominated monarchy, although when
Hutus rose up in mass violence in 1959, it organized elections in which it sup-
ported Hutu rule in that portion of the colony. Tutsis have historically made up
only 10 to 14 percent of the population in both Burundi and Rwanda, with 85 to
90 percent Hutus and 1 percent Twa (Uvin 1999, 253). There is no systematic
religious difference between these ethnic groups, and religion has not been a
source of violent mobilizations.

Rwanda’s 1959 Violence, and War in the Early 1960s

The violence in Rwanda during 1959 illustrates what would become a pattern in
both countries’ organized violence: a single or localized attack that involves per-
petrators of a different ethnic group than the victim leads to expanded violence
along ethnic lines, and then reprisals that are far greater in scale than the initial
attacks. The mass reprisals, often led by the state, sometimes constitute “selective
genocide” (Lemarchand 2008, 2), which invokes terror and mass ethnic flight to
neighboring states, which in turn sows the seeds for future cross-border incur-
sions by armies comprising refugees. Mass atrocities have often overshadowed
organized instances of combat between rebels and state forces. Weinstein (1972,
23) makes this point and shows the importance of analyzing the interaction
between these two countries: “The knowledge among Tutsi in Burundi about
Rwanda has traumatized them and the repression after each Hutu revolt attempt
has been intensified in geometric proportions.”
In November 1959 a group of Tutsis assaulted a Hutu subchief, which in
turn led Hutu groups to attack Tutsi officials. Reprisals carried out by Tutsis led
to the deaths of several hundred, with some 160,000 Tutsis fleeing to adjacent
countries by 1966 (Des Forges 1999; US Department of State 2008). Among
these “fifty-niner” refugees were the future rebel commander Paul Kagame and
his family.
The war of the early 1960s commenced in the shadow of the 1959 violence. In
1961, refugees in Burundi began to launch cross-border attacks in Rwanda, lead-
Nonseparatist Recurrences of Civil War 153

ing Hutus to target Tutsis still within Rwanda as accomplices. Ethnic violence
would continue through independence in 1962 to 1966. Only the attack of De-
cember 1963 threatened the country’s stability, but roughly 20,000 people died
in these attacks in the early 1960s (Des Forges 1999). Alison Des Forges (1999,
36) reports that “Hutu leaders used them all [the attacks] to bolster the sense of
Hutu solidarity, to solidify their own control and to eradicate the last vestiges of
respect for Tutsi authority.”
Rwanda’s 1959 ethnic violence had transformative consequences for politics
in Burundi, because it deepened Tutsi determination to hold onto power more
tightly (Weinstein 1972). Thus, although a multiparty constitutional monar-
chy was inaugurated upon Burundi’s independence in 1962, fearful Tutsis as-
sassinated the prodemocracy King Rwagasore (who had married a Hutu) and
in 1965 carried out a coup that ushered in one of the most ethnically exclusive
regimes on earth (Ndikumana 1998). Uvin (1999, 256–57) is worth quoting
at length:

Following the events in Rwanda, state control [in Burundi] became the sole
vehicle for Tutsi to retain their privileges, while conversely it was the sole means
of rapid social advancement for Hutu. . . . From 1966 to 1993 political and by
extension economic power in Burundi was tightly held by three military regimes
(Micombero, 1966–82; Bagaza, 1982–87; Buyoya 1987–93) that used their
military might to keep their privileges. All three presidents were Tutsi-Hima
[lower-caste Tutsi] from the same village in the Bururi region, born within
two miles of each other. . . . Almost all positions of importance in Burundi
were monopolized by the Tutsi minority. They included the higher levels of the
single party, . . . the full command structure of the army, the police and security
forces, and the judicial system (even in 1994, only 13 out of 241 magistrates
were Hutu).

Analysts virtually unanimously concur on the role of exclusion as a key driver

of violence of the region. Lemarchand (2004, 62) states that “political, economic
and social exclusion are seen as the principal dimensions that need to be ex-
plored if we are to grasp the dynamics of domestic and interstate violence in
the Great Lakes.” The extreme and brutal treatment of ethnic majorities (along
with some minorities) do not map perfectly onto the concept of exclusion that
has prevailed in the case studies up to this point in the book. In cases such as
Liberia, East Timor, Tibet, and South Ossetia, a shift toward a more exclusion-
ary state policy played a role in sparking renewed armed conflict. However, in
Burundi and Rwanda, there was what I call “chronic” exclusion, extreme social
discrimination coupled with persistent state structures and policies that rested
on excluding the “other.”
154 Examining the Cases

War and Mass Atrocities in Burundi in 1972

The warfare and mass killings of 1972 in Burundi illustrate the altered pattern
of violence and the role of chronic, rather than precipitant, exclusion. The vio-
lence began on April 29, when bands of 10 to 30 Hutus and Mulelists carried
out simultaneous attacks in the capital and in southern provinces that cost some
2,000 lives (Lemarchand 1975). Lemarchand (2008, 2) reports that “the trig-
gering factor behind the bloodbath was a Hutu-led rural insurrection aimed at
seizing power from the ruling Tutsi minority.” Most of the radicalized intellectual
refugees leading the uprising came from Tanzania. The Tutsi-dominated gov-
ernment denounced the uprising as an attempt to destroy Tutsis altogether and
killed on the order of 150,000 to 300,000 persons (Lemarchand 2008, 2). The
longtime researcher of violence in the region René Lemarchand commented
(1975) that “the pattern of counterviolence initiated by the government and the
army was more systematic and hence more efficient in terms of human destruc-
tiveness. . . . The repression took on the qualities of a selective genocide directed
at all the educated or semi-educated strata of Hutu society.” The reprisal’s “aveng-
ing furor swept across the entire country and lasted for months after it had been
brought under control” (Lemarchand 2008, 2).
The consequences of the 1972 violence deepened a sense of ethnicity-based
insecurity and oppression, of which political exclusion was only one component.
As Uvin (1999, 258) explains, “[The 1972] rampage created sufficient fear to
suppress Hutu unrest for two decades. . . . For many years to come Hutu parents
would not send their children to school for fear of making them targets in future
pogroms. These events constitute the defining moments in independent Burun-
di’s history. They crystallized Hutu and Tutsi identities and created a climate of
permanent mutual fear.”
In 1972, Hutus within the civil service and higher education had been es-
pecially targeted, and afterward Tutsi-led governments de facto excluded Hutu
from these positions. Token representation existed for Hutus, and by 1987, “they
held four government ministries out of 20; seven of the 65 seats in the National
Assembly and two places of 65 on the Central Committee; one provincial gov-
ernor out of 15” (Gale Group 2008, 4). This exclusion was not a policy decision
that precipitated the 1988 war but a prevailing condition during the ten years or
more before the war.

Burundi’s 1988 Recurrence

In the context of this persistent exclusion, Major Pierre Buyoya took power
through a coup in 1987, signaling a commitment to greater inclusion of Hutus.
He appointed more Hutus to government posts and freed political prisoners.
Nonseparatist Recurrences of Civil War 155

Even so, the Tutsi dominance persisted. By the time of the outbreak of violence
in 1988, three-quarters of the Cabinet and the legislature, thirteen of fifteen gov-
ernors, and all army officers and 96 percent of the enlisted soldiers were Tutsis
(Gale Group 2008, 4). Rather than this book’s argued state exclusion, state inclu-
sionary measures presaged the 1988 recurrence. Some allege that Buyoya’s mea-
sures toward inclusion generated heightened expectations among Hutus, whose
frustration fueled conflict (Gale Group 2008, 4). The Burundi case suggests that,
in the context of extreme chronic exclusion, steps to greater exclusion may not
provoke armed conflict. However, it would be a mistake to say that Buyoya’s pre-
violence measures constituted powersharing, because they rested on informal,
inclusionary measures. Indeed, steps toward powersharing only emerged after
the violence. Buyoya then appointed the country’s first Hutu prime minister and
created a commission, composed equally of Hutus and Tutsis, to investigate the
violence and make recommendations for national unity (Gale Group 2008, 4).
The 1988 violence began when Hutu farmers, resentful of corrupt Tutsi local
rule, killed up to 3,000 Tutsi in two northern villages of Burundi (Uvin 1999).
The Tutsi-dominated army responded directly and with heavy force, killing be-
tween 20,000 and 150,000 Hutu and driving out tens of thousands of refugees
(Uvin 1999; US Department of State 2008). Given the mass atrocities in the
government response, violence eventually diminished. Buyoya’s measures of in-
clusion did not amount to powersharing, but they signaled a willingness to grant
Hutus a greater participation in governance that would still be clearly dominated
by Tutsis.

Civil War Recurrence in Rwanda in 1990

As in Burundi in 1988, the civil war that rocked Rwanda in 1990 with the inva-
sion by the Tutsi-led, refugee-based Rwandan Patriotic Front (RPF) emerged
not from any direct exclusionary steps but from a lessening of the oppressive and
chronic exclusion, and from other international factors. The wave of democrati-
zation that quickened around the world in the late 1980s touched much of Africa,
including the Great Lakes region. France put pressure on Rwandan president
Juvenal Habyarimana to hold multiparty elections, leading to the legalization of
new parties. Several of these, under Hutu leaders who decried the peace process,
put pressure on Habyarimana and mobilized critics from among Hutus that he
had to take into account. The peace process, in conjunction with democratiza-
tion, seems to have also instilled fear and insecurity among hard-line Hutus, lead-
ing to preparations for mass killings.
Rwanda’s 1990 civil war recurrence emblemizes the myriad ways that regional
and international factors shape the Great Lakes conflicts. We have already seen
the signaling and “learning” by Tutsi and Hutu constituencies across borders
156 Examining the Cases

after waves of violence, and the importance of refugee communities as the source
of mobilization of political organization and armies. “Refugees provide the con-
ceptual link among all three forms [political, economic, and social] of exclusion,”
according to Lemarchand (2004, 67). Their role derives not solely as direct pro-
tagonists of violence but also as sources of economic and social stress on the
host society, which provokes scapegoating by host nations’ authorities and fuels
cycles of discrimination if not demonization.
Such was the case of Tutsi refugees who had first fled by the tens of thou-
sands after the 1959 violence, and then had become central actors in the civil
war of 1990. Given the persistent Hutu political domination in Rwanda through-
out the 1960s and 1970s, the Tutsi refugees in Uganda became key recruits and
even commanders of the rebel National Resistance Army (NRA) organized by
Museveni to overthrow Milton Obote. The Obote government expelled thou-
sands of Rwandan refugees, who were subsequently permitted to return in 1985
to NRA-controlled areas. When they took Kampala in 1986, one-fourth of the
16,000 rebels were Banyarwandans. Among these was Paul Kagame, the future
leader of the Rwandan Patriotic Front, who was named acting head of intelli-
gence in Uganda (Mamdani 2001).
Mamdani (2001) and Prunier (1995) argue that the RPF invasion cannot be
understood outside the context of political events in Uganda. In the words of one
RPF commander, “If the NRM [National Resistance Movement, the political
arm of the NRA] could liberate Uganda, the RPF began to ask why it could not
do the same in Rwanda” (quoted by Mamdani 2001, 173). In addition, Ugandan
political events also pushed the RPF to launch an armed attempt to challenge
the Hutu-led government in Kigali. According to the new Ugandan ambassador
to the United States, two decisions by the new Kampala government pushed the
Uganda-based leadership of the RPF to decide to invade in 1990: the decision to
ban nonindigenous Banyarwandans from owning land in Uganda, and the deci-
sion to ban them from holding state offices there (Mamdani 2001).
The RPF invaded from Ugandan bases on October 1, 1990, with ongoing
combat halted only in July 1992 when a cease-fire was signed amid international
pressure. That cease-fire failed within eight months, leading to renewed fight-
ing and a new cease-fire forced on the government of Juvenal Habyarimana in
August 1993. The downing of the airplane carrying the presidents of Rwanda
and Burundi prompted the genocide in April 1994, and the taking of Kigali by
the RPF three months later. Rwanda’s recurrence is distinctive insofar as it de-
parts from the pattern identified by numerous analysts of the Great Lakes re-
gion of a limited uprising or assassination leading to the murder of one ethnic
group by the hundreds or thousands, followed by massive reprisals costing hun-
dreds of thousands of lives. However, Burundi in 1993 represents a return to
this pattern.
Nonseparatist Recurrences of Civil War 157

Burundi’s 1993 Recurrence

In Burundi, Buyoya also embraced reforms leading to multiparty elections.
Uvin (1999, 261) suggests that the end of the Cold War and the recognition
that single-party minority rule was not sustainable contributed to this decision.
A Hutu prime minister was appointed; a “government of national unity” was
formed, with equal numbers of Hutus and Tutsis in the Cabinet; and a “Charter
of Unity” was adopted that emphasized national rather than ethnic loyalties. The
state even obliged educated Burundians to explain the notion of ethnic unity to
poor farmers (Uvin 1999, 261). A new Constitution was adopted in 1992, which
led to the election of a Hutu president for the first time in June 1993.
The hope was that this process could overcome the pattern of the prior three
years, and prior decades, described here by Uvin (1999, 259): “All these cases
[of violence in 1972, 1988, 1991, and 1992] presented the same pattern: in re-
sponse to rumors and fear, Hutu peasants attacked and killed local Tutsi, power
holders, and even ordinary people. The army was then sent in to restore order
and indiscriminately killed vastly more people in retaliation.” Weinstein (1972,
23) remarks similarly, “the repression after each Hutu revolt attempt has been
intensified in geometric proportions.” As Lemarchand (2004, 63) indicates,
Hutu rebellions had ended in “dismal failure, each time resulting in extremely
brutal repression,” until the 1993 election of a Hutu to the presidency, inspiring
hope that they may finally be lifted from this chronic discrimination and exclu-
sion. Unfortunately, their hopes were crushed when the “all-Tutsi army killed the
newly elected president” and the leaders of the Assembly, sparking renewed war-
fare and violence. Within three months an estimated 50,000 to 100,000 persons
were killed, and millions fled as refugees (Uvin 1999, 262). The new war would
last into the twenty-first century; there was a pause, with a cease-fire among im-
portant rebel groups in 2002, but the war ended only in 2006 when the last of
the rebel groups agreed to a negotiated settlement.
Buyoya’s experiment in the early 1990s marks the failure of informal inclu-
sion and democratization as a means of preventing war recurrence. Note that
this inclusionary mechanism was one of “power dividing” rather than power-
sharing, because the 1992 Constitution banned ethnically based parties and re-
quired parties to gather signatures in every province as an ethnic inclusionary
measure (Uvin 1999, 261). In line with the theories of Horowitz (1985) and
Roeder (2005), the Constitution eschewed powersharing in favor of efforts to
draw ethnic groups into a competitive electoral system that rewarded intergroup
cooperation. This case suggests that inclusionary behavior does not always work.
It also suggests, however, that powersharing may be necessary as an interim mea-
sure, backed with external guarantees where required, to provide security reas-
surance to major groups.
158 Examining the Cases

Moreover, these cases should not lead one to eschew exclusionary behavior
as a significant factor. On the contrary, the scholarly consensus is that exclu-
sion has played the most important role in the patterns of organized violence in
Rwanda and Burundi during the past four decades. Rather, the cases suggest that
where chronic, extreme exclusion persists, inclusionary efforts may in fact trig-
ger armed uprisings through raised expectations. The experience of Burundi’s
peace process since 2002, though not taken up here, suggests the potential that
inclusionary mechanisms, including powersharing, can have in such a context
with the presence of international guarantors and programs aimed at facilitating
dialogue and mutual understanding.

Alternative Explanations and Conclusions

What economic factors and other explanations can be discerned in these cases?
Although Rwanda exhibits many of the risk factors for civil war recurrence—
poverty, competition for scarce resources, weak state institutions, and a high
portion of mountainous territory (though not excessive reliance on exports of
natural resources)—specialists in the region’s violence subordinate these factors
to political exclusion. In Rwanda and Burundi, the dearth of valuable exports or
domestic production has made access to state power even more important. As
Rwantabagu (2001, 49) notes, “Because the country is poor in terms of limited
investment opportunities in the private, industrial, and services sectors, the main
avenue for economic prosperity is through political activity, the civil service, or
joining the armed forces. . . . After 1972, a notable rift was created between the
Hutu and Tutsi communities in terms of access to wealth and privilege. . . . At
another level, overpopulation and an acute land shortage in the countryside has
heightened intercommunal tensions and precipitated violent clashes. Thus, the
Burundian conflict has clear economic undertones as two groups compete for
the control of limited resources.”
Although Lemarchand (2004, 66–67) refers to political, social, and eco-
nomic exclusion, political exclusion plays the most significant role in the Great
Lakes violence:18 “If there is little doubt that the 1959 Hutu revolution in Rwanda
received its impetus from the political exclusion of Western-educated Hutu
elites, it is equally clear that economic exclusion had relatively little to do with
the Hutu-Tutsi conflict.” In addition, Prunier (1995) acknowledges another eco-
nomic factor. The drop in commodity prices that hit Rwanda so hard made it dif-
ficult for the government to accept the refugee return that was under negotiation
with the RPF in the early 1990s.
In conclusion, recurrent violence in Rwanda and Burundi holds four impor-
tant implications for this study. First, although included in this study of civil
wars, their distinctive pattern of violence (shared by other regions and countries,
especially in Sub-Saharan Africa) bolsters recent calls (Cederman and Gleditsch
Nonseparatist Recurrences of Civil War 159

2009) to disaggregate the study of civil wars. The pattern of mass-based vio-
lence, organized by rebel armies operating across borders in waves of violent
reprisals led by the state, offers one pattern of what are classified as “civil wars”
but that merit separate explanations.
Second, the cases of Burundi, Rwanda, and the entire Great Lakes region
(certainly including Rwandan and Ugandan involvement in postgenocide war-
fare in Zaire) exemplify the importance of a regional level of analysis and the
salience of transnational factors. Third, both cases support claims that democ-
ratization can fuel mass ethnic violence. As Uvin (1999, 267) says, “The most
recent and most extreme rounds of violence are the direct result of processes of
democratization set in motion in large part by pressure from the international
community.” In the absence of coherent prodemocratic political actors, democ-
ratization itself may prompt insecurity and mass violence. And fourth, the role of
chronic, rather than precipitating, exclusion stands out, despite instances where
inclusionary measures seemed to provoke mass violence. These cases point to
the need for further conceptualization of types of exclusion and of the processes
whereby it enervates war and violence.
The cases of nonseparatist recurrence of armed conflict reveal important di-
vergences in the character of recurrence, in the aims of insurgents, and in causal
factors. The levels of violence diverged greatly, ranging from hundreds of thou-
sands dead in Africa’s Great Lakes region to only dozens of combat deaths in
Haiti and East Timor’s recurrences. The rebels’ specific motives also differed in
these cases.
Many hypothesized factors played a role in the recurrence of armed conflict
in these cases. These include state weakness in all the cases examined in this
chapter. Indeed, countries like the Central African Republic, Haiti, and East
Timor have come to embody weak state institutions to such an extent that it is
difficult to distinguish those fragile institutions from the very violence that ac-
companies the frequent political changes of government, often through coups
that may constitute internal armed conflicts. In addition, regional factors play
an important role in recurrences in several countries. Although these countries
are all extremely poor, this poverty was not directly related to the instigation of
war recurrences without regard to the more unusual factors emphasized here—
namely, exclusion.
Exclusion by new postwar governments characterized all these cases of non-
separatist recurrences. These cases reinforce the important role identified in
separatist cases and in Liberia of exclusionary behavior. In every case actors un-
derstood their own renewed violence mainly as a reaction to the exclusionary
positions adopted vis-à-vis state power by postwar elected governments.
In addition the cases of East Timor and Zimbabwe show that the exclusion of
former enemies is not the only dimension along which warfare may recur. The
exclusion of former allies or coalition partners after a successful rebellion also
160 Examining the Cases

can prompt a recurrence of armed conflict. In both Zimbabwe and the recent
case of East Timor, the perceived marginalization from state power represents
the main armed violence experienced subsequent to independence. Thus exclu-
sion may precipitate renewed armed conflict when postwar governments move
to marginalize, exclude, or even persecute their former rebel partners.
Finally, the cases of Rwanda and Burundi broaden our understanding of ex-
clusion to reinforce an argument that has been prevalent for many years: that
chronic oppression involving exclusion from political and economic power can
elicit armed violence on a massive scale. This chronic exclusion constitutes a
standing factor that must be combined with other explanations to fully account
for the recurrence of wars in these cases. In both cases, political competition
in democratization efforts represents an important trigger for ethnicity-based
violence, as did the approaching implementation of the 1993 Arusha agreement
in Rwanda.
However, these five nonseparatist recurrences are not the sole cases of recur-
rence. Chapter 6 examines the causes of war’s recurrence in the four cases that
seem to contradict the book’s central argument.

Notes
1. United Nations Office of Coordination of Humanitarian Affairs, “Central African
Republic: IRIN Background Notes,” June 11, 1998, www.africa.upenn.edu / Hornet /
irin_61198.html.
2. On the “predatory” state, see North (1981); and on its existence in Haiti, see
Maguire (1995).
3. This despite the presence of darker-skinned heads of state, such as Francois Du-
valier and Jean-Bertrand Aristide.
4. National Coalition for Haitian Refugees / Human Rights Watch / Jesuit Refugee
Service, “Fugitives from Injustice: The Crisis of Internal Displacement in Haiti,” Au-
gust 1994 report, Washington, D.C. Also see Perusse (1995, 35).
5. UN Statistics Division, “National Accounts Estimates of Main Aggregates,”
http: // data.un .org / Data .aspx?q=haiti&d=SNAAMA&f=grID percent3A101
percent3BcurrID percent3AUSD percent3BpcFlag percent3A1 percent3BcrID
percent3A332#SNAAMA.
6. International Monetary Fund, “World Economic Outlook Database, October
2010: Nominal GDP List of Countries,” data for 2009.
7. Personal communication to the author from Shepard Forman, October 14, 2004.
8. The successive UN presences are as follows: UNAMET, June–October 1999,
provided electoral assistance for the referendum; UNTAET, October 1999–May
2002, administered the territory until independence; UNMISET, May 2002–May
2005, supported strengthening of the state and security with lower numbers of troops;
UNOTIL, May 2005–August 2006, a political mission with only thirty-four unarmed
Nonseparatist Recurrences of Civil War 161

military observers supporting state building; and UNMIT, August 2006 until this
writing.
9. For the details of the events of April and May 2006, I draw heavily on “UN Spe-
cial Commission 2006” report, officially the “Report of the UN Independent Special
Commission of Inquiry for Timor-Leste.”
10. UN Department of Peacekeeping, “UNMISET Home Page: Facts and Figures,”
www.un.org / en / peacekeeping / missions / past / unmiset / facts.html.
11. Australian Defence Force website on the International Stabilization Force in
East Timor, www.defence.gov.au / op / eastTimor / general.htm.
12. Republica Democratica de Timor-Leste and UN Country Team, “Timor-Leste
Millennium Development Goals Report,” United Nations, Dili, 2004, 15.
13. On the history of the independence struggle, see Stedman (1993), Sibanda
(2005), and Bhebe and Ranger (1995).
14. See Stedman (1993), Sibanda (2005), and Bhebe and Ranger (1995).
15. On Nkomo and ZAPU, see Sibanda (2005).
16. For more details, see Lawyers Committee for Human Rights (1986).
17. Ibid.
18. Lemarchand (2004, 66) defines “political exclusion” in a liberal manner, as “the
denial of political rights to specific ethnic or ethnoregional communities, most notably
the right to vote, organize political parties, freely contest elections.”
6
Recurrences That Defy
the Argument

NO SINGLE ARGUMENT or factor can plausibly lay claim to playing the crucial
causal role in recurrences of internal armed conflict. Despite my making a strong
case for decisions by a state that render it exclusionary and insufficiently legiti-
mate, exclusion is neither sufficient nor necessary for war recurrence. Of the fif-
teen core cases of recurrence, four cases do not support the role of exclusionary
behavior in the recurrence of armed conflict. This chapter examines those four
cases. It first explores Lebanon and Mali, two cases that directly defy the hypoth-
esized role of exclusion. In both cases inclusionary behavior was adopted and
maintained, yet dissident groups among former armed insurrectionary groups
took up arms once again. The chapter identifies which factors these cases suggest
are causally important, and why inclusionary behavior failed.
The chapter then turns to Nicaragua and Peru, two postwar states that en-
gaged in exclusionary behavior, but where other factors seem to trump the
impact of government exclusion. These two cases suggest, respectively, the im-
portance of external actors and of the rents from natural resources (viz., coca
products). Despite the preponderance of evidence that shows the significance of
exclusionary behavior in accounting for postwar recurrences of armed conflict,
these cases represent exceptions. They show that other factors, including those
hypothesized in the literature drawn from quantitative works, can play a salient
role in war recurrence.
Two cases show the limits of an explanation that relies solely on exclusionary
behavior. In the case of Lebanon, a 1943 powersharing arrangement failed in
1958, and it was then restored after that initial civil war. Analysts concur that the
failure of that powersharing arrangement constituted the precipitating factor in
the second civil war of 1975–90. Other factors are explored below in the mini–
case study. Of the entire pool of forty-two countries that experienced either war
recurrence or nonrecurrence, Mali offers the only post-1988 case where power-
sharing failed to prevent recurrence. In both cases dissident groups reverted to
armed conflict, despite the adherence to powersharing by some elements of the
Recurrences That Defy the Argument 163

prior insurrectionary armed forces. Whereas religious difference constituted the

lines of mobilization in the Lebanese case, in Mali tribal / ethnic rather than reli-
gious differences posed the lines for the mobilization of an armed revolt.

Lebanon in 1958 and 1975–90: Failed Powersharing

Lebanon’s civil war of 1975–90 has been widely studied, especially as an example
of the shortcomings of powersharing. Largely in response to the weakened hold
that occupied France held on its colonies, Lebanese elites in 1943 negotiated
a governance pact that smoothed the way to independence from the mandate
system in 1946. The 1943 pact also served as the foundation for powersharing
arrangement that succeeded until an armed conflict in 1958, and then resumed
until the second civil war of 1975.

The 1943 National Pact and the 1958 Civil War

The National Pact represents one of the most oft-cited instances of powersharing,
which makes it worthwhile to sketch its main features. Christians (including
Maronites, Greek Catholics, and Greek Orthodox), Muslims (Sunni and Shia),
and Druze are the most prevalent of the seventeen religions found in Lebanon.
Ethnic identity—including self-identification by some Christians and others as
Phoenician or Canaanites, as against the majority of Lebanese, who self-identify
as Arab—is not nearly as salient a feature of identity as religion. Under the 1943
pact, Maronites held the presidency, Sunni held the premiership, and cabinet
posts were allocated according to the relative prevalence of the religions (e.g.,
including a Druze defense minister and a Greek Orthodox deputy premier) (Za-
har 2005, 228). The pact also allocated legislative seats according to a Christian-
to-Muslim ratio of 6 / 5. Although these seats were allocated by district and not
religious community, they became vehicles for religion-based voter mobilization
and patron–client networks that served elites and offered little responsiveness to
the population (Zahar 2005, 229).
The parties to the 1958 armed conflict raised the same issues that underlay
the religious tensions of the 1940s. Most Lebanese Muslims opposed the appar-
ent Lebanese Christian policy preference for increased proximity to the West
and instead preferred closer ties to Syria and the Arab world. The pact signified
a compromise between these two paths (Zahar 2005, 228). After the United
States took Britain’s side in the Suez Crisis, Lebanese president Camille Nimr
Chamoun, a Christian, refused to break ties with the West and also eschewed
the United Arab Republic, which had been proclaimed by Nasser in early 1958.
In the meantime, Muslims by the mid-1950s were discontent with the power-
sharing arrangement. Economic growth bypassed the agriculture sector, prompt-
164 Examining the Cases

ing migration to the cities, which benefited Christian and Sunni communities
but not Shia Muslims in the South (Makdisi and Sadaka 2005; 60; Zahar 2005,
229–30). This “created a never-before experienced socioeconomic gap that
closely mapped onto religious affiliation” (Zahar 2005, 229). In light of this dis-
satisfaction with Muslims’ share of state power, they questioned the degree of
power President Chamoun enjoyed to make decisions. Specifically, the Lebanese
National Movement (LNM) objected to Chamoun’s joining the Baghdad Pact
and thereby setting back Arab unification. With urging by radio broadcasts from
Syria and Egypt to rise up and overthrow the government, the LNM launched a
series of attacks. The army chief refused to send the army against the rebels for
fear of mutiny, and American marines waded ashore among sunbathers to back
the government against a feared loss of the country to Syrian control and the
United Arab Republic. Roughly two to four thousand persons died in the fight-
ing (Collelo 1989, 186). The president resigned, but the powersharing arrange-
ment was revived amid new leadership.
In this first crisis of powersharing, external influences were apparent. More
than 140,000 Palestinians, 10 percent of them Christians, became permanent
residents of Lebanon in conjunction with the creation of a Jewish state on Leba-
non’s southern border (Picard 1996, 79). Their presence not only constituted
a haven for armed resistance to Israel, and thus a hostile target of the Israeli
Defense Forces, but also invited Arab and other meddling. Syria persistently
attempted to exercise power in Lebanon, emboldened by the Arab nationalist
movement that was embodied in the United Arab Republic. The West, especially
the United States, also exercised its influence at this time. In contrast to most
cases of civil war recurrence examined here, these divisions fell along religious
rather than ethnic lines. External factors enhanced a sense of marginalization
along religious lines to ignite the 1958 armed conflict. Marginalization was ap-
parent, but not through a deliberate disruption, reversal, or reneging of the in-
clusionary approach taken in the 1943 National Pact.

Civil War Recurrence, 1975

The customary marker of the beginning of the civil war is the exchange of attacks
on April 13, 1975, between Palestinian militias and the predominantly Maronite
Kataeb (Lebanese Phalange) of Pierre Gemayel (Picard 1996, 102). Tensions
and attacks between the two communities had intensified in the previous years,
especially after the 1967 war and the related strategic decline of the United Arab
Republic.
The second civil war’s factions divided along conservative / progressive lines
and also religious lines. On one side lay the Christian-dominated forces defend-
Recurrences That Defy the Argument 165

ing the status quo, including the Kataeb that would become part of the Lebanese
Front (LF) and its allies, the Israeli-influenced Southern Lebanon Front, and the
“Tigers” military wing of Camille Chamoun’s National Liberal Party. The state
army, which initially stayed on the sidelines, would by 1976 break into fragments.
Thus, the LF became the main protagonist defending the status quo.
On the other side lay a less cohesive group led mainly by the Lebanese Na-
tional Movement (LNM), under the Druze Kamal Jumblatt, and its allies in a
rough coalition of Muslims, Palestinians refugees, and leftists (including roughly
10,000 Palestinians from the Rejectionist Front drawn partly from the Popular
Front for the Liberation of Palestine, plus the pro-Iraq Arab Liberation Front, the
Independent Nasserite Movement, the Syrian National Socialist Party, and the
Communist Party). Although some Christian leftists belonged to the progressive
coalition led by Jumblatt, very few Muslims were able to stay in the conservative
forces led by the LF.
After the attacks of April 13, 1975, led to a series of mutual recriminatory at-
tacks against civilians, Jamblatt refused to participate in a government to which
the Kataeb belonged. No Sunni political leaders would cross him, but neither
would any Christian-dominated groups concede to form a government without
the Kataeb. Hence a stalemate led to the formation of a Christian-favored mili-
tary cabinet by Maronite president Suleiman Frangieh, leading to a summer of
relative calm, aided by Palestine Liberation Organization (PLO) chairman Yas-
ser Arafat’s success in keeping PLO factions out of the fighting through the end
of the year (Cobban 1987, 128). However, continued wrangling and maneuver-
ing led to renewed and massive fighting among militias. By the fall of 1975, most
of Beirut’s port and the city itself were destroyed, and within two years roughly
600,000 to 900,000 Lebanese had fled the country. The war would eventually
result in the death of some 130,000 persons and cost $100 billion in damage
(Collelo 1989, 181).
The government became largely dysfunctional in early 1976, at which point
many Palestinians joined in the warfare on the side of the LNM. Ostensibly to
quell the violence, Syria launched a full-scale invasion in May 1976. At that point,
the LNM and its Palestinian and other allies occupied 70 percent of the coun-
try. The fighting continued, however, with Israel and Syria tacitly cooperating to
prevent the Palestinian groups and their LNM allies from gaining control of the
country (Winslow 1996, 208–10). With Syrian advances, the Lebanese Forces
made gains, taking back territory from the LNM and convening the Parliament
to install a new government in September. Throughout 1976, Syria and other
Arab states had engaged in dialogue, and after its invasion, Syria sought to legiti-
mize its presence through the Arab League, which deployed troops (Winslow
1996, 210–12). A peace conference in Riyadh in October 1976 led to a significant
166 Examining the Cases

reduction in armed actions, leaving Syria dominating the politico-military scene.

Although the war continued through new evolutions, influenced heavily by out-
side actors, those latter phases are less relevant to the present inquiry.

Analysis of the Lebanon Case

The Lebanese case runs counter to two generalizations found in this study. First,
it represents a failure of powersharing. That failure is not simply a correlation,
because the powersharing mechanisms fueled conflict recurrence and impeded
conflict resolution in the early weeks of both the 1958 and 1975 wars. Specifi-
cally, the faith-based allocation in the legislature permitted mutual vetoes that
led to a deadlock once Jamblatt refused to work with Kataeb early in the fight-
ing. Moreover, religiously based cabinet postings led certain ministers to favor
specific militias in 1975–76. Once specific army commanders refused to obey
orders out of a concern that the religious divisions would accelerate, the state’s
armed forces broke apart. More broadly, Muslim dissatisfaction with the fixed
allocation of power based on outmoded demographics certainly undercut the le-
gitimacy of the state, which came to be seen as overly representative of wealthier
Christian interests, especially once economic divisions deepened. Thus, Zahar
(2005, 231) notes that “dissatisfaction with the powersharing formula that privi-
leged the Christians was a major cause.”
Second, the case calls into question the contribution of international troops
to the maintenance of stability. Foreign troops departed Lebanon in 1946, and
they were absent throughout the 1958 conflict and through the beginning of the
1975 war. The 1976 Syrian invasion did ultimately lead to a decline in violence
in 1977–80. However, that incursion did not produce stability. It did prevent
the emergence of one feared outcome, a radical Palestinian-controlled state in
Lebanon, along with the feared alternative, the creation of a pro-Israeli Chris-
tian enclave state. Yet it enmeshed Syria in a more serious war and occupation
than expected. The Arab Deterrent Force (which replaced a brief Arab Security
Force after three months), largely a fig-leaf for the Syrians, did not bring endur-
ing stability and departed in 1982. Perhaps because foreign troops, even when
under the rubric of “peacekeeping” or “deterrent,” have generally deployed in-
dependent of a broad internationally agreed-on framework, their deployments
have not yielded the sort of stability found in other cases. A UN Interim Force
in Lebanon, deployed in 1978, helped bring some stability to southern Lebanon
but played only a secondary role in fostering negotiations (Day 1986). Zahar
(2005, 235–40) concludes that the presence of “multilateral protectorates,” such
as after the 1989 Ta’if Accord, has provided the backing for powersharing to suc-
ceed. She finds that bilateral interventions (e.g., the invasions of Syria and Israel)
fueled political crisis and conflict.
Recurrences That Defy the Argument 167

Conversely, Lebanon’s war recurrence confirms other arguments put forth in

this book. It ratifies the role of religious fractionalization but not that of ethnic
fractionalization. In their explicit testing of the model of Collier and Hoeffler
(2002, 2004) for the Lebanese case, Makdisi and Sadaka (2005, 59, 69) make
this their central conclusion. In sorting out the complex religious panorama
of Lebanese society, they follow Collier and Hoeffler in applying a Christian /
Muslim distinction that maps onto a 45 / 55 ratio, which rendered Christians a
market-dominant minority (Chua 2002). In addition, they find the economic
motivations “weak” in the recurrence of civil war in 1975, and they “immediately
dismiss” the role of natural resources. However, they amass considerable data
showing that the various Lebanese factions received considerable rents—per-
haps $15 billion locally and $30 billion from foreign supporters—that consti-
tuted an important incentive to persist in warfare and thus delayed a negotiated
settlement (Makdisi and Sadaka 2005, 73–81).
The case of Lebanon also reinforces the need to examine regional dynamics
and external actors. Civil war and its recurrence in Lebanon were heavily influ-
enced by external actors, especially the PLO, Israel, Syria, Iran, the United States,
and other Arab states. The migration of Palestinians changed the politics, secu-
rity, and demographics of the society in ways that fueled internal armed conflict.
Syria and Israel at times brought periods of stability but also fueled warfare by
supporting and even constituting some of the multiple armed factions to the
conflict. One of the consequences of the extensive and at times sporadic involve-
ment of external actors was an almost constant series of peace initiatives, mainly
by Syria, but also by other bilateral and multilateral actors. The Arab League ne-
gotiated a number of cease-fires during the recurrence. Syria contributed to the
elaboration of the Constitutional Document’s revisions to powersharing in 1976,
to further modifications to powersharing in the Fourteen Points of Reconcilia-
tion in 1980 and again in 1984 in talks in Switzerland, and to the Ta’if Accord in
1989 that eventually ended the war (Zahar 2005, 237).
The case of Lebanon does not represent an example of exclusionary behavior
prompting civil war recurrence. The 1975 civil war reflects the shortcomings of
powersharing, but this does not mean that powersharing is a failed mechanism
for sustaining peace. Powersharing sustained stability for multiple decades, and
some form of powersharing has been seen as necessary for reaching agreement
on a stable political order by both internal and external actors, even up to the
present. In fact the case shows both the necessity and the perils of powersharing.

Mali in 1990–96 and 2007: Failed Powersharing

Mali is unusual as a case of a civil war that ended with a powersharing agree-
ment that was largely respected, but where armed conflict nevertheless recurred.
168 Examining the Cases

The conflict reflects exclusionary behavior at the outset, including chronic ex-
clusion from before independence through the first civil war of the early 1990s.
Exclusion and inclusion were based not on religion but on ethnic, regional, and
other distinctions involving the pastoral, nomadic Tuaregs who inhabit mainly
the northern, desertified portion of Mali and the western area of Niger. Recur-
rent armed conflict in the period 2006–9 reflects not exclusion so much as the
initiative of a minority of former rebels pressing for more rights, power, and state
benefits. Regional factors were important both because a Tuareg rebellion in Ni-
ger preceded that of Mali in the early 1990s, and interaction continued between
Tuaregs in both countries. In addition, Libyan and Algerian governments both
deployed troops and became active, along with other African states, in seeking
to mediate a resolution. Control of smuggling routes played a role in the 2006
recurrence but was not the primary factor, and natural resources did not figure
prominently.
Despite its reputation as a peaceful West African country, Mali has experi-
enced internal armed conflicts, mostly concentrated in what is generally called
northern Mali, comprising the three (out of eight) regions of Tombouctou, Gao,
and Kidal. These three remote, desert regions constitute 59 percent of Mali, a
portion larger than Afghanistan or Texas. Uprisings against French rule were put
down, as was an uprising by Tuaregs and Moors after independence in 1960.
Tuaregs and Moors make up 5 to 10 percent of the population of Mali, and differ
from other ethnic groups in the country as they generally have lighter skin and
are considered to be related to North African Berbers.1 They numbered 621,000
of Mali’s 12.5 million inhabitants in the 1990s (Keita 1998, 6).
Tuaregs (the term sometimes includes Moors) have episodically sought
greater autonomy from empires and states during the past several decades. The
former commander of a key base in the counterinsurgency of the 1990s described
“a conviction among the Tuaregs” after independence in 1960 that “they were
singled out for particular discrimination, and were more neglected than others in
the distribution of state benefits” (Keita 1998, 10). From 1962 to 1964, Tuaregs
carried out a number of hit-and-run raids that did not constitute a civil war and
that were suppressed by the state. Keita (1998, 10) describes how Tuaregs at the
time felt threatened by land reforms aimed at modernization by southern Mali-
ans unsympathetic to the northern tribes’ migratory pastoral lifestyle. As a result
of the uprising, the government placed Tuaregs under a “repressive” military ap-
paratus (Keita 1998, 10).

The Initial Civil War

The main civil war that Mali has experienced was waged by a few rebel groups
dominated by Tuaregs from 1990 to 1996. In 1990 the spread of democratization
Recurrences That Defy the Argument 169

in Eastern Europe and Africa also led to pressures in Mali for greater political
and press freedoms, but then–president Moussa Traoré rejected calls for multi-
party electoral system. In 1991 Traoré was ousted and arrested. A transitional
government headed by Lieutenant Colonel Amadou Toumani Touré organized
a constitutional reform process that led to the approval of a new charter in 1992
that provided for multiparty elections and the freedom of political parties except
where based on ethnicity, religion, region, or gender (US Department of State
2010c). Alpha Oumar Konaré of the Alliance for Democracy became the coun-
try’s first democratically elected president in 1992, which ended eighteen years
of one-party rule.
In June 1990 the Popular Movement for the Liberation of the Azaouad
(Mouvement Populaire de Libération de l’Azaouad, MPLA, “Azaouad” being the
Tuareg term for Northern Mali) carried out attacks on two towns in Kidal that
marked the beginning of the war, almost simultaneously with the onset of armed
conflict by Tuaregs in Niger (Dillon 2007, 28; Vinding 2002, 355). These rebels
sought greater autonomy and resources from the state. In contrast to the smaller
rebellion of the early 1960s, the attacks spread across the northern regions and
intensified in 1990–92. One factor here was the harsh repression that the Ma-
lian army visited upon not just the communities in areas of conflict but also on
lighter-skinned peoples throughout the northern areas, which prompted more to
flee and an intensification of the warfare (Poulton and ag Youssouf 1998). Other
rebel groups emerged, some of which splintered. Under Algerian pressure four
of these groups united as the Unified Movements and Fronts of the Azaouad
(Mouvements et Fronts Unifiés de l’Azaouad, MFUA).
Virtually every analysis of the Tuareg Rebellion of the early 1990s refers to
social marginalization, exclusion, and cultural and political suppression. Krings
(1995, 57) argues, for instance, that among the decisive factors for the “margin-
alisation and deprivation of the Tuareg” that led to war were the “negation of the
existence of the Tuareg ethnic community” by black African leaders in Mali and
Niger, along with “exclusion of the Tuareg from political participation.” Dillon
(2007, 2) cites as a “root cause” of the war the “social exclusion of Touaregs from
military and civil service positions.”
Some have argued that severe drought in the North was a major factor in the
Tuareg rebellion. Severe droughts struck the region in the years 1973–74 and
1983–84 (Poulton and ag Youssef 1998, 33; Somerville 1986). These droughts
reduced grasslands and animal herds, made livelihoods more difficult for no-
mads, and forced mass migration (Dillon 2007, 47–49). One detailed study of
the effects of drought concluded that this was not the main cause of the war
(Benjaminsen 2008). In addition, these droughts cannot explain the timing of
the onset of the 1989–96 war. The uncertainties unleashed by democratization
better account for this timing.
170 Examining the Cases

International factors played a role in the initial conflict. The onset of conflict
came in June 1990, one month after rebel attacks began in adjacent Niger af-
ter 18,000 Tuareg refugees returned from Algeria. The Nigerian attacks, which
were related but not coordinated, forced the MPLA’s leader, Iyad ag Ghali, to
launch attacks earlier than he had hoped (Krings 1995). Libya also played a role.
Although Libya’s Qaddafi in 1980 declared his country as the “homeland” for all
Tuaregs and offered them citizenship, one researcher who interviewed Tuareg
leaders found that they dismissed these claims, seeing Libya instead as having
“abandoned” Tuaregs during the first civil war (Dillon 2007, 55). Vinding (2002,
355–56) emphasizes the role of unemployed youth who, affected by the drought
and poverty in the 1980s, joined Qaddafi’s Islamic Legion and received military
training, and then returned to Mali or Niger to fight.

Peace Agreements and Powersharing

Peace negotiations commenced early in the war. An initial accord fell apart with
the ouster of President Traoré in 1991. In April 1992, coincident with Konaré’s
election as president, Lieutenant Colonel Touré successfully reached a settle-
ment with the MFUA. The National Pact, or “Bamako Accord,” provided for
territorial, economic, and military forms of powersharing. The accord created
new insurgent-majority units for a period of one year; stipulated that all minority
populations would have a place in the local police, security forces, and civil ser-
vice; and dismantled certain army posts in northern Mali. It also created a new
commissioner for the North of Mali to report directly to the president, directed
reconstruction resources to the region, and established local and interregional
assemblies that would have authority over the local police forces and administra-
tive decisions.
Some of the agreed-on measures were implemented but others were not. In
April 1993 610 ex-combatants were integrated into the army. Local assemblies
assumed greater authority, as did the new commissioner for the North; however,
the redistribution of power took time and created new uncertainties. There was
little investment, and permitting instruction in the Tuareg language (Tamashek)
took time to implement. Although the levels of attacks declined, some attacks
continued, especially in 1994.
Divisions within the rebels also undermined peace implementation. Arab
groups within the rebel movements had formed their own insurgent group, con-
cerned that negotiations would enhance Tuareg autonomy but not address Arab
concerns. Other splits occurred over caste lines, between older traditional clan
elders and younger men who sought chances for advancement, and between
those demanding independence versus autonomy. Around 1994 Songhoi com-
munities began forming militias that carried out attacks on Tuaregs, sometimes
Recurrences That Defy the Argument 171

with the blessing of the Malian army, which fueled society-based warfare. In ad-
dition, some distrustful Tuaregs in the army mutinied and killed several of their
colleagues (Keita 1998, 18).
Violence continued through 1996, and it then dissipated to such an extent
that the war was considered over. Ag Ghani’s group (split off from the MPLA and
renamed first the FPLA and then the MPA) did not abandon the National Pact
and helped defeat two other MFUA groups, the ARLA and the FIAA (Hum-
phreys and ag Mohamed 2003, 21). The money of another group, the FPLA,
dried up. Foreign states pressed for implementation of the National Pact dur-
ing this time, and the United Nations Development Program funded a program
working with local chiefs to reinforce peace. In 1996, a “Flame of Peace” cer-
emony in Timbuktu marked an official end to the war, because roughly 3,000
rebel weapons were destroyed. In October 1996 1,620 additional ex-combatants
were integrated into the army, the various police and customs forces, and the civil
service, including 77 at the officer level or equivalent (Dillon 2007, 100, based
on Keita 1998).2

Recurrence of Armed Conflict

In 2006, disaffected ex-combatant Tuaregs deserted the army and attacked and
sacked a few villages in the North. A former ex-combatant, Ibrahim ag Bahanga,
led what was called the Alliance Démocratique du 23 mai pour le Changement
(ADC). Lieutenant Colonel Amadou Touré, the former interim leader, had
won as an independent candidate for president and established a broad-based
government in 2002. His government issued statements emphasizing its con-
tinuing commitment to an inclusive state, and negotiated a new peace agree-
ment in 2006. Many of the ADC rebels demobilized and joined (or rejoined)
the Malian army.
However, in 2007, ag Bahanga led a further splinter group called the Alliance
Touareg Nord Mali pour le Changement (ATNMC), which broke off and initi-
ated attacks that killed dozens and displaced hundreds, constituting a renewed
internal armed conflict. Attacks intensified in mid-2007 and were concentrated
in the extreme northeastern region (Kidal), adjacent to Niger, where related but
uncoordinated renewed attacks also took place.
Tuaregs remained split, and some fighting continued in 2008–9, but not
enough for the Peace Research Institute Oslo (PRIO) to score the armed con-
flict as continuing. An agreement in August 2008 was brokered by the Alge-
rians, which led to a further incorporation of ex-combatant fighters back into
self-contained units of the army deployed in Mali. However ag Bahanga still de-
manded greater autonomy (though not independence) for the Kidal region and
the withdrawal of Malian army troops, which the government resisted. His forces
172 Examining the Cases

continued attacks from the Tigharghar Mountains, and he took refuge in Algeria,
from where he persisted in his struggle through 2009, even though the Malian
army made important gains and won massive defections of the rebels.
These splinter groups, mainly ex-combatant army deserters, claimed that the
poverty and neglect of the Kidal region underlay their struggle, so it would be
erroneous to see the recurrence as reflecting a reneging on the National Pact,
certainly not its core elements. Instead, Mali’s recurrent armed conflict, albeit a
small-scale recurrence, defies the argument of this book. What factors does this
case suggest are important for the failure of peace? Mali points to the role that
splintered groups and incohesion among insurgent forces pose for the consoli-
dation of peace.3 The increasing involvement of these groups in smuggling also
suggests an economic opportunity that fueled the recurrence. Smuggling had in-
tensified in northern Mali over the previous years, ironically helped by the reduc-
tion of the state’s coercive forces there as part of the National Pact (Dillon 2007).
The integration of ex-combatants was seen as both significant and successful,
despite the problems that developed among some ex-combatants. Dillon (2007,
99–102), in 2006 interviews with Tuareg leaders and ex-combatants about the
stability of peace, found that most Tuaregs brought up integration, citing it as ei-
ther well implemented, helpful to peace, or beneficial to Tuaregs, or all the above:
“Several [interviewees] saw the endeavor as the key factor that helped end the
fighting and ‘establish peace’ in northern Mali. . . . By granting Touaregs access
to these jobs, the integrations went a long way towards addressing the social ex-
clusion that had contributed to the start of the rebellion” (Dillon 2007, 99, 101).
In addition, economic aid to the North and greater power for the newly de-
centralized structures were functioning within a few years. That aid was less than
hoped for and expected, partly due to economic downturn.
Mali’s armed conflict recurrence resembles that of Peru in that many former
combatants abandoned their struggle. However, in Mali, a series of negotiations
played the crucial role, whereas in Peru intelligence feats that led to the capture
or killing of crucial commanders played a key role. Moreover, the Malian state
adopted inclusionary measures that were largely implemented and were not re-
versed by the Malian state. This case offers the most serious challenge to the
argument that exclusionary behavior underlies war recurrence, and that the in-
corporation of combatants into the state army and police forces is a crucial ele-
ment of consolidating peace.

Analyzing the Cases of Failed Powersharing

The cases of Lebanon and Mali clearly challenge the notion that exclusion is nec-
essary for a recurrence of internal armed conflict after a period of stability. Inclu-
Recurrences That Defy the Argument 173

sionary measures were deemed inadequate by some dissident groups of former

insurgent forces, who took up arms once again against the state. In both cases
the state had incorporated significant groups of former rebels as part of power-
sharing, and these largely remained opposed to taking up arms again. Doyle and
Sambanis (2000, 2006) point to the importance of rebel cohesion as a factor in
successful peacebuilding. These cases suggest that where a lack of cohesion ex-
ists, then powersharing can prove insufficient and war can recur.
However, certain factors may be seen as qualifying the failure of powershar-
ing here. First, the migration of Palestinian refugees into Lebanon certainly
paved the way for social tensions and stress upon the carefully negotiated 1943
powersharing National Pact. Second, these refugees became a central element
of regional and cross-border factors that played an important role in stimulat-
ing dissent from powersharing in both countries. In Mali the rebellion in Niger,
which was supported by Libya and others, also helped spark insurrection. And
in Mali expectations of inclusion may have been met in some areas, but resources
promised were not forthcoming, which provided support for the complaints of
dissidents who eventually took up arms. Weak institutions in both societies and
a divided and dysfunctional state in Lebanon also played roles. Finally, it bears
noting that, despite the problems of powersharing and conflict recurrence, in
both societies powersharing and inclusionary measures are still viewed as indis-
pensable to a sustained peace.

Nicaragua in 1972–79 and 1981–90: Externally Driven Recurrence

Apart from the cases of Lebanon and Mali that contradict the argument of this
book, two other cases also represent nonconforming recurrences. The differ-
ence here is that both cases exhibited exclusionary conduct toward enemies. The
policies of the Sandinista government after its revolutionary victory in Nicaragua
in 1979, and those of the Peruvian government after the effective defeat of the
Shining Path insurgency in 2000, were more exclusionary than inclusionary. De-
spite democratizing moves later on, the early days of the Sandinista government
saw the departure of key allies, including political leaders of moderate political
parties associated with the middle class, along with former rebel allies who be-
came disillusioned with their exclusion (much along the lines of former rebels
in Zimbabwe and East Timor). In Peru the state continued to treat the Shining
Path as a terrorist organization, and no hint of powersharing or inclusion issued
forth from the governments of presidents Fujimori, Paniagua, Toledo, or Garcia
from 2000 to 2007.4
Given that Peru and Nicaragua engaged in exclusionary policies, why are these
cases considered not to conform to the book’s argument? The answer is that
174 Examining the Cases

exclusion seems to be ancillary to war recurrence. The secondary literature on

these two Latin American cases does not identify this exclusionary behavior as a
precipitating factor in the recurrence of armed conflict. As such, I do not empha-
size exclusionary behavior in these four cases, but explore other causal factors.

Nicaragua’s Successful Revolution

Nicaragua experienced the first successful Communist revolution in Central
America in 1979. The Marxist-Leninist Sandinista National Liberation Front,
the “Sandinistas,” toppled the right-wing United States–backed dictator, Anas-
tasio Somoza, whose family had ruled since 1932. The Sandinista party had been
founded in 1961 to fight the dictatorship, but its military operations and threat
became serious in the mid-1970s. Underlying this revolutionary struggle was re-
sentment at a personalistic dictatorship that accumulated wealth for itself, sup-
pressed any form of democratic competition, and repressed its opponents. The
insurgency achieved victory only once the middle classes joined in protests. Nei-
ther ethnicity nor religious differences were sources of mobilization, although
indigenous peoples and Creoles of African descent found mainly on the Atlantic
coast make up roughly 5 percent and 9 percent of the population, respectively.
Once Somoza fled the country in July 1979, roughly 5,000 to 7,000 National
Guardsmen fled, mainly to neighboring Honduras (Prevost 1987, 7). A combi-
nation of exiled former Somocistas (as the former supporters of Somoza were
called) and disillusioned Sandinista peasants and elites, organized and funded
by the United States, launched a counterrevolution (leading them to be called
“Contras” in Spanish) by late 1981.

Nicaragua’s Civil War Recurrence

As with most internal wars, the exact date of the start of the renewed civil war in
Nicaragua is not wholly clear, but it is generally considered to be a year, perhaps
two. Sporadic, uncoordinated attacks, especially at the hands of disgruntled peas-
ants over land reform, occurred in the first half of 1980 (Padro-Maurer 1990, 2).5
During this time, former National Guardsmen and wealthy exiles began to meet
and organize, but nothing beyond isolated, quick attacks by small groups on San-
dinista forces or civilian targets. In this period peasants who spoke against the
conversion of private farms to cooperatives were labeled “counterevolutionaries”
and were harassed or arrested by the new authorities (Brown 2001, 14–16). And
in April 1980 the first resignations of Sandinista cabinet members occurred, fol-
lowed by a steady trickle that by 1981 included future president Violeta Cha-
morro (the widow of a popular opposition political figure) and one of the most
charismatic and well-known commanders, Eden Pastora.
Recurrences That Defy the Argument 175

Most analysts cite the attacks of 1981 in their accounts as significant open-
ings to the counterrevolutionary war (Booth 1985; Padro-Maurer 1990; Dil-
lon 1991). Sam Dillon (1991, 54), for instance, describes an initial attack led by
Commander Tigrillo on June 13, 1981, in which six people died in an attack on
a Sandinista jeep with civilians aboard. He describes one rebel’s sentiment that
they had finally attacked and captured weapons “after starving in the hills for a
year.” Prevost (1987, 8) reports that bands of former guardsmen and business-
men formed two groups, the Fifteenth of September Legion and the National
Democratic Union, along the Honduran border, although these resulted only in
“a few isolated attacks” on literacy workers in 1980.

Invasive Neighbors: The Crucial Role of the United States

To what extent was the origin of the recurrence instigated by Nicaraguan actors
versus external actors, namely, the Americans with the support of the Argentina
military? Although bands of resistance emerged in 1980, it was only with the
United States’ organization, funding, encouragement, and pressure that a rebel
army came together and began launching military operations that could consti-
tute an internal armed conflict. In a perhaps self-serving statement, Contra com-
mander Enrique Bermudez stated in the fall of 1981 that he knew of organized
resistance in Honduras, but that without international backing it would not gain
traction (Dillon 1991, 59).
The initiative of the Americans in organizing the rebellion is omnipresent,
even though the Contras are not entirely an American invention. In late 1979
Bermudez himself was initially recruited to join the rebels by US general Chuck
Boyd, who was in charge of the Western Hemisphere for the US Air Force (Dil-
lon 1991, 62). Within a few months, he quit his job as a truck delivery driver for
Newsweek to go on the payroll of the US Central Intelligence Agency (CIA). He
reached out in 1981 to other ex-guardsmen in the United States, Honduras, and
Guatemala. In March 1981 newly inaugurated President Ronald Reagan autho-
rized covert activities in Nicaragua, and $20 million was allocated to the task
(Dillon 1991, 64). In 1981 the Argentines also began funding anti- Sandinista
groups, and the United States paid for Argentine advisers to the nascent rebels.
In August 1981 CIA officers arranged for the merger of the Fifteenth of Septem-
ber Legion (approximately sixty men) and two other groups. Bermudez assisted
in this process, and he also moved their headquarters to Tegucigalpa (Dillon
1991, 65). Although former Somoza guardsmen wanted the National Guard in
the name of the new rebel army, the Americans insisted instead on something
more appealing to the public, and the Nicaraguan Democratic Force was born:
“CIA agents drafted documents for the signing ceremony” (Dillon 1991, 65).
Even the former Sandinista hero, Eden Pastora, was contacted by the CIA and
176 Examining the Cases

eventually commanded a unit based in Costa Rica with funding from the CIA
and the US Army (Dillon 1991, 86).
Political exclusion played an important role. The Sandinistas made no effort
to include their former enemies in the new government. People from the prior
government who posed a potential threat (including National Guard officers)
were pursued and assassinated rather than integrated into the government or
armed forces (Dillon 1991). Moreover, the Sandinistas quickly showed intol-
erance for those who were former commanders of impeccable revolutionary
credentials but who sought a less radical approach to both land reform and po-
litical life. Once land reform moved beyond taking the large haciendas of the
wealthiest to taking private medium-sized farms, these Sandinistas objected,
which led to their marginalization, to their treatment as potential subversives,
and ultimately to their joining the opposition and even the armed resistance
(Brown 2001, 15). The departure of many Sandinistas echoes, on a less influen-
tial scale, the split of insurgent coalitions in the proindependence movements in
East Timor and Zimbabwe.
Nevertheless, it would be misleading to say that exclusionary policies consti-
tuted the primary explanation for recurrence. Exclusionary behavior drove many
Sandinista supporters to support the insurgency, including objections to their
political and economic policies. Ideological concerns and fear for their lives led
many National Guardsmen to flee and join the rebels. The foreign policy de-
cision by the Reagan administration to contest the Sandinistas on ideological
grounds was a key factor in elevating these disparate elements into an organized
fighting force capable of any serious military operations. The causes of war recur-
rence in Nicaragua lie with an “invasive neighbor” as much as Sandinista exclu-
sionary policies or the objections to them.

What of Ethnicity and Economics?

Because many Miskito and other indigenous groups joined the Contras, some
analysts pondered whether the counterrevolution could be considered an ethnic
war (Stoll 2005). Certainly the recurrence mobilized Native Americans more
than the initial revolution. Thousands of Indians fled from their communities
once the Sandinista reforms threatened their traditional forms of authority, and
many indigenous resisted recruitment into the Sandinista army. Yet this war re-
currence reflected principally ideology and class rather than ethnicity. The mo-
bilization was rooted in anticommunist claims, and very few of the command-
ers or troops were indigenous. Only 3 percent of the Contras who demobilized
reported their origin as the Atlantic coast region, where indigenous and Creoles
predominate (Horton 1998, vii).
Recurrences That Defy the Argument 177

Natural resources played no role whatsoever in the origins of the civil war or
its recurrence, except insofar as the banana and coffee export-oriented economy
is seen by some scholars as associated with the one-man dictatorship that So-
moza represented (Robinson 1996; Paige 1997).
In summary, once in power, the Sandinistas set about consolidating their grip
on control of the state. Even though they would later engage in limited inclu-
sionary measures, including competitive elections, their policies excluded both
former enemies and former commanders who criticized government policies or
behavior. Nevertheless, that exclusionary conduct was not the principal trigger
for the recurrence of 1981. The United States’ decision to provide funding and
other support for organizing the remnants of the National Guard was the prin-
cipal factor in recurrence.

Peru in 1980–2000 and 2007–Present:

Exclusion, Coca, and Rebel Resurgence
Peru’s recurrent armed conflict presents another case that does not primarily re-
flect exclusionary behavior, even though the former guerrilla army was not in any
way included in governance. Instead the case supports the agency of rebel groups
(more than the state), in contradiction to this book’s main argument about the
crucial role of agency on the part of the postwar national government. The case
is also one of the very few core recurrences where natural resources, illicit in this
case, play a vital role in the resurgence of internal armed conflict.

Peru’s Civil War

Peru was one of several Latin American countries to experience civil war in the
1970s and 1980s. Its war began in 1980 with the emergence of a revolutionary
movement called the Shining Path (Sendero Luminoso) under the leadership of
a Maoist-Marxist academic Abimael Guzman. Although the movement operated
mainly in the central and southern highlands, which are populated heavily by
indigenous tribes, peasants were mobilized principally on class issues (and coer-
cion).6 Guzman and most of the insurgent leadership were ladinos (i.e., of mixed
European and indigenous heritage). A second less deadly and less successful in-
surgency emerged at the same time, with its first action in 1982, from the more
conventional Marxist-Leninist Tupac Amaru Revolutionary Movement. These
wars killed an estimated 70,000 persons, mainly civilians, and both ended in the
late 1990s with a counterinsurgency campaign characterized by rights abuses
and intelligence breakthroughs that resulted in the capture or death of most of
the leaders.
178 Examining the Cases

Peace via Strategic Victory

The termination of Peru’s first civil war ended not via negotiations but via a pe-
tering out of military operations, which were considered a strategic victory by
virtually all observers. By the mid-1990s, casualties and combat had diminished.
One source found that the conflicts between the government and the two rebel
groups cost 117 military and civilian lives in 1998, 60 in 1999, none in 2000, and
only 5 in 2001.7 Another source, PRIO, lists between 50 and 100 battle deaths
for 1999, and then in 2003 reported the conflict as terminated due to too few ca-
sualties from 2000 onward, despite one or two attacks in 2002 and 2003 (PRIO
2003a, 2003b, 2003c). By 2003 about 1,200 former Shining Path members re-
mained in Peruvian jails (Taylor 2006). Peace was widely perceived to have been
consolidated, despite an awareness that conditions in the indigenous-dominant
areas home to the Shining Path had not changed significantly.

Recurrence of Armed Conflict

However, a number of Shining Path militants resisted capture. In 2007 renewed
Shining Path armed attacks rose to a level that placed the country back on the
list of active armed conflicts. Security experts estimated that the Shining Path
numbers had reached 300 by the end of 2006. Peru engaged in no inclusionary
behavior toward the former guerrillas, and it continued to prosecute and try its
combatants from the 1990s to the present. In this sense, exclusion rather than
inclusion characterizes the government’s posture toward the Shining Path. Con-
ditions were not significantly altered for the population most associated with the
insurgency—a region of small cities but mainly rural towns and remote peasant
communities primarily of strong indigenous influence. However, the recurrence
of armed conflict did not derive primarily from exclusionary policies by the state
vis-à-vis the armed opposition or those it purported to represent. Nor did recur-
rence derive from external exertion of force or from religious differences.
Instead, the gradual resurgence of the Shining Path beginning in 2007 derived
from the insurgent initiative, as reconstituted and surviving units and leaders of
the movement launched sporadic attacks. Those attacks targeted mainly state
military and police forces, which represented twenty-two of twenty-six battle
deaths in 2008 (Romero 2009). This reactivation came under the same rubric
of the original goals of the guerrilla organization. A December 2008 report stated
that the head of the rebel group, Comrade Artemio, “is the sole top-level [Shin-
ing Path] leader who has not yet been killed or captured” (Hyland 2008, 2).
Although repressive actions and human rights violations have characterized the
Peruvian army’s response to specific incidents and communities where guerrilla
presence was apparent, state repression was not a major factor in this resurgence.
Recurrences That Defy the Argument 179

High-rent exports of natural resources constitute another important cause

of the recurrence. Following the example of the reinvigorated Revolutionary
Armed Forces of Colombia (known as the FARC) of the 1990s in adjacent Co-
lombia, the Shining Path made the processing of coca into cocaine a focus of
its operations. It engaged in extortion of those involved in the trade and itself
undertook trafficking and processing (Romero 2009). According to one news
account, the Shining Path, “while still professing to be a Maoist insurgency at
heart, is now in the business of protecting drug smugglers, extorting taxes from
farmers and operating its own cocaine laboratories” (Romero 2009). According
to a 2002 report from the US State Department, membership in the Shining Path
had already begun to grow, due partly to the group’s greater involvement in the
cocaine trade (US Department of State 2002). One Peruvian security analyst
estimated that by 2009 the Shining Path had rebounded to some 500 workers
in the cocaine trade and an additional 350 armed combatants (Romero 2009).
Even incarcerated Shining Path founder Guzman referred in 2010 to the leaders
of the movement as “mercenaries.”8
Mountainous geography facilitated this war and its recurrence in the high-
lands, as did ethnic (but not religious) differences. Most of the recruitment of
the Shining Path occurred in highland areas populated mainly by indigenous
peoples. However, the main leadership are ladinos, not indigenous, and the ideo-
logical basis for mobilization eschews ethnicity for ideology.
Like the recurrences in Mali, the Central African Republic, Azerbaijan, and
South Ossetia, Peru in 2007 and afterward experienced very few battle-related
deaths—just enough to count as internal armed conflicts. Like Mali, Peru rep-
resents a case where insurgent agency played a more important role than any
specific action of exclusion undertaken by the government. In these two cases no
violation of expectations on the part of the insurgents explains the taking up of
arms again. As such, the case does not confirm this book’s argument; nor does it
disconfirm it, given the absence of any inclusionary behavior. This case supports
theories that emphasize insurgent agency and the role of illicit resources in civil
war recurrence.
Peru’s recurrence offers a strong case for considering the violence as part of
a continuous struggle—this would be the view of the insurgents. Even though
most Peruvians saw the Shining Path as defeated in the early 2000s, the insur-
gents did not share this understanding.

Conclusion
What do these four cases tell us about the causes of recurrence of internal armed
conflicts? They first underline the lack of a single factor or approach that can fully
account for war recurrence. The presence of exclusionary behavior by postwar
180 Examining the Cases

states is neither necessary nor sufficient to elicit renewed warfare. It is worth not-
ing the extenuating circumstances that help account for the failure, in particular,
of powersharing in Lebanon and Mali. The massive influx of Palestinians into
southern Lebanon transformed the demographic foundation upon which power-
sharing was based. It also aggravated political tensions, both in the expressed
preferences of this refugee population and by drawing in regional interests and
actors to a greater degree whose policies shifted accordingly. The postwar Ma-
lian government did not deliver on many of its promises of inclusion of Tuaregs,
paving the way for complaints of persistent exclusion. This experience should
serve as a cautionary lesson for cases (see Nepal and Bangladesh in chapter 7)
where authorities are not moving to fulfill pledges made as part of powersharing
arrangements. These cases nevertheless show the vulnerability of powersharing
arrangements to changing circumstances and to perceptions of exclusion.
These two cases of failed powersharing also suggest a new hypothesis that
draws on the theoretical claim that rebel cohesion is important for peacebuilding
success. They specifically point to the importance of rebel cohesion as a factor in
the success of powersharing arrangements. One can also see the element of rebel
cohesion in the Timorese and Zimbabwean recurrences, where victorious armies
experienced divisions and then exclusion, which led to relapse into conflict. Atlas
and Licklider (2005) point to the presence of a similar process of postindepen-
dence violence in Chad and Sudan.
The cases of Nicaragua and Peru illustrate show that exclusion can exist but
also be eclipsed by other factors in recurrence of armed conflict. The Peruvian
case especially lends support to theories that emphasize natural resources as a
central factor in civil wars and their recurrence. Coca products clearly helped
provide the fuel for a resurgence of the Shining Path, much as they have given
oxygen to the growth and persistence of the FARC. Nicaragua’s recurrence calls
into question the level of analysis common to civil war studies: the state. The role
of external states or other actors in not just abetting but also constituting civil wars
and their recurrence requires reconceptualizing civil wars, armed conflicts, and
their analysis. In particular, regional actors and dynamics, as well as extraregional
actors and dynamics, merit systematic and integrated analysis when seeking to
understand today’s armed conflicts.
The three prior cases have utilized a particular method that bears reflection
at this point. The method of “structured, focused” case comparison is not new.
Numerous books and articles have utilized the method, which was initially de-
veloped by the insightful and personable Alex George (Bennett 2008). What
is unusual is the degree to which these have been condensed, considering only
those factors identified in the qualitative and quantitative literature reviewed
initially. These “mini–case studies” constitute one of the contributions of this
Recurrences That Defy the Argument 181

book. In contrast to the wholly acontextual research of quantitative methods,

they offer the reader a sweeping glance of the entire pool of cases of recurrence of
internal armed conflicts. This method offers a foundation from which the main
hypothesized causes of a civil war and its recurrence can be explored and, at a
minimum, found implausible, as occurred for so many variables in so many cases.
Short of fuller historical accounts of each conflictive country, these case studies
offer the most comprehensive understanding of each case.
At the same time the method has its shortcomings. Relevant context is neces-
sarily abridged, and important factors undoubtedly do not receive the attention
they deserve, partly because available theories may omit relevant variables. This
blindness to relevant variables is partly compensated by a careful and faithful
adherence to the primary and secondary sources about the conflicts by those
who are more specialized. Nevertheless, a generalist researcher does not pos-
sess enough knowledge of the context and history of each case to adequately
sift through competing accounts or to discern where the secondary literature
has omitted important factors. More likely is an oversensitivity to hypothesized
variables in a necessarily brief account of these cases. Although I have sought to
avoid a skewed interpretation of the origins of each armed conflict and its recur-
rence (or not), the mini–case studies may fall into this trap.
Having analyzed these structured, focused “mini–case studies,” the entire pool
of qualifying recurrences of internal armed conflicts from 1944 to 2007, I now
turn to examine the pool of cases of nonrecurrence. These cases, twenty-seven
in number, are too numerous to make mini–case studies a practical option as a
research method. Consequently, the next chapter draws on a structured, focused
examination of each case but presents it in a more synthesized format, though I
still seek to be consistent, rigorous, and fair in assessing the relative importance
of various factors.

Notes
1. Some 52 percent of Mali’s national population is of the Mande ethnicity (Bam-
bara or Malinke), 11 percent are Fulani, 7 percent are Songhai, 7 percent are Saracolé,
5 percent are Moor and Tuareg, 4 percent are Mianka, and 14 percent are other ethnici-
ties (US Department of State 2010c). The Songhai, Moors, and Tuaregs predominate
in the North, where Tuaregs numbered about 621,000 in the 1990s (Keita 1998, 6).
2. This brought total integration of ex-combatants to 2,230, including 1,311 in the
army, 348 in the National Guard, 301 in the gendarmerie and police, 100 in the customs
service, 50 in the water / forest service, and 120 in the civil service (Dillon 2007, 100).
3. Doyle (2002) and Doyle and Sambanis (2006) emphasize this variable as one of
several that influence successful peacebuilding.
182 Examining the Cases

4. After debate among Peruvian officials, members of Shining Path were permitted
to run for office for the first time in the October 2010 elections (Briceño 2010). See
also Longmire (2009).
5. Brown (2001, 1–19) offers an account of the origins of anti-Sandinista actions
that begins with disgruntled peasants, many former Sandinistas, that fled to hills in the
face of land confiscations. His interviews with former rebels date the earliest attack as
November 1979, when a group of “Milpistas” (People’s Anti-Sandinista Militias, for-
merly a component of the Sandinista army called the People’s Anti-Somoza Militias)
attacked a Sandinista base with only five deaths among the attackers (Brown 2001, 14).
6. On the context of the war and the Shining Path, see McClintock (1998), Fumer-
ton (2003), Fox and Starn (1997), Gorriti (1999), Campbell (1973), Taylor (2006).
In the highlands, poverty correlated heavily with indigenous identity; see Arnson
(2005).
7. See Ploughshares’ “Armed Conflict Reports: Peru,” www.ploughshares.ca /
libraries / ACRText / ACR-Peru.html, which relies in part on data from Peruvian non-
governmental organizations, US State Department estimates, and data from Human
Rights Watch, and was updated in February 2002,.
8. “Minister: Shining Path Remains a Threat in Peru’s Jungles,” Latin American Her-
ald Tribune, August 6, 2009.
7
Making Peace Stick
Inclusionary Politics and Twenty-Seven Nonrecurrent
Civil Wars

HAVING EXAMINED CASES of recurrence, what do cases of successful peace-

building tell us? Is inclusionary behavior in some way associated with persis-
tent peace? In turning to the cases of nonrecurrence, we will see that they offer
more robust findings, ones that inspire some hope. Recall that peace is defined
here narrowly as “negative peace” and that all cases of nonrecurrence refer to
full-scale civil wars (not lesser armed conflicts), including those with the active
involvement of other states or intergovernmental forces in combat roles. These
twenty-seven cases of nonrecurrence (listed in table I.1 in the introduction) ex-
perienced no recurrence at all up until 2007. The first inference I draw from this
pool of cases is that historical context is extremely important. Interests and ideas
do not exist in a historical vacuum. The extent to which people see the resort to
arms as viable or attractive partially reflects the global and regional ideational
context. National elites’ choices about whether to negotiate a peace and about
what governance models are possible reflect those ideas that seem to be viable
and prevalent in a given time and place.
The Cold War specifically provided legitimacy for ideas about governance
and warfare that were backed by the interests of the two superpowers. The viabil-
ity of Communism and single-party socialism was pronounced after 1917, and
especially after 1945. Just as World War II transformed notions of sovereignty
and colonialism, undercutting fascist totalitarianism as a viable regime type, so
also the onset of the Cold War legitimated regime types ranging from Commu-
nist totalitarian to authoritarian regimes to democratic or electoral regimes. As
mentioned, diplomatic pressure, foreign aid, military action, and development
programs all advanced specific models of governance among new and less power-
ful countries. Diffusion, as argued by new institutionalists (Powell and DiMaggio
1991; Meyer et al. 1997), occurred via both strategic and interest-based decisions
by local elites and global sociocultural ideas and understandings about what
is normal and even trendy. Although such diffusion rarely took place without
significant modification based on national or local realities, interests, ideas, and
184 Examining the Cases

institutional trajectories in those less powerful or peripheral states, the impact of

expectations and discourse about regimes was pronounced.
The twenty-seven cases of nonrecurrence illustrate that the practice of nego-
tiating an end to civil wars was not prevalent during the Cold War but became
the norm after 1988. This empirical development has been noted elsewhere
(Fortna 2004a; Wallensteen and Sollenberg 1997; Hartzell and Hoddie 2007).
As table 7.1 notes, of the eleven cases of civil wars that concluded by 1988, nine
ended in strategic victory for one side or another. Only two cases—Laos in
1973 and Cyprus in 1974—ended in negotiated settlements during the Cold
War. Conversely, of the sixteen cases of nonrecurrence that ended after 1988,
all but one ended in negotiated settlements, as coded by various sources. An-
gola in 2002, one month after Jonas Savimbi was killed by government forces,
represents the only case of a strategic victory after the Cold War (there was also
a negotiated accord that formally concluded this war, but the war represented
more of a strategic victory).
The changing interests of the United States and the USSR in the 1980s,
in accord with realist thought, certainly shaped the trend toward negotiated
settlements. The shift by both superpowers toward support for negotiated settle-
ments and away from fueling armed conflicts contributed significantly to settle-
ments in Cambodia, Nicaragua, El Salvador, and Guatemala. Mediated by the
shifting strategic regional role of South Africa, the decline of the Cold War also
opened the way for ending wars in Namibia, Mozambique, and South Africa
itself.
The end of the Cold War also provided a historical window for the emergence
of new internal armed conflicts, several of which spawned secondary armed con-
flicts from secessionist movements within seceding states. Thus the world wit-
nessed new wars with the disintegration of the Soviet Union and the former Yu-
goslavia in Georgia, Chechnya, Moldova, Azerbaijan, Bosnia, and Croatia, which
also eventually spread to Kosovo, Dagestan, and Ingushetia. In some of these
cases, Russian pressure subsequently helped secure a cease-fire. In the region of
West Africa, the end of the Cold War also heralded new civil wars in Mali, Sen-
egal, and Liberia in 1989 and in Sierra Leone in 1991.
These cases illustrate not only the regional dynamics and dimensions of the
onset and termination of civil wars but also the role of powerful states in both
fostering such wars and concluding them. At the same time, interests alone can-
not account for clusters of cases of internal armed conflicts either spatially (i.e.,
in geographic regions) or temporally (i.e., concentrated within a few years). It is
very difficult to maintain that this dramatic shift toward negotiated settlements
stems entirely from interests of the great powers. Ideas and notional expecta-
tions vary over time about why and how armed conflict generally is or is not
warranted, whether negotiation is generally seen as an acceptable or worthwhile
Making Peace Stick 185

Table 7.1 Mode of Civil War Termination by Historical Period, All Nonrecurrent Civil Wars,
1944–2007
Period Victory Negotiated Settlement
Ended before 1988 Bolivia, 1952 Cyprus, 1974
Costa Rica, 1948 Laos, 1960–73
Cuba, 1958–59
Greece, 1945–49
Jordan, 1970
South Korea, 1949–50
Paraguay, 1947
South Vietnam, 1960–75
N=8 N=2
Ended in or after 1988 Angola, 1975–2002 Bangladesh, 1976–97
Bosnia, 1992–95
Croatia, 1992–95
Djibouti, 1993–94
El Salvador, 1979–92
Guatemala, 1968–96
Moldova, 1992
Morocco, 1975–88
Mozambique, 1976–95
Nepal, 1997–2006
Papua New Guinea, 1988–98
Sierra Leone, 1991–2000
South Africa, 1983–94
Tajikistan, 1992–97
United Kingdom / Northern
Ireland, 1969–99
N=1 N = 15
Sources: Table I.1, drawn from Fearon and Laitin (2003), augmented by data from the Peace Research
Institute Oslo, as described in the introduction to the present volume. Determinations of the mode of war
termination (victory vs. negotiated) are the author’s.

practice, and whether international organizations are seen as having potential

utility in mediation. In the wake of the Cold War, growing norms toward peaceful
settlement of civil wars and the enhanced role of international institutions like
the United Nations in settling such disputes contributed to and reflected a nor-
mative shift. Peace processes became more accepted and expected, and the role
of international mediators also became more widely appreciated. On the side of
interests, the incentives for warring parties to participate in peace talks increased
in the 1990s, as did the costs of refusing to participate.
In sum, a confluence of interests and ideas help account for a striking shift
in the mode of civil war termination, from outright victory before 1989 to ne-
gotiated settlements beginning in that year. This finding is important because
186 Examining the Cases

it demonstrates the possibility that normative changes could be an important

factor in civil wars and peacebuilding. Negotiated settlements have also been
shown to correlate with higher rates of reversion to internal armed conflict. Of
course, they usually signify at least a temporary reduction in mass violence.

Inclusion, Powersharing, and Peacebuilding Success

I now turn to another empirical conclusion suggested by an examination of all
cases of nonrecurrent civil wars: the importance of inclusionary behavior. As
discussed above, inclusionary behavior is defined broadly to include not just
powersharing arrangements of political, military, and other forms but also po-
litical systems that meet the expectations of former rebels to have the chance to
obtain political power either through elections or through fixed access to state
posts. Thus both power-dividing systems and powersharing systems are compat-
ible with inclusionary behavior, so long as the expectations of the parties for such
political power are not violated.
In this chapter I examine all cases of powersharing among the pool of cases
of both recurrence and nonrecurrence. This is not a pool of all cases of power-
sharing, even for the pool of country cases listed above. Among cases of recur-
rence, it refers only to instances of powersharing associated with the termination
of a prior war, and not to either earlier or subsequent powersharing arrange-
ments. Thus the powersharing “pact” that emerged in the 2000s in Nicaragua
is omitted because the relevant period for that country is between the end of its
initial war in 1979 and the commencing of its recurrence in 1981. Similarly, this
set of cases excludes analyzing powersharing arrangements that occurred well
after the scored war termination (e.g., Djibouti 2001) or that were linked to the
end of the most recent recurrent war termination, which thus means omitting
such important recent powersharing experiences as Burundi (2004–5), Rwanda
(2003), and Liberia (after 2003). In addition, this list excludes the many power-
sharing arrangements that have existed in the countries that have experienced
multiple, simultaneous, or intractable wars (e.g., Afghanistan, Iraq, Colombia,
Somalia, the Democratic Republic of the Congo, and Yemen).
Table 7.2 contains the list and types of powersharing for the twenty-seven
cases of recurrence and nonrecurrence that undertook a powersharing arrange-
ment.1 Of the fifteen cases of civil war recurrence, related powersharing arrange-
ments were undertaken in ten. And of the twenty-seven cases of nonrecur-
rence, powersharing arrangements were undertaken in sixteen. The list given in
table 7.2 shows that the most common types of powersharing are political and
security; and the latter generally involves the integration of rebel forces into the
state armed forces or police forces. In the following subsections I use specific
Making Peace Stick 187

Table 7.2 Powersharing by Type, All Cases of Recurrence and Nonrecurrence, 1944–2007
Country (relevant years) Political Military Territorial Economic
Angola (2002–) Yes Yes
Azerbaijan (1994–2005) Yes Yes Yes
Bangladesh (1997–) Yes Yes Yes
Bosnia (1995–) Yes Yes Yes
Burundi (1988–93) Yes ®
Central African Republic (1997–2001) Yes ® Yes ®
China / Tibet (1951–56) Yes
Croatia (1995–) Yes
Djibouti (1994–) Yes Yes
El Salvador (1992–) Yes Yes
Georgia / South Ossetia (1994–2004) Yes Yes
Guatemala (1996–) Yes Yes
Laos (1973–) Yes ®
Lebanon (1958–75) Yes Yes
Liberia (1996–99) Yes Yes
Mali (1994–2007) Yes Yes
Moldova (1992–) Yes Yes Yes
Mozambique (1995–) Yes Yes Yes Yes
Nepal (2006–) Yes Yes Yes
Papua New Guinea (1998–) Yes
Russia / Chechnya (1996–99) Yes ® Yes Yes ®
Sierra Leone (2000–) Yes Yes Yes
South Africa (1994–) Yes Yes Yes Yes
Tajikistan (1997–) Yes Yes
United Kingdom (1998–) Yes Yes
Zimbabwe (1979–83) Yes ® Yes Yes
N = 26
Note: Where power sharing failed, the “Yes” is followed by “®” for “reversal.”
Source: Author’s determinations, based initially and largely on Hoddie and Hartzell (2005).

cases of each of the four types of powersharing to illustrate its role in successfully
consolidating peace.

Political Powersharing
In several cases, political powersharing was deemed indispensable to the success
of a peace process. Powersharing was crucial in ending Djibouti’s civil war, for in-
stance, which began in late 1991 when disgruntled members of the minority Afar
ethnic group launched attacks as the Front for the Restoration of Unity and De-
mocracy (FRUD). The government and the FRUD signed a peace accord in De-
cember 1994, ending the conflict, although a splinter group persisted in fighting.
188 Examining the Cases

As part of the deal, several Afars and two FRUD members were appointed to the
cabinet. In 1999 the FRUD campaigned in support of the governing People’s
Rally for Progress party. This political powersharing was bolstered by security
powersharing that led to seven hundred ex-FRUD being incorporated into the
army by 1996.2 As one analyst put it, “Opting for an inclusive approach to tack-
ling political differences seems to have helped in bringing an end to the Djibouti
civil war” (Yoh 2003, 86).
Tajikistan offers another example of a country where political powersharing
has proven important to making peace stick. As per the 1997 agreement that
ended the war, 30 percent of cabinet posts went to the rebel United Tajik Op-
position (UTO). Although the elections of 2000 and 2005 yielded only two of
the sixty-three lower house seats for either of the two parties to emerge from the
rebel movements, the access to the system for UTO leaders seemed to make a
difference. Again, security powersharing bolstered the political arrangement.3 By
2000, 4,498 UTO combatants had been integrated into the armed forces, mainly
in stand-alone units with their own officers.
Other cases of political powersharing also bear mention. Liberia’s recurrent
civil war, from 1999 to 2003, ended in a powersharing accord that brought the
main armed factions together in cabinet posts and other ministerial appoint-
ments during the Transitional National Government before elections in 2005.
Despite serious corruption among senior officials, the political outcome of this
more inclusionary experience has proven much more successful than the exclu-
sionary Charles Taylor administration. Cases such as Bosnia, Northern Ireland,
South Africa, Mozambique, and Sierra Leone provide better-known examples of
political powersharing that have been successful to date.
Conversely, some cases of political powersharing have resulted in failure and
war recurrence. Lebanon was widely considered to have been a success for many
years between its adoption of powersharing in 1958 through the renewed civil
war in 1975. As mentioned in the discussion of powersharing’s weaknesses, most
analysts point to the inflexibility of that particular arrangement in the face of
shifting demographics that left Muslims, especially poorer Shiite Muslims, feel-
ing disenfranchised as their numbers grew. The arrival of thousands of Palestin-
ians in southern Lebanon also placed stress on the system. After 1989 Lebanon
adopted a revised formula of powersharing for the executive based on revised
population estimates.
In addition, powersharing agreements have sometimes not been imple-
mented, but the ongoing negotiations and / or promise of reaching an agreement
seems to hold recurrence at bay. Such are the cases of Bangladesh (1997) and
Nepal (2006), where some elements have been adopted but many others have
not. Most important in these cases, outright violation, rejection, or reversal of
inclusionary powersharing arrangements has not occurred.
Making Peace Stick 189

Two other cases of an apparent failure of political powersharing are not as

damaging as they might seem. First, as discussed above, Laos is an unusual case
in which regime change seemed to have occurred without warfare. If there had
been sufficient resistance and armed opposition, powersharing’s demise amid
the Pathet Lao Communist takeover of the state would have represented a re-
currence of warfare. Instead, given that the Pathet Lao were heavily dependent
on North Vietnam, after Saigon and Phnom Penh fell to Communist forces in
early 1975, the ruling Lao noncommunists decided to cede the capital rather
than see it taken by force. In other words, what would have been an ideologi-
cally driven war was preemptively ceded. The regional level of analysis—namely,
the toppling of regimes to Communism and the end of the neighboring wars—
played a crucial role here in leading to the failure of powersharing, but helps
account for why war did not recur.
Second, Burundi’s reversion to mass violence and armed conflict in 1993
was more a cause for the adoption of new powersharing mechanisms than it was
caused by the failure of the minimal arrangements that had prevailed after the
1988 armed conflict. As argued above and posited persuasively by Lemarchand
(1994, 2004) and Uvin (1999), Burundi and Rwanda’s waves of ethnic violence
from the 1950s to the 1990s represent interrelated episodes of exclusion, rebel-
lious attacks, massive counterattacks, and continued exclusion (Lemarchand
1994, 2004). The first serious attempts at inclusion of Hutus occurred with the
presidency of Pierre Buyoya. Just before a coup brought Major Pierre Buyoya to
power, Hutus held four of twenty government ministries, seven of sixty-five leg-
islative seats, and two of sixty-five seats on the central committee.4 Buyoya gave
more posts to Hutus in government after he came to power in 1987, but Tutsis
still dominated his cabinet and the army. The ethnic killings and recurrence of
armed conflict in October 1988 principally reflected the extreme exclusionary
behavior that persisted in access to services and lower-level government posts.
Powersharing did not fail in 1988 so much as it emerged at Buyoya’s instance
from the ashes of the prior war.
In 1993, the assassination by Tutsi soldiers of the country’s first Hutu presi-
dent sparked a decade of warfare and ethnic violence. Weak state institutions
contributed to the 1993 civil war, as did the initiation of multiparty elections
pressed upon then–president Pierre Buyoya starting in 1990. In response to
the earlier 1988 conflict, Buyoya had initiated a series of steps toward informal
sharing of power with Hutus. As described above, these included a commission
composed equally of Hutus and Tutsis to make recommendations toward an
inclusive state and a Charter of National Unity. But these steps toward power-
sharing ultimately were unable to stave off the 1993 civil war.
However, it was the failure of power-dividing measures, more than powersharing,
that provoked Tutsis to interrupt the democratization process by assassinating
190 Examining the Cases

the Hutu president, Melchior Ndadaye. The assassination of Ndadaye could, of

course, be seen as an abrogation of powersharing, in this case by the dominant
Tutsis threatened with the potential loss of their access to state power. Former
US ambassador Robert Krueger charged in his memoirs that Buyoya engineered
the coup that led to Ndadaye’s death (Krueger and Krueger 2007).
In any case, politics in Burundi after 1988 reflected a mix of small inclusion-
ary measures amid widespread exclusion of the vast majority of the population.
Inclusion also reflected partisan affiliations as much as ethnic allegiances (Uvin
1999, 266). Before the 1993 war, measures aimed at power dividing included the
ban adopted in the 1992 Constitution on ethnicity-based political parties and
the requirement that parties receive signatures from each of the country’s prov-
inces (Uvin 1999). It was fears of a reversal of the minimal inclusion achieved,
more than any setback in representation, that provided the context for the en-
raged outbreak of mass violence by Hutus once Ndadaye had been assassinated.
Powersharing agreements followed in 1994, 1998, and again in 2004–5 in con-
junction with the demobilization of important rebel groups.
Burundi thus serves as an example of the reversal of powersharing in the recur-
rence of the 1993 civil war. Some inclusionary measures aimed at dividing power
were clearly at work in the recurrence of warfare in 1993. More generally, how-
ever, persistent exclusion underlay recurrent episodes of violence and warfare.

Security Powersharing
As mentioned above, exclusion from the military and police forces seems to be a
very provocative factor in prompting war recurrence. Conversely, the incorpora-
tion of former enemies is especially important because it provides concrete, and
usually persistent (at least through several years), guarantees to rebels that sacri-
fice their military organization and weapons. It signifies a curb on the exercise of
the arbitrary or deliberate targeting of former enemies by the state. As discussed
above, a recent study shows that military integration is a statistically significant
factor in the consolidation of peace.
Security powersharing proved vital to consolidating peace in several cases.
Although several contextual variables favored the successful peace process in
Mozambique in 1992, the incorporation of former RENAMO rebels into the
Mozambican armed forces was crucial. By 1995, three years after the negotiated
agreement and cease-fire, 3,662 of 12,195 soldiers (30 percent) in the Mozambi-
can armed forces were former RENAMO combatants (Young 1994, 2).5 As is ex-
plained above, several cases of political powersharing (e.g., Djibouti, Tajikistan)
were accompanied by crucial security powersharing as well.
El Salvador shows that security powersharing can also prove important
within police rather than military forces. Although the majority of the members
Making Peace Stick 191

of the newly created civilian-run national civilian police were not former com-
batants, up to 20 percent of the force could be former guerrillas along with an
equal number of members of the former military-controlled national police (Call
1999, chap. 4). The rebel participation in the new force, along with its participa-
tion in electoral processes, gave important security guarantees to the Farabundo
Marti National Liberation Front.6 Similarly, the creation of a new police service
to replace the Royal Ulster Constabulary in Northern Ireland via the 1998 Good
Friday Agreement proved important in alleviating both sides’ fears about poten-
tial armed actions by spoilers. As part of the reform process, 50 percent of all new
recruits for the Police Service of Northern Ireland have been Catholics.
At the same time, security powersharing at times seems epiphenomenal, not
essential to the maintenance of peace. Angola’s 2002 peace accord, for instance,
included the incorporation of ex-UNITA combatants into the armed forces, but
the killing of UNITA leader Jonas Savimbi only weeks earlier is seen as the single
biggest factor in drawing the bulk of the rebels to lay down their weapons. The
unification of the three armies in the new Bosnian state encountered numerous
obstacles and remains incomplete at the operational level. However, the politi-
cal unification, intertwined with territorial autonomy, served as the foundation
for the successful cease-fire, not the slow and partial unification of armed forces
between the two entities.

Territorial Powersharing
Conflicts over territorial autonomy or secession merit special attention. Despite
the prevailing sense that such territorial conflicts are very difficult to resolve, the
data are unclear. Eight of the twenty-seven cases of nonrecurrence involved sa-
lient issues of territorial autonomy or secession. Of these eight successfully re-
solved armed conflicts, five (Bangladesh, Bosnia, Papua New Guinea, and North-
ern Ireland) included powersharing provisions regarding territorial governance.
Territorial powersharing also may lead to war recurrence. Of the fifteen cases
of recurrence, territorial autonomy or secession was salient in six. Of these six,
five countries reached some sort of powersharing involving territory (usually in
conjunction with other types of powersharing) before experiencing war recur-
rence. In such cases, powersharing was not a panacea. Cases such as Chechnya,
Azerbaijan, Tibet / China, and South Ossetia / Georgia show that providing some
measure of autonomy may help secure a cease-fire, but it does not ensure consoli-
dated peace. In most such cases, an autonomy arrangement only kicked the can
of final status down the road, permitting other factors to reignite warfare. In some
cases inactivity on the part of a powerful state whose sovereignty is recognized
over the territory seems common and may well have occurred before other trig-
gers renew the armed conflict.
192 Examining the Cases

Economic Powersharing
The definition of economic powersharing is wide-ranging, sweeping from mini-
mal consultations with former rebel parties on economic development deci-
sions to property guarantees to specified shares of oil or other high-rent natural
resources. In South Africa, for instance, the guarantees of private property re-
sponded to a key insecurity among whites, given the apparent dominance that
black South Africans would exercise politically. In Guatemala guarantees in the
1996 peace accords provided for land reform and for expanded recognition of
social and economic rights among indigenous populations. The 2006 Nepalese
peace accord also contained measures to reform the feudal land tenure system,
to provide remedial development to the landless, and to protect private property.
Unfortunately, these provisions lay dormant for the transitional period, and thus
were neither abrogated nor fulfilled. At the same time, economic powersharing
may have a relatively small impact on whether or not peace sticks. In Lebanon,
for instance, the disenchantment over political offices far outweighed the rela-
tively minimal influence of the stipulation that all religious communities would
benefit equally from development aid.
In addition, the sharing of natural resources as a means of conflict resolution
can perversely aggravate civil wars. Such was the case with Sierra Leone’s 1999
Lomé powersharing accord. The agreement made Foday Sankoh, the leader of
the Revolutionary United Front, a vice president and chair of a commission over-
seeing the country’s diamonds and other natural resources, a position he used
to fortify his rebel group’s capabilities and attack peacekeeping forces in 2000.
Ironically his participation in the government via the powersharing arrangement
had brought him into the capital and made him more vulnerable.7 Government
troops and intervening British forces were consequently better able to locate and
capture him and his lieutenants, and thus bring a more decisive end to the civil
war by 2002. Nevertheless, Sierra Leone’s 1996 Abidjan and 1999 Lomé power-
sharing agreements show that powersharing is not always successful, and can in
the midst of war portend continued waves of violent warfare.

Powersharing and Peace Consolidation:

Examining the Pool of Cases
Above I examined the impact of the four types of powersharing, drawing illus-
tratively on selected cases of nonrecurrence. Here I examine how powersharing
correlates with both recurrence and nonrecurrence, drawing from the pool of
forty-two cases of civil wars that either never recurred or that clearly recurred as
internal armed conflicts. Of these forty-two cases, twenty-six experienced some
form of powersharing.
Making Peace Stick 193

Table 7.3 Incidence of Powersharing among All Recurrences and Nonrecurrences, 1944–2007
Status of
Powersharing Nonrecurrence Recurrence
Abrogated Laos, 1973 Burundi, 1972, 1988, 1993–2006
Central African Republic, 1996–97, 2001–2
China / Tibet, 1950–51, 1956–59
Liberia, 1989–96, 1999–2003
Russia / Chechnya, 1994–96, 1999
Zimbabwe, 1972–79, 1983–87
N=1 N=6
Not abrogated Angola, 1975–2002 Azerbaijan, 1992–94, 2005
Bangladesh, 1976–97 Lebanon, 1958, 1975–90
Bosnia, 1992–95 Mali, 1989–94, 2007
Croatia, 1992–95
Djibouti, 1993–94
El Salvador, 1979–92
Guatemala, 1968–96
Moldova, 1992
Mozambique, 1976–95
Nepal, 1997–2006
Papua New Guinea, 1988–98
Sierra Leone, 1991–2000
South Africa, 1983–94
Tajikistan, 1992–97
UK / Northern Ireland, 1969–99
N = 15 N=3
Source: Author’s case-by-case determinations.

Table 7.3 includes twenty-five of these cases, based on whether or not these
powersharing arrangements were abrogated in the key few years after conflict
termination (Georgia is an ambiguous case, and so I have set it aside). The as-
sessment of whether a powersharing arrangement is violated is based on whether
any of the four types of powersharing was perceived to have been reneged upon
or withdrawn by the state. Delays in meeting targets to incorporate former rebels
into the armed forces, as has occurred in Nepal, for instance, is not sufficient to
be deemed “abrogation.” Conversely, the ouster of hundreds of former rebels
from the army’s ranks would be considered “abrogation.” Table 7.3 draws directly
on table 7.2.
Table 7.3 shows the high correlation between the abrogation of powersharing
arrangements, on the one hand, and civil war recurrence, on the other hand.
Fifteen cases of preserved powersharing did not experience a recurrence, and
only three such cases (Azerbaijan in 2004, Lebanon in 1975, and Mali in 2007)
did experience recurrence of internal armed conflict. As such, 87 percent of
194 Examining the Cases

powersharing arrangements that were not violated resulted in the preservation

of peace. All cases of peace have held for at least ten years except the recent cases
of Angola (2002) and Nepal (2006).8
Among the seven cases where powersharing was abrogated, all but one case
experienced a recurrence of internal armed conflict. These six cases include Li-
beria in 1999, ethnic warfare in Burundi and the Central African Republic, sepa-
ratist struggles in Chechnya and Tibet, and the falling out among formerly allied
proindependence rebels in Zimbabwe. Laos is the only case of nonrecurrence,
and the circumstances of the choice not to contest the regionally supported
Communist takeover in 1975 were discussed above.
It is worth recalling here that powersharing is neither a surefire nor a per-
manent ticket to peace. Table 7.3 does not include many cases of powersharing
experiments that have failed. As mentioned above, Sierra Leone’s two experi-
ences with powersharing in earlier peacemaking efforts did not yield success,
although they arguably may have laid the groundwork for subsequent resolution
and success. Aside from the cases analyzed here, powersharing efforts have failed
to bring about an end to civil wars in the Democratic Republic of the Congo, Af-
ghanistan, Georgia (Abkhazia and South Ossetia), Sudan, and Yemen. In Colom-
bia, successful powersharing is blamed for creating the conditions for the emer-
gence of the leftist insurgency in the 1960s that persists today (Hartlyn 1988;
Oquist 1980; Bergquist, Peñaranda, and Sánchez 1992; Call 1995). The overall
record of powersharing in the midst of civil wars—that is, not solely as power-
sharing is linked to periods of sustained peace, as defined here—has been the
subject of much scholarship. Although contested, this scholarship increasingly
sees powersharing as a positive and influential mechanism in resolving conflicts,
despite its flaws and depending on the types of sharing and the context (Sriram
2008; Hoddie and Hartzell 2005).
Therefore, only careful and qualified claims are possible based on the evi-
dence examined here. Powersharing is not sufficient to ensure sustained peace, as
cases like Mali and Lebanon show. However, table 7.3 suggests that powersharing as
one instance of inclusionary behavior highly favors sustained peace. A total of 83 per-
cent of cases where powersharing is observed correspond to the instances of
nonrecurrence of civil wars or even internal armed conflict. These cases include
not just celebrated successes such as Northern Ireland, El Salvador, Mozam-
bique, and South Africa but also Angola, Tajikistan, Djibouti, and Papua New
Guinea. Although powersharing has been marginal in some of these successes,
as detailed above, in most cases observers unanimously argue that powersharing
has proven essential to maintaining peace over a period of many years.
Despite its flaws in concept and practice, powersharing shows a remarkable
affinity with consolidated peace around the world. As stated, it is not sufficient
to ensure peace, but it is highly favorable for peace. We now turn to instances
Making Peace Stick 195

of nonrecurrence that show how powersharing is not only insufficient but not
necessary for peace.

Beyond Powersharing: Inclusionary Behavior and Peace

As elaborated above, powersharing is one form of what may be considered “in-
clusionary behavior” by political actors. It is conceptualized here to include state
and regime structures whose design permits the chance for diverse political par-
ties, especially those associated with the prior combatant forces, to participate
in governance through appointed or elective state offices. Inclusionary political
behavior thus includes also instances of “power dividing,” whereby civil liber-
ties are minimally guaranteed; electorally competitive regimes that permit the
contestation for office; and decentralization of authority in ways that permit ac-
cess to local power. Although territorial autonomy could be considered a form of
power dividing, it is addressed as a form of powersharing above.
What conclusions might one reach regarding the role not just of power-
sharing but also of inclusionary behavior more broadly in whether civil wars re-
cur as internal armed conflicts? Recall from table 7.3 that among all core cases of
recurrence plus those cases of the absence of any civil war recurrence, only eigh-
teen instances of powersharing were maintained, of which fifteen corresponded
to nonrecurrence and only three led to recurrent armed conflicts. Apart from
these cases, I find only two cases of inclusionary behavior that do not represent
powersharing arrangements. Both these cases—the Dominican Republic (after
its war of 1965) and Cyprus (after 1974)—experienced political arrangements
that provided some degree of participation in their own governance to the armed
faction that contested state power and failed to find victory.
Inclusionary behavior does not need to stem from heartfelt goodwill or a
strategy aimed at integration or reconciliation. The crucial element here is that
those insurrectionary forces that did not win were not killed, persecuted, or
excluded from their own governance. In the case of Northern Cyprus, Turkish
Cypriots have exercised significant self-rule, even as their aspirations of inde-
pendence have remained deferred. Thus Northern Cyprus operates as a de facto
state, electing its own president and Parliament, managing its own 5,000-person
defense force as well as a larger reserve force, and issuing official documents like
birth certificates and passports.
In the Dominican Republic the 1965 US intervention halted the warfare and
installed an interim government that organized elections in 1966, even as the
killings of dozens, perhaps hundreds, of social democrat “constitutionalists” per-
sisted.9 Jorge Balaguer, a protégé of the anticommunist former dictator Rafael
Trujillo, was elected president and reelected twice (Torrijos 2008, 73–74). How-
ever, the constitutionalist former president Juan Bosch was a candidate in the
196 Examining the Cases

elections, and his party gained five of twenty-seven Senate seats and twenty-six
House seats in 1966 (Maríñez 1990). His party continued to compete in subse-
quent elections and elected presidents in 1978, 1982, and 2000. This form of de-
mocracy was illiberal and curtailed; it reflected the anticommunist foreign policy
of the United States and its allies, and it thus inhibited participation by those who
had risen up against a military government in 1965. Yet it was not exclusionary.
Former combatants had the chance to participate in governance and did so, even
though the inclusionary character of political authority was not benevolent nor
perhaps even voluntary in this case or that of Northern Cyprus. If it were not for
external powers and / or intergovernmental bodies using, and then threatening to
use, force, then exclusionary behavior could well have ensued.

Peace and Exclusionary Behavior?

Powersharing is not a necessary condition for making peace stick, because a num-
ber of long-standing cases of nonrecurrence had no powersharing arrangement.
On the contrary, it is striking how many cases of nonrecurrence reflect quite exclu-
sionary behavior—totalitarian or authoritarian regimes that have jailed, killed, or
repressed political opponents. Such cases include Cuba in 1959, Jordan in 1970,
Bolivia in 1952, Vietnam in 1975, South Korea in 1950, and Paraguay in 1947.10

Victorious Socialist Regimes

One such group includes revolutionary Communist regimes that had little inter-
est in the inclusion of their former enemies. In Cuba, for instance, hundreds of
former soldiers and police officers were tried and executed, although many who
had been deemed sufficiently nonpolitical were retained (Clark 1992). Large
farms and private productive facilities were nationalized without guarantees for
property owners, and opposition political leaders were harassed, monitored, and
arrested rather than included in governance. In a statement that applies to the
several socialist and Communist regimes of the Cold War era, “Within the logic
of the revolution, there was no legitimate opposition to la patria and socialism,
and opponents suffered repression or exile” (Perez-Stable 1999, 105). Such was
the politics of Vietnam after its revolutionary victory in 1975. Bolivia’s 1952 post-
war government represented a socialist military government that engaged in the
repression and exclusion of its former enemies.

Victorious Anticommunist Regimes

Among counterrevolutionary cases of victory, exclusionary behavior was equally
egregious in the initial years of their respective postwar periods. In Paraguay, for
Making Peace Stick 197

instance, the point of the internal conflict was precisely to keep leftists and Com-
munists out of power, which the regime successfully did for decades until 1989.
The Greek Civil War similarly ended with the Communist Party being outlawed
(until 1974), many of its leaders tried and executed, and the remainder jailed or
in exile (Samatas 1986). Even Costa Rica, which had been peaceful since 1948,
represents exclusionary politics because victorious center-right rebels reneged
on their pledge to permit the Communist Party to compete electorally and jailed
some of its leaders.
South Korea’s civil war offers another instance of postwar anticommunist
exclusionary behavior technically followed by no recurrent internal armed con-
flict. Its civil war is not the major war that divided North Korea and South Ko-
rea, which were already separated by 1948 (and thus were considered to have
fought an interstate war, one that has also not recurred amid heavy international
troop presence). Instead, the South Korean nonrecurrent civil war that appears
in the Fearon and Laitin data set is the insurrection of outraged residents of Jeju,
an island where up to 30,000 were killed in the period 1948–50. Months after
police fired on a demonstration commemorating Korea’s struggle against Japa-
nese occupation in 1947, increasingly frustrated residents attacked twelve gov-
ernment police posts. Because many Koreans considered the government and
police to have collaborated with the Japanese occupiers, the socialist South Ko-
rean Workers Party found fertile ground in organizing resistance on Jeju, joined
(and armed) by several hundred defectors in mutinied units of the Korean armed
forces. In this place, the lingering effects of World War II occupation politics
coupled with Cold War ideological competition to fuel protests that grew into
organized armed actions and deadly reprisal massacres by government troops
and anticommunist youth volunteers. The rebels, who eventually received exten-
sive material support from North Korea and began flying its flag, were brutally
suppressed by anticommunist government forces.11
Exclusion followed the war. In 1948 the South Korean legislature banned the
South Korean Workers Party and made it a crime to even mention the Jeju upris-
ing and subsequent massacres. An official truth commission into the Jeju upris-
ing found that, rather than any sort of inclusion of those who had arisen against
the government, at least 14,373 had been killed, mostly amid the “wanton de-
struction” of 95 percent of the villages on a central mountain and the elimination
of most of their inhabitants.12 In April 2006 President Roh Moo-hyun officially
apologized for the massacres associated with the Jeju uprisings and granted the
island’s long-standing wish for greater autonomy, denominating it “Jeju Special
Self-Governing Province.”13
These cases of exclusionary postwar regimes that successfully averted sub-
sequent civil wars reveal that excluding former enemies is compatible with sus-
tained negative peace. However, they also reveal other insights. They show that
198 Examining the Cases

historically specific and normative contexts shape the ideas and patterns of war-
fare, peace, regime type, and governance structures and behavior. Seven of these
nine cases (77 percent) are best seen through a Cold War ideological prism, even
where local power struggles and identity politics underlay wars exhibiting East /
West ideological divides. Admittedly, locally specific contests were subjected to
the manipulation of images, fears, and symbols to fuel the fears or interests of
transnational movements for or against Communism.
Nevertheless, transnational norms in historical moments worked to render
acceptable and even laudable the abrogation of liberal democratic politics and
the adoption of Communist or socialist one-party models. Those historical mo-
ments were global—the end of World War II, the delegitimation of fascism, and
the rise of two models of democratic capitalism and Soviet-style Communism—
as well as regional in nature—the wave of Communist regimes that swept South-
east Asia, the insurgencies that swept from Cuba to South and Central America,
and the legitimation of military institutional authoritarian regimes in South
America. These geographic contexts and historical moments are not determi-
nate, but they are overlooked by both large-N quantitative methods that treat
only countries as the unit of analysis and by qualitative studies that have a depth
of only one or two cases.
Inherent in the philosophical, political, and economic thought underlying
these regimes was not just the marginalization of formerly dominant bour-
geois political parties and economic elites but also their active suppression
through physical elimination, incarceration, reeducation, and / or exile. Con-
versely, anticommunist governments dehumanized Communists and sought to
destroy their organizations. For both Communists and anticommunists, torture
and abuse were at times used as a public display of official rejection, humiliation,
and terror.
Analyzing the cases of exclusionary behavior among countries that experi-
enced one civil war but no recurrence reveals the importance of historical con-
text. Roughly two-thirds of the twenty-seven cases exhibited inclusionary be-
havior, and one-third showed exclusionary conduct. At first glance, exclusionary
politics is not incompatible with sustained peace.
However, every single case of nonrecurrence under exclusionary behavior oc-
curred before 1989. Conversely, inclusionary behavior characterizes successful
terminations of wars after 1988 (see table 7.4). Fifteen of seventeen cases (88 per-
cent) of nonrecurrent wars that ended with inclusionary behavior occurred after
the Cold War’s conclusion in 1989. Only two cases of nonrecurrence with inclu-
sionary behavior occurred before 1989. In both those cases, Cyprus in 1974 and
the Dominican Republic in 1965, external armies intervened to stop fighting on
behalf of an ally, which led to the forging of a political transition that precluded a
victory against the status quo ante but forestalled outright exclusionary behavior
by the state’s armed forces.
 Making Peace Stick 199

Table 7.4 Inclusion and Peace after the Cold War: All Cases of Recurrence and Nonrecurrence, by
Political Behavior, 1988–2007
Type of Behavior Nonrecurrence Recurrence
Exclusionary Burundi, 1988, 1993–2006
Central African Republic, 1996–97, 2001–2
East Timor (Indonesia), 1975–99, 2006
Haiti, 1991–95, 2004
Liberia, 1989–96, 1999–2003
Peru, 1981–95, 2007
Russia, 1994–96, 1999–
Rwanda, 1962–65, 1990–2002
N=0 N=8
Inclusionary Angola, 1975–2002 Azerbaijan, 1992–94, 2005
Bangladesh, 1976–97 Mali, 1989–94, 2007
Bosnia, 1992–95
Croatia, 1992–95
Djibouti, 1993–94
El Salvador, 1979–92
Guatemala, 1968–96
Moldova, 1992
Mozambique, 1976–95
Nepal, 1997–2006
Papua New Guinea, 1988–98
Sierra Leone, 1991–2000
South Africa, 1983–94
Tajikistan, 1992–97
UK / Northern Ireland, 1969–99
N = 15 N=2
Source: Author’s case-by-case determinations.

These cases suggest that the end of the Cold War marks a historic shift in
historical contexts that dramatically shape the chances and character of sustain-
able peace in three important ways. First, ideology was the mobilizing principle
and point of reference for most of the conflicts during the Cold War, but not
afterward. Of the nine cases of exclusionary cases that did not recur, ideological
battles over Communism were central to all of them except Jordan in 1970. Com-
munist rebellions or anticommunist counterrevolutionary efforts lay at the core
of efforts to mobilize support and articulate stances, whether it was Paraguay or
Bolivia (with ethnic overtones) in South America; the Caribbean Basin cases of
Costa Rica and the Dominican Republic; the Asian cases of Vietnam, Laos, and
South Korea; or Europe’s Greece. And although several cases of postwar inclu-
sionary behavior occurred in tandem with conflicts that began as ideologically
centered wars during the Cold War, only one armed conflict of the nonrecur-
rences commenced after 1988 (Nepal in 1997).
200 Examining the Cases

Second, global norms of war termination also shifted. As noted earlier in the
chapter, outright victory had been the most common form of war termination
during (and before) the Cold War. However, negotiated settlements became far
more prevalent in conjunction with the end of the Cold War. We know this from
other research on civil war settlements (Fortna 2004a). This shift reflected at-
titudes among armed factions as well as neighboring powerful states and inter-
national organizations.14
Third, just as international actors have played a more important role in fa-
cilitating and mediating peace agreements among warring parties since the Cold
War, foreign troops have also played a notably more prominent role since 1988.
In none of the pre-1989 exclusionary cases of nonrecurrence did international
peacekeeping troops appear. Certainly foreign armed forces arrived and played
combat roles in several cases. These include the assistance from US armed forces
in putting down rebellions in Greece and South Korea. In addition, neighboring
states such as Nicaragua and North Korea helped one or both sides in the civil
wars of Costa Rica, Laos, Vietnam, and South Korea. The closest thing to inter-
national intervention in preserving peace in these cases of exclusion is the role
of teams of diplomats and military experts from the Organization of American
States in facilitating the resolution of conflicts launched by exiles from Nicara-
guan soil against the government of “Pepe” Figueres in late 1948 and again in
1955. However, those were small missions with no combat component (Herz
2008, 6, 13). They also occurred after Figueres had dissolved the country’s
standing army, failing not because of international deployments but because
Costa Rica’s former insurgent combatants regrouped and joined civilians in tak-
ing up arms.
Among cases after 1988, international troop deployments played a salient role
in several. External combat troops helped reach and / or maintain cease-fires in
Bosnia, Croatia, Cyprus, Mozambique, and Sierra Leone. In the first three of
these cases, many observers maintained for years that a withdrawal of foreign
troops would result in war recurrence. In addition, unarmed UN military observ-
ers (usually no more than three hundred) were deployed to strengthen peace
processes in El Salvador, Guatemala, Tajikistan, and Nepal. Still others—such
as South Africa, Bangladesh, and Djibouti—showed that external troops are not
necessary at all to advance. Interestingly, the two cases of inclusionary behavior
before 1989 are also the only two pre-1989 cases in the pool being considered
here in which international troops deployed. These include the US and Orga-
nization of American States deployments to the Dominican Republic and the
involvement of Turkey’s armed forces in forcing an end to hostilities in Cyprus
in 1974. The latter two cases suggest that foreign troop deployments helped avert
the normal outcome of exclusionary behavior during the Cold War. We will re-
turn to the issue of foreign troop presence and “frozen” conflicts.
Making Peace Stick 201

The dramatic turning point of 1988–89 marks important transformations of

global norms and the interests of major powers. As such, I examined the cases
of post–Cold War cases separately. Table 7.4 presents the ten cases of civil war
recurrence after 1988, and the fifteen cases of nonrecurrence after 1988.15
Of those twenty-five cases, twenty-three (92 percent) conform to the argu-
ment here that exclusionary behavior is the key causal factor in recurrence, and
that inclusionary behavior helps prevent recurrence. Only two cases of inclu-
sionary behavior—Azerbaijan in 2004 and Mali in 2007—led to recurrence.
In both cases persistent dissatisfaction with the status quo underlay recurrences
that were triggered partly by external factors. In the case of Mali, Tuareg upris-
ings in adjacent Niger helped foment unrest and resentment among Tuaregs in
Mali. In both cases casualties from the recurrence were fewer than a hundred
deaths, and do not warrant designation as a full civil war.
What are the implications of table 7.4 and the dramatic changes that occurred
with the end of the Cold War? Most important, given the normative context of
the post–Cold War era (as dated as that may sound), exclusion is no longer a vi-
able path to negative peace. Whereas before the Cold War regimes could suppress
dissent, jail opponents, and through repression cow armed opposition or nip it
in the bud, since the Cold War’s end such post–civil war exclusionary behavior
has led without exception to armed conflict. In none of the cases examined has
exclusionary behavior produced sustained peace.
This finding does not mean that exclusionary behavior is incompatible with
suppressing dissent. In “grandfathered” cases from the Cold War, such as Cuba
and Equatorial Guinea, a sustained exclusion of regime opponents has coexisted
with a negative peace. In cases of multiple or simultaneous armed conflicts—
such as Somalia, Yemen, and the Democratic Republic of the Congo—exclu-
sionary behavior persists alongside conflicts, despite instances of inclusionary
behavior toward some groups.
Furthermore, the sort of repression and exclusion common during the Cold
War has changed in character and content. Repressing ideological enemies has
not been as common since 1988. Instead exclusion is much more often seen in
terms of identity groups presented and perceived as threats to other identity
groups. In addition, authoritarianism and exclusionary behavior are not equiva-
lent. Thus some authoritarian or semiauthoritarian regimes like those of Burma
and Rwanda have managed countries without internal armed conflicts, but with
some degree of inclusionary behavior toward former enemies. Even as demo-
cratic regimes generally now include sufficient mechanisms of inclusion, the ab-
sence of democracy is not an automatic signal of exclusionary politics.
Nevertheless, the importance of this finding about exclusion since the 1989
turning point in history should not be lost. The changed international context
means that governments are not free to flagrantly exclude their former enemies
202 Examining the Cases

after war and expect to be able to enjoy stability. Inclusionary behavior toward
former enemies is the surest path to nonrecurrence; exclusionary behavior is
the route to renewed armed conflict. Inclusionary behavior need not be power-
sharing, certainly not in the inflexible, timeless formula that facilitated the return
to warfare in Lebanon in 1975. Yet former rebels and their supporters require
some hope of exercising control over their own lives through institutions of au-
thority, especially the state, either nationally or locally. Providing them with ei-
ther a guaranteed share of that power or a regular guaranteed opportunity to seek
it through elections with sufficient guarantees is the sine qua non of sustainable
peace in the contemporary era.
Several components of the international context are important here. Among
them are new transnational norms of elections as legitimating mechanisms and
global human rights standards. A form of authoritarianism that leaves large eth-
nic or religious groups with no hope of inclusion in the polity is likely to confront
armed challenges. Even elected, formally democratic regimes that limit contes-
tation for formerly mobilized populations can expect to face renewed armed
conflict. This is precisely what occurred with governments that were elected in
Liberia, Haiti, and East Timor.
In addition, these cases suggest that postconflict societies are especially sus-
ceptible to the impact of transnational norms, institutions, and international in-
fluence in ways that render exclusionary politics unsustainable. The international
peacebuilding “industry” emphasizes the importance of inclusion, competitive
electoral systems, powersharing, and certainly some measure of responsiveness
to former enemies. Governments that engage in outright exclusion can expect to
receive a colder international reception and even a drop in international aid, as
occurred in both Liberia under Charles Taylor and Haiti under President Aris-
tide. Excluded groups and former combatants perceive these incentive structures
and norms. Once they see that their political or physical fate is threatened, the
international environment encourages rejection and justifies protest in the face
of such injustice. In some cases external assistance for renewed armed conflict in
the face of exclusion is more direct.

International Troops and “Frozen” Conflicts

A survey of the cases of nonrecurrence points to an important element of non-
recurrence: the presence of international peacekeeping forces, either from an in-
ternational organization or from interested bilateral powers. The positive impact
of third parties, especially armed troops, is much celebrated in the literature and
has reached near consensus as a factor in policy-oriented work.
“Frozen conflicts” is a term that emerged in the 1990s to refer to the several
wars that achieved cease-fires without any perceived permanent political settle-
Making Peace Stick 203

ment. The term was applied principally to the several wars and secondary civil
wars related to the breakup of the Soviet Union: Moldova (Transnistria), Geor-
gia (Abkhazia and South Ossetia), Azerbaijan (Nagorno-Karabakh), and the
Caucasus (Chechnya). The moniker has been extended to a number of other
conflicts, including Bosnia from 1995 to the present, Kosovo since its 1999 war,
and Cyprus since 1963. The “frozen” character of these conflicts corresponds in
all cases to the presence or threats of invasion by international troops stationed
specifically to deter renewed combat. In Transnistria, for instance, some 1,300
troops (mainly Russian) of the Commonwealth of Independent States form part
of a Joint Control Commission designed to prevent flare-ups of attacks. Even
where the numbers of foreign troops have been reduced significantly (as in the
Balkans), their value derives from their deterrent effect and the promise that
more forces could be deployed if necessary.
Scholars have still not fully fleshed out our understanding of the concept or
the wars to which it refers. Lynch (2004, 41) has written that these wars are
“events [that] have developed dynamically” in these conflicts, changing the ob-
jective conditions and subjective perceptions of these conflicts (Lynch 2004;
2005, 192–95). Most important, he suggests that the empirical administration
of territory by de facto states has increased the practice, perception, and expecta-
tion of sovereignty among elites of contested separatist territories.
An examination of the cases of recurrence and nonrecurrence of civil wars
here suggests three important findings with regard to frozen conflicts. First, the
concept of “frozen conflicts” calls into question whether “peace” can be said to
prevail in these societies. I have counted most such societies as “peaceful” given
a widespread perception that armed conflict is over, at least for the time being
and the near future. It is not clear that all such societies conform to this defini-
tion, which again raises the question about my definition of “peace” and its op-
erationalization. By including these cases, I have sought to cast a wide net on the
diversity of manifestations of “peace.” However, one must acknowledge that the
quality of perceived peace is less deep and sure in such frozen conflicts than in
other postwar societies. Episodic skirmishes in Georgia, Moldova, and Azerbai-
jan, for instance, show that peace was fragile even as no internal armed conflict
was recorded.
Second, these cases confirm a prominent, oft-cited claim in the literature on
civil wars and peacebuilding: the importance of third-party forces in deterring
renewed armed conflict. Of the twenty-seven conflicts that did not recur, five
had such external troops serving in a deterrent peacekeeping role: Cyprus (af-
ter 1964), the Dominican Republic (1965), Moldova (1992–), and Bosnia and
Croatia (1995–). In some cases that force was more welcome (e.g., South Ko-
rea) than others (the Dominican Republic). In several of these cases, observers
widely believe that renewed armed conflict would recur if not for the presence of
204 Examining the Cases

these external third-party troops providing a third-party deterrent and verifica-

tion role. In Bosnia, for instance, the presence of the NATO-commanded Imple-
mentation Force (IFOR) (initially called the Stabilization Force, or SFOR) was
indispensable in preventing a likely recurrence of war in the years after the 1995
Dayton Peace Accords. The peak deployment of IFOR troops was 60,000 in
1996, and the size had declined to some 12,000 by 2005, when the mission came
under command of the European Union. However, the presence of these troops
remains vital to the maintenance of the cease-fire and the beginnings of creating
a multicultural state.
Further evidence of the importance of the presence of external or third-party
military forces is the timing of recurrences. In the three main cases studied ear-
lier in the book—Liberia, Haiti, and East Timor—recurrence occurred within
several months of the withdrawal of international combat-ready peacekeeping
troops. In Liberia, the UN’s troops departed only three months after Taylor’s
1997 election, and the peacekeeping mandate of the Economic Community of
West African States’ Cease-Fire Monitoring Group (ECOMOG) concluded in
February 1998. Although 5,000 ECOMOG troops remained after that time to
conduct capacity-building tasks, the number had diminished by early 1999 as a
result of tensions between the Taylor government and the regional organization
(Tuck 2000, para. 7). The UN created a special peacebuilding support office
with many of the same tasks as the concluding peacekeeping operation but with-
out any troops. Liberia’s 1999 recurrence occurred within eighteen months of
the termination of the peacekeeping mandate, and just as the last of ECOMOG’s
capacity-building effort was being drawn down. The last troops left in October of
that year just as the second civil war was beginning in earnest.
Similarly, the UN’s successive peacekeeping operations to Haiti fully departed
in 2000, shortly before the second election of Aristide to the presidency. Within
three years of his February 2001 swearing-in, Aristide was beset by armed attacks
that constituted an internal armed conflict by late 2003. With the departure of
Aristide on Leap Day and the redeployment of international troops in March
2004, the threat to stability subsided. In East Timor the UN peace operation
closed in 2002, and it was followed by a political mission that did not possess
a troop component. Instead, Australian troops remained as a stabilization force
until their full withdrawal in that role in June 2005, when they left only a handful
of troops in a training function. Within one year, the tensions within the armed
forces and the police had erupted into the recurrent internal armed conflict ana-
lyzed above.
The Central African Republic offers another example where the recurrence
of internal armed conflict seemed to follow on the heels of the withdrawal of
international peacekeeping troops. As detailed above, the recurrence of 2001
involved first a coup attempt against then–president Patassé for his perceived
Making Peace Stick 205

bias in treatment of Yakoma soldiers, which led to clashes throughout that year
and 2002. However, the presence of international troops corresponds to rela-
tive stability. First, deployments by the Mosques and Imams National Advisory
Board helped prevent violent attacks (including on the board’s forces) from
growing into broader civil war. The subsequent deployment of peacekeeping
troops by the UN Mission in the Central African Republic saw little violence
between the 1997 Bangui Agreements and the end of the peacekeeping mission
and the withdrawal of French forces from the Central African Republic in 2000.
Second, however, violence emerged in early 2001, roughly one year after these
troops withdrew.16 The Central African Republic, Liberia, Haiti, and East Timor
all show the susceptibility of postconflict countries—especially poor countries
with weakly institutionalized polities—to recurrent conflict in the early years
after international peacekeeping deployments are withdrawn. They affirm the
findings of research cited above by Fortna (2008), Walter (2004), Doyle and
Sambanis (2006), and Sambanis (2008) that the presence of peacekeeping
troops correlates with a longer duration of stability.
Third, territorial powersharing may not be fruitful in the absence of political
or military powersharing in more integrated, permanent arrangements. Indeed,
territorial powersharing can create frozen conflicts with perverse incentives to
achieve more sustainable, long-term solutions. Ultimately, once territorial de
facto administration commences, the population looks to the de facto authority
for services, shifts its expectations, and confers greater legitimacy on that de facto
state and away from the de jure state. In this way, territorial powersharing is the
most divisive form of powersharing, given that it does not require the separate
pieces of a whole to communicate, much less cooperate. Minority members and
even units within the armed forces or police must at least communicate and tac-
itly cooperate in order for the institution to function effectively and the leaders to
receive accolades. Cabinets divvied up according to powersharing arrangements
still must work together under the guidance of national leaders to function. In
contrast, the incentives for cooperation between separatist regions that have
won some degree of autonomy generally use it to institutionalize that autonomy.
Moreover, their inhabitants become more accustomed to and influenced by the
discourse of sovereignty and independence—making it extremely hard to re-
verse de facto quasi-statehood. Given these dynamics, unless powerful economic
or symbolic capital is available to sway inhabitants of a separatist region not to
separate, then consistent pressure for independence will probably persist, leaving
armed action as more necessary to alter the status quo on the ground.
To recap, “frozen” conflicts represent an important subset of nonrecurrent
civil wars. The absence of recurrence in these cases may be due to the inclu-
sionary behavior of the state and elites, but it may also (or instead) stem from
the altered incentives that derive from third-party armed guarantors. Frozen
206 Examining the Cases

conflicts also offer a number of hypotheses, some of which have been advanced
and even explored by researchers. But one important and novel finding regarding
these cases is the strong suggestion that the presence of international military
forces—acting as a buffer, deterrent, or enforcer of a cease-fire—is a powerful
factor in the prevention of recurrence of internal armed conflict. The cases ex-
amined, admittedly without systematic controls, in this book affirm the recent
scholarly emphasis on the presence of international peacekeepers in preventing
armed conflicts.
Indeed, the cases of “frozen” conflicts may share an important trait with the
Cold War–era cases of repressive authoritarian regimes that have been able to
prevent recurrence. They both point to the dissuasive power of coercive force
in the calculation of regime opponents about whether to contest state power
through the force of arms. I would not wish to carry this comparison too far,
because there are many instances of repressive, authoritarian regimes that insur-
gents have successfully challenged. However, the socialist cases of nonrecurrence
suggest that the mobilizing, all-encompassing, society-penetrating projection of
state power may have had some success in precluding the emergence of armed
challenge. As suggested above, unless non-Western alternatives gain new promi-
nence, the normative contemporary global environment seems very intolerant
of such stifling and suppressing behavior, rendering this option null—as the em-
pirics of this chapter testify.

Conclusion
What can one conclude from having extended this analysis of the failure of peace
to the full pool of cases of civil wars that seem to have ended? As noted, exclu-
sionary behavior played a central role not only in Liberia and the other two cases
where peacekeeping troops had to be reintroduced, Haiti and East Timor, but
also in several armed conflicts over the pool of civil wars since 1944. However,
this is neither a necessary nor sufficient condition for recurrence. Exclusion can
lead to decades without civil war, just as it is often the trigger for conflict.
When examining the cases of nonrecurrence, however, exclusion assumes
new significance. After 1989 exclusion never produces stability, but it always
leads to the recurrence of internal armed conflict. The cases of nonrecurrence
also reveal the importance of inclusionary political behavior for sustaining peace.
Among the pool of core cases examined here of war recurrence or nonrecur-
rence, 85 percent of instances of inclusionary behavior are associated with sus-
tained peace, in some cases lasting for decades. In only three cases (Azerbaijan,
Mali, and Lebanon) does inclusionary behavior result in the recurrence of inter-
nal armed conflict. This finding also contrasts with a prevalent claim that victory
Making Peace Stick 207

in warfare produces stable outcomes, because it seems that inclusionary behav-

ior, after either victory or negotiated settlements, is key.
To clearly summarize the main findings of this chapter, it is useful to enumer-
ate them:

• Among cases of nonrecurrence, two alternative paths stand out: in-

clusionary behavior, especially with regard to politics and the security
forces; and exclusionary behavior accompanied by the rigid repres-
sion of regime critics and opponents.
• These two paths reflect the context-dependent character of factors
influencing peacebuilding success. The factors that shape success and
failure in peacebuilding change over time, reflecting the prevailing
normative environment. Norms, which both shape and are shaped by
interests of powerful states and nonstate actors, constitute and con-
strain the character of regimes and armed conflict, especially in these
ways: (1) Whereas all-out victories were the norm for war termina-
tion before 1989, negotiated settlements have become the norm since
the Cold War; (2) transitions from peace since 1989 have rarely led to
repressive, authoritarian regimes, but rather to elected regimes that,
if not democratic, are semiauthoritarian or electoral authoritarian
regimes that do not engage in sustained, egregious repression of their
former enemies; and (3) powersharing of at least one form has be-
come far more prevalent since the Cold War, and revolt or resistance
thus may have become less what forms of resistance and rebellion
are viewed as viable or acceptable, what kinds and degrees of exclu-
sion and repression are acceptable, and what weapons and means of
warfare are viable or acceptable.
• Exclusionary political behavior by elites was certainly a path of
prevention, but only before 1989, in the decades of the Cold War.
Since that point, several normative changes have combined to make
it extremely likely that exclusionary behavior will not prevent former
combatants and the parties they claimed to represent from taking up
arms again.
• Throughout the 1944–2007 period examined, inclusionary be-
havior correlates highly with the consolidation of peace. Power-
sharing, which also reflects normative shifts insofar as it has become
much more prevalent since 1988, remains the single most common
manifestation of inclusionary behavior. Since the Cold War, nonre-
currence has correlated highly with the adoption of powersharing
mechanisms. The cases suggest that the most common and influential
208 Examining the Cases

forms of powersharing are of political offices and the integration

of former enemies into the security forces, either the military or
police.
• At the same time, important constraints accompany the claims sug-
gested by the cases examined here. Inclusionary behavior, including
powersharing, can and does fail. Powersharing experiments have
failed before, both during and after cases of war recurrence and of
multiple, simultaneous, or unrelated civil wars.
• Power-dividing models can also constitute inclusionary behavior,
but only if they specifically meet the expectations of the vanquished
parties and the groups they represent to find political representation.
Thus strategies such as open competitive elections may also facilitate
sustained peace. Yet the absence of democracy does not ipso facto
mean exclusionary politics, because salient elites from previously
mobilized guerrilla groups or social groups may find preserved seats
of power in a semiauthoritarian structure. Such illiberal arrangements
may not be desirable, and they may sow the seeds of new, different
forms of protest and even armed conflict. But the evidence shows that
the inclusion of former combatant elites helps prevent their return
to combat.
• The cases ratify the importance of the presence of third-party interna-
tional peacekeeping troops. In a number of cases (especially “frozen”
conflicts), the presence of international troops has facilitated the
maintenance of a negative peace, even as the underlying causes of
warfare fester. This factor may be at work in some cases where power-
sharing correlates with nonrecurrence.
• Despite this finding, a number of cases of nonrecurrent peace con-
solidation have occurred without the sustained presence of interna-
tional combat units serving as peacekeepers: El Salvador, Guatemala,
Tajikistan, Papua New Guinea, and (to date) Nepal. This finding
suggests that where internal factors on the ground favor the consoli-
dation of peace, then international peacekeepers are secondary or
epiphenomenal.

These points are intended to clarify the findings of a complex and lengthy
attempt to extrapolate from the findings of the three main case studies. By exam-
ining the full pool of cases, I hope to provide a more substantive picture of the
causal mechanisms at work than large-N quantitative studies yield. I also seek to
provide something more comprehensive and wide-ranging than detailed case
studies that often show too little awareness of and relation to the broader range
of cases of civil wars and internal armed conflicts.
Making Peace Stick 209

The findings of this study draw on and advance theories of civil war recur-
rence. They offer a critique of the prevalent views that emphasize economic
factors and natural resources in civil war occurrence and recurrence. What are
the full implications of these findings for peacebuilding theory? The following
chapter takes up this question.

Notes
1. I have drawn initially here on the data of Hoddie and Hartzell (2005), excluding
only the Dominican Republic, where cease-fires and demobilization seemed the sole
criterion for qualifying it as “security” powersharing. I then conducted my own catego-
rization of powersharing types for all other cases.
2. Office of the UN High Commissioner for Refugees, “Chronology for Afars in
Djibouti,” RefWorld, www.unhcr.org / refworld / country,,,CHRON,DJI,,469f3882c,0
.html.
3. International Federation of Electoral Systems, “Election Guide,” www.electionguide
.org / election.php?ID=100. Ex-rebel parties gained none of the upper house seats in
2000.
4. World Directory of Minorities, “Hutu and Tutsi of Burundi,” www.faqs.org /
minorities / Sub-Saharan-Africa / Hutu-and-Tutsi-of-Burundi.html.
5. Anthropologists have highlighted the role of tribal rituals for cleansing and rec-
onciliation of ex-fighters; see Honwana (2006).
6. Personal interviews with Farabundo Marti National Liberation Front com-
mander Shafick Handal, 1996, San Salvador; and with Farabundo Marti National Lib-
eration Front political adviser Salvador Samayoa, 1996, San Salvador.
7. See Dupuy and Benningsbo (2008, 2).
8. Note that one case of war recurrence enjoyed peace under powersharing for more
than ten years: Lebanon from 1958 to 1975.
9. On these events, see Atkins and Wilson (1998) and Torrijos (2008).
10. Jordan’s war was essentially King Hussein’s reestablishing control over the mo-
nopoly of coercive force in the territory after tens of thousands of Palestinian refugees
entered the country and carried weapons openly and began exercising some degree of
self-rule. Upon the defeat of the Palestinian forces, no powersharing was contemplated
or adopted with these noncitizens.
11. “National Committee for Investigation of the Truth about the Jeju April 3 Inci-
dent,” report released by the government on December 15, 2003, www.jeju43.go.kr /
english / sub05.html.
12. Ibid.
13. See “Special Act for a ‘Jeju Special Self-Governing Province’ Enacted,” on the
website of the Korean Ministry of Public Administration and Security, www.mopas
.go.kr / gpms / ns / mogaha / . Also see the entry on the Jeju uprising, in The New World
Encyclopedia, www.newworldencyclopedia.org / entry / Jeju_Uprising.
210 Examining the Cases

14. See Human Security Research Group (2005).

15. The case of Georgia is complex. First, I have sought to retain it as it comes close
to refuting my argument as a case of nonexclusion that involved recurrence. Note that
the recurrence scored here is the 2004 flare-up of fighting in South Ossetia (1992–94,
2004), in which only twenty-seven persons died, rather than the war that took place in
2008 involving domestic and external Ossetian nationalist forces versus the Georgian
armed forces with the virtually immediate incursion and occupation of all of South
Ossetia and Abkhazia, as well as much of Georgia itself, by Russian troops eager to abet
independence for these separatist territories. Although I could have used the civil war
in Abkhazia as grounds to omit Georgia due to complex or simultaneous conflicts, the
nonrecurrence of Abkhazia before this date and the core role played by Ossetia in the
2008 recurrence permitted me to count it as a recurrence.
Second, it is one of two cases of civil war recurrence (the other is China in Tibet,
1950–51, 1956–59) that I abstain from classifying as exhibiting either inclusionary
or exclusionary behavior. In both cases the state had reached a deal to grant some
de facto self-governance and to eventually specify a more concrete powersharing ar-
rangement. In both cases the recognized state army initiated efforts to solidify support
for integration that then led to protests and armed resistance. Its conduct did violate
some Ossetians’ (and in the case of China, some Tibetans’) expectations but did not
explicitly violate an agreement, and was not inclusionary.
16. The deployment of a new Central African Multinational Force of African troops
under the Economic and Monetary Community of Central Africa helped provide sta-
bility after 2003, despite occasional skirmishes that increased in intensity in 2006.
Part III
Implications for Theory and Practice
This page intentionally left blank
8
Conclusions for Theory
Legitimacy-Focused Peacebuilding

HAVING EXAMINED THE POOL of the clearest cases of both successful and failed
postwar peacebuilding, what can one say about the implications of these findings
for the theory of civil wars and their recurrence, and for theories of postconflict
peacebuilding? This chapter revisits the initial analysis of what scholars have
found about the causes of civil wars and of consolidated postconflict peace. It
first recaps the central findings of the five previous, empirical chapters (part II)
of the book. Here the focus is on synthesizing the findings of the case studies,
tempered by an examination of the range of cases of war recurrence and non-
recurrence that have informed the study. I then turn to theory, focusing on the
principal overarching approaches to peacebuilding that dominate contempo-
rary debate on postwar peacebuilding. The next section sets forth the concept of
“legitimacy-focused peacebuilding” and its potential to address shortcomings of
the four extant approaches. Then, informed by the perils of extrapolating policy
implications from the theoretical findings, the subsequent chapter turns to what,
if anything, these findings mean for the ability of internal and external actors to
shape peacebuilding outcomes.

The Main Findings of the Book

The first important implication of this book is that elite political behavior—
more directly than economic risk factors, state capacity, natural resource depen-
dence, or democracy per se—lies at the heart of the success of efforts to con-
solidate peace after civil wars. Political factors, shaped by national elites and to a
lesser extent external diplomacy and leverage, condition the chances of whether
peace can be consolidated or not. These findings underscore the focus on “spoil-
ers” and the need to understand and manage them (Stedman 1997; Zahar 2003).
Such factors figure in much of the literature on postconflict reconstruction and
peacebuilding, but economic factors have been emphasized more recently, espe-
cially in the quantitative literature that has achieved a certain stature in political
214 Implications for Theory and Practice

science. Moreover, the extensive literature on peacekeeping does not adequately

address these challenges (Durch 1996, 2006; Howard 2008).
Specifically, inclusion and exclusion are crucial factors in many cases of war
recurrence, and should be the first place to look for evidence of potential rever-
sion to internal armed conflict in postwar societies. This exclusion takes the form
(as defined in this book) of perceived or actual deprivation of an opportunity to
participate in state administration in ways that defy the expectations of former
warring parties or the social groups associated with them. Exclusion is prevalent
in many cases of recurrence. However, it is important to acknowledge the pres-
ence of several other pathways and variables that are associated with failure. I am
calling special attention to the role of exclusion and inclusion but am not saying
these are the sole causal variables in war recurrence. Recurrence cases point to
other factors in explaining outcomes such as meddlesome or invasive neighbors;
dissatisfaction in “frozen” conflicts with the status quo after some time; and the
resurgence of rebels organizing after apparent defeat and without significant
changes in government policy. Nevertheless, in many of these cases (recall Ti-
bet, Nicaragua, and Peru), exclusion also played a significant role. Even cases
that were omitted from the analysis as not enjoying sufficient levels of “interwar”
peace, like Cambodia and Senegal, exhibited exclusionary behavior.
Moreover, in cases where many of the postulated factors prevalent in current
theory—natural resource dependency, poverty, and economic malaise—were
present, these factors were not considered key factors in the recurrence of armed
conflict. Recall the cases of Liberia, the Central African Republic, and Sudan
(natural resource dependency) and the poor economies of Nicaragua and Mali.
Conversely, here exclusionary behavior proved the most helpful for understand-
ing these recurrent wars. And even where exclusion was not the predominant
factor, neither were these more commonly touted variables. Consider again the
cases of Nicaragua, Burundi, and Rwanda, where such factors cannot distinguish
these recurrences from nonrecurrence in cases like Djibouti, Laos, Nepal, Sierra
Leone, and Paraguay. These cases instead point to the importance of other sorts
of factors—certainly to political exclusion, but also to a regional level of analysis
(both for meddlesome neighbors and for the patterns of violence) and to the pre-
vailing global and regional norms about state repression. Such factors have not
received the same level of attention as poverty, inequality, and natural resources.
This is not to say that other hypothesized variables are unimportant. Many
cases examined here—Liberia, Haiti, the Central African Republic, Sudan, Mali,
Nicaragua, East Timor, Burundi, and Zimbabwe—are poor countries charac-
terized by very weak formal institutions, including state institutions. Despite
the problems associated with conceptualizing and measuring state “fragility” or
“weakness,” countries that have a low per capita income or whose formal institu-
tions are irrelevant, penetrable, or unreliable appear to be more prone to recur-
Conclusions for Theory 215

rent armed conflict. The robust quantitative correlation between poverty and
the onset of civil war is not disconfirmed in these cases.
The causal import of natural resource dependency is less persuasive based
on the universe of cases, although this factor seems quite important in a subset
of cases of civil war recurrences and especially in a subset of countries that have
experienced long-standing, nonrecurrent internal warfare. For instance, specific
export commodities play a salient role in the persistence of civil wars in Afghani-
stan, Colombia, and the Democratic Republic of the Congo. Conversely, such
export dependency is not apparent in other long-lasting wars in Burma, Pakistan,
Sri Lanka, and the Philippines. Liberia shows how the availability of high-rent
exports under a weakly institutionalized state facilitated the organization of in-
surgency and defections from government military forces. Yet even in this case,
only in combination with exclusionary behavior did natural resources or low ca-
pacity contribute to war recurrence.
Admittedly, the definition of “peace” used here is a narrow one focused on
the nonrecurrence of internal armed conflict. This definition omits the dignity
of life of a society, and the presence or absence of justice for its individuals and
social groups. It falls well short of the preferred content of peacebuilding that
has emerged in the scholarly literature and even policy statements by interna-
tional organizations working on peace and development. Setting aside the lack of
shared conception and measurement of “positive peace,” some substantive rea-
sons warrant a study of negative peace. It has been the main operational indicator
of failure for international peacekeeping and peacebuilding efforts, and the main
yardstick for the success of the UN Peacebuilding Commission. Without nega-
tive peace, positive peace is only a fleeting hope. And discerning how to achieve
negative peace is a necessary first step in determining the steps to advance posi-
tive peace. Surprisingly little consensus exists about the causes of consolidated
peace in international scholarship.
Turning to the cases of nonrecurrence, this study reinforces, to an unexpected
degree, the correlation between negative peace and inclusionary behavior, prin-
cipally by the state toward its former enemies. Powersharing is the most preva-
lent form of inclusionary behavior. It takes many forms beyond allocation of the
presidency or the cabinet to former warring parties, although such guaranteed
quotas remain very prevalent. Especially important among alternative types of
powersharing is the provision of security posts in the military or police forces,
which provides crucial guarantees to alleviate fears of physical violence against
demobilized fighters. This issue constituted a trigger for war recurrence in Li-
beria, the Central African Republic, and East Timor, and to a lesser extent in
Zimbabwe, Rwanda, and Haiti. Although powersharing is neither necessary nor
sufficient to prevent war recurrence, its presence corresponds greatly to peace
consolidation, whereas its abrogation correlates with failure.
216 Implications for Theory and Practice

Yet powersharing is not the only form of inclusionary conduct that helps
peace to “stick.” The vanquished parties’ expectations of being included in po-
litical life can be met via the realistic hope of electoral representation, whether at
the local, provincial, national, or supranational level. To the extent that electoral
results are not rendered meaningless through either vacating the authority of
public office or an outright reversal of voting outcomes, then such democratic
processes constitute inclusionary politics. Yet democratic political forms can
also coexist with exclusionary political behavior, as we saw in Liberia, Rwanda,
Burundi, the Central African Republic, East Timor, and Haiti. These cases dem-
onstrate precisely the limits of liberal peacebuilding models, to which I return
just below.
What are the implications of this book for the civil war literature, with its
quantitative tendencies and its focus on onset rather than the prevention of re-
currence? This study strongly suggests that the substantive significance of exclu-
sionary or inclusionary behavior exceeds that of poverty, resource dependency,
or state capacity in war recurrence. This research calls into question the strong
focus by policymaking entities and scholars on fighting poverty as the most ur-
gent means to advance peace, and it raises questions about our excessive heed to
quantitative methods rather than drawing on some more contextualized analysis.
Poor countries are at a greater risk of civil war and its recurrence, but inclusionary
conduct seems to have much more analytic power in accounting for recurrence.
One illustration of the limits of quantitative methods is what the cases revealed
about the salience of religious fractionalization. Although this variable emerged
as highly statistically significant for the recurrence of civil wars, it figured in only
a few of the clearest cases of recurrence, represented by the fifteen core cases.
Ethnic fractionalization also figured in several, but not in several others. Those
findings are difficult to interpret, but they suggest that the way such social divi-
sions are drawn on by elites or masses is complex. The interplay of these variables
warrants more consideration.
The picture here of war recurrence is one where risk factors elevate the likeli-
hood of exclusionary behavior, often by elected governments reneging on the
perceived commitments to losing factions or groups. It ratifies the findings of
those who find that powersharing can play a positive role for peace, specifically
in the sustainability of postwar peace. It suggests that elite agency is the locus for
examining the chances of recurrence of internal armed conflict after civil war.
The findings call into question a prevailing view that victory in warfare facili-
tates postwar stability, because how postwar polities—shaped by their own po-
litical contests, civil society organizations, and international actors—respond
to power after warfare is more commonly the crucial factor. Including one’s en-
emies, rather than killing them, is the most frequent path to peace.
But it is also vital to reconsider the nature of scholarship on civil wars and
their occurrence and recurrence in light of the historically contextual nature of
Conclusions for Theory 217

the phenomenon. Constructivists contend that shared, intersubjective under-

standings among subjects constrain and even constitute interests, rendering
norms important causal factors in shaping outcomes of those subjects’ interac-
tion—in this case, collective notions of what sorts of regime structures and be-
havior are relevant. In the past two decades, norms of elections as a mechanism
for validating political power have spread to the vast majority of states. Even
some decades-old holdout nondemocratic regimes like Cuba and Saudi Arabia
have adopted elements of electoral ratification.
Because wars—and their termination—often constitute important windows
of opportunity for new actors, ideas, and structures to be adopted, it is more dif-
ficult for postwar regimes to embrace norms that have been eclipsed by emerging
notions associated with or prevalent under new historical periods (Ikenberry
2000; Call 2002). Given the post–Cold War environment, those peoples who
are excluded or repressed are more likely to resist such behavior and to call on
international norms, incentives, and allies. It is difficult to imagine postwar coun-
tries that experience a regime transition in the twenty-first century adopting
nonelective authoritarian regimes that engage in exclusionary behavior. Where
such regimes emerge victorious, then they may retain their characteristics, but
these are the exception.
Norms are neither created nor advanced wholly independent of power. Given
the generally increased power exercised by major donors, Western-influenced
intergovernmental organizations like the United Nations and the World Bank,
and regional organizations, postwar states are more susceptible to these pres-
sures and transnational social expectations. Post–Cold War expectations facing
postconflict countries include the following: negotiations amid warfare, intense
international pressure for a settlement, domestic awareness of the possibilities of
international assistance to postconflict societies, widespread domestic awareness
of a strengthened human rights regime that focuses on conflict countries, and
pressure to suppress terrorists that may contradict the impetus to include for-
mer enemies. These practices-cum-expectations modify what postwar politico-
military elites see as possible, desirable, and acceptable.
The crucial element of this normative environment is not that it ensures con-
formity with these expectations but that the norms are sufficiently prevalent to
ensure that former parties to warfare will (perhaps selectively in their favor) refer
to such norms, advocate for them, and draw on them to mobilize armed opposi-
tion in the face of flagrant violation of those norms in an exclusionary fashion
by the postwar state. As such, the record of postwar peacebuilding shown here
reinforces constructivism’s call for understanding the social construction of war
termination and political regimes, and how these shape peace. Heuristically, the
cases examined here demonstrate the limits of quantitative methods that com-
pare all wars across decades without regard for social context. They also suggest
the limits of case studies that ignore the causal impact of transnational normative
218 Implications for Theory and Practice

environment and how norms constrain the choices available to individual or col-
lective agents within a particular country or specific war. In this study I have not
sought to specify which norms are at work in these cases or to formalize their
role in outcomes—work that has begun elsewhere. Nor do I analyze or seek to
anticipate what sort of normative changes are afoot in the international system. A
number of potential trends regarding human rights, transnational activism, truth
and justice, and peace itself suggest possible future norms, but these are beyond
the scope of the study.

Rethinking the Aims and Approaches of Peacebuilding

Given the importance of ideas in both constraining and empowering institutions
and choices by postwar decision makers, what do the findings of this study mean
for the conceptual frameworks for postconflict peacebuilding? How should the-
orists rethink prevailing approaches to peacebuilding and state building? This
chapter begins by reexamining the liberal peacebuilding model, explicitly explor-
ing alternatives in light of the findings of the research presented in the previous
chapters.

Abandoning the Liberal End State?

The evolution of scholarship and the empirical experiences of successful and
unsuccessful peace consolidation presented above call into question the pursuit
of “liberal peacebuilding.” They do not necessarily call into question the cor-
relations between democracy and internal peace or between economic growth /
prosperity and internal peace. Crucially, it is the process of political and economic
change, more than the attained existence of liberal politics and economies, that
jeopardizes internal stability and peace. If one accepts this claim, then it may bol-
ster the argument to salvage liberal peace. Indeed, Richmond and Franks (2009)
lament that this is precisely the case with Paris (2004) and other reformers of
liberalism, who seek to save liberalism through their critiques.
However, the predominance of the liberal peacebuilding model induces in-
ternal and external actors to settle into pursuing liberalizing policies without ac-
knowledging their contradictions and dangers. Richmond and Franks (2009, 185)
note that the liberal model is the only framework upon which donors and interna-
tional organizations can agree. Consequently, despite the reservations many dip-
lomats feel about neoliberal economic prescriptions, and despite the skepticism
many development specialists hold about elections, these liberal hallmarks by
default attract support as a minimal consensus, especially in the absence of other
frames. In addition, both national and international officials risk accusations of
being “authoritarian” if they depart from democratizing conventional wisdom.
Conclusions for Theory 219

Therefore, it seems sensible to seek to modify the conceptual end state—in-

cluding the language of “democracy” and “liberalism”—that dominates the in-
dustry (Ponzio 2010). Transforming the liberal end state is necessary to modify
the mindsets, frameworks, assumptions, policy models, and operational guide-
lines that prevail among donors and external peacebuilders. Without refocusing
the strategic goal of peacebuilding, liberal assumptions and formulas will con-
tinue to squeeze out more inclusionary or participatory approaches.
But what feasible and desirable alternative frameworks are there to liberal
peacebuilding? Even recent articulate and thoughtful critics have difficulty com-
ing up with viable alternatives that can attract the support of scholars, much less
the more constrained policymakers and practitioners of international organiza-
tions (Newman, Paris, and Richmond 2009).

Authoritarianism—and Repression
Let us quickly dispense with the most obvious alternative to liberal peacebuild-
ing: authoritarian or nondemocratic regimes or exclusionary state structures.1
Stepping away from “democratization” as an end or means of state legitimation
would seem to require that international actors be open to some authoritarian
strains of governance. Paris (2004, 180) explicitly considers this alternative, and
he rejects authoritarianism because “it would rely on the permanent, forcible
suppression of contestation” that would engender resistance. In fact, we saw in
chapter 6 that many authoritarian states offer some of the most durable instances
of nonrecurrence. The use of demonstrably violent repression against politi-
cal dissidents can both eliminate armed opponents and remove the incentive
to rebel, in some cases for decades. Authoritarianism is unappealing because it
embodies a lack of even the possibility of fulfilling either individual (liberal) or
group (communitarian) aspirations.
Embracing authoritarian governance is not appealing for another reason: It
requires repression. It suppresses contestation and breeds resentment and alien-
ation. Obviously, these sentiments erode legitimacy, but they may not lead to
armed rebellion in the short or medium run. Many empirical experiences of
postwar authoritarianism—consider Vietnam and Syria—involve the repres-
sion and exclusion of wartime foes. Only through repression did these regimes
survive and avoid war recurrence. It is both unrealistic and objectionable on
humanitarian grounds to expect international actors to embrace severe repres-
sion as the route to avoiding renewed warfare. It is one thing to downplay the
authoritarian traits of elected postconflict regimes, as donors have done in places
like Cambodia and Guatemala. It is quite another to openly embrace severe re-
pression in authoritarian regimes as a path to stability. Such policies fly in the
face of the long-standing normative commitments of donors and international
220 Implications for Theory and Practice

organizations, and would open Western states adopting them to scathing inter-
nal criticism.

Revisiting Republican Peacebuilding

Illiberal peacebuilding, however, is not necessarily authoritarian. Michael Bar-
nett (2006) proposes a “republican” alternative to liberal peacebuilding. His ap-
proach comes closer to providing the scaffolding for an inclusionary approach to
peacebuilding and thus offers valuable remedies for liberalism’s neglect of issues
of exclusion and legitimacy. For instance, Barnett (2006, 94) argues that modern
republicanism (represented by Niccolò Machiavelli and James Madison, rather
than their republican predecessor Aristotle) is concerned largely with stability
in the face of contentious factions. Drawing on republican theory, Barnett sug-
gests three core elements of republican postconflict peacebuilding: deliberation,
representation, and constitutionalism.
On deliberation, Barnett (2006, 94) praises Machiavelli’s view of “the domes-
ticating influence of public deliberation; by compelling individuals to speak in
the language of community they might develop a greater sense of patriotism, that
is, a love of country.” In line with the need for multivalent forms of participation
that I advanced above, he suggests that representation go beyond democratic
elections to include alternative forms of participation (p. 102). Such participa-
tion may include consultations, but it rests principally with representatives that
eschew direct participation by the masses. Finally, Barnett calls for constitutional
constraints that divide power and regulate the behavior of representatives. He
contends that “all constitutional systems have (1) an agreement over the rules of
the game and the underlying principles that are to maintain the political order;
(2) rules and institutions that limit the exercise of power; and (3) rules that are
relatively difficult to amend” (p. 105). In this Barnett falls not far from liberal
theory and its concern with the social contract.
Yet the assumptions inherent in republican thought are of questionable ap-
plication to poor, weak states. Republican theory seems to assume that individu-
als and social groups already enjoy security, presumably within a single, statelike
entity that provides the locus for debate and exchange of views. Patten (1996,
26) sets forth “the central republican problem, which is to identify the condi-
tions under which a society can maintain the institutions of its freedom, despite
[a] tendency to corruption.” Republican theory emerged from a Western context
that took the state system for granted and assumed certain state functions and
attendant capacities that exhibited sustained security from organized violence.
Patten describes republicanism’s concern with reinstating “the proper value of
community and public service.” Barnett’s call for deliberation echoes this con-
cern. But how can one choose to dedicate one’s time to community or public
Conclusions for Theory 221

service without some sense that one is safe and that such “service” will not lead
to personal danger, group peril, or enhanced social conflict?
The republican worldview emphasizes the dangers of factions, but its solu-
tions seem blind to the ever-present possibility of armed violence and a perilous
aggravation of social hatreds. Thus Barnett’s call for constitutionalism rests on an
assumption of highly formalized institutions that constrain actors. It fallaciously
presumes, rather than helps constitute, the requisite consensus on the rules of
the game by society’s main groups. Madison and Hamilton emphasize the need
for checks and balances (mechanisms that can certainly enhance internal state le-
gitimacy and may reflect power dividing) to curb authoritarian tendencies. Nev-
ertheless, the assumed milieu does not correspond to a fragile cease-fire amid
intense identity conflict. Republican theory fails to recognize the intensity of
group affiliation and loyalty. As Patten (1996, 26) says, “By assuming that there is
some single good—such as civic friendship or self-government—that is a good
for everyone, republicans seem insensitive to difference.”
In this sense Barnett’s proposals for extended debates to reach shared under-
standings and views seem somewhat archaic, oddly reminiscent of the security
enjoyed by eighteenth-century diplomacy. They appear incognizant of the sim-
mering, hate-laced social relations among so many postwar societies. The no-
tion of convening a series of conversations whereby shared understandings may
emerge is attractive, and one that critical theorists like Richmond and Franks
(2009, 212) also seem to favor. However, without some attention to security—
to who will ensure safety against roving and (almost always) still-armed fac-
tions—such conversations are likely to lead to assassinations, threats, circum-
scribed debate, or mass violence in the extreme.

Revisiting State Building as Peacebuilding

The state-building agenda has much to offer the study and practice of postwar
peacebuilding. It remedies some long-standing deficiencies in liberal theories
and donor practices in postconflict societies. Nevertheless, scholars have identi-
fied at least three important challenges or shortcomings vis-à-vis state building.
First, the empirical testing of the relationship between state strength and civil
war recurrence, or even a sustained quality of peace, is difficult and in its early
stages. The crux lies in the conceptualization, definition, and operationalization
of state institutions and of peace, and it is difficult to identify measures enjoying
widespread agreement.
Second, the approach yields policies that are too blindly supportive of states
with too little regard for whether they exclude, abuse, repress, or cheat their pop-
ulations. For instance the recent emphasis by the Development Assistance Com-
mittee (DAC) of the Organization for Economic Cooperation and Development
222 Implications for Theory and Practice

on channeling resources through the state is a welcome antidote to the inter-

national tendency to channel money through external organizations.2 Without
the means to provide for and distribute resources, states have little means by
which to garner legitimacy, and political representation means little. However,
the DAC’s Paris Declaration on Aid Effectiveness treats all states as if they are
equally legitimate and without serious problems. It fails to recognize that there
are sometimes good reasons not to channel international aid resources through
the state, because they will only line the pockets of a few corrupt individuals,
enabling unrepresentative leadership and failing to advance development.
Third and related, the state-building agenda remains too focused on capacity.
Despite paying lip service to state legitimacy, most of the works on state building
in postwar societies focus on enhancing institutional capacity. Scholars like Fu-
kuyama, Krasner, Zartman, Ghani and Lockhart, and Chesterman address how
state structures and institutions and processes are important to institutionalizing
bargains designed to include or exclude specific social groups and armed actors.
However, this nod to legitimacy is overshadowed by a privileging of advancing
the professionalism, service delivery, and efficiency of state institutions. Ghani
and Lockhart (2008, 124), for instance, focus on the need to foster ten core func-
tions of the state.3 Most of these functions are technical in nature, such as admin-
istrative control, public finance, infrastructure provision, public borrowing, and
managing public assets: “We maintain that consensus on a specific set of func-
tions provides the vital agreement for international agreement on . . . creative
responses to the challenge of state-building.”
Similarly, the World Development Report 2011 (World Bank 2011, 270) calls for
“investing in the provision of citizen security, justice and jobs” as the key to pre-
venting repeated cycles of violence. The report’s welcome emphasis on a broader
understanding of postconflict violence emphasizes judicious enhancements of
national and international capacities. It calls for fostering “inclusive-enough” co-
alitions that can provide buy-in, especially from leaders and factions that could
prove to be spoilers of a peace process (World Bank 2011, 139; Stedman 1997;
Gates and Strøm 2007; Nilsson 2008). However, it has less to say about how
external actors can build inclusive coalitions. Without attention to legitimacy,
capacity building can deepen illegitimacy and the likelihood of war by leaving a
state better able to mistreat, exclude, or swindle its population. One reason for
this focus on capacity is that it is both more difficult and more politically peril-
ous for external actors to try to build state legitimacy than to build capacity. The
former implicitly calls into question the authority of those who control the state,
whereas the latter strengthens this authority.
Scholars of state building have especially neglected horizontal legitimacy in
postconflict societies. Sustained peace requires more than legitimacy via either
well-articulated long-term state–society relations or broad calls for popular par-
Conclusions for Theory 223

ticipation or simplistic prescriptions of democracy or elections. Adequate ad-

dress of state legitimacy also includes the representational element of the armed
protagonists of civil war and especially of the social groups that they purport to
represent. Horizontal legitimacy should lie at the center of analysis of the role of
the state in postconflict peacebuilding. In addition, the state-building focus on
institutionalization and professionalization of the organs of the state downplays
the role of civil society. The crucial vectors of state–society relations and trust at
the local level remain deemphasized.

Revisiting Critical Peacebuilding Theory

Recall that critics of the liberal peacebuilding approach score the international
pressure for neoliberal economic policies and liberal democratic political mod-
els. They also reject the emphasis of state-building theorists on the state itself
and “governmentality.” As we saw in chapter 1, Richmond and Franks (2009, 6),
criticize the prevailing convergence of “balance of power, hegemony, institution-
alism and constitutionalism, and civil society” in an externally imposed, West-
ern self-serving, and one-size-fits-all approach to postconflict societies (see also
Autesserre 2010). Although cogent alternative approaches are not widely shared,
an emphasis on broader and deeper participation on the part of the population
in postconflict societies is a central element of this vision.
The contribution of this approach is that it leads us away from single mod-
els of any sort of approach to postconflict societies. Critical theorists focus on
context-specific analysis and policymaking. They also expose the interests that
underlie external actors. Peacebuilding projects represent an exercise of power
by powerful states and serve foreign economic interests. Being mindful of power
in these interactions permits a clearer understanding of how they may contradict
the interests of the population at large and of specific groups within postwar
countries, along with the interests of their elites. This approach also usefully
highlights the need for long-term legitimation.
The shortcoming of these critical theorists is their lack of recognition of the
need for minimum interests of potential spoilers to be addressed. This study
finds that state institutions, and the participation of elites in them, are vital to the
maintenance of stability. Neither justice nor affluence is possible without stabil-
ity, which is a chief concern of poor and marginalized populations. “Governance”
is crucial to making peace stick, even though that governance serves the interests
of powerful external actors and is often implemented in ways that harm the eco-
nomic well-being and political participation of the broader populace.
Stepping back to examine these four principal theoretical approaches to
postwar peacebuilding, one can draw some generalizations. Each of these theo-
retical approaches to peacebuilding offers important insights and contributions
224 Implications for Theory and Practice

to the phenomenon. As shown in chapter 1, they often respond to deficien-

cies in previously articulated approaches. These contributions are genuine and
substantial.
Nevertheless, each of the main approaches fails to adequately recognize the
causal factors identified in this study. Horizontal legitimacy, expressed most di-
rectly through some inclusion of former combatants and their associated popula-
tions in state offices, is the touchstone of peace consolidation. This book’s find-
ings support the argument that the dominant approaches to peacebuilding do
not place issues of political exclusion and inclusion sufficiently at the center of
theory. Instead we have liberal democracy, which overemphasizes vertical legiti-
macy and underemphasizes (or explicitly eschews) horizontal legitimacy. One
sees a republican peacebuilding that underrecognizes intense ethnic identities
and loyalties and that places too much faith in the possibility of a dialogue that
will not suffer attacks or lead to renewed violence. State-building approaches
emphasize the functions and services of the state and thus implicitly focus exces-
sively on its capacity rather than its legitimacy. And critical peacebuilding theo-
rists eschew elite politics and deal making, pointing toward a greater involvement
of civil society as the ticket to consolidating peace. The main contributions and
shortcomings of each of these theoretical approaches are identified in table 8.1,
which synthesizes the preceding analysis.
Given the shortcomings of these principal theoretical approaches to peace-
building, can one formulate an alternative? In the next subsection I turn to this
question.

Legitimacy-Focused Peacebuilding
Two central findings of this book are that exclusion is highly correlated with civil
war recurrence, and that inclusion is very highly correlated with the absence of
recurrence, or “negative” peace. This finding suggests that national elites and
international actors can foster sustained peace if they seek to meet the parties’
expectations about their role in governance even after initial elections and into
the future. Various sorts of inclusionary behavior can meet this expectation.
Powersharing in its various guises may do so, but so also may access to power
by holding mayorships or a few posts in cabinet ministries or on local councils
as a result of elections rather than fixed arrangements. In short, inclusion can
include powersharing, power dividing, or simply access to compete in elections
if poor results do not violate expectations concerning representation in the state.
Once spoilers have been mollified in the early stages of war termination, then
medium-term stability requires managing those spoilers by including some of
them in accord with their expectations.
However, minimal inclusion in government represents a highly mechanistic
view of participation and the state. Although the cases point to inclusion as a
Conclusions for Theory 225

Table 8.1 Key Contributions and Shortcomings of the Main Theoretical Approaches to
Peacebuilding
Approach Contributions Shortcomings
Liberal Focus on political nature of Overemphasis on vertical legitimacy,
peacebuilding peacebuilding especially democracy and
Focus on rights as basis for stable elections, versus horizontal
polity legitimacy
Too little concern with quality of
elected governments
Overly individualistic, insufficient
focus on group interests and
insecurities
Republican Focus on finding common ground Presumes single common “good”
peacebuilding among factions Too blind to group identities and
Focus on tolerance loyalties
Focus on intergroup dialogue Neglects possibility of “virtuous”
inclusionary policies in the
absence of dialogue or interaction
State building as Focus on state arena Overemphasis on state capacity over
peacebuilding Offers framework of state legitimacy legitimacy
for addressing inclusion Overemphasis on state rather than
legitimate authority at substate
and suprastate levels
Critical Emphasis on long-term sustainable Undervalues elite politics and deal
peacebuilding legitimacy making
theorists Eschews one-size-fits-all formulae Neglects threat posed by spoilers
Focus on civil society participation Rejects statism, conflating it with
liberalism

crucial element, the historical context and linked factors in these cases suggest
that a broader understanding of how new postwar governments relate to the main
groups of the population is necessary. If stability requires inclusion of the main
social factions in the state, it is because state institutions require internal legiti-
macy. To be sustained, that inclusion must be understood as part of state–society
relations that are minimally institutionalized and accepted—that is, legitimate.
State legitimacy, especially horizontal legitimacy, offers the conceptual mecha-
nism for a more holistic understanding of the relationship for which inclusion
serves as the most visible manifestation. Consequently, a concept of peacebuild-
ing that encompasses not just inclusion of elites but also focuses more broadly
on legitimation is appropriate. This book’s findings suggest a greater focus on
legitimacy in peacebuilding.
I have here used the cases examined in a heuristic manner to formulate an ap-
proach called “legitimacy-focused peacebuilding.” This term constitutes neither
226 Implications for Theory and Practice

a well-developed theory nor a wholly original approach to peacebuilding. I draw

freely on the positive features identified in existing approaches. And this ap-
proach is not as theoretically developed or philosophically grounded as the ap-
proaches described above.
Legitimacy-focused peacebuilding sees the consolidation of peace as depen-
dent foremost on a state whose main social groups see the state as offering them
acceptable levels of representation and participation. The state and its institu-
tions are important, but the state’s relationship to society, including its com-
position, occupies a privileged position over capacity in this view. Again, state
legitimacy and capacity are interrelated insofar as the capacity of the state to
deliver basic services influences legitimacy—but service delivery is not the sole
or even the primary criterion for overall legitimacy. State legitimation reflects
tumultuous historical processes. Recalling the sources of internal legitimacy, the
“embedded” element of the legitimacy of a particular state is extremely difficult
to modify in the space of a few years. The delivery of specific goods and services
reflects state capacity and is important for performance legitimacy. However, the
cases analyzed above demonstrate that performance legitimacy per se is not as
important as the perception that the state and its leaders are acting in a way that
will not exclude, repress, or silence minority (or majority) ethnic, religious, or
class-based groups, especially those involved in the prior armed conflict.
Process legitimacy is the principal form of legitimacy that is relevant for an
inclusionary pattern of conduct by the state. Of course, exclusionary behavior
can be reflected in the delivery of goods and services, especially where the state
controls a high portion of income via natural resources. Governments often fla-
grantly exclude certain social or political groups in providing access to revenues
that are channeled through the state. In most such circumstances, however, those
social groups are also excluded from political processes.
Crucially, elections do not alone constitute process legitimacy internally.
They may provide a certain degree of legitimacy for the population and even un-
derrepresented groups that “lose” in the process. However, if social groups, even
after elections, perceive that their access to power is inordinately curtailed—via
exclusion from the military or police forces, from appointments to cabinet posi-
tions, from alternative levels of authority, or from territorial autonomy—then
political processes have not conferred sufficient legitimacy. Indeed all three cases
of failed peace processes after the close of successful peacekeeping operations
(Liberia, Haiti, and East Timor) turned on the exclusionary practices of govern-
ments empowered by internationally supported and celebrated elections.
Two crucial challenges to this notion of inclusion merit attention. First, what
of the critique that elite-brokered inclusionary deals, especially where power-
sharing ensues with its built-in marginalization of third parties and the masses,
can breed resentment and new exclusion that sparks new armed conflict? Do not
Conclusions for Theory 227

these elite-brokered inclusionary arrangements inherently exclude others who

may revert to combat—resulting in a “locked-out masses” problem? Second,
are there some factions or leaders whose inclusion would undermine internal and
external legitimacy? Might the inclusion of such unsavory actors endanger peace
if they enjoy or acquire a reputation as illegitimate, prompting a rejection by
some factions or social groups of the entire political process and even the post-
war state? Here one can think of combatant commanders accused of war crimes,
mass atrocities, or terrorism—producing a backlash against elite-brokered deals
that “embrace the reviled.” Just as the Paris Declaration’s imperative to channel
aid through the state seems blind to the problem of corrupt, abusive, or illegiti-
mate states, is the imperative to include criminal or brutal individuals or factions
in the postwar polity equally blind?
These two problems are alleviated by emergent norms governing postwar
political conduct. Here I utilize the same principle identified above about his-
torically contingent options and constraints that reflect new normative environ-
ments. Just as the practices of mediated settlements grew once the Cold War
ended, so too have new norms emerged regarding the “locked-out masses” prob-
lem and the “embracing-the-reviled” problem.
First, it is increasingly difficult for internal elites and international actors to
foster a negotiated agreement in which some broader popular participation is not
involved. Certainly elite-brokered pacts remain a prevalent and positive tool of
conflict resolution and mediation processes. However, given the spread of global
civil society, transnational demands for greater popular voices in state-level gov-
ernance, and increased pressures by bilateral powers and intergovernmental orga-
nizations, virtually all such mediation processes encounter demands for broader
public engagement. Civil society organizations seek a seat at negotiating tables,
and they have increasingly found such a seat since the innovative involvement of
civil society in Guatemala’s transition in the mid-1990s. External development
actors at least pay lip service to taking transitional political processes beyond
narrow elites to include some level of popular input and participation. Critical
theorists’ critique of the absence of local-level or mass voices and participation in
peacebuilding remains well founded, but elites can no longer get away with cut-
ting static deals among themselves without a recognition of the need to involve
the mass level of various constituencies.
Consequently, legitimacy-focused peacebuilding conceptually encompasses
long-term legitimacy that reflects broader participation. Wider horizontal legiti-
macy includes the perceptions of the social groups mobilized in warfare. The in-
clusion of the elites in political bargains may not be sustainable if it comes at the
cost of the exclusion of the broader population in the longer term. Legitimacy-
focused peacebuilding must ensure that the interests—and therefore the voices
and participation—of the masses of the population also have a stake in postwar
228 Implications for Theory and Practice

peacebuilding. The march of thinking has expanded from including armed fac-
tions to including economic elite interests to including civil society and the
broader population in postwar governance.
As for the second problem of “embracing the reviled,” regimes governing in-
ternational human rights and transnational crime make it somewhat more diffi-
cult for war criminals and alleged perpetrators of atrocities to become part of the
political order in internationally supported transitions. The universal jurisdic-
tion of the International Criminal Court and the increased codification of major
violations of human rights help constitute a much stronger and broader web for
catching the most egregious transgressors. Of course this regime and the trans-
national crime regime remain highly selective, inconsistent, and politicized in
practice. For example, the persistence in the Afghan Cabinet of warlords accused
of war crimes, such as those involved in transnational crime there and elsewhere,
demonstrates the limits of this norm. As seen in chapter 9, candidate selection is
one window whereby international actors can influence the inclusion of political
leaders who may undermine the overall legitimacy of a transition or a postwar
state; however, such action may well be seen as outside interference and prompt
political backlash.
Genuine trade-offs sometimes arise where inclusionary behavior may result
in the accession to power of a faction or individual whose presence undermines
state legitimacy. Under such circumstances, external and internal actors may
perforce accept the temporary inclusion of these unsavory and delegitimizing
leaders with the hope that criminal sanctions or popular sentiment may eventu-
ally politically incapacitate them. Human rights and criminal justice mechanisms
that have shown their reach and tenacity well after transitions have involved se-
nior state offices for the accused, as shown by the cases of Liberia’s Charles Tay-
lor, Serbia’s Slobodan Miloševic, and Kosovo’s Ramush Haradinaj. Whereas one
segment of the population usually intensely supports such individuals, another
reviles them, which leads to a conundrum for horizontal legitimacy. Even with
strengthened mechanisms of global criminal justice, however, they generally take
time and do not adequately or consistently address the problem of embracing
reviled leaders.
Here I present legitimacy-focused peacebuilding as a skeletal framework for
orienting postwar peacebuilding activities. The following statements are in-
tended to clarify the main traits of this admittedly underdeveloped approach:

1. Legitimacy-focused peacebuilding sees the political process as cru-

cial because it lies at the crux of economic, social and institutional
processes.
2. Legitimacy-focused peacebuilding focuses on the relationship
between state and society, encompassing the state and its legitimacy,
Conclusions for Theory 229

but going beyond it to other forms of authority and emphasizing their

articulation with society.
3. Legitimacy-focused peacebuilding acknowledges the importance of
state services and state capacity but subordinates it to the process of
legitimation of authority.
4. Legitimacy-focused peacebuilding emphasizes the importance of per-
ceived legitimacy of authority among both potential spoilers (elites)
and the entire population.
5. Legitimacy-focused peacebuilding acknowledges the role of electoral
processes as legitimizing events, but it sees alternative forms of con-
sultation and participation as useful vehicles for legitimacy.
6. Legitimacy-focused peacebuilding sees the postelection moment as
especially perilous for peace. It urges sustained, creative engagement
to enhance inclusionary state behavior after elected governments
take office.
7. Legitimacy-focused peacebuilding eschews two hallmarks of lib-
eralism: the framework of democratization and neoliberal eco-
nomic policies.

Legitimacy-focused peacebuilding serves here as a conceptual orientation

to respond to the empirics of the cases of postconflict peacebuilding failures.
It also draws on an examination of the pool of successes and failures of postwar
peacebuilding. The principles described above provide a skeleton around which
better understandings of the peacebuilding process, and its multiple outcomes,
can be analyzed. It also offers clear conceptual alternatives to the most prevalent
theoretical approaches to peacebuilding.
However, the development of this conceptual approach by no means answers
such crucial policy questions as the following:

1. Who should undertake such legitimacy-building efforts?

2. If internal actors should undertake these efforts, then which internal
actors are best positioned to do so?
3. Will the internal actors with the capacity to act have the will, and vice
versa? If not, what incentives might change their calculations?
4. What can external actors do to advance legitimate authority in post-
conflict societies? Will such efforts have the intended effect?

These last questions are important in light of the disjuncture and even con-
tradictions that have come to light between certain scholarly causal claims and
policy responses. More simply put, the discovery or claim in the social sciences
that factor A is correlated with outcome B does not necessarily mean that policy
230 Implications for Theory and Practice

actions intended to enhance or stimulate A actually lead to B. Why not? Be-

cause other variables intervene, and because the action to stimulate A may itself
produce other consequences that undermine or interfere with the desired im-
pact on B.
One prominent and relevant example took international relations by storm
in the 1980s and 1990s: the democratic peace thesis. The thesis—democratic
countries do not go to (interstate) war with one another—has extremely strong
empirical support. Critics debate not the degree of empirical support for the
thesis but whether there are any exceptions to it (Doyle 1986; Russett 1993;
Russett and O’Neal 2001; Layne 1994; Maoz and Abdolali 1989). However, Sny-
der (2000) and Mansfield and Snyder (2005) show that efforts to initiate de-
mocracy in certain polities often have the unintended consequence of sparking
armed conflict and thus undermining democracy and peace. Even if one agrees
that democracy favors interstate peace, democratization does not foster peace
unproblematically. As we saw above, democratization in Burundi and Rwanda
coincided with renewed civil wars in those countries.
Given the potential disjuncture between scholarly claims and prudent poli-
cies, what are the implications of the findings of this book for those who seek to
consolidate peace? Who should undertake legitimacy-focused peacebuilding?
Is it possible to “build” legitimacy, and to do so without causing harm? 4 The
next chapter takes up these questions and thus offers conclusions for policy and
practice. But before turning to these conclusions, it is important to acknowledge
and reflect on some of this study’s main shortcomings or limitations.

Addressing Limitations
The study makes ambitious claims about theory, method, and policy approaches
to postwar peacebuilding. At the same time, the study exhibits several short-
comings, including the limits of the methodology, the concepts used, the em-
pirics, and the argument. These limitations simultaneously suggest further lines
of inquiry and thinking. This section addresses three of these weaknesses and
some of their implications for the study of war and peacebuilding. Broadly speak-
ing, I present each of these lines of inquiry in the order of its potential scope
and significance.

The Argument
First, the argument advanced here—that sustained peace depends upon the in-
clusion of former enemies and the social groups associated with them—is not
new. Others have pointed to the concept of exclusion as well as the main mecha-
nism of inclusion, powersharing, in accounting for civil war outcomes, if not the
Conclusions for Theory 231

less-researched concept of war recurrence. However, this study uses quantita-

tive methods as a jumping off point for more in-depth qualitative research on a
pool of cases. Even if one contests the criteria for creating the pool of cases, any
adjustment of the pool would be unlikely to reverse the prevalent role of exclu-
sionary behavior in recurrences, and of inclusionary mechanisms in the cases
of nonrecurrence.
Yet the principal manifestation of inclusion—powersharing—exhibits well-
known weaknesses. As analyzed in chapter 1, these weaknesses are (1) inflex-
ibility; (2) a tendency to marginalize broader participation, at the expense of
“vertical legitimacy”; and (3) exclusion of the most marginal or tertiary ethnic,
social, or political groups. The Lebanese case exemplifies the problems of inflex-
ible powersharing, because fixed quotas of state offices did not allow for evolving
population growth, and much less for the influence of large numbers of nonciti-
zen refugees whose presence altered society. These weaknesses underscore the
possibility that powersharing will end in mass violence.
However, this study’s empirics suggest that the likelihood of recurrence under
powersharing is overstated. Moreover, comparable to measures that can mitigate
the tensions inherent in inclusion of unsavory leaders, scholars and policy ana-
lysts have in recent years emphasized a number of ways to embrace the utility of
powersharing while reducing its negative attributes. For instance, powersharing
can be limited in time to a few years, after which a more flexible system of power
(or elections) comes into effect, possibly including majoritarian elements. Simi-
larly, the scope of powersharing can be scaled back over time, which gradually
opens state offices to electoral processes or unrestricted appointments. Alter-
natively, powersharing arrangements can be revisited after a number of years,
either through new elite negotiations or through a referendum or electoral pro-
cess. One salient example of both these last mechanisms is the multifaceted (ter-
ritorial, security, economic, and political) powersharing arrangement that ended
Sudan’s second civil war in 2005. That agreement provided for elections after
four years to replace a number of apportioned government posts, and for a refer-
endum on South Sudanese independence after a six-year period.
Another means of diminishing the shortcomings of powersharing is by adopt-
ing a system whose referents are “generic” minorities, determined through elec-
toral outcomes, rather than concrete ethnic or religious groups. Thus the top
four parties may be associated with shifting ethnic or social groups over time,
which injects a flexibility that undermines rather than reinforces the reigning
identity groups at the time of transition. Powersharing advocates increasingly
also acknowledge the utility of these various means to mitigate the downsides
of powersharing. Finally, it bears noting that powersharing, which figures promi-
nently in my findings, has not been tested in this study. Although powersharing
emerges as the most salient form of inclusion in sustaining peace, I did not use
232 Implications for Theory and Practice

a method that permits me to determine the extent to which powersharing cor-

relates with success or failure. That approach, carried out by others with regard
to negotiated settlements (Hartzell and Hoddie 2007) and democracies, would
have required examining the pool of instances of powersharing. That question is
different from the one this study explores about why peace fails or succeeds, but
its urgent relevance is made apparent by my findings.

The Method
Second, this study seeks to do something novel methodologically. It uses both
quantitative methods and a more in-depth case study, but it combines these with
mini–case studies of the entire pool of civil wars that experience recurrent armed
conflict. It also draws heuristically on the pool of civil wars that do not recur.
This tripartite method—seeking not only quantitative methods and case studies
but also a middle ground of extremely short, focused case studies—has some
deficiencies in both concept and execution.
Conceptually, it is impossible for anyone to perform numerous mini–case
studies with the degree of contextual vibrancy and fullness that in-depth case
studies offer. Case specialists will always find themselves dissatisfied with the
omission of key elements and causal relationships that did not make the cut. By
drawing on a structured, focused tablet of hypothesized independent variables,
any researcher will omit the unique or case-specific explanatory factors that have
not been proposed or hypothesized. Using primary source material on such a
large pool of cases is impossible except in large, well-funded collective research
projects—which are discouraged by academic career incentives. Furthermore,
comparativists (among whom I count myself) may find the comparison of
widely disparate cases like Russia, East Timor, and Sudan excessively ahistorical
and decontextualized, especially in the absence of regional similarities that act as
controls. Although these sorts of disparate comparisons are at the core of quan-
titative work, such methods purport to control for potentially relevant variables,
which renders the causal relationship more reliable than when such controls are
not able to be exercised.
In terms of execution here, the linear regressions presented would have
been more robust and persuasive if they had included more complex methods.
They could also have utilized armed conflicts themselves as the unit of analy-
sis, rather than a country whose inclusion is clouded by multiple or unrelated
armed conflicts, which unduly reduced the data set. In addition, the single case
study of Liberia relies on secondary sources for judgments about things such as
the quality if interwar peace and the role of various factors in recurrence. And
experts may criticize the analyses contained in the mini–case studies, even in
condensed form.
Conclusions for Theory 233

Despite these flaws, the method of short, focused case studies also shows
promise for future research on related topics. The presentation of such case stud-
ies, especially where the entire universe of cases can be covered, addresses the
main deficiencies of the two most prevalent methodological approaches: large-N
quantitative studies and a few in-depth case studies. As described in chapter 1,
large-N quantitative studies cannot address variables that do not lend themselves
to useful proxies or measurement. Case studies, including mini–case studies, can
address such hypothesized factors. Furthermore, covering the full scope of cases
gives the reader a sense of understanding of the dynamics at work for all the
cases, which permits the identification of patterns and an understanding of inter-
action among different variables as well as the role of different levels of analysis.
In-depth case studies, of course, offer all these advantages, but it is simply not
possible to construct a dozen or two dozen such in-depth case studies in a way
that is either feasible for researchers or digestible for the general reader.

Concepts and Terms

Third, and finally, this study shows the limits of the concepts of “civil war” and
even “armed conflict,” which suggests the fruitfulness of exploring concepts that
capture the complexities of contemporary organized violence. For most of the
twentieth century, scholars of international security focused excessively on inter-
state warfare, a situation that changed only with the end of the Cold War and the
decline of interstate warfare. During the 1990s scholarship focused increasingly
on “civil wars,” and some specialists in interstate warfare brought their concepts
to bear on these civil wars (Posen 1993; Walter and Snyder 1999; Walter 2004).
The reality of organized violence in the past twenty years (and perhaps de-
cades earlier) calls for scholars to explicitly recognize the limits of these concepts
of warfare, on three grounds. First, the clean demarcation of cross-border warfare
and intrastate warfare is obsolete. Many of the cases of intrastate recurrence ex-
plored here—Georgia, Lebanon, Sudan, Nicaragua, Burundi, Rwanda, Liberia,
the Central African Republic, and Haiti—involved extrastate armed actors or
sponsors of armed actors. The Peace Research Institute Oslo–Uppsala Conflict
Data Program data set shows that only three armed conflicts of the most recently
coded decade (1999–2008) were interstate wars that involved no intrastate
armed forces (viz., Djibouti / Eritrea, 2008; Ethiopia / Eritrea, 1998–2000; and
India / Pakistan in Kashmir, 2003).
A new category of warfare—proxy wars—also deserves more attention.
Some have suggested that the reason Africa records very few interstate wars is
that these take the form of proxy wars instead (Craig 2011). As Salehyan (2008,
2010) shows, rebels are agents who seek out support from external sponsors,
take action across borders, and attack across borders to draw more supporters
234 Implications for Theory and Practice

and resources, often ensnaring other states as enemies. Conversely, states can
constitute rebel organizations in other countries or inside their own borders to
attack other states and to feed, arm, train, advise, and provide legitimacy and
other communications support to such proxies. The relationship between state
sponsors and their proxies requires much more research, and it falls through the
cracks of most of our conceptual maps.
Second, if civil wars are wars between two formed armies, then what happens
when one side is not an army at all? Scholars must increasingly recognize the var-
ious and increasingly important roles played by civilians—as agents of violence,
as victims, or as active supporters—in relation to armed combatants. There are
various dimensions to this relationship. One is that scholars need to devote more
attention to analyzing “one-sided violence” and its relation to more conventional
notions of warfare. Andrew Mack and the Human Security Research Group have
made pathbreaking progress in emphasizing this form of violence as distinct from
warfare and going beyond genocide (Human Security Research Group 2005,
2009). It is clear that one-sided violence is a salient subject in its own right. What
is not clear is how one-sided violence interacts with more conventional warfare.
Many of the cases included in the Fearon and Laitin (2003) data set would not
be counted as “civil wars” by many observers. Some, like the Central African
Republic, come closer to coups involving dozens of deaths, whereas others rep-
resent high levels of one-sided violence. The cases of Rwanda, Burundi, Haiti,
and others demonstrate the need for more discriminating concepts of what have
been called “civil wars.” They also show the importance of analyzing how “one-
sided violence” interacts over territory and across borders with organized armed
factions (civil, interstate, or proxy war). When does one-sided violence breed
insurgent movements or trigger their attacks? Does one-sided violence draw in
third-party armies / peacekeepers more often than what are seen to be civil wars?
Do external states react differently once a civil war becomes one-sided violence?
In addition to one-sided violence, the role of civilians further challenges our
understanding of war. The most obvious is that civilians frequently comprise the
bulk of both “combatants” and casualties. Attacks are often carried out by “civil-
ians” who take on combat roles for only a few hours. Personally, for instance, I
recall my experience in the late 1980s living in a village in El Salvador, where the
guerrillas sadly pressured children to participate in brief nighttime excursions
to burn buses in neighboring villages. And civilians conduct training and indoc-
trination for warfare, not to mention financing. The Human Security Research
Group (2009) has also drawn attention to the role of civilians as the principal,
often uncounted, victims of warfare in places like the Democratic Republic of
the Congo, where their flight from combat leads to disease and hunger-related
deaths. Because these are not combat-related deaths, they are not convention-
ally counted as casualties of war. As the events in Libya of 2011 illustrate, civil-
Conclusions for Theory 235

ian demonstrators and protesters can also spark internal armed conflicts, and
even civil wars, if armies split and states militarize such contests. More broadly,
other nonstate civilian actors—including members of diasporas, multinational
corporations, nongovernmental organizations, and regional intergovernmental
organizations—also play increasing roles in armed conflicts. These multifaceted
roles of civilians and how they relate to combat have received increased atten-
tion in recent years, but this scholarship has been poorly integrated into broader
understandings of civil wars and their recurrence.
In summary, the study of civil wars, their recurrence, and warfare more gen-
erally requires more nuanced concepts for the contemporary world. By delving
into the specifics of the cases that are captured by some of the best data sets,
data that are used to draw high-profile and widely read scholarly works on civil
wars, this book reveals the diversity of these cases. Rather than assuming that
more numerous and discriminating control variables can ensure that comparable
events are being compared, scholars should press to develop more discriminat-
ing concepts that differentiate these kinds of violence in ways that might lead to
better understandings, enhanced theory, and improved policy. The implications
for this last arena—policy and practice—are especially important in a field so
intertwined with those who actively seek to advance peace, justice, and develop-
ment. The final chapter turns to these implications.

Notes
1. Paris (2004, 181–82) also addresses partition as an alternative. However, parti-
tion is one solution that parties may reach through dialogue, and renegotiating state
borders is not incompatible with liberal traditions.
2. See the 2005 Paris Declaration and the 2008 Accra Agenda for Action, both at
www.oecd.org / document / 18 / 0,3343,en_2649_3236398_35401554_1_1_1_1,00&
&en-USS_01DBC.html.
3. Zartman’s works are useful to consult—see Zartman (1977, 1995a, 1995b,
1995c) and Zartman and Berman (1982).
4. On “doing no harm” in postconflict state building and peacebuilding, see Call
(2008a); DAC-OECD (2007).
9
Conclusions for Policy
and Practice
Can External Actors Build Legitimacy after War?
PAST EXPERIENCE INDICATES THAT the enterprise of consolidating peace
fails in a high percentage of cases, often within the first few years of a cease-fire.
Transforming states and their relationships with societies is even more chal-
lenging, and the role of external actors is complex and fraught with risks. The
previous chapter explains how postwar inclusion of elites is tied to theoretical
notions of state legitimacy, and advocates a “legitimacy-focused peacebuilding”
that contains elements of, but differs from, liberal, democratizing, republican,
state-capacity-focused, and critical approaches to peacebuilding. Given how dif-
ficult it seems for outside actors to foster inclusionary behavior and the state
legitimacy that seems so central to making peace stick, is there a role for external
legitimacy building? What role, if any, can outsiders play in fostering inclusion
and legitimacy?
The behavior of elites toward different ethnic or religious groups is shaped
by numerous factors that are historically rooted, economically driven, and resis-
tant to outside pressures. National elites and their own leadership (Machiavelli’s
virtu) have historically proven decisive in the process of laying the foundation for
a legitimate state in a war-torn society, one in which divided groups eventually
consent to conform to the political order and acknowledge, albeit begrudgingly,
the authority of political structures, rules, norms, and governments. Exclusion-
ary behavior occurs foremost through the agency of national elites, especially of-
fice holders. The ability of international actors to influence such behavior is not
straightforward. A less politically engaged role for donors and the UN system—
focusing on state capacity, basic needs like health and education, social dialogue,
or verifying peace agreements—might seem to make more sense than trying to
engage the complex and perilous process of enhancing horizontal legitimacy in
war-torn societies.
This final chapter argues the contrary. I contend that there is a positive role
for external actors in enhancing inclusionary behavior and legitimacy-focused
peacebuilding, despite the limits and perils of their influence. Those perils are
Conclusions for Policy and Practice 237

extraordinary. External actors face systematic and, in most ways, growing con-
straints in their ability to shape the legitimacy of postwar states through delib-
erate efforts. The chapter begins with the formidable obstacles facing external
actors to engage in the tasks of enhancing inclusionary governance and building
legitimate authority. It then returns to the earlier point that no single formula
or model of inclusionary governance will guarantee legitimacy. However, I do
find a set of persistent challenges, trade-offs, or decision points common to most
postwar societies.
The chapter then breaks down the process of transition from warfare into
four “moments,” or temporal junctures, whose dynamics differ from one another,
but that hold some common challenges across most postconflict societies. These
four moments are (1) the decision on what structures and which individuals
will rule for an interim period, (2) the long-term “design” of the state, (3) elec-
tions and the end of an interim administration, and (4) the postelectoral period.
Although the first three phases are important, it is the “postelection moment”
where exclusionary behavior has led to recurrence in the case studies examined
here, and the phase that has been neglected in research. The chapter outlines
the particular challenges and possible steps to overcome them for each of these
phases in a postconflict environment. In doing so, it advances several principles
for peacebuilders that maximize the chances to overcome the obstacles to inclu-
sion and state legitimacy described in the previous chapter.

Why Legitimacy Building Is Exceptionally Difficult

Given the centrality of political legitimacy for peace consolidation, external ac-
tors must recognize up front the severe constraints on their ability to foster this
legitimacy. It is difficult for multilateral or bilateral actors to deliberately enhance
the legitimacy of a specific regime, and nigh impossible to do so consistently
across the gamut of postconflict or fragile states. Why is that the case?

The Inherent Conundrum of Enhancing Legitimacy from Without

External efforts to enhance legitimacy represent an inherent conundrum or con-
tradiction. To the extent that outside actors use either resources or leverage to
influence the state or its relations with society, they substitute for or interfere
with the relationship between state and society. The extensive literature on the
informal state, the “shadow” state, and the “quasi-state” in Africa shows that
the very effort to legitimize a state can and often does readily undermine that
state’s legitimacy in important ways (Herbst 2000; Reno 1999; Jackson 1990).
It creates or strengthens a constituency (viz., the outside actor) other than its
core constituency (its own people). Even a “sincerely autonomous” request by
238 Implications for Theory and Practice

national-level authorities for external assistance or recognition means that the

state has sought out a competing constituency and, perhaps unwittingly, under-
cut its direct relationship with its own population.
Of course, the recognition and support from international actors contrib-
utes in important ways to the legitimacy of a given state as well. Admission
into regional or multilateral organizations confers more legitimacy on a polity
than exclusion. Yet states that acquire a good portion (some receive more than
50 percent) of their revenues from donors must answer to these donors, often
with fiscal and monetary policies that may attract foreign investment but have
differential internal economic and political consequences. They may adversely
affect not only the poor but also ethnic groups whose members see themselves
as losing in such policies.
Moreover, where populations know that nonstate outsiders provide services,
they have no reason to respect, invest in, or heed national authorities. Incentives
to follow electoral campaigns, vote, or lobby legislators evaporate when deci-
sions about services and resource provision come from external actors. Politi-
cal office seekers then have fewer grounds on which to make appeals, and thus
turn to more symbolic politics, and perhaps identity politics. To the extent that
elections and interest group relations with office holders function as sources of
accountability or mediation of state–society relations, this accountability and
mediation is weakened by the exercise of external leverage.

Increased Resistance to the West

Resistance to liberal policies and to perceived Western manipulation has increased
in recent years. Boosted lately by resentment in the Muslim world at US wars in
Afghanistan and Iraq, the global South is articulating and acting on rejection of
US and Western liberal notions. In 2001, for example, UN member states opted
not to elect the United States to the Human Rights Commission for the first time
since its creation in 1947. China and other “nontraditional donors” increasingly
urge poor countries to reject the liberalizing and “democratization” conditional-
ity that often accompanies aid from the West. In Latin America, the “pink tide”
of governments aligned with Venezuela’s Hugo Chávez signaled a surge of anti-
neoliberal and, to a lesser extent, anti–United States stances in that region. The
election of President Obama blunted the sharpest edges of this trend, but has not
reversed the more structural elements of this reaction to US hegemony.
In terms of Western efforts to build state legitimacy abroad, the perception
that outside, hegemonic, Western, or imperial interests are orchestrating state
reform or backing a particular set of elites can produce a backlash. Such a per-
ception can lead to nationalist resentment, anger, and rejection of the national
elites allied with the external interests. In effect, even if legitimacy builders are
Conclusions for Policy and Practice 239

heeding national elites’ wishes and deferring to their judgments, they could still
undermine the latters’ legitimacy solely by virtue of their external character. This
“legitimacy-building conundrum” presents especially difficult hurdle in contem-
porary international relations in the wake of intensification of global Southern
rhetorical resentment of Western dominance, especially in the Arab world.

The Curious Strength of the Sovereignty Norm

In recent years, scholars have increasingly proclaimed the demise, or at a mini-
mum the attenuation, of the sovereignty norm that underlay the postwar inter-
national system. Even as some analysts have argued that Westphalian notions
of sovereignty have never existed in reality, many analysts have argued that re-
cent developments have eroded the sovereignty enjoyed by states since World
War II.1 Several factors are allegedly at work: increased economic interdepen-
dence and activity; globalized communications; transnational regimes for trade,
human rights, and transnational crime; and heightened expectations of states’
responsibilities to “protect,” to govern ungoverned spaces, and to deliver basic
services (Rosecrance 2000; Keohane and Nye 1977; Ruggie 1993). Transna-
tional sources of power, in this view, are inevitably supplanting sovereignty.
And yet recent developments have shown that, to paraphrase Mark Twain, re-
ports of state sovereignty’s death are greatly exaggerated. Many states, especially
the very strongest and very weakest, cling to state sovereignty in ways that sys-
tematically undermine the ability of international bodies or coalitions to engage
domestic affairs within states, including some of the central state-building tasks.
Diplomatic leverage respects certain boundaries much more than one might ex-
pect with the talk of globalization and the International Criminal Court. Oil-rich
and corrupt dictators’ rule and riches remain unchallenged by Western states’
trade and foreign policies. The unwillingness of the world to enter into combat
in Darfur emblemizes the strength of sovereignty in the face of perhaps the most
widely agreed upon exception to its rules: genocide. Indeed, resource-rich states
like Venezuela and Saudi Arabia have in many ways increased their leverage in
the twenty-first century. Within the United Nations, the persistent success of the
permanent five members of the Security Council in blocking a more representa-
tive council, of states that weaken the enforcement mechanisms of the UN Hu-
man Rights Council, and of poor states that oppose putting a “responsibility to
protect” into practice all reflect a rejection of incursions on sovereignty.
The sovereignty norm persists among weak states in a variety of ways that
complicate international efforts. First, Western and international officials have
limited leverage in dealing with sovereign states. They could use military force,
but this is an option that is only used in the extreme, and often with unforeseen
negative consequences—as the Iraq war reminded us. They can also exercise
240 Implications for Theory and Practice

leverage by threatening to cut things like travel privileges, trade preferences, and
diplomatic relations. Yet such steps are disproportionate for trying to stimulate
political will to reform the judiciary or to seat a representative on an electoral
tribunal. Of course, the aid for specific programs themselves can be threatened
(or offered as incentives), but all too often sovereign states can enjoy the benefits
of aid without in the end complying with the strings attached. And donors are
reluctant to follow through on such threats, because it leaves them with no le-
verage and little to show to their constituents back home. As mentioned above,
some poor states are also expressing a clearer rejection of political conditionality
because they have been encouraged by competing nontraditional donors.
In addition, development actors tend to see themselves as providing resources
even to illegitimate regimes, emphasizing “getting along” with the government in
order not to be cut off or kicked out of the country by sovereign states. They have
historically been reluctant to threaten to withhold resources over “politics” (e.g.,
exclusionary behavior or a lack of transparency) that lie at the heart of legitimacy.
Many development officials also resent the idea that peacekeeping organizations
or diplomats might leverage their “neutral” development aid for political ends
such as incentivizing spoilers or bolstering powersharing.

The Negative Effects of “Democracy Promotion”

Efforts to foster legitimacy also seem to face the same challenges that democ-
racy promotion has faced. Indeed, given that democratic regimes are the main
standard for legitimacy in the international system, why focus on “legitimacy
building” at all, rather than democratization? After examining the contributions
of democratization to the study of peacebuilding, this section sets forth several
important advantages of legitimacy building over democracy promotion. How-
ever, both enterprises confront some of the same obstacles.
First, it is important to recognize that democracies themselves are more in-
ternally stable and more prosperous than nondemocratic or anocratic regimes.
They virtually never go to war with one another (Doyle 1986; Russett and O’Neal
2001; Ray 1998). Democracies are generally less likely to experience civil war
than nondemocracies (Krain and Myers 1997; Henderson and Singer 2000).2
Moreover, democratic forms of authority enjoy support from oppressed and ex-
cluded populations (Ponzio 2010), and they hold an initial level of legitimacy
within postauthoritarian societies, as reflected in the Arab Spring of 2011.
However, the process of abetting democracy differs in its risks and conse-
quences from the presence or existence of democracy. As mentioned above, even
though democracy is positive, steps to reach it—namely, democratization and
democracy promotion—are perilous and often harmful. Below I provide a more
detailed analysis of elections; however, a few brief observations are warranted
Conclusions for Policy and Practice 241

here. Although elections have come to epitomize the institution whereby rulers
can supplant one another nonviolently, they can also spark social division and
conflict, especially where uncertainty due to past authoritarianism or warfare
reigns. As Kumar and Ottaway (1998, 230) admit in a volume that advocates
the positive impact of elections for postconflict societies, “In the absence of a
tradition of democratic contestation, elections in the aftermath of civil war can
be highly divisive.”
Multiparty democracy injects a chance at power, but also a chance for oth-
ers to gain unfairly. Societies experiencing fair and competitive elections for the
first time also face uncertainty about who might cheat and how, especially in the
absence of an effective third-party arbiter of election behavior (be it the courts,
a trusted electoral tribunal, or even an international third party). Unfortunately,
international actors are often not authorized or are ill prepared to confront the
worst cheating and election-related violence. Therefore, much like the situation
in which a commitment problem can lead to violence between two parties whose
relationship undergoes a transition once an overarching authority is weakened
(Fearon 1998), so also a transition from a single-party regime to a multiparty
one can spark violence centered on the electoral timetable. For external actors,
promoting elections may channel attention and resources away from inclusion.
As Carothers (2009, 10) says, “A narrow focus on the basic institutions of politi-
cal competition, especially elections, may not help aid providers to arrive at ways
of broadening inclusion, representation, and participation.”
Most notable here is the role of elections in sparking mass violence. Mans-
field and Snyder (2005, chap. 5) find that in countries undergoing transitions
to democracy, the risk of interstate warfare goes up by a factor of four to fifteen.
Snyder (2000) documents a number of instances where democratization, and
specifically the introduction of competitive, multiparty elections, played a lead-
ing role in the onset of civil war. Prominent cases include the internal wars that
became cross-border wars in Yugoslavia after 1991, Rwanda in 1990, Burundi in
1993, Nagorno-Karabakh in 1992, and Georgia in 1992 (Snyder 2000; Mansfield
and Snyder 2005). Mansfield and Snyder (2005) describe the polarizing effect
of elections via the stirring up of nationalism.
The poster child for the violent potential of postwar national elections is An-
gola in 1991. During peace negotiations, UNITA rebel leader Jonas Savimbi used
the international fascination with a presidential election as a means to regroup,
buy time, and test the opportunity to gain legitimate access to state power, with-
out a commitment to abide by the results. Once his loss was apparent, he resumed
warfare. Whereas the war before the elections had cost 100,000 to 350,000 in
battle deaths, with hundreds of thousands displaced, the postelectoral violence
was by most accounts even more deadly (Tvedten 1997, 111; also see Duke 1998;
Stedman 2002). Between 300,000 and 500,000 people died just between the
242 Implications for Theory and Practice

1992 elections and 1994, which illustrates the dangers of an ill-conceived rush
to national-level winner-take-all elections.3
Since that tragedy, most policymakers and scholars have aggressively dis-
avowed prompt presidential elections as a sufficient marker for democracy (US
Agency for International Development 2005; United Nations Development Pro-
gram 2007). They have instead championed the demobilization of insurgents
before national elections (e.g., Liberia after 2003); the delay of elections until
state institutions are sufficiently in place to manage potential instability (e.g.,
Kosovo); and the holding of municipal and / or legislative elections before presi-
dential elections, a point to which I return below (Paris 2004). Each of these
modifications of the crudest form of postwar democracy promotion signals a
greater flexibility and context-specific approach to postconflict assistance.
Even with these alterations to the “liberal” or “democratic” peacebuilding
model (see Paris 2004; Ottaway 2000; Richmond and Franks 2009), however,
elections continue to serve as the focal point for international legitimation ef-
forts in postconflict settings. Thus the holding of a procedurally free and fair
election is deemed to settle the question of internal and external legitimacy. Ex-
ternal actors are understandably reluctant to broach indicators of broader illegiti-
macy, because these might be interpreted as stepping back from the principles
of democracy. Within the “democratization” framework, national actors have no
clear alternative framework within with which to challenge exclusionary behav-
ior. The presumptive answer, aside from resort to violence, is to wait until the
next election. And the international community is unlikely in the medium term
to agree on an alternative regime type as preferable to some form of competi-
tive multiparty electoral system (on democracy assistance, see Carothers 1999,
2009; McFaul 2004).4 In posttransition states that are even minimally reliant on
external assistance, some form of electoral regime remains the sine qua non of
interstate legitimacy.
If democracy promotion can be harmful, why is legitimacy building any less
perilous? Given that state legitimacy inevitably involves some form of electoral
democracy, it would be deceptive and naive to pretend that it is possible to con-
ceptualize “legitimacy building” or to engage in it in a way that is wholly indepen-
dent of many of the challenges posed by democratization. The inherent contra-
diction of using outside influence to empower “rule by the people” applies to both
“legitimacy building” and democracy promotion. The tensions between elite-
driven bargains and popular inclusion and participation also afflict both these
enterprises. And the greater a state’s legitimacy, the less influence or leverage en-
joyed by outsiders, whether elected in a shallow democracy or selected through
a nonelectoral form of consultation or participation.
Nevertheless, the concept of legitimacy building offers some advantages of
democracy promotion. As discussed in chapter 2, state legitimation is distinct
Conclusions for Policy and Practice 243

from a particular regime type, although some form of popular consultation and
participation constitutes a minimal foundation for broader state legitimacy. First,
state legitimacy encompasses more multivalent political relationships than the
rules of the political game and their procedures connoted by democracy. Among
those political relationships, “legitimacy” more explicitly includes broader elite
bargains than the more formal constitutional or other pacted agreements that
appear in the democratization literature and are connoted by democracy pro-
motion (Karl 1990; Mainwaring, O’Donnell, and Valenzuela 1992; O’Donnell,
Schmitter, and Whitehead 1986; Przeworski 1991; Hagopian 1991). The politi-
cal relationships connoted by “legitimacy” also encompass a broader array of
forms of participation for society, including various social sectors or the “masses.”
Furthermore, state legitimacy derives not solely from the representational
element of society in the state, although this is one core element. The “legiti-
macy” of a state and its authorities may derive from forms of consultation or
participation that differ from formal party-based elections that are the hallmark
of democracy. In this sense, legitimacy is more difficult to define, to measure, and
to influence. But it is also more flexible than the specific features or structures
associated with democratic regimes.
Second, the terms “state legitimacy” and “strengthening legitimacy” do not
raise the hackles of the global South nearly as much as the language of “democ-
racy” and “democracy promotion.” Thomas Carothers (2009, 6) speaks of a
“growing backlash against both democracy promotion and democracy more
generally.” The spread of elective, one-party regimes in parts of Africa (Ottaway
2000), together with a renewed wave of coups and coup attempts (LeVan 2011),
calls into question the unbridled march of multiparty competitive regimes and
liberal rights (Carothers 2006). In addition, China and other nontraditional do-
nors now deliberately seek to offer an alternative to conditions related to human
rights and democratic governance placed by Western donors on their foreign
aid. These trends, coupled with the backlash against US hegemony surround-
ing the aftermath of the September 11, 2001, terrorist attacks and the perceived
entanglement of the United Nations and other international organizations with
Western power, deepen the mixed reception of much of the world to Western
democracy-promotion efforts. Although the global South does not embrace the
concept of state legitimacy, it remains more amenable to the concept than to
democratization.
Legitimacy has the advantage of including a crucial element of state–society
relations: the ability of the state to meet social expectations about delivering ba-
sic services. These basic services at a minimum include ensuring security from
external and internal threats, minimal regulation for the economy, and minimal
public goods such as infrastructure, education, and health. One study, by the
economist Philip Keefer (2008), finds that the ability of leaders to make credible
244 Implications for Theory and Practice

commitments, especially to service delivery, is crucial in whether armed chal-

lengers emerge.
This conceptual link also represents a disadvantage relative to democracy, be-
cause it inextricably ties capacity to legitimacy. State capacity, therefore, is some-
what endogenous to legitimacy. This complicates the useful distinction between
legitimacy and capacity, given that very capable states can also suffer legitimacy
deficits, just as legitimate states (think of postcolonial states at the dawn of inde-
pendence) can also be quite weak.
In sum, legitimacy building confronts some of the same challenges as democ-
racy promotion, plus others, such as less denotative specificity and its inextri-
cable partial reflection of state capacity. However, it also offers important ad-
vantages. It focuses more squarely on the relationship between state and society.
Most important, in an era of increasing resistance to the discourse of democracy
and democracy promotion, the term “legitimate authority” is less threatening to
national elites, enjoys broader and wider acceptance as a normative end for poor
and war-torn countries, and encompasses mass aspirations at least to the same
extent. In addition, it maintains a focus on the state and its institutionality, but it
is flexible enough to pertain to other levels of suprastate and substate authority.
Such levels of authority reflect changing governance models amid globalization,
a challenge to which I now turn.

Globalization’s Added Hurdles to Legitimacy Building

Globalization has had multiple positive effects on the ability of states to foster
legitimate states abroad. First, the spread of global norms surrounding core el-
ements of legitimacy is quicker, more intense, and farther reaching into non-
elite populations and remote territories than ever before. These include norms
about elections, an independent media, popular empowerment, autonomy for
both judicial and legislative branches, and related accountability checks on the
executive. Economic liberalism has spread even more widely, as emblemized by
Communist Cuba and China. The Internet has empowered poor, remote com-
munities to access models of governance around the world, and given voice to
many who were previously voiceless.
At the same time, globalization poses serious challenges for would-be legiti-
macy builders. The global reach of the Internet and communications provides
new and quicker means for spoilers to undermine peace processes. The diaspora
can channel funds in ways that advance inclusive relations, but various diasporas
too often act in favor of exclusionary elites back home. The association of global-
ization with the West also undermines its effectiveness. The Internet and social
networks can thus be mobilized more readily against external actors solely on the
Conclusions for Policy and Practice 245

basis of their status as external. Globalization thus presents both opportunities

and obstacles.
These obstacles constitute systematic and serious challenges facing external
actors that would undertake efforts to seek to consolidate peace by enhancing
legitimate authorities in postconflict societies. Given these obstacles, can and
should external actors seek to enhance legitimacy in war-torn societies? Why
should they have a positive impact? One crucial factor is the distinction between
horizontal and vertical legitimacy. International actors tend to focus on vertical
legitimacy and capacity in their peacebuilding, state-building, and democracy-
promotion efforts, but to give short shrift to horizontal legitimacy. The cases of
Liberia, Haiti, and East Timor show the perils of this approach, especially after
new governments have been elected.
Yet we have seen numerous instances where external states, international do-
nors, and intergovernmental organizations have been able to abet the inclusion
of former enemies in postwar political regimes. Verification and monitoring of
inclusionary agreements—namely, when accompanied by a multidimensional
deployment of troops and civilian assistance—has played a positive role in sev-
eral cases of nonrecurrence of civil wars. Prominent examples include the wars
of Southern Africa and Central America. In other cases, such as Tajikistan and
Papua New Guinea, external support for legitimacy took the form of diplomatic
and economic roles, but not troops, in support of inclusionary behavior building
on peace agreements. Are there lessons from prior experiences that might guide
external actors? The remainder of this chapter addresses this question.

Beyond Blanket Inclusionary Formulas: Four “Moments”

for Key Choices and External Strategy
Neither powersharing nor power dividing should be embraced as “models” to
be advanced in postconflict societies. As described earlier, legitimacy-focused
peacebuilding eschews blanket formulas suggested by either elites or outsiders.
Instead, the most sustainable processes of consolidating peace have emerged
from consultative processes that mollified the main spoilers and simultaneously
met the expectations of the members of broader society, including their desires
for representation and participation.
Sustained long-term peace depends on at least the bare outlines of a social
contract with the populace that also recognizes the diversity of society as seen
in important ethnic, religious, or social groups whose existence and power can-
not be denied, and whose salience in identity among segments of the popula-
tion has been heightened by war. Yet in postconflict societies, these processes
are fraught with difficulties and dangers. In the conundrum of local ownership,
246 Implications for Theory and Practice

someone must decide who the deciders will be. Actors who exercise control by
virtue of contested sovereign claims or because they can deploy military force
(including external actors like neighboring states, African Union forces, or UN
peacekeepers) will influence at some initial point the determination of who are
the “genuine” spokespersons for society and for major social groups. Someone
must decide which social groups merit a place at the table to begin with.
Such decisions will of course vary from society to society, and national ac-
tors usually fulfill that determining role. In exceptional cases (e.g., Bangladesh in
1997 and Djibouti in 1994), international actors will play no direct or influential
role in such decisions. Where international actors play such a role, it may well be
possible for the major international and national stakeholders to agree virtually
unanimously on the composition of a consultative process, even where no elec-
tions validate the composition. Yet the power of the groups with guns will always
play an oversized role in the constitution of the earliest foundations of a process
of determining who is a legitimate representative of a particular ethnic group
or of society as a whole. External actors perforce shape such a decision where
their armed forces are involved. This is certainly true of processes without elec-
toral validation (speaking here of processes in the earliest stages of consultation
to determine when, whether, and what sort of transitional or political process
should be adopted), but also true even where elections are held in the immediate
aftermath of warfare.
Here I turn away from specific formulas for enhancing legitimate authority
and from given processes that should be followed in postconflict societies. In-
stead, I suggest that thinking about particular decision “moments” in the evolu-
tion of postwar political processes offer clusters of decisions that will shape the
incentives for exclusionary behavior by national actors during and after a transi-
tion from war. Rather than suggest processes to be followed, I suggest that com-
mon choices confront international and national actors at specific moments in
most transitions from armed conflict. Thinking sequentially about possibilities
should supplant thinking mechanistically about implanting a particular form of
governance-like powersharing.
When addressing the ability of international actors to make enhanced legiti-
macy a cornerstone of peacebuilding policy, it is useful to focus on four moments
or phases. International actors face different challenges in each of these phases.
There is no formula to help international actors address legitimation challenges
in any of these moments. However, it is possible to provide midlevel generaliza-
tions (George and Bennett 2005) about what has tended to produce problem-
atic—or positive—outcomes. These points of guidance can serve as cautionary
tales or encouraging tales for possible lessons to be adapted. These moments are
intended to help organize our thinking (i.e., conceptual approaches) to postcon-
flict international efforts. In particular, some of these moments have received
Conclusions for Policy and Practice 247

insufficient attention in favor of others. Finally, the identification of these four

moments does not imply that there are ready formulas or correlations of success
found in any of them. On the contrary, none lends themselves to easy answers.
In this discussion, I draw on episodic illustrative examples of recent transitions
from warfare.

The First Moment: Deciding on Transitional Arrangements and Rulers

The first moment is the decision on a transitional political arrangement and tran-
sitional rulers. Unless an existing regime successfully defeats insurgents, wars
end with either a regime change or a negotiated settlement. Regime changes may
result in immediate instauration of a new regime, as in revolutionary Nicaragua
and Iran in 1979. Yet even when a regime falls or withdraws—as in Afghani-
stan, Kosovo, and East Timor—interim arrangements may emerge until more
permanent arrangements can be negotiated among the victors (and sometimes
the vanquished). Similarly, negotiated settlements may retain the existing regime
(as in El Salvador and Guatemala), but they usually involve some transitional
confidence-building arrangements.
Transitional governments constrain permanent arrangements in three ways.
If such interim arrangements are blatantly corrupt, unrepresentative, or abusive,
they can contribute to stagnant or negative legitimacy. The impact of an interim
administration is neither necessary nor permanent, but it is not negligible. Con-
versely, some of the legitimacy of an interim arrangement can rub off on a more
permanent arrangement if citizens feel included or represented in the interim
arrangement, and / or if the interim administration acts competently and cleanly.
Finally, transitional governments, which often reflect the historical institutional
arrangements inherited from the past, also tend to provide the most salient de-
parture point for choices about permanent institutions. Bargaining over such in-
stitutions naturally converges on existing arrangements as departure points for
debate and choice. In this sense interim arrangements shape the choices available
for permanent governance.
Past experiences tend to separate transitional arrangements into two catego-
ries: (1) countries that have experienced a robust international transitional au-
thority (e.g., Kosovo, East Timor, and Iraq) and (2) those whose experience has
been one where domestic actors never lost sovereignty, even where they had to
negotiate with powerful rebel forces and with international actors during a transi-
tion where foreign forces may have been present. The latter societies include the
three Central American cases, Mozambique, Angola, South Africa, and North-
ern Ireland. In between lie a number of experiences where national sovereignty
was formally maintained but with either a high degree of formal international
prerogatives (e.g., Cambodia and Bosnia) or a low degree of formal international
248 Implications for Theory and Practice

prerogatives but a high degree of informal external influence (e.g., Sierra Leone,
Liberia, Burundi, Haiti, and Afghanistan). In most of the latter cases (e.g., Haiti
and Afghanistan), external control is high in the immediate aftermath of war but
becomes more circumscribed once a national interim government is formed.
Less common are cases (e.g., Bosnia) where international prerogatives grow af-
ter an initial period.
One of the most striking bifurcations among the staff members of bilateral
actors and international organizations (and some academics) is between those
whose experience lies in cases where international actors exercised intrusive, ro-
bust prerogatives and those where they did not. The former group exhibits a
striking tendency in postconflict peacebuilding discussions to presume as their
starting point that international actors will or should exercise such heavy author-
ity. Conversely, those analysts shaped by the latter experiences tend to assume a
norm that national actors will remain the primary transitional agents. The reality
is that most transitional arrangements are chosen and led by national actors, even
where international troops, aid agencies, and diplomats exercise important influ-
ence on the selection of rules and / or individuals. In the next subsections I dis-
tinguish between transitions from war where international actors had a decisive
influence on decisions about interim authority and those where they did not.

Interim Governance: Cases of a Decisive International Role

What do recent cases indicate about what actions foster more legitimate politi-
cal arrangements, especially interim authorities? Let us begin with transitions
where international actors play a robust, even determinative role in deciding on
transitional authorities and rulers. Such a role may arise out of the defeat of a
state’s armed forces and / or the toppling of its regime. It may also result from the
withdrawal by a sovereign from a particular territory after a struggle for national
liberation, as occurred in East Timor.
First, the conditions of a war’s termination shape the role that international
actors will play in deciding on the subsequent interim administration of a state or
territory. NATO’s victories in forcing the Serbian withdrawal from Kosovo and
in toppling the Taliban regime in 2001, for instance, left NATO states at the cen-
ter of decision making about subsequent governance arrangements. When the
state withdraws or disappears, the victors become the protagonists in selecting a
successor regime. As the cases of Afghanistan, Iraq, and Haiti (in both 1994 and
2004) illustrate, international forces that constitute or support the victorious
side are quite influential.
At the same time, the conditions on the ground—literally—generally place
national rebel armies in a position of influence (or at least of veto power) vis-à-
vis decisions about both interim rulers and any redesign of the state. National
rebel armies often play the key ground troop role supported by (or supporting)
Conclusions for Policy and Practice 249

international air attack campaigns. Thus in Afghanistan in 2001, East Timor

and Kosovo in 1999, and Haiti in 2004, national forces were virtually the sole
ground army. The Northern Alliance defied NATO’s bidding not to enter Kabul
upon the withdrawal of the Taliban regime in 2001, and thus it took and held the
capital on the ground and guaranteed itself an important role in the negotiations
that produced the Bonn agreement on interim governance in Afghanistan. The
Kosovo Liberation Army was the sole ground invasion force in defeating the
Serbian forces, because NATO was so concerned with force protection that it
refused to send in ground forces and flew at such a high altitude that its bombing
accuracy suffered. It is useful to briefly consider these cases in more detail.
Afghanistan. The wisest route to fostering sustainable legitimate authority will
generally be reliance on relevant national actors to deliberate over and decide
on an interim government. Such a path was followed, with initial but ephemeral
success, by the international actors involved in negotiating the Bonn agreement
for Afghanistan in December 2001. After Kabul fell in mid-November, NATO
deferred to the special representative of the UN secretary-general, Lakhdar Bra-
himi, to convene a process to decide on an interim authority that would enjoy the
support of international donors, NATO itself, and the armed factions that had
fought against the Taliban (Chesterman 2002, 2007).
Brahimi rejected suggestions for an international protectorate and adhered to
the “light footprint” of retaining national sovereignty and minimizing interna-
tional civilian presence, as he had urged to no avail for the East Timor mission.
Instead, he drew on advice from those with intimate knowledge of Afghan poli-
tics to convene a select group combining both potential spoiler warlords with
key civilian political figures. That group, which had been picked by international
actors in consultation with the main political and armed figures, was by no means
representative. (Nixon and Ponzio 2007; Deledda 2006). It included warlords
who had committed the worst atrocities of the prior decade, thereby assuring
that justice for those abuses would not be punished in the near term.5 It also
excluded those vanquished by the war, the Taliban, and the allied factions that
had fought NATO’s forces. Nevertheless, it drew in the main potential spoilers
from among the victors. Their role was to define, without any formal role for
external actors, a process whereby an interim government would be chosen. This
interim government would then facilitate the process defining the selection of a
permanent government.
The process defined by Brahimi and his colleagues was praised by many as
wise and prudent. In the Afghan context, a heavier exercise of external authority
would likely have elicited a violent backlash and thus jeopardized the fragile co-
existence among the various elements of the victorious Northern Alliance. The
light footprint approach was also criticized as being undemocratic, especially
when a traditional council mechanism (the Loya Jirga) used to determine the
250 Implications for Theory and Practice

interim government structure and personnel featured warlords who intimidated

civilian political representatives (Thier 2002; Johnson et al. 2003). As discussed
later in the chapter, the presidency of Hamid Karzai neither ended the ongoing
war against the Taliban nor resulted in an internally legitimate or responsive state.
Nevertheless, this process did prevent the feared fragmentation of Afghanistan at
the hands of the various armed factions that toppled the Taliban regime in 2001.
Iraq. The experience of Iraq, considered only briefly here, marks a contrast
with the consultative, inclusive interim government fostered in Afghanistan.
The administration of George W. Bush opted initially to appoint a US general
to head an interim authority in Iraq. When the tasks proved overwhelming, it
appointed a diplomat to head the interim administration, with a shadow interim
Iraqi administration. Before the US invasion, the Pentagon deliberately ignored
the advice of an expert panel convened by the State Department on postcon-
flict reconstruction and instead imposed a handpicked group of mainly exiles on
this interim administration. Not only were internal Iraqi factions not consulted;
neither were other US federal departments, allied governments, or the United
Nations (Diamond 2005). Having flaunted lessons of multiple transitions from
warfare, the Bush administration also purged all Baathists from the state appa-
ratus, leaving them angry, alienated, and jobless, while depriving the state of a
vast number of experienced civil servants. Before the constituent assembly, each
phase of decision making about who would lead the interim administration re-
flected centralized, ill-founded decisions by Bush administration officials rather
than a previously announced clear process of consultation with Iraqi factions. As
in Afghanistan, the defeated regime was entirely snubbed, which facilitated the
recruitment of their followers into the armed resistance. These choices, along
with al-Qaeda strategies and other factors unrelated to the interim government,
contributed to the emergence of the fierce armed insurgency and the civil war
that cost more than a hundred thousand lives, including more than four thou-
sand among the US and other occupying forces.6
Haiti and East Timor. Two cases illustrate the limits facing international actors
even where war ends through the capitulation of the opposing force. In 1994,
when a UN-authorized US-led intervention restored President Jean-Bertrand
Aristide to power, it was unthinkable that the international community would
have supported someone else as even interim leader. If it had done so, confusion
and popular outrage would have ensued. In 2004, after helping oust Aristide, the
United States had considerably more sway in the selection process of an interim
administration, because the insurgent forces were inchoate, relied on interna-
tional support, and mainly shared a fervent opposition to Aristide rather than a
coherent policy program.
In East Timor, upon Indonesia’s withdrawal following the violence after the
1999 referendum on independence, international actors (especially the Aus-
Conclusions for Policy and Practice 251

tralians, Americans, Europeans, and the UN) played an important legitimating

role in a process that would clearly lead to a new state. However, the protago-
nists remained the national liberation forces—the Timorese National Council
of Resistance (CNRT) and its armed wing, FALANTIL, both under the lead-
ership of Xanana Gusmao—which could now claim victory after a twenty-five
year struggle that had left about two hundred thousand dead. International ac-
tors considered several options, including a collaborative relationship with the
CNRT; outright CNRT governance; and the path eventually chosen, a wholly
sovereign international administration under UN leadership. Although Gusmao
initially favored an executive mandate for the United Nations mission, he later
came to feel that the UN had marginalized Timorese leaders excessively in its
interim administration (Forman 2005).
The choices made about the scope of international authorities in Timor,
where the CNRT leaders enjoyed very high legitimacy both internally and ex-
ternally, offer insights into the dangers of robust international mandates. Analysts
are remarkably divided about what international actors, mainly the UN, should
have done differently in East Timor. In accord with some findings of this study,
some argue that the armed violence in 2006 would have been prevented if the in-
ternational community had maintained a longer-term and more robust presence,
especially with respect to security. Others argue, also in accord with this study,
that the UN assumed excessive authority and included popularly supported Ti-
morese leaders in governing too little and too slowly (Chopra and Hohe 2004;
Caplan 2004b). One way to square these findings would have been to agree on
more robust authorities for Timorese nationals earlier and more explicitly. At the
same time, the international community could have extended a robust technical
support role, especially one that could have helped provide for institutional co-
herence, transparency, and accountability in the security and justice sectors for
several years. In other words, a less intrusive but meaningful mandate to provide
checks on interim administration might have proven wiser.
As in Haiti in the 1990s, the widespread and intense support for national
leaders rendered it virtually unthinkable that any Timorese besides the CNRT
could have been chosen as interim leaders in 1999. International actors’ role in
selecting who would serve as interim rulers was constrained, even as important
choices remained about the types of power that national and international actors
would hold. Similar constraints confront most transitions from wars that end
through a negotiated settlement. Below I turn to these cases.
In concluding this discussion of international interim authority, note that ex-
ternal actors, especially intergovernmental bodies, influence what role they will
play. This includes whether their own authorities will be robust and decisive, or
more subordinate to national interim or permanent authorities. Of particular
importance, international actors tend to prefer greater power and authorities.
252 Implications for Theory and Practice

They do so for both principled beliefs in their enhanced ability to foster peace,
prosperity, and justice and for self-interested goals of power and profit.
Despite the allure of greater prerogatives and powers, external actors should
resist this temptation unless it is absolutely necessary. The starting assumption
and the preferred option for external actors’ optimal role should always be mini-
mal unless circumstances dictate otherwise. If legitimation of internal state au-
thority is the linchpin for success in peace consolidation, then it makes no sense
to usurp this authority unless compelled by circumstances. And even where cir-
cumstances do seem to dictate a heavier hand for international actors, the long-
term formation of legitimate authority in a territory may suffer. Such a scenario
unfolded in Bosnia, where the Office of the High Representative accrued greater
powers to disempower spoilers but in the process undermined the emergence
of an unmediated relationship between the Bosnian state and the population
(Cox 2008).

Interim Governance: Cases of a Decisive Role by Domestic Actors

In negotiated settlements, domestic actors virtually always determine who shall
rule and under what interim arrangements. In many cases—including Mozam-
bique, Angola, El Salvador, Guatemala, and Burundi—the sovereign state and
the government in power retain authority during a transition, often accepting
confidence-building measures, constitutional changes, and / or some degree of
interim powersharing. In other settlements—such as Liberia, Nicaragua, South
Africa, and Sierra Leone—the constitutional order is preserved with modest
reforms and a change in regime or government. In such cases, national actors
determine both who and what form the transitional authority shall assume. In-
ternational actors face even more limits in the selection of interim rulers and
interim arrangements than in cases where their own forces have been involved
in interventions.
Dividing these cases of domestic-actor predominance into two further cate-
gories is useful: negotiated settlements and outright victory. Where the parties to
warfare have reached a negotiated settlement, international actors tend to priori-
tize encouragement of adherence to the terms of the agreement.7 They also often
advance powersharing in peace processes and press for the inclusion of poten-
tial spoilers ( Jarstad 2008, 2009). The main logic of international engagement
therefore becomes support for the peace agreement, pressure to maintain any
powersharing elements, and efforts to keep potential spoilers on board. Often
sacrificed are investigation and prosecution for past atrocities, consultation and
communication with the broader public, accountability for ongoing misconduct
(including corruption), and responsiveness to popular preferences. Admittedly,
knowing which national actors are representative is difficult, and external actors
inevitably exercise their own judgment or preferences in selecting who will partic-
Conclusions for Policy and Practice 253

ipate in the determination of a participatory process. At the end of the day, peace
tends to trump justice, participation, accountability, and “good governance.”
Here external actors are generally supporting a more inclusionary peace pro-
cess of former combatants. Because peace agreements tend to privilege the secu-
rity and interests of the warring parties, promoting adherence to them generally
helps alleviate any uncertainty and insecurity that the armed parties may have.
International privileging of the warring parties’ interests over broader social de-
mands is understandable, given that concern in the first few years about rever-
sion to conflict is paramount.
Unfortunately, such privileging has a downside. Negotiated settlements often
do not include sufficient incorporation of civil society and the broader public,
not just in their preparation but also in their substantive terms. Third parties
need to balance the concerns of the potential spoilers in decisions about interim
authority against those of the broader population. One might think of this as
balancing horizontal legitimacy (to the extent that this derives from the interests
of the parties to an accord) with vertical legitimacy. From another angle, priori-
tizing the demands of the warring parties may be good for short-term stability
but may work against the long-term legitimation of state authority. Many peace
processes reflect a marginalization of civil society participation and demands,
such as Liberia’s peace agreements of 1996 and 2003, as well as Bosnia’s Day-
ton Accords.
One example of a peace agreement that successfully incorporated civil soci-
ety into the negotiations process and pacts was Guatemala. The multiple agree-
ments of 1995–96 included many of the issues and demands of civil society. The
weakness of the Guatemalan insurgents, who had been strategically defeated by
the mid-1980s, provided the permissive condition for civil society to have such a
weighty role in the negotiations with the government. Many civil society actors
enjoyed greater internal and external legitimacy than the rebels.
Unfortunately, the peace process in Guatemala is regarded as flawed in the
extent to which civil society’s interests actually were met (Stanley and Holiday
2002). Most of the important elements of the peace agreements went unful-
filled. However, the flaw was not in a failure to emphasize the interests of the
parties sufficiently vis-à-vis the vertical legitimacy. Instead, the problem lay in
the unwillingness or inability of external actors to extract from Guatemalan en-
trenched interests enough concessions to put teeth into the peace agreement and
its implementation (Stanley and Holiday 2002; Prado 1998; Alvarez and Prado
2002). This case reflects the problem identified above of how difficult it is to
wrest concessions from undefeated factions and how military reality constrains
the range of options.
The other situation of selecting interim arrangements is where one side
emerges victorious. Here again the victorious domestic faction holds the cards
254 Implications for Theory and Practice

in fashioning an interim or permanent authority. And these victorious forces

tend to select an interim leader from the ranks of the leadership of the military
(or politico-military) ranks. Upon recognition (and at times relief) that national
actors have determined what the interim form of governance and leader shall
be, the impulse of international actors is often to seek to curb exclusionary be-
havior on the part of these actors. This tendency also conforms to the dictates
of strengthening the legitimacy of national actors and ensuring that they do
not alienate potential spoilers. As with coups, postwar interim leaders may re-
nege on initial promises to relinquish power, especially if no threat to continued
rule appears.

The Second Moment: State Design

The second moment concerns decisions about state design (e.g., federalism, re-
gional autonomy, the relative strength of the central state vis-à-vis society) and
adoption of regime norms, rules, and procedures. Some transitions, especially
where the ancien regime emerges victorious, involve no regime alterations. The
old rules of the political game persist, perhaps with very minor changes. The
least disruptive transitions were the Latin American transitions from authori-
tarianism—not war—that involved only lifting the ban on Communist parties
alongside a de facto cessation of repression. Most contemporary transitions from
war, however, involve important alterations of the regime. Interim governments
are often tasked with rewriting the rules and procedures of governance. These, of
course, may be prescribed in a prior Constitution. Even where a new Constitu-
tion is adopted, the “old ways” of organizing the state tend to endure.
State design refers to where and how state powers are allocated or arranged.8
States can vary in different dimensions. First, they can distribute power over the
territory differentially. They can reflect federal or confederal systems whereby
the component provinces or entities wield the most power, and the internation-
ally recognized federal entity has more limited power than in centralized states.
They can also grant ethnic or social groups a degree of territorial autonomy in
a certain region. Second, state power can vary by levels of government over the
territory, such that local or provincial governments might hold more (or less)
budgetary and regulatory powers.
Finally, states vary in their degree of intrusiveness in social life, that is, in
the range of roles they play.9 Some states, such as the American and Swiss, are
relatively unintrusive, leaving services that states provide in many industrial
countries to the private or nongovernmental sector. The social expectations of
this latter dimension of state design are reflected in the cultural aspect of the
state, as mentioned above. In fact, all such arrangements shape popular expecta-
tions of the role of the state over time, as people come to expect the state to act
Conclusions for Policy and Practice 255

more or less intrusively, or to restrict its actions in certain arenas or geographical

territories.
State design may also distribute power across different branches of govern-
ment (e.g., legislatures vs. executives). In this sense, it is difficult to distinguish
state configurations from regime configurations. However, the allocation of state
power refers more to territory and levels of power, whereas regimes vary more
with regard to the relative powers of different branches of government and the
selection criteria for rulers.
The extraordinary circumstances of war and its termination both compel and
provide windows of opportunity for national elites (or external powers) to revisit
the state’s organization, for better or worse.10 The design of the state is often the
most crucial element of a negotiated settlement of a separatist struggle.
In addition, state design issues are linked to questions of extrastate author-
ity—both “below” the state with regard to local or tribal authorities and “above”
the state in relation to regional or transnational arrangements that usurp some
state functions (e.g., the European Union and World Trade Organization) (Reno
2008; Van den Brande et al. 2008; Call 2008b). Governance can be achieved
without state institutions, something that discussions of state design should take
into account. It is virtually impossible for decisions about state structures not to
favor certain social or political groups, as well as the process of the selection of
rulers, thus influencing regime type.11
How can international actors advance legitimacy via state design? First and
foremost, they can avoid appearing to control or engineer the process of state
design. The perception of colonial or imperial interference in domestic matters
has always elicited resentment and often armed resistance. Nothing is more cen-
tral to a political community’s identity than how it achieves expression in, or is
suppressed by, state power. The American policies in post-Saddam Iraq offer a
recent example of the perils of foreign control of state design, but direct foreign
occupation is not necessary. Even supposedly benign, UN-endorsed multilateral
efforts to shape political institutions can spark umbrage and resistance, as seen in
Bosnia, Kosovo, East Timor, Haiti, and Rwanda. In the current historical phase
of self-determination and anti-Western sentiment, fostering legitimate national
authority requires deference to national processes and actors, even while seeking
to ensure that these do not lead to illegitimate outcomes.
Second, external actors can seek measures that provide checks on exclusion-
ary behavior. Where such power-dividing mechanisms as multiple levels of au-
thority (especially the decentralization of power, revenue, and expenditures) and
checks and balances are possible, then such measures are likely to reduce the
scope for abuse. During the Salvadoran war, for instance, the government ac-
ceded to international pressure and introduced the election of mayors and may-
oral councils, as well as the dedication of certain state funds for municipalities
256 Implications for Theory and Practice

(Tulchin and Bland 1992). Once the peace process concluded in 1994, the leftist
guerrillas initially enjoyed electoral success only at the local level and as a mi-
nority party in the legislature, through proportional representation. Given that
the former guerrilla parties only won the presidency in 2009, seventeen long
years after laying down their arms, some analysts believe that decentralization in
El Salvador was crucial to keeping the former insurgents from resuming armed
conflict (Tulchin and Bland 1992). Lyons (2002, 222) laments that too many
countries concentrate power and resources in the presidency rather than decen-
tralized systems. Because of the potential for patronage and abuse, mechanisms
of accountability and transparency at lower levels of governance are just as ur-
gent as at the national level.
To the extent that governance may be distributed at different levels above the
state, including allocating functions to the suprastate level, then dividing power
may also help prevent war recurrence. The most salient success of such measures
has been achieved in Europe. The allure of accession to the European Union not
only elicited the institutionalization of improved human rights mechanisms in
the Balkans and parts of Eastern Europe but also served as an incentive for more
inclusionary state design and behavior. The Organization of American States and
the African Union, as well as subregional African organizations, offer the most
promising avenues for the creative design of legitimate authority at the supra-
state level.
However, as indicated above, the willingness of armed factions to diffuse
power and create checks on state power may be limited. The guarantees afforded
by powersharing may be the best (or only) option for convincing armed factions
to cease their armed struggle. Similarly, electoral systems that seem to smooth
the path to stability may be necessary not just to bridge a transition but also to ex-
ist into the future in order to alleviate uncertainties in the short term. Two skep-
tics of proportional representation for its reifying effects in the long term, Reilly
and Reynolds (1999, 29–31), nevertheless acknowledge that “large-scale [pub-
lic relations] appears to be an effective instrument” for smoothing democratic
transitions. To the extent that those deciding on both interim and permanent
structures can build in mechanisms to alleviate the weaknesses of powersharing,
then all the better. South Africa, El Salvador, Liberia, and other countries have
placed sunset provisions on their powersharing mechanisms with some success
thus far. However, there have been too few cases to draw firm conclusions about
adopting interim mechanisms of powersharing that end in winner-take-all sys-
tems or decision points. For instance, once the ten-year phase before a vote on
independence takes place in Southern Sudan, armed conflict may well resume,
just as the protracted negotiations on Kosovo’s final status have not resulted in a
resolution embraced by all parties.
Conclusions for Policy and Practice 257

The Third Moment: Elections and the End of Interim Governance

The third moment concerns the decisions about how an interim government is
terminated. This moment generally involves the convening of local and national
elections, sometimes through alternative consultative processes. Analysts have
devoted considerable attention to the role of elections in postconflict countries,
often in the context of the relationship between democratization and peace-
building.12 An interim arrangement may also end with a coup d’état or with a
resumption of warfare. In many cases interim arrangements are extended with
the claim that the country is not ready—because of political, security, or techni-
cal preparedness factors—for a national election.

Putting Elections in Context

As was made plain above, international scholars and policymakers have lent
elections excessive emphasis in postconflict legitimation strategies. Elections
are fraught with the problem of sparking renewed conflict, and they offer a po-
tential flashpoint for violence and duplicitous politics, along with negotiation
(Collier 2009; Mansfield and Snyder 2005). Of course, elections are not the
sole mechanism for consultation. However, for permanent national administra-
tions, it is difficult to imagine a selection process that does not involve some sort
of nationwide election. It may be, as in East Timor, that a national president or
head of government is named by an elected parliament or other body. Or it may
be that some parliamentarians are appointed as a result of a powersharing agree-
ment or other arrangement.
Yet external legitimacy is likely to be withheld without some form of national
election whose winners represent or are responsible for naming the head of
government and / or head of state. Moreover, internal legitimacy increasingly re-
quires elections as ratification of a political regime. The level, scope, timing, and
sequencing of elections are variable. But in any internationally assisted transition
from warfare, internal critics will seek to undermine a regime’s legitimacy if some
permanent political authority is initiated without any election process whatso-
ever. Some type of national election is therefore necessary to signal the end of an
interim process and the initiation of a political regime.
Nor do electoral processes need to monopolize or terminate informal or tra-
ditional means of consultation by elite representatives or the transitional state,
on the one hand, and with the broader population, on the other hand. Con-
tinued involvement with and participation in governance by society, through
context-specific mechanisms at the local or provincial level, can be important.
Consultative consensual selection of local leaders may also persist as a means of
state legitimation. Society may participate in the formulation and debates over
258 Implications for Theory and Practice

policy in specific functional areas of the state, such as health, education, security,
hunger response, disaster preparedness, and justice.
Before turning to some of the salient experiences and their implications for
the electoral process in transitioning to a long-term new political authority, it
is important to recall that election day itself is usually a small piece of the elec-
toral process. Because of heightened domestic scrutiny, security presence, and
international presence, violence is often much less than in the preparatory or the
postelection period. One study of electoral violence in East Timor, for instance,
showed that election day was the least violent phase of the electoral cycle, from
preparation to the campaign through to the counting and immediate postcount-
ing periods (Timor-Leste Armed Violence Assessment 2009).
As with all electoral processes, the preparation and campaign season is crucial
in whether the process will be seen as a free, fair, and accurate indicator of popu-
lar preference. The politics of that preparation, including the electoral tribunal
or commission tasked with overseeing the processing and tabulation of results,
will reflect the most divisive and serious threats to political coexistence among
the main parties. Too often, transitional governments have monopolized control
over electoral tribunals, undermining the credibility of the process, not to men-
tion the institutionalization, of open competitive elections.

The Importance of Elections

However, elections remain extremely important for several reasons. Most ob-
vious is their value for democracy broadly, because they represent a means by
which social and political interests and groups can resort not to coups or vio-
lence but to nonviolent processes in the quest and exercise of power. They do
so (assuming they operate as stipulated by the Constitution or convention) by
providing a future date by which the possibility of a change in government will
occur, and by generally creating uncertainty about who shall emerge victorious.
A surprising number of rulers (e.g., Chile’s Pinochet and Nicaragua’s Ortega in
1990) have convened elections confident that their tenure will be ratified, only
to be turned out of office. In terms of uncertainty, elections are the gift that keeps
on giving. In democracies generally, elections also offer a means for popular par-
ticipation in governance.
In transitions from war, elections can play additional roles. First, in theory
they offer the chance to settle control over the state, often the chief aim of armed
struggle, without violence. Of course, elections often spark armed conflict, as we
have seen. At a minimum, elections indicate where popular preferences lie with
regard to the programs offered by former armed factions. Second, elections also
offer an inducement for parties to adhere to the peace process during an interim
phase. For instance, demobilization, disarmament, and reintegration processes
have a different value if the hope of political representation lies at their termi-
Conclusions for Policy and Practice 259

nus. Third, and on the downside, elections offer the possibility of definitively
excluding losing parties. The tenure of elected officials marks a period of cer-
tainty about who shall rule, clarifying who has lost and for how long. The end of
uncertainty here offers disincentives for engagement with the political process
in the short term after elections.
Finally, elections also offer flash points for international engagement. Most
saliently, they offer external actors the chance, in the name of newfound sov-
ereignty rooted in the popular will, to defer to legitimately elected national au-
thorities by reducing their financial, troop, and other commitments. Insofar as
elections become the exit strategy for international actors, they earn the resent-
ment of many internal and external actors that believe the continued engage-
ment and resources are necessary to conclude a positive transition from war. In
addition, they offer a somewhat arbitrary, and often misleading, marker for the
end of “transitional” economic, security, political, or humanitarian processes. As
such, elections figure into timetables, often in ways that distort context-based
assessments of how well internationally supported internal processes have met
needs and expectations.

Electoral Experiences Worth Considering

What do cases of war recurrence and nonrecurrence tell us about decisions on
elections and peacebuilding? That literature tends to focus on two issues: elec-
toral systems, and the sequencing and timing of elections vis-à-vis other pro-
cesses (Lyons 2005; Bjornlund, Cowan, and Gallery 2006; Sisk 2009).13 Hav-
ing addressed the conceptual issues underlying debates about electoral systems
in the discussion of powersharing above, here I focus on issues of sequencing
and timing.
Demobilization before national elections: the Angolan experience. One of the
cardinal lessons of peacebuilding in the past two decades is that demobilization
must precede national elections, as emblemized by Angola’s failed peace process.
The Angolan experience included not only the 1992 election described earlier
but also a powersharing agreement that would forge a new army out of the prior
government army and the UNITA rebels. Hedging his bets, UNITA leader Jonas
Savimbi held back thousands of his troops and his arms until the outcome of
the election—his own loss—was apparent and produced a bloodier phase of
the war. A similar powersharing agreement that was reached in 1994 failed again
when some 30,000 UNITA troops were held back from a demobilization and
disarmament process.
As a result of the Angolan experience, international organizations have sought
full demobilization before a national election. Although rebel groups rarely de-
mobilize and disarm totally during a transition, a much more substantial demo-
bilization than that allowed by Savimbi is necessary to make it difficult to return
260 Implications for Theory and Practice

to warfare and to send a signal of commitment to the other side and to one’s
own militants that political contestation is replacing the battlefield as the locus
of struggle. International actors have sought to infuse peace agreements with this
sequencing, with a good degree of success.
Delayed elections: the Bosnian experience. A second generalization is central
to the state-building approach to peacebuilding, especially Paris’s “institutional-
ization before liberalization”: delaying national elections until the institutional
skeleton is strong enough to ensure their validity, acceptance, and translation
into effective governance. Here the Angolan experience was also informed by
that of Bosnia, where the outcome of national elections surprised international
actors by dealing a blow to their peacebuilding project.
Bosnia’s Dayton Agreement of 1995 “put the first round of elections on an
exceptionally tight schedule,” sparking heated debate (Cousens 2002, 554). It
provided for national elections within nine months of the accord, permitting but
not requiring local elections. The Bosnian experience elicited a critique of quick
elections not only for the technical difficulties but also for the security and politi-
cal problems posed by such a rushed timetable. Despite the insecurity perceived
by many voters and the technical challenges in mounting registration and bal-
loting, the elections took place in September 1996. As many feared, they served
only to legitimate the nationalist wartime leaders of their respective ethnic con-
stituencies. Elites who consented to Dayton under duress and amid warfare now
could impede its implementation with the imprimatur of being democratically
elected. The nationalist parties triumphed not just at the presidential level but
also in all three entity-level parliamentary bodies (Cousens 2002, n57). The
moderates made slight gains in four subsequent elections before winning sub-
stantial power in the 2000 elections, which opened the door to progress on more
provisions of the Dayton Agreements. Although the nationalists made gains in
the 2002 elections, the moderates once again surged in the 2006 voting.
One factor in the critique of prompt elections is the widespread perception
in the 1990s that quick timetables were driven not by context-specific factors
on the ground but by political and financial factors in Western capitals. In the
United States, which mediated the Dayton Agreements, President Bill Clinton
faced a reelection campaign in 1996 and pledged to withdraw troops by Decem-
ber. Thus the September elections became an urgent exit strategy for US troops.
Moreover, financial concerns about extended peacekeeping operations and their
costs have historically driven members of the UN Security Council to seek to
conclude operations promptly. Here again, elections became the sine qua non
not of democracy (as put by Huntington 1991, 9), but of a swift exit.
What are the problems of early elections? First and foremost, demobilization
and disarmament (or whatever degree of disarmament occurs, which is never
100 percent) require enough time to proceed. Schedules for demobilization and
Conclusions for Policy and Practice 261

disarmament—not to mention reintegration, which usually takes a year or more

to implement—are often not met. Retention of troops is the ultimate guarantor
of security and survival for rebel armies, as well as for governments where they
are demobilizing. In addition, the process of transforming armies—their leaders,
their internal organization, their goals, and their organization throughout the
country—into political parties takes time.
Second, early elections provide a focal point for violence before other issues
have been settled and before the parties and the population have adjusted to
a cease-fire. An interim administration can make significant progress in imple-
menting a peace agreement and can reassure minorities that their rights and
opportunities will be respected. State institutions can show that they are func-
tioning at least minimally, instilling confidence in the process. In addition, wars
can reinforce identity rather than policy preference as the basis for voting, and a
prompt election can reinforce the built-in advantage that armed factions enjoy
by having the best-developed organizations already in place. Of course, lengthy
periods before elections may lead to violence because of a loss in confidence in
either the peace process (because agreements are unmet) or the electoral process
(because polls show a likely loss). Yet such cases are exceptional.
Third, early elections may disenfranchise one or more parties to a peace pro-
cess, leaving them less incentive to cooperate in a peace process or in democratic
politics. To the extent that elections lead to the winner holding all the levers of
power in a society, political marginalization may then prompt desperation and
violence. Elections, more than any other single event except disarmament, sig-
nify a gain in power for some and a loss for others. Thus it seems prudent to put
off elections until other gains of peace can be recognized and valued. Of course,
it may not be possible to delay elections indefinitely if some domestic actors
demand them. This type of situation unfolded in Iraq in 2005, when the Ameri-
cans sought to hold elections but the Shiite Ayatollah Ali al-Sistani insisted on
prompt elections that favored Shiites and reduced the leverage of the Americans
in Iraqi politics.
Local and legislative elections before presidential elections: Kosovo and East Timor.
If the dangers of concentrating power in the hands of a single party represent part
of the perils of elections, then one strategy has been to emphasize local and leg-
islative elections rather than the national election of a head of state. Rather than
a dominant party emerging as the sole winner (e.g., of a presidency), multiple
parties representing minorities can accede to electoral offices. Such sequencing
offers concrete political gains for former rebel forces and minority ethnic or reli-
gious groups. As these groups gain some measure of political power, even if it is
in a minority in parliament or a minority of the country’s districts, they become
more wedded to the political process and more aware that the presidency may
not signify power.
262 Implications for Theory and Practice

This process was followed in Kosovo, but with a relatively quick sequence of
elections. Under Kosovo’s constitutional framework, municipal elections were
held in 2000, only a year after Serb forces capitulated and withdrew in June 1999,
followed by territory-wide legislative elections in 2001. These elections were
organized by the Organization for Security and Cooperation in Europe under
the UN’s authority.14 The Kosovan Assembly’s 200 members are elected for a
three-year term, and half of them are elected by proportional representation and
20 are reserved for ethnic minorities, including 10 for Kosovan Serbs. Under the
framework, the Assembly names a president and prime minister, which it first did
in early 2002. Municipal elections and legislative elections have been held every
three years since then, without being interrupted by the declaration of indepen-
dence by the Assembly in February 2008. Kosovo’s Albanian community has ex-
perienced relative stability under this system that encourages a plethora of politi-
cal parties. However, the process has not satisfied the members of Serb and other
minority groups, who have been divided between participating and abstaining
from participation in elections, pending their desire for unification with Serbia.15
In East Timor, presidential elections were not seen to hold the potential
for destabilizing violence, because support for the revolutionary leadership of
Xanana Gusmao was widespread and evident. Nevertheless, elections for a con-
stituent assembly that would become a parliament preceded presidential elec-
tions. Elections of the eighty-eight-member body, seventy-five of whose seats
were by national proportional representation and the remaining in single-seat
districts, took place in August 2001, two years after Indonesia’s withdrawal.16 In
April 2002, Xanana Gusmao won 87 percent of the territory-wide direct vote
as president, the new Constitution was then finalized, and independence was
declared and internationally recognized in May 2002.
Local-level elections were also held. In 2000, a World Bank project organized
the very first electoral processes under the authority of the UN Transitional Ad-
ministration in East Timor, with the election of 6,400 hamlet-level village de-
velopment councils composed of a dozen or more men and women (equally
divided; see Ospina and Hohe 2001, chap. 4). Those councils were apart from
traditional councils of elders that had largely survived the violence. They were
given decision making over how to spend tens of thousands of dollars in grant
monies for local projects, and they operated in consultation with both the CNRT
informal administrators across the territory and the traditional elders. Then, af-
ter the first parliamentary and presidential elections, during a ten-month period
from 2004 to 2005, village (suco) leaders and councils were elected and exer-
cised control over some development resources. There were few violent inci-
dents, and these were isolated in the 2001, 2002, and 2004–5 elections, but the
parliamentary elections of 2007 resulted in two deaths, dozens wounded, and
Conclusions for Policy and Practice 263

7,000 families displaced.17 President Jose Ramos-Horta was also elected in a run-
off second-round election in 2007.
As in Kosovo, despite the swift holding of elections, analysts consider the
fact that parliamentary elections preceded presidential elections helpful in East
Timor’s political transition. Holding parliamentary elections is a way to distrib-
ute power early, to emphasize centers of power other than the presidency, and
to offer representation to multiple and minority parties. In each of these cases,
both a president and a prime minister held important executive powers, and pro-
portional representation served as the main basis for parliamentary elections to
a single chamber. Decentralization, largely at the instance of international actors,
resulted in some important gains in control over resources at the local level, fur-
ther lowering the stakes of a presidential victory. This approach seems promis-
ing for fostering legitimacy without a recurrence to warfare; however, in some
contexts, local elections could well aggravate tensions, threaten a peace process,
or spark violence.
Maneuvering to favor the “right” candidate: Liberia in 1997. One controversial
aspect of international strategies in postwar elections is their role in shaping
which candidates compete and which ones emerge victorious in postwar elec-
tions. What reasonable conclusions, rooted in past experiences, can one draw
about such an external role? From the perspective of the long-term legitimation
of elections as instruments of domestic legitimation, external interference would
seem self-defeating. This is true as it relates to a specific country and in terms of
the general utility of elections. Nothing crushes the ideals of young supporters
of self-determination and democratic processes like seeing powerful countries
manipulate such processes in ways that pervert the will of the populace, espe-
cially when they get away with it. Such external manipulation applies not just to
self-interested foreign countries (like former colonial powers) but also to inter-
national organizations that claim the mantle of neutrality and popular empower-
ment yet still favor certain candidates.
Even where a candidate may enjoy popular support, the perception that his or
her power derives from external support or endorsement may be self-defeating.
Because of the inherent contradiction discussed above, external support as an
instrument of legitimation can backfire and thus undermine internal support or
legitimacy. The American experience in Iraq exemplifies the backlash against
the leadership proposed initially by the George W. Bush administration. That
group of exiles, whose highest-profile member was the controversial Ahmed
Chalabi, enjoyed little support among Sunnis or Shiites who had endured the
regime of Saddam Hussein (Diamond 2005, 28–33). The growing resentment
during 2003 of the extended American-led occupation further eroded support
for this group.
264 Implications for Theory and Practice

At the same time, it is tempting and sometimes sensible for external actors,
whether they are states or international organizations, to seek to influence the
selection or performance of particular candidates. Most obviously, when candi-
dates widely known for their human rights abuses or atrocities run for office in
the wake of warfare, international organizations have a principled obligation to
withhold their support. The United Nations now is precluded from lending its
support for the candidacy of persons accused of war crimes, and most interna-
tional organizations now shy away from lending support to such candidates, even
where these candidates benefit from externally funded peace processes involv-
ing elections.
Nevertheless, from Cambodia to Afghanistan to Bosnia to Liberia and be-
yond, the United Nations and regional bodies have supported both transi-
tional and permanent governments that have included those implicated in war
crimes. That is because peace processes so frequently include commanders or
warlords accused of international war crimes in interim powersharing or gov-
ernance roles.18 The justification of diplomats who negotiate and support such
peace agreements is that an outright ban on dealing with thugs and human rights
violators would preclude the possibility of a cease-fire and peace. With a peace
process, the mechanisms of justice—whether domestic, third-party, or interna-
tional—can begin to function and bring abusers to account for their crimes.
Even in peace processes, international actors may have good reason to sup-
port one candidate over another. Liberia offers an example. As was discussed in
chapter 3, international actors felt compelled to support a peace agreement that
offered a chance for success after more than a dozen failures in 1996. Charles
Taylor was the leading candidate in opinion polls, and no other candidate ap-
peared to enjoy much electoral support. And yet by supporting that process,
rather than delaying elections and seeking avenues to undermine Taylor’s repu-
tation, the international community implicitly endorsed his candidacy and was
implicated in his administration after the 1997 election. Taylor’s past record and
subsequent corrupt and exclusionary rule, which led to renewed warfare, suggest
that international actors might in some circumstances want to seek to weigh in
against a candidate, especially one accused of war crimes.
Powerful countries have long sought to influence electoral processes in less
powerful countries, even as the former profess commitment to open, democratic
processes as expressions of the will of the people of the latter. Of course, how
international actors express their favoritism shapes the extent to which it is ef-
fective. Where external favoritism takes the form of recruiting candidates who
might otherwise appear reluctant, and where much of the campaign appears
overtly funded by external actors, the outcome can be the opposite of the desired
effect. This is especially true for candidates who are perceived to be solely the na-
tive tool of a military occupation, as emblemized by Russia-occupied Chechnya’s
Conclusions for Policy and Practice 265

fixed elections in 2003, whereby the Russian-appointed candidate won, only to

be assassinated by separatists the following year. In this era of strong national
identity and sovereignty, the backlash against the idea of outside interference is
virtually universal.
In general, the most prudent measure for outside actors confronted with an
electoral process that is likely to result in the election of a spoiler is not to seek
to favor one or more candidates over others but to shape the timing and circum-
stances of the elections in a more evenhanded manner. This may mean urging a
delay, seeking more consultation internally in deciding on the electoral format
and timetable, or exploring other options that might offer the chance to alter the
electoral dynamics.
In the contemporary climate, where sensitivity to outside meddling in elec-
tions is extremely high, even the slightest form of influence may be perceived
as undue meddling. For instance, statements that aid or investment might be
endangered (or, conversely, that more might be forthcoming) if a particular can-
didate wins are unmistakable signals to the electorate, which are often resented
but perhaps are effective in shifting support in the desired direction. Threats of a
lack of external recognition (or even of external support for armed resistance) if a
particular candidate emerges victorious may also distort the democratic process
in the desired direction and complicate the legitimation of the eventual victor.
Often the international community’s preference for a candidate is apparent
without any visible demonstration. The background of the particular candidate—
for instance, that he or she has worked abroad for a multinational corporation
or intergovernmental organization—will suggest an affinity with specific inter-
national interests. In cases like Chechnya (Russia) or Afghanistan, the pertinent
relationships lie elsewhere, with neighboring powers. More important, the stated
policies of such candidates make clear their likely affinity with outside actors.
Such a convergence of external preferences favoring a candidate who com-
mands support internally would seem a happy coincidence. Few postconflict
cases have represented such a candidate. Some who might fall into this cate-
gory include the rightist presidential candidate of the ruling party in El Salva-
dor in 1994, Calderon Sol; Xanana Gusmao of East Timor, in 2002; and Ellen
Johnson- Sirleaf of Liberia, in 2005. Many candidates like these have enjoyed
success. In other cases, candidates who specifically are not favored by most in-
ternational actors have emerged triumphant. Examples include the nationalist
parties in Bosnia’s first election after the Dayton Agreements, and Hamas’s 2006
victory in Palestine.
Unfortunately, the perception of an implicit or explicit international prefer-
ence for a candidate may have perverse consequences for the consolidation of
peace or democracy, or both. Consider the example of Hamid Karzai in Afghan-
istan. The sort of candidate favored by international actors is often one more
266 Implications for Theory and Practice

beholden to foreign interests than national interests, and whose disposition is

more that of an administrator than a visionary political leader capable of forg-
ing nationhood and statehood (Thier 2002). Karzai was not a military com-
mander, and he represented a compromise candidate who was largely brought
to prominence because of international preferences. He was reliant on warlords
for the main elements of the new national army, and was without control over
the international security forces, so he faced limits on his ability to extract re-
sources from provincial leaders and weaken political rivals who sought a weak
central state (US Department of Defense 2009). A plethora of factors explain
the Afghan state’s inability to provide stability and development during the past
decade. However, one factor is this peculiar affinity between reliance on interna-
tional political support and lukewarm leadership capacity and internal support.

The Fourth Moment: State Legitimacy after Elections

The fourth and final, and least analyzed, moment of engaging state legitimacy oc-
curs after the first national election that brings an end to a transitional authority
(Sisk 2009; Wallensteen 2008). This postelectoral period is especially important
in postwar countries because of the intensity of loyalties and identities that are
fostered by armed conflict. The work involved in mobilizing people to risk their
lives requires creating intense political and / or identity-based loyalties. Wars re-
quire an “other,” especially civil wars among people who have lived within the
same state for decades or more. Creating the “us” versus “them” of armed conflict
involves loyalties that are difficult to dismantle. A transitional period is generally
insufficient to see these loyalties and preferences, much less identities, change.
As a consequence, one common notion is illusory: that a national presidential
election marks the end of a transition from warfare as well as the beginning of
a democracy that resembles other rocky, new, and poorly institutionalized de-
mocracies that will “mature” over time. Although social groups in non-war-torn
societies grappling with legitimacy crises are often deeply divided to the point of
potential armed conflict (consider Zimbabwe, as of this writing), the dynamics
and challenges of war-torn societies are more acute. People are quicker to take
up arms, even though war weariness has taken its toll. Collier and others argue
that postwar countries face a higher risk of civil war, even though the degree of
that risk remains in dispute. Although elections offer the important possibility
of an alternation of power, these intense affiliations require attention beyond an
electoral event every four to six years.
In many ways, the postelectoral period is the most dangerous for exclusionary
behavior. The trappings of sovereignty confer a crucial role on national elections
as the main marker for a return to “normal” status among bilateral, regional, and
intergovernmental actors outside the country. UN agencies, international finan-
Conclusions for Policy and Practice 267

cial institutions, and donors tend to return to a mode of treating the state as its
sole and most important partner, the legitimate voice and representative of the
territory’s entire society. The “development” moment has arrived, closing the
“peacekeeping” (and “peacebuilding”) moment for too many actors. And devel-
opment agencies and nongovernmental organizations (NGOs) are loath to jeop-
ardize their good relations with the host government, which they now can em-
brace as their crucial partner and counterpart. Even where international actors
embrace development projects aimed at easing tensions between former warring
groups, at enhancing the sense of belonging of those marginalized groups, and
at dissuading the state from behaving in an exclusionary manner, the mentality
underlying the planning and consultation by external actors prioritizes the state
to the detriment of those marginalized sectors.
In postwar Sierra Leone, for instance, leaders of the opposition party ex-
pressed their surprise that after their electoral loss in 2002, international embas-
sies’ contacts and invitations to meetings dropped off precipitously. They were
left in the cold in terms not only of domestic governance but also of shaping
the views and business of the United Nations, bilateral actors, and international
NGOs. These external actors concentrated on their counterpart—the state—
and organized their strategies, planning processes, and implementation regimens
with the state, leaving the political parties that emerged from the conflict to sort
out their own relations with one another. When the opposition gained power in
the 2007 elections, its leaders felt once again besieged by the flood of attention
from international organizations.
Of course external actors confront a dicey challenge in seeking to engage op-
position or minority parties after a recognized national election has occurred.
There is no set template or protocol for international actors to maintain rela-
tions, dialogue, or involvement with opposition or minority parties. After all, the
former insurgent forces (or state armies) are no longer armies that merit atten-
tion as security threats from international forces. Instead, they have presumably
become political parties subordinating themselves to the rules of the political
regime, not above the power of the state. Elected governments are likely to be
uneasy or threatened when external actors try to involve former warring parties
or other, third parties in governance—whether in dialogue over state priorities,
budgetary questions, planning processes, sector specifics, implementation, or
the creation or modification of mechanisms for accountability and transparency.
They will quickly decry outside interference and hark back to state sovereignty.
But it is precisely in this phase when external actors must seek to retain a
peace-oriented engagement and thus prioritize the prevention of exclusion-
ary policies and practices. Speaking of postconflict societies, Peter Wallensteen
(2008, 238) argues for “long-term readiness for the international community to
remain engaged, for the benefit of democracy and peace.” As described above,
268 Implications for Theory and Practice

this is extremely challenging, especially once elected governments can claim the
mantle of sovereignty and legitimation via elections. However, there are a num-
ber of means to overcome the barriers to emphasis on legitimation via inclusion-
ary processes.
The groundwork for such a role to be played by international actors must
be laid before an election is held for a posttransition head of state. Where UN
Security Council mandates for verification and monitoring of a peace agreement
exist, an extension of this mandate through the tenure of the first elected head of
state is one means of providing the legal and political basis for continued interna-
tional verification and monitoring. Because external monitoring roles generally
offer informal leverage over the parties to the accord, some means for influencing
elite and governmental behavior persist, especially where foreign troops remain
present. The practice of extended peace operations with a gradually reduced
peacekeeping troop presence seems to have provided useful leverage to the UN
missions in Sierra Leone and Liberia, as well as to international efforts in Bosnia
and Kosovo.
In addition, a newly elected government can extend transitional powersharing
or other co-governance mechanisms past the swearing-in of a government to
continue providing voice to those who potentially could be marginalized. Such
groups may include the armed groups that did not emerge triumphant; the eth-
nic, religious, or social groups associated with those armed groups; and vulner-
able minority groups that were not represented by armed groups or in a peace
deal. The inclusion of other marginalized groups (e.g., representing women,
moderate third parties, and ethnic groups that no army claimed to represent) can
help extend horizontal legitimacy. Extended built-in consultation and voice (one
can think of them as “extended transitional” measures) may, of course, be embed-
ded in a political or electoral system in one of the powersharing mechanisms
described above. Powersharing brings the attendant benefits and limitations of
an institutionalized role for particular social groups in the state and governance.
Conversely, extended consultation mechanisms could be fashioned which
represent temporary or ad hoc arrangements that expire either on a specific date
or after certain predetermined events have occurred (e.g., the transfer of power
from one party to another). Here, I call such arrangements that extend past the
swearing-in of a nontransitional elected government “extended transitional”
measures. Rather than the result of renegotiation processes or Pierre DuToit’s
(2003) “postsettlement settlements,” these are negotiated earlier and persist
beyond elections. Because elected governments are generally reluctant to cede
power immediately after having won it at the polls, extended transitional mea-
sures are unlikely to emerge unless a prior agreement has been reached on them.
One of the most promising areas for external actors to support inclusionary
peacebuilding is in fostering such agreements before the beginning of the cam-
Conclusions for Policy and Practice 269

paign season for elections of a government under a permanent regime. Nego-

tiating such extended transitional arrangements as part of a peace agreement
will build in a greater acceptance and expectation of this guarantee. Possible ex-
amples include a preagreed temporary role for former guerrilla forces in deter-
mining how state resources are spent in particular regions of the country where
insurgents enjoyed strength. They include temporary powersharing measures
like quotas for an initial hiring of the security forces or for seats in the legisla-
ture, or consultation with other political parties in major decisions on things
like natural resources, declaration of an emergency, or defense expenditures.
Equally important, the international financial institutions, especially the World
Bank, can emphasize the inclusion of former enemies and other political repre-
sentatives in the project design, assessment, and execution phases and structure
them to persist through an initial postelection period, though not permanently.
World Development Report 2011 (World Bank 2011, 285) points also toward a
role for external donors in pressing for greater budgetary oversight and transpar-
ency from postconflict states as a means to curb transnational crime and other
obstacles to stability.

Civil Society
Beyond the inclusion of political parties and former insurgents, civil society or-
ganizations offer the possibility of advancing broad participation and shunting
exclusion. Here the concerns of critical peacebuilding theorists converge with a
legitimacy-focused approach that includes both elite inclusion and broader mass
inclusionary practices. Civil society is already incorporated into the formulation
of the main national economic planning document used by the World Bank,
the Poverty Reduction Strategy Paper. Projects undertaken by the international
financial institutions also routinely consult with NGOs and other civil society
elements. Donors channel many of their funds through national-level NGOs,
although more pass through UN agencies and international NGOs.
However, consultation of civil society tends to be project centered and not
provide the sort of broad participation and accountability that are useful to re-
duce exclusionary governance. Haitian NGOs, for instance, complained even
after participating in consultations all over the country that their views were
largely ignored when the government and international technocrats drafted the
Interim Cooperation Framework in 2004.19 It is rare that civil society groups
are formally part of state decision making about funding priorities or about re-
source management and allocation in ways that affect specific regions and eth-
nic groups. Externally supported state reforms in specific sectors, such as justice
reform and education reform, have increasingly included civil society represen-
tatives. Yet civil society has not generally substituted for the sense of participa-
tion and ownership that comes with powersharing arrangements. Furthermore,
270 Implications for Theory and Practice

governments are adept at co-opting civil society organizations, converting them

from agents of oversight, accountability, and diverse types of participation
into fig leaves for advancing the interests of the governing party or dominant
ethnic group.

“Layered” Legitimacy Building

The manner in which international actors operate can shape their ability to pre-
vent exclusionary practices and to overcome the inherent contradictions of ex-
ternal efforts to enhance internal legitimacy. Chief among these is the degree
to which individual countries act in a unilateral as opposed to multilateral way.
Despite the stated commitment of many donors like the United States and the
United Kingdom to multilateralism, advances have been slow and difficult. Uni-
lateral action persists, especially in the policies of donor development agencies
and occasionally in military operations.
Unilateral action persists for several compelling reasons. Every international
actor has divergent interests, making it rare for the objectives of particular poli-
cies or projects to fully coincide. Donor countries, which are mostly democratic,
are also held accountable by the public and by legislative bodies for their devel-
opment aid and foreign engagements. They need to exercise sufficient control
over their programs to achieve their objectives. In addition, the complex logistics
of planning, executing, and funding projects becomes even more complicated
when done in concert with other actors. Finally, bilateral actors often feel that
other donors, NGOs, and regional and international organizations do not en-
gage in peacebuilding or development projects as effectively or accountably as
they do. As a result, they often eschew harmonized approaches because of the
corruption, inefficiencies, slow pace, or high transaction or administrative costs
(read: headaches) of working with others.
If, however, the legitimacy of both international efforts and internal state in-
stitutions lies at the core of the peacebuilding process, then the importance of
perceptions of the motives and modes of external interventions is paramount.
Many bilateral actors have largely discounted the benefits of multilateral har-
monized efforts for this legitimation process (e.g., as seen in local perceptions).
These advantages must be weighed against the transaction costs, inefficiencies,
and risks of concerted multilateral efforts. Acting in concert with others mini-
mizes two of the hurdles to legitimacy building identified above: (1) the per-
ceived Western bias and interests underlying these programs, and (2) the notion
that external actors act in their own interests at the expense of the interests of the
population of a postwar society.
Fortunately, the twenty-first century has seen a serious weakening of unilat-
eralism in postconflict peacebuilding. Bilateral and other international actors
can deepen this multilateralism in three ways, which I call “layered” legitimacy
Conclusions for Policy and Practice 271

building, that will enhance efforts to build legitimacy. The first of these layers is
enhanced coherence among bilateral donors. Even the most unilateral aid efforts
engage in coordination exercises and proclaim a commitment to coordinated
development assistance. The Organization for Economic Cooperation and De-
velopment’s Development Assistance Committee (OECD-DAC) was erected by
the main donors to help coordinate and harmonize aid efforts, and to develop
and advance their conformity with broadly agreed-on standards. In the area of
postconflict countries and fragile states, the OECD-DAC has developed a good
number of standards to enhance not just coherent development efforts among its
members but also mutual accountability with the host country.20
In terms of military action, unilateral interventions or wars remain the excep-
tion rather than the norm. Most international deployments of troops are under-
taken not by a sole power but in multilateral form. For example, even the epitome
of recent “going-it-alone” wars—the US war in Iraq—involved not just the United
States but also an assortment of ancillary coalition partners. Of course, greater co-
herence within each donor country in “whole-of-government” approaches (en-
compassing development, diplomacy, defense, and more) is also important for
the coherence of external efforts.
The second “layer” of legitimacy building is the United Nations system. Most
obviously, military efforts to end (or start) wars can conform to the UN Charter’s
demand that military actions beyond self-defense must be authorized by the UN
Security Council. Although most UN member states saw the 2003 US war on
Iraq as a violation of this legally enshrined principle, the US experience in that
country seems to have ironically produced a backlash in favor of the wisdom of
this international standard. Beyond providing authorization for military inter-
ventions into unconsenting countries, the UN Security Council also approves
political missions and postconflict peacebuilding support missions. Individual
countries can coordinate their diplomacy with these missions, can provide ma-
terial support to them, and can coordinate their funding priorities in concert
with other regional actors. When elected governments tend toward exclusion-
ary policies and practices, the most influential states can help curb this behavior
by coordinating conditionality and leverage with one another, with neighboring
states, and with the senior representatives of the United Nations and other re-
gional organizations.
The third and final layer of legitimacy building is composed of regional and
subregional organizations. The most important of these bodies are the African
Union, the Organization of American States, and the European Union, but other
groups like the Association of Southeast Asian Nations and the Organization of
the Islamic Conference also have potentially important roles. Regional organi-
zations are playing an increasingly important role in peace operations, includ-
ing postconflict peacebuilding. One reason is the desire of Western powers and
272 Implications for Theory and Practice

donors to foist off the financial and diplomatic responsibilities for international
support for these processes to others, especially in Africa. The African Union’s
deployment of troops to Burundi in 2003 and 7,000 troops to Darfur by 2005
reflects this trend as well as the union’s distrust of the UN and its interest in en-
hancing its stature and impact (Bah 2009).
A factor favoring the increased activities of regional and subregional organiza-
tions is the notion that these states can provide technical civilian assistance that
is more culturally appropriate than that provided by civilian staff members from
other regions. Judicial advisers, security advisers, and advisers for institutional
capacity building and demobilization, disarmament, and reintegration programs
will be more effective if they speak common languages and share common his-
tories. In places like Africa, such regional civilian assistance will presumably cost
less than if deployed from a wide range of wealthier UN countries.
However, a more compelling reason for providing an enhanced place for
regional organizations (and subregional ones like the Economic Community
of West African States, ECOWAS; the Intergovernmental Authority on Devel-
opment; and the Common Market for Eastern and Southern Africa) is the le-
gitimating function they can play. World Development Report 2011 (World Bank
2011, 288) urges “better aligning international assistance behind regional gover-
nance efforts” and thus points to the crucial legitimating role those organizations
can play. If curbing exclusionary behavior by elected governments is crucial for
peacebuilding success, then diplomatic leverage remains a key tool for discourag-
ing such behavior. The ability of Western governments to exercise this leverage
without sparking a backlash against outside interference is limited to the extent
that it smacks of either unilateralism or Western self-interest. Working with and
through the United Nations system, and its various organs and agencies, helps
mute this sort of objection. In early March 2011, the European Union and the
Obama administration both followed this approach to regional organizations’
role in legitimating international actions. In stating a position also embraced by
the United States, a European diplomat stated: “The position even of European
countries who theoretically support a no-fly zone is clear that it would require
the full support of the Arab League or African Union, and a UN Security Coun-
cil Resolution.”21
Of course, regional organizations themselves or their most influential mem-
bers are often contributors to the causes or persistence of warfare. In virtually all
the civil wars examined in this book—from the Central American wars of the
1980s to the Central African wars of the 1990s, and from the West African wars to
the Southeast Asian wars of the 2000s—meddlesome neighbors have spawned
or abetted war. Powerful regional states have a self-interested and perverse influ-
ence not only on what happens in neighboring states but also on the decisions
of regional and subregional organizations, which often act in pathological ways
Conclusions for Policy and Practice 273

(Barnett and Finnemore 2004). The dominant role of the United States in the Or-
ganization of American States and NATO, of Nigeria in ECOWAS, and of South
Africa in the Southern African Development Community emblemize this ten-
dency. Nigeria’s leadership of the Economic Community of West African States
in Liberia, for instance, was seen in the region as corrupt and advancing both the
strategic interests of Nigeria and the personal interests of the troops it deployed.
Nevertheless, nothing blunts the charges of Western interference more than
a regional or subregional organization taking the lead in diplomatic steps or even
adopting certain sanctions to lean on governments to eschew exclusionary prac-
tices. In this sense, regional and subregional organizations represent the gold
standard for the legitimate exercise of outside leverage. Regional institutions can
also embrace regional-level norms of good governance that signal the need to
conform to standards of inclusionary conduct. As the World Bank (2011, 288)
claims, “supporting regional platforms to discuss the application of governance
norms is an effective way to increase ownership.” The legitimacy given to such
measures does not ensure that they will achieve their aims, especially in Africa,
where states are notoriously reluctant to lean on one another. But it does remove
one of the principal shields behind which national governments hide when they
flout international standards of conduct: sovereignty. For these reasons, both
regional and subregional organizations are likely to play a much more salient role
in international affairs in the coming decades.

Conclusion
National-level leaders and organizations hold responsibility for making deci-
sions that can alienate and exclude salient groups from power. They can also cre-
ate state structures and adhere to them in practice so that these groups, including
potential spoilers, feel that their expectations of participation are being met. I
have offered here four different “moments” in which national leaders confront
different points in a postconflict transition to make such decisions. Many others
have written on specific aspects of these moments, including the design and con-
tent of interim administrations; the different constitutional, powersharing, and
other options in state design; strategies about the timing and systems of elections
that bring transitional periods to an end; and the postelection period that proved
so harmful for peace in Liberia, Haiti, East Timor, and other cases. In this chapter
I hope to have structured the contingent choices that confront both national and
international actors in these processes.
It is difficult for outside actors to respond to the imperative to enhance inclu-
sionary behavior and horizontal legitimacy. However, in this chapter I have set
forth some ways in which international actors can engage these processes con-
structively. Engaging processes that strengthen inclusion is more productive than
274 Implications for Theory and Practice

advocating specific structures. No one-size-fits-all blanket recommendations ap-

ply to all postwar societies. International actors can play important roles, using
leverage on aid and diplomatic strategically to support participatory processes.
Just as globalization can undermine efforts to build legitimacy, so it can also abet
the need for external legitimation and the role of external actors, including both
multilateral and nongovernmental actors.
Given the complex globalized context of peacebuilding, legitimation depends
not solely on elite inclusion but also on the inclusion of the broader population.
Although this book’s empirics point mainly to elite accommodation in inclusion-
ary practices, the shifted norms regarding popular voice and participation in pro-
cesses determining postwar polities require broader participation. Such broader
inclusion is imperative if the negative peace standard examined in this study is to
be transformed into at least a consolidated and self-sustaining peace, if not “posi-
tive peace.” And beyond their role in enhancing national horizontal authorities’
legitimacy, external actors must also pay heed to their own legitimacy as actors
in these societies (Rubin 2008). By working in conjunction with other external
actors, and ensuring meaningful partnerships with domestic actors and groups,
international actors can enhance their own effectiveness.
In seeking to fully explore the potential and perils of this book’s findings for
policymakers, this chapter may not appeal to the theoretically oriented reader.
However, scholars frequently invest inordinate time and effort in research that
points to particular policy pathways but fails to analyze the specific likely contri-
butions and pitfalls of those implicit recommendations. The policy implications
of an inclusionary peacebuilding approach are intimately tied to state legitimacy.
One implication of the forty-two cases examined in this study is that the be-
havior of a postwar national government—especially during the first years after
the election, which are seen to mark the end of a transition from war—is a mo-
ment of special danger. The most underrecognized moment of vulnerability to
conflict-sparking exclusionary behavior is when a new government is installed
after an internationally accepted election. At this moment, when the trappings of
sovereignty, the illusion of a terminated transition, and the exhaustion of external
actors are all ascendant, domestic actors count on international deference and
negligence to nudge out (or worse) their former enemies. The hour of celebrat-
ing an externally facilitated postwar transition is a time of heightened danger
of war recurrence. Rather than turning away their attention and resources after
the postelection moment, external actors should instead remain engaged with
national actors and do what is prudent to encourage inclusionary behavior. Al-
though national actors remain in the driver’s seat of processes that may spark
renewed conflict, international actors can influence this behavior through judi-
cious, farsighted, and careful diplomatic leverage. This road is a more prudent
one than assuming that electoral democracy will produce a responsive regime
Conclusions for Policy and Practice 275

and ultimately a legitimate and sustainable state. Peace and good governance
depend on numerous factors, but they are likely to depend heavily on sound
judgment and sustained engagement.

Notes
1. See Krasner (2001). Even so, the international system reflected greater coher-
ence, coverage, and comparability of states in the late 1800s and early 1900s than be-
fore or after.
2. Hegre et al. (2001) found that countries experiencing regime transitions of any
sort are more inclined to civil wars than stable democracies or autocracies.
3. See Tvedten (1997, 111); also see Duke (1998); Stedman (2002), plus UN esti-
mates of 450,000 to 500,000 dead between October 1992 and December 1993, includ-
ing secondary deaths due to starvation, given by SIPRI (1994).
4. Most regimes considered among the less democratic (e.g., Cuba, North Korea,
Turkmenistan, and Burma) hold elections, and some (e.g., Uzbekistan and Iran) in-
clude multiple parties. Two potential alternatives, Chinese-style single-party open
economy and Islamic republics, may attract additional adherents.
5. NATO’s continued alliance with warlords who had overseen the worst atrocities,
along with its refusal to risk its forces to detain alleged perpetrators, assured that transi-
tional justice would be on the back burner for many months, which became many long
years (and still pending as of mid-2010). In his departure speech as special representa-
tive of the secretary-general in January 2004, Brahimi rued this decision.
6. See “Iraq Body Count,” www.iraqbodycount.org.
7. Where human rights violators are part of an agreement and interim arrangement,
external actors may also seek prosecution for atrocities, or at least exclusion of said
violators from power. Outside powers also may seek to favor one side over another in
a peace process.
8. For various perspectives on what is here called state design and its relation to war
termination, see Elazar (1994), Burton and Higley (1998), Horowitz (1985), Bermeo
(2003), and Fukuyama (2004).
9. Fukuyama (2004, 6–14) refers to this as “scope” of the state.
10. See Call (2002). For an analogous view at the interstate systemic level, see Iken-
berry (2000).
11. On the well-trodden distinction between states, regimes, and governments, see
Lawson (1993); Eagles, Holoman, and Johnston (2004); O’Donnell, Schmitter, and
Whitehead (1986); and Fishman (1990).
12. See, e.g., Kumar (1998); Lyons (2002, 2004, 2005); Ottaway and Carothers
(2000); Jarstad and Sisk (2008); Diamond (2005); Doyle and Sambanis (2006); and
Paris and Sisk (2009).
13. The elections literature also addresses technical challenges of electoral observa-
tion and monitoring, which I do not take up here.
276 Implications for Theory and Practice

14. Organization for Security and Cooperation in Europe, “Kosovo Municipal Elec-
tions 2000–Final Results,” 2000. Reports from the Organization for Security and Co-
operation in Europe for municipal and assembly elections for these and other years can
be found at www.osce.org / kosovo / 13208.html.
15. See Kosovar Institute for Policy Research and Development (2009).
16. On the parliamentary process and results, see International Foundation for
Electoral Systems (2005).
17. Timor-Leste Armed Violence Assessment, “Elections Violence in Timor-Leste:
Mapping Incidents and Responses,” Issue Brief 3, June 2009.
18. Sometimes the most prominent candidates to emerge from the populace associ-
ated with the former rebels have not been military commanders but instead political
opposition figures not associated with armed action. Examples include the election of
Ibrahim Rugova as Kosovo’s first president in 2002 and Violeta Chamorro in Nicara-
gua in 1990.
19. See open letter from sixteen Haitian nongovernmental organizations, “Haitian
Civil Society Organisations’ Declaration on the Interim Cooperation Framework,”
June 14, 2004, stating that “the process of drafting the ICF is controlled by external ac-
tors with the complicity of the current government in the framework of a technocratic
approach. This excludes all real participation of the majority and vulnerable sectors of
our country” (para 1.3), www.grassrootsonline.org / node / 723.
20. See the 2005 Paris Declaration and the 2008 Accra Agenda for Action, both at
www.oecd.org / document / 18 / 0,3343,en_2649_3236398_35401554_1_1_1_1,00&
&en-USS_01DBC.html.
21. Anonymous quotation from a European “knowledgeable diplomat,” given
by James Kitfield, “The Libyan Crisis: Grounding No-Fly Talk,” National Journal,
March 10, 2011, www.cnas.org / node / 5963.
References

Adebajo, Adekeye. 2002a. Building Peace in West Africa: Liberia, Sierra Leone, and
Guinea-Bissau. Boulder, CO: Lynne Rienner.
———. 2002b. Liberia’s Civil War: Nigeria, ECOMOG, and Regional Security in West
Africa. Boulder, CO: Lynne Rienner.
Alao, Abiodun, John MacKinlay, and Funmi Olonisakin. 2000. Peacekeepers, Politicians,
and Warlords: The Liberian Peace Process. Tokyo: United Nations University Press.
Alexander, Jocelyn, and JoAnn McGregor. 2004. “War Stories: Guerrilla Narratives of
Zimbabwe’s Liberation War.” History Workshop Journal 57 (1): 79–100.
Ali-Dinar, Ali B. 1999. IRIN Update 448. Abdijan: United Nations Office for the Coor-
dination of Humanitarian Affairs Integrated Regional Information Network for West
Africa. www.africa.upenn.edu / Newsletters / irinw448.html.
Alvarez, Enrique, and Tania Palencia Prado. 2002. “Guatemala’s Peace Process: Context,
Analysis and Evaluation.” Accord 13:38–43.
Anderson, G. Norman. 1999. Sudan in Crisis: The Failure of Democracy. Gainesville: Uni-
versity Press of Florida.
Arnson, Cynthia J. 2005. “The Evolution of Internal War in Peru: The Conjunction of
Need, Creed and Organizational Finance.” In Rethinking the Economics of War, edited
by Cynthia J. Arnson and I. William Zartman. Baltimore: John Hopkins University
Press.
Ashour, Omar. 2004. “Security, Oil, Internal Politics: The Causes of the Russo-Chechen
Conflicts.” Studies in Conflict and Terrorism 27 (2): 127.
Assefa, Hizkias. 1987. Mediation of Civil Wars: Approaches and Strategies—The Sudan
Conflict. Boulder, CO: Westview Press.
Atkins, G. Pope, and Larman Wilson. 1998. The Dominican Republic and the United States
Athens: University of Georgia Press.
Atkinson, Philippa. 1997. The War Economy in Liberia: A Political Analysis. London:
Overseas Development Institute.
Atlas, Pierre M., and Roy Licklider. 1999. “Conflict among Former Allies after Civil War
Settlement: Sudan, Zimbabwe, Chad, and Lebanon.” Journal of Peace Research 36 (1):
35–54.
Autesserre, Severine. 2010. The Trouble with the Congo: Local Violence and the Failure of
International Peacebuilding. Cambridge: Cambridge University Press.
278 References

Baev, Pavel K. 2005. “Chechnya and the Russian Military: A War Too Far?” In Chechnya:
From Past to Future, edited by Richard Sakwa. London: Anthem.
Baev, Pavel, Jan Koehler, and Christoph Zurcher. 2005. “Civil Wars in the Caucasus.” In
Understanding Civil War: Evidence and Analysis, Volume 2: Europe, Central Asia, and
Other Regions, edited by Paul Collier and Nicholas Sambanis. Washington, DC: World
Bank.
Bah, A. Sarjoh. 2009. “The Evolution of Regional and Subregional Collective Security
Mechanisms in the Post-Cold War World.” In Cooperating for Peace and Security, ed-
ited by Bruce Jones, Shepard Forman, and Richard Gowan. Cambridge: Cambridge
University Press.
Ballentine, Karen, and Jake Sherman. 2003. The Political Economy of Armed Conflict: Be-
yond Greed and Grievance. Illustrated ed. Boulder, CO: Lynne Rienner.
Barnett, Michael N. 2006. “Building a Republican Peace: Stabilizing States after War.”
International Security 30:87–112.
Barnett, Michael N., and Martha Finnemore. 2004. Rules for the World: International Or-
ganizations in Global Politics. Ithaca, NY: Cornell University Press.
BBC. 2010. “Central African Republic Timeline.” May 10. www.bbc.co.uk / news /
world-africa-13150044.
Beauvais, Joel C. 2000. “Benevolent Despotism: A Critique of U.N. State-Building in
East Timor.” New York University Journal of International Law and Politics 33:1101.
Beidas, Sandra, Colin Granderson, and Rachel Neild. 2007. “Haiti.” In Constructing Jus-
tice and Security after War, edited by Charles T. Call. Washington, DC: US Institute of
Peace Press.
Benjaminsen, Tor A. 2008. “Does Supply-Induced Scarcity Drive Violent Conflicts in
the African Sahel? The Case of the Tuareg Rebellion in Northern Mali.” Journal of
Peace Research 45 (6): 819–36.
Bennett, Andrew. 2008. “Building Communities, Bridging Gaps: Alexander George’s
Contributions to Research Methods.” Political Psychology 29 (4): 489–507.
Berdal, Mats. 2009. Building Peace after War. London: Routledge.
Berdal, Mats, and Richard Caplan. 2004. “The Politics of International Administration.”
Global Governance 10:1.
Bergquist, Charles W., Ricardo Peñaranda, and Gonzalo Sánchez G., eds. 1992. Violence
in Colombia: The Contemporary Crisis in Historical Perspective. Wilmington, DE: SR
Books.
Bermeo, Nancy. 2003. “What the Democratization Literature Says, and Doesn’t Say,
about Post-War Democratization.” Global Governance 9 (2): 159–77.
Bhebe, Ngwabi, and Terrence Ranger, eds. 1995. Soldiers in Zimbabwe’s Liberation War.
London: James Currey.
Binningsbø, Helga Malmin. 2006. “Consociational Democracy and Postconflict Peace:
Will Power-Sharing Institutions Increase the Probability of Lasting Peace after Civil
War?” Paper presented at annual meeting of International Studies Association, Town
and Country Resort and Convention Center, San Diego, March.
References 279

Bjornlund, Eric, Glenn Cowan, and William Gallery. 2006. “Elections Systems and Par-
ties.” In Governance in Post-Conflict Societies: Rebuilding Fragile States, edited by Der-
ick W. Brinkeroff. London: Routledge.
Booth, John A. 1985. The End and the Beginning: The Nicaraguan Revolution. Boulder,
CO: Westview Press.
Boutros Ghali, Boutros. 1992. An Agenda for Peace. New York: United Nations.
Bowers, Stephen. 1994. “Tibet since Mao Zedong.” Journal of Social, Political, and Eco-
nomic Studies 19 (4): 409.
Bowles, Edith, and Tanja Chopra. 2008. “East Timor: Statebuilding Revisited.” In Build-
ing States to Build Peace, edited by Charles T. Call with Vanessa Wyeth. Boulder, CO:
Lynne Rienner.
Brabazon, James. 2003. Liberia: Liberians United for Reconciliation and Democracy. Lon-
don: Royal Institute of International Affairs, www.chathamhouse.org / publications /
papers / view / 107611.
Brahimi, Lakhdar. 2007. “State Building in Crisis and Post-Conflict Countries.” http: //
unpan1.un.org / intradoc / groups / public / documents / un / unpan026305.pdf.
Briceño, Franklin. 2010. “Ex-Peru Rebels, Out of Prison, Run for Elected Office.” San
Francisco Examiner, September 30.
Brinkerhoff, Derick W., ed. 2007. Governance in Post-Conflict Societies: Rebuilding Fragile
States. London: Routledge.
Bromley, Mark. 2007. UN Arms Embargoes, Their Impact on Arms Flows and Target Behav-
ior: Case Study Haiti 1993–94. Stockholm: Stockholm Peace and Research Institute.
Brown, Timothy C. 2001. The Real Contra War, Highlander Peasant Resistance in Nicara-
gua. Oklahoma City: University of Oklahoma Press.
Brubaker, Rogers, and David Laitin. 1998. “Ethnic and Nationalist Violence.” Annual Re-
view of Sociology 24:423–52.
Bull, Hedley. 1977. The Anarchical Society: A Study of Order in World Politics. New York:
Columbia University Press.
Burton, Michael, and John Higley. 1998. “Political Crises and Elite Settlements.” In Elites,
Crises, and the Origins of Regimes, edited by Mattei Dogan and John Higley. Lanham,
MD: Rowman & Littlefield.
Call, Charles T. 1995. “The U.S. Military.” In U.S.–Latin American Policymaking: A Refer-
ence Handbook, edited by David W. Dent.
———. 1999. “From Soldiers to Cops: ‘War Transitions’ and the Demilitarization of
Public Security in Latin America and the Caribbean.” PhD diss., Stanford University.
———. 2002. “War Transitions and the New Civilian Security in Latin America.” Com-
parative Politics 35 (1).
———, ed., with Vanessa Wyeth. 2008a. Building States to Build Peace. Boulder, CO:
Lynne Rienner.
———. 2008b. “The Fallacy of the ‘Failed State.’” Third World Quarterly 29:1491–1507.
———. 2008c. “Knowing Peace When You See It: Setting Standards for Peacebuilding
Success.” Civil Wars 10 (2): 173–94.
280 References

———. 2011. “Beyond the Failed State: Some Conceptual Alternatives.” European Jour-
nal of International Relations 17 (2): 302–26.
Call, Charles T., and Susan E. Cook. 2003. “On Democracy and Peacebuilding.” Global
Governance 9 (2).
Call, Charles T., and Elizabeth Cousens. 2008. “Ending Wars and Building Peace.” Inter-
national Studies Perspectives 9 (1): 1–21.
Call, Charles T., and John Schmitt. 2009. “Why Peace Fails: Explaining Civil War Re-
currence.” Paper presented at annual conference of International Studies Association,
New York.
Campbell, Leon G. 1973. “The Historiography of the Peruvian Guerrilla Movement,
1960–1965.” Latin American Research Review 8 (1): 45–70.
Caplan, Richard. 2004a. “International Authority and State Building: The Case of Bos-
nia and Herzegovina.” Global Governance 10:53.
———. 2004b. “Partner or Patron? International Civil Administration and Local
Capacity-Building.” International Peacekeeping 11 (2): 229–47.
———. 2005. International Governance of War-Torn Territories. Oxford: Oxford Univer-
sity Press.
Carothers, Thomas. 1999. Aiding Democracy Abroad. Washington, DC: Carnegie En-
dowment for International Peace.
———. 2006. “The Backlash against Democracy Promotion.” Foreign Affairs 85 (March–
April): 55–68.
———. 2009. “Democracy Assistance: Political versus Developmental.” Journal of De-
mocracy 20 (1): 5–19.
Catholic Commission for Justice and Peace in Zimbabwe. 1997. Report on the 1980s Dis-
turbances in Matabeland and the Midlands. Harare: Catholic Commission for Justice
and Peace in Zimbabwe.
Cederman, Lars-Erik, and Kristian Skrede Gleditsch. 2009. “Introduction to Special
Issue on ‘Disaggregating Civil Wars.’ ” Journal of Conflict Resolution 53 (4): 487–95.
Chandler, David. 2006. Empire in Denial: The Politics of State-Building. London: Pluto.
Chandra, Kanchan, and Steven Wilkinson. 2008. “Measuring the Effect of ‘Ethnicity.’”
Comparative Political Studies 41:515–63.
Chatterjee, Partha. 1993. The Nation and Its Fragments: Colonial and Postcolonial Histories.
Princeton, NJ: Princeton University Press.
Chesterman, Simon. 2001. Kosovo in Limbo: Statebuilding and “Substantial Autonomy”.
New York: Transitional Administrations Project at International Peace Institute. www
.ipinst.org / media / pdf / publications / kosovo_in_limbo.pdf.
———. 2002a. “Ownership in Theory and in Practice: Transfer of Authority in UN
Statebuilding Operations.” Journal of Intervention and Statebuilding 1 (1): 3–26.
———. 2002b. “Walking Softly in Afghanistan: The Future of UN State-Building.” Sur-
vival 44 (3): 37–45.
———. 2004. You, the People: The United Nations, Transitional Administration, and State-
Building. Oxford: Oxford University Press.
References 281

———. 2007. “Ownership in Theory and in Practice: Transfer of Authority in UN State-

building Operations.” Journal of Intervention and Statebuilding 1:3.
Chesterman, Simon, Michael Ignatieff, and Ramesh Chandra Thakur, eds. 2005. Making
States Work: State Failure and the Crisis of Governance. Tokyo: United Nations Univer-
sity Press.
Chopra, Jarat. 2000. “The UN’s Kingdom of East Timor.” Survival 42 (3): 27–39.
———. 2002. “Building State Failure in East Timor.” Development and Change 33 (5):
979–1000.
Chopra, Jarat, and Tanja Hohe. 2004. “Participatory Intervention.” Global Governance
10:289–305.
Chua, Amy. 2002. World on Fire: How Exporting Free Markets and Democracy Breeds Eth-
nic Hatred and Global Instability. New York: Doubleday.
CIA (Central Intelligence Agency). 2009. “Central African Republic.” CIA World Fact-
book. www.cia.gov / library / publications / the-world-factbook / geos / ct.html.
———. 2010. “East Timor.” In CIA World Factbook. Washington, DC: Central Intelli-
gence Agency. www.cia.gov / library / publications / the-world-factbook / geos / tt.html.
CIC and IPI (Center on International Cooperation and International Peace Institute).
2006. Next Steps for the Peacebuilding Commission: Seminar Report. New York: Center
on International Cooperation and International Peace Institute.
———. 2008. Taking Stock, Looking Forward: A Strategic Review of the Peacebuilding
Commission. New York: Center on International Cooperation and International Peace
Institute.
Clark, Juan. 1992. Cuba; Mito y Realidad. Miami: Saete Ediciones.
Cobban, Helena. 1987. The Making of Modern Lebanon. Boulder, CO: Westview Press.
Cohen, David. 2006. Indifference and Accountability: The United Nations and the Politics
of International Justice in East Timor. East-West Center Special Report 9. Honolulu:
East-West Center.
Cohen, Frank S. 1997. “Proportional versus Majoritarian Ethnic Conflict Management
in Democracies.” Comparative Political Studies 30 (5): 607–31.
Coleman, Katharina Pichler. 2007. International Organisations and Peace Enforcement:
The Politics of Legitimacy. Cambridge: Cambridge University Press.
Collelo, Thomas. 1989. Lebanon: A Country Study, 3rd ed. Prepared by the Library of
Congress, Federal Research Division. Washington, DC: US Department of the Army
Press.
Collier, Paul. 2002. “Understanding Civil War.” Journal of Conflict Resolution 46 (1): 3–12.
———. 2007. The Bottom Billion: Why the Poorest Countries Are Failing and What Can Be
Done about It. New York: Oxford University Press.
———. 2008. “Postconflict Economic Policy.” In Building States to Build Peace, edited by
Charles T. Call. Boulder, CO: Lynne Rienner.
———. 2009. Wars, Guns and Votes. New York: HarperCollins.
Collier, Paul, and Anke Hoeffler. 2002. “On the Incidence of Civil War in Africa.” Journal
of Conflict Resolution 46 (1): 13–28.
282 References

———. 2004. “Greed and Grievance in Civil War.” Oxford Economic Papers 56:563–95.
———. 2006. “Military Expenditure in Post-Conflict Societies.” Economics of Gover-
nance 7 (1): 89–107.
Collier, Paul, Anke Hoeffler, and Måns Söderbom. 2006a. Aid, Policies and Risks in Post-
Conflict Societies. Working paper for Oxford Centre for the Study of African Econo-
mies. Oxford: Oxford University.
———. 2006b. Post-Conflict Societies. Working paper for the Oxford Centre for the
Study of African Economies. Oxford: Oxford University.
Collier, Paul, V. L. Elliott, Håvard Hegre, Anke Hoeffler, Marta Reynal-Querol, and
Nicholas Sambanis. 2003. Breaking the Conflict Trap: Civil War and Development
Policy. Washington, DC: World Bank. www-wds.worldbank .org / external / default /
WDSContentServer/IW3P/IB/2003/06/30/000094946_0306190405396/Rendered/
PDF / multi0page.pdf.
Cook, Nicholas. 2003. Liberia: 1989–1997 Civil War, Post-War Developments, and US
Relations. Washington, DC: Congressional Research Service.
———. 2004. Liberia: Transition to Peace. Washington, DC: Congressional Research
Service.
Cornell, Svante E. 2001. Small Nations and Great Powers: A Study of Ethnopolitical Conflict
in the Caucasus. London: Curzon.
Cousens, Elizabeth. 2002. “From Missed Opportunities to Overcompensation: Imple-
menting the Dayton Agreement on Bosnia.” In Ending Civil Wars: The Implementation
of Peace Agreements, edited by Stephen John Stedman, Donald S. Rothchild, and Eliza-
beth M. Cousens. Boulder, CO: Lynne Rienner.
Cox, Marcus. 2008. “Bosnia and Herzegovina: The Limits of Liberal Imperialism.”
In Building States to Build Peace, edited by Charles T. Call. Boulder, CO: Lynne
Rienner.
Craig, Dylan. 2011. “The Wars in the Walls: Sovereign Interstices and African Conflict.”
Paper presented at International Studies Association conference, Montreal, March.
Curtis, Glenn E., ed. 1995. Armenia, Azerbaijan, and Georgia. Washington, DC: Federal
Research Division, Library of Congress.
Cvetkovski, Nikola. 2001. The Georgian–South Ossetian Conflict. Copenhagen: Danish
Association for Research on the Caucasus. www.caucasus.dk / publication5.htm.
DAC-OECD (Development Assistance Committee of Organization for Economic Co-
operation and Development). 2007. Principles for Good International Engagement in
Fragile States and Situations. Paris: OECD.
Dailey, Peter. 2003. “Haiti: The Fall of the House of Aristide.” New York Review of Books
50 (4).
———. 2004. “Who Got Rid of Aristide?” Letter printed in London Review of Books,
May 20.
Daly, M. W. 1993. “Broken Bridge and Empty Basket: The Political and Economic Back-
ground of the Sudanese Civil War.” In Civil War in the Sudan, edited by M. W. Daly and
Ahmad Alawad Sikainga. London: British Academic Press.
References 283

Davis, Michael C. 2008. “Establishing a Workable Autonomy in Tibet.” Human Rights

Quarterly 30 (2): 227–58.
Dawisha, Karen, and Bruce Parrott. 1997. Conflict, Cleavage, and Change in Central Asia
and the Caucasus. New York: Cambridge University Press.
Day, Arthur R. 1986. “Civil Conflict in Lebanon.” In Escalation and Intervention: Multilat-
eral Security and Its Alternatives, edited by Arthur R. Day and Michael Doyle. Boulder,
CO: Westview Press.
De Soto, Alvaro, and Graciana del Castillo. 1994. “Obstacles to Peacebuilding.” Foreign
Policy 94 (Spring): 69–83.
del Castillo, Graciana. 2008. Rebuilding War-Torn States: The Challenge of Post-Conflict
Economic Reconstruction. Oxford: Oxford University Press.
Deledda, A. 2006. “Afghanistan: The End of the Bonn Process.” Transition Studies Review
13 (1): 155–71.
Deng Deng, William. 2001. A Survey of Programs on the Reintegration of Child Soldiers.
Tokyo: Ministry of Foreign Affairs of Japan. www.mofa.go.jp / policy / human / child /
survey / index.html.
Des Forges, Alison. 1999. Leave None to Tell the Story. New York: Human Rights
Watch.
Desmangles, Leslie G. 1992. The Faces of the Gods: Vodou and Roman Catholicism in Haiti.
Chapel Hill: University of North Carolina Press.
Deutsch, Karl W. 1953. “The Growth of Nations: Some Recurrent Patterns of Political
and Social Integration.” World Politics 5 (2): 168–95.
Diamond, Larry. 2005. Squandered Victory: The American Occupation and the Bungled
Effort to Bring Democracy to Iraq. New York: Times Books.
Dillon, Karin. 2007. “Mali’s Touareg Rebellion.” MA thesis, International Peace and
Conflict Resolution Program, American University.
Dillon, Sam. 1991. Comandos: The CIA and Nicaragua’s Contra Rebels. New York: Henry
Holt.
Dowley, Kathleen M., and Brian D. Silver. 2000. “Subnational and National Loyalty:
Cross-National Comparisons.” International Journal of Public Opinion Research 12 (4):
357–71.
Doyle, Michael W. 1986. “Liberalism and World Politics.” American Political Science Re-
view 80 (4): 1151–69.
———. 2002. “Strategy and Transitional Administration.” In Ending Civil Wars: The
Implementation of Peace Agreements, edited by Stephen John Stedman, Donald S. Roth-
child, and Elizabeth M. Cousens. Boulder, CO: Lynne Rienner Publishers.
Doyle, Michael W., and Nicholas Sambanis. 2000. “International Peacebuilding: A The-
oretical and Quantitative Analysis.” American Political Science Review 94:779–801.
———. 2006. Making War and Building Peace: United Nations Peace Operations. Prince-
ton, NJ: Princeton University Press.
Duke, Lynne. 1998. “Angola’s Peace Withers Again under Fire,” Washington Post, Decem-
ber 15.
284 References

Dunlop, John B. 1998. Russia Confronts Chechnya: Roots of a Separatist Conflict. Cam-
bridge: Cambridge University Press.
Dunning, Thad. 2005. “Resource Dependence, Economic Performance, and Political
Stability.” Journal of Conflict Resolution 49:451–82.
Dupuy, Alex. 2007. The Prophet and Power: Jean-Bertrand Aristide, the International Com-
munity, and Haiti. Lanham, MD: Rowman & Littlefield.
———. 2009. “Indefensible: On Aristide, Violence, and Democracy.” Small Axe 13 (3):
161–73.
Dupuy, Kendra, and Helga Malmin Binningsbø. 2008. Power-Sharing and Peace-
Building in Sierra Leone: Power-Sharing Agreements, Negotiations and Peace Processes.
Oslo: Peace Research Institute Oslo / Centre for the Study of Civil War. www.prio
.no / sptrans / -42188479 / Sierra_Leone_full_report.pdf.
Dupuy, Kendra, and Julian Detzel. 2007. Peace-Building and Powersharing in Liberia.
Oslo: Centre for the Study of Civil War.
Durch, William J., ed. 1996. UN Peacekeeping, American Politics, and the Uncivil Wars of
the 1990s. New York: St. Martin’s Press.
———. 2006. Twenty-First-Century Peacekeeping Operations. Washington, DC: US In-
stitute of Peace Press.
DuToit, Pierre. 2003. “Why Post- Settlement Settlements?” Journal of Democracy 14
(3):104–18.
Eagles, Munroe, Christopher Holoman, and Larry Johnston. 2004. Politics, Introduction
to Democratic Government, 2nd ed. Peterborough, Ont.: Broadview.
Eckstein, Harry. 1975. “Case Study and Theory in Political Science.” In Handbook of Po-
litical Science, edited by Fred Greenstein and Neslon Polsby. Reading, MA: Addison-
Wesley.
Elazar, Daniel J. 1994. Federalism and the Way to Peace. Kingston, Ont.: Institute of Inter-
governmental Relations.
Elkins, Zachary, and John Sides. 2007. “Can Institutions Build Unity in Multi-Ethnic
States?” American Political Science Review 101:693–708.
Elliott, Paul D. 1978. “The East Timor Dispute.” International and Comparative Law
Quarterly 27 (1): 238–49.
Ellis, Stephen. 2001. The Mask of Anarchy: The Destruction of Liberia and the Religious
Dimension of an African Civil War. New York: New York University Press.
Englebert, Pierre. 2000. State Legitimacy and Development in Africa. Boulder, CO: Lynne
Rienner.
Esty, Daniel C., Jack A. Goldstone, Ted Robert Gurr, Pamela T. Surko, and Alan N. Unger.
1995. “Working Papers: State Failure Task Force Report.” Prepared on contract for the
US Central Intelligence Agency’s Directorate of Intelligence.
Evangelista, Matthew. 2002. The Chechen Wars: Could Russia Go the Way of the Soviet
Union? Washington, DC: Brookings Institution Press.
Eziakonwa, Ahunna. 1998. “Donors Back Liberia’s Reconstruction.” Africa Recovery 12.
Farmer, Paul. 2004. “Who Removed Aristide?” London Review of Books 26 (8): 28–31.
References 285

Fasano, Ugo. 2009. “Haiti’s Economic Development since 2004 / 05 and Macroeco-
nomic Outlook.” PowerPoint presentation by former resident representative for In-
ternational Monetary Fund, Port-au-Prince, July 1.
Fatton, Robert. 2002. Haiti’s Predatory Republic: The Unending Transition to Democracy.
Boulder, CO: Lynne Rienner.
Fayemi, J. ‘Kayode. 2004. “Governing Insecurity in Post-Conflict States: The Case of
Sierra Leone and Liberia.” In Reform and Reconstruction of the Security Sector, edited by
Alan Bryden and Heiner Hänggi. Geneva: Geneva Center for the Democratic Control
of Armed Forces.
Fearon, James D. 1998. “Commitment Problems and the Spread of Ethnic Conflict.” In
The International Spread of Ethnic Conflict, edited by David Lake and Donald Roth-
child. Princeton, NJ: Princeton University Press.
———. 2005. “Primary Commodity Exports and Civil War.” Journal of Conflict Resolu-
tion 49:483–507.
Fearon, James D., and David D. Laitin. 2003. “Ethnicity, Insurgency, and Civil War.”
American Political Science Review 97:75–90.
Feigon, Lee. 1996. Demystifying Tibet. Chicago: Ivan R. Dee.
Fishman, Robert M. 1990. “Rethinking State and Regime: Southern Europe’s Transition
to Democracy.” World Politics 42 (3): 422–40.
Forman, Shepard. 2005. “Working Better Together: Implementing the High Level Pan-
el’s Recommendations on Peacebuilding.” Presentation at conference on “New Multi-
lateralism” sponsored by L20 Forum, Brussels, May 12.
Fortna, Virginia Page. 2004a. Peace Time: Cease-Fire Agreements and the Durability of
Peace. Princeton, NJ: Princeton University Press.
———. 2004b. “Where Have All the Victories Gone?” Paper presented at conference on
order, conflict and violence, Yale University, New Haven, CT, April.
———. 2008. Does Peacekeeping Work? Shaping Belligerents’ Choices after Civil War.
Princeton, NJ: Princeton University Press.
Fox, Richard G., and Orin Starn, eds. 1997. Between Resistance and Revolution: Cultural
Politics and Social Protest. New Brunswick, NJ: Rutgers University Press.
Fukuyama, Francis. 2004. State-Building: Governance and World Order in the 21st Century.
Ithaca, NY: Cornell University Press.
Fumerton, Mario. 2003. From Victims to Heroes: Peasant Counter-Rebellion and Civil War
in Ayacucho, Peru, 1980–2000. West Lafayette, IN: Purdue University Press.
Gagnon, V. P. 1994. “Ethnic Nationalism and International Conflict: The Case of Serbia.”
International Security 19 (3): 130–66.
Gale Group. 2008. “Hutu and Tutsi of Burundi.” In The World Directory of Minori-
ties. www.faqs .org / minorities / Sub- Saharan-Africa / Hutu-and-Tutsi-of-Burundi
.html.
Gall, Carlotta, and Thomas de Waal. 1998. Chechnya: Calamity in the Caucasus. New
York: New York University Press.
Galtung, Johan. 1964. “An Editorial.” Journal of Peace Research 1 (1).
286 References

Gates, Scott, and Kaare Strøm. 2007. Power-sharing, Agency and Civil Conflict: Power-
sharing Agreements, Negotiations and Peace Processes. Centre for the Study of Civil Wars
Policy Brief. Oslo: Peace Research Institute Olso.
Geertz, Clifford. 1963. Old Societies and New States: The Quest for Modernity in Asia and
Africa. Glencoe, IL: Free Press.
Gellner, Ernest. 1991. “Nationalism and Politics in Eastern Europe.” New Left Review
189:127–43.
George, Alexander, and Andrew Bennett, 2005. Case Studies and Theory Development in
the Social Sciences. Cambridge, MA: MIT Press.
German, Tracey. 2003. Russia’s Chechen War. New York: Routledge.
Ghani, Ashraf, and Clare Lockhart. 2008. Fixing Failed States: A Framework for Rebuilding
a Fractured World. Oxford: Oxford University Press.
Ghani, Ashraf, Clare Lockhart, and Michael Carnahan. 2005. Closing the Sovereignty
Gap: An Approach to State-Building. London: Overseas Development Institute.
Gleason, Gregory. 2001. “Power Sharing in Tajikistan: Political Compromise and Re-
gional Instability.” Conflict, Security & Development 1:125.
Gleditsch, Kristian Skrede, and Andrea Ruggeri. 2010. “Political Opportunity Structures,
Democracy, and Civil War.” Journal of Peace Research 47 (3): 299–310.
Goldsmith, Andrew. 2009. “‘It Wasn’t Like Normal Policing’: Voices of Australian Po-
lice Peacekeepers in Operation Serene, Timor-Leste 2006.” Policing & Society 19 (2):
119–33.
Goldstein, Melvyn C. 2007. A History of Modern Tibet, 1951–1955: The Calm before the
Storm. Berkeley: University of California Press.
Goldstone, Anthony. 2004. “UNTAET with Hindsight: The Peculiarities of Politics in
an Incomplete State.” Global Governance 10 (1): 83–98.
Gorjao, Paolo. 2008. “UNMISET’s Hasty Replacement by UNOTIL in East Timor.”
www.etan.org / et2005 / may / 08 / 00unmiset.htm.
Gorriti, Gustavo. 1999. The Shining Path: A History of the Millenarian War in Peru. Chapel
Hill: University of North Carolina Press. Orig. pub. in Spanish in 1990.
Greenfeld, Liah. 1992. Nationalism: Five Roads to Modernity. Cambridge, MA: Harvard
University Press.
Grunfeld, A. Tom. 1996. The Making of Modern Tibet. Armonk, NY: East Gate Books.
Gurr, Ted Robert. 1970. Why Men Rebel. Princeton, NJ: Princeton University Press.
———. 1993. Minorities at Risk. Washington, DC: US Institute of Peace Press.
———. 2000. Peoples versus States: Minorities at Risk in the New Century. Washington,
DC: US Institute of Peace Press.
Gurses, Mehmet, Nicolas Rost, and Patrick McLeod. 2008. “Mediating Civil War Settle-
ments and the Duration of Peace.” International Interactions 34 (2): 129–55.
Hagopian, Frances. 1991. “Democracy by Undemocratic Means? Elites, Political Pacts,
and Regime Transition in Brazil.” Comparative Political Studies 23 (2): 147–70.
Hampson, Fen Osler. 1996. Nurturing Peace: Why Peace Settlements Succeed or Fail. Wash-
ington, DC: US Institute of Peace Press.
References 287

Harbom, Lotta, Stina Högbladh, and Peter Wallentsteen. 2006. “Armed Conflict and
Peace Agreements.” Journal of Peace Research 43 (5): 617–31.
Harrington, Andrew. 2007. “Ethnicity, Violence and Land and Property Disputes in East
Timor.” East Timor Law Journal 4 (1).
Hartlyn, Jonathan. 1988. The Politics of Coalition Rule in Colombia. Cambridge: Cam-
bridge University Press.
Hartzell, Caroline, and Matthew Hoddie. 2003. “Institutionalizing Peace: Power Shar-
ing and Post–Civil War Conflict Management.” American Journal of Political Science
47:318–32.
———. 2007. Crafting Peace: Power-Sharing Institutions and the Negotiated Settlement of
Civil Wars. University Park: Pennsylvania State University Press.
———. 2010. “Economic Liberalization via IMF Structural Adjustment: Sowing the
Seeds of Civil War?” International Organization 64:339–56.
Havermans, Jos. 2000. “Central African Republic: Ethnic Strife in a Democratic Setting.”
www.conflict-prevention.net/page.php?id=40&formid=73&action=show&surveyid=61.
Hegre, Håvard, Tanja Ellingsen, Scott Gates, and Nils Petter Gleditsch. 2001. “Toward
a Democratic Civil Peace? Democracy, Political Change, and Civil War, 1816–1992.”
American Political Science Review 95 (1): 33–48.
Hegre, Håvard, and Nicholas Sambanis. 2004. “Sensitivity Analysis of the Empirical Lit-
erature on Civil War Onset.” Paper presented at meeting of Polarization and Conflict
Network, Voksenåsen, Oslo, July 2–4.
Henderson, Errol, and Singer, J. David. 2000. “Civil War in the Postcolonial World,
1946–92.” Journal of Peace Research 37 (3): 275–99.
Herbst, Jeffrey. 2000. States and Power in Africa. Princeton, NJ: Princeton University
Press.
Herz, Monica. 2008. Does the Organisation of American States Matter? Crisis States Work-
ing Paper Series 2, no. 34. London: Crisis States Research Centre.
Hill, Fiona. 1995. Russia’s Tinderbox: Conflict in the North Caucasus and Its Implications for
the Future of the Russian Federation. Report for the Strengthening Democratic Institu-
tions Project. Cambridge, MA: Harvard University.
Hoadley, J. Stephen. 1975. The Future of Portuguese Timor: Dilemmas and Opportunities.
Singapore: Institute of Southeast Asian Studies.
Hoddie, Matthew, and Caroline Hartzell. 2005. “Power Sharing in Peace Agreements.”
In Sustainable Peace: Power and Democracy after Civil Wars, edited by Philip G. Roeder
and Donald Rothchild. Ithaca, NY: Cornell University Press.
Hoffman, David. 2000. “Miscalculations Paved Path to Chechen War.” Washington Post,
March 20.
Hohe, Tanja. 2002. “The Clash of Paradigms: International Administration and
Local Political Legitimacy in East Timor.” Contemporary Southeast Asia 24 (3):
569–89.
Holsti, Kalevi J. 1996. The State, War, and the State of War. Cambridge: Cambridge Uni-
versity Press.
288 References

Honwana, Alcinda. 2006. Child Soldiers in Africa. Philadelphia: University of Pennsyl-

vania Press.
Hood, Ludovic. 2006. “Security Sector Reform in East Timor, 1999–2004.” Interna-
tional Peacekeeping 13 (1): 60–77.
Horowitz, Donald L. 1985. Ethnic Groups in Conflict. Berkeley: University of California
Press.
Horton, Lynn. 1998. Peasants in Arms: War and Peace in the Mountains of Nicaragua,
1979–1994. Athens: Ohio University Center for International Studies.
Howard, Lise Morje. 2008. UN Peacekeeping in Civil Wars. Cambridge: Cambridge Uni-
versity Press.
Howe, Herbert. 1996. “Lessons of Liberia: ECOMOG and Regional Peacekeeping.” In-
ternational Security 21:145–76.
Human Rights Watch. 1993. Silencing a People: The Destruction of Civil Society in Haiti.
New York: Human Rights Watch and National Coalition for Haitian Refugees.
———. 2002a. Back to the Brink: War Crimes by Liberian Government and Rebels. New
York: Human Rights Watch.
———. 2002b. World Report 2002. New York: Human Rights Watch.
Human Security Report Project. 2005. Human Security Report 2005: War and Peace in the
21st Century. Burnaby, BC: Human Security Research Group, School for International
Studies, Simon Fraser University. www.hsrgroup.org / human-security-reports / 2005 /
text.aspx.
———. 2009. Human Security Report 2009 / 10: The Causes of Peace and The Shrink-
ing Costs of War. Burnaby, BC: Human Security Research Group, School for Interna-
tional Studies, Simon Fraser University. www.hsrgroup.org / human-security-reports /
20092010 / text.aspx.
Humphreys, Macartan. 2005. “Natural Resources, Conflict, and Conflict Resolution:
Uncovering the Mechanisms.” Journal of Conflict Resolution 49:508–37.
Humpreys, Macartan, and Habaye ag Mohamed. 2003. “Senegal and Mali.” www.columbia
.edu / %7Emh2245 / papers1 / sen_mali.pdf. (Longer version in Understanding Civil
War, edited by Paul Collier and Nicholas Sambanis. Washington, DC: World Bank,
2005.)
Humphreys, Macartan, and Jeremy M. Weinstein. 2007. “Demobilization and Reintegra-
tion.” Journal of Conflict Resolution, 51 (4): 531–67.
Hunter, Boyd, and Kirrily Jordan. 2010. “Explaining Social Exclusion: Towards So-
cial Inclusion for Indigenous Australians.” Australian Journal of Social Issues 45 (2):
243–65.
Huntington, Samuel P. 1991. The Third Wave: Democratization in the Late Twentieth Cen-
tury. Norman: University of Oklahoma Press.
Hyland, Frank. 2008. “Peru’s Sendero Luminoso: From Maoism to Narco-Terrorism.”
Terrorism Monitor ( Jamestown Foundation), December 8.
Ikenberry, John G. 2000. After Victory: Order and Power in International Politics. Prince-
ton, NJ: Princeton University Press.
References 289

International Crisis Group. 2002. Liberia: The Key to Ending Regional Instability. Wash-
ington, DC: International Crisis Group.
———. 2004. Georgia: Avoiding War in South Ossetia. Europe Report 59. Brussels: In-
ternational Crisis Group.
———. 2006a. Managing Tensions on the Timor Leste / Indonesian Border. Asia Briefing
50. Brussels: International Crisis Group.
International Foundation for Electoral Systems. 2005. “Electoral System for East Timor.”
www.electionguide.org / country.php?ID=63.
International Monetary Fund. 2010. World Economic Outlook Database, September 2011.
Washington, DC: International Monetary Fund.
IRBC (Immigration and Refugee Board of Canada). 2001. “Liberia: Madison Wion
and Treatment of His Relatives by Liberian Authorities.” www.unhcr.org / refworld /
docid / 3df4be5d20.html.
IRIN. 2004. “Liberia: LURD Leader Drops Demand for Bryant to Go.” http: // www
.irinnews.org / report.aspx?reportid=48466.
Itano, Nicole. 2003. Liberating Liberia: Charles Taylor and the Rebels Who Unseated Him.
ISS Paper 82. Johannesburg: Institute for Security Studies.
Iyob, Ruth, and Gilbert Khadiagala. 2006. Sudan: The Elusive Quest for Peace. Boulder,
CO: Lynne Rienner.
Jackson, Robert H. 1990. Quasi-States, Sovereignty, International Relations, and the Third
World. Cambridge: Cambridge University Press.
Jackson, Robert H., and Carl G. Rosberg. 1986. “Sovereignty and Underdevelopment: Ju-
ridical Statehood in the African Crisis.” Journal of Modern African Studies 24 (1): 1–31.
Jardine, Matthew, and Constancio Pinto. 1997. East Timor’s Unfinished Struggle. Boston:
South End Press.
Jarstad, Anna. 2008. “Post-Accord Elections, Power Sharing and Conflict.” Paper presented
at 49th Annual Convention of the International Studies Association, San Francisco.
———. 2009. “The Prevaslence of Power- Sharing: Exploring the Patterns of Post-
Election Peace.” Africa Spectrum 44 (3): 41–62.
Jarstad, Anna, and Timothy Sisk. 2008. From War to Democracy: Dilemmas of Peacebuild-
ing. Cambridge: Cambridge University Press.
Jean, François. 2000. “Chechnya: Moscow’s Revenge.” Harvard International Review 22
(3): 16–21.
Johnson, Chris, William Maley, Alex Their, and Ali Wardak. 2003. “Afghanistan’s Politi-
cal and Constitutional Development.” Report for UK Department for International
Development and Humanitarian Policy Group at Overseas Development Institute,
London.
Johnson, Douglas H., and Gerard Prunier. 1993. “The Foundation and Expansion of the
Sudan People’s Liberation Army.” In Civil War in the Sudan, edited by M. W. Daly and
Ahmad Alawad Sikainga. London: British Academic Press.
Jones, Bruce, Rahul Chandran, Elizabeth Cousens, Jenna Slotin, and Jake Sherman.
2008. Concepts and Dilemmas in State Building in Fragile Situations. Report based on
290 References

paper by Center on International Cooperation. Paris: Organization for Economic Co-

operation and Development.
Kaldor, Mary. 1999. New and Old Wars: Organized Violence in a Global Era. Stanford, CA:
Stanford University Press.
Kalyvas, Stathis N. 2006. The Logic of Violence in Civil War. New York: Cambridge Uni-
versity Press.
Karl, Terry Lynn. 1990. “Dilemmas of Democratization in Latin America.” Comparative
Politics 23 (1).
———. 1997. The Paradox of Plenty: Oil Booms and Petro-States. Berkeley: University of
California Press.
Kaufmann, Chaim. 1996. “Possible and Impossible Solutions to Ethnic Civil Wars.” In-
ternational Security 20 (4): 136–75.
———. 1998. “When All Else Fails: Ethnic Population Transfers and Partition in the
Twentieth Century.” International Security 23 (2): 120–56.
Keating, Thomas F., and W. Andy Knight. 2004. Building Sustainable Peace. Tokyo:
United Nations University Press.
Keefer, Philip. 2008. “Insurgencies and Credible Commitments in Autocracies and De-
mocracies.” World Bank Economic Review 22 (1): 33–61.
Keita, Kalifa. 1998. “Conflict and Conflict Resolution in the Sahel: The Tuareg Insur-
gency in Mali.” MA thesis, US Army War College.
Keohane, Robert O., and Joseph S. Nye. 1977. Power and Interdependence: World Politics
in Transition. Boston: Little, Brown.
King, Gary, Robert Keohane, and Sidney Verba. 1994. Designing Social Inquiry: Scientific
Inference in Qualitative Research. Princeton, NJ: Princeton University Press.
Kingsbury, Damien. 2009. East Timor: The Price of Liberty. New York: Palgrave Mac-
millan.
König, Marietta. 2005. “The Georgian–South Ossetian Conflict.” In OSCE Yearbook
2004, edited by Institute for Peace Research and Security Policy at University of Ham-
burg. Baden-Baden: Institute for Peace Research and Security Policy .
Kosovar Institute for Policy Research and Development. 2009. Decentralization in Kosovo
I: Municipal Elections and the Serb Participation. Policy Brief 15. Prishtinë: Kosovar
Institute for Policy Research and Development.
Krain, Matthew. 2000. Repression and Accommodation in Post-Revolutionary States. New
York: St. Martin’s Press.
Krain, Matthew, and Marissa Myers. 1997. “Democracy and Civil War: A Note on the
Democratic Peace Proposition.” International Interactions 23 (1): 109–18.
Krasner, Stephen D. 2001. “Rethinking the Sovereign State Model.” Review of Interna-
tional Studies 27:17–42.
———. 2004. “Sharing Sovereignty: New Institutions for Collapsed and Failing States.”
International Security 29:85–120.
Krings, Thomas. 1995. “Marginalisation and Revolt among the Tuareg in Mali and Ni-
ger.” GeoJournal 36 (1): 57–63.
References 291

Krueger, Robert, and Kathleen T. Krueger. 2007. From Bloodshed to Hope in Burundi: Our
Embassy Years during Genocide. Austin: University of Texas Press.
Kumar, Krishna. ed. 1998. Postconflict Elections, Democratization and International Assis-
tance. Boulder, CO: Lynne Rienner.
Kumar, Krishna, and Marina Ottaway. 1998. “General Conclusions and Priorities for
Policy Research.” In Postconflict Elections, Democratization and International Assistance,
edited by Krishna Kumar. Boulder, CO: Lynne Rienner.
Lawson, Stephanie. 1993. “Conceptual Issues in the Comparative Study of Regime
Change and Democratization.” Comparative Politics 25 (2): 183–206.
Lawyers Committee for Human Rights. 1986. Zimbabwe: Wages of War. New York: Law-
yers Committee for Human Rights.
Layne, Christopher. 1994. “Kant or Cant: The Myth of the Democratic Peace.” Interna-
tional Security 19 (2): 5–49.
Lemarchand, René. 1975. “Ethnic Genocide.” Issue: A Journal of Opinion 5 (2): 9–16.
———. 1994. “Managing Transition Anarchies: Rwanda, Burundi, and South Africa in
Comparative Perspective.” Journal of Modern African Studies 32 (4): 581–604.
———. 2004. “Exclusion, Marginalization and Political Mobilization: The Road to Hell
in the Great Lakes.” In Facing Ethnic Conflicts: Toward a New Realism, edited by An-
dreas Wimmer, Richard J. Goldstone, Donald Horowitz, Ulrike Joras, and Conrad
Schetter. Lanham, MD: Rowman & Littlefield.
———. 2008. “The Killings of 1972.” In Online Encyclopedia of Mass Violence. Paris:
SciencesPo.
LeVan, Carl. 2011. “Power Sharing and Inclusive Politics in Africa’s Uncertain Democra-
cies.” Governance 24 (1).
Levitas, Ruth. 2005. The Inclusive Society?: Social Exclusion and New Labour, 2nd ed. New
York: Palgrave.
Licklider, Roy. 1995. “The Consequences of Negotiated Settlements in Civil Wars,
1945–1993.” American Political Science Review 89:681–90.
Lieven, Anatol. 1998. Chechnya: Tombstone of Russian Power. New Haven, CT: Yale Uni-
versity Press.
Lijphart, Arend. 1977. Democracy in Plural Societies. New Haven, CT: Yale University
Press.
Longmire, Sylvia. 2009. “Peru: Controversy over Potential for Shining Path Terror-
ists to Participate in Elections.” South American Policy Examiner, October 28. www
.examiner.com/south-america-policy-in-national/peru-controversy-over-potential-for
-shining-path-terrorists-to-participate-elections.
LURD (Liberians United for Reconciliation and Democracy). 2002. “Blue Print for Sta-
bility in Liberia.” The Perspective, October 4. www.theperspective.org / lurdblueprint
.html.
Lynch, Dov. 2004. Engaging Eurasia’s Separatist States: Unresolved Conflicts and De Facto
States. Washington, DC: US Institute of Peace Press.
———. 2005. “New Thinking about Frozen Conflicts.” Helsinki Monitor 16 (3).
292 References

Lynch, Marika. 2003. “Haiti’s Ex-Chief of Police Cites Corruption.” Miami Herald,
June 28.
Lyons, Terrence. 1998. “Peace and Elections in Liberia.” In Postconflict Elections, Democ-
ratization, and International Assistance, edited by Krishna Kumar. Boulder, CO: Lynne
Rienner.
———. 2002. “The Role of Postsettlement Elections.” In Ending Civil Wars: The Imple-
mentation of Peace Agreements, edited by Stephen John Stedman, Donald Rothchild,
and Elizabeth M. Cousens. Boulder, CO: Lynne Rienner.
———. 2004. “Post-Conflict Elections and the Process of Demilitarizing Politics: The
Role of Electoral Administration.” Democratization 11 (3): 36–62.
———. 2005. Demilitarizing Politics: Elections on the Uncertain Road to Peace. Boulder,
CO: Lynne Rienner.
———. 2006. “Diasporas and Homeland Conflict” In Globalization, Territoriality, and
Conflict in an Era of Globalization, edited by Miles Kahler and Barbara Walter. Cam-
bridge: Cambridge University Press.
Maguire, Robert E. 1995. “Defanging the Predatory State.” Hemisphere 7 (1):14–16.
Mahdavy, Hossein. 1970. “The Patterns and Problems of Economic Development in
Rentier States: The Case of Iran.” In Studies in the Economic History of the Middle East:
From the Rise of Islam to the Present Day, edited by M. A. Cook. Oxford: Oxford Uni-
versity Press.
Mainwaring, Scott, Guillermo O’Donnell, and J. Samuel Valenzuela. 1992. Issues in Dem-
ocratic Consolidation: The New South American Democracies in Comparative Perspective.
Notre Dame, IN: University of Notre Dame Press.
Makdisi, Samir, and Richard Sadaka. 2005. “The Lebanese Civil War, 1975–90.” In Un-
derstanding Civil War: Evidence and Analysis, vol. 2, edited by Paul Collier and Nicholas
Sambanis. Washington, DC: World Bank.
Malan, Mark. 2008. Security Sector Reform in Liberia. Carlisle Barracks, PA: Strategic Stud-
ies Institute, US Army War College. www.strategicstudiesinstitute.army.mil / pdffiles /
pub855.pdf.
Malone, David. 1998. Decision-Making in the UN Security Council: The Case of Haiti. Ox-
ford: Clarendon Press.
Mamdani, Mahmoud. 1996. Citizen and Subject: Contemporary Africa and the Legacy of
Late Colonialism. Princeton, NJ: Princeton University Press.
———. 2001. When Victims Become Killers: Colonialism, Nativism and the Genocide in
Rwanda. Princeton, NJ: Princeton University Press.
———. 2002. When Victims Become Killers. Princeton, NJ: Princeton University Press.
Mann, Michael. 1993. The Sources of Social Power, Volume II: The Rise of Classes and
Nation-States, 1760–1914. New York: Cambridge University Press.
Mansfield, Edward, and Jack Snyder. 2005. Electing to Fight: Why Emerging Democracies
Go to War. Cambridge: Belfer Center for International Science and Affairs.
Maoz, Z., and N. Abdolali. 1989. “Regime Types and International Conflict, 1816–1976.”
Journal of Conflict Resolution 33:3–35.
References 293

Maríñez, Pablo A. 1990. “El proceso electoral dominicano: difícil camino hacia la de-
mocracia.” Secuencia 18:57–82.
Martin, Ian, and Alexander Mayer-Rieckh. 2005. “The United Nations and East Timor:
From Self-Determination to State-Building.” International Peacekeeping 12 (1):
125–45.
Martinetti, Irene. 2007. “The Peacebuilding Commission: A Year in Review.” Center for
UN Reform Education. www.globalpolicy.org / component / content / article / 228 /
32578.html.
McClintock, Cynthia. 1998. Revolutionary Movements in Latin America: El Salvador’s
FMLN and Peru’s Shining Path. Washington, DC: US Institute of Peace Press.
McFaul, Michael. 2004. “Democracy Promotion as a World Value.” Washington Quarterly
28 (1): 147–63.
McGarry, John, and Breandan O’Leary. 2003. The Politics of Conflict Regulation. London:
Routledge.
———. 2004. The Northern Ireland Conflict: Consociational Engagement. London: Ox-
ford University Press.
Meyer, John W., John Boli, George M. Thomas, and Francisco O. Ramirez. 1997. “World
Society and the Nation-State.” American Journal of Sociology 103 (1): 144–81.
Milante, Gary, and Stergios Skaperdas. 2009. “Powersharing under the Threat of Con-
flict.” http: // siteresources.worldbank .org / INTCONFLICT / Resources / Milante
Skaperdas090517.pdf.
Minahan, James. 1996. Nations without States: A Historical Dictionary of Contemporary
National Movements. Westport, CT: Greenwood Press.
Mitchell, Christopher R. 1989. Conflict Resolution and Civil War: Reflections on the Su-
danese Settlement of 1972. Working Paper 3 Fairfax, VA: Center for Conflict Analysis
and Resolution, George Mason University. http: // icar.gmu.edu / wp_3_mitchell.pdf.
Mitchell, Paul, Geoffrey Evans, and Brendan O’Leary. 2009. “Extremist Outbidding in
Ethnic Party Systems Is Not Inevitable: Tribune Parties in Northern Ireland.” Political
Studies 57 (2): 397–421.
Moxham, Ben. 2008. State Making and the Post-Conflict City: Integration in Dili, Disinte-
gration in Timor-Leste. Crisis States Working Paper Series 2, no. 32. London: Crisis
States Research Centre.
Mukherjee, Bumba. 2006. “Why Political Power- Sharing Agreements Lead to Endur-
ing Peaceful Resolution of Some Civil Wars, But Not Others.” International Studies
Quarterly 50 (2): 479–504.
Murithi, Tim. 2008. “Peacebuilding or ‘UN-Building’? African Institutional Responses
to the Peacebuilding Commission.” Journal of Peacebuilding and Development 4 (2):
89–94.
Ndikumana, Léonce. 1998. “Institutional Failure and Ethnic Conflicts in Burundi.” Afri-
can Studies Review 41 (1): 29–47.
Newman, Edward, Roland Paris, and Oliver P. Richmond, eds. 2009. New Perspectives on
Liberal Peacebuilding. Tokyo: United Nations University Press.
294 References

Nilsson, Desirée. 2008. “Partial Peace: Rebel Groups Inside and Outside of Civil War
Settlements.” Journal of Peace Research 45 (4): 479–95.
Niner, Sarah. 2001. “A Long Journey of Resistance: The Origins and Struggle of CNRT.”
In Bitter Flowers, Sweet Flowers: East Timor, Indonesia, and the World Community, edited
by Richard Tanter, Mark Selden, Stephen R. Shalom. London: Pluto Press.
Nixon, Hamish, and Richard Ponzio. 2007. “Building Democracy in Afghanistan: The
Statebuilding Agenda and International Engagement.” International Peacekeeping 14
(1): 26–40.
Norris, Pippa. 2008. Driving Democracy: Do Power-Sharing Institutions Work? Cambridge:
Cambridge University Press.
North, Douglass. 1981. Structure and Change in Economic History. New York: W. W.
Norton.
Norton, Augustus Richard, 1991. “Lebanon after Ta’if: Is the Civil War Over?” Middle
East Journal 40 (3): 457–73.
OAS (Organization of American States). 2000. “Chief of Mission Report to the OAS
Permanent Council.” July 13, Washington, DC.
O’Ballance, Edgar. 2000. Sudan, Civil War, and Terrorism. New York: St. Martin’s Press.
O’Donnell, Guillermo, Philippe C. Schmitter, and and Laurence Whitehead. 1986. Tran-
sitions from Authoritarian Rule. Baltimore: Johns Hopkins University Press.
Ohlson, Thomas, and Stephen John Stedman. 1994. The New Is Not Yet Born: Conflict
Resolution in Southern Africa. Washington, DC: Brookings Institution Press.
Olonisakin, Funmi. 2007. Peacekeeping in Sierra Leone: The Story of Unamsil. Boulder,
CO: Lynne Rienner.
Oquist, Paul H. 1980. Violence, Conflict, and Politics in Columbia. New York: Academic
Press.
Orleans, Leo A. 1966. “A Note on Tibet’s Population.” China Quarterly 27:120–22.
Ospina, Sofi, and Tanja Hohe. 2001. “Traditional Power Structures and the Community
Empowerment and Local Governance Project: Final Report.” Report to the World
Bank, September, Dili.
Ottaway, Marina. 2000. Africa’s New Leaders: Democracy or State Reconstruction? Wash-
ington, DC: Carnegie Endowment for International Peace.
Ottaway, Marina, and Thomas Carothers. 2000. Funding Virtue: Civil Society Aid and
Democracy Promotion. Washington, DC: Carnegie Endowment for International
Peace.
Outram, Quentin. 2004. “Economy.” In Regional Surveys of the World: Africa South of the
Sahara 2004. London: Europa Publications.
Padro-Maurer, Rogelio. 1990. The Contras 1980–1989: A Special Kind of Politics. New
York: Praeger.
Paige, Jeffrey. 1997. Coffee and Power: Revolution and the Rise of Democracy in Central
America. Cambridge, MA: Harvard University Press.
Palmer, Robin. 1990. “Land Reform in Zimbabwe, 1980–1990.” African Affairs 89:
163–81.
References 295

Paris, Roland. 2004. At War’s End: Building Peace after Civil Conflict. Cambridge: Cam-
bridge University Press.
Paris, Roland, and Timothy Sisk. 2009. The Dilemmas of Statebuilding: Confronting the
Contradictions of Postwar Peace Operations. New York: Routledge.
Patten, Alan. 1996. “The Republican Critique of Liberalism.” British Journal of Political
Science 26 (1): 25–44.
Patterson, George N. 1960. Tibet in Revolt. London: Faber & Faber.
Perez- Stable, Marifeli. 1999. The Cuban Revolution: Origins, Course, and Legacy. New
York: Oxford University Press.
Perusse, Roland I. 1995. Haitian Democracy Restored. Lanham, MD: University Press of
America.
Pham, John-Peter. 2004. Liberia: Portrait of a Failed State. New York: Reed Press.
Picard, Elizabeth, 1996. Lebanon: A Shattered Country—Myths and Realities of the Wars
in Lebanon. New York: Holmes & Meier.
Poggo, Scopas S. 2009. The First Sudanese Civil War: Africans, Arabs and Israelis in the
Southern Sudan, 1955–1972. New York: Palgrave.
Political Instability Task Force. 2008. PITF Dataset. http: // globalpolicy.gmu.edu / pitf / .
Ponzio, Richard J. 2010. Democratic Peacebuilding: Aiding Afghanistan and Other Fragile
States. New York: Oxford University Press.
Posen, Barry R. 1993. “The Security Dilemma and Ethnic Conflict.” Survival 35 (1):
27–47.
Poulton, Robin-Edward, and Ibrahim ag Youssouf. 1998. A Peace of Timbuktu: Demo-
cratic Governance, Development and African Peacemaking. New York: United Nations
Publications.
Powell, Walter W., and Paul J. DiMaggio, eds. 1991. New Institutionalism in Organizational
Analysis. Chicago: University of Chicago Press.
Prado, Tania Palencia. 1998. “Advocates and Guarantors: Establishing Participative De-
mocracy in Post-war Guatemala.” Accord: An International Review of Peace Initiatives:
Negotiating Rights: The Guatemalan Peace Process. www.c-r.org / cr / acc_guat / prado
.htm.
Prevost, Gary. 1987. “The ‘Contra’ War in Nicaragua.” Conflict Quarterly 7 (3): 5–21.
PRIO (Peace Research Institute Oslo). 2003a. “Conflict Summaries: Central African
Republic.” Available from the author.
———. 2003b. “Conflict Summaries: Europe, Georgia (Abkhazia) and Georgia (South
Ossetia).” Available from the author.
———. 2003c. “Conflict Summaries: Peru.” Available from the author.
Prunier, Gerard. 1995. The Rwanda Crisis: History of a Genocide. New York: Colombia
University Press.
Przeworski, Adam. 1991. Democracy and the Market. Cambridge: Cambridge University
Press.
Quinn, Jason M., T. David Mason, and Mehmet Gurses. 2007. “Sustaining the Peace:
Determinants of Civil War Recurrence.” International Interactions 33:167–93.
296 References

Ramos-Horta, Jose. 1987. Funu: The Unfinished Saga of East Timor. Trenton, NJ: Red
Sea Press.
Ray, James Lee. 1998. “Does Democracy Cause Peace?” Annual Review of Political Science
1:27–46.
Rees, Edward. 2004. Under Pressure FALANTIL: Three Decades of Security Force Develop-
ment in Timor Leste: 1975–2004. Working Paper 139. Geneva: Geneva Centre for the
Democratic Control of Armed Forces.
Regan, Jane. 2003. “Haiti No. 2 on Danger List for Media.” Miami Herald, August 16.
Reilly, Ben, and Andrew Reynolds.1999. Electoral Systems and Conflict in Divided Societies.
Washington, DC: National Academies Press.
Reno, William. 1999. Warlord Politics and African States. Boulder, CO: Lynne Rienner.
———. 2008. “Anti-Corruption Efforts in Liberia: Are They Aimed at the Right Tar-
gets?” International Peacekeeping 15 (3): 387–404.
Reynal-Querol, Marta. 2002. “Ethnicity, Political Systems, and Civil Wars.” Journal of
Conflict Resolution 46 (1): 29–54.
Richardson, H. E. 1962. Tibet and Its History. Oxford: Oxford University Press.
Richmond, Oliver P., and Jason Franks, eds. 2009. Liberal Peace Transitions: Between
Statebuilding and Peacebuilding. Edinburgh: Edinburgh University Press.
Robinson, Randall. 2007. An Unbroken Agony: From Revolution to the Kidnapping of a
President. Basic Civitas.
Robinson, William I. 1996. Promoting Polyarchy: Globalization, US Intervention and He-
gemony. Cambridge: Cambridge University Press.
Roeder, Philip G. 2005. “Power Dividing as an Alternative to Ethnic Power Sharing.” In
Sustainable Peace: Power and Democracy after Civil Wars, edited by Philip G. Roeder
and Donald Rothchild. Ithaca, NY: Cornell University Press.
Roeder, Philip G., and Donald S. Rothchild, eds. 2005. Sustainable Peace: Power and De-
mocracy after Civil Wars. Ithaca, NY: Cornell University Press.
Romero, Simon. 2009. “Cocaine Trade Helps Rebels Reignite War in Peru.” New York
Times, March 18.
Ron, J. 2005. “Paradigm in Distress? Primary Commodities and Civil War.” Journal of
Conflict Resolution 49:443–50.
Rosecrance, Richard. 2000. The Rise of the Virtual State: Wealth and Power in the Coming
Century. New York: Basic Books.
Ross, Michael L. 1999. “Review: The Political Economy of the Resource Curse.” World
Politics 51:297–322.
Rothchild, Donald, and Philip G. Roeder. 2005. “Power Sharing as an Impediment to
Peace and Democracy.” In Sustainable Peace: Power and Democracy after Civil Wars,
edited by Philip G. Roeder and Donald Rothchild. Ithaca, NY: Cornell University
Press.
Ruane, Christopher. 2003. “Central African Republic: There May Be Trouble Ahead.”
Africa Today 417 (April).
Rubin, Barnett R. 2006. “Peace Building and State-Building in Afghanistan: Construct-
ing Sovereignty for Whose Security?” Third World Quarterly 27:175.
References 297

———. 2008. “The Politics of Security in Internationalized State-building Operations.

In Building States to Build Peace, edited by Charles T. Call with Vanessa Wyeth. Boul-
der, CO: Lynne Rienner.
Ruggie, John Gerard. 1993. “Territoriality and Beyond: Problematizing Modernity in
International Relations.” International Organization 47 (1): 139–74.
Russell, John. 2005. “Terrorists, Bandits, Spooks and Thieves: Russian Demonisation of
the Chechens before and since 9 / 11.” Third World Quarterly 26 (1): 101.
Russett, Bruce R. 1993. Grasping the Democratic Peace. Princeton, NJ: Princeton Univer-
sity Press.
Russett, Bruce R., and John O’Neal. 2001. Triangulating Peace: Democracy, Interdepen-
dence and International Organizations. New York: W. W. Norton.
Rwantabagu, Hermenegilde. 2001. “Explaining Intra-state Conflicts in Africa: The Case
of Burundi.” International Journal on World Peace 18 (2): 41–53.
Saideman, Stephen M., and R. William Ayres. 2000. “Determining the Causes of Irre-
dentism: Logit Analyses of Minorities at Risk Data from the 1980s and 1990s.” Journal
of Politics 62 (4): 1126–44.
Sakwa, Richard. 2005. Chechnya: From Past to Future. New York: Anthem Press.
Salehyan, Idean. 2008. “No Shelter Here: Rebel Sanctuaries and International Conflict.”
Journal of Politics 70 (1): 54–66.
———. 2010. “The Delegation of War to Rebel Organizations.” Journal of Conflict Resolu-
tion 54 (3): 493–515.
Samatas, Minas. 1986. “Greek McCarthyism: A Comparative Assessment of Greek Post–
Civil War Repressive Anticommunism and the US Truman-McCarthy Era.” Journal of
the Hellenic Diaspora 13 (3): 5–75.
Sambanis, Nicholas. 2000. “Partition as a Solution to Ethnic War: An Empirical Critique
of the Theoretical Literature.” World Politics 52 (4) ( July): 437–83.
———. 2004. “What Is Civil War? Conceptual and Empirical Complexities of an Op-
erational Definition.” Journal of Conflict Resolution 48 (6): 814–58.
———. 2007. “Short-Term and Long-Term Effects of UN Operations.” World Bank Eco-
nomic Review 22 (1): 9–32.
———. 2008. “Short- and Long-Term Effects of United Nations Peace Operations.”
World Bank Economic Review 22 (1): 9–32.
Sawyer, Amos. 1990. “Proprietary Authority and Local Administration in Liberia.” In
The Failure of the Centralized State: Institutions and Self-Governance in Africa, edited by
James S. Wunsch and Dele Olowu. Boulder, CO: Westview Press.
Scott, Amy. 2008. “The United Nations Peacebuilding Commission: An Early Assess-
ment.” Journal of Peacebuilding and Development 4 (2): 7–19.
Serrill, Michael S. 1991. “General without an Army.” Time, January 21.
Shacochis, Bob. 1999. The Immaculate Invasion. New York: Viking Press.
Shakya, Tsering. 2000. The Dragon in the Land of Snows: A History of Modern Tibet since
1947. New York: Penguin Press.
Shermatova, Sanobar. 2003. “The War for Chechen oil.” Capitalism, Nature, Socialism
14 (1): 113.
298 References

Shillington, Kevin. 2005. “Arms, Armies: Post-Colonial Africa.” In Encyclopedia of African

History, Volume 1. Boca Raton, FL: CRC Press.
Shils, Edward. 1957. “Primordial, Personal, Sacred, and Civil Ties: Some Particular Ob-
servations on the Relationships of Sociological Research and Theory.” British Journal
of Sociology 8 (2): 130–45.
Shoesmith, Dennis. 2003. “Timor-Leste: Divided Leadership in a Semi-Presidential Sys-
tem.” Asian Survey 43 (2): 231–52.
Sibanda, Eliakim M. 2005. The Zimbabwe African People’s Union 1961–1987: A Political
History of Insurgency in Southern Rhodesia. Trenton, NJ: Africa World Press.
Simonsen, Sven Gunnar. 2006. “The Authoritarian Temptation in East Timor: Nation-
building and the Need for Inclusive Governance.” Asian Survey 46 (4): 575–96.
SIPRI (Stockholm International Peace Research Institute). 1994. SIPRI Yearbook 1994.
Stockholm: SIPRI.
Sisk, Timothy. 1996. Power Sharing and International Mediation in Ethnic Conflicts. Wash-
ington, DC: US Institute of Peace Press.
———. 2009. “South Africa: Negotiating Democracy after Apartheid.” In International
Mediation in Civil Wars: Bargaining with Bullets. London: Taylor and Francis.
Skocpol, Theda. 1979. States and Social Revolutions. Cambridge: Cambridge University
Press.
Snyder, Jack. 2000. From Voting to Violence: Democratization and Nationalist Conflict. New
York: W. W. Norton.
Somerville, Carolyn M. 1986. Drought and Aid in the Sahel: A Decade of Development
Cooperation. Boulder, CO: Westview Press.
Sriram, Chandra Lekha. 2008. Peace as Governance: Power-Sharing, Armed Groups and
Contemporary Peace Negotiations. London: Palgrave Macmillan.
Stanley, William, and David Holiday. 2002. “Broad Participation, Diffuse Responsibil-
ity: Peace Implementation in Guatemala.” In Ending Civil Wars: The Implementation of
Peace Agreements, edited by Stephen John Stedman, Donald S. Rothchild, and Eliza-
beth M. Cousens. Boulder, CO: Lynne Rienner.
Stedman, Stephen John. 1993. Peacemaking in Civil War: International Mediation in Zim-
babwe, 1974–1980. Boulder, CO: Lynne Rienner.
———. 1996. “Conflict and Conciliation in Sub- Saharan Africa. In The International
Dimensions of Economic Conflict, edited by Michael Edward Brown. Cambridge, MA:
Center for Science and International Affairs.
———. 1997. “Spoiler Problems in Peace Processes.” International Security 22:5–53.
———. 2002. “Introduction.” In Ending Civil Wars: The Implementation of Peace Agree-
ments, edited by Stephen John Stedman, Donald S. Rothchild, and Elizabeth M.
Cousens. Boulder, CO: Lynne Rienner.
Stedman, Stephen John, Donald S. Rothchild, and Elizabeth M. Cousens, eds. 2002. End-
ing Civil Wars: The Implementation of Peace Agreements. Boulder, CO: Lynne Rienner.
Steele, Jonathan. 2002. “Nation Building in East Timor.” World Policy Journal 19 (2): 76–87.
Stewart, Frances, and Graham Brown. 2006. The Implications of Horizontal Inequality for
Aid. CRISE Working Paper 36. Oxford: CRISE, University of Oxford.
References 299

Stoll, David. 2005. “The Nicaraguan Contras: Were They Indios?” Latin American Poli-
tics and Society 45 (3): 145–57.
Street, Anne M., Howard Mollet, and Jennifer Smith. 2008. “Experiences of the United
Nations Peacebuilding Commission in Sierra Leone and Burundi.” Journal of Peace-
building and Development 4 (2): 33–46.
Strohmeyer, Hansjörg. 2001. “Collapse and Reconstruction of a Judicial System: The
United Nations Missions in Kosovo and East Timor.” American Journal of International
Law 95 (1): 46–63.
Suchman, Mark C. 1995. “Managing Legitimacy: Strategic and Institutional Approaches.”
Academy of Management Review 20 (3): 571–609.
Suhrke, Astri. 2001. “Peacekeepers as Nation-Builders: Dilemmas of the UN in East
Timor.” International Peacekeeping 8 (4): 1–20.
Suhrke, Astri, and Ingrid Samset. 2007. “What’s in a Figure? Estimating Recurrence of
Civil War.” International Peacekeeping 14 (2):195–203.
Suliman, Mohamed. 1997. “Civil War in Sudan: The Impact of Ecological Degrada-
tion.” Contributions in Black Studies 15:99–121. http: // scholarworks.umass.edu / cgi /
viewcontent.cgi?article=1100&context=cibs.
Taylor, Ian. 2010. “Liberal Peace, Liberal Imperialism: A Gramscian Critique.” In Palgrave
Advances in Peacebuilding, edited by Oliver P. Richmond. London: Palgrave Macmillan.
Taylor, Lewis. 2006. The Shining Path: Guerrilla War in Peru’s Northern Highlands. Liver-
pool: Liverpool University Press.
Thier, Alex. 2002. “The Road Ahead: Political and Institutional Reconstruction in Af-
ghanistan.” Third World Quarterly 23 (5): 893–907.
Thomas, Lowell. 1959. The Silent War in Tibet. Garden City, NY: Doubleday.
Thomson Reuters Foundation. 2001. “Lucrative Timber Trade Fuels Liberian War Mis-
ery.” http: // reliefweb.int / node / 86035.
Tilly, Charles. 1975. “Reflections on the History of European State-Making.” In The For-
mation of National States in Western Europe, edited by Charles Tilly. Princeton, NJ:
Princeton University Press.
Timor-Leste Armed Violence Assessment. 2009. Elections Violence in Timor-Leste:
Mapping Incidents and Responses. Issue Brief 3. www.timor-leste-violence.org / pdfs /
Timor-Leste-Violence-IB3-ENGLISH.pdf.
Tishkov, Valery. 1999. “Ethnic Conflicts in the Former USSR: The Use and Misuse of
Typologies and Data.” Journal of Peace Research 36 (5): 571–91.
Torrijos, Giancarlo Soler. 2008. Under the Shadow of the United States: Democracy and
Regional Order in the Latin Caribbean. Boca Raton, FL: Brown Walker Press.
Trimberger, Ellen Kay. 1978. Revolution from Above. Piscataway, NJ: Transaction.
Trouillot, Michel-Rolph. 1989. Haiti: State against Nation: The Origins and Legacy of
Duvalierism. New York: Monthly Review Press.
———. 1994. “Culture, Color and Politics in Haiti.” In Race, edited by Steven Gregory
and Roger Sanjek. Piscataway, NJ: Rutgers University Press.
Tuck, Christopher. 2000. ““Every Car or Moving Object Gone”: The ECOMOG Inter-
vention in Liberia.” African Studies Quarterly 4 (1): 1–16.
300 References

Tulchin, Joseph, and Gary Bland. 1992. Is There a Transition to Democracy in El Salvador?
Boulder, CO: Lynne Rienner.
Tvedten, Inge. 1997. Angola: Struggle for Peace and Reconciliation. Boulder, CO: West-
view.
United Nations Department of Public Information. N.d. “Liberia: UNOMIL, Back-
ground.” Section on “December 1995–November 1996.” www.un.org/Depts/DPKO/
Missions / unomil_b.htm.
United Nations Development Program. 2000. Human Development Report 2000. New
York: United Nations.
———. 2007. Democratization and Civil Society Empowerment Programme. New York:
United Nations Development Program.
United Nations Secretary-General, Office of. 1997a. Final Report of the Secretary General
on the UN Mission to Liberia. New York: United Nations. http: // daccess-ods.un.org /
TMP / 287225.9.html.
———. 1997b. Twenty-Second Report on the UN Observer Mission in Liberia. New York:
United Nations. www.undemocracy.com / S-1997-237.
———. 2007. Fifteenth Progress Report of the Secretary-General on the UN Mission in Libe-
ria. New York: United Nations.
United Nations Special Commission of Inquiry for Timor-Leste. 2006. “Report of the
United Nations Independent Special Commission of Inquiry for Timor-Leste.” Ge-
neva: Office the UN High Commissioner for Human Rights.
US Agency for International Development. 2005. At Freedom’s Frontiers: A Democracy
and Governance Strategic Framework. Report PD-ACF-99. Washington, DC: US
Agency for International Development.
US Department of Defense. 2009. Progress toward Security and Stability in Afghanistan.
Report to the Congress. Washington, DC: US Government Printing Office.
US Department of State. 1998. Background Notes: Georgia. Washington, DC: US Govern-
ment Printing Office.
———. 1999. Liberia: Country Report on Human Rights Practices for 1998. Washington,
DC: US Department of State. www.state.gov / g / drl / rls / hrrpt / 1999 / 254.htm.
———.2002. Patterns of Global Terrorism, 2002. Washington, DC: US Department of
State. www.state.gov / s / ct / rls / crt / 2002 / .
———. 2008. “Background Note: Burundi.” www.state.gov / r / pa / ei / bgn / 2821.htm.
———. 2010a. “Background Note: Central African Republic.” www.state.gov / r / pa / ei /
bgn / 4007.htm.
———. 2010b. “Background Note: Mali.” www.state.gov / r / pa / ei / bgn / 2828.htm.
Utas, Mats. 2005. “The Reintegration and Remarginalization of Youth in Liberia.” In No
Peace, No War, edited by Paul Richards. Athens: Ohio University Press.
Uvin, Peter. 1998. Aiding Violence: The Development Enterprise in Rwanda. West Hartford,
CT: Kumarian Press.
———. 1999. “Ethnicity and Power in Burundi and Rwanda: Different Paths to Mass
Violence.” Comparative Politics 31 (3): 253–71.
References 301

Van den Brande, K., with S. Happaerts and H. Bruyninckx. 2008. “The Role of the Sub-
national Level of Government in Decision-Making for Sustainable Development.”
Institute for International and European Policy.
Vinding, Diana. 2002. The Indigenous World, 2001–2002. Copenhagen: International
Work Group for Indigenous Affairs.
Wakoson, Elias Nyamlell. 1993. “The Politics of Southern Self-Government.” In Civil
War in the Sudan, edited by M. W. Daly and Ahmad Alawad Sikainga. London: British
Academic Press.
Wallensteen, Peter. 2008. “International Responses to Crises of Democratization in War-
Torn Societies.” In From War to Democracy, edited by Anna K. Jarstad and Timothy D.
Sisk. New York: Cambridge University Press.
Wallensteen, Peter, and Margareta Sollenberg. 1997. “Armed Conflicts Conflict Termi-
nation, and Peace Agreements, 1989–96.” Journal of Peace Research 34 (3): 339–58.
Walter, Barbara F. 1997. “The Critical Barrier to Civil War Settlement.” International Or-
ganization 51:335–64.
———. 2002. Committing to Peace: The Successful Settlement of Civil Wars. Princeton, NJ:
Princeton University Press.
———. 2004. “Does Conflict Beget Conflict? Explaining Recurring Civil War.” Journal
of Peace Research 41 (3): 371–88.
Walter, Barbara F., and Jack L Snyder, eds. 1999. Civil Wars, Insecurity, and Intervention.
New York: Columbia University Press.
Weinstein, Warren. 1972. “Conflict and Confrontation in Central Africa: The Revolt in
Burundi.” Africa Today 19 (4): 17–37.
Welt, Cory. 2010. “The Thawing of a Frozen Conflict: The Internal Security Dilemma
and the 2004 Prelude to the Russo-Georgian War.” Europe-Asia Studies 62 (1):
63–97.
West, Ronald. 2007. “Lawyers, Guns, and Money: Justice and Security Reform in East
Timor.” In Constructing Justice and Security after War, edited by Charles T. Call. Wash-
ington, DC: US Institute of Peace Press.
Whitman, Jim. 1999. Peacekeeping and the UN Agencies. Portland: Frank Cass.
Williams, Gabriel I. H. 2006. Liberia: The Heart of Darkness. Victoria: Trafford Publishing.
Wimmer, Andreas. 2002. Nationalist Exclusion and Ethnic Conflict. Cambridge: Cam-
bridge University Press.
Wimmer, Andreas, Lars-Erik Cederman, and Brian Min. 2009. “Ethnic Politics and
Armed Conflict: A Configurational Analysis of a New Global Data Set.” American So-
ciological Review 74 (2): 316–37.
Winslow, Charles. 1996. Lebanon: War and Politics in a Fragmented Society. London:
Routledge.
Wolff, Stefan. 2009. “Complex Power- Sharing and the Centrality of Territorial Self-
Governance in Contemporary Conflict Settlements.” Ethnopolitics 8 (1): 27–45.
Wood, Elizabeth Jean. 2003. Insurgent Collective Action and Civil War in El Salvador. Cam-
bridge: Cambridge University Press.
302 References

World Bank. 2002. Haiti: Country Assistance Evaluation. Operations Evaluation Depart-
ment Report 23637. Washington, DC: World Bank.
———. 2011. World Development Report 2011: Conflict, Security and Development. Wash-
ington, DC: World Bank.
Yoh, John G. Nyuot. 2003. “Peace Processes and Conflict Resolution in the Horn of
Africa.” African Security Review 12 (3).
Young, Crawford. 1965. Politics in the Congo: Decolonization and Independence. Princeton,
NJ: Princeton University Press.
———. 1994. The African Colonial State in Comparative Perspective. New Haven, CT:
Yale University Press.
Zahar, Marie-Jöelle. 2003. “Reframing the Spoiler Debate in Peace Processes.” In Con-
temporary Peacemaking: Conflict, Violence and Peace Processes. New York: Palgrave
Macmillan.
———. 2005. “Power Sharing in Lebanon: Foreign Protectors, Domestic Peace, and
Democratic Failure.” In Sustainable Peace: Power and Democracy after Civil Wars, edited
by Philip G. Roeder and Donald Rothchild. Ithaca, NY: Cornell University Press.
Zartman, I. William. 1977. “Negotiation as a Joint Decision-Making Process.” Journal of
Conflict Resolution 21:619–38.
Zartman, I. William, ed. 1995a. Collapsed States. Boulder, CO: Lynne Rienner.
———. 1995b. Collapsed States: Disintegration and Restoration of Legitimate Authority.
Washington, DC: US Institute of Peace Press.
———. 1995c. Elusive Peace: Negotiating an End to Civil Wars. Washington, DC: Brook-
ings Institution Press.
Zartman, I. William, and Maureen R. Berman. 1982. The Practical Negotiator. New
Haven, CT: Yale University Press.
 Index

Tables are indicated by t following page numbers. Arabic surnames starting with “al-” are alphabetized by the fol-
lowing part of the name.

Abacha, Sani, 83 anticommunist regimes, 196–202

Abkhazia as separatist struggle, 46, 110, 111, 210n15 Anti-Terrorism Unit (Liberia), 80–81
Abuja I accord (1995, Liberia), 80, 84 Anya-Nya (Sudan), 103, 105
Abuja II accord (1996, Liberia), 75, 78, 80, 84 APODETI (Popular Democratic Association of
Accra Peace Agreement (2003, Liberia), 89, 90 Timor), 137
Addis Ababa agreement (1972, Sudan), 103–4 Arab Deterrent Force (Lebanon), 166
Adebajo, Adekeye, 73, 74, 81 Arab League, 165, 167
Afghanistan: election candidates in, 265–66; Arab Spring (2011), 240
exporting from, 215; failed powersharing in, 194; Arafat, Yasser, 165
“failed state” in, 34; informal external influ- Argentina and Contras (Nicaragua), 175
ence on, 248; international decision making on Aristide, Jean-Bertrand, 129–36, 150, 160n3, 202,
interim government for, 249–50; national forces 250. See also Haiti
on ground in, 249; NATO in decision-making Aristotle, 33, 220
role for, 248; no peace agreement in, 2; state armed conflicts, defined, 11
building in, 35, 48n9; war criminals within Armed Forces for the Liberation of East Timor
Afghan Cabinet, 228 (FALINTIL), 137–38, 140, 145
African Union, 82, 95n23, 256, 271, 272 Armed Forces of Liberia, 75, 76, 79
Ag Bahanga, Ibrahim, 171 Armed Forces Revolutionary Council (AFRC),
Ag Ghali, Iyad, 170–71 77, 83
Akosombo Agreement (Liberia), 80 Armée Populaire pour la Restauration de la Répub-
Algeria: and Mali, 168; war for independence in, lique et la Démocratie (APRD; Central African
21n16 Republic), 129
Alkatiri, Mari, 141 Artibonite Resistance Front for the Overthrow of
Alliance Démocratique du 23 mai pour le Change- Jean-Bertrand Aristide, 133
ment (ADC), 171 Association of Southeast Asian Nations, 271
Alliance for Democracy, 169 Atlas, Pierre M., 123, 149, 180
Alliance Touareg Nord Mali pour le Changement Australia’s peacekeeping in East Timor, 141, 144,
(ATNMC), 171 145, 204
Angola: interim governance, role of domestic ac- authoritarianism. See repression
tors in, 252; peace accord (2002) attributed to Avril, Prosper, 131
killing of rebel leader in, 191; peace agreement Azerbaijan: fragile peace in, 203; inclusionary
and failure to secure peace in, 34; peacekeeping behavior, yet recurrence in, 201
troops withdrawn from, 9; postwar national elec-
tions leading to violence in, 241–42, 259–60; Baathists (Iraq), 250
sovereignty of, 247; strategic victory in, 184 Baev, Pavel, 106–7, 109
Annan, Kofi, 85, 144 Baghdad Pact, 164
304 Index

Balaguer, Jorge, 195 case-study approach, 9–15, 232–33; “structured,

Bangui Agreements (1997, Central African Repub- focused” case comparison, 180–81
lic), 125–26 Catholic Commission for Justice and Peace (Zim-
Barnett, Michael N., 33, 220–21 babwe), 147
Basaev, Shamil, 108 cease-fires: and “frozen” conflicts, 202–6; nature of,
Bennett, Andrew, 16 2, 12; third-party guarantors of, 28. See also peace
Bermudez, Enrique, 175 agreements
Binningsbø, Helga Malmin, 40 Cederman, Lars-Erik, 58–59
Boley, George, 73, 76, 79 Cedras, Raoul, 132
Bolivia, 196 CEMAC (Economic Community of Central African
border disputes, 46 States), 127–28
Bosch, Juan, 195–96 Central African Republic, 1, 124–29; abrogation
Bosnia: elections in, 260–61, 265; formal interna- of powersharing arrangements in, 99, 194, 215;
tional influence on, 247–48; international troop Bangui Agreements (1997), 125–26, 205;
deployments to, 200, 203–4; marginalization coup in, 10, 128–29, 234; economic factors in,
of civil society in, 253; political powersharing 214; ethnic and religious composition of, 125,
in, 188, 191; state building in, 35; withdrawal of 128–29; exclusion precipitating recurrence in,
peacekeeping troops from, 33–34 98, 124–25, 149; French and regional troops in,
Boutros-Ghali, Boutros, 31 150; initial conflict (1996–97), 125; institu-
Boyd, Chuck, 175 tional weakness in, 99, 124–25, 127, 159, 214;
Boyle, Francis, 108 recurrence in, 18, 126–28; subsequent armed
Bozizé, François, 126–29 conflict and coup attempts (2002–present),
Brahimi, Lakhdar, 249 124, 128–29; withdrawal of peacekeeping forces,
Brown, Timothy C., 182n5 effect of, 149, 204–5
Burundi, 1, 150–52, 154–55; abrogation of pow- Central Intelligence Agency (CIA; US), 118, 132,
ersharing arrangements in, 99, 194; alternative 175–76
explanations for recurrences, 150, 158–60; Central Revolutionary Council (Liberia), 73
background of, 152; chronic exclusionary Chad, relationship with Central African Republic,
behavior in, 123, 154, 155, 160, 190; democrati- 127
zation in, 157, 159, 230, 241; economic factors in, Chalabi, Ahmed, 263
158; ethnic violence in, 154–55, 189; informal Chamorro, Violeta, 174, 276n18
external influence on, 248; institutional weakness Chamoun, Camille Nimr, 163–65
in, 189, 214; interim governance, role of domestic Chandler, David, 35
actors in, 252; multiple recurrences in, 18, 21n19; Chechyna, 1, 106–10; abrogation of powersharing
recurrence and mass atrocities (1988), 154–55; arrangements in, 99, 194; alternative explana-
recurrence as failure of inclusion (1993–2006), tions for recurrence in, 109–10; economic
157–58, 189; regional factors in recurrences, 159; factors in, 109; election candidates, Russian
severity of exclusionary behavior in, 99; war and interference with, 264–65; first Chechen war
mass atrocities (1972), 154 (1994–96), 12, 106–7; international factors
Bush, George W., 77, 90, 135, 250, 263 in recurrence in, 109–10; Khasavyurt Accords
Buyoya, Pierre, 154–55, 157, 189–90 (1996), 107–8, 119; recurrence precipitated by
exclusion in, 98; second Chechen war (1999),
Call, Charles T., 66n6 12, 18, 108–9, 119; territorial autonomy, reneging
Cambodia: exclusionary behavior in, 214; formal on, 119, 191
international influence on, 247; interwar peace, checks and balances, 221
perceptions of, 12, 15; negotiated settlement in, Chesterman, Simon, 222
184; postwar repression in, 219; resource curse Chevron, 105
in, 27 China fighting in Tibet. See Tibet
Cannibal Army (Haiti), 133 chronic exclusion. See exclusion
capacity vs. legitimacy, 7, 47, 222, 244 Cisse, Lamine, 126
Carothers, Thomas, 241, 243 Civilian Mission in Haiti, 132
Carr, Burgess, 103 civilians participating in wars, 234–35
Carter, Jimmy, 132 civil society organizations, 227, 253, 267, 269–70
Index 305

civil wars: defined, 11; devastation of, 2; extrastate Contras (Nicaragua), 174, 175–76
aspects of, 233; new civil war in same country Correlates of War Project, 9, 21n11, 27
or territory, 14t; number of battle-related deaths Costa Rica: regional assistance in civil war in, 200;
for internal armed conflict to qualify as, 12; victorious anticommunist regime in, 197
scholarship on, 59–65, 233–35; theories about, Côte d’Ivoire and Liberian unrest, 71, 73, 81, 82
26–30. See also nonrecurrence; onset of civil war; coups, 10, 128–29. See also specific countries
recurrences critical peacebuilding theorists, 35–36, 44, 223–24,
classification of conflicts, 13–15, 14–15t, 21n14, 51 225t, 227
Clinton, Bill, 132, 135, 260 Croatia, international troop deployments to, 200,
CNRT. See National Council of Timorese Resis- 203
tance Cuba: nonrecurrence in, 19; victorious Communist
coercion, 28, 130 regime in, 196
Cold War: and anticommunist regimes, 196–202; Cyprus: international troop deployments to, 200,
breakup of Soviet Union, wars related to, 203; 203; negotiated settlement in, 184; no power-
and ideology as mobilizing principle, 199; nego- sharing, but inclusionary behavior in, 195–96,
tiated settlements during, 184, 200; post–Cold 198
War era vs., 5, 184, 201, 207; and regime types,
183; and victorious Communist regime in Cuba, Dailey, Peter, 135
196 Dalai Lama, 116–18
Collier, Paul: adhering to neoliberal macroeconom- Daly, M. W., 104, 105, 106
ics, 63; on “conflict trap,” 124, 127; on economic Dayton Accords (1995, Bosnia), 253, 260
factors, 6, 72, 105, 130–31, 135; on ethnicity DDR. See disarmament, demobilization, and
factor, 48n2; on postwar countries’ risk level for reintegration
recurrence, 266; on rate that armed conflicts decision “moments” in postwar political processes,
recur, 61; on rebel leaders hiding their true mo- 246–73; elections and end of interim gover-
tives, 95n21; on religious factor, 167; on resource nance, 257–66; interim governance, decisive
curse and civil war, 26–28, 127; on variables and role of domestic actors in, 252–54; interim gov-
regressions, 52–54 ernance, decisive role of international actors in,
Colombia: natural resources in, 27, 215; renewal of 248–52; state design, 254–56; state legitimacy
offensive actions in, 48n7; successful powershar- after elections, 266–73
ing in, 194 Defense Forces of Timor-Leste (FDTL), 140–43
colonialism, effect of, 45 Del Castillo, Graciana, 32
Commission of Inquiry (Sudan), 102 demobilization. See disarmament, demobilization,
Committee to Protect Journalists, 135 and reintegration (DDR)
Common Market for Eastern and Southern Africa, “democracy promotion” and democratization:
272 consociational democracy, 40; negative effects of,
Community of Sahel-Saharan States, 127 7, 240–44; and prevention of recurrence, 28. See
comparativists, 232 also specific countries
Condominium Agreement (Sudan), 101 democratic peace thesis, 230
conflict diamonds, 84, 87, 90 Democratic Republic of the Congo: civilian victims
conflict resolution. See negotiated settlements; in, 223; failed powersharing in, 194; multiple
peace agreements; peacebuilding armed conflicts in, 201; natural resources in,
Congo, Democratic Republic of. See Democratic 215
Republic of the Congo Des Forges, Alison, 153
Conneh, Sekou, 77 De Soto, Alvaro, 32
Conselho Nacional de Resistência Maubere Detzel, Julian, 90
(CNRM; East Timor), 138 Development Assistance Committee (OECD-DAC),
consociational democracy, 40 221–22, 271
consolidation of peace, 3, 4, 207, 226. See also Development Research Group (World Bank), 26
peacebuilding De Waal, Alex, 106
constitutionalism, 221 diffusion, 183–84
constructivist theory, 217 Dillon, Karin, 169, 172
Contract and Monopolies Commission (Liberia), 91 Dillon, Sam, 175
306 Index

disarmament, demobilization, and reintegration Economic Community of Central African States

(DDR), 28, 84–85, 89, 172, 181n2, 193, 209n5, (CEMAC), 127–28
242, 259–60 Economic Community of West African States
disillusion of recurrence, 3 (ECOWAS), 71, 74–75, 80, 82–84, 90, 95n23,
Djibouti, powersharing in, 187–88 272–73
Doe, Samuel, 73, 80–81 economic factors, 6, 8, 27, 72, 119–20, 144–45, 214,
Dokie, Sam, 78 216. See also natural resources; specific countries
Dominican Republic: international troop deploy- economic liberalism, 244
ments to, 200, 203; no powersharing, but economic powersharing, 192
inclusionary behavior in, 195, 198 ECOWAS Cease-Fire Monitoring Group
Doyle, Michael W.: on capacity, 7, 20n5; civil wars (ECOMOG), 74–75, 83–85, 87, 95n23, 204
data set of, 9, 51, 67n16; on international aid, electoral representation and elections: candidate
63–64; on liberal peacebuilding, 32; on peace- maneuvering, 263–66, 276n18; decision “mo-
keeping troops’ role in maintaining stability, 205; ments” in postwar political processes, 257–66;
proxy use by, 62; on rebel cohesion, 173, 181n3; delayed elections in Bosnia, 260–61; demobi-
on severity of war, 28; on UN peace efforts’ ef- lization prior to national elections (lesson from
fectiveness, 30, 90 Angola), 259–60; importance of, 258–59; as
Dudayev, Djokhar Musayevich, 107 inclusionary process, 216; in less democratic
Dupuy, Kendra, 90, 134 regimes, 217, 275n4; local and legislative elec-
duration of prior war as factor in recurrence, 28 tions, 255–56, 261; as process legitimacy, 226;
DuToit, Pierre, 268 sequencing and timing of, 259–66; violence
Duvalier, Jean-Claude “Baby Doc,” 130–31 linked to, 241–42
elite political behavior. See exclusion; inclusion
East Timor, 1, 136–46; abrogation of powersharing El Salvador: interim governance, role of domestic
arrangements in, 99, 215; background of, 136– actors in, 252; international pressure and elec-
37; conflicts with few deaths in, 10, 159; divisions tions in, 255–56, 265; negotiated settlement in,
among former allies leading to recurrence in, 149, 184; nonrecurrence in, 19; organized criminal
159–60; economic factors in, 144–45; electoral violence in, 20n7; security powersharing in,
violence in, 258; end of first war in, 138–39; 190–91; sunset provisions on powersharing in,
exclusionary behavior in, 98, 122, 142–43, 146, 256; UN mediation of peace in, 32, 200
149, 202; institutional weakness in, 99, 143–44, embedded legitimacy, 45, 226
149, 159, 214; interim government, international ethnicity: and bloody wars, effect on peace duration,
decision making on, 250–51; internal armed con- 64; and civil war onset, 26, 28; recurrence and
flict (2006), 140–41; international and regional ethnic fractionalization, 28–29, 48n7, 54–59,
factors in, 46, 145–46; international transitional 55–56t, 66n5, 216; social contract with diverse
authority in, 247; international troops and ethnic groups as basis for sustained long-term
stability in, 144; languages used in, 136–37, 145; peace, 245–46. See also specific countries
local and legislative elections before presidential European Union, 204, 256, 271–72
elections in, 261–63; no peace agreement in, 2; Evangelista, Matthew, 109
peacekeeping considered successful, yet conflict exclusion: agents (who do excluding), 38; chronic
recurred in, 2, 9; postwar government reneging exclusionary behavior, 122–23, 150, 153–54,
on commitments in, 150; rebel cohesion in, 180; 160; as decisive factor, 4, 6, 20n3, 100, 199t,
religious and ethnic factors in, 136; UN’s role 206, 213–14; defined, 4, 37, 161n18; forms of,
in, 3, 139–40, 141, 144–45, 146, 251, 255; war 38–39; history of concept, 37–39; interna-
for independence followed by civil war in, 13, tional actors’ ability to influence national elites’
18, 137–38; withdrawal of peacekeeping troops behavior, 236, 255; nonseparatist recurrences
from, 33–34, 149–50, 204 precipitated by, 122, 123–24; not as decisive
East Timor Transitional Authority, 140 factor, 162–63, 179–80; separatist recurrences
Eckstein, Harry, 16, 20n10, 97 precipitated by, 98; severity of, 38, 99; subjective
ECOMOG. See ECOWAS Cease-Fire Monitoring perceptions among former warring factions, 4–5;
Group targets (who are excluded), 38. See also specific
Economic and Monetary Community of Central countries
Africa, 210n16 Ex-Liberators (Central African Republic), 128–29
Index 307

external actors’ role, 236–37; deciding on transi- Garang, John, 101, 105
tional arrangements and rulers, 247–54, 273–74; Gemayel, Pierre, 164
and elections, 259; shaping legitimacy from genocide, 48n10, 150, 151, 154, 239
without, 45–47, 237–38, 270–73; strategy of, George, Alexander, 16, 180
245–47. See also peacekeeping operations Georgia and South Ossetia, 110–14; abrogation of
powersharing arrangements in, 99; coding in
“failed peace,” 1, 25–49; civil wars and ethnic data bases for, 110–11; economic factors in, 114;
conflict, 26–30; defined, 8, 11; exclusion, inclu- ethnic factors in, 112; failed powersharing in,
sion, and legitimacy, 36–47; reasons for, 25; 194; first conflict (1991–93), 111; fragile peace
recurrence as universally accepted indicator of, 8; in, 203; international factors in, 46, 233; recur-
standards for, 8–9 rence precipitated by exclusion in, 98; second
failed states, 6 conflict (2004), 18, 112–13, 210n15; separatist
FALINTIL. See Armed Forces for the Liberation of recurrences in, 97; territorial autonomy, reneging
East Timor on, 119, 120, 191
Famille Lavalas party (Haiti), 133–34 German, Tracey, 109
F&L. See Fearon and Laitin data sets Ghani, Ashraf, 7, 35, 222
Farabundo Marti National Liberation Front (El Gleditsch, Kristian Skrede, 61
Salvador), 191 globalization and state legitimacy, 244–45
Fearon and Laitin data sets, 74; on Burundi and Goldstein, Melvyn, 115
Rwanda, 151; on Central African Republic, 124; governance: interim governance, decisive role of
civil war at end for at least one year, 12–13; civil domestic actors in, 252–54; interim gover-
wars data set of, 9–10, 20n1, 51, 66nn1–6, 67n7, nance, decisive role of international actors in,
67n13; classification of civil wars not always 248–52; as key to peace, 33–35, 233. See also
clear, 234; collapse of Soviet Union, effect of, state building
121n4; colonial wars for independence included, Governance Reform Commission (Liberia), 91
21n16; conditions favoring insurgency, 27; ethnic Great Lakes region (Africa). See Burundi; Rwanda
fractionalization, 29, 54–56, 55–56t; GDP as greed / grievance debate, 26–27, 58
factor, 53–54; proxies in, 61–62; and religious Greek Civil War, 197, 200
fractionalization, 54–56; and resource curse, 74; grievance-based approaches: and civil wars’ onset,
on South Korea’s nonrecurrent civil war, 197; on 26–27; in recurrence, 72
South Ossetia and Abkhazia, 110, 111; variables gross domestic product (GDP). See economic factors
and regressions in, 52–54, 53t Guatemala: civil societies, involvement in peace ne-
Fifteenth of September Legion (Nicaragua), 175 gotiations, 253; economic powersharing, 192; in-
Fifth Brigade (Zimbabwe), 147 terim governance, role of domestic actors in, 252;
Figueres, “Pepe,” 200 negotiated settlement in, 184; nonrecurrence,
Firaku (Timorese people), 136–37 19; organized criminal violence, 20n7; postwar
FOMUC (Central African Republic), 128 repression, 219; UN military observers, 200
foreign troops. See international troops; peacekeep- Guinea and Liberian unrest, 71, 76, 82, 150
ing operations Gurr, Ted Robert, 26, 37, 41
Forman, Shepard, 139–40 Gurses, Mehmet, 64
Fortna, Virginia Page, 28, 30, 48n5, 64, 90, 205 Gusmao, Xanana, 138–41, 143, 251, 262, 265
France: aid to Central African Republic, 128, 150; Guzman, Abimael, 177, 179
role in Haiti, 133
Frangieh, Suleiman, 165 Habibie, B. J., 138
Franks, Jason, 35, 36, 218, 221, 223 Habyarimana, Juvenal, 155–56
FRETILIN (Revolutionary Front of East Timor), Haiti, 1, 129–36; abrogation of powersharing
137–38, 140–43, 146 arrangements in, 99, 215; background of, 129;
Front for the Advancement and Progress of Haiti, coup attempts in, 131; economic factors in,
132 130–31; elections of Aristide, 135–36; exclusion
Front for the Restoration of Unity and Democracy precipitating recurrence in, 98, 122, 123, 129,
(FRUD; Djibouti), 187–88 134, 135, 202; explanatory factors for recurrence
“frozen conflicts,” 6, 202–6, 214 in, 130, 134–36; first armed conflict (1990–95),
Fukuyama, Francis, 3, 35, 222 130–32; informal external influence on, 248;
308 Index

Haiti (continued) international actors. See external actors’ role

institutional weakness in, 99, 134, 159, 214; international aid, 63–64, 82–83, 133, 135, 202, 222,
interim government, international decision 240
making on, 250; international forces’ influence International Court of Justice on border disputes, 46
on, 248; no peace agreement in, 2; onset of peace International Criminal Court: and human rights
in, 132; peacekeeping considered successful, yet violators, 228; and state sovereignty, 239
conflict recurred, 2, 9; political upheaval in, 10; International Crisis Group, 111
recurrence in, 18, 132–34; religious factors in, International Force for East Timor (INTERFET),
129; UN peacekeeping in, 3, 134, 136; US role in, 144
130–35, 150; withdrawal of peacekeeping troops International Monetary Fund, 32, 63, 105, 131
from, 33–34, 204 international troops: deployments to prevent or
Hamas, 265 end war, 144, 200, 208; and “frozen” conflicts,
Hamilton, Alexander, 221 6, 202–6; in legitimacy building, 271. See also
Haradinaj, Ramush, 228 peacekeeping operations
Harrington, Andrew, 142–43, 145 Iraq: elections in, 261, 263; international forces’
Hartzell, Caroline, 28, 40, 63, 64, 130–31, 209n1 influence on, 248; international transitional
Hegre, Håvard, 28 authority in, 247; multiple recurrences in, 21n19;
Hoddie, Matthew, 28, 40, 63, 64, 130–31, 209n1 separatist recurrences in, 97; US control of state
Hoeffler, Anke, 48n2, 167 design for, 255; US decision making on interim
Holsti, Kalevi, 43–44, 49n15 government for, 250
horizontal legitimacy, 43–44, 222–23, 225, 227, Islam: in Lebanon, 163–64, 188; in Sudan, 103, 104
244, 245, 268 Islamic Legion (Libya), 170
Horowitz, Donald L., 40, 157 Israel: and Lebanon, 164, 167; and legitimacy of
Human Rights Watch, 84 state, 44
Human Security Report Project, 21n13, 67n12 Itano, Nicole, 81, 94n13
Human Security Research Group, 234 Iyob, Ruth, 106
Humphreys, Macartan, 28
Hussein (King of Jordan), 209n10 Jackson, Robert H., 46
Hutus. See Burundi; Rwanda Jarstad, Anna, 40
Johnson, Douglas H., 105
IFOR (NATO) troops, 204 Johnson, Roosevelt, 73, 76, 78–79
inclusion, 39–41, 195–96; and avoiding recurrence, Johnson, Yormie, 73
38, 199t, 202, 206, 214, 216; and consolidation Johnson-Sirleaf, Ellen, 77, 79, 89, 265
of peace, 4, 207, 230; defined, 39, 186; external Joint Control Commission (South Ossetia), 114
actors’ influence on, 236–37; legitimation Joint Peacekeeping Force (South Ossetia), 111,
depending on, 274; monitoring of inclusionary 113–14
agreements, 245; policy implications of, 274; Jones, Bruce, 45
reneging on, 148–50; war criminals, problematic Jordan: nonrecurrence in, 19; Palestinian refugees
inclusion in political order, 228, 275n7. See also in, 209n10; totalitarian government in, 199
exclusion; powersharing Jumblatt, Kamal, 165
“inclusive-enough” coalitions, 222 Jungle Fire (Liberia), 81
Independent NPFL, 73
Indonesia: and Aceh independence struggle, 10; Kabbah, Ahmed, 76, 83
East Timor involvement of, 137, 140, 145, 150; Kabila, Laurent-Désiré, 126
separatist recurrences in, 97 Kagame, Paul, 156
informal states, 237 Kaladi (Timorese people), 136–37
institutional weakness / strength. See state building Kant, Immanuel, 36
Inter-African Mission to Monitor the Bangui Agree- Karzai, Hamid, 250, 265–66
ment, 126 Keefer, Philip, 243–44
Intergovernmental Authority on Development, 272 Keita, Kalifa, 168
Interim Cooperation Framework, 269 Khadiagala, Gilbert, 106
interim governance. See governance Khasavyurt Accords (1996, Chechnya), 107–8, 119
internal armed conflicts, defined, 11 al-Khattab, Ibn, 108
Index 309

Koehler, Jan, 106–7 Lemarchand, René, 151, 153–54, 156, 157, 158,
Kokoity, Eduard, 113 161n18, 189
Kolingba, André, 125–26, 149 liberal peacebuilding, 19, 31–33, 44, 47, 218–19,
Konaré, Alpha Oumar, 169–70 224, 225t
König, Marietta, 111 Liberia, 71–95; abrogation of powersharing
Kosovo: international factors and separatist struggle arrangements in, 99, 194, 215; constitutional
in, 46; international transitional authority order, preservation of, 252; demobilization and
in, 247; local and legislative elections before reintegration process in, 84–86, 89; election
presidential elections in, 261–63; national candidates in, 263–66; exclusionary behavior
ground forces in, 249; NATO bombing of, 13, of Taylor regime, 72, 78–81, 98, 100, 123–24,
110; NATO in decision-making role for, 248; no 202; first civil war, 72–74; “greed” and natural
peace agreement in, 2; withdrawal of peacekeep- resources in, 73–74, 86–88, 214, 215; informal
ing troops from, 33–34 external influence on, 248; international factors
Kosovo Liberation Army, 13, 249 in, 82–84; marginalization of civil society in,
Krahns (Liberia), 73, 77, 79, 80, 92 253; media suppression by Taylor, 79; onset of
Krain, Matthew, 149 peace (1996), 1, 12, 74–75; peacekeeping con-
Krasner, Stephen D., 222 sidered successful, yet conflict recurred in, 2, 9,
Krings, Thomas, 169 204; political powersharing in, 188, 256; political
Kromah, Alhaji G. V., 73, 76, 78–79, 85 repression under Taylor, 78–79; regional factors
Krueger, Robert, 190 in, 81–82, 150; second civil war, 1, 16, 71, 76–78,
Kumar, Krishna, 241 188; second postwar peace process (2003–pres-
ent), 88–91; security forces, exclusion in, 79–81;
Lafontant, Roger, 131 UN peacekeeping in, 3, 16, 204
Laitin. See Fearon and Laitin data sets Liberia Coalition for Reconciliation and Democ-
Laos: negotiated settlement in, 184; nonrecurrence racy, 76, 88
in, 194; regime change without warfare in, 189; Liberian Forest Development Company, 86
regional assistance in civil war, 200 Liberian Peace Council, 74, 76, 86
Latortue, Gerard, 134 Liberians United for Reconciliation and Democracy
Lavalas Family (Haiti), 133–34 (LURD), 71, 76–77, 80, 86, 88, 92, 94n3
Lebanese Front (LF), 165 Liberia Peace Council, 73–74, 76, 94n3
Lebanese National Movement (LNM), 164–65 Libya: civilian participants in uprising (2011),
Lebanon, 1, 163–67; analysis as anomaly to 234–35; and Mali, 168, 170, 173
study’s generalizations, 162, 166–67; economic Licklider, Roy, 123, 149, 180
powersharing in, 192; failed powersharing in, Lieven, Anatol, 107
162, 164, 166, 167, 172–73, 180, 188, 202, 231; Lijphart, Arendt, 40
institutional weakness in, 173; international Lobato, Rogelio, 141
troops to maintain stability in, 166, 233; multiple Locke, John, 36
recurrences in, 21n19; National Pact (1943) and Lockhart, Clare, 7, 35, 222
civil war (1958), 163–64; recurrence (1975), 18, Lofa Defense Force (Liberia), 73, 76
164–66; religious factors and marginalization in, Lord’s Resistance Army (Uganda), 128
163–64; Ta’if Accord (1989), 166, 167 Loya Jirga, 249
Lebed, Alexsandr, 107, 109 LURD. See Liberians United for Reconciliation and
legitimacy building, 36–37, 237–45; and democ- Democracy
ratization, 240, 242; external actors enhancing, Lynch, Dov, 203
237–38, 270–73; importance of, 222–23; Lyons, Terrence, 75, 256
increased resistance to Western manipulation,
238–39; “layered,” 270–73; and negative ef- Machiavelli, Niccolò, 33, 220, 236
fects of “democracy promotion,” 240–44; and Mack, Andrew, 234
sovereignty norm, 239–40. See also horizontal Madison, James, 33, 220–21
legitimacy; state legitimacy al-Mahdi, Sadiq, 104
legitimacy-focused peacebuilding, 6–7, 19, 213, Makdisi, Samir, 167
224–30; and blanket formulas, 245; framework Making War and Building Peace (Doyle & Samba-
for, 228–29; and policy questions, 229–30 nis), 32
310 Index

Mali, 167–73; analysis of, 162; background of, 168; National Council of Timorese Resistance (CNRT),
droughts in, 169; economic factors, 214; ethnic 138–40, 142, 262
factors in, 163, 168, 181n1; exclusionary behavior National Democratic Union (Nicaragua), 175
in, 168; failed powersharing in, 162, 172–73, 180; National Liberal Party (Lebanon), 165
inclusionary behavior, yet recurrence in, 172–73, National Patriotic Front of Liberia (NPFL), 73–77,
201; initial civil war, 168–70; institutional weak- 80, 85–86, 94nn5–6
ness in, 99, 173, 214; peace agreements and pow- National Police of Timor-Leste (PNTL), 141
ersharing in, 170–71, 180; recurrence (2006–08), National Reconciliation (Sudan), 104
18, 171–72; reintegration of ex-combatants in, 172 National Reconciliation Commission (Liberia), 78
Malu, Victor, 83 National Resistance Army (NRA ; Uganda), 156
Mamdani, Mahmoud, 151, 156 NATO, 13, 34, 110, 248–49, 273, 275n5
Mandingos (Liberia), 73, 74, 77, 92 NATO-commanded Implementation Force
Mansfield, Edward, 230, 241 (IFOR), 204
Mao Zedong, 18, 116–17, 119 natural resources: high-rent exports of, 179; leverage
Maskhadov, Aslan, 107, 108–9 of resource-rich states, 239; precipitating warfare,
Massaquoi, François, 76 26–27; sharing as means of conflict resolution,
mass violence: classification as civil war, 158–59; 192; significance as factor, 180, 215. See also
failed powersharing resulting in, 231; genocide specific countries
vs., 151 Ndadaye, Melchior, 190
McLeod, Patrick, 64 N’dayen, Didace, 125
mediation. See negotiated settlements; peace negotiated settlements, 184–86, 185t, 200, 207, 253
agreements Nepal: economic powersharing, 192; ongoing nego-
Metayer, Amiot, 133 tiations in, 188; UN military observers, 200
methodology, 6, 15–16; limitations of, 230, 232–33 NGOs (nongovernmental organizations). See civil
Miloševic, Slobodan, 228 society organizations
Min, Brian, 58–59 Nicaragua, 173–77; Communist revolution in,
Minorities at Risk data set, 37 174; constitutional order, preservation of, 252;
Mitchell, Christopher R., 104, 105 economic factors in, 177, 214; ethnic factors in,
MODEL (Liberia), 71, 77, 92 176; exclusionary behavior in, 123, 162, 173, 176;
Mosques and Imams National Advisory Board international intervention in, 200, 233; invasive
(Central African Republic), 205 US role in, 175–76; negotiated settlement in,
Mouvement Populaire de Libération de l’Azaouad 184; powersharing pact in, 186; recurrence in,
(MPLA), 169–70 1, 12, 18, 174–75; religious and ethnic factors
Mouvement pour la Libération du Congo (MLC; in, 174
Central African Republic), 126 Nicaraguan Democratic Force, 175
Mouvements et Fronts Unifiés de l’Azaouad Niger, role in Mali unrest, 168, 169–70, 171, 173, 201
(MFUA; Mali), 169–71 Nigeria, relationship with Taylor in Liberia, 75
Movement for Democracy in Liberia (MODEL), Nimeiri, Gaafar Muhammad, 103–5, 119
71, 77, 92 Nkomo, Joshua, 147–48
Mozambique: end of war in, 184; interim gover- nonrecurrence, 19, 183–210; and economic
nance, role of domestic actors in, 252; interna- powersharing, 192; and exclusionary behavior,
tional troop deployments in, 200; nonrecurrence 199, 199t; and inclusionary behavior, 186–92,
in, 19; political powersharing in, 188; security 199t; paths to, 207; and political powersharing,
powersharing in, 190; sovereignty maintained 187–90; pool of cases, 14t; and security power-
in, 247 sharing, 190–91; and territorial powersharing, 5,
Mugabe, Robert, 147–48 191. See also inclusion; peacebuilding
multiparty democracy, 241 nonseparatist recurrences, 18, 122–61; Burundi,
150–58; Central African Republic, 124–29;
Nagorno-Karabakh: elections sparking mass vio- East Timor, 136–46; exclusionary behavior pre-
lence in, 241; ethnicity and recurrence in, 48n7; cipitating, 123–24; Haiti, 129–36; reneging on
exclusion from study, 120n1 inclusion, 148–50; Rwanda, 150–58; Zimbabwe,
National Council of Maubere Resistance (Timor), 146–48
138 norms, creation after termination of wars, 217–18
Index 311

Norris, Pippa, 40 207; peacekeeping distinguished from, 20n9;

Northern Alliance, 249 pool of cases, 183–85, 185t; in postconflict soci-
Northern Ireland: nonrecurrence in, 19; political eties, 25; republican peacebuilding, 33, 220–21,
powersharing in, 188; security powersharing in, 224, 225t; rethinking aims and approaches of,
191; sovereignty maintained, 247 218–30; Sambanis on, 64; as state building,
NPFL. See National Patriotic Front of Liberia 33–35, 221–23, 225t. See also consolidation of
peace; liberal peacebuilding; powersharing
O’Ballance, Edgar, 106 peacekeeping operations, 2, 3, 6, 9, 127, 132,
Obama, Barack, 272 149–50, 200, 208. See also UN peacekeeping
Obote, Milton, 156 operations
Office of the High Representative (Bosnia), 252 peacemaking. See peacebuilding
Ohlson, Thomas, 148 Peace Research Institute Oslo (PRIO), 9, 13, 94n4,
oil exporters, 67n9, 99, 105, 109, 127 110, 120n1, 124, 171, 178; Uppsala Conflict Data
one-sided violence, 234 Program data set, 233
onset of civil war: and ethnicity, 26, 28; grievance in, People’s Liberation Army (PLA; China), 115–17
26–27; reasons for, 27–30, 29t; recurrence vs., People’s Rally for Progress party (Djibouti), 188
17–18, 50–67 perceptions about peace between periods of war,
Operation Camp Johnson Road (Liberia), 78 12–13, 15, 97
Organization for Security and Cooperation in performance legitimacy, 45
Europe (OSCE), 111–12, 114, 262, 276n14 Peru, 177–79; civil war (1980–2000), 177; conflicts
Organization of African Unity, 127 with few deaths in, 179; ethnic factors in, 179;
Organization of American States (OAS), 132–33, exclusionary behavior in, 162; natural resources
136, 200, 256, 271, 273 in, 179, 180; peace via strategic victory in, 178;
Organization of the Islamic Conference, 271 recurrence (2007–present), 18, 99, 178–79
organized criminal violence, 20n7, 27 Philippe, Guy, 133
Ottaway, Marina, 241 plausibility probe, 16, 97
Poggo, Scopas S., 102
Palestine Liberation Organization (PLO), 165, 167 Police Service of Northern Ireland, 191
Palestinians, 164, 165, 167, 173, 209n10 political exclusion. See exclusion
Paris, Roland: and delayed elections, 260; on liberal political inclusion. See inclusion; powersharing
peacebuilding, 31–32, 218–19; on partition, political powersharing, 187–90
235n1; on state building as condition precedent, Popular Movement for the Liberation of the
34, 36 Azaouad, 169
Paris Declaration on Aid Effectiveness, 222, 227 “postsettlement settlements,” 268
partition and secession: as alternative, 235n1; and postwar norms, 217–18
ethnic group identity, 29. See also territorial poverty. See economic factors
powersharing poverty reduction, questioning focus on, 216
Pastora, Eden, 174–75 Poverty Reduction Strategy Paper (World Bank),
Patassé, Ange-Félix, 124–26, 128–29, 150, 204 269
Patriots for Renewal / Ex-Liberators (Central Afri- powersharing: abrogation of arrangements, 99, 193–
can Republic), 128 94, 193t, 215, 216; brokered deals for, 226–27;
Patten, Alan, 220–21 consultation and voice in, 268; criticisms of, 41,
peace, defined, 183, 215 49n13, 122, 231–32; defined, 39–40, 48n12;
peace agreements: during Cold War vs. post–Cold economic powersharing, 192; failed power-
War era, 5, 184–86; and deepening divisions sharing, 162, 164, 166, 167, 172–73, 180, 208;
among former allies, 149; and duration of peace, multifaceted approach to on multi-year schedule,
64–65; expectations shaped by, 5; focus on war- 231; need to test for effect on failure or success,
ring parties, downside of, 253; no peace agree- 231–32; not essential for nonrecurrence, 196;
ment and one side as victor, 2. See also negotiated ongoing negotiations, effect of, 188; political
settlements powersharing, 187–90; pool of cases, 186, 187t,
peacebuilding, 17, 30–36; critical theories of, 192–95, 193t; power-dividing arrangements,
35–36, 44, 223–24, 225t; defined, 8, 215; 5–6, 28, 186, 189–90, 208; prevalence of, 207,
legitimacy-focused, 6, 19, 213, 224–30; paths to, 231; and prevention of recurrence, 28, 64, 194,
312 Index

powersharing (continued) reintegration programs. See disarmament, demobili-

215; security powersharing, 4, 188, 190–91, 215; zation, and reintegration (DDR)
sunset provisions on, 256; territorial powershar- religious fractionalization, 54–56, 216; social
ing, 5, 191, 205; types of, 40. See also specific contract with diverse religious groups as basis
countries for sustained long-term peace, 245–46. See also
pragmatic legitimacy, 42 specific countries
Preval, René, 133–34 RENAMO rebels (Mozambique), 190
Prevost, Gary, 175 Reno, William, 87
PRIO. See Peace Research Institute Oslo repression, 214, 219–20
process legitimacy, 45, 226 republican peacebuilding, 33, 44, 220–21, 224, 225t
Provisional Electoral Commission (Haiti), 133 research design, 7–17; more nuanced concepts
Provisional Government of East Timor, 137 needed in, 235; standards for “failed peace,” 8–9.
proxy wars, 233–34 See also case-study approach; methodology
Prunier, Gerard, 105, 151, 156, 158 resistance to Western manipulation, 238–39
Putin, Vladimir, 107, 109, 119 resource curse. See natural resources
Revolutionary Armed Forces of Colombia (FARC),
Qaddafi, Muammar, 170 179–80
“quality of life” indicators as predictors of recur- Revolutionary United Front (RUF; Sierra Leone),
rence, 27–28 76–77, 82–83, 87, 192
quantitative methods, 59–65; arbitrary or question- Reynolds, Andrew, 256
able estimation, 60–61; limits of, 216, 233; Richmond, Oliver P., 35, 36, 218, 221, 223
omitted variable bias, 62; policy applicability, Roeder, Philip G., 28, 41, 48n12, 49n13, 157
62–65, 231; proxies, use of, 61–62; on recur- Roh Moo-hyun, 197
rences, 27–30; too few observations, 60; unit Rosberg, Carl G., 46
heterogeneity, 60 Rose Revolution (South Ossetia), 113
quasi-states, 237 Rost, Nicolas, 64
Quiwonkpa, Thomas, 94n5 Rothchild, Donald S., 28, 41, 48n12, 49n13
Royal Ulster Constabulary, 191
Ramos-Horta, José, 141, 144, 263 Ruggeri, Andrea, 61
Reagan, Ronald, 175–76 Russia: and “frozen” conflicts, 203; role in cease-
rebel cohesion, 173, 180 fires, 184. See also Chechyna
recurrences: criteria for study, 13; duration of prior Rwagasore (King of Tutsis), 153
war as factor, 28; and ethnicity, 28–29, 48n7, Rwanda, 150–58; alternative explanations for recur-
54–59, 55–56t, 66n5; list of core cases, 15, 15t; rence in, 150, 158–60, 215; background of, 152;
not exclusionary behavior, 162–82; not fitting democratization in, 157, 159, 230, 241; economic
the model of exclusionary causes, 162–63; onset factors in, 158; ethnic violence in, 152–53, 189;
of civil war vs., 17–18, 50–67; pool of cases, 2, exclusion precipitating recurrence in, 123, 153;
13–15, 14–15t, 51; “quality of life” indicators genocide in, 150, 151; multiple recurrences in,
as predictors of, 27–28; quantitative research 21n19; peace agreement and failure to secure
on, 27–30, 29t; refining analysis of, 54–56; peace in, 34, 152, 156; recurrence (1990) in, 18,
regression analysis of, 18, 51–59, 232; territorial 155–56; regional factors in recurrence in, 159;
autonomy, reneging on, 118–20; variables and severity of exclusionary behavior in, 99; violence
regressions, 52–54. See also nonseparatist recur- (1959) and war (early 1960s) in, 152–53
rences; separatist recurrences; specific countries Rwandan Patriotic Front (RPF), 152, 155, 156
Regan, Jane, 142 Rwantabagu, Hermenegilde, 158
regime changes, 247–54
regime legitimacy, 43. See also state legitimacy Saakashvili, Mikheil, 112–13, 119, 121n7
regional factors, 46, 81–82, 127–28, 159, 214. See Sadaka, Richard, 167
also specific countries Salehyan, Idean, 233
regional organizations, role in legitimacy building, Sambanis, Nicholas: on capacity, 7, 20n5; civil wars
271–72 data set of, 9, 51; on international aid, 63–64;
Reilly, Ben, 256 on liberal peacebuilding, 32; on mass atrocities
Reinado, Alfredo, 141 in Burundi, 151; on peacekeeping troops’ role in
Index 313

maintaining stability, 205; proxy use by, 62; on Somoza, Anastasio, 174, 177
rebel cohesion, 173, 181n3; shortcomings of re- South Africa: constitutional order, preservation of,
search of, 63–64; on South Ossetia, 110, 121n5; 252; economic powersharing in, 192; end of war
on sustainability of peacebuilding, 64; on UN in, 184; nonrecurrence in, 19; organized criminal
peace efforts’ effectiveness, 30, 90; on variables violence in, 20n7; political powersharing in, 188,
leading to recurrence, 28–30 256; sovereignty of, 247
Sandinistas. See Nicaragua Southern African Development Community, 273
Sankoh, Foday, 87, 192 Southern Lebanon Front, 165
Savimbi, Jonas, 85, 184, 191, 241, 259 Southern Party (Sudan), 101
Schevardnadze, Eduard, 110 South Korea: nonrecurrent civil war in, 197;
Schmitt, John, 15, 50, 66n6 regional assistance in civil war in, 200; US role
“security dilemma” approach to internal armed in, 200
conflict, 26 South Korean Workers Party, 197
security powersharing, 4, 188, 190–91, 215 South Ossetia. See Georgia and South Ossetia
Sendero Luminoso. See Shining Path (Peru) South Ossetian Popular Front, 111
Senegal: exclusionary behavior in, 214; interwar South Ossetian Supreme Soviet, 111
peace in, 12, 97 South Sudan Liberation Movement, 103
separatist recurrences, 96–121; Chechyna, 106–10; sovereignty norm, 239–40
China and Tibet, 115–18; exclusionary behavior Special Security Service (Liberia), 78, 80–81
precipitating, 98; Georgia and South Ossetia, Sriram, Chandra Lekha, 41, 49n13
110–14; and political exclusion, 4; pool of cases, stability of postwar societies: and deepening divi-
97–98, 98t; Sudan, 100–106 sions among former allies, 149; and inclusion of
Seventeen-Point Agreement (Tibet, 1951), 116, 118 state’s main social factions, 225; and third-party
“shadow” states, 237 military troops, 6
Shevardnadze, Eduard, 111, 121n7 state building, 6–7; to consolidate peace, 3; insti-
Shining Path (Peru), 99, 173, 177–78, 180, 182n4 tutional weakness, 58, 99–100, 109, 120, 149,
Shoesmith, Dennis, 140 159, 214–15; as more conservative than liberal,
Sierra Leone: conflict resolution and sharing of 36; peacebuilding as, 33–35, 221–23, 225t;
natural resources in, 192, 194; constitutional shortcomings of, 224
order, preservation of, 252; informal external State Department. See US State Department
influence on, 248; international troop deploy- state design, 254–56
ments to, 200; and Liberian unrest, 71, 76, 77, 81, state legitimacy, 41–47, 226, 243–44; after elec-
82, 83, 150; Lomé powersharing accord (1999), tions, 266–73; bases of, 243; defined, 41–43;
192, 194; political powersharing in, 188; resource and democratization, 243; embedded, 45, 226;
curse in, 27, 87; state legitimacy after elections and globalization, 244–45; international sources
in, 267; withdrawal of peacekeeping troops from, of, 45–47; performance legitimacy, 45; prag-
33–34 matic, 42; process legitimacy, 45, 226; and social
Sierra Leone Expert Panel, 94n18 expectations, 243–44; vertical, 43–45, 49n15,
Simonsen, Sven Gunnar, 142 245; Western efforts to build, 239–40. See also
Singer, David J., 9, 21n11, 51 horizontal legitimacy; legitimacy building
al-Sistani, Ali, 261 Stedman, Stephen John, 148
Small, Melvin, 9, 21n11, 51 Stepashin, Sergei, 109
Smith, Ian, 146 “structured, focused” case comparison, 180–81
smuggling, 172 Suchman, Mark, 41–42
Snyder, Jack L., 28, 230, 241 Sudan, 100–106; abrogation of powersharing ar-
Sochi Agreement (1992, South Ossetia), 111–12 rangements in, 99; background of, 101; exclusion
social exclusion, 37, 48n8 precipitating recurrence in, 98, 105, 123; failed
social expectations and state legitimacy, 243–44, powersharing in, 194; first civil war (1955–72),
254–55 100, 101–3; institutional weakness in, 99, 214;
socialist regimes, 196 international factors in recurrence in, 106, 233;
Somalia: interwar peace, perception of, 12, 15, 97; natural resources in, 105–6, 119, 214; peace
multiple armed conflicts in, 201; peacekeeping in, 103–4; powersharing after 2005 involving
troops withdrawn amid failure, 9 elections after 4 years and South Sudanese inde-
314 Index

Sudan (continued) ULIMO-J, 73–74, 76, 94nn2–3

pendence after 6-year period, 231; relationship ULIMO-K, 71, 73–74, 76, 86, 94nn2–3
with Central African Republic, 127–28; second UN Development Program, 7, 171
civil war (1983–2005), 18, 104–6; separatist UN Human Rights Commission, 238
recurrences in, 97; territorial autonomy, reneging UN Human Rights Council, 239
on, 119; world unwillingness to enter into Darfur, Unified Movements and Fronts of the Azaouad
239 (Mali), 169
Sudan African National Union (SANU), 102–3 UN Integrated Mission in East Timor, 144
Sudan People’s Liberation Army, 105 UN Interim Force in Lebanon, 166
Sudan People’s Liberation Movement, 105 Union of Democratic Forces for Unity (UFDR;
Suez Crisis, 164 Central African Republic), 128–29
Suharto, 138 UNITA (Angola), 191, 241, 259
Suliman, Mohamed, 105 United Arab Republic, 163, 164
Syria: involvement in Lebanon, 165–66, 167; post- United Kingdom: sanctions on Liberia, 83–84;
war authoritarianism in, 219 Zimbabwe’s war for independence against Brit-
ain, 146. See also Northern Ireland
Ta’if Accord (1989, Lebanon), 166, 167 United Liberation Movement of Liberia for Democ-
Tajikistan: external support for legitimacy in, 245; racy (ULIMO), 73, 92
powersharing in, 188; UN military observers United Nations’ role: in legitimacy building, 271; as
in, 200 military observers, 200; negative reaction to role
Taliban. See Afghanistan in shaping political institutions, 255; in peaceful
Tate, Joe, 81 settlements, 185; support of election candidates
Taylor, Charles, 71–93, 94n12, 188, 202, 228, 264. prohibited, 264. See also UN peacekeeping
See also Liberia operations
territorial autonomy, reneging on, 118–20 United States: Dalai Lama support in Tibet, 118;
territorial powersharing, 5, 191, 205 dominance in NATO and OAS, 273; putting
third-party military troops, 6 down foreign rebellions, 200; resistance to
Thomas, Lowell, 118 manipulation of, 238–39; role in Dominican Re-
Tibet, 115–18; abrogation of powersharing arrange- public, 195–96; role in Haiti, 130–35, 150, 250;
ments in, 194; first Chinese war over (1950–51), role in Iraq interim government, 250; role with
115–16; recurrence precipitated by exclusion in, Contras (Nicaragua), 174, 175–77; sanctions on
98, 118; second Chinese war over (1956–60), 18, Liberia, 83–84
97, 115; territorial autonomy, reneging on, 119, United Tajik Opposition (UTO; Tajikistan), 188
120, 191 UN Mission in Liberia (UNMIL), 89–90, 268
Timorese National Council of Resistance (CNRT), UN Mission in the Central African Republic, 126,
251 205
Tishkov, Valery, 111 UN Observer Mission in Liberia (UNOMIL), 84,
Tonton Macoutes (Haiti), 130 85, 89
Touré, Amadou Toumani, 126, 169–71 UN Panel of Experts on Liberia, 94n18
transitional arrangements and rulers, 247–54, UN Peacebuilding Commission, 3, 8, 215
274. See also decision “moments” in postwar UN Peacebuilding Support Office (Liberia), 84, 89
political processes; “democracy promotion” and UN peacekeeping operations, 3, 32, 96, 112, 132. See
democratization also specific missions
Traoré, Moussa, 169–70 UN Security Council: East Timor actions, 139;
Trujillo, Rafael, 195 monitoring post-national elections, 268; reject-
Tuaregs. See Mali ing incursions on sovereignty, 239; resolution
Tupac Amaru Revolutionary Movement (Peru), 177 on Haiti, 132; sanctions on Liberia, 83–84, 87,
Tutsis. See Burundi; Rwanda 90, 95n24
UN Special Commission of Inquiry for Timor-
UDT (Timorese Democratic Union), 137 Leste, 142, 143
Uganda, role in Rwanda wars, 156, 159 UN Stabilization Mission in Haiti, 134, 136
ULIMO. See United Liberation Movement of Libe- UN Support Mission in East Timor (UNMISET),
ria for Democracy 140, 144, 160n8
Index 315

UN Transitional Administration in East Timor women, exclusion of, 48n11

(UNTAET), 139–40, 144–45, 160n8, 262 World Bank, 7, 32, 63, 105, 133, 262, 269, 273
Uppsala Conflict Data Program, 9–10, 21n13, 233 World Development Report (World Bank), 2, 222
US State Department, 78–80, 179, 250
Utas, Mats, 85–86 Yeltsin, Boris, 107–9, 111
Uvin, Peter, 151, 153, 154, 157, 159, 189 Yemen: failed powersharing in, 194; multiple armed
conflicts in, 201
vertical legitimacy, 43, 49n15, 245 Yugoslavia, former: difficulty coding wars of, 13;
victory as method of war termination, 2, 178, elections sparking mass violence in, 241
196–97, 200, 207, 216, 253–54
Vietnam: nonrecurrence in, 19; postwar authoritari- Zahar, Marie-Jöelle, 166
anism in, 219; regional assistance in civil war in, Zartman, I. William, 222
200; totalitarian regime in, 196 Zhou Enlai, 115
Vinding, Diana, 170 Zimbabwe, 146–48; abrogation of powersharing ar-
rangements in, 99, 194, 215; divisions among for-
Wakoson, Elias Nyamlell, 105 mer allies leading to recurrence in, 149, 159–60;
Wallensteen, Peter, 267 exclusionary behavior in, 123, 148, 149; initial
Walter, Barbara F., 21n18, 27–28, 29, 51, 62, 85, war for independence (1972–79), 146–48;
121n5, 205 institutional weakness in, 99, 149, 214; postwar
war criminals, problematic integration into political government reneging on commitments in, 150;
order, 228, 275n7 rebel cohesion in, 180; recurrence precipitated by
weak state approach. See state building exclusion in, 98
Weinstein, Jeremy M., 28 Zimbabwe African National Union (ZANU & its
Weinstein, Warren, 152, 157 armed wing ZANLA), 146–48
Welt, Cory, 112–13 Zimbabwe African People’s Union (ZAPU & its
Wimmer, Andreas, 58–59 armed wing ZIPRA), 146–48
Wion, Madison, 78, 94n11 Zurcher, Christoph, 106–7