University of Southern California Dissertations and Theses

Shaking the foundations of violent civil conflict: institutions, disasters, and the political economies of state-rebel interaction

(USC Thesis Other)

Shaking the foundations of violent civil conflict: institutions, disasters, and the political economies of state-rebel interaction

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content SHAKING THE FOUNDATIONS OF VIOLENT CIVIL CONFLICT:
INSTITUTIONS, DISASTERS, AND THE POLITICAL ECONOMIES OF
STATE-REBEL INTERACTION

by

Jason S. Enia

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(POLITICS AND INTERNATIONAL RELATIONS)

August 2009

Copyright 2009 Jason S. Enia

ii
Dedication
To Monica. For so many reasons.

iii
Acknowledgements
A number of people are deserving of great thanks for their continuing
encouragement and support. My words here are unlikely to be sufficient. Nonetheless…
I have been blessed with good teachers. From the earliest point in my education
right through graduate school, my teachers pushed me to think. They pushed me to think in
ways that not only made me a better scholar but also in ways that I believe made me a better
person. I will always be grateful for their enthusiasm, energy, sacrifice, and tough love.
The sum of one’s graduate school experience is as much (if not more) dependent on
what goes on outside the classroom as within. In this respect, I’ve been lucky enough to be
part of a great cohort of students at USC, and I am thankful to each and every one of them
for being good conversationalists, good listeners, and (most of all) good tailgaters. I wish
them all the best in the future.
My dissertation committee deserves special mention. Peter Gordon performed
above and beyond the call of duty as my outside member. John Odell has always been the
best combination of cheerleader and critic. From an early point, he asked the tough
questions in an attempt to push me to make this project as good as possible. Saori Katada
has been a mentor from my first day on campus in 2003. Much of what I’ve learned about
what it takes to be a successful academic occurred while I was Saori’s teaching and then
research assistant. Finally, when Pat James arrived on campus in 2005, he brought with him
a legendary (and almost unbelievable!) reputation as a tremendous mentor to graduate
students. It’s all true. Selfless, disciplined, good-spirited, and on-and-on it goes. He truly is
the perfect role model. My deepest thanks go out to all of you.

iv
Nothing I have accomplished would have been possible without the support of my
family. My parents were my earliest teachers, and whether they know it or not, it’s a role
that they continue to play very well. They have been a constant source of encouragement
throughout my life. My brother continues to keep me grounded in ways that only a brother
can. Words cannot fully express my thanks for these gifts.
And finally my greatest thanks and deepest love go to Monica. Always there, you
told me, “You can do it.” You were right.

v

Table of Contents
Dedication ii
Acknowledgements iii
List of Tables vi
List of Figures vii
Abstract viii
Chapter 1: Introduction 1
A Theory of Institution Capture 6
Potential Implications of This Research 11
Outline of the Dissertation 17
Chapter 2: Developing the Theory and Hypotheses 20
Predicting Civil Violence: From Discrete to
Overlapping Explanations 21
Deriving the Preferences of Institution
Capture in a Post-Disaster Context 39
Chapter 3: Natural Disasters, Data, Methods and Cases 47
Using Natural Disasters as an Analytical Tool 48
Methodology and Case Selection 58
Conclusion 60
Chapter 4: Testing the Relationship between Institutional
Quality and Post-Disaster Violence 62
Empirical Design 63
Results and Analysis 76
Conclusion 95
Chapter 5: Exploring the Cases of Indonesia & Sri Lanka 96
Indonesia 98
Sri Lanka 116
Analysis and Implications for Theory 131
Conclusion 137
Chapter 6: Conclusion 139
The Project and Its Findings 139
The Significance of the Findings and Future Work 143
Bibliography 153

vi
List of Tables
Table 1: Disconfirming Cases and the Collier-Hoeffler Model 30
Table 2: Global Natural Disasters by Type, 1984-1999 68
Table 3: Logistic Regression of Violent Civil Conflict, 1984-1999 79
Table 4: Logistic Regression of Violent Civil Conflict, 1984-1999 81
Table 5: Freedom House Scores for Indonesia and Sri Lanka, 2002-2003 90

vii
List of Figures
Figure 1: Geographic Location of Earthquakes
Magnitude 5.5 or Greater (1975-2002) 57
Figure 2: Receiver Operator Characteristic (ROC) Plot 77
Figure 3: Predicted Probabilities with Variables at their Means/Medians 85
Figure 4: Predicted Probabilities with IQ Index Interaction Variable
-1 Standard Deviation from Mean 86
Figure 5: Predicted Probabilities with IQ Index Interaction Variable
+1 Standard Deviation from Mean 88

viii
Abstract
What is the effect of natural disasters on political violence? Despite the widespread
speculation about the possibilities for peace that typically occurs after a natural disaster
strikes in a conflict zone, relatively little is known about the way disasters impact political
mechanisms. In this project, I find that post-disaster strategic dynamics around control of
disaster relief institutions are likely to impact levels of grievance and feasibility and thus have
important implications for the likelihood of post-disaster violence. Using time-series, cross-
sectional data, I find that post-disaster violence is more likely where pre-disaster institutional
quality is relatively weak, as weak institutions create incentives for institution capture. A
qualitative comparison of Indonesia and Sri Lanka following the 2004 tsunami suggests that
post-disaster violence also depends on patterns of institutional control at the time of the
disaster. The disaster results in a “lock-in” effect, accelerating and magnifying the material
and non-material benefits of institutional control. These changes alter levels of grievance
and the opportunities for rebellion.
This project makes two theoretical contributions. First, in contributing to the
research on “disaster diplomacy,” it suggests that political variables are affected by natural
disasters but not necessarily in transformative ways. Instead, natural disasters help lock-in
and even accelerate political processes already in place in the time before the disaster.
Second, viewing natural disasters as “crises” or “focusing events,” suggests that battles over
institutional control can define and re-define the micro-foundations, and often the dynamic
evolution, of intrastate conflict as much as they can define the endpoints of conflict. The
material and signaling benefits of institutional control change the incentive structures of
both rebels and states around rebellion.

1
Chapter 1: Introduction

On 26 December 2004 at 07:58 local time, a massive earthquake occurred in the
Indian Ocean off the northwestern coast of the Indonesian island of Sumatra. The quake
measured 9.1 on the moment magnitude scale
1
and struck at a depth of about 30 km (Bilham
2005; Charles et al. 2005; Lay et al. 2005). The energy released by the underwater
earthquake
2
, and particularly the vertical rise of the seabed by several meters, displaced an
enormous volume of water generating a massive tsunami.
In less than 30 minutes, the tsunami’s waves reached the shallow waters off the
western shores of Sumatra and Java, crashing on shore in successive retreat and rise cycles
that moved miles inland and destroyed virtually everything in their path. Approximately one
hour after the eastbound waves reached Indonesia, waves traveling in a westward direction
reach the eastern seaboard of the island nation of Sri Lanka. It is estimated that between
these two countries, 200,000 people were killed that day—approximately 165,000 in
Indonesia (Badan Rehabilitasi dan Rekonstruksi (BRR) and International Partners 2005) and
35,000 in Sri Lanka (DeVotta 2005). One million others were displaced—approximately
500,000 in each country.

1
The moment magnitude scale (MMS)—like its predecessor, the Richter scale—is a
scientific means of measuring the size of earthquakes in terms of the energy released. As
was the case with the Richter scale, the MMS is a logarithmic measurement, where an
increase of 1 point on the scale is indicative of approximately 30 times more power. Unlike
the Richter or other scales with a range of 1-10, the MMS has no upper limit. For more
information, see Hanks and Kanamori (1979).
2
The 2004 Great Sumatra-Andaman earthquake was the second largest earthquake ever
recorded on a seismograph, causing the entire planet to vibrate as much as 1 cm and
releasing the energy of a 100 gigaton bomb (Bilham 2005).

2
Indonesia and Sri Lanka quickly became the primary focal point for international
attention on the disaster. In large part this had to do with the fact that they had the highest
casualty rates. However, an immediate point of focus—and speculation—was how the
disaster would affect the separatist conflicts that had been ongoing in these countries for
more than 30 years.
In Indonesia, the Free Aceh Movement (Gerakan Aceh Merdeka or “GAM”)
3
had
been fighting for autonomous control over the region of Aceh since 1976. Their claims for
independence had evolved over time but were strongly influenced by a mixture of identity
politics, the political economies associated with natural resources in the region, and a series
of human rights abuses committed during a government military offensive against the GAM
in the 1990s. At the time of the tsunami in 2004, much of the military leadership of the
GAM was hiding in the mountains surrounding the provincial capital, Banda Aceh, and
many of the key GAM political leaders had been living in exile in Sweden since the 1970s.
In 2002, a tentative ceasefire was reached between the two sides; however, in 2003, the truce
broke down when both sides complained about how the other was interpreting the terms of
the agreement. In May 2003, the Indonesian government imposed a state of emergency in
Aceh. This resulted in a blockade, preventing entry into Aceh by almost all outsiders,
including diplomats, journalists, and humanitarian groups. During this time, there was very
little diplomatic interaction between the GAM and the Indonesian government. In the two

3
The label GAM is at times used only to refer to the militant wing of the larger Acehnese
separatist movement, using Aceh Merdeka to refer to the broader movement. However, in
most of the literature and the popular press, the label GAM is used to refer to the broader
movement. Thus, this project will follow this more widespread usage. Any reference to a
militant aspect of the movement as separate from other aspects will be made explicit
throughout.

3
years prior to the tsunami, tensions between the GAM and the Indonesian government—but
particularly the Armed Forces of Indonesia (Tentara Nasional Indonesia or “TNI”)—were
running extremely high.
The status of Sri Lanka’s separatist conflict bared some similarities to that in Aceh at
the point in time of the tsunami. The Sinhalese-speaking and largely Buddhist majority Sri
Lankan government had been battling the Tamil-speaking and largely Hindu minority Tamil
Tiger rebels for more than 20 years, resulting in approximately 60,000 deaths. The goal of
the campaign was to achieve an independent, separate state (“Eelam”) for the Tamil
minority on the island.
4
Decades of systematic exclusion fostered wide spread resentment,
which fed into Tamil nationalistic sentiments. Although the Sinhalese majority controlled
the Sri Lankan capital and national government in Colombo, the most prominent Tamil
separatist group—the Liberation Tigers of Tamil Eelam (LTTE), also known as the “Tamil
Tigers”—controlled a 30 square mile patch of land in the northeast corner of the island. In
fighting to extend their autonomous rule to other parts of the northeast where Tamils
resided, the rebels had perfected the tactic of suicide bombing. Throughout the war, they
carried out many successful attacks on Sinhalese politicians and civilians as well as more
moderate Tamils. In 2002, a Norwegian-brokered ceasefire put a tentative end to the
violence. However, in May 2003, talks between the two parties broke down and in the

4
For the purpose of this project, the term Tamil will be used to refer to the Tamil-speaking
population on the island. However, it is important to recognize that within this population
of Tamil-speaking people, there is a wide diversity of religions, geographic dispersions, and
communities that speak Tamil, but only as a second language after English or Sinhala. For a
concise overview of the diversity within the Tamil region of Sri Lanka see Kearney (1985)
and Stokke and Ryntveit (2000: 288).

4
months leading up to the tsunami, the tensions between the two parties seemed to be
steadily increasing.
Shortly after the tsunami hit these two countries, there were immediate declarations
of hope made within the media, the international community, and the leaders of the
respective parties that the disaster might be able to contribute to a peaceful resolution to
these conflicts (e.g., McNeil 2005; New York Times 2005; Thakur 2005). These sentiments
regarding the potential political effects of disasters were nothing new and continue to be
invoked regularly.
5
Many were evoking memories of the politically transformative effects of
past natural disasters with hopes that the current disaster could somehow lead to peace in
conflict-ridden countries.
Despite these sentiments and the anecdotal similarities between these two cases pre-
tsunami, Indonesia and Sri Lanka experienced different levels of post-disaster conflict. In
August 2005, the Indonesian government signed a tentative peace agreement in Helsinki
with the GAM. The agreement, among other things, called for the GAM rebels to surrender
their arms in return for a promise of amnesty from the Indonesian government. Almost
immediately, the agreement was hailed in the press as a victory for “disaster diplomacy.”
(e.g., Aglionby 2005). In Sri Lanka, however, the story was quite the opposite. In the
months after the tsunami, the tensions between the Sinhalese government and the Tamil
Tigers increased dramatically. A series of conflicts around the provision of disaster relief aid
and charges by the Tamil-speaking population of the government’s distribution biases,

5
For a representative but by no means exhaustive sample, see Reynolds (2003) following the
earthquake in Iran in 2003; see Maddox (2005) after the October 2005 earthquake in
Kashmir; see Allen (2008) and Kurlantzick (2008) after Cyclone Nargis struck Myanmar in
2008.

5
contributed to a context in which civil war had reemerged by the end of 2005 (Jagtiani 2005;
Sengupta 2005a).
In addition to the variant empirical outcomes of these two cases, the idea of disaster
diplomacy also generates a number of important theoretical questions about whether and
how the processes associated with a natural disaster might alter the probabilities of post-
disaster violence. What is the role, if any, that disasters play in fostering peaceful diplomatic
interaction between conflicting parties? If there is a role, what is the political mechanism
through which diplomacy emerges? Is the natural disaster merely conduit for more
intensified cooperation on issues unrelated to the conflict, which then generates a kind of
spillover with respect to more peaceful diplomacy? If so, is there something specific about
recovery and relief from a natural disaster, as opposed to other types of cooperation, which
makes it a more likely context in which these spillovers might be produced? Finally, to the
extent that the natural disaster might play some role in engendering diplomacy, how lasting
are its effects? Alternatively, are there any political mechanisms affected by disasters that
could lead to an increase in conflict rather than an increase in peace?
This project offers a first step in exploring these questions by applying the wider
research on civil wars and intrastate violence to this more specific post-disaster context. In
particular, I argue that in the aftermath of natural disasters
6
, political and economic
institutions become important (although, of course, not sufficient) determinants of changes

6
Disasters are defined in EM-DAT as “a situation or event which overwhelms local capacity,
necessitating a request to the national or international level for external assistance, or is
recognized as such by a multilateral agency or by at least two sources, such as national,
regional or international assistance groups and the media” (Guha-Sapir, Hargitt, and Hoyois
2004: 16). More information about the EM-DAT data is provided in chapter 4.

6
in the probability of violence.
7
Specifically, institutional quality at the time of the disaster
matters, as relatively low quality institutions make them susceptible to capture by rebel or
insurgent groups. To the extent that these groups are successful at gaining control of
institutions in the period immediately following the disaster—particularly those institutions
associated with disaster relief processes—they improve the opportunity structures of
rebellion in such a way that makes future violence more likely. The exploration of this
argument is likely to have some important theoretical and empirical implications, not only
for the study of natural disasters but also for the study of intrastate violence more broadly.

A Theory of Institution Capture
As a means of gaining theoretical traction on the questions of post-disaster violence,
I turn to the literature on civil wars and intrastate violence. Within this literature, three
developments are important: the effects of both opportunities and opportunity costs on
various aspects of the feasibility of rebellion, the recognition of the role of political
institutions and institutional quality in contributing to violence, and the recent importance
placed on the strategic interaction between states and rebels. In combination, these three
factors form an important basis for a plausible theory of post-disaster violence.
First, any changes in the opportunity structures associated with rebellion have an
important impact on rebels’ ability to fight effectively. In this respect, changes in sources of

7
Throughout this project, I use the concept institutions in the same way as Doug North
(1990): “Institutions are the rules of the game in a society or, more formally, are the humanly
devised constraints that shape human interaction” (3). By conceiving of institutions as
distinct from organizations, he allows for the study of their interaction (5).

7
financing or access to potential sources of financing can be critical. However, changes in the
ability of a rebel group to procure human resources—members of society willing to fight
against the government or alternate groups—can be equally important. Here, an individual
citizen’s calculations about whether to participate in rebellion and/or violence involve a
variety of factors, including the costs associated with forgoing any other economic
opportunities that may be available as well as any other means of political expression. As
these opportunity costs are lower, rebel groups are likely to have an easier time gaining the
human and monetary resources that are necessary for a rebellion to be feasible.
Second, political institutions play an important role in intrastate violence, both as
sources of constraint and opportunity. Formal institutional features, including the level of
democracy, the inclusiveness of the system, and the characteristics of regimes have been
shown to impact the likelihood of intrastate violence in a variety of ways. In particular,
institutions that impact state capacity—such as the effectiveness of a state bureaucracy,
corruption, the rule of law and order, and resource extraction procedures—are important, as
they affect the ability of a state to repress insurgencies.
Aside from the importance of the presence or absence of these institutions, recent
research also demonstrates that changes in institutional quality have important effects on the
onset of civil war both through the broad avenues of grievance and opportunity. As there
becomes a lack of good alternatives for redressing grievances, this lowers the opportunity
costs associated with rebel recruitment of monetary and human support. In addition, lower
quality institutions have been associated with lower-levels of long-run economic
development, contributing to a lack of economic alternatives, which also serve to lower the
opportunity costs associated with gaining rebel support and resources.

8
Finally, the literature on civil wars and intrastate conflict more broadly has only
recently begun to incorporate strategic concerns, bringing the state back into discussions of
the conditions that contribute to violence. On the one hand, governments are careful to
build the anticipated risks and costs of future challenges into their calculations of how they
deal with insurgencies. On the other hand, the insurgents’ calculations are based on a similar
logic, as they try to balance their strategic options in a way that maximizes their ability to
gain support and resources from the population while at the same time maximizing their
ability to send costly signals to their government opponents.
Together, these three insights provide a plausible argument for the importance of
these factors in the aftermath of natural disaster. Disasters present an interesting
opportunity in this regard as they represent a point in time where institutional quality should
matter most. In addition, the effects of natural disasters tend to be worse in areas that are
relatively under-developed (Pelling 2003). As such, disasters present a moment in which the
institutions of disaster relief may be susceptible to “capture.” The ability to capture
institutional control results in both material and non-material benefit, which impact levels of
grievance and opportunity and therefore have important implications for civil violence.
Specifically, state control over disaster relief institutions generates a number of
material benefits. First, the state can capitalize on the variety of rent-seeking opportunities
and government consumption that often follows an influx in foreign aid. Additionally, the
aid allows the government to spend on disaster relief without raising taxes or initiating some
other source of revenue generation. Third, control over these institutions allows the state to
determine the rules and procedures associated with fund disbursement. Finally, control over
disaster relief institutions—especially monopoly control—has various benefits associated

9
with transaction and enforcement costs. Each of these material benefits that accrue to the
state due to relative institution capture should decrease levels of post-disaster violence, as the
state is able to increase its own capacity to combat rebellion, decrease relative societal
grievances, and positively alter the opportunity structures (resources and opportunity costs)
toward successful rebellion.
States also stand to gain critical and strategic non-material benefits from their relative
abilities to control disaster relief institutions, in whole or in part. To the extent that states
are able to control the disaster relief process, the state is able to send a signal that the
government continues to be the best provider of services for the people. This positive signal
theoretically should result in increased support for the legitimacy of the current government
by both outsiders—countries and non-governmental organizations (NGOs)—as well as the
government’s own citizens. Strategically, this signal also has the effect of weakening the
rebel claims to legitimacy and in so doing, making it more difficult for rebels to recruit
resources in the form of people and money from both within and outside the population.
Rebels possess the same material incentives for capturing disaster relief institutions,
although they are magnified given their asymmetries vis-à-vis the state. If the rebels gain
marginal control over these institutions, the viability of rebellion should increase. First, the
inflow of foreign aid presents a potential source of revenue, increasing the rebels’ capacity to
fund the logistics of rebellion, ceteris paribus. Second, the capacity of the state to fend off
rebellion is negatively affected by this development, as it has a relatively smaller pool of
resources to spend fighting any rebel movements. Third, the fact that the state is not as able
to provide services could lead to higher levels of grievance among the general population.
Finally, given these post-disaster services for citizens, the individual opportunity cost of the

10
choice to rebel should be lower, as the quality of the life forgone by a choice to rebel is
relatively low. In this environment, it becomes easier for rebel leaders to recruit and the
hurdles of collective action are relatively easier to clear given the changes in opportunity
structures.
In addition, rebels also stand to gain critical and strategic non-material benefits from
their relative abilities to capture disaster relief institutions, in whole or in part. For rebels
relative control over these institutions sends related signals to citizens and outsiders about
the inadequacy of the government and legitimacy of the rebel group’s claim to be a credible
alternative. In the battle over state control, this signal is a powerful one, and the rebels’
relative ability to send such a signal would have important consequences for recruiting new
members and gaining outside support and financing. These signals about the functional
legitimacy of the state should also increase post-disaster violence through the avenues of
grievance and opportunity.
Finally, if the battle over institutional control leads to changes in political violence
amidst a disaster, then violence should be (indirectly, at least) affected by the quality of
institutions. States with higher levels of institutional quality at the time of a disaster should
be better able to respond to natural disasters, making the opportunities for institutional
capture by the rebels small. Thus, when controlling for the context of natural disasters, one
would expect to see a lower likelihood of the onset of violent civil conflict when there is a
higher level of institutional quality and vice versa.
This project uses multiple methods to explore three important hypotheses derived
from this theory. Specifically, I conduct a quantitative test of the relationship between

11
institutional quality and post-disaster violence and a qualitative comparison of the critical
cases of Indonesia and Sri Lanka following the 2004 tsunami. The findings are likely to
impact not only our understandings of post-disaster violence, but also our understanding of
intrastate violence more broadly.

Potential Implications of this Research
The findings of this project contribute to a recent turn in the broader research on the
politics of natural disasters. Specifically, this project adds to the budding research on so-
called “disaster diplomacy.” This state-rebel interaction over disaster relief institutions offers
a political mechanism through which disaster and changing levels of conflict might be linked.
In addition, by viewing natural disasters as “crises” or “focusing events,” they present an
opportunity to better understand the relationship between variables that impact intrastate
conflict and violence more widely. In this way, this project on the relationship between
disasters and violence contributes to the literature and policy-making around intrastate
violence as much as it gains from it.

Disaster Diplomacy
The tendency to conceptualize natural disasters as political events has ebbed and
flowed over time. Several early studies in political science made inroads exploring the
relationship between natural disasters and political change, especially focusing on the
political consequences of great American natural disasters such as the Mississippi River
flooding during the early 1920s and the Dust Bowl during the 1930s. However, over time,

12
this earlier work was marginalized as disaster research became almost explicitly apolitical.
New research was increasingly driven by the needs of the disaster relief community, and as
such, “politics” became associated with potential negative outcomes or impediments to
successful policy implementation such as corruption or obstruction.
However, over the past decade, especially as disasters have increased the quantity
and complexity of the demands on an affected state or region’s political system, they have
begun to be examined as political crises (Olson 2000: 267), “focusing events,” (Birkland
1996, 1998) or “critical junctures” (Olson and Gawronski 2003) in which significant changes
occur in the community and through which major policy shifts can take place. Within this
context, recent work has attempted to focus on the ways in which natural disasters might
affect pre-existing political economies (e.g., Albala-Bertrand 1993a, 1993b; Simile 1995;
Gendron 1998) and security environments (e.g., Olson and Drury 1997; Drury and Olson
1998; Shefner 1999; Brancati 2007; Nel and Righarts 2008) within and between affected
states.
Despite this shift in focus, little empirical or theoretical consensus has developed
around the specific conditions under which disasters contribute to greater or lesser political
unrest and/or violence (c.f., Oliver-Smith 1977, 1979; Blocker, Rochford, and Sherkat 1991;
Olson and Drury 1997; Drury and Olson 1998; Tavera-Fenollosa 1999). In this vein, the
emergence of a literature on disaster diplomacy (e.g., Comfort 2000; Glantz 2000; Kelman
and Koukis 2000; Ker-Lindsay 2000; Kelman 2005, 2006; Akcinaroglu, DiCicco, and
Radziszewski 2008; Enia 2008) has been an important development.
Since the reconciliation between Greece and Turkey following a series of
earthquakes in the two countries in 1999, almost every major natural disaster that has

13
occurred within a zone of conflict has been followed by expressions of hope that some
diplomatic good—often referred to as “disaster diplomacy” or sometimes more specifically,
“earthquake diplomacy”—might emerge from the tragedy. However, greater exploration of
the case of Greece and Turkey reveals that the earthquakes themselves played a secondary
role at best in the peace process (Ker-Lindsay 2000; Evin 2004). In addition, much of the
early literature on this topic has been case-specific and has only hinted at the particular
mechanisms through which natural disasters might lead to peace or conflict.
8
By utilizing a
comparative approach and keeping the connection between political economies and conflict
at the fore, this project contributes to filling these theoretical and empirical gaps.

Disasters as “Critical Junctures”
The idea that relatively exogenous shocks provide good opportunities for more
clearly illuminating the relationships between variables is common in economic approaches
to the study of politics. For example, the existing literature in international political
economy provides a number of examples of how various shocks affect political alignments
(e.g., Olson 1982; Rogowski 1989; Frieden 1991). Such studies have been important for
theoretical understandings of interest group formation and function around issues of trade,
finance, and the business cycle. In addition, the study of crises as particular types of shocks
has been prevalent in other aspects of international relations research. The idea that crises
often better illuminate the mechanisms at work with respect to the relationship between

8
Important exceptions are emerging; cf. Kelman (2006), Brancati (2007), and Akcinaroglu,
DiCicco, and Radziszewski (2008).

14
variables has been central to the International Crisis Behavior (ICB) Project (Brecher and
Wilkenfeld 1982).
As natural disasters represent similar types of crises, they allow us to study political
phenomena within the context of a natural experiment. This project argues that rapid-onset
natural disasters provide a relatively exogenous, but certainly not arbitrary, entry point for
thinking about how rapid changes to political economies affect intrastate conflict.
Specifically, the argument and empirics presented herein point to the fact that even relative
changes in political and economic institutional control have important implications for
violence.
This is a new development for the study of intrastate conflict, particularly separatist
conflicts. To some extent, all separatist conflicts are about a battle to control political
and/or economic institutions. But often this conflict and its outcome are portrayed as an
“all or nothing” proposition. Conflicts are studied in a way that focuses on onset (the
beginning of attempts to control institutions) or duration (the end of the process once
institutions are either captured or once rebellion is suppressed). However, this approach this
does not adequately describe the battles over institutions that occur throughout the conflict
at varying levels of violence.
The study of violence in the context of natural disasters underscores the fact that
institutions, in this sense, are not merely prizes that are awarded the victors at the end of the
contest; they are battled over constantly throughout the contest with different sides
controlling them at different points in time given different contexts.
9
These battles can

9
Here, North’s distinction between institutions and organizations becomes critical (North
1990). Many intrastate conflicts are about the battle to control the organization and
institution-making apparatus of the state. But my argument is that underlying institutional

15
define and re-define the micro-foundations, and often the dynamic evolution, of intrastate
conflict as much as they can define the endpoints of conflict.

The Likely Policy Implications of the Disaster-Conflict Nexus
In addition to their analytical utility for the study of intrastate conflict, the response
to natural disasters has been and will likely continue to be an important policy concern,
particularly where disasters and conflict overlap. This co-incidence of natural disasters and
zones of conflict has historically occurred more often than might be expected. A
preliminary investigation of the sample set includes a number of important examples: aside
from the 2004 tsunami in Indonesia and Sri Lanka, there was an earthquake in Guatemala
that occurred in 1976 during a budding conflict between indigenous Maya and the
Guatemalan government; the cyclone Daimone in Mozambique (1984) which occurred
during the conflict between anti-communist RENAMO rebels and the Marxist government;
hurricane Joan in Nicaragua (1988) which occurred during the conflict between the
Sandinistas and the Contras; a volcano in Goma in eastern Congo (2002) which occurred
during the ongoing conflict between the rebel forces and the national government in
Kinshasa. Most recently, in May 2008, cyclone Nargis ripped through Burma (Myanmar), a
country with long-standing and ongoing conflict between the ruling military junta and a
number of citizen groups.
It is likely that this trend toward the co-incidence of disaster and conflict will
continue and may even accelerate. Recent scientific studies have expressed concern that the

battles occur throughout the conflict, as both sides attempt to control the rules of the game
and thereby structure interaction within the society in their favor.

16
intensity of violent, weather-related, natural events will increase in the future due to the
current pace of global climate change (Ekstrom, Nettles, and Tsai 2006; Intergovernmental
Panel on Climate Change 2007). In particular, scientific models predict that if sea surface
temperatures continue to increase, tropical cyclones, including hurricanes and typhoons,
could become more extreme.
10
This increased intensity when combined with rising sea levels
could result in storm surges that are more frequent and more severe (Schneider et al. 2007:
795; Trenbert et al. 2007: 308-15).
In these future disasters, external parties (both countries and NGOs) looking to
provide disaster relief aid will be bedeviled by a particularly acute dilemma: balancing the
rapid provision of disaster relief assistance with the need to tread carefully around the
potential consequences any assistance may have for pre-existing conflicts. Thus, it will be
important to clearly understand the incentives facing the parties to the conflict. Despite any
initial cooperative rhetoric, the post-disaster context could be utilized to strengthen their
relative bargaining positions. Given these incentives, any external post-disaster relief and
reconstruction efforts may have important consequences for the acceleration or cessation of
violence. In fact, a rhetorical goal of “complete neutrality” with respect to disaster relief
provision may do more harm than good. A better understanding of the specific relationship
between natural disasters and conflict is critical lest future natural disasters generate more
negative indirect consequences than direct ones.

10
It is important to note that the frequency and overall quantity of these storms remains
debated, as a number of different scientific models on the effects of climate change continue
to provide different results due mainly, for example, to changes in the way that tropical
storms have been measured over time (Trenbert et al. 2007: 299-305). Nonetheless, the
intensity of the storms that do occur in the models appears to be universally stronger, given
continued rising sea levels and temperatures (Schneider et al. 2007: 795).

17
Outline of the Dissertation
In chapter 2, I develop the theoretical argument that underpins this project and
establish three hypotheses that will be subject to empirical testing. After considering the
positive and negative aspects of the “greed versus grievance” framework that has defined the
civil war literature for most of the past decade, I argue that a number of important advances
aid our understanding of intrastate conflict and violence. First, the study of intrastate
violence must move beyond its focus on civil wars as discrete episodes of onset, duration
and termination and instead focus on the dynamic and complex processes of conflict and
violence that often occur within states. Second, while the distinction between greed and
grievance has been enlightening with respect to the motives of rebels, both approaches have
the effect of underestimating the role of the state. New approaches need to focus on the
strategic interaction—and in particular, the signaling—that takes place between states and
rebels, especially in light of the preferences, motivations and opportunity structures of
various actors. Third, one important source of strategic interaction is the control over
political and economic institutions. Control for either party results in both material and
non-material benefits, including an increase in each party’s ability to credibly signal its
functional ability to govern. While each group’s ability to control institutions does define the
possible outcomes of conflict, the struggle over this control is dynamic; thus, it defines the
processes of the conflict itself as much as it defines its endpoints. This dynamic fight over
institutional control needs to be more carefully analyzed and articulated.
Building on these conclusions, I establish political economies of state-rebel
interaction over the institutions of disaster relief. The preferences of each actor—states and
rebels—are carefully established as to possible outcomes in the struggle to control political

18
and economic post-disaster relief institutions. I use the derived preferences of each to
generate testable hypotheses about how this particular interaction might contribute to the
likelihood of post-disaster intrastate violence.
Chapter 3 establishes the methodology of the project. Crucially, most of this chapter
is devoted to the justification of using rapid-onset disasters as a type of natural experiment.
Specifically, I argue that a natural disaster in this context is a type of exogenous shock that
helps to illuminate the strategic battles over political and economic institutions that are
crucial to the story of intrastate conflict. In addition, this chapter justifies the mixed-method
empirical design utilized in project. Large-n quantitative analysis is used to test a hypothesis
regarding the link between pre-disaster institutional quality and post disaster violence. A
comparative case study of Sri Lanka and Indonesia following the 2004 Indian Ocean tsunami
is used to explore the implications of the quantitative study for the theory of institution
capture.
The results of the quantitative study are presented in chapter 4. Using individual
measures of institutional quality—corruption, the rule of law and order, and bureaucratic
effectiveness—as well as an additive index, I find that countries with weaker levels of
institutional quality a priori a rapid-onset disaster are more likely to experience the
occurrence of violent civil conflict in the year following a disaster. Countries with relatively
higher levels of institutional quality a priori a rapid-onset disaster are less likely to experience
the occurrence of violent civil conflict in the following year. However, when the critical
cases of Indonesia and Sri Lanka, which are not included in the large-n dataset, are used to
check the study, they present some empirical challenges. A brief discussion of these
challenges is used to motivate the qualitative comparison that follows.

19
The mechanisms that underpin the theory of institution capture are explored more
systematically in chapter 5, using a structured, focused comparison. First, the histories of
Indonesia’s and Sri Lanka’s battles with their respective separatist groups, as those groups
attempted to seize control of local and/or regional political institutions, are established as
pre-tsunami baselines. The cases are then compared analyzing the ways in which the
tsunami changed the incentives for either the GAM or LTTE to further seize control over
institutions by capturing the processes of disaster relief, the extent to which these groups
were successful, and the implications of this relative success for the ongoing patterns of
conflict or peacemaking in the year following the tsunami. The cases of Indonesia and Sri
Lanka demonstrate that although low institutional quality at the time of the disaster may
contribute to post-disaster battles over institutional control, they do not, in these cases,
affect the outcomes of these battles. Ultimately, it is the relative success of the separatist
groups in these cases that contributes to post-disaster peace or violence.
In chapter 6, I conclude by briefly reviewing the findings and then highlighting the
implications of this research. In addition to contributing to the budding research agenda
that explores the relationship between natural disasters and conflict, this project lends
another voice to arguments for a more strategically oriented and institution-centered
understanding of civil war. In this respect, future work that further analyzes the role of
institutional quality and separately the role of strategic signaling in intrastate conflict should
yield important theoretical and empirical insights. Finally, the findings will be used to
reassess a number of the policy concerns introduced here.

20
Chapter 2: Developing the Theory and Hypotheses

The phenomenon of intrastate violence is most often studied by asking, “What are
the conditions that make civil war onset more or less likely?” Much of the study of this
question has focused on the motivations for rebellion, broadly categorizing them as either
“greed” or “grievance.” Although this research has fundamentally advanced our
understanding of civil wars, it generally discounts the role of states and their institutions.
While recent research has attempted to rectify this deficiency, even more robust explanations
of intrastate violence are possible.
This chapter proceeds as follows: First, I briefly review the key cross-national and
comparative studies of civil war, arguing that the debate over greed versus grievance has the
effect of overestimating the importance of the role of rebels in determining civil wars. While
a number of more recent studies have focused on the role of the state, the focus tends to be
on state characteristics and the extent to which they foster or inhibit rebel decision-making.
Future work should reconsider the role that states play as actors with agency and particularly
the types of strategic interactions between states and rebels that this view engenders.
Proceeding from this argument, the second part of the chapter argues that one of the
key strategic interactions between states and rebels is the game played around control over
political and economic institutions. Rebel success or failure regarding institution capture has
important consequences for intrastate violence. Institutional capture for the rebels, or
maintenance of control for the state, even on the margins, results in material benefits as well
as non-material benefits, the most important of which is a credible signal of a functional
ability to govern. Accrual of these benefits for rebels should lead to greater ease in recruiting

21
new members as well as procuring outside support and financing. This positive change in
the rebels’ opportunity structures is likely to lead to further conflict and the onset or renewal
of violence. This is particularly likely to be true when coupled with increased grievances due
to the government’s relative inability to provide disaster relief services. However, if the
government is able to better manage or prevent the rebels from gaining control of these
institutions, the signal sent and resulting changes in the opportunity structures of rebellion,
along with the relative decrease in grievances, would be more likely lead to cessation of
violence or prevent its onset.
This chapter develops the political economies of this state-rebel interaction over
institutions. The model derives the preferences of the government and the rebels as to
whether the government controls institutions, the rebels control institutions, they share
control, or whether there is effectively no control. I use the derived preferences of each to
generate testable hypotheses about how this particular interaction might contribute to the
onset or cessation of intrastate violence.

Predicting Civil Violence: From Discrete to Overlapping Explanations
Grievances have long been the focus of political scientists and sociologists studying
the motivations of rebels. The idea of relative deprivation (Gurr 1970) has been a key aspect
of this approach. Perceived gaps between expectations and outcomes with respect to the
provision of public goods coupled with the rent-seeking, predatory behavior of elites can
produce grievances that trigger collective action and possibly violence (Gurr 1970; Eckstein
1980; Auvinen and Nafziger 1999; Gurr 2000; Nafziger and Auvinen 2002; Thorbecke and
Charumilind 2002; Regan and Norton 2005). Often, these differences in public goods

22
provision produce socioeconomic inequalities that are vertical (between individuals) and/or
horizontal (between groups) leading to a number of possible paths of intrastate violence
(Huntington 1968; Sigelman and Simpson 1977; Eckstein 1980; Stewart 2002; Esteban and
Ray 2008; Østby 2008). Finally, these grievances have been demonstrated to be particularly
acute motivations for political violence under conditions of real or perceived structural
impediments to socioeconomic mobility (Huntington 1968; Eckstein 1980; Zimmermann
1980) such as group exclusion and discrimination (e.g., Nel 2006; Urdal 2006).
11
Empirically,
these grievance-producing inequalities—social, economic, and political—have been linked to
violence in cases such as the Maoist insurgency in Nepal (Murshed and Gates 2005), the
conflict between the separatist Mouvement des Forces Démocratiques de Casamance
(MFDC) and the government in Senegal (Humphreys and Mohamed 2005), the conflict
between indigenous peoples in Chiapas, Mexico and the Mexican government (Stewart
2002), and the civil wars in Uganda and Sri Lanka (Stewart 2002).
However, grievances alone are not likely to be sufficient predictors of the onset of
political violence. In fact, as a set of theoretical and probabilistic explanations, grievances
are likely to over-predict the onset of political violence, leading to a higher expectation of
observable cases of actually existed historically (Kahl 1998: 83; Urdal 2006: 610). In attempt
to explain this deficiency, Sambanis (2002: 223) points out that if grievances can trigger
collective action, the redressing of grievances is essentially a collective good whose provision
presents typical and significant collective action challenges (Olson 1965; Tullock 1974).

11
However, see Lichbach’s work (1989, 1990, 1994) for the important counter-argument
that inequality leads to greater suppression of dissent, creating greater collective action
problems for rebels that may actually lead to less violence, not more.

23
Public choice scholars began to build on these ideas during the late 1990s by
exploring the cost-benefit calculation that underpins individual decisions to engage in
rebellion. Facing the costs associated with collective action, the cost of organizing becomes
an important factor in determining the viability of rebel groups. The work of sociologists
and political scientists studying social movements and contentious politics seemed to
confirm that the opportunity structures facing rebels are as important to explaining rebellion
as are grievances (Eisinger 1973; Oberschall 1973; Gamson 1975; McCarthy and Zald 1977;
Tilly 1978; McAdam, Tarrow, and Tilly 2001).
Building on these ideas, a number of other economists began to model rebellions in
a similar fashion as industries (Grossman 1991, 1995, 1999). In these models, rebels are
“indistinguishable from bandits or pirates,” and their principal objective is to generate profits
from looting (Grossman 1999: 269). Thus, political rebellion is fostered by the opportunities
for the rebels to monopolize resources and utilize violence as a means of extracting wealth.
Within this model, both opportunity and viability become important theoretical conditions
for the existence of civil war (Hirshleifer 1995, 2001; Collier and Sambanis 2005a).

Rebellion, Civil War and the Motive of “Greed”
The empirical work of Paul Collier and Anke Hoeffler has been seminal to this
economic approach to rebellion and civil war onset (Collier and Hoeffler 2001, 2004). The
authors posit that rebel financing—an indicator of viability—comes from three sources:
subventions from other governments, donations from diasporas in other countries, and
from the extortion of natural resources. In addition, they argue that the opportunities for
rebellion can arise from low costs, specifically when rebels’ foregone income is relatively low;

24
when military equipment or other capital specific to conflict is particularly cheap; when
government military capability is particularly weak; and finally when social cohesion within
the country is weak. In order to compare their approach to others, they factor in the
objective grievances that are often given as causes of civil war onset: ethnic/religious hatred,
political repression, political exclusion, and economic inequality.
The regressions run by Collier and Hoeffler find that the variables measuring rebel
opportunity seem to perform better than the variables for grievance in predicting the risk of
civil war outbreak. In particular, they find that natural resources, a variable they measure as
the share of gross domestic product taken by primary commodity exports, are a significant
factor in explaining civil war onset. The argument is that primary commodity resources
provide greater opportunities for extortion, making rebellion both more feasible and
attractive (Collier and Hoeffler 2001; Collier, Hoeffler, and Sambanis 2005: 17).
Another important factor related to the risk of civil war is the opportunities of
rebellion. The Collier and Hoeffler model predicts that as opportunity costs, the earnings
forgone by rebellion, are relatively low—a measure proxied using education enrollment, per
capita income, and growth rates—the risk of civil war should increase. Finally, there are
other important opportunities in the form of military advantages, and the risk of civil war is
mildly higher when populations are dispersed and when a country contains more
mountainous terrain. With the exception of ethnic dominance, none of the measures of
grievance has statistically significant effects within the Collier-Hoeffler model (Collier and
Hoeffler 1998, 2001; Collier, Hoeffler, and Sambanis 2005).
Given the relative novelty of their findings, a large amount of work followed the
publication of Collier-Hoeffler’s, and a number of important studies seemed to offer

25
confirmation (e.g., Collier and Hoeffler 2002; Collier, Hoeffler, and Soderbom 2004; Ross
2004, 2006). The model seemed to fit a number of cases, correctly indicating a high risk of
civil war prior to actual outbreak of war in Mozambique 1976 (Weinstein and Francisco
2005), Algeria in the early 1990s (Lowi 2005), Indonesia in Aceh (Ross 2005), Colombia
(Sanchez, Solimano, and Formisano 2005), and Nigeria from 1980-1984 (Zinn 2005) among
others. In addition, the model indicates a lower than average risk of civil war onset in
several countries which either do not have a war or where the violence does not cross the
definitional threshold: Kenya (Kimenyi and Ndung'U 2005) and Northern Ireland
(Woodwell 2005) are two important examples.

Criticizing the Greed-Grievance Distinction
Despite the widespread attention garnered by Collier and Hoeffler’s findings (e.g.,
Rosenberg 2002), a number of critiques have been articulated. While generally aimed at the
Collier-Hoeffler finding on the relative importance of primary commodity exports, these
arguments also attack the foundations of the purely economic approach to the study of civil
wars by pointing out: 1) there are a large number of cases that do not seem to fit the Collier-
Hoeffler model; 2) the lack of fit with many of the cases can be explained by the fact that the
economic approaches to the study of civil wars have yet to adequately theorize the political
mechanisms at work; 3) the distinction between greed and grievance as discrete explanations
is a theoretically and empirically unhelpful one; 4) both the greed and grievance approaches
underestimate the role of the state in explaining civil war; and 5) civil wars are better
understood as dynamic and complex processes of conflict and violence rather than as a
collection of discrete episodes of onset, duration and termination.

26
First, despite the empirical basis on which Collier and Hoeffler derive their
argument, there are important cases that seem to provide disconfirming evidence either as
false positives or false negatives. In a series of case studies chosen and designed to test and
refine the Collier and Hoeffler model (Collier and Sambanis 2005a, 2005b), there are a
number of cases where the model predicts a low probability of civil war but the cases
demonstrate its occurrence: Burundi (Ngaruko and Nkurunziza 2005), Nigeria between
1965-1969 (Zinn 2005), Senegal (Humphreys and Mohamed 2005), Mali (Humphreys and
Mohamed 2005), and Lebanon 1975-1990 (Makdisi and Sadaka 2005). There are also a
number of false positive cases where the model predicts a high probability of war but where
the cases demonstrate no war (or the violence that does occur does not meet the threshold
definition for civil war
12
): Macedonia 1990-2001 (Lund 2005) and Nigeria 1985-1999 (Zinn
2005). Particularly in light of the false positive cases, one is left searching for an explanation
as to why the critical conditions in the Collier-Hoeffler model seem to lead to civil war in
some cases but not others.
13

The second criticism of this approach to understanding civil war is that despite any
apparent correlation between the greed variables and civil war onset, the actual political
mechanisms through which the greed variables operate is unclear. For example, are the
primary commodity export variables really capturing an aspect of grievance or are they
actually capturing nothing more than viability? Much of this confusion arises from a

12
On the theoretical and methodological controversies regarding the definition of “civil
war,” see Sambanis (2002, 2004).
13
Snyder (2006) is motivated by a very similar question, pointing out that among the 42
countries that produced alluvial diamonds, tropical timber, and/or illicit drugs between 1960
and 1999, 24 of them (57%) did not experience civil war.

27
comparative exploration of specific cases. There are cases where a country has some of
elements that would to put them at risk for civil war but unexpectedly lacks a number of
other seemingly critical elements. Algeria, for example, has high primary commodity export
levels which would place it a high risk for civil war onset but relatively low income from its
diasporas which would place it at low risk (Lowi 2005). Which of the two (if any) then best
explains the decade of civil war that began in 1992? The case of Bosnia is similar in that its
primary commodity exports were relatively low in 1990-1992, a fact that makes it
inconsistent with the Collier-Hoeffler model; however, the Bosnia case does appear
consistent with a number of other explanatory variables, such as geographic dispersion and
mountainous terrain (Kalyvas and Sambanis 2005: 222-23).
There are also cases where despite the model’s accurate prediction of the presence or
absence of civil war, the actual mechanisms leading to war are not adequately captured. In
several cases, natural resources and other primary commodities appear to be important
explainers of the onset of civil wars, but they do so in different ways from each other and
from the logic present in the Collier-Hoeffler model. For example, in the Democratic
Republic of Congo, natural resources do seem to have been a significant determinant of the
civil wars; however, the geographic concentration and resulting unequal distribution of those
resources among ethnic groups seems to have mattered more than the state’s dependency on
the export of those resources (Ndikumana and Emizet 2005). In Indonesia, the effects of
primary commodities do not appear to manifest themselves in easier financing for rebels as
Collier and Hoeffler predict but instead in a set of grievances over resource distribution, in a
larger military, and in a lack of government credibility (Ross 2005).

28
Other qualitative studies of specific cases of civil war point to an altogether different
set of causal mechanisms for war onset. For example, while the Collier-Hoeffler model does
correctly predict a high risk of civil war in Nigeria between 1980-1984, the Maitatsine War
occurred absent a significant source of rebel financing. The Maitatsine rebels fought with
very primitive weapons and received very little funding from outside diasporas or
governments or from natural resource looting (Zinn 2005: 104). In addition, the conflict
occurred in a particular political context where the Kano state government was initially
reluctant to completely repress the Maitatsine movement due to constitutional human rights
guarantees. The aerial and ground assaults by the government that did eventually occur
when the Maitatsine uprising began in 1980 demonstrate that a lack of political will shaped
the Kano state government’s response to the rebels more than capability (Zinn 2005: 105).
With respect to the 1976 civil war in Mozambique, the model fails to capture the
important effects of the interaction between the external backing of the Rhodesian
government in organizing the Mozambican National Resistance (Resistência Nacional
Moçambicana, RENAMO) rebels and the emergence of internal grievances that were at the
heart of the rebels’ ability to grow and expand (Weinstein and Francisco 2005). The
capability of RENAMO depended heavily on external financing, and to some extent this
confirms the logic of the Collier-Hoeffler model. However, over time, the Marxist policies
of the Front for the Liberation of Mozambique (Frente de Libertação de Moçambique,
FRELIMO) government created a number of grievances that provided RENAMO a base of
support among Mozambican peasants (Bowen 1989; Pitcher 1998; Bowen 2000). The
interaction of these external and internal forces is something that is not adequately captured
in the Collier-Hoeffler model (Weinstein and Francisco 2005). Table 1 summarizes a

29
number of cases that lend empirical credence to these first two criticisms of the Collier-
Hoeffler finding.

30
Table 1: Disconfirming Cases and the Collier-Hoeffler Model
14

Case Years False Negative/Positive, Mechanism Research
Burundi 1965-1972;
1988; 1991;
1993-present
False negative (Ngaruko and
Nkurunziza 2005)
Nigeria 1965-1969 False negative (Zinn 2005)
Senegal 1983-present False negative (Humphreys and
Mohamed 2005)
Mali 1990-1996 False negative (Humphreys and
Mohamed 2005),
Lebanon 1975-1990 False negative (Makdisi and Sadaka
2005)
Macedonia 1990-2001 False positive (Lund 2005)
Nigeria 1985-1999 False positive (Zinn 2005)
Burundi 1972-1988 False positive (Ngaruko and
Nkurunziza 2005)
Algeria 1992-2002 Wrong mechanism—primary commodity
exports; but no external funding
(Lowi 2005)
Bosnia 1990-1992 Wrong mechanism—low primary
commodity exports; other factors more
important
(Kalyvas and
Sambanis 2005)
Dem. Rep.
of Congo
1964-1965;
1990s
Wrong mechanism—natural resources, yes;
but internal grievances they create more than
state’s export dependency
(Olsson and Fors
2004; Ndikumana
and Emizet 2005)
Mozambique 1976 Wrong mechanism—mixture of external
financing and grievances
(Weinstein and
Francisco 2005)
Indonesia 1976-1979,
1989-1991,
1999-2005
Wrong mechanism—natural resources, yes;
but internal grievances over distribution and
gov’t legitimacy
(Ross 2005; Aspinall
2007)
Nigeria 1980-1984 Wrong mechanism—no financing for
Maitatsines; lack of state political will
(Zinn 2005)
Sierra Leone 1991-2003 Wrong mechanism—some grievances
important; role of coercion not adequately
captured
(Humphreys and
Weinstein 2008)

14
The cases and years coded as false negatives in the table represent years where the Collier-
Hoeffler model predicts a low probability for civil war, but civil war actually occurred. For
those cases and years coded as false positives, the model predicts a high probability for civil
war, but in reality civil war either does not break out or the violence that does break out is
not enough to meet the threshold definition of 1,000 battle-related deaths that is commonly
used to identify civil wars.

31
The third criticism of the Collier-Hoeffler approach to studying civil wars aims at the
idea of studying civil wars using discrete and competing frameworks of greed and grievance.
Indeed, a number of scholars have begun to return to the idea that the dichotomous nature
of this argument is largely unhelpful (Kalyvas 2001; Regan and Norton 2005).
15
The
response has been to properly contextualize the ways in which both greed and grievance
motivate rebellion (Gates 2002; Weinstein 2003; Humphreys 2005; Korf 2005; Weinstein
2005; Korf 2006; Weinstein 2007; Humphreys and Weinstein 2008). Weinstein’s work, for
example, focuses on the micro-foundations of rebel recruitment and violence in a way that
accounts for both the types of grievances which motivate rebellion and the quantity and
quality of resources with which they can rebel (Weinstein 2007). Both play important roles
in the rebels’ ability to organize and recruit as well as they level of violence groups are willing
employ.
The fourth critique of the Collier-Hoeffler approach begins by criticizing the use of
primary commodity exports as an empirical proxy for rebel financing opportunities. While
the general theory about the importance of rebel financing might be correct, there is no a
priori reason to think primary commodity exports is the most adequate variable for testing
the theory (Fearon 2005). In fact, Fearon’s empirical work (Fearon and Laitin 2003; Fearon
2005) has generally found little support for any independent effect of primary commodity
exports on civil war onset. The differences between the two studies appears due to the fact

15
To be fair, even those that gravitate toward grievance explanations recognize that “it is
simplistic to argue that one kind of motivation is primary and the others subsidiary” and that
“protest and rebellion are consequences of complex interactions among collective
experience, normative commitments, contention for power, and strategic assessments about
how best to promote individual and collective interests” (Gurr 2000: 66). Gurr’s arguments
have contained this nuance for some time (Gurr 1993).

32
that the effect of primary commodity exports in the Collier-Hoeffler models seems largely
driven by the subcategory of oil exports (Fearon and Laitin 2003). As a primary commodity,
oil has been found trigger conflict more through the political consequences that flow from
rent-seeking behavior than through purely economic effects (Karl 1997; Ross 1999).
This suggests the need to bring the state back into the study of civil wars (Ron 2005:
448). The rebel-centric nature of the greed versus grievance turn in the literature effectively
ignores any active role of the state as a participant in the outbreak of intrastate conflict
(DeRouen and Sobek 2004). Consideration of the state requires an exploration of
institutions–both their effects on structure and process (Taydas, Peksen, and James 2008) as
well as the specific ways in which the strategic interaction between states and citizens
produce different levels of intrastate conflict.
The fifth argument levied against the Collier-Hoeffler approach is that it is too
focused on civil war onset as opposed to duration or termination. While the distinction
between onset and duration is empirically useful,
16
civil wars might be better understood if
they were studied as the outcomes of complex patterns of conflict, violence and criminal
activity (Andrienko and Shelley 2005; Sanchez, Solimano, and Formisano 2005). Attempting
to study these patterns in discrete phases can lead to selection bias where the determinants
of the phases are similar (Elbadawi and Sambanis 2002). If, for example, a particular cause is
associated with civil war onset and a sample of conflict-affected countries is used to study
civil war duration, than it may be that no statistical relationship is found between the cause

16
The distinction allows for the possibility that the set of factors that spark the beginning of
a civil war may, in fact, be different from the set of factors that condition the continuation of
war over a period of time. On issues related to data coding of civil wars, see Sambanis
(Sambanis 2002, 2004).

33
and duration. However, this finding could be due to the lack of variance that emerges as a
result of the bias inherent in the chosen cases. This statistical confusion can mute the
complexity of the interdependence between phases and causes.
This critique also suggests that the study of the broader category of intrastate
violence might be preferable to the more narrowly defined subcategory of civil war.
Empirically, it is difficult to arrive at an operational definition of civil wars without placing
arbitrary boundaries around the concept that allow it to be distinguished from other forms
of violence. This operationalization has important consequences for statistical findings
(Sambanis 2004). Theoretically, it is difficult to imagine civil wars as having a distinct causal
logic apart from other forms of violence or “coordinated destruction” (Tilly 2003: 14).
Each of these five reactions to the greed versus grievance turn in the civil war
literature, in its own way and as a collective, portends the need for an approach to the study
of intrastate conflict that goes beyond the analytical structures of greed and grievance,
beyond the structures of onset, duration and termination, and even beyond the narrowly-
defined operationalization of civil war. A clearer examination of rebel and state strategic
decision-making around all phases and levels of intrastate conflict and violence offers one
possible path toward these goals.

Bringing Back Institutions and Interaction
An approach to intrastate conflict that recognizes the active role of the state, the
importance of institutions, and the strategic interaction between states and rebels is
consistent with the recent trajectory of the literature that builds on the critiques previously
outlined (cf. Berdal and Malone 2000; De Soysa 2002; Ballentine and Sherman 2003; Collier

34
and Hoeffler 2004; Korf 2005; Regan and Norton 2005). Political institutions play an
important role, both as sources of constraint and opportunity, in the processes associated
with intrastate conflict (Fearon 2005). A number of studies have already moved in this
direction, focusing on the more formal features of political institutions such as the level of
democracy (Ellingsen 2000; Hegre et al. 2001; Sambanis 2001; Elbadawi and Sambanis 2002;
Reynal-Querol 2002a, 2002b), the general level of inclusiveness of the political system
(Reynal-Querol 2005), the characteristics of political regimes (Ellingsen 2000), and electoral
rules (Reynal-Querol 2002b, 2002a; Carey 2007). While the general intuition behind many of
these studies is that democracies should be more peaceful than autocracies, the empirical
evidence shows that both long-standing, stable autocracies and long-standing, stable
democracies are more peaceful than democracies in transition (Ellingsen 2000; Hegre et al.
2001; Sambanis 2001; Reynal-Querol 2002a, 2002b, 2005). Institutionally inconsistent
regimes, containing a mix of autocratic and democratic features, seem to drive this result as
the lack of self-reinforcing institutional aspects tend to move a country away from a stable
equilibrium (Gates et al. 2006).
Other studies have explored the importance of institutions that reflect on state
capacity in determining the likelihood of conflict. Most of these posit a negative relationship
between state capacity and civil war, arguing that the greater the ability of a state to repress
insurgencies, for instance, the less likely is an outbreak of full-blown civil war (Mason,
Weingarten, and Fett 1999; De Soysa 2002; Fearon and Laitin 2003; Snyder 2006). In this
light, a state’s capacity to defeat an insurgency is a function of the overall effectiveness of the
state bureaucracy (DeRouen and Sobek 2004). Fearon and Laitin (2003: 75-76) point out
that central governments that are organizationally, financially and/or politically weak make

35
insurgencies relatively more attractive as such central governments have poor local policing
and inept or corrupt counterinsurgency practices.
A variety of economic institutions have also been offered to explain how variance in
state capacity affects intrastate conflict. Economic institutions that support extraction in
countries with large lootable resources such as alluvial diamonds or opium provide either a
base of revenue for the state or a base of financial support for the rebel group (Snyder 2006).
However, in states that have an abundance of natural resources, particularly in oil, the
abundance often leads to weak extractive apparatuses, as the incentive to tax citizens is
lessened due to the reliance on oil revenues (Karl 1997; Ross 1999). This has the effect of
stunting the development of the state’s administrative capacity, a weakness which often
manifests itself in weaker rule of law (Ross 2004: 338) and poor governance, more generally,
including corruption, poor provision of services, and poor management of the economy
(Sachs and Warner 2000; Collier and Hoeffler 2004: 567). These relatively poor capabilities
create opportunities for rebels providing them with a set of grievances (Collier and Hoeffler
2004) as well as with an opportunity to seize control of extractive apparatuses (Fearon and
Laitin 2003; Fearon 2005; Snyder and Bhavnani 2005; Snyder 2006). The ease with which
rebels can access sources of financing has an important impact on their ability to fight as
well as to recruit new members (Gates 2002; Weinstein 2007).
In addition to extractive capacity, a number of measures of government defense
capacity have been explored. The most straightforward measure of effectiveness in this
regard is the extent to which various governments are able to mobilize their militaries to
combat a rebellion. This appears to be a function of factors such as whether the rebel threat
is primarily internal or external (Herbst 2004) and the extent of fungible resources provided

36
to the government by the international community (Azam 2002; Herbst 2004). All of these
different approaches to studying political institutions and their resulting outcomes suggest
that specific institutional attributes, such as institutional quality, play important roles in the
prevention or onset of intrastate violence.
Taydas, Peksen and James (2008) find some preliminary empirical support for the
idea institutional quality should impact the likelihood of civil war onset. They argue that
measures of quality such as corruption, the rule of law and order, and the effectiveness of
the bureaucracy represent the specific components of an “effective governance structure.”
Where these components are lacking (i.e., where institutional quality is low) they are likely to
affect the onset of civil war both through the broad avenues of grievance and opportunity.
With low levels of institutional quality, the perceived lack of governmental efficacy provides
signals of state weakness, opportunities to question legitimacy, and a set of grievances
(Schock 1996: 107; La Porta et al. 1997). In the case of high levels of corruption (Della
Porta and Pizzorno 1996; Sandholtz and Koetzle 2000; Warren 2002), lack of the rule of law
and order (Diamond and Morlino 2004; Rogers 2004), and low levels of bureaucratic
effectiveness (Knack and Keefer 1995; Knack 2001; DeRouen and Sobek 2004), the
resulting political grievances can lead to political instability.
In addition, these components of institutional quality are likely to impact the
likelihood of violent civil conflict by changing the opportunity structures of rebellion,
specifically the opportunity costs associated with rebel recruitment. For example, as the
perception increases that there is a lack of good alternatives for redressing grievances, this
lowers the opportunity costs associated with rebel recruitment of monetary and human
support. In addition, lower quality institutions have been associated with lower-levels of

37
long-run economic development (e.g., North 1990; Shleifer and Vishny 1993; Knack and
Keefer 1995; Mauro 1995; Clague et al. 1996; Rodrik 1999; Knack 2001). In this case, a lack
of economic alternatives for potential rebel recruits also serves to lower the opportunity
costs associated with rebel support. These changes in the opportunity cost structures impact
the overall capability of the rebels to effectively combat the state.
In addition to the focus on the role that institutions play in fostering and/or
preventing intrastate violence, a number of studies have begun to turn away from the
exclusive focus on either states or rebels fostered by the greed versus grievance debate. A
different approach focuses on how the particulars of the strategic interaction between states
and rebels might lead to peace or conflict. Both governments and rebels act based on
perceptions of what the other party has done in the past and what it is likely to do in the
future. This brings into play the possibility that factors such as misperception and credibility
are important determinants of intrastate conflict as are both parties’ relative signaling
abilities.
Governments and rebels tend to think more about the future than current theories
might anticipate. They are careful to build into their strategic calculations the probable
outcome and expected duration of the current conflict (Mason, Weingarten, and Fett 1999;
DeRouen and Sobek 2004: 304) as well as the anticipated risks and costs of future challenges
(Walter 2006a: 324; 2006b: 106). Overall, the relative strength of both parties has important
consequences for signaling their respective abilities to commit and build on-going
reputations in the conflict bargaining game (Fearon 2004; Walter 2006a, 2006b).
Signaling by the respective parties to the conflict (or crisis) has long been recognized
as an important component of this strategic interaction (Fearon 1994, 1997, 1998). Signals

38
sent by either party provide the other side with information about relative strengths in
capabilities or interests, and this information can influence strategic decision-making. In the
literature on civil wars, several studies focus on the motivations underpinning particular
ethnic groups seeking self-determination. These motivations include the extent to which the
group feels relatively disadvantaged within a country compared to other ethnic groups (Gurr
2000; Marshall and Gurr 2003), the extent to which it feels relatively incapable of achieving
its particular economic goals (Gurr 2000), and/or the extent to which it perceives the ruling
government’s central authority to be in decline (Bartkus 1999). In addition, from the
perspective of governments, they may be more or less willing to fight against the insurgency
on the basis of the relative capabilities as well. The government’s choice to fight the
insurgency (as opposed to ignoring it) is likely based on the economic, political and/or
psychological value of the territory in question (Diehl 1999; Toft 2003).
Both of these advances in the study of intrastate violence—the focus on institutions
and the focus on the strategic interaction between states and rebels—influence the
hypotheses derived in the next section. The outcomes of the strategic battles between states
and rebels over institutional control have important implications for civil violence. Each
group stands to gain both material and non-material benefits from institutional capture, and
these benefits impact levels of greed, grievance, and opportunity. As such, they lead to a
number of testable hypotheses.

39
Deriving the Preferences of Institution Capture in a Post-Disaster
Context

Immediately following a natural disaster, there are a number of possible outcomes
with respect to institutional control. Both states and rebel groups have material and non-
material incentives to seize control over the institutional mechanisms of disaster relief.

Possible Outcomes for Post-Disaster Institutions
Within the context of post-disaster relief provision, I distinguish among five
different outcomes. State control of disaster relief institutions means that the state is largely
able to control the processes and functions of disaster relief provision. In this outcome the
state’s ability to control these institutions could be achieved through a lack of rebel capability
or through an ability of the state to suppress any rebel attempts at control. A recent example
of state control exists in Myanmar (Burma) after cyclone Nargis. The ruling military junta
was able to maintain control over nearly all of the political and economic institutions of
disaster relief provision. Any opposition groups had little capacity to take control of these
institutions, and where they attempted to, the government was able to block their attempts
(e.g., Mydans 2008). This outcome is contrasted with rebel control of disaster relief
institutions. Here the rebels are able to seize control of the institutions of disaster relief
provision and in conjunction with either the state’s inability or ineptitude at controlling these
institutions or with their own ability to prevent the state from controlling them.
Another possible outcome is shared control of disaster relief institutions and processes.
In this scenario, the state and the rebels each share some control of the institutions and
provision of disaster relief. This can occur without a formal agreement, as both parties are

40
able to capture the institutions that service their more localized populations and neither is
able to capture the institutions in the other’s territory. Alternatively, there might be a formal
agreement between the parties about how and by whom disaster relief will be distributed.
After the 2004 tsunami in Sri Lanka, the Sri Lankan government and the Liberation Tigers
of Tamil Eelam (LTTE, commonly referred to as the “Tamil Tigers”) were able to come to
an agreement that allowed the LTTE a greater role in the distribution of over $3 billion in
aid. The aid-sharing agreement, known as the Post-Tsunami Operational Management
Structures (P-TOMS), provided that review committees formed of LTTE rebels,
government officials as well as Muslims and other non-LTTE Tamil groups would be
allowed to recommend and monitor projects in areas hit by the tsunami.
17

A fourth possible outcome in the interaction over post-disaster relief institutions is
outside control. Here, neither the government nor the rebels is adequately able to capture the
institutions of disaster relief, and outsiders direct most of the processes of disaster relief.
These outsiders might include other states, intergovernmental organizations (IGOs), or non-
governmental organizations (NGOs). Finally, the outcome of no control is a possibility. In
this outcome, neither the government nor the state is able to capture the institutions of relief
provision, and at the same time neither is any outside state or organization. This lack of
capture could be in conjunction with outside disinterest or with the state or rebels’ (or both)
successful attempts to thwart outside intervention.
Assuming that 1) states want to remain in power, 2) rebels want to capture power,
and 3) institution capture serves as a means of gaining governmental power both directly and

17
The P-TOMS aid-sharing agreement, which was ultimately ruled to be unconstitutional by
the Sri Lankan Supreme Court, will be discussed in more detail in chapter 5.

41
indirectly as a signal of one’s functional ability to govern, then it is possible to derive a set of
hypotheses that explain the relationship between institution capture and conflict, particularly
in the post-disaster context.

Deriving Hypotheses around the Benefits of Institution Capture
These benefits that accrue from institutional capture should have consequences for
the level of violent civil conflict in the country. Specifically, an increase in rebel control over
disaster relief institutions should lead to greater conflict in a country, as the opportunity
structures of rebellion change and the set of domestic grievances about the government
increase. A similar increase or maintenance of control over institutions by the state should
lead to lower levels of violent civil conflict within a society. In this situation, we would
expect a relative decrease in domestic grievances and a relative increase in the opportunities
of rebellion due to the material and nonmaterial benefits accrued (or lost) as a result of the
state maintaining solid institutional control.
Maintenance or relative capture of institutional control by the state generates both
material and non-material benefits. Monopoly control over these important institutions
allows states first and foremost to control the influx of foreign aid. Aid flows carry a variety
of rent-seeking opportunities (Svensson 2000) and often result in an increase in government
consumption that benefits a wealthy political elite (Boone 1996). The receipt of the aid itself
allows the government to spend on disaster relief without raising taxes or initiating some
other source of revenue generation. Control over these institutions allows the state to
determine the rules and procedures associated with fund disbursement, affecting outcomes
such as the companies or organizations that receive the rebuilding contracts, the geographic

42
locations which are rebuilt most quickly, and the implementation of measures intended to
prevent or mitigate the effects of future disasters. Finally, monopoly control of its disaster
relief institutions also has various transaction cost benefits, as the state does not have to
worry about enforcement and other collective action costs that inevitably come with
institution-sharing agreements that could be signed with rebels or with outsiders (Keohane
1984; Levi 1988).
These material benefits that accrue to the state due to relative institution capture
should decrease levels of post-disaster violence through the channels of grievance and
opportunity. If the state maintains control over these institutions, the viability of rebellion
should decrease. First, the capacity of the state to fend off rebellion is positively affected by
this development. The state is able to effectively utilize the post-disaster resources and aid
flows without diverting resources from other spending. The net affect of this is greater
resources to spend fighting any rebel movements. Second, the material benefits allow the
state to provide better services to the citizens resulting in relatively fewer grievances among
the general population. Finally, given these post-disaster services for citizens, the individual
opportunity cost of the choice to rebel should be higher, as the quality of the life forgone by
a choice to rebel is relatively high. In this environment, it becomes difficult for rebel leaders
to recruit and the problems of collective action become more difficult to overcome.
States also stand to gain critical and strategic non-material benefits from their relative
abilities to control disaster relief institutions, in whole or in part. To the extent that states
are able to control the disaster relief process, the state is able to send a signal that the
government continues to be the best provider of services for the people. This positive signal
theoretically should result in increased support for the legitimacy of the current government

43
by both outsiders, countries and NGOs, as well as the government’s own citizens.
Strategically, this signal also has the effect of weakening the rebel claims to legitimacy and in
so doing, making it more difficult for to rebels to recruit resources in the form of people and
money from both within and outside the population.
These signals about the functional legitimacy of the state should also decrease post-
disaster violence through the avenues of grievance and opportunity. As states are able to
send a signal of their legitimacy, the level of grievance among the general population should
decrease. Citizens see their state as relatively capable of providing the goods and services
they expect, particularly in a critical state-of-emergency situation like a natural disaster. In
addition, this decrease in grievances should make the viability of rebellion more difficult as
rebel leaders find it harder to generate out-of-country support, rhetorical and financial, from
diasporas, third-party countries, and/or the international community writ large.

HYPOTHESIS 1: In the context of natural disasters, relative institution capture—or maintenance of
institutional control—by the state should decrease the risk of post-disaster violence.

Rebels have the same material incentives as the state in gaining monopoly control
over disaster relief institutions. For rebels, however, these benefits are magnified given their
resource and revenue-raising asymmetries vis-à-vis the government. Control over the
provision of disaster relief aid provides the rebels with similar benefits as the state in terms
of directing the recipients of rebuilding contracts and the geographic locations that are
rebuilt most quickly. Each of these outcomes can be directed in such a way as to increase
support for the rebel cause.

44
These material benefits that accrue to the rebels due to relative institution capture
should increase levels of post-disaster violence through the channels of grievance and
opportunity. If the rebels gain marginal control over these institutions, the viability of
rebellion should increase. First, the inflow of foreign aid presents a potential source of
revenue for the rebels in the same way that primary commodities do in the Collier and
Hoeffler models. Any increase in revenue should increase their capacity to fund the logistics
of rebellion. Second, the capacity of the state to fend off rebellion is negatively affected by
this development. Because the state is forced to spend a higher proportion of its own funds
on the processes associated with disaster relief, it has a relatively smaller pool of resources to
spend fighting any rebel movements. Again, this increases the potential viability of rebellion.
Third, the fact that the state is not as able to provide services could lead to higher levels of
grievance among the general population. Finally, given these post-disaster services for
citizens, the individual opportunity cost of the choice to rebel should be lower, as the quality
of the life forgone by a choice to rebel is relatively low. In this environment, it becomes
easier for rebel leaders to recruit and the hurdles of collective action are relatively easier to
clear.
In addition, rebels also stand to gain critical and strategic non-material benefits from
their relative abilities to capture disaster relief institutions, in whole or in part. For rebels
relative control over these institutions sends two related signals to citizens and outsiders: 1)
the current government cannot adequately provide both private and public goods and
services to the people, particularly in times of crisis, and 2) the rebel group presents a
potential alternative that can provide these items. In the battle over state control, this signal
is a powerful one, and it is expected that the rebels’ relative ability to send such a signal

45
would have important consequences for its ability to recruit new members from amongst the
citizenry, members who might offer political support for the rebels and/or be willing to fight
along side them. Signaling a functional ability to govern could also have positive
consequences for the rebels’ outside support. It is likely that other governments, inter-
governmental organizations (IGOs), and NGOs would be more willing to support the
rebels’ cause if it becomes apparent that the rebels represent a legitimate alternative to a state
government that has become ineffectual.
These signals about the functional legitimacy of the state should also increase post-
disaster violence through the avenues of grievance and opportunity. First, as states appear
less capable of providing expected goods and services, particularly in a critical state-of-
emergency situation like a natural disaster, the level of grievance among the general
population should increase. Second, this increase in grievances should make the viability of
rebellion more likely as rebel leaders should find it relatively easy to generate out-of-country
support, rhetorical and financial, from diasporas, third-party countries, and/or the
international community writ large.

HYPOTHESIS 2: In the context of natural disasters, relative institution capture by rebel groups should
increase the risk of post-disaster violence.

Finally, if the battle over institutional control leads to changes in political violence
amidst a disaster, we might expect to see violence indirectly affected by the quality of
institutions. States with higher levels of institutional quality should be better able to respond
to natural disasters, making the opportunities for institutional capture by the rebels small.

46
Thus, when controlling for the context of natural disasters, one would expect to see a lower
likelihood of the onset of violent civil conflict when there is a higher level of institutional
quality and vice versa.

HYPOTHESIS 3: In the context of natural disasters, lower quality political institutions increase the risk of
the onset of violent civil conflict.

Before proceeding with a test of these hypotheses, some specification of the research
design and variables are in order. In the next chapter, I begin with an exploration of the
analytical utility of placing this explanation of intrastate violence within the context of post-
disaster situations. This is followed by a quantitative empirical test of hypothesis 3—the
relationship between institutional quality at the time of the disaster and post disaster violence
(chapter 4). As the testing the first two hypotheses is problematic using quantitative analysis,
a qualitative comparison of two case studies (chapter 5) allows for more concrete
illumination of the causal processes. In the midst of each of these chapters, a number of
important variables will be specified, particularly the types of political and economic
institutions utilized in the process of disaster relief, the measurements used to assess levels of
intrastate violence, and the data utilized to control for various contexts such as level of state
capacity and the type of regime.

47
Chapter 3: Natural Disasters, Data, Methods and Cases

As stated in the introductory chapter, this project has two goals with respect to
natural disasters and intrastate conflict. One is obviously to establish the theoretical and
empirical relationship between the two events themselves. The other, however, is to
consider the period of time after a natural disaster as an analytical window through which we
might better view ongoing strategic interactions between warring groups (states and rebels),
particularly as they battle for institutional control. Thus, post-disaster behavior may be able
to provide insight into state-rebel interaction more generally. Before proceeding to the
empirical tests of the hypotheses established in the previous chapter, it is necessary to
establish the theoretical and empirical utility of using natural disasters as a means of
achieving this latter goal.
This chapter aims to justify a number of different aspects of this project’s research
design by discussing its merits and potential drawbacks. First, I discuss the use of natural
disasters as contextualizing events. I argue that their use in this project constitutes an
exogenous shock creating a type of natural experiment in which the interaction between
states and rebels over political and economic institutions might be seen more clearly. The
wider literature on natural disasters themselves has begun to explore the ways in which they
affect political outcomes. Second, I discuss the research design of the project, specifically
focusing on the use of both quantitative and qualitative approaches. The large-n,
quantitative study is used to analyze the relationship between institutional quality and post-
disaster violence, while the qualitative case studies are used to explore the specific
mechanisms presented in hypotheses 1 and 2 in the previous chapter.

48
Using Natural Disasters as an Analytical Tool

How can natural disasters assist researchers in better understanding the nature of
intrastate conflict? In this section, I develop three arguments for the use of natural disasters:
1) they are increasingly considered by researchers to be events that have political
consequences among others; 2) they serve analytically as a type of exogenous shock which
allows for analysis in the context of a natural experiment; and 3) they are likely to continue—
with possibly worse effects in the future—making the relationship between natural disasters
and conflict one with important policy consequences and worth understanding.

Disasters as Political Events
The tendency to conceptualize natural disasters as political events has ebbed and
flowed over time. Several early studies in political science made inroads exploring the
relationship between natural disasters and political change. This research focused on the
United States and the relationship between a more general category of weather-related
phenomenon and various state or regional political developments. Taken as a whole, it
demonstrates that the negative effects of natural disasters can alter the contextual landscape
in which important political changes can occur. Scholars observed a number of important
post-disaster political developments such as the emergence of particular political parties
(Barnhart 1925), the adaptation of particular institutional features (Walker and Hansen
1946), and changes in political culture (Abney and Hill 1966).
However, this earlier work was largely ignored as disaster research became almost
explicitly apolitical. Richard Stuart Olson (2000: 266) argues that much of this newer
research was conducted in the public policy realm and thus tended to focus on normative

49
assessments of prevention, relief, and reconstruction policies. In this sphere, the notion of
“politics” was associated with potential negative outcomes or impediments to successful
policy implementation such as corruption or obstruction (cf. Freudenheim 1979; Cuny 1983;
Albala-Bertrand 1993b). In addition, the discipline of political science as a whole began to
wrongly treat natural disasters as merely epiphenomenal or as problems to be left for
engineers or urban planners (Olson 2000: 265).
Aiming to rectify this deficiency, several recent and noteworthy studies attempt to
bring the politics back into the study of natural disasters. As disasters increase the quantity
and complexity of the demands on an affected state or region’s political system, they can
easily turn into political crises (Olson 2000: 267). Thus, some have returned to exploring the
potential of natural disasters as “focusing events” (Birkland 1996, 1998) or “critical
junctures” (Olson and Gawronski 2003) in which significant changes occur in the
community and through which major policy shifts can take place. In this sense, the
connections between politics and disaster in all facets should be more fully explored for a
wide range of possible outcomes both positive and negative (Wolensky and Miller 1983;
Blocker, Rochford, and Sherkat 1991; Belgrad and Nachmias 1997; Olson and Drury 1997;
Drury and Olson 1998; Platt 1999; Olson 2000; Olson and Gawronski 2003; Van Belle,
Rioux, and Potter 2004; Drury, Olson, and Van Belle 2005).
In light of developments in the literature and reaction to specific events, recent work
on the political effects of natural disasters focused on the ways in which natural disasters
might affect pre-existing political economies and security environments within affected
states. Much of the work on political economies has focused on the political economies
associated with the processes of relief and reconstruction. For example, several scholars

50
have noted that the massive influx of foreign aid into the affected state following a natural
disaster often distorts government incentives to engage in disaster prevention spending
(Cohen 2006). In addition, disaster relief aid is often a source of other political distortions
due to the rent-seeking behavior it engenders (Albala-Bertrand 1993b). Working outside the
area of disaster relief aid, other scholars have noted that natural disasters can have a
transformative effect on public-private partnerships (Gendron 1998) and can affect pre-
existing social organizations’ ability to respond through contentious collective action (Simile
1995).
On the question of natural disasters’ effects on the security environments of affected
states, most of the academic literature is ambivalent about whether natural disasters lead to
greater conflict, greater peace, or simply have no effect at all (Quarantelli and Dynes 1976).
For example, some recent literature focuses on the fact that natural disasters may be more
likely to engender conflict than create the space for conflict resolution. Albala-Bertrand
(1993b: 105) argues that a highly positive correlation exists between the number of primary
victims in a disaster and the amount of political unrest that follows. Drury and Olson (1998)
test this statement and demonstrate quantitatively that the severity of natural disasters, as
measured by the number of fatalities, has a positive correlation to political and social unrest
following the disaster. Recent work shows that post-disaster political unrest may even
translate to civil war onset (Brancati 2007; Nel and Righarts 2008). Shefner (1999) attempts
to further specify the linkage between disaster and political instability by demonstrating how
prior instability specifically affects post-disaster instability and how various parties to the
conflict politicize the disaster in such a way to make it advantageous for their respective
causes.

51
However, other studies suggest that natural disasters can be an impetus for positive
social change. Oliver-Smith argues that the process of disaster rehabilitation following the
1970 earthquake in Peru led to important social changes in the town of Yungay, which had
been destroyed in the quake (Oliver-Smith 1977). Blocker, Rochford, and Sherkat argue that
natural disasters often result in community solidarity, which is a necessary ingredient for
facilitating social movement (Blocker, Rochford, and Sherkat 1991). Along these same lines,
Tavera-Fenollosa demonstrates how the series of earthquakes that rocked Mexico City in
1985 sparked the type of social mobilization, in the form of the Movimiento de Damnificados
(the Earthquake Victims’ Movement), necessary to stimulate an important series of
democratic reforms within the country (Tavera-Fenollosa 1999).

Disasters as an Exogenous Shock—Creating a Natural Experiment
Part of the analytical usefulness of natural disasters comes from their ability to act as
a relatively exogenous shock, which can more clearly illuminate the relationships between
variables. This type of analysis is common in economics per se and in economic approaches
to the study of politics. For example, the existing literature in international political
economy provides myriad examples of how various shocks—with admittedly varying levels
of exogeneity—affect interest group formation and function. The early work of Mancur
Olson is seminal in this regard, arguing that business cycle shocks, particularly unexpected
inflation or deflation, shake up the sclerotic effects of interest groups and are therefore
bound to have large effects on both political and economic performance (Olson 1982). In
addition, changes in trade (e.g., Rogowski 1989) and finance patterns (e.g., Frieden 1991)
have been shown to affect domestic interest group formulation and thus have important

52
consequences for domestic governments. When these changes occur rapidly, the resulting
consequences for political stability and security are likely to be particularly acute.
The study of crises as particular types of shocks has been prevalent in other aspects
of international relations research as well. The idea that crises often better illuminate the
mechanisms at work with respect to the relationship between variables has been part of the
core of the International Crisis Behavior (ICB) Project (Brecher and Wilkenfeld 1982). A
number of studies focus on economic crises. For example, the Asian (and subsequently
global) Financial Crisis of the late 1990s spawned a large literature exploring the crisis’
effects on a number of important variables such as state revenue-generating capacity as well
as state capacity in other important functions.
18

To the extent that the shock is more exogenous than not, it allows us to study the
phenomenon in question within the context of a natural experiment. In a pure randomized
controlled experimental design, the responses of subjects exposed to a treatment are
compared to the responses of subjects exposed to the absence of a treatment, the control
group. The hallmarks of this design are that subjects are assigned to the control and the
treatment groups randomly and the treatment is manipulated by the researcher (Babbie
1995). In the social sciences, however, the manipulation of the treatment variable is not
typically something the researcher can control. Nonetheless, a type of “natural experiment”
is possible where the researcher can make valid causal inferences if the exposure of subjects
to treatment and control conditions is determined to have occurred “as if random”
(Dunning 2008a: 283).

18
While this literature is too large to cite completely here, see Wade (1998) and Noble and
Ravenhill (2000) as important examples.

53
The credibility of the claim that the treatment occurred “as if random” rests on a
number of key elements. The first is the question of subject self-selection. If it can be
argued that the subjects “plausibly self-select into treatment and control groups, in ways that
are unobserved or unmeasured by the analyst but that are correlated with the outcome of
interest,” then the researcher’s inferences made on the basis of a natural experimental setting
must be questioned (Dunning 2008a: 290).
For example, recent work that explores the relationship between income levels and
other variables such as political attitudes uses lottery winnings to create a natural experiment
(Doherty, Green, and Gerber 2006). Lottery winnings, in this case, becomes a randomizing
variable with a known probability distribution that allow the researchers to control for any
unmeasured factors that might affect both political attitudes and income such as level of
education. These unmeasured factors should be statistically independent of lottery winnings.
However, if these studies had surveyed and made inferences about all those that played the
lottery—whether they won or not—the quality of the natural experiment would be
diminished. Those who play the lottery are a self-selected group. Using lottery winners
increases the validity of the “as if random” claim and makes for a stronger natural
experiment (Dunning 2008a).
The second important question is whether policy makers or other political actors
have intervened in ways that are correlated with the potential behavioral responses of the
citizens (Dunning 2008a: 290). For example, Brady and McNulty (2005) use changes in
election districting as a natural experiment to study the relationship between voter turnout
and the cost of voting. Here, the validity of the “as if random” nature of the changes in
voting districts on which the experiment rests was thrown into question as further

54
explorations revealed that local politicians may have altered particular districts in ways that
were correlated with voter turnout (Brady and McNulty 2005: 15-16). In these situations, it
is important to establish the level of pre-treatment difference between the experimental and
control groups in order to establish validity.
The third important consideration in establishing the validity of the “as if random”
assignment of the treatment in natural experiments is the extent to which other variables
explain the differences in the average outcomes across groups. Again, at the very least, the
treatment and control groups need to be balanced in this respect prior to the application of
the treatment in order to control for these other possible explanatory variables (Dunning
2008a: 290-91). However, even in the presence of relative empirical balance, these types of
concerns about unobserved differences are always present in observational studies.
One way of satisfying the “as if” random criterion of a natural experiment is to find a
relatively exogenous and randomly distributed treatment. Natural disasters and other
weather-related phenomena provide relatively important contexts in this regard. In a recent
study, Brancati (2007) argues that earthquakes—in particular, because of their rapid-onset
and random distribution with respect to conflict—create a natural experimental setting in
which one can observe the ways in which resource scarcities affect intrastate conflict.
Often times, even absent a specific effort to achieve a natural experimental setting,
instrumental variables are used to proxy for variables that have endogeneity problems with
the outcome(s) of interest. Natural disasters and other weather-related phenomena have
been utilized in this way as well. For example, the endogeneity between income and conflict
makes the study of their relationship difficult, especially as there are few good instrumental
variable proxies for income or gross domestic product (GDP). A number of studies have

55
attempted to overcome this difficulty using average rainfall as a proxy for income. The
results have been used to explore the relationship between income and a number of levels of
intrastate violence including murder (Miguel 2005) through civil war (Miguel, Satyanath, and
Sergenti 2004).
19
Older studies have used rainfall as an instrumental variable for income in
order to study the relationship between transitory income and savings (Paxson 1992).
This project argues that rapid-onset natural disasters provide a relatively exogenous,
but certainly not arbitrary, entry point for thinking about how rapid changes to political
economies affect intrastate conflict. Olson (2000: 268) distinguishes between disasters
which are rapid-onset (e.g., earthquakes, hurricanes, typhoons, floods, etc.) and those which
are slow-onset (e.g., drought, desertification, famine, etc.). Several scholars have pointed to
the problems associated with generalizing across the two different disaster types: inherent
differences in severity, differences in both physical and sectoral impacts, and differences in
the ability of affected states to respond and recover (Vogel 1996). In addition, several types
of disaster that are often considered “natural”—drought, famine, and desertification for
example—are often created, hastened, or worsened by the particular social, political, and/or
economic environments in which they occur (Wisner et al. 2004). This presents a potential

19
Miguel, Satyanath, and Sergenti demonstrate that there is a positive correlation in sub-
Saharan Africa between the percentage change in rainfall from one year to the next and
economic growth. In addition, they argue that average rainfall as an instrumental variable
plausibly satisfies the exclusion restriction—that is, that average rainfall should only affect
violent conflict through economic growth (Miguel, Satyanath, and Sergenti 2004: 735, 45-
46). Using average rainfall as an instrumental variable, their analysis, then, demonstrates a
relationship between economic growth shocks and the likelihood of a civil war, finding “a
five-percentage-point negative growth shock increases the likelihood of a civil war the
following year by nearly one-half” (746). However, despite the fact that their instrumental
variable satisfies the exclusion restriction, the Miguel, Satyanath, and Sergenti model implies
a constant relationship across all components of economic growth and civil conflict which
may not accurately depict the complexity of the relationship (Dunning 2008b: 297).

56
endogeneity problem. Thus, this study will attempt to mitigate some of these inherent
differences between disaster types by focusing solely on rapid-onset disasters such as
earthquakes (and earthquake-generated tsunamis) and wind-related events (cyclones,
hurricanes, typhoons).
Given this subset of natural disasters, a number of the criteria of the “as if random”
assumption that is part of the foundation of natural experiments are fulfilled. While
disasters such as earthquakes do occur in some geographic regions of the world more than
others, there is not likely to be a self-selection problem that may correlate with this study’s
variable of interest. For example, Brancati (2007: 718-19) points out that the places where
earthquakes occur around the world are not necessarily more conflict prone than others and
that earthquakes occur on average as often in countries belonging to the Organization for
Economic Co-Operation and Development (OECD) as they do in non-OECD countries.

57
Figure 1: Geographic Location of Earthquakes
Magnitude 5.5 or Greater (1975-2002)
20

Because of these factors, it should also be possible to achieve relative balance in
regards policy interventions and controlling for alternative explanations, which are the two
other concerns of the natural experimental design. In all, one of the foundational premises
of this project is that the shock of the natural disaster presents an analytical window through
which one might be able to more clearly see the relationships between various institutional
political economies and conflict. However, there are also a number of other important
benefits that flow from a better understanding of the relationship between natural disasters
and conflict and violence.

20
Map is reprinted from Brancati (2007: 719) and based on data from Engdahl and
Villaseñor (2002).

58
Methodology and Case Selection
This project utilizes a multi-method research design, proceeding from a large-n
quantitative analysis of the third hypothesis stated in chapter 2 toward a qualitative
comparison of two cases where the aim will be to specifically explore the first two
hypotheses developed in chapter 2. This section develops a justification of each piece of the
research design. Much of the discussion of specific tools (e.g., specific methods of
quantitative analysis, robustness checks, etc.) will be left to the individual chapters in which
the analysis takes place.

Large-n Empirical Analysis
The quantitative analysis in chapter 4 will explore the relationship between
institutional quality at the time of a natural disaster and the incidence of post-disaster
violence. As a test of H3, I utilize a large-n pooled cross-section research design. Here, the
country-year is the unit of analysis and the sample size consists of 2,573 observations on 188
political units and encompasses 16 years between 1984 and 1999. The dependent variable is
a dichotomous coding of violent civil conflict, which the Uppsala/PRIO data set labels
“domestic armed conflict.” Given the dichotomous nature of this variable, the analysis
utilized will be a standard logistic regression. The primary explanatory variable will be an
interaction term that incorporates the incidence of natural disaster with a measure of pre-
disaster institutional quality.
21

21
Further information on the employed variables and regression analysis is provided in
chapter 4.

59
Case Selection for Qualitative Comparison
While the large-n quantitative comparison will allow us to develop the empirical
connection between institutional quality and violent conflict in a post-disaster context, it will
not provide us with a specific test of the first two hypotheses developed in chapter 2. In
addition, a cursory qualitative exploration reveals that the cases of Sri Lanka and Indonesia
following the 2004 tsunami present an interesting set of challenges to the quantitative
findings in chapter 4. In order to explore the first two hypotheses in chapter 2 and rectify
the challenges these cases present to the quantitative findings of chapter 4, the causal logic is
best traced out through comparative case studies (George and Bennett 2005; Gerring 2007;
Seawright and Gerring 2008).
Thus, in chapter 5, I attempt to evaluate the hypotheses established in chapter 2
through a systematic structured, focused comparison of violent conflict in Indonesia and Sri
Lanka following the Indian Ocean earthquake and resulting tsunami in December 2004. The
cases of Indonesia and Sri Lanka represent puzzling cases, as similar pre-disaster civil
conflicts seemingly moved in opposite directions in the aftermath of the same disaster. By
August 2005, some eight months after the tsunami that killed almost 130,000 Indonesians,
the Indonesian government signed a tentative peace agreement in Helsinki with the GAM.
In Sri Lanka, however, the story was quite the opposite, as eighteen months after the
tsunami, which killed over 30,000 Sri Lankan citizens, the majority Sinhalese-led Sri Lankan
government and the LTTE renewed their civil war. Any proposed theory of the relationship
between natural disasters and violence must be able to reconcile these two variant outcomes.
In addition, the cases of Indonesia and Sri Lanka make for a systematic comparison
because their similarities allow one to exert relative control over a number of competing

60
explanations. First, the length of the preexisting conflict with their respective governments
was approximately two decades in both cases. In neither of the cases was the separatist
movement new. Second, both of the movements were separatist movements seeking
autonomous control over a particular portion of existing state territory. Third, prior to the
tsunami, both countries were operating amidst quickly crumbling ceasefires with their
respective separatist movements. In fact, in both cases, in the months leading up to the
tsunami, the tensions between the state governments and the separatist movements seemed
to be increasing. Finally, both countries experienced the same natural disaster, and it had
drastic effects on the separatist communities in question. Both the Aceh region of Indonesia
and the Tamil region of Sri Lanka bore significant percentages of their country’s overall
tsunami-related casualties.
Given these similarities and their divergent outcomes, these cases present an
interesting opportunity to explore the previously established hypotheses. Thus, chapter 5
utilizes Alexander George’s method of “structured, focused comparison” (George 1979;
Haney 1997; George and Bennett 2005: 69-72), which explores the cases by “asking a set of
standardized, general questions of each case” (George and Bennett 2005: 69).

Conclusion
The relationship between natural disasters and violent civil conflict serves two related
purposes in this project. First, it has an analytical utility in that it allows us to more clearly
witness particular strategic interactions that take place even absent a natural disaster: ongoing
battles between states and rebels to control political and economic institutions. Second, the
relationship between disaster and conflict is interesting in its own right, as the historical and

61
possible future coincidence of these two phenomena make the relationship an important
policy concern, particularly as countries tread carefully in attempting to provide disaster
relief assistance without being drawn directly into preexisting or budding conflicts. The next
chapter attempts to explore the relationship between institutional quality and post-disaster
violence more clearly on a large cross-national scale.

62
Chapter 4: Testing the Relationship between Institutional Quality
and Post-Disaster Violence

The hypotheses set forth in chapter 2 suggest marginal gains in control over disaster
relief institutions have important consequences for violent civil conflict inside an affected
country. Specifically, an increase in rebel control over institutions will lead to greater conflict
in a country, as the material and nonmaterial gains from institutional control will positively
affect rebels’ relative power. A similar increase or maintenance of control over institutions
by the state should lead to lower levels of violent civil conflict within a society, again due to
changes in the relative power of the groups given the material and nonmaterial benefits of
institutional control.
Given the difficulties inherent in quantitatively testing the first two hypotheses, a
more valid exploration of the causal processes will be left to the qualitative comparison in
chapter 5. However, hypothesis 3 lends itself to a large-n, quantitative empirical test.
Namely, if the battle over institutional control leads to changes in political violence amidst a
disaster, we might expect to see violence indirectly affected by the quality of institutions.
States with higher levels of institutional quality should be better able to respond to natural
disasters, making the opportunities for institutional capture by the rebels small. Thus, when
controlling for the context of natural disasters, one would expect to see a lower likelihood of
the onset of violent civil conflict when there is a higher level of institutional quality and vice
versa.
The chapter continues with the discussion of the variables employed in the analysis.
The set of variables is chosen both because of their previously established relationship with
violent civil conflict as well as their empirical utility for exploring the implications and

63
robustness of the arguments at hand. A standard logistic estimation technique is employed.
The chapter concludes with an analysis and discussion of the results and their implications
for the forthcoming case studies.

Empirical Design
As a preliminary exercise in testing the above hypothesis, this paper uses a large-n
pooled cross-section research design. Here, the country-year is the unit of analysis and the
sample size consists of 2,573 observations on 188 political units and encompasses 16 years
between 1984 and 1999.
22
This section introduces the variables used as well as the logistic
regression estimation technique employed.

Dependent Variable
In keeping with recent empirical work on the relationship between natural disasters
and conflict (Nel and Righarts 2008: 615), the dependent variable is violent civil conflict, which
the Uppsala/PRIO data set labels “domestic armed conflict.” The Uppsala/PRIO dataset
on the onset of intrastate conflict contains observations of all members of the international
system, defined by Gleditsch and Ward (1999), and is itself derived from the larger Uppsala
Conflict Data Project. This project defines armed conflict as any “contested incompatibility
that concerns government or territory or both, where the use of armed force between two
parties results in at least 25 battle-related deaths in a year” and where at least one of the two

22
The relatively small time period, and thus sample size, of this study is governed by the
International Country Risk Group (ICRG) data, which contains measurements of
institutional quality and covers the years indicated.

64
parties involved is the government of a state (Gleditsch et al. 2002; Harbom and Wallensteen
2007: 632).
23

Incidence of a violent conflict is coded dichotomously, with an observation recorded
for any conflict-year that produces a minimum of 25 conflict-related deaths. As such, there
are 501 observations of the incidence of violent civil conflict between the years 1984-1999.
Within these cases, major civil conflicts are also distinguished. A major conflict is coded
when a conflict crosses the threshold of 1,000 conflict-related deaths. Of the 501 conflicts
between 1984 and 1999, 394 are coded as major conflicts. In order to account for changes
in the incidence of violent civil conflict that occur after a disaster, I make use of a t+1
version of the variable.
One of the main criticisms of the Uppsala/PRIO data as well as other sets that
contain cross-national data on conflict and violence is that they lack any sub-national
specificity (Miguel, Satyanath, and Sergenti 2004). These limitations do restrict the empirical
claims that research can actually make about the relationship between natural disasters and
violent conflict, as the effects of most disasters are contained to specific geographic locations
and discrete time periods (Nel and Righarts 2008: 167).
24

Nonetheless, the Uppsala/PRIO data set remains a preferred measure because of its
lower threshold. The 25-battle-related-deaths threshold captures more instances of violent

23
While Gleditsch et al. (2002) is the original source of the data, there have been a number
of revisions to the data set since that time. Harbom and Wallensteen (2007) provide the
most recent update to the data through 2006 as well as add four alternative onset variables
based on Strand’s definitions (Strand 2006; Buhaug 2007).
24
See chapter 6 for a brief discussion of a new dataset solves many of these problems by
coding instances of intrastate conflict at specific sub-state, geographic points (e.g., Buhaug
and Gates 2002; Buhaug and Rød 2006).

65
civil conflict than other databases that are focused on the study of civil war more specifically
and use the minimum 1,000 battle-related deaths threshold (e.g., Sarkees 2000; Fearon and
Laitin 2003). This is consistent with the argument previously advanced that the study of the
broader category of intrastate violence might be preferable to the more narrowly defined
subcategory of civil war so as to avoid arbitrary definitional, theoretical, and empirical
boundaries (Tilly 2003; Sambanis 2004).

Explanatory Variables
The data on natural disasters are derived from EM-DAT: The OFDA/CRED
International Disaster Database (Office of U.S. Foreign Disaster Assistance (OFDA) Centre
for Research on the Epidemiology of Disasters (CRED) 2008). Disasters are defined in EM-
DAT as “a situation or event which overwhelms local capacity, necessitating a request to the
national or international level for external assistance, or is recognized as such by a
multilateral agency or by at least two sources, such as national, regional or international
assistance groups and the media” (Guha-Sapir, Hargitt, and Hoyois 2004: 16). As such, in
the CRED database, events are classified as disasters when at least one of the following
criteria is met: 1) ten or more people are reported as killed by the event, 2) one hundred or
more people are reported as being affected by the event, 3) the event leads to the declaration
of a state of emergency, and 4) the event leads to requests for international assistance (Guha-
Sapir, Hargitt, and Hoyois 2004: 16).
While the EM-DAT database is generally accepted to be the most comprehensive
collection of global data on natural disasters, a number of criticisms and limitations are
noteworthy. Collection of the data by EM-DAT is done through a culling of public sources

66
that in their original forms were not primarily intended for statistical use (Guha-Sapir,
Hargitt, and Hoyois 2004: 15).
25
Given these multiple sources, the data are prone to
definitional inconsistencies. Since every natural disaster involves a “physical energy
expenditure” that is followed by a “human reaction” (Alexander 1997: 289) it is often
difficult to distinguish whether measurements such as deaths, value of destruction, and
overall social impacts are a result of the physical event or the human response (Tschoegl,
Below, and Guha-Sapir 2006). These definitional issues are not unique to the EM-DAT
data; they plague all available disaster statistics (Albala-Bertrand 1993b: 39; Quarantelli 2001;
Stallings 2003; Tschoegl, Below, and Guha-Sapir 2006). Nonetheless, the EM-DAT
database remains one of the most comprehensive and most transparent about its
methodology (Tschoegl, Below, and Guha-Sapir 2006: 7). Thus, it is the most commonly
employed database.
For the period of analysis (1984-1999), there are 4,163 instances of natural disasters
listed in the EM-DAT database. Table 2 shows the breakdown of disasters by type and lists
the definitions for subcategory. Natural disasters are classified by CRED as having five main
subtypes: geophysical, meteorological, hydrological, climatological, biological, and extra-
terrestrial. I use the data created by Nel and Righarts (2008: 167) which sums three
categories—hydro-meteorological, geological, and other—into a variable labeled all natural
disasters. This variable is then limited to rapid onset natural disasters in order to exclude
famine, drought, desertification and other events with lower levels of exogeneity. In the

25
As of 2003, approximately 27.9% of the EM-DAT data came from US government
disaster agencies, 27% came from insurance companies, 18.1% from various press agencies,
and the remaining 7% from a variety of humanitarian organizations (Guha-Sapir, Hargitt,
and Hoyois 2004: 15).

67
dataset, the variable takes the value of a 1 in a country year in which a rapid-onset natural
disaster occurs and a 0 where one does not occur.

68
Table 2: Global Natural Disasters by Type, 1984-1999
26

Type of Disaster Events Deaths Total affected
Drought
27
213 6,610 697,890,390
Earthquake (seismic activity)
28
379 147,536 54,303,793
Epidemic
29
479 127,284 14,943,590
Extreme temperature
30
120 12,364 7,171,986
Flood
31
1,212 124,593 1,780,188,069
Insect infestation
32
59 0 2,200
Mass movement dry
33
26 1,561 2,152
Mass movement wet
34
209 12,073 4,756,554
Storm
35
1,255 246,295 331,349,489
Volcano
36
71 24,564 2,213,996

26
The data in this table are from the EM-DAT database (Office of U.S. Foreign Disaster
Assistance (OFDA) Centre for Research on the Epidemiology of Disasters (CRED) 2008).
Definitions are from EM-DAT database and from Scheuren et al. (2008).
27
An extended period of time characterized by a deficiency in a region’s water supply
resulting from constantly below average precipitation (OFDA-CRED 2008)
28
The shaking and displacement of the ground due to seismic waves; this classification does
not include secondary effects (OFDA-CRED 2008)
29
Either an unusual increase in the number of cases of an already existing infectious disease
or the appearance of a previously absent infection (OFDA-CRED 2008)
30
Heat wave is a prolonged period of excessively hot/humid weather relative to normal
climate patterns. A cold wave can be either a prolonged period of excessively cold weather
or the sudden invasion of very cold air over a large area (OFDA-CRED 2008).
31
The significant rise of water level in a stream, lake, reservoir or coast (OFDA-CRED
2008)
32
The pervasive influx and development of insects or parasites affecting humans, animals,
crops and materials (OFDA-CRED 2008)
33
This broad category includes the subcategories of subsidence (“motion of Earth’s surface
as it shifts downward relative to sea level”), rock fall (“rock or stone falling freely from a cliff
face”), avalanche (“snow or ice that slides down a mountainside under the force of gravity”),
and landslide (“movement of soil or rock controlled by gravity”) (OFDA-CRED 2008)
34
This includes the same subcategories as mass movement dry, but events here are triggered
and categorized by hydrological phenomenon rather than geophysical (OFDA-CRED 2008).
35
Storms are events with wind with speed between 48 and 55 knots (OFDA-CRED 2008).

69
Table 2, Continued
Wildfire
37
140 1,139 3,618,032
Total 4,163 70,4019 2,896,459,351

Following recent work (Knack and Keefer 1995; Chong and Zanforlin 2000;
Acemoglu, Johnson, and Robinson 2001; Easterly 2001; Knack 2001; Bräutigam and Knack
2004; Dietz, Neumayer, and De Soysa 2007; Taydas, Peksen, and James 2008), the variables
that attempt to measure institutional quality are taken from the International Country Risk
Group (ICRG) database.
38
The first variable, corruption in government, is a measure of the
extent to which bribes affect government decision-making at all levels. These bribes may
take the form of special payments, bribes connected to import and export licenses, exchange
controls, policy protection, and/or special tax assessments. The value of the variable is 0-6
with higher numbers indicating lower corruption.
The second measure of institutional quality is bureaucratic quality. This variable
attempts to capture state effectiveness. ICRG measures this variable on a scale of 0-4,
39
and

36
This category includes all volcanic activity such as rock fall, ash fall, lava streams, gases,
etc. (OFDA-CRED 2008).
37
An uncontrolled burning fire, usually in wild lands (OFDA-CRED 2008)
38
While there are other datasets that measure various aspects of institutional quality, they
have a variety of problems associated with them. Taydas, Peksen, and James (2008: 15, fn 7)
point out that Transparency International’s corruption perception index, for example, ranks
more than 150 countries in terms of expert assessments and opinion surveys on perceived
levels of corruption. But this indicator only measures corruption, a single aspect of
institutional quality. While the World Bank’s data has a number of other indicators (i.e.,
voice and accountability, political stability and absence of violence, government
effectiveness, regulatory quality, rule of law, and control of corruption) its time span is small
(1996-2004) and is limited to recent years.
39
Following Taydas, Peksen, and James (2008), this variable has been rescaled in order to
match the other two ICRG measures of institutional quality.

70
higher scores indicate that the state bureaucracy is relatively immune from political pressure.
Higher scores are also indicative of a state’s higher ability to govern without major
disruptions in government services, especially during changes in government leadership.
Lower values indicate an inferior bureaucratic system that lacks the training and/or expertise
to carry out the basic duties of governance without interruptions and/or influence.
The third ICRG component of institutional quality used here is a variable that
measures rule of law and order. With this variable, the ICRG measures law and order
separately, combining them into a 0-6 scale. A score close to 6 indicates a country with a
relatively strong and impartial legal system as well as strong institutions that provide orderly
transitions of political power and well-respected courts for settling legal claims. Scores
closer to 0 indicate that the country has a history of relying on physical violence to settle
claims rather than on courts.
40

Following the lead of others that have utilized these measures of institutional quality
(Easterly 2001; Bräutigam and Knack 2004; Huther and Shah 2005; Taydas, Peksen, and
James 2008), I also utilize an overall institutional quality index by adding together the three
component measures. Thus, the scale for this variable is 0-18. States with scores closer to
18 indicate a higher quality of institutions, and states with scores closer to 0 indicate a lower
quality of institutions.
41

40
In order to control for the high degree of correlation between these three separate
measures of institutional quality (r > 0.63 in all cases), I will introduce them into the logistic
regression one at a time.
41
Each of the separate measures of institutional quality is highly correlated with the overall
index of institutional quality (r > 0.82 in all three cases). As such, the index will also be
introduced separately.

71
Finally, in order to test the specific hypothesis in question, an interaction term is
necessary. Thus, I create a new variable, interacting each of the measures of institutional
quality with the all natural disasters variable. The IQ index variable is also interacted with
disasters. Each of these interaction terms is entered into the regression separately from the
others, in the same fashion of each of the individual measures of institutional quality. This
interaction term takes the value of the ICRG measure in a country year where a natural
disaster exists (the ICRG measure x 1) and a zero in a country year where a natural disaster
does not exist (essentially, the ICRG measure x 0). This variable controls for the fact that
this project seeks to test the hypothesis that changes institutional quality should impact the
likelihood of violent civil conflict in the context of a natural disaster.

Control Variables
A country’s level of economic development has been widely established in the
quantitative study of civil wars to be one of the most important determinants of probability
of the onset of domestic armed conflict (e.g., Collier and Hoeffler 1998; Sambanis 2002;
Collier et al. 2003; Collier, Hoeffler, and Sambanis 2005; Collier and Sambanis 2005a).
There are several theories about why this is the case, ranging from wealthier states being
more easily able to suppress opposition (De Soysa 2002), to wealthier states being able to
more easily combat grievances through a more extensive provision of public goods
(Henderson and Singer 2000), to wealth’s effect on female fertility rates and thus
demographic shifts (Urdal 2006, 2007), and finally higher levels of economic development
raising the opportunity costs associated with rebel recruitment (e.g., Weinstein 2007). GDP
per capita is widely accepted as a proxy for a variety of development indicators in the study of

72
civil wars (Hegre and Sambanis 2006). Following others, I use the natural log of GDP per
capita, adjusted for purchasing power parity (Heston, Summers, and Aten 2008). I also
control for economic growth using Urdal’s (2006) calculation of the 5-year average of the Penn
World Tables measure of the growth in per capita real GDP adjusted again for purchasing
power parity.
Theoretically, income inequality should be a particular economic variable that has a
positive relationship to the onset of violent civil conflict, as one would expect grievances to
increase as a result of increasing disparities. However, there are a number of difficulties
associated with existing direct measures of income inequality.
42
Thus, GDP per capita is
often used as proxy, given its close correlation with income inequality. However, several
scholars have made the point that the linear relationship between GDP per capita and
violent civil conflict does not reflect the inverted U-shape relationship between violent civil
conflict and income inequality has been previously demonstrated; thus, other proxies for
income inequality are in order (Nagel 1974; Lichbach 1989; Nel 2006; Nel and Righarts
2008).
Sen (1998) argues that infant mortality is a good proxy for a country’s economic
development more generally as it depends on factors such as level of education, living
standards, and the quality of the health care system (Anand and Ravallion 1993). In
addition, there appears to be a strong correlation between infant mortality rates and existing
but flawed direct measures of income inequality (Nel and Righarts 2008: 169). I follow

42
In his assessment of the empirical work on the relationship between economic growth and
income inequality, Knowles (2005: 143-44) points out that the many ways income inequality
has been measured are highly inconsistent—with differing sample sizes and often focusing
on only highly developed countries. This makes comparability across findings difficult.

73
Urdal (2006) and utilize his data on infant mortality rate constructed from the UN’s World
Population Prospects and the UN’s Demographic Yearbook. This variable is measured as each
country’s fraction of children, born-live, that die before age one.
43

Regime type has been pegged as an important predictor of the risk of violent civil
conflict (e.g., Gates et al. 2006). Two important findings are noteworthy. First, the
relationship between regime type and violent civil conflict takes the shape of an inverted U,
indicating that both extreme autocracies and significantly consolidated democracies
experience very little violent civil conflict. Most of the violence seems to occur in regimes
with Polity scores somewhere in the middle (Hegre et al. 2001; Urdal 2006). Second, a
number of solid theoretical reasons have been advanced as to why these middle, or mixed
regimes, should be more likely to experience violent civil conflict. Again, these run the
gamut of grievance- and opportunity-focused explanations. For instance, despite the fact
that mixed democracies may contain some institutions for voicing grievances, the
government may be so dominated by an elite few that it is not able to respond effectively
(Merkel 2005; Hegre and Sambanis 2006). Taking into account both of these advancements,
I follow the lead of Nel and Righarts and utilize their mixed regime variable, a dummy variable
designed to capture inconsistent democracies and autocracies on the same Polity scale (2008:
169).

43
While the data on violent civil conflict account for consecutive years, the data on natural
disasters do not. Thus, in order to escape the endogeneity problem with infant mortality
rates, I utilize a lagged version of the variable. In addition, following Nel & Righarts (2008:
169), I employ a squared version of the infant mortality rate variable in order to account for
the fact that economic inequality may intervene in a nonlinear fashion. Finally, because of
relationship between infant mortality rates and development, the infant mortality rate
variable and the GDP per capita and GDP growth variables are introduced into the
regressions separately in order to avoid potential problems of multicollinearity.

74
Recent work by Henrik Urdal (2006, 2007) finds that developing states whose
populations contain relatively large cohorts of youths are at higher risk for intrastate political
violence. These youth bulges are measured as the percentage of 15-24-year-olds in the total
adult population of a country, those that are 15 years and older.
44
Theoretically, youth bulges
can impact intrastate violence through the mechanisms of motive (grievance) and
opportunity. That is, youths are more likely to become aggrieved due to high
unemployment, “institutional bottlenecks” (Urdal 2006: 610), and urban over-crowding
(Choucri 1974; Braungart 1984; Goldstone 1991, 2001; Cincotta, Engelman, and Anastasion
2003). At the same time, a large pool of unemployed youths is likely to significantly reduce
the costs associated with rebel recruitment, thereby changing opportunity structures of
rebellion (Collier 2000: 94). Theoretically, natural disasters should impact both sets of
conditions (Goldstone 2001), and some empirical evidence supports the claim (Nel and
Righarts 2008). Thus, youth bulges are included here as an appropriate control variable.
45

I include the size of the political unit, measured as the natural log of the total square
kilometers, in order to account for differences in violence civil conflict that might be due to
differences in the size of states. In many studies of civil war, the natural log of the total
population is employed as a control variable in this respect (Fearon and Laitin 2003; Hegre
and Sambanis 2006); thus, it is also included here.

44
Urdal is careful to separate his measurement from others (Huntington 1996; Fearon and
Laitin 2003; Collier and Hoeffler 2004) that measure youth bulges relative to total population
as opposed to total adult population. Countries with high fertility rates – and therefore large
populations under 15 – tend to exhibit deflated youth bulge measures (Urdal 2006: 615).
45
Given the highly positive correlation between infant mortality rate and youth bulge (r > 0.6), I
follow the lead of Nel and Righarts (2008: 170, fn 24) and introduce these variables into the
estimation separately in order to avoid collinearity.

75
Because logistic regression is the basis of this analysis, some attempts must be made
to control for possible time dependency in the variables. In particular, countries that
experience previous conflict have a higher conflict probability than countries that have
experienced no conflict (Raknerud and Hegre 1997). In countries that have experienced
multiple years of conflict, any subsequent years of conflict will appear to be statistically
dependent on the first year of the conflict. The measures of the incidence of violent civil
conflict in the PRIO data do not control for time dependency related to prior conflicts, but
they also do not control for the fact that multiple years of peace also display the same
problem of time dependence (Urdal 2006: 617). In order to control for time dependency, I
employ the variable brevity of peace, which specifies the estimated likelihood of conflict given
the number of years of peace following the previous conflict (Urdal 2006: 617). The variable
assumes that the risk of new conflict is highest immediately following a conflict and that the
risk declines over time (Hegre et al. 2001).
46
Countries very close to the end of a previous
conflict have brevity scores of close to 1; as the time from the end of a conflict grows longer,
the brevity value moves close to zero.
Finally, a number of other independent variables have been included that have been
found to be important predictors of the likelihood of civil war onset. Territorial contiguity and
mountainous terrain are measures used to indicate particular geographic conditions that might
be favorable to rebels or other insurgent groups (Fearon and Laitin 2003). Territorial contiguity

46
This likelihood is measured as exp[( - years in peace)/α] where “years in peace” is the
number of years since a given country has experienced an armed conflict and α indicates the
rate at which the effect of a previous armed conflict on any new armed conflict diminishes as
time passes (Hegre et al. 2001: 37; Urdal 2006: 617). Urdal follows the assumption that the
risk of conflict is halved every three years; therefore, α takes a value of 4 (Toset, Gleditsch,
and Hegre 2000).

76
takes a value of 1 when the territorial base in the country is geographically separated from
the capital city by land or water and a value of 0 otherwise. Mountainous terrain is the logged
share of the country’s geographic area that is covered by mountains. Oil is a dichotomous
variable that captures the argument that certain natural resources are important predictors of
violent civil conflict because, if captured, they can be used to finance rebellion. The variable
is coded 1 if more than one third of the country’s export revenues come from oil exports
and 0 otherwise. Finally, a measure of ethnic fractionalization is included to account for the
positive relationship that has been argued between ethnic diversity and civil war (Ellingsen
2000). The ethnic fractionalization variable is an index ranging from 0, indicative of total ethnic
homogeneity, to 1, indicative of total ethnic heterogeneity (Fearon and Laitin 2003).

Results and Analysis
Figure 2 presents the “Receiver Operator Characteristics” or “ROC” curve. This is a
goodness-of-fit curve for logistic regression that compares in-sample results to out-of-
sample results.

77
Figure 2: Receiver Operator Characteristic (ROC) Plot
47

For this particular ROC curve, the solid line indicates the proportion of 1s and 0s predicted
following natural disasters when measures of institutional quality (here, the ICRG variables)
are omitted. The dotted line indicates the proportion of 1s and 0s predicted following
natural disasters when the disaster*IQ Index interaction term is added. The curve indicates

47
The solid line represents the logit on incidence of violent civil conflict (t+1) in disasters
without ICRG; the dotted line represents the logit on the same with the addition of the
ICRG (IQ Index) interaction variable. Figure corresponds with Table 3, Model 4.

78
slightly higher predictive power in the model that includes the IQ Index interaction variable,
particularly in the model’s ability to predict those instances where violent civil conflict does
not occur in the year after a natural disaster.
Tables 3 and 4 present the results of the logistic regression. The dependent variable
in both tables is the incidence of violent civil conflict in the year immediately following a
rapid onset natural disaster, and each of the models in the tables interacts one of the
institutional quality variables with natural disasters and the final model interacts the IQ index
variable. Table 3 uses GDP per capita adjusted for purchasing power parity in order to
control for economic development. Table 4 uses the measures of infant mortality rate to
proxy for economic development.

79
Table 3: Logistic Regression of Violent Civil Conflict, 1984-1999
48

(violent civil conflict measured at time t+1)
49

Model 1 Model 2 Model 3 Model 4
Rapid natural
disasters x ICRG
-0.095 ***
(0.032)
-0.057 *
(0.032)
-0.036
(0.027)
-0.029 **
(0.012)
Corruption -0.002
(0.082)

Bureaucratic quality -0.058
(0.120)

Law and Order -0.266 ***
(0.082)

IQ index -0.061 *
(0.035)
Rapid natural
disasters
0.230 ***
(0.089)
0.095
(0.074)
0.093
(0.094)
0.188 *
(0.098)
GDP change
50
0.002
(0.017)
0.003
(0.017)
0.009
(0.017)
0.006
(0.017)
GDP per cap PPP
(ln)
-0.043
(0.158)
-0.081
(0.164)
0.086
(0.161)
Brevity (t+1) 4.976 ****
(0.234)
4.897 ****
(0.230)
4.767 ****
(0.229)
4.863 ****
(0.231)
Mixed regime -0.081
(0.194)
-0.058
(0.194)
-0.127
(0.194)
-0.113
(0.194)
Mountains 0.159 **
0.069
0.153 **
(0.068)
0.089
(0.068)
0.113
(0.069)
Oil 0.259
(0.261)
0.280
(0.264)
0.318
(0.261)
0.203
(0.263)
Ethnic
fractionalization
0.319
(0.360)
0.344
(0.361)
0.287
(0.362)
0.339
(0.361)
Contiguity 0.323
(0.299)
0.418
(0.298)
0.505
(0.308)
0.416
(0.304)
Size (ln) -0.194 **
(0.082)
-0.206 **
(0.082)
-0.196 **
(0.083)
-0.186 **
(0.083)
Youth bulge 0.040 *
(0.022)
0.037 *
(0.022)
0.046 **
(0.022)
0.045 **
(0.222)
Population (ln) 0.351 ***
(0.112)
0.374 ***
(0.115)
0.461 ****
(0.113)
0.427 ****
(0.115)

48
Logit regression coefficients with robust standard errors in parentheses: *p<.10, **p<.05,
***p<.01, ****p<.001; estimations performed using Zelig statistical package for R (Imai,
King, and Lau 2007).
49
All violent civil conflict = the incidence of violent conflict with minimum of 25 battle-
related deaths.
50
All time variant independent variables are lagged at t-1.

80
Table 3, Continued
Constant -5.700 ***
(2.016)
-5.293 **
(2.076)
-6.981 ****
(1.937)
-7.148 ****
(2.072)
N 2573 2573 2573 2573

Model 1 in Table 3 indicates that there appears to be a statistically significant relationship
between corruption following natural disasters (Rapid natural disasters x ICRG) and the
incidence of violent civil conflicts. The sign of the coefficient is in the expected direction,
indicating that as levels of corruption decrease
51
in countries that experience natural
disasters, the likelihood of violent conflict decreases. The finding is significant at the 0.01
level, and appears to be robust across alternative means of controlling for economic
development. The results of Table 4, Model 1 demonstrate the same relationship between
corruption and violent civil conflicts in the year after natural disasters, using infant mortality
rates to proxy for economic growth.

51
Again, higher values of this particular ICRG measure indicate lower levels of corruption.

81
Table 4: Logistic Regression of Violent Civil Conflict, 1984-1999
52

(violent civil conflict measured at time t+1; infant mortality rate)
53

Model 5 Model 6 Model 7 Model 8
Rapid natural
disasters x ICRG
-0.094 ***
(0.032)
-0.062 *
(0.032)
-0.044 *
(0.027)
-0.030 ***
(0.012)
Corruption 0.055
(0.077)

Bureaucratic quality 0.018
(0.108)

Law and Order -0.221 ***
(0.079)

IQ index -0.029
(0.032)
Rapid natural
disasters
0.241 ***
(0.088)
0.114
(0.074)
0.136
(0.093)
0.223 **
(0.098)
GDP change 0.007
(0.017)
0.008
(0.017)
0.019
(0.017)
0.014
(0.017)
Brevity (t+1) 4.920 ****
(0.234)
4.860 ****
(0.232)
4.720 ****
(0.230)
4.820 ****
(0.231)
Mixed regime 0.109
(0.200)
0.130
(0.200)
0.077
(0.199)
0.090
(0.200)
Mountains 0.205 ***
(0.070)
0.197 ***
(0.069)
0.134 *
(0.069)
0.167 **
(0.070)
Oil 0.466 *
(0.257)
0.467 *
(0.259)
0.559 **
(0.261)
0.445 *
(0.260)
Ethnic
fractionalization
0.201
(0.360)
0.181
(0.359)
0.152
(0.362)
0.174
(0.360)
Contiguity 0.293
(0.296)
0.387
(0.299)
0.396
(0.304)
0.343
(0.301)
Size (ln) -0.218 **
(0.085)
-0.230 ***
(0.084)
-0.198 **
(0.086)
-0.199 **
(0.085)
Infant morality rate 0.001
(0.008)
0.003
(0.008)
-0.006
(0.009)
-0.003
(0.009)
Infant mortality rate
(sq)
0.000
(0.000)
0.000
(0.000)
0.000 *
(0.000)
0.000
(0.000)
Population (ln) 0.363 ***
(0.112)
0.389 ****
(0.112)
0.454 ****
(0.112)
0.411 ****
(0.112)
Constant -5.320 ****
(0.829)
-5.360 ****
(0.833)
-5.220 ****
(0.820)
-5.370 ****
(0.828)
N 2573 2573 2573 2573

52
Logit regression coefficients with robust standard errors in parentheses: *p<.10, **p<.05,
***p<.01, ****p<.001; estimations performed using Zelig for R (Imai, King, and Lau 2007).
53
All violent civil conflict = the incidence of violent conflict with minimum of 25 battle-
related deaths. All time variant independent variables are lagged at t-1.

82
The other models in Table 3 and 4 also indicate a relationship between the quality of
institutions and violent civil conflict amidst natural disasters. Each of the other individual
measures of institutional quality demonstrates the expected sign, and most are statistically
significant. The variable for bureaucratic quality, for example, indicates that countries with
higher levels of bureaucratic quality—that is, those that are better able to provide goods and
services without major interruptions—are less likely to experience the incidence of violent
civil conflict in the aftermath of natural disasters. This conforms to the theoretical
expectations discussed earlier.
Finally, the findings regarding the relationship between the overall index of institutional
quality and the incidence of violent civil conflict appears to confirm the theoretical
expectations. Given the expectation of the individual component variables, one would
expect that countries with a higher level of institutional quality would have lower likelihoods
of violent civil conflict. In both models, utilizing different controls for economic
development, the sign of the IQ index variable is in the expected direction, indicating this
negative relationship. In addition, the finding is statistically significant at the 0.05 and 0.01
levels respectively across various specifications of economic development.
A number of the control variables conform to the findings of previous research. For
example, there is a strongly significant relationship between the length of time between prior
conflicts and the onset of new violent civil conflict. As one might expect and as others have
demonstrated previously, there is a higher likelihood of the onset of new conflict if the time
since the end of the previous conflict is short. Across both specifications of economic
development and each type of measure of institutional quality, the brevity variable displayed
the expected sign and in most cases is statistically significant at the 0.001 level. In addition,

83
the size of a country also follows earlier findings that larger countries have a greater
likelihood of the onset of violent civil conflict; this relationship was significant at the 0.05
level and was a robust finding across the models.
There were some variables that displayed some surprising findings. Although the
proxies for economic development displayed the anticipated sign in all cases, the majority
was not statistically significant. This, of course, goes against some very well-established
empirical findings that economic development is one of the strongest predictors of the onset
of violent civil conflict (e.g., Elbadawi and Sambanis 2000, 2002; Sambanis 2002; Collier et
al. 2003; Collier and Sambanis 2005a). Among other possibilities, these differences in
findings could be due to the limitations of the comparatively small sample size given the
available ICRG data.
54

However, the substantive effects of logistic regression are difficult to determine
merely by looking at the significance and signs of the coefficients (King 1998: 102-08) and
this is even more difficult when the regression includes an interaction term (Braumoeller
2004; Brambor, Clark, and Golder 2006). Calculating the first differences—how much the
likelihood of a particular outcome would change given a particular change in a single
explanatory variable—becomes important for interpreting the substantive results of the
regression. Figures 3, 4, and 5 present the results of these calculations on the models that
include the IQ Index interaction term.

54
Future work will attempt several more robustness checks on these findings across a
number of types of natural disasters, using a variety of economic development indicators,
and a number of different measures of institutional quality. In addition, newly acquired
ICRG data will allow me to expand the dataset to include 1984-2006.

84
Figure 3 demonstrates that when all of the independent variables are held at their
means (or medians for the dichotomous variables), and the predicted probabilities are
generated based on 1000 simulations of the logistic regression, the probability of y=1 is
approximately 9.5%. More substantively, the probability of the incidence of violent civil
conflict in the year following a natural disaster is about 9.5% with a 95% confidence interval.

85
Figure 3: Predicted Probabilities with Variables at their Means/Medians
55

Figure 4 shows that if we shift the variable of interest—here, the interaction of
natural disasters and the IQ Index variable—downward from its mean by one standard
deviation while holding all other explanatory variables at their means, there is a mean 5.7%
increase in the probability of the incidence of violent civil conflict in the year after a natural
disaster. Another way of stating this is that the risk of violent civil conflict is almost 1.6

55
Figure corresponds with the predicted probabilities from Table 3, Model 4. Figure
generated using Zelig statistical package for R (Imai, King, and Lau 2007).

86
times higher where the interaction term is 1 standard deviation below its mean, indicating
lower levels of pre-disaster institutional quality.

Figure 4: Predicted Probabilities with IQ Index Interaction Variable
-1 Standard Deviation from Mean
56

56
Figure corresponds with the predicted probabilities from Table 3, Model 4. Figure
generated using Zelig statistical package for R (Imai, King, and Lau 2007).

87
Figure 5 displays the same calculations if the interaction variable is shifted 1 standard
deviation above its mean, a shift that would indicate higher levels of pre-disaster institutional
quality. When the other variables are held at their means, the result is a 3.4% mean decrease
(with a 95% confidence interval) in the probability of the incidence of violent civil conflict in
the year following a natural disaster. The risk of violent civil conflict in the year after a
natural disaster is 0.63 times lower where pre-disaster institutional quality—as captured by
the IQ index measure—is relatively high.

88
Figure 5: Predicted Probabilities with Interaction Variable
+1 Standard Deviation from Mean
57

These results seem to provide preliminary confirmation of the hypothesis that pre-
disaster institutional quality is an important predictor of the likelihood of the incidence of
post-disaster violent civil conflict. However, the puzzling cases of Sri Lanka and Indonesia,
which will be more fully explored in chapter 5, provide an interesting set of challenges to

57
Figure corresponds with the predicted probabilities from Table 3, Model 4. Figure
generated using Zelig statistical package for R (Imai, King, and Lau 2007).

89
these findings. The next section discusses these cases and the issues they present for the
quantitative empirics.

Returning to the Puzzle: Indonesia & Sri Lanka
A brief investigation of the cases of Indonesia and Sri Lanka after the 2004 tsunami
yields qualified confirmation of the above cases. First, at the time of the tsunami, both
countries had very low levels of institutional quality. While the ICRG data used in the
analysis above are only available through 1999, the assessments of Freedom House are
revealing. Both countries are included in the 2004 edition of Countries at the Crossroads, an
annual Freedom House report assessing the quality of institutions in 30 countries
determined to be at critical moments in their political histories (Freedom House 2004).

90
Table 5: Freedom House Scores for Indonesia and Sri Lanka, 2002-2003
58

Indonesia Sri Lanka Mean Median Standard
Deviation
FH “basic
standard”
Civil liberties
59
3.38 4.59 3.64 3.68 0.67 5.0
Rule of law
60
2.95 4.49 3.11 3.24 0.82 5.0
Anti-corruption and
transparency
61

2.46 3.97 2.53 2.467 0.71 5.0
Accountability and
public voice
62

4.03 4.45 3.17 3.25 0.95 5.0

Table 5 displays the Freedom House scores of both countries compared to the 30
total countries contained in the report. In the categories that most closely resemble the
ICRG data utilized above, “Rule of Law” and “Anti-corruption and Transparency,”
Indonesia’s scores were below the sample mean and median, while Sri Lanka’s were above.
However, for both Indonesia and Sri Lanka, none of the scores in any of the categories

58
Data are taken from Countries at the Crossroads 2004, covering events between 1 October
2002 and 30 September 2003 (Freedom House 2004). Freedom House’s Crossroads scores
are based on a scale of 0 to 7, with 0 representing the weakest performance and 7
representing the strongest performance. The mean, median, and standard deviations are
calculated based on the 30 countries included in the Crossroads 2004 report. The Freedom
House “basic standard of effective performance” score is 5.0 across all categories as defined
in the Crossroads 2004 report (Freedom House 2004).
59
Defined as the protection from state terror, unjustified imprisonment, and torture; gender
equity and minority rights; freedom of conscience and belief; freedom of association
(Freedom House 2004).
60
Defined as an independent judiciary; primacy of rule of law in civil and criminal matters;
accountability of security forces and military to civilian authorities; equal treatment under the
law (Freedom House 2004).
61
Extent to which there is an environment to protect against corruption; existence of laws,
ethical standards, and boundaries between private and public sectors; enforcement of
anticorruption laws; governmental transparency (Freedom House 2004).
62
Measure of existence of free and fair electoral laws and elections; effective and accountable
government; civic engagement and civic monitoring; media independence and freedom of
expression (Freedom House 2004).

91
meets the Freedom House “basic standard of effective performance” of 5.0. This cursory
exploration of the quality of institutions in these countries indicates relative weaknesses
compared to basic standards.
Following the causal logic, the weaknesses of these institutions in both countries
should make them susceptible to capture by sub-state groups and thus make the incidence of
post-disaster violent civil conflict relatively more likely in both countries. The Sri Lanka case
provides cursory confirmation, as in the year after the tsunami the country descends into
civil war. However, the case of Indonesia requires further exploration. On its surface the
Indonesian case appears to run counter to the findings here, particularly given the successful
conclusion of a peace agreement in August 2005. However, the Uppsala/PRIO dataset
actually codes 2005 as a conflict year for Indonesia. In between the tsunami in December
2004 and the peace agreement in August 2005, there were a number of low-level minor
conflicts resulting in casualty rates, particularly amongst the GAM, that easily crossed the 25
battle-related-death threshold.
Yet, the conclusion of a peace agreement is an important political reality in 2005 that
is not captured in the dichotomous measure of the incidence of violent civil conflict in the
PRIO dataset. As argued in chapter 2, civil wars need to be understood as dynamic and
complex processes of conflict and violence rather than merely as a collection of discrete
episodes of onset, duration and termination. In addition, it is important to acknowledge the
perception within Indonesia—of both GAM rebels and the Indonesian government—that
the tsunami played an important role in setting the stage for peace in 2005. The GAM
negotiators remain convinced that the tsunami gave Jakarta strong reasons to come to the
negotiating table (Djuli and Rahman 2008: 28). The chief negotiator for the Indonesian

92
government argues that the tsunami gave the government and the rebels “common ground”
throughout the negotiating process (Awaluddin 2008: 26). In this light, it becomes
important to further explore this case as a disaster followed by peace, despite the PRIO
coding of the data.
In addition to exploring the post-disaster context of these two cases as dynamic
processes rather than dichotomous data points, there may be an intermediate step in the
theory or causal logic that is not captured in the empirical testing. Lower levels of pre-
disaster institutional quality may make it more likely that states and rebels battle over post-
disaster control over institutions, but the outcome of that post-disaster battle is not
necessarily pre-determined. Thus, for these two cases, we would expect battles to control
post-disaster relief institutions in the aftermath of disaster because the pre-disaster
institutional quality was low in both cases. But there are other factors that ultimately
determine the outcome of those battles. Thus, the Indonesian case may present a challenge
to the above theory and empirics. A cursory exploration of the two cases highlights these
issues and sets the stage for a more structured comparison in chapter 5.
In the immediate aftermath of the disaster in Indonesia, the GAM rebels came down
from their mountain hideouts and offered supplies to the local Acehnese villagers. They
assisted in burying the dead. They brought aid and whatever medial supplies they could to
assist the survivors; they built shelters and first aid stations (Nakashima 2005b). From the
beginning, the Indonesian military countered these efforts by issuing reports that GAM
rebels were actually attacking supply lines, attempting to replenish their own supplies that
had been severely diminished in the two years that Aceh had been under a military state of
emergency. Additionally, the military argued that any good done by the GAM rebels was

93
only an attempt to shift public opinion toward their side. The argument was that the rebels
wanted to place themselves in the best light for both the Acehnese people and the
international aid workers with whom they were in contact (Nakashima 2005b). The rebels
knew that international pressure would be necessary for GAM to be able to force the
Indonesian government to accept their demands of independence at the bargaining table.
However, the rebels were never able to gain control of any of the aid distribution. In
the end, the Indonesian government, but particularly the military, was successful in
preventing the GAM rebels from establishing any ability to autonomously provide services
to the people of Aceh. This had the effect of negatively altering the GAM’s opportunity
structures associated with rebellion due to their inability to capture a potential source of
revenue as well as the government’s ability to send a general signal of its relative ability to
provide the services of government. These changes in opportunity structures set the context
in which the GAM had little choice but to negotiate an end to the conflict in August 2005.
In Sri Lanka, however, the LTTE was able to capture some of the key economic and
political institutions of disaster relief. Almost immediately following the crisis, the Tamil
Tigers began to complain that the government was discriminately distributing aid and
purposefully keeping it from the largely Tamil areas of the country. In keeping with their
core demands that they be treated as an autonomous region, the rebels continued to insist
that international relief aid earmarked for Tamil regions of Sri Lanka be allowed to be
delivered directly to those regions, rather than get filtered through the national government
in Colombo.
However, as time passed, the Sri Lankan government was forced to deal with the
reality that the LTTE had established itself as more than just a guerrilla movement. The

94
Tigers had instituted a taxation system and used the revenue it generated as well as donations
from Tamils abroad to establish both a police force as well as a legal system. This budding
political and economic infrastructure allowed the LTTE to begin to effectively cooperate
with international aid agencies such as the United Nations Human Rights Commission
(Pocha 2005a).
As the government came to terms with this reality, it became more flexible in its
stance on the possibility of sharing the distributional responsibilities with respect to
international aid. At the end of June, the two sides reached an agreement that allowed the
LTTE a greater role in the distribution of over $3 billion in aid. According to the aid-
sharing agreement, known as the Post-Tsunami Operational Management Structures (P-
TOMS), review committees formed of rebels, government officials and Muslims would be
allowed to recommend and monitor projects in areas hit by the tsunami. The devastation of
the tsunami, in effect, created a situation in which the LTTE was able to improve its
opportunity structures associated with rebellion. It was able to capture key sources of
revenue and at the same time signal its ability to effectively manage the Tamil-majority areas
of Sri Lanka. In addition, the devastation of the tsunami and the appearance that the
government was neglecting the Tamil areas provided another set of grievances that could be
used to motivate potential recruits.
Under these conditions, the reemergence of widespread violence is not surprising.
When the concluded P-TOMS agreement broke down in the Sri Lankan Supreme Court, the
Tigers became emboldened and began to harden their position with an ultimatum. Within
the next six months, the LTTE and the government descended into a renewed state of civil
war.

95
Conclusion
It appears that in situations of natural disasters, countries with lower levels of
institutional quality may have a higher likelihood of the onset of violent civil conflict.
However, given the difficulties presented by the Indonesian case in particular and the need
to explore intrastate violence as a dynamic process beyond only static measures of onset,
duration and cessation, exploration of the causal mechanisms presented in hypotheses 1 and
2 becomes critical. Thus, this project now turns to a qualitative comparison of the battles
over post-disaster institutional control in Indonesia and Sri Lanka following the 2004
tsunami.

96
Chapter 5: Exploring the Cases of Indonesia & Sri Lanka

As discussed in chapter 3, the cases of Indonesia and Sri Lanka represent important
cases both theoretically and empirically and thus must be explored for their insights on the
relationship among disasters, political institutions and conflict. Given the current
importance of these cases and the challenges that a cursory exploration presented for the
quantitative findings in chapter 4, it is critical that they be explored more systematically.
Thus, this chapter utilizes Alexander George’s method of “structured, focused comparison”
(George 1979; Haney 1997; George and Bennett 2005: 69-72), which explores the cases by
“asking a set of standardized, general questions of each case” (George and Bennett 2005:
69).
In order to explore the hypotheses in chapter 2, the questions asked of each case will
need to explore the status of each conflict prior to the 2004 tsunami. Beyond simply status,
however, each case needs to be understood in the context of pre-disaster institutional
control. What are the histories of Indonesia’s and Sri Lanka’s battles with their respective
separatist groups as those groups attempted to seize control of local and/or regional political
institutions? After establishing these baselines, the cases will be used to explore the ways in
which the tsunami changed the incentives for either the Gerakan Aceh Merdeka (GAM) or
Liberation Tigers of Tamil Eelam (LTTE) to further seize control over institutions by
“capturing” the processes of disaster relief, the extent to which these groups were successful,
and the implications of this relative success for the ongoing patterns of conflict or
peacemaking in the year following the tsunami.

97
In order to accomplish these tasks, this chapter proceeds with a qualitative
comparison of these crucial cases, focused on the following five questions: 1) What were the
statuses of the conflicts prior to the tsunami? 2) During pre-tsunami history of the conflict
how had the various parties seized control of local, regional, and/or national institutions? 3)
How did the tsunami impact the key region in question? 4) What attempts were made by the
rebel groups to gain control of the institutions of disaster relief? How were these attempts
“reacted to” by the government? 5) How did the outcome of this interaction lead to further
developments with respect to the overall conflict? These questions allow for systematic
comparison of the two cases—specifically, the status of institutional control pre-disaster, the
effects of the disaster itself, the battles to control disaster relief institutions in the post-
disaster period, and the connections that these battles had with changing levels of violence.
After the structured exploration of each of the cases, the chapter concludes with a
discussion of results. Particular attention is paid to the differences between the quantitative
findings in chapter 4 and the findings from the qualitative comparisons here. The cases of
Indonesia and Sri Lanka demonstrate that although low institutional quality at the time of
the disaster may contribute to post-disaster battles over institutional control, they do not, in
these cases, affect the outcomes of these battles.
Ultimately, it is the relative success of the separatist groups in these cases that
contributes to post-disaster peace or violence. The GAM is relatively unsuccessful, as the
Indonesian military in particular maintains solid control over disaster relief institutions. The
result is an extremely weakened GAM that is forced to give up claims for separate statehood
by mid-August 2005. The LTTE, on the other hand, is relatively successful at capturing the
processes of disaster relief, and emboldened with new resources and new support by the fall

98
of 2005. Much of this success is the result of pre-disaster control over institutions and the
“lock-in” effect that the tsunami seems to have had with respect to this control. These
dynamics contribute to the context in which either renewed violence or the movement
toward peace becomes possible.

Indonesia
After a brief history of the roots of the conflict, this section establishes the recurring
pattern of violence and tentative ceasefire that characterizes the relationship between the
GAM rebels in Aceh and the Indonesian national government between 1976 and 2005. This
pattern places the tsunami in context and establishes its similarities to the context in which
the tsunami struck Sri Lanka. The bulk of this section focuses on the battles between rebels
and the Indonesian government—particularly the military—to gain relative control of
various disaster relief institutions in the aftermath of the tsunami. The outcomes of this
interaction contribute to the conditions under which the peace agreement is negotiated
between the rebels and the Indonesian government in Helsinki in August 2005.

The Status of the Acehnese Separatist Movement Prior to the Tsunami
The conflict between the Acehnese and the government in Jakarta dates back to
Aceh’s colonization by the Dutch. The Acehnese War, which resulted from the Dutch
invasion of Aceh in 1873, lasted for 30 years and claimed tens of thousands of lives on both
sides. During this period, any sense of unity between the Acehnese and the Indonesian
nationalists on Java was driven by their shared anti-colonialists attitudes toward the Dutch—
not by any particular shared affinity toward Indonesian nationalism per se (Aspinall and

99
Berger 2001: 1016). After a brief period of colonization by the Japanese, Indonesia was re-
colonized by the Dutch and eventually granted independence in 1949. With this granting of
independence, control over the territory of Aceh fell to the new Indonesian central
government in Jakarta.
The rise and continued motivations of the separatist movement—Gerakan Aceh
Merdeka—have generally been traced to three factors: identity politics within Aceh, the
political economies associated with resource extraction in the region, and the behavior of the
Indonesian military and the human rights abuses that occurred during the Military
Operations Area (Daerah Operasi Militer, “DOM”) period between 1990 and 1998 (Barron
and Clark 2006). The relative strength of each of these factors has ebbed and flowed over
time, with certain factors being more helpful for explaining the motivations of the GAM and
of local Acehnese support for the GAM during certain time periods.
A sense of difference of identity for the Acehense people vis-à-vis other
Indonesians sowed the seeds of the conflict in the early 1950s, and the political economies
associated with the Aceh’s natural resources were particularly important to the uprisings in
1976 and 1989. The identity-related grievances have their roots in the 1949 grant of
independence. The Acehnese agreed to the membership in the new Indonesian state, but
did so under the auspices that they would be granted local autonomy due to the importance
of local Islamic culture and as a reward for the region’s contributions in the nationalist fight
against the Dutch. The regional autonomy was short-lived, and by 1953, the sense of
betrayal amongst Muslims in Aceh especially was very high (Schulze 2004). A rebellion led
by Islamic religious leader (ulama), Teungku M. Daud Beureueh, was managed by the
Sukarno government throughout the late 1950s using a variety of compromises that granted

100
“special region” status to the territory. In 1961, the Indonesian government granted Aceh
the ability to enforce Islamic Shari’ah law for Muslims living in the region (Miller 2008: 13-
14). These promises of limited autonomy were never really upheld, and the result was an
even greater divide in the way that the Acehnese people viewed their identity vis-à-vis the
rest of Indonesia.
Over time, the GAM came to mythologize the fact that the Sumatra-Aceh region
was included in the transfer of control to the Javanese as an illegal transfer of their own
sovereignty (ASNLF 1976). They viewed their territorial claim in terms of a singular national
identity that they based on ethnicity, language, culture, history, religion and geography (Kell
1995; Reid 2004; Schulze 2004; Miller 2008). Many of these differences between the
Acehnese and the rest of Indonesia were real. Michele Ann Miller notes, for example, that
“Aceh’s strategic location along the Malacca Straits trading route also led to the development
of a Malay-Islamic written and cultural tradition, setting the Acehnese apart from many other
ethnic groups in the archipelago closer geographically to the island of Java” (Miller 2008:
12).
63
In addition, the deep Acehnese roots in Islam were at odds with the explicitly secular
nation-building policies of Sukarno (Schulze 2004: 1).

63
However, Miller and others (e.g., Schulze 2004: 6) also argue that some of this GAM view
of a unique Acehnese identity and of an illegal transfer of sovereignty is contrary to the
actual history. She points out that a full 20 percent of people living in Aceh are not ethically
Acehnese at all, but instead members of seven other linguistically and culturally distinct
ethnic groups. With respect to sovereignty, the Acehnese willingly volunteered, cooperating
with—and at times supplying critical resources to—the Indonesian nationals in Java (2008:
13). Nonetheless, these perceptions as the basis for grievance are important aspects of
understanding the roots of the separatist claims that lie at the heart of the conflict.

101
Suharto’s ascendancy in 1966 brought with it the “New Order”
64
regime, and a
variety of factors during his 32-year rule contributed to an even stronger sense of identity-
based independence within Aceh. Suharto attempted to use his regional development
initiatives in Aceh to counter, at least in part, the local influence of the religious ulama. While
political and economic decision-making during this time became increasingly concentrated in
Jakarta, Suharto fostered a new class of bureaucratic technocrats in Aceh. These were
mainly indigenous Acehnese, and while their rising political influence did foster divisions
within Aceh—particularly with the older religious order (Sulistiyanto 2001: 440)—it also
contributed to an even greater sense of Acehnese identity, particularly relative to the rest of
Indonesia and the leadership in Jakarta (Sulistiyanto 2001: 440; Miller 2008: 14).
Resentment around these issues only increased during the 1970s with the passage of
Law No. 5/1974. The law stripped Aceh of any autonomy previously promised and placed
even greater power in the hands of the Indonesian government. It established gubernatorial
control over local government in Aceh and provided the leadership in Jakarta the power to
appoint those governors (Miller 2008: 14).
Yet, the contemporary separatist conflict’s historical roots in cultural, ethnic, and
religious differences must be understood in the context of the political economies associated
with the region’s natural resources (Kell 1995: 13-28; Robinson 1998: 135-39; Bowen 2002:
340; Aspinall 2007). Aceh is a region rich in oil, natural gas, mineral deposits, and timber,

64
Suharto used the label “New Order” to characterize his regime as it seized power in 1966
and specifically to contrast it with the regime of his predecessor Sukarno. While the label
was originally used in reference to the transition between regimes and to signify separation
from all frustrations with the Sukarno regime, it came to be associated with the entire period
in which Suharto ruled Indonesia (1966-1998). For a good overview of these and other
developments in Indonesia and the region, see Tarling (1999).

102
and much of these were discovered during the late 1960s and early 1970s. During the mid-
1970s, the New Order government’s heavy extraction of these resources was accompanied
by two key institutional developments that contributed to an uneven distribution of the
resource revenues. First, due to the material benefits associated with massive corruption
(i.e., bribes, kick-backs, etc.), the regime came to rely heavily on foreign capital (Robinson
1998: 134). Thus, profits from resource industries in Aceh were pulled out of the province
and accrued to Indonesian and foreign businesses rather than Acehnese people (Dawood
and Sjafrizal 1989: 115; Kell 1995: 14-18; Robinson 1998: 134-39; Aspinall and Berger 2001:
1017; Sulistiyanto 2001: 439).
Second, the New Order regime’s approach to economic decision-making regarding
the province was highly centralized (Kell 1995: 27; Robinson 1998: 134). This concentration
of decision-making authority over large matters of policy as well as bureaucratic agencies that
issued licenses and made decisions on smaller matters resulted in what Kell calls “Java-
centric effects.” In particular it assured the result that members of the military and
government officials in Java would be the greatest beneficiaries from the negotiation of
contracts regarding the extraction of resources in Aceh (Kell 1995: 27). These discrepancies
were felt quite acutely in Aceh and resulted in the growth of support for the Aceh Merdeka
movement (Morris 1983: 300; Dawood and Sjafrizal 1989; Kell 1995: 54-55), particularly as
this view of a “shared fate” reinforced ideas about Acehnese identity separate from the rest
of Indonesia (Robinson 1998: 139; Aspinall 2007).
In mid-1990, the government’s approach to combating the separatist movement
shifted to a full-scale military response, and the Aceh region officially became a Military

103
Operations Area (Daerah Operasi Militer, “DOM”).
65
Reports issued during this time
demonstrate the wide latitude given to the Armed Forces of Indonesia (Tentara Nasional
Indonesia or “TNI”) to make political decisions in their areas of operation, all in the name of
national security (Amnesty International 1993; Robinson 1998; Sulistiyanto 2001). This level
of militarization paved the way for massive abuses of power during 1990-1998. The TNI
employed a campaign of terror throughout the region, burning down houses of suspected
GAM supporters (Robinson 1998: 141), raiding homes at night, arresting citizens arbitrarily,
torturing and raping suspected sympathizers and finally utilizing public executions (Amnesty
International 1993: 18). In addition, there is evidence military utilized civilians as a shields in
a process known as “fence of legs,” where “ordinary villagers were compelled to sweep
through an area ahead of armed troops in order to both flush out rebels and inhibit them
from returning fire” (Amnesty International 1993: 12).
66

These tactics resulted in more than 1,000 deaths, 2,000 disappearances, and 3,000
people reported to have been tortured during the nine years of DOM (Sulistiyanto 2001:
442).
67
However, despite the fact that by 1991 the GAM had been effectively removed from
Aceh (Schulze 2004: 5), the continued military-led violence during the next 7 years
contributed to a strengthening of public sentiment toward support for the GAM movement.

65
Despite the increase in military presence, the government did not officially acknowledge
DOM as a counter-insurgency operation against a separatist movement. In fact, it refused to
classify the GAM as anything other than a “gang of peace disturbers” that had “no political
motives whatsoever” (Sukma 2004: 9).
66
Robinson (1998: 143) notes that this strategy was not unique to Aceh, having been used
widely by the TNI in East Timor and other counter-insurgency operations.
67
Approximately 7,000 cases of human rights abuses have been documented to have
occurred during the DOM period (Schulze 2004: 5).

104
A number of the key leaders of the GAM had escaped to Malaysia during DOM and had
been supported by various Acehnese diasporas (Barber 2000: 34; Schulze 2004: 5). While
earlier discourse on the topic had been based on the foundational disputes over economics
and Acehnese distinctiveness, the actions of the New Order government during the 1990s
became the object of frustration and outrage that fueled many GAM supporters (Robinson
1998: 145; Aspinall and Berger 2001: 1017; Schulze 2004).
In the years following the downfall of the Suharto, the new Wahid government
attempted to bring about a more peaceful resolution of the conflict and allowed international
aid to begin to flow into the region. In May 2000, the government signed a deal with the
GAM for a “Humanitarian Pause” (jeda kemanusiaan) in the conflict. During the period of
time that followed—as it had done with previous ceasefire periods—the GAM capitalized
on its newly energized support gained as a result of the human rights abuses committed
during DOM.
During the post-DOM period between 1998 and 2004, a recurring pattern of
hostility emerged: a period of ceasefire, followed by renewed GAM activities, followed by a
renewed military counter-insurgency campaign, ultimately leading to another period of
ceasefire. The TNI offensive during 2001 weakened GAM significantly, producing GAM
casualties to such an extent that the GAM was forced to change its command structure at
the local level (International Crisis Group 2005: 4-5). In 2002, the signing of the Cessation
of Hostilities Agreement (“CoHA”) created zones of peace in the region, created a joint
GAM-government security committee, and a government agreement to phased relocation of
the TNI (Cessation of Hostilities: Framework Agreement between Government of the
Republic of Indonesia and the Free Aceh Movement 2002; Schulze 2003).

105
However, in May 2003, the CoHA collapsed, mainly due to the lack of progress by
the government on the demilitarization process.
68
A 2003 declaration of martial law by the
new Sukarnoputri government was followed one year later by a declaration of civil
emergency. Again, during this period, the TNI was able to make significant gains disrupting
a GAM that had taken advantage of the CoHA in order to rebuild (Kipp 2004; International
Crisis Group 2005: 4-5; McCulloch 2005). Despite changing levels of violence, these
particular patterns associated with ceasefire and military response had important effects for
each side’s ability to gain control of important local institutions, particularly in the run-up to
and aftermath of the 2004 tsunami.

Pre-Disaster Institutional Control
The ebb and flow of the level of military response throughout the years has
correlated with the GAM’s ability to gain control of specific localities and wield power
through local political institutions (Schulze 2003). During times of relative peace, where the
Indonesian government relied more on non-military means to deal with the separatists, the
GAM was able to strengthen institutional control. During times when government relied
more heavily on direct military counter-insurgency operations, the GAM rebels became very
dispersed and went into hiding in the forests and hills. As a result, their control over local
village and district institutions virtually disappeared.

68
For a comprehensive overview of the negotiation and subsequent failure of the CoHA, see
International Crisis Group (2002, 2003b, 2003a, 2005), Schulze (2003, 2004), Aspinall and
Crouch (2003), Reid (2004), and McCulloch (2005), among others.

106
In the mid-1970s when GAM was established as an organization, its membership
was limited.
69
However, during the next 24 years, the GAM was able to make large
geographic and institutional gains during two key time periods: 1986-89 and 1999-2001. In
the 1980s GAM was able to recruit locally and send new guerrillas off to Libya to be trained
(Schulze 2004: 30-31). With newly trained forces, the GAM expanded into Greater Aceh,
North Aceh and East Aceh (Schulze 2003: 15).
Early in its existence, the GAM had little control over any political institutions in
Aceh. To the extent that it did exert control in various localities, that control was often the
result of the fractionalization that occurred within the GAM as it expanded its membership.
New recruits had a variety of motivations, not the least of which was economic gain through
ad hoc cooptation and extraction of local tax revenues (Aspinall 2000; Schulze 2004: 20-21,
24).
In the period of Humanitarian Pause, the GAM was able to rebuild its military and
political capacity and effectively disable and/or co-opt various structures and institutions of
local government (Schulze 2003: 24). By the summer of 2001, the GAM had considerable
political influence in Aceh, particularly relative to its military capability. A 2001 report of the
International Crisis Group listed the membership of the militant wing of the GAM at
between 15,000 and 27,000 people. However, the number of firearms amongst this group
numbered only a few thousand (International Crisis Group 2001: 11). Despite this, the
GAM at the time controlled some 75 percent of the territory of Aceh, mostly at the village

69
Schulze (2004: 14) argues that this was due in part to the secular nature of the original
members. Most of the initial 70 or so rebels that joined Hasan di Tiro in the mountains
were motivated more by loyalty to di Tiro than by any particular issues associated with
identity. In addition, much of the early leadership of the group was university educated.

107
level and particularly in the districts of North Aceh, East Aceh, Pidie and Bireun
(International Crisis Group 2001: 11).
In the midst of the GAM’s control, many of the Indonesian government
representatives in these areas had fled their posts at both the village and sub-district levels,
and the legal system in these areas had virtually ceased to function (International Crisis
Group 2001: 12-13). In this vacuum, two sets of new institutional gains were made. First,
on the economic front, the GAM increased its ability to collect taxes. Throughout its history
GAM has utilized “taxes” as a means of financing its own activities. These taxes have
involved seeking direct payments on both profits and project values, demanding that local
constituencies donate parts of their incomes, and skimming off the top of various
humanitarian and development aid, coming both from international NGOs and the
Indonesian central government (Schulze 2004: 24-27).
However, a second key development was the reemergence of traditional Acehnese
institutions for managing local villages. The GAM attempted to create structures of local
governance that mirrored those controlled by the Indonesian government (Barre et al. 2006:
10). During this time, for example, the tuha peut—a group of village elders—reemerged and
began to take over the function of village management and dispute adjudication
(International Crisis Group 2001: 13). This group, in accordance with the GAM, presided
over minor legal issues such as the issuance of birth and marriage certificates (Schulze 2004:
12).
The creation of parallel structures of governance was a critical part of the GAM’s
strategy, particularly following their reemergence during the Humanitarian Pause. Schulze
cites several interviews with leading GAM officials, such as Tiro field commander Amri bin

108
Abdul Wahab, who in 2003 argued “the crucial element is how to establish a GAM
government so we can exercise control and society does not have to deal with the
Indonesian structure. That strengthens our relationship with society and we can spread our
ideology” (Schulze 2004: 34). The GAM minister of state, Malik Mahmud, echoed these
sentiments in 2002 stating, “In two years [since the end of DOM] we have taken over 60 to
80 percent of the administration of the Indonesian government in Aceh. We make use of
Indonesian officials. We know they have a job with Indonesia, but now we are in power in
Aceh, and we want them to change so what you see is positive.” (Schulze 2004: 35). In
2001, independent reports argued that the GAM was in control of as much as 75 percent of
the region of Aceh (International Crisis Group 2001: 11).
In 2001, the government began another military offensive to try to counter the
political, economic, and military gains the GAM had made since 1998. Again, this strategy
proved effective, as the TNI was able to reduce the GAM’s gains by approximately half.
The GAM’s self-reported percentages as to the amount of the province of Aceh it controlled
in 2002 fell to 30-40 percent, down from 60-70 percent during the preceding year
(International Crisis Group 2001: 12-13; 2002: 2).
Aside from their effectiveness in dismantling the GAM institutions and local control
structures, these counter-insurgency campaigns also had the important effect of entrenching
control over many local institutions within the hands of the TNI. Starting with Suharto’s
New Order, the military had been granted dual function (dwifungsi) over both external
matters (national defense) as well as internal matters (domestic security). This dual
functionality created the conditions in which the military grew to play a very large political
role within the country. It also allowed the military to establish local control, giving them

109
license to set up their own businesses for profit (McCulloch 2005).
70
These practices
associated with self-financing continued into the 1990s and were especially prevalent during
the DOM period where they became linked to a number of violations of power, human
rights abuses and a high level of corruption (Large 2008: 65).
During DOM, the military initiated and managed a number of local development
and infrastructure programs (Barber 2000: 35). The goal of these territorial operations
(operasi territorial) was to counter the gains in public support that had been made by the GAM
during the late 1980s by winning loyalties of the Acehnese people (Sukma 2004: 11). The
military engaged in a similar practice during the declarations of martial law and civil
emergency during 2003, attempting to provide humanitarian aid in a way that “separate[ed]
GAM from the civilian population” and destroyed institutional gains the GAM had made
during the Humanitarian Pause (Sukma 2004: 23). The government was able to make some
gains during this time, restoring local structures of governance—particularly sub-district
heads (camats) and village heads (keuchik)—that had previously fallen under the GAM’s
control (Sukma 2004: 25).
Thus, in the years leading up to the tsunami, the GAM’s structures of parallel
governance at the local level were countered, and in many cases overturned, by the
Indonesian military’s own efforts at local institutionalization. In the aftermath of the 2004
tsunami, this pre-disaster military control of institutions would play a large role in the
Indonesian government’s decision to place much of the control of local disaster relief in

70
McCulloch (2005) points out that this revenue raising has always been important for the
regular function of the TNI, particularly as there was little funding left to direct to the
military following the major post-independence and post-war projects initiated by Sukarno
and Suharto.

110
Aceh under the oversight of the TNI. In this context, the military was effectively able to
prevent the GAM from making any post-disaster institutional gains, a crucial contribution to
the context in which the GAM agreed to a peace deal in August 2005.

The Impact of the 2004 Tsunami
The huge moment magnitude 9.1 earthquake that occurred on 26 December 2004
near the island of Simeulue off the cost of Sumatra was located at a depth of about 30 km
and immediately generated a tsunami that took only 30 minutes to hit the northwestern coast
of Sumatra (Bilham 2005; Charles et al. 2005; Lay et al. 2005). The northern tip of Sumatra
was hardest hit by the tsunami with 165,000 estimated to have lost their lives.
71
The tsunami
was devastating for the GAM as well. Approximately 200 GAM rebels drowned as they
were being held as prisoners when the floodwaters swept threw a prison in Banda Aceh.
72

Another 180 GAM rebels died in a similar manner, as they were held captive in a prison in
the town of Lhoknga (Tanner 2005).
In the immediate aftermath of the disaster, both the GAM and the Indonesian
government made conciliatory gestures toward one another. Indonesian President Susilo
Bambang Yudhoyono lifted the state of emergency in Aceh in order to facilitate the flow of
aid and relief workers into the area. Rebel leaders from the GAM offered a tentative

71
The original estimates in the months immediately following the disaster were as high as
250,000 deaths (McCulloch 2005; Soesastro and Atje 2005). However, more recent, and
believed accurate, estimates put the number killed between 165,700 (Rofi, Doocy, and
Robinson 2006) and 167,000 people (BRR and International Partners 2005). See Gaillard et
al. (2008) for a discussion.
72
One of the escapees from the Banda Aceh prison was the GAM commander Yusuf
Irwandi, who was elected governor of Aceh in 2006 (Marshall 2007).

111
unilateral ceasefire. This rebel ceasefire was not matched by the TNI. Within a week, it
appeared that events in the aftermath of the tsunami would increase conflict between the
Indonesian government and the GAM.
During the following weeks, the Indonesian government, the military, and the GAM
all had radically different interpretations of what was happening on the ground. Toward the
middle of January, the chief of the Indonesian Army, Ryamizard Ryucudu, announced that
208 rebels had been killed by the military in the three weeks following the tsunami (Powell
2005). Since May 2003, the rate at which GAM rebels were being killed had been about 115
per month; 208 in three weeks represented a significant increase. The military argued that
the rebels had been armed; the GAM countered that most of those killed were not rebels at
all but unarmed civilians. The military claimed that rebels had been attacking aid supply
convoys; the GAM rebels claimed that this was not the case and that the military had been
attacking indiscriminately (Powell 2005).

The Battle over Control of Disaster Relief Institutions
The extent of destruction in many of Aceh’s provinces was such that many local
government institutions were either physically destroyed or rendered completely ineffective
because of lack of personnel (McCulloch 2005; Barron 2008). In the immediate aftermath of
the tsunami one source reported that over one-third of the 4,312 villages were left without
functioning government (McCulloch 2005: 25). In addition, much of the public
infrastructure in the region was destroyed, including an estimated 3,600 miles of roads and
490 bridges (McCulloch 2005: 25). Given this devastation and the preexistence of the
Indonesian military in the region, many of the institutions of disaster relief were immediately

112
folded into the military structures and processes that had been established to handle the pre-
existing civil emergency.
The overall management of the relief effort was handed over to the National
Coordinating Board for Disaster Management (Bakornas), under the chair of the Vice-
President, Yusaf Kulla. However, within this scheme, the implementation, coordination,
and oversight of these relief processes were handled by the TNI (McCulloch 2005: 25). In
this capacity, the TNI created and assisted in the creation (by NGOs and international aid
organizations) of many poskos—local emergency aid distribution centers—which provided
medical aid, food, water, and information to the wounded and displaced. Nearly 80 percent
of the military personnel in Aceh were deployed to the western coast in the days following
the tsunami. They were tasked with collecting bodies and rebuilding critical infrastructure
such as water and sanitation facilities and bridges (McCulloch 2005: 28).
For their part, in the immediate aftermath of the tsunami, the GAM rebels came
down from their mountain hideouts and offered supplies to the local Acehnese villagers.
They assisted in burying the dead. They brought aid and whatever medial supplies they
could to assist the survivors; they built shelters and first aid stations. They crafted makeshift
stretchers and transported many of the most critically injured to Banda Aceh for medical
care. In some of the more remote sections outside of Banda Aceh, it was reported that the
GAM’s medical care was the only care available for four or five days after the tsunami struck
(Nakashima 2005a).
Despite these efforts, two factors contributed to GAM’s relative inability to
participate in the disaster relief process. First, in the days following the tsunami, many of the
GAM’s humanitarian motives and efforts were countered through their own doing and

113
through a TNI propaganda campaign. In a couple of instances, the GAM attempted to use
the tentative ceasefire initiated by the disaster in the same way they had used previous
ceasefires: to regroup, rebuild, and attempt to regain institutional control. For example, in
the first week following the disaster, rebels captured an aid convoy, seizing all of the goods
(Schulze 2005). Such occurrences seriously undermined the GAM humanitarian credibility.
The TNI seized every opportunity to make this argument to the people of Aceh and
Indonesia and discredit the GAM (e.g., Burns et al. 2005; Huxley 2005). The military
severely restricted movement of international disaster relief organizations and NGOs
throughout Aceh, requiring notification and registration as conditions of operation (Howden
2005; McCawley 2005). The chief of the TNI, General Endriartono Sutarto, cited individual
security, aid security, and attempts by the GAM to steal aid supplies as justifications for the
policies (McCulloch 2005: 26). No workers were allowed to travel outside of Banda Aceh or
Meulaboh without clearance by the TNI. Various international NGOs reported difficulty in
getting access to areas of Aceh—particularly the internal regions—where the heaviest
fighting was reported (Burke and Afnan 2005).
Second, there is evidence that the TNI used the opportunity of the disaster to
conduct offensive operations against the remaining GAM rebels. There were reports that
Acehnese citizens, in the midst of receiving relief services from the military, were forced to
identify any GAM rebels or GAM-sympathizers (Kearney and Spencer 2005). In addition,
while the majority of TNI troops were involved in the provision of disaster relief aid
throughout the country, a number of military-supported militias appeared and conducted
devastating attacks on rebel strongholds and areas sympathetic to the GAM (Kearney and
Spencer 2005).

114
The result was that the Indonesian government—and particularly the military—was
able to maintain and even augment institutional control in the months following the tsunami
(Barron 2008: 61). Part of this institutional arrangement was dictated out of necessity given
the magnitude of the disaster and the pre-disaster role of the military in the local governance
of many of the villages of Aceh. However, another part this reality was driven by the
GAM’s own attempts at institution capture, the government’s worry that the GAM might try
to seize control of aid distribution, and the TNI’s ability to rhetorically spin the GAM’s
motives as anything other than altruistic.

The August 2005 Peace Agreement and its Aftermath
In late January, the Indonesian government announced that it would be willing to
talk to the Acehnese leaders exiled in Sweden. Shortly before the scheduled meeting,
Indonesian President Yudhoyono offered the GAM rebels amnesty and greater autonomy in
exchange for a ceasefire. A presidential spokesperson was quoted as saying, “We are coming
with an olive branch…Let’s move forward to rebuilding Aceh within the framework of the
Republic of Indonesia” (Nakashima 2005b).
Despite their history, by the end of the January meetings, the GAM leaders had
relaxed of their previous insistence on complete Acehnese independence. While laying the
groundwork for future talks, the GAM leaders stated that they were ready to accept an
agreement that would not provide Aceh’s complete independence (Donnan and Bergstrom
2005). Despite this progress, the meetings between GAM and the Indonesian government
continued in July 2005. In moving towards a more specific agreement, a number of key
issues continued to divide the two sides. The GAM initially proposed that in return for its

115
giving up on Acehnese independence, it be allowed to become an official political party.
The government initially resisted this request, fearing that allowing the GAM to run in local
elections might lead to a referendum such as the one that ended Indonesian rule in East
Timor in 1999. Another critical and historically divisive issue was the specific way in which
demilitarization would take place.
Eventually, an agreement was reached. The government agreed to allow the GAM
to form a political party as long as it would do so within bounds of Indonesian election laws
(The Government of the Republic of Indonesia and The Free Aceh Movement 2005). Such
laws only allow for the existence of nationally based political parties for fear that parties
operating only within a particular locality could stimulate separatism. As for demilitarization,
both sides agreed that an international mission made up of two hundred European and one
hundred Association of South East Asian Nations (ASEAN) observers would be allowed to
oversee a process in which amnesty would be offered to all rebels willing to lay down their
arms. In addition, seventy percent of revenues generated from the oil-rich province would
go directly to the provincial government. Finally, both sides agreed to establish a Human
Rights Court and Truth and Reconciliation commission so that all aspects of the thirty-year
conflict could be uncovered and grievances addressed. The agreement was officially signed
on August 15, 2005.
By the middle of September, the GAM rebels were turning in their weapons
(Donnan and Hidayat 2005), political prisoners were starting to be released (Donnan 2005),
and Indonesian troops were slowly starting to leave the region. During the following years,
the reconciliation and reconstruction processes were not been without serious difficulty.
There remained many difficult and unresolved issues including keeping the peace

116
(International Crisis Group 2007a; Jones 2008; Schulze 2008), the reintegration of the GAM
into Indonesian society (Frödin 2008), the legal transition from the August 2005 Memorandum
of Understanding to the more specific, autonomy-granting Law on the Governing of Aceh (May
2008), the prosecution of military officers for human rights abuses during DOM (Hadi
2008), the role of Islam in the current and future governance of Aceh (International Crisis
Group 2006a), and the ongoing processes of reconstruction in the aftermath of 30 years of
civil violence and a major natural disaster (BRR and International Partners 2005; Barron
2008; Large 2008). However, the election of former imprisoned GAM commander, Irwandi
Yusuf, as governor of Aceh in December 2006 is evidence that the August 2005 peace
process has been at least partially consolidated (International Crisis Group 2007b; Kingsbury
2007; Aguswandi and Large 2008).

Sri Lanka
In Sri Lanka, the years preceding the 2004 tsunami were quite similar to those in
Indonesia. A twenty-year civil war between the Sinhalese-speaking and largely Buddhist
majority Sri Lankan government and the Tamil speaking and largely Hindu minority Tamil
Tiger rebels had resulted in approximately 60,000 deaths. Although the Sinhalese majority
controlled the Sri Lankan capital and national government in Colombo, the Tamil Tigers
controlled a 30 square mile patch of land in the northeast corner of the island. In fighting to
extend their autonomous rule to other parts of the northeast where Tamils resided, the
rebels had perfected the tactic of suicide bombing. Throughout the war, they carried out
many successful attacks on Sinhalese politicians and civilians as well as more moderate

117
Tamils. In 2002, a Norwegian-brokered ceasefire put a tentative end to the violence.
However, in May 2003, talks between the two parties broke down and in the months leading
up to the tsunami, the tensions between the two parties seemed to be steadily increasing.
In the context of the 2002-2003 ceasefire, the LTTE regenerated a process that it
had begun in the 1990s, hedging its strategy by expanding from a purely military organization
to one which would attempt to consolidate its territorial control through the cooptation
and/or establishment of political institutions. This strategic choice resulted in an attempt to
use its tax revenues to provide domestic security, build local procedures for the rule of law,
and provide a number of social services. Some of this was done directly by the LTTE, but
much of it—especially the service provision—was done through local Tamil-controlled
NGOs. While the Sri Lankan government attempted to combat these attempts at
institutional control, much of the northern and eastern territories of Sri Lanka were
operating under either Tamil-controlled or dual administrative structures in the months
preceding the tsunami. Consequently, in the aftermath of the disaster, the LTTE was able to
gain control of many of the processes and resulting institutions of disaster relief. These
developments contributed to the context in which civil war reemerged in the late fall of
2005.

The Status of the Tamil Separatist Movement Prior to the Tsunami
The contemporary violence on the island of Sri Lanka has its roots in the mid-1970s
as the Liberation Tigers of Tamil Eelam (LTTE) began their military campaign against the
Sri Lankan government. The goal of the campaign was to achieve an independent, separate
state (“Eelam”) for the Tamil minority on the island. Tamil views on self-autonomy and

118
nationalism were initially constructed during the politics of the late 19
th
century and then
cemented during the post-colonial politics of the 20
th
(Kearney 1985; Stokke and Ryntveit
2000: 292). After British rule ended, the Tamils were often excluded from Sri Lankan
politics, disqualified, for example, from any bureaucratic or administrative posts and not
allowed to participate in the judiciary or serve in the Sri Lankan army (Manogaran 1987;
Chandrakanthan 2000; Flanigan 2008: 500-01). Decades of systematic exclusion fostered
wide spread resentment, which fed into Tamil nationalistic sentiments (Kearney 1985: 904;
Stokke and Ryntveit 2000).
In 1976, the Tamil United Liberation Front (TULF) declared that the only way to
insure the survival of the Tamil Nation was for it to become an independent state, making
separatism a newly stated goal (Kearney 1985: 904-05). In this context, the LTTE, founded
by Vellupillai Prabhakaran in 1972, began to position itself as the self-proclaimed leader of
the Tamil minority community in Sri Lanka (Wilson 2000; DeVotta 2004; Flanigan 2008:
501).
73
Tamil clashes with the Sri Lankan government grew increasingly violent throughout
the 1970s and 1980s.
74

In 2002, the Sri Lankan government and the LTTE signed a Memorandum of
Understanding (“MoU”), which laid important groundwork for future Tamil institutional
control. The substance of the memorandum involved an end to the government embargo of

73
This is not to say that the LTTE has complete acceptance from the population of its role
as leader of the Tamils. There are a number of reports that indicate that in many of the
northern and eastern provinces claimed by the LTTE, much of the community support is
due to the LTTE’s ability to either directly silence dissent or coerce (e.g., Wilson 2000;
Human Rights Watch 2004; La 2004; Philipson and Thangarajah 2005; Stokke 2006;
Sarvananthan 2007; Stokke 2007; Flanigan 2008).
74
For a comprehensive overview, see Philipson et al. (1998).

119
the northeast and a reopening of the A-9 highway
75
by the LTTE (Shastri 2003: 216).
However, the MoU had the effect of legitimizing the political objectives of the LTTE by
designating them as “partners.” In so doing, the government ceded control of northeast Sri
Lanka, allowing the LTTE to keep its arms and military bases in those areas. The Tigers
were also authorized to move unarmed into government-held areas in the North and East
provinces in order to open up political offices (Shastri 2003: 216)
After a year of further negotiations, the LTTE and the Sri Lankan government
agreed to a formal ceasefire in early 2003 (Agreement on Ceasefire 2003). Much of what had
been agreed to in the more formal MoU was now institutionalized in the formal ceasefire
agreement. In July, the Sri Lankan government presented a more permanent solution to
questions of governance in the northern and eastern provinces. The proposal called for a
two-tier structure of governance that left the LTTE’s specific roles relatively undefined: it
allowed the LTTE to manage development institutions but denied it any control over
revenue, land, and security (DeVotta 2004: 52).
The LTTE rejected this proposal outright. In October 2003, they countered with a
governance proposal of their own. The Tigers’ proposal called for an LTTE-led “Interim
Self-Governing Authority (ISGA)” in the northeastern portion of Sri Lanka (Liberation
Tigers of Tamil Eelam 2003). This ISGA would be granted powers over development,
revenue generation and disbursement, trade, foreign aid, the adjacent seas, and all
administrative structures including police and courts (Liberation Tigers of Tamil Eelam

75
The A-9 Highway is one of the main roads linking the center and south of the island with
the north. It had been under control of the LTTE—and thus effectively closed—in the
years prior to the MoU (Shastri 2003: 216).

120
2003; DeVotta 2004: 52).
76
This measure was rejected by the government but was
subsequently used by the LTTE anyway as the basis for their claim to be the rightful
government of Tamil regions.
By early 2004, both the government and the LTTE were internally fragmented, and
this contributed to greater levels of external conflict with each other (DeVotta 2005). The
disputes amongst Tamils over the defection of Vinayagamoorthy Muralitharan (Colonel
Karuna) from the Vellupillai Prabhakaran-controlled LTTE, for example, culminated in a
number of small skirmishes. The violence escalated as Karuna attempted to start a new
party—the Tamil Eelam People's Liberation Tigers (TEPLT)—in an attempt to undercut the
power and legitimacy of the LTTE (DeVotta 2005: 102-03). In this context, the Sri Lankan
military attempted to work with the TEPLT in order to split the Tamils further (DeVotta
2005: 102-03). Thus, although the tsunami technically occurred within the time frame of the
ceasefire agreement, the peace in Sri Lanka in December 2004 was tenuous at best.

Pre-Disaster Institutional Control
The LTTE spent much of 2003 attempting to consolidate the political gains made
through the peace process (DeVotta 2004: 50; Flanigan 2008: 501). This resulted in a
qualitative shift away from strictly military means and toward a greater role for political
structures within the LTTE controlled areas (Stokke 2006: 1022). The LTTE continued to
use their system of taxation in order to generate local revenues. They consolidated the

76
Shastri (2003: 218 fn 6) notes that at this point in time, the LTTE already had a level of
administrative control in the northeast with their own judicial system, civilian courts, and 19
police stations. Thus, the proposed ISGA structure was merely an attempt to gain a stamp
of legal legitimacy from the Sri Lankan government.

121
power of their police forces, courts and other administrative structures. However, they also
became aware that their relative ability to fill perceived institutional voids in Tamil regions
could provide a key means to securing the support and trust of the Tamil public (Flanigan
2006; 2008: 503). Thus, the LTTE began to create several organizations, focused on the
economic development of the region and the provision of services to the wider population.
Some of these organizations were explicitly LTTE-run while others were operated under the
guise of an NGO moniker.
Stokke (2006) argues that by the end of 2003, the Tamil-controlled areas of Sri Lanka
were being operated as if they were a completely separate country.
77
The Tigers had created
a “Code of Law” for these regions that established a court system, a police force, a military
78

and a variety of administrative procedures. Much of this governance was paid for using a
system of “taxation” that involved collecting taxes from both the local population as well as
NGOs operating within its territory (Chandrakanthan 2000; Stokke 2006; Flanigan 2008).
These structures of governance had broad reach, overseeing institutions of education,
economic development, health care, and arts and culture (DeVotta 2004).

77
He argues “traveling from government-controlled to LTTE-controlled areas resembles a
border crossing between two nation states” (Stokke 2006: 1022). However, it is important to
point out that Stokke’s characterization of the level of LTTE institutional control remains
somewhat controversial. For a heated and politicized exchange on these issues, see the
rejoinder by Sarvananthan (2007) and the response by Stokke (2007). While a number of
Sarvananthan’s concerns about methodology are legitimate—and acknowledged—Stokke’s
response provides satisfactory (at least to this author) explanations of his methods and
motives. Thus, his descriptions of LTTE institutional control are heavily relied on in this
chapter and this section in particular.
78
The military branch of the LTTE included an intelligence unit, a navy (the “Sea Tigers”), a
small air force (the “Sky Tigers”), and an elite unit known as the “Black Tiger Squad,” which
was responsible for carrying out most of the terrorist attacks (mostly suicide bombings)
within Sri Lanka (Hogg 2006: 6).

122
A number of NGOs were created by the LTTE in order to provide services to the
Tamil population. For example, the Tamils Rehabilitation Organisation (TRO) is often
described as the “humanitarian arm of the LTTE” (Flanigan 2008: 512). Initially, the TRO
was established in 1985 as a means of supporting Tamil refugees in South India (Hogg
2006). However, the TRO now provides a wide variety of services to local Tamil
communities such as construction of housing, education services, sanitation projects, and
health and medical relief (Tamils Rehabilitation Organization (TRO) 2005a).
In addition to its primarily local presence, the TRO also established wide
connections with the greater international NGO community, setting up offices in 18
countries around the world and establishing partnerships with a number of prominent
NGOs such as Save the Children and UNICEF (Hogg 2006; Flanigan 2008). While there
has always been debate over the exact relationship between the TRO and the LTTE
(Flanigan 2008: 512-13), most Tamils viewed the operations of the TRO as reflective the
“wishes of the LTTE leadership” (Wayland 2004: 422). They popularly pointed to the fact
that the LTTE had appointed the eastern head of the TRO as its own political chief for the
eastern part of Sri Lanka (Philipson and Thangarajah 2005).
In addition to this international presence, the LTTE created and operated a number
of local NGOs that were able to serve as local coordinators for the TRO (Philipson and
Thangarajah 2005; Hogg 2006; Flanigan 2008). As a means of coordinating the activities of
a number of organizations, the LTTE established the Planning and Development Secretariat
(PDS) in 2004. The LTTE declared that the PDS would be “the pivotal unit that [would]
identify the needs of the people and formulate plans to carry out quick implementation with
the assistance of experts from the Tamil Diaspora” (Hogg 2006: 9).

123
The establishment of a social service infrastructure in Sri Lanka served two purposes
for the LTTE. First, these organizations provided the Tigers with a needed source of
revenue. The TRO and other LTTE-created NGOs became important revenue-generating
mechanisms for the LTTE. In addition to money donated by the worldwide Tamil diaspora
(La 2004; Wayland 2004), these NGOS would often engage in a process of overbidding on
service-provision contracts in order to send excess money to the LTTE (Flanigan 2008: 513).
Second, the provision of social services was clearly aimed at gaining support for the LTTE
within the local communities (Chandrakanthan 2000; Philipson and Thangarajah 2005). The
LTTE actively countered any local development initiatives that increased local dependencies
on and affinities toward any relief organization that was not LTTE-backed (Raheem and
Gosselin 2003; Philipson and Thangarajah 2005).
And yet, despite the large amount of institutional and administrative control in the
region, there was still an important level of local influence of the Sri Lankan government in
Colombo. Although some locations were exclusively under LTTE control, there were a
number of locations where Sri Lankan institutions continued to exist, creating parallel
institutional structures (Uyangoda 2005: 347). In addition, although some of the LTTE
structures of governance were paid for through their taxation practices, some of the bill for
these structures was actually paid for by the Sri Lankan government (Flanigan 2008: 511).

The Impact of the 2004 Tsunami
The successive waves of the tsunami on in December 2004 killed approximately
38,000 Sri Lankan citizens and left approximately 700,000 homeless (Wickramasinghe and
Fernando 2006). In one highly publicized story, a train traveling along the Sri Lankan coast

124
was literally swept off the tracks by one of the waves, killing approximately 2,500 people
(Uyangoda 2005: 347). In the Tiger-controlled areas in the northern and eastern parts of the
island, the tsunami is estimated to have killed over 9,000 people (Bennett et al. 2006).
As happened between the Indonesian government and the GAM, the tsunami
resulted in immediate conciliatory statements from the Sri Lankan government and the
LTTE. Yet, these statements quickly fell by the wayside as the two sides began to publicly
disagree on the control and management of disaster relief institutions. Again, similar to
Indonesia, the LTTE and the Sri Lankan government had different interpretations regarding
the motives of the rebels and the quantity and distribution of aid flow into Tamil regions of
the country.

The Battle over Control of Disaster Relief Institutions
Almost immediately following the disaster, the Tamil Tigers began to complain that
the government was discriminately distributing aid and purposefully keeping it from the
largely Tamil areas of the country. The government quickly sought to provide assurances
that there was no discrimination. Sri Lanka’s president, Chandrika Kumaratunga, attempted
to send an immediate message on this point, pointing out in a speech at a national mourning
service: “Nature has treated us all equally. Can’t we treat each other likewise?” (Steele 2005b:
2). However, LTTE forces continued to complain. One negotiator for the Tamil Tigers,
who was having little success in coming to an agreement with the Sri Lankan government on
aid management, stated: “We’re two-thirds of the casualties and damage, but the government
is creating roadblocks to us receiving aid” (Pocha 2005c: A12). The LTTE position on the

125
management of disaster relief flows was founded on the idea that it be considered an equal
partner with the Sri Lankan state (Uyangoda 2005: 344).
This debate continued in the early months following the tsunami. The Sri Lankan
government continued to insist that it was doing everything in its power to ensure that relief
aid reached the Tamil-controlled areas of the country. Any problems, according to the
government, were the result of the Tigers mismanaging the relief aid. The LTTE on the
other hand, rejected this claim, arguing that the Sinhalese government was doing everything
in its power to discredit them (Pocha 2005b). Rather than view the LTTE as an equal
partner, the government saw the post-tsunami context as one in which they could reassert
their sovereign control over the entire country, particularly given the reports of military
losses incurred by the LTTE during the tsunami (Uyangoda 2005: 344).
The Sri Lankan government attempted to end the war of words between the two
sides, and President Kumaratunga extended an invitation to the LTTE to participate in a
multi-party disaster management task force. However, the Tigers rejected this offer. The
political head of LTTE argued, “Since no constructive steps have been taken to help the
north, we believe it is a propaganda trick…We think she wants the international community
to believe that she is not discriminating against the north. Only a small amount of aid has
started arriving here compared to the relief pouring into Colombo airport” (Steele 2005c).
In keeping with the core demands that they be treated as an autonomous region, the
rebels continued to insist that international relief aid earmarked for Tamil regions of Sri
Lanka be allowed to be delivered directly to those regions, rather than get filtered through
the national government in Colombo. However, as several key countries in the international
community—including the United States and the United Kingdom—labeled the Tamil

126
Tigers a terrorist organization, the Sri Lankan government found it politically impossible to
meet that demand or any other which would appear to give the Tigers any autonomous
legitimacy. These tensions played out symbolically when the Sri Lankan government refused
to let Kofi Annan, the Secretary General of the United Nations, visit any of the areas in the
northeast part of the country in the days following the disaster (Pocha 2005a).
Despite these disputes among the leadership, on the ground the realities of disaster
relief provision were dictated by the government’s general lack of institutional capacity and
the LTTE’s extensive network of NGOs. The government was caught completely off guard
by the tsunami and was quite slow to react (Uyangoda 2005: 347; Bennett et al. 2006). In
addition, the political and bureaucratic culture of centralized decision-making (e.g., Matthews
1982) made it extremely difficult to assess and meet the needs of local communities, needs
which differed greatly depending on the pre-disaster level of infrastructure and development
and the actual extent of damage done by the tsunami (Uyangoda 2005: 348).
In this void, the institutional infrastructure of the LTTE allowed it to immediately
begin to manage the disaster relief processes and institutions in northeastern Sri Lanka as
well as coordinate with international aid agencies such as the United Nations Human Rights
Commission (Pocha 2005a; Bennett et al. 2006). The PDS released a “Needs Assessment
Report” within the first month following the tsunami. The report identified three phases of
relief and reconstruction tasks and argued for decentralized and local processes for meeting
disaster relief needs (Planning and Development Secretariat of the Liberation Tigers of
Tamil Eelam 2005: 39).
Rather than control these institutions directly, the LTTE utilized their preexisting
NGO structures. During the period after the tsunami, the TRO used their existing service

127
provision infrastructure to provide emergency shelter, establish and provide administration
within refugee camps, provide food and medical aid, and trauma counseling (Tamils
Rehabilitation Organization (TRO) 2005a, 2005b; Hogg 2006; Flanigan 2008: 511). In
addition, the LTTE pressured international NGOs working in the northern areas of Sri
Lanka to work directly through the TRO to coordinate aid flows and service provision
(Culbert 2005; Philipson and Thangarajah 2005; Amnesty International 2006; Bennett et al.
2006).
The LTTE used the opportunity to control these disaster relief institutions as a
means of increasing their own revenues and support within Sri Lanka. Flanigan notes that
virtually every NGO interviewed in her study was subject to some form of taxation on its
provision of goods and services within the region. She offers anecdotal evidence that close
to 50 percent of the disaster relief aid flowing into northeastern Sri Lanka through NGOs
after the tsunami may have been captured by the LTTE in the form of taxes (Flanigan 2008:
511). In addition, the LTTE recognized the potential benefits from cultivating the
appearance that disaster relief service provision was largely LTTE generated. The legitimacy
of the organization would be strengthened and the Tamil public would be more likely to be
supportive of their agenda (Philipson and Thangarajah 2005; Bennett et al. 2006; Flanigan
2008).
Indeed, many local citizens recognized and commented on the speed of the LTTE
disaster response compared to that of the Sri Lankan government (c.f., Pocha 2005b, 2005c).
The TRO was able to legitimize its claim to be able to provide services within a variety of
political and social conditions (Stokke 2006: 1029-30). Many in the international community
took note as well, as a number of international NGOs praised the LTTE for their role in

128
coordinating and providing rapid disaster relief in a situation in which the Sri Lankan
government was relatively ineffective (Bennett et al. 2006).
As the Sri Lankan government came to terms with the realities of LTTE control, it
became more flexible in its stance on the possibility of sharing the distributional
responsibilities with respect to international aid. By May 2005, the government showed a
willingness to sign off on a plan to allow the Tigers to control much more of its own aid
distribution. At the end of June, the two sides reached an agreement that allowed the LTTE
a greater role in the distribution of over $3 billion in aid. According to the aid-sharing
agreement, known as the Post-Tsunami Operational Management Structure (P-TOMS),
review committees formed of rebels, government officials and Muslims would be allowed to
recommend and monitor projects in areas hit by the tsunami (P-TOMS Agreement 2005).
The agreement created a “high-level” committee comprised of one government nominee,
one LTTE nominee, and one Muslim-party nominee. This committee was designated to
oversee relief and reconstruction processes throughout the affected areas. The P-TOMS
Agreement specified that decisions would be made on a consensus basis (P-TOMS
Agreement 2005: 3-4).
79

Many in the Sri Lankan parliament were opposed to the concession. They argued
that the pact threatened the country’s sovereignty by legitimizing the Tiger rebels.
Representatives from the Marxist People’s Liberation Front (JVP) party—long opposed to
providing any autonomy to the Tamils—withdrew from the governing coalition in protest

79
The P-TOMS Agreement also specified an opt-out provision if consensus could not be
reached: “In the event that consensus can still not be reached the nominating parties may,
after having followed the consultation procedure laid down in Section 5(g, i and ii) and after
having given 14 days notice, suspend the cooperation in the High-Level Committee” (P-
TOMS Agreement 2005: 4).

129
(Hogg 2006: 5). President Kumaratunga attempted to counter these objections, telling an
assembly of Buddhist monks: “Some sections question how a sovereign government could
have dealings with a terrorist organization...but why can’t we give them [Tigers] an
opportunity to reform themselves? Peace is right before us to see. It will be a major blunder
if we don’t take this opportunity just because a few oppose it” (Bedi 2005).
The agreement, however, was short-lived. In early July 2005, the Sri Lankan
Supreme Court issued an injunction on the P-TOMS agreement, following a petition by the
JVP arguing that the aid sharing deal with the rebels as unconstitutional. The basis for their
argument was that the LTTE was a terrorist organization that did not possess any legal
legitimacy; therefore it was illegal to locate P-TOMS headquarters in a rebel-controlled area
(Sengupta 2005b; Hogg 2006: 5). In the weeks following the injunction, the Sri Lankan
government did little to resurrect the deal. The failure of P-TOMS and the widespread view
that the government was covertly but directly involved in scuttling the deal (e.g., Kaldor
2005) had a negative impact on the LTTE’s willingness to participate in the peace process
moving forward (Hogg 2006: 5).

2005: Return to Civil War
On August 15, 2005, the Sri Lankan Foreign Minister, Lakshman Kadirgamar, was
assassinated at his home while getting out of his swimming pool. Immediately, the Sri
Lankan government accused the LTTE of being responsible for the killing. The Tigers
denied any involvement, pointing the blame instead at a faction within the Sri Lankan
establishment that wanted the 2002 ceasefire to fail. However, it is clear that Kadirgamar
presented an opportune target for the Tigers. Supporters of the LTTE had denounced

130
Kadirgamar, the highest-ranking ethnic Tamil in the government, as a traitor. It was largely
his international campaigning that had led to the Tigers being labeled as a terrorist
organization in the United States, the United Kingdom, and elsewhere (Huggler 2005).
As Sri Lanka moved toward national elections in November 2005, the violence
increased. Violations of the ceasefire agreement were commonplace, and assassinations
became an almost daily occurrence (McDougall 2005; Sengupta 2005a; Amnesty
International 2006). Norwegian officials monitoring ceasefire violations documented 190
killings between February 2005 and the end of October 2005, an average of more than five a
week. By comparison, during all of 2004, they counted sixty killings (Sengupta 2005a). The
elections came and went without the explosion of violence that was anticipated; however,
the results of the elections created even more tension and a wider sense of discouragement
within the country, as the new Sri Lankan president, Mahinda Rajapakse, expressed a much
firmer stance on the LTTE key demand for independence (International Crisis Group
2006b).
Whereas past talks between the government and the Tigers had focused on giving
the group some autonomy within their region—a type of federalism as a possible
compromise solution—Mr. Rajapakse insisted that former governmental negotiators had
ceded too much to the rebels. In an angry response after the presidential election, Velupillai
Prabhakaran, the Tigers’ leader, threatened a return to war if no political settlement emerged
within the next year. The LTTE reinforced its position with a show of power that included
the orchestration of an election boycott in November, and a walkout by government
employees, students, and shopkeepers in December 2005 (Steele 2005a). The violence
continued to build, and by the end of December, the LTTE had provoked the Sri Lankan

131
into full-scale conflict, once again through a series of targeted terrorist attacks and
assassinations (International Crisis Group 2006b).

Analysis and Implications for Theory
In chapter 2, this project posited 3 hypotheses for the relationship between natural
disasters and the probability of post-disaster violence. The first and second hypotheses
argued that the ability to capture disaster relief institutions in a post-disaster context should
have important implications for the likelihood of post-disaster violence. The third
hypothesis argued that lower levels of institutional quality create incentives to engage in
these battles over institutional control; thus, relative levels of pre-disaster institutional quality
affect the likelihood of post-disaster violence because. The quantitative analysis in chapter 4
offered preliminary support for this third hypothesis.
However, on the surface, the cases of Sri Lanka and Indonesia in the aftermath of
the 2004 tsunami presented some empirical challenges for this argument. As discussed in
chapter 4, both countries had relatively low measures of institutional quality prior to the
disaster;
80
yet, the dynamic levels of violence seemed to go in opposite directions in the year
following the tsunami. Indonesia negotiated a peace agreement with the GAM rebels while
the LTTE and Sri Lankan government reignited a new round of civil war. This suggests a
need to delve more deeply into these crucial cases in an attempt to explore the dynamics
associated with hypotheses 1 and 2. The structured, focused comparison presented here

80
Both countries scored below the Freedom House “basic standard of effective
performance” in “Rule of Law” and “Anti-corruption and Transparency.” See chapter 4 for
details.

132
demonstrates that while the quality of institutions in the pre-disaster moment surely matters
for incentives toward post-disaster attempts at capture, the actual outcome of those attempts
is truly what impacts the likelihood of post-disaster violence.

Questions 1 and 2: The Pre-Disaster Conflict Status and Institutional Control
In Indonesia, the TNI controlled many of the local institutions in Aceh in the time
prior to the disaster. Many of the institutional gains that had been made by the GAM during
the Humanitarian Pause in 2002 were wiped out during the martial law and civil emergency
periods of 2003 and 2004. By December 2004, the TNI had restored a number of local
structures of governance and were providing development services in a way that assured the
Acehnese were separated from the GAM as much as possible. In addition, the TNI
reignited a relatively dormant military campaign during the period of martial law, ensuring
that at the time of the tsunami, much of the critical GAM leadership was either hiding in the
mountains of central Aceh or in prison.
Alternatively, in Sri Lanka the LTTE displayed a much stronger level of pre-disaster
institutional control. The rebels had established a variety of processes of governance within
the key Tamil regions of northeast Sri Lanka, including a system of taxation, a judicial
system, a police system, and an NGO-based system for the provision of a variety of social
services. Despite the fact that the Sri Lankan government was contributing to the funding
of some of these LTTE-controlled institutions, most of the parallel structures of Sri Lankan
governance were crowded out.

133
Questions 3 and 4: The Impact of the 2004 Tsunami and Post-Disaster Battles over
Institutional Control

The fact that the TNI controlled many of the local institutions as the time of the
tsunami made it sensible for the Indonesian government to turn over control of the aid
distribution processes and institutions to the TNI in the months after the tsunami. The fact
that the military controlled many of the disaster relief processes and institutions made it
relatively easy to physically suppress the GAM’s attempts to provide goods and services,
discredit any of the GAM’s successes as not based on altruistic motives but on attempts to
improve its power base as it had done during previous ceasefire arrangements, and finally
utilize any attempts by the rebels to gain institutional control as excuses to not merely
prevent institutional capture but to attack, further decimating their remaining power base.
In Sri Lanka, the LTTE was able to capitalize on its existing network of local
institutions and NGOs. The TRO, in particular, became a useful tool for the Tigers, as it
pressured a number of international NGOs to distribute aid and disaster relief within its
preexisting development and social infrastructure, namely regional and local NGOs. The
PDS conducted a post-disaster needs assessment for all northeastern Sri Lanka in the first
month after the disaster. The TRO was able to coordinate the execution of much of this
plan within its territories, providing emergency shelter, food and medical supplies, and
assisting in the immediate reconstruction of critical infrastructure.

Question 5: The Effects on Post-Disaster Conflict
In both of these cases, the ability to maintain institutional control in a post-disaster
context brought the controlling groups needed material and nonmaterial gains. In
Indonesia, the military was able to use its control of aid distribution to continue its practices

134
of local revenue generation. Control over disaster relief aid offered a number of
opportunities for the TNI to funnel projects toward its own profit-generating businesses. In
addition, the TNI was able to use its close contact with the international NGO community
in order to discredit any rebel attempts at institution capture. The TNI insured that the
international NGO community and the GAM remained physically separated, denying—or
making extremely difficult—international access into areas that were known GAM
territories. Finally, the TNI was able to use the post-disaster as an opportunity to conduct
offensive operations aimed at further destroying the GAM’s capabilities, again done
rhetorically in the name of keeping the GAM from capturing disaster aid. All of this had the
effect of seriously weakening GAM’s ability to be involved in any institution capture. Any
control that GAM had of local institutions, built up for a time during the Humanitarian
Pause, was destroyed. With it, their ability to credibly signal their functional ability to govern
Aceh as an autonomous territory was severely damaged as well, contributing to a context in
which they had little choice but to approach the peace negotiations in August 2005.
In Sri Lanka, the LTTE was able to capitalize on its capture, gaining material and
non-material benefits. As aid flowed into the region, the ability to direct the flows of aid and
channel it through their own LTTE-influenced (if not, controlled) organizations allowed the
Tigers to siphon aid money for their own purposes. It gave them direct control of revenue
streams and ensured that the most damaged pieces of their own LTTE infrastructure could
be rebuilt most quickly. In addition, post-disaster institutional capture gave the rebels the
opportunity to send a signal of their functional ability to govern. There is evidence that the
rebels recognized this opportunity, used their disaster relief organizations to foster it, and

135
that the signal was sent successfully, as a number of Tamil citizens and international aid
organizations noted the LTTE success.
The government had little choice but to recognize these realities and agree to an
arrangement that further institutionalized the sharing of disaster relief: P-TOMS. When the
Sri Lankan Supreme Court declared the P-TOMS agreement unconstitutional, the LTTE was
in a relatively good position. Their access to resources and general (but by no means
universal) good standing within the country’s Tamil population made the renewal of full
scale violence in the name of rebellion and separatism more feasible.

Implication for Hypotheses
The hypotheses generated in chapter 2 argue that the ability to capture disaster relief
institutions in a post-disaster context should have important implications for the likelihood
of post-disaster violence. The principle mechanisms through which this occur are that of
grievance and feasibility, which are widely accepted to have important effects on the
probably of intrastate conflict more generally. The relative ability for governments to
provide the disaster relief services they are expected to provide should affect the level of
grievance within the country. At the same time, the material and non-material benefits that
accrue to those that are able to capture disaster relief institutions—even marginally—should
have important effects on the opportunities for rebellion.
The comparison in this chapter is generally supportive of this story of the
relationship between disasters and intrastate conflict. Clearly, in both Indonesia and Sri
Lanka, control of disaster relief institutions established a critical part of the context in which
peace and war emerged respectively. In Indonesia, military control of these institutions gave

136
them the continued source of local revenue—historically important as a means of financing
their own activities—and allowed them the space to discredit any of the GAM attempts to
engage in the provision of goods and services to the Acehnese population. In Sri Lanka, the
LTTE’s control of institutions gave them needed revenue streams and the ability to signal
their functional capabilities. In both cases, there is evidence that the particulars of
institutional control affected both grievances and feasibility, thus contributing to the context
in which peace or war emerged.
This idea of “contributing to context” is important, here as, despite their noteworthy
pre-disaster similarities, there are a number of important differences in these two cases that
also may have contributed to their divergent outcomes. Among other important differences,
these countries had two different histories with respect to their political systems. Indonesia
was relatively new to the processes and effects of democratization after the 1998 ousting of
Suharto. At the time of the disaster, Sri Lanka had been exposed to a variety of levels of
democracy and democratic processes throughout its history. Each of these differences may
have contributed to the outcomes, particularly in Indonesia where the institutions associated
with the military and TNI were important vestiges of the preexisting dictatorships of
Sukarno and Suharto.
81

81
Historical experiences with different political systems comprise one set of other factors
that may have contributed to the context of different post-disaster levels of violence in these
two countries. Other noteworthy differences between these two cases included: the religious
and ethnic diversity within the separatist region (particularly Sri Lanka), the historical ability
of the separatist group to carry out violence in the capital city (high in Sri Lanka, low in
Indonesia), and the post-disaster role of the U.S. Navy (and other international military
forces) in Indonesia but not Sri Lanka. Future work will be needed to compare these factors
in order to make any credible assessments of causal weight in the creation of the context for
changes in post-disaster levels of violence. In addition, future work that expands the case set
to include a greater diversity of cases will be helpful in teasing out these potentially
competing factors.

137
Finally, it is important to note that the respective groups’ institutional control was
already set at the time of the disaster. In both Indonesia and Sri Lanka the TNI and the
LTTE were largely in control of institutions prior to the disaster itself. Instead of the
tsunami creating the space to battle over institutions, the disaster seems to have created a
“lock-in” effect, whereby pre-disaster institutional control got cemented and in part legalized
in the post-disaster period. The pre-disaster control, in both cases, created high capacity to
handle this massive civil emergency, especially relative to each group’s opposing party. In
Indonesia, the GAM had virtually no capacity at the time of the disaster to assist in relief
services. Likewise in Sri Lanka, the LTTE controlled virtually all of the northeastern part of
the country, contributing to the Sri Lankan government’s low capacity in these areas.
Thus, in the immediate aftermath of the disaster, domestic and international relief aid
naturally flowed through the pre-existing mechanisms and institutions that possessed the
necessary capacity. In this way, the disaster itself seems to have not drastically changed the
institutional context, but locked-in the pre-disaster status quo. This suggests that disasters
might be more likely to push internal conflicts in the direction that they were already headed
at the time of the disaster than they would be to drastically change the internal political
dynamics and mechanisms.

Conclusion
Closer comparison of the cases of Indonesia and Sri Lanka yields support for the
notion that institutions matter in the aftermath of natural disasters, particularly when they
occur in conflict zones. In particular, disasters seem to have a “lock-in” effect, cementing
institutional control of the groups (the government or the rebels) who had it at the moment

138
that the disaster occurs. In this way, it seems that natural disasters are not likely to drastically
change a conflict situation but instead push it more rapidly in the direction it was heading
anyway in the time prior to the disaster. The concluding chapter reviews the arguments and
findings of this project and discussed their implications for future research and policy on the
relationship between disaster and civil conflict.

139
Chapter 6: Conclusion
As a means of conclusion, I begin by briefly reviewing the project and restating its
major findings. Following this, the theoretical and empirical significance of the findings will
be restated in light of the questions asked in the introductory chapter, and these will be
discussed for their theoretical and policy-making implications. I end with some
consideration of the future research that might flow from this project.

The Project and its Findings
This dissertation attempted to answer the question: How do the political effects of
natural disasters contribute to the likelihood of post-disaster violence? After many years of
research that has downplayed the importance of the politics associated with natural disasters,
recent work has analyzed disasters as events that could have important political effects. In
addition, recent cases of natural disasters occurring in zones of conflict have spawned a rash
of commentary on the possibility of disaster diplomacy—the notion international or
intranational peacemaking might be more likely in a post-disaster environment.
Yet, very little of this new disaster diplomacy research has developed and
comparatively tested a theoretical mechanism through which this might occur. In addition,
there are a variety of reasons why the post-disaster environment is likelier to be more
violent, not less. In these relatively uncharted waters, a better understanding of the cases of
Indonesia and Sri Lanka following the 2004 Indian Ocean tsunami becomes crucial. In the
year following the tsunami, Indonesia was able to bring its 30-year separatist conflict to a
relatively peaceful (and thus far, stable) conclusion with the negotiation of a peace agreement
with the GAM rebels in August 2005. Alternatively, systematic violence reemerged in Sri

140
Lanka in 2005, with the government in Colombo clashing with the LTTE and reigniting a
civil war that continued over the next three and a half years. In many respects, these crucial
cases are ideal for theory testing.
In an attempt to gain some traction on the possible political mechanisms that may be
impacted by natural disasters, I began this project by reviewing the recent literature on
violent civil conflict. Our current understanding of the processes that affect the likelihood
of violent civil conflict presents three important and disaster-relevant ideas. First, the
likelihood of civil war rests as much on the rebels’ or insurgents’ ability to be successful as it
does any particular mix of grievances. Rebellions must be feasible. Therefore, it is
important to understand the micro-foundations of rebel groups’ attempts to gain sources of
funding, recruit new members, and gain nominal support from citizens and outside parties.
In this dynamic, the role of opportunity costs become important, as individual calculations
about whether to participate in rebellion include the costs associated with forgoing any other
economic opportunities or forms of political expression. As these opportunity costs are
lower, rebel groups are likely to have an easier time gaining the human and monetary
resources that are necessary for a rebellion to be feasible.
Second, changes in political institutions—and particularly, institutional quality—are
likely to impact the probability of violent civil conflict. The effects of these changes impact
and alter levels of grievance and opportunity structures. As institutions change for the
worse, or as institutional quality decreases, there becomes a lack of good alternatives for
redressing grievances. This is likely to make grievances worse. It is also likely to lower the
opportunity costs associated with rebel recruitment of monetary and human support. In
addition, lower quality institutions have been associated with lower-levels of long-run

141
economic development, contributing to a lack of economic alternatives, which serve to
lower the opportunity costs associated with gaining rebel support and resources. As the
opportunity costs of rebel recruitment decrease, the rebellion becomes more feasible and
violent civil conflict more probable.
Finally, while the feasibility of rebellion is important, recent literature argues that the
state plays an important role in contributing to violence as well. Specifically, the strategic
interaction between states and rebels is critical in determining whether contextual factors
that may contribute to the likelihood of violence are actually translated into violent
interaction. In this respect, governments are careful to build the anticipated risks and costs
of future challenges into their calculations of how they deal with insurgencies. On the other
hand, the insurgents’ calculations are based on a similar logic, as they try to balance their
strategic options in a way that maximizes their ability to gain support and resources from the
population while at the same time maximizing their ability to send costly signals to their
government opponents.
Each of these three factors is likely to be impacted by the shock of a natural disaster,
particularly one that is rapid onset. The effects of natural disasters tend to be worse in areas
that are relatively under-developed, and as such, they represent a point in time where
institutional quality should matter most. Shocks in underdeveloped areas are likely to have
significant impacts on the opportunity structures that impact decisions on rebellion and are
likely to heighten the strategic interaction between states and pre-existing rebel groups.
In combination, these three factors form an important basis for a plausible theory of
post-disaster violence. Disasters present a moment in which the institutions of disaster relief
may be susceptible to “capture.” The ability to capture institutional control results in both

142
material and non-material benefit, which impact levels of grievance and opportunity and
therefore, have important implications for civil violence.
Given these impacts, this project developed three hypotheses. First, post-disaster
violence is expected to be more likely when rebels are able to gain marginal control of
disaster relief institutions. Second, post-disaster violence is less likely when states are able to
maintain control of disaster relief institutions. Finally, given the role of institutional quality
in the theory, post-disaster violence is more likely where pre-disaster measures of
institutional quality are relatively weak.
Using time series, cross-sectional data, including individual measures of institutional
quality—corruption, the rule of law and order, and bureaucratic effectiveness—as well as an
additive index, I found that countries with weaker levels of institutional quality a priori a
rapid-onset disaster are more likely to experience the onset of violent civil conflict in the year
following a disaster. Countries with relatively higher levels of institutional quality a priori a
rapid-onset disaster are less likely to experience the onset of violent civil conflict in the
following year.
However, given that the cases of Indonesia and Sri Lanka were part of the primary
motivation for the project but are not included in the time series data, it is important to
analyze whether these cases fit the quantitative findings. A preliminary investigation of the
cases at the end of chapter 4 yielded some complex results, particularly with respect to
Indonesia. One the one hand, the Indonesia case seems to provide further confirming
evidence for the theory, as the Uppsala/PRIO dataset codes 2005 as a conflict year.
However, this data point needs to be balanced with the reality that perceptions within

143
Indonesia—of both GAM rebels and the Indonesian government—are that the tsunami
played an important role in setting the stage for a peace agreement in 2005.
Thus, chapter 5 presented a systematic comparison of the cases of Sri Lanka and
Indonesia following the 2004 tsunami. Using the method of structured, focused
comparison, I analyzed the pre-disaster levels of institutional control in the separatist
regions, the ways in which the tsunami changed the incentives for either the GAM or the
LTTE to further seize control over institutions by “capturing” the processes of disaster
relief, the extent to which these groups were successful, and the implications of this relative
success for the ongoing patterns of conflict or peacemaking in the year following the
tsunami. The cases of Indonesia and Sri Lanka demonstrated that although low institutional
quality at the time of the disaster may contribute to post-disaster battles over institutional
control, they do not, in these cases, affect the outcomes of these battles. Ultimately, it is the
relative success of the separatist groups in these cases that contributes to post-disaster peace
or violence.

The Significance of the Findings and Future Work
In the introductory chapter, I argued that this project contributed to the budding
research on disaster diplomacy as well as to the literature and policy-making around
intrastate violence more widely, beyond the context of natural disasters. In light of the
findings in the project and as a means of reinforcing this argument, I conclude by pointing
to a number of interesting implications that this project has for future research.

144
Theoretical Implications for the Study of Violent Civil Conflict
The study of intrastate violence in the context of a natural disaster allows us to
analyze the way in which relative changes in political economies associated with institutions
have important implications for violence. In this respect, two important implications emerge
from the findings that institutional capture carries material and non-material benefits. The
first implication is the need to expand the empirical work to more consistently reflect the
role of institutional quality. The second implication is the need for more theoretical and
empirical work on the strategic role of signaling in intrastate conflicts.
First, institutional quality must continue to be studied and the quantitative data-set
expanded, using a greater time series and alternative measures of quality in order to check
the robustness of these findings. The ICRG institutional quality measures used in this
project can be substituted using a variety of similar measures developed by Freedom House,
Transparency International, and the World Bank.
In addition, recent additions to the Uppsala/PRIO data set offer a stronger empirical
basis on which to evaluate the disaster-conflict nexus. Using Geographic Information
Systems (GIS), much of the data on violent civil conflict has recently been disaggregated
from the state level and given a specific geographic location using either a fixed center point
for the conflict (Buhaug and Gates 2002) or by sub-dividing a region into 100 km x 100 km
geographic squares in order to assess specific geographic locations (Buhaug and Rød 2006).
This new data has the potential to solve a number of ecological inference problems
that have plagued the quantitative study of civil war and violent civil conflict. Sub-state data
that is geographically specific allows for a better understanding of the relationship between
violent civil conflict and variables that are location specific such as access to natural

145
resources, distances to borders (Buhaug and Gates 2002; Buhaug and Rød 2006), distances
to other conflicts (Buhaug and Gleditsch 2008), and the physical distance between
disadvantaged ethnic groups and the center of local power (Buhaug, Cederman, and Rød
2008). Future extensions of this project will involve a more rigorous mapping of the
geographic overlaps of conflict and disaster in order to produce a more exact and
generalizable empirical analysis.
However, further case study work is also necessary. This study suggests that
although intrastate conflicts—and particularly separatist ones—are often about a battle to
control the large institution of government in a country or region, there are many battles
over smaller institutions that occur constantly throughout an ongoing conflict. As different
sides control these institutions at different points in time given different contexts, control
over these smaller institutions defines and re-defines the micro-foundations and often the
dynamic evolution of intrastate conflicts. In this respect, scalar measures of institutional
quality and dichotomous quantitative measures of intrastate violence—while very useful in
yielding important empirical and theoretical insights—will only take our understanding of
intrastate conflict so far. Intrastate violence needs to continue to be studied both as discrete
episodes of onset, duration and cessation as well as dynamic processes of conflict.
Second, this project contributes to the idea that signaling plays an important role in
the strategic aspects of political violence. Signaling games have been used throughout
political science and economics to explore strategic interaction that occurs when both
players have incomplete information. Neither player can be sure about the other player’s
preference ordering with respect to possible outcome of the game; thus, the players use
information communicated both directly and indirectly throughout the game in order to

146
update their individual beliefs.
82
Such games have been used to study the initiation of
international conflict (Fearon 1995; Powell 1999, 2002), the strategic aspects of deterrence
(Zagare and Kilgour 2000), the end of the Cold War (Kydd 2000), among a number of other
areas of international politics (Walsh 2007).
The literature on civil wars and intrastate conflict has only recently begun to
incorporate strategic concerns more broadly (e.g., Walter 2006a, 2006b). Both governments
and rebels or insurgents act based on perceptions of what the other party has done in the
past and what it is likely to do in the future. On the one hand, governments tend to think
more about the future than current theories might anticipate, and they are careful to build
the anticipated risks and costs of future challenges into their calculations of how they deal
with insurgencies (Walter 2006a: 324). On the other hand, the insurgents’ calculations are
the result of “a complex strategic interaction where governments are actively seeking to deter
separatists, and separatists are carefully seeking to uncover if and when the government will
grant concessions” (Walter 2006b: 106).
Throughout the signaling literature—both domestic and international—the
information problem at the heart of the game is most often mitigated through the use of a
costly signal. This notion goes at least as far back as Schelling, who recognized that a costly
signal was a much more effective way of communicating preferences than “cheap talk”
(Schelling 1960[1980]). Talk itself is often considered cheap because groups—whether
states, sub-state actors, or trans-national actors such as terrorist groups—have incentives to
disguise their beliefs and preferences. Doing so may cause the opposing group to choose a

82
Often this belief updating is modeled using Bayes’ Rule. For an overview of the formal
representation of signaling games in political science, see Morrow (1994) and McCarty &
Meirowitz (2007) among others.

147
sub-optimal strategy based on misinformation. Because of the incentives to verbally
misrepresent preferences, costly action is considered a much stronger signal (Kydd 2005;
Kydd and Walter 2006).
Costly action is most clearly defined as an action that would not necessarily be taken
by a group attempting to misrepresent itself (Riley 2001). In an international context, force
mobilization, going to war, or undertaking some action that “ties hands” with respect to
strategic options later are all examples of this type of costly signaling (Fearon 1994, 1997;
Powell 2002, 2004). In the space of domestic violence, the concession of territory or other
military gains can serve as a costly signal that can contribute to the success of power-sharing
provisions in peace agreements (Jarstad and Nilsson 2008). Finally, terrorist groups use the
tactic of terrorism as a specific form of costly signaling that can contribute to either public
support or fear as well as the enemy’s (often the state’s) belief that terrorists are serious in
their commitment and capabilities (Kydd and Walter 2006).
But in the context of rebellion or insurgency, costly signaling need not involve
violence or the threat of violence (Stephan and Chenoweth 2008). Sub-state groups—
particularly ones underpinned by a separatist agenda or a goal of control of the apparatuses
of the state—can send peaceful costly signals that serve to communicate the seriousness of
their claim and their relative capability. Often, these groups attempt to provide goods and
services to the population at costs that would not be borne by a group merely operating on a
strategy of deception. The recent examples of Hamas in the Palestinian territory and
Hezbollah in Lebanon provide anecdotal support for this strategy. Hamas’ rise to power in
Palestine has been largely built on its ability to provide critical services to Palestinians.
Hezbollah is currently engaged in reconstruction projects in Beirut neighborhoods that were

148
hit hard by Israeli bombing campaigns in their 2006 conflict. Each of these groups is
utilizing the provision of goods and services as a means of signaling for support amongst
their domestic audiences.
Given that this nonmaterial signaling benefit had important implications in Sri Lanka
and Indonesia, how does this logic work outside the context of a natural disaster? More
specifically, what are the strategies through which insurgent groups signal for support
domestically and internationally and how do they balance these strategic choices with the
need to send costly signals to their opponents, presumably governments? All types of
signaling need to be more completely understood and modeled with the aim of a better
understanding of the strategic rationales of intrastate violence.

Significance for Disaster Diplomacy and the Policies of Disaster Relief Aid in
Conflict Zones

The finding that natural disasters help lock-in and even accelerate the political
processes already in place in the time before the disaster is consistent with the budding
thinking on disaster diplomacy (e.g., Kelman and Koukis 2000; Ker-Lindsay 2000; Kelman
2006). However, future work must begin to go beyond these cases. There may be
circumstances where natural disasters have a positive impact on violent civil conflict or at
least provide an impetus for positive social change. The 1970 earthquake in Peru and the
series of earthquakes that rocked Mexico City in 1985 among others are reminders of the
fact that any future work should continue to remain open to the possibility of alternative
empirical outcomes.
In addition, the cases of Indonesia and Sri Lanka are, of course, internal, separatist
conflicts. However, natural disasters have often occurred in areas that affect two

149
neighboring countries that have a preexisting rivalry.
83
They can also occur in one country
that finds itself in conflict or rivalry with a country (or countries) that is not necessarily a
neighbor.
84
Although the aspect of the hypothesis that deals with rebels’ functional ability to
govern disputed territory may not translate to international issues, the aspect of signaling as
it pertains to the bargaining relationship between two or more parties in conflict could
certainly be assessed. Does the promise and/or delivery of relief aid send any kind of
credible signal? If so, how do both parties—the donor and recipient of relief aid—factor
these signals into their strategic calculations regarding the conflict in the aftermath of the
disaster?
However, to the extent that the preliminary hypothesis suggested by these cases
holds up to further comparative testing, several policy implications are clear. First, the policy
implications for each of the parties to the conflict, whether that conflict is intrastate or
interstate, are clear. Neither group should assume the period after a disaster to be one where
the prior conflict should be put completely out of the minds of policy makers. The period
after a disaster could create an entrenchment of a bad situation, if one existed prior to the
disaster. Given the importance of signaling in conflict bargaining, and each party to the
conflict should take care to view it as such.

83
As an example, the aforementioned 1999 case of Turkey and Greece might be compared
to the 2005 earthquake in Kashmir and how it affects the ongoing tension between India and
Pakistan.
84
A prominent recent example in this regard would be the 2003 earthquake in the city of
Bam, Iran. For a case-study exploration of the way in which this disaster impacted the
ongoing tensions between Iran and the United States and in turn how those tensions might
have impacted international disaster relief see Enia (2006).

150
Related, the second policy implication of the different outcomes in Sri Lanka and
Indonesia is that parties should not assume that the opponent in the conflict is completely
ignoring the conflict issue in order to focus on relief and reconstruction. Strategically, one
must assume that the opponent party has similar information and faces similar incentives
regarding the potential material and non-material benefits of institution capture following
disasters. Despite any rhetoric that may occur about “putting aside differences in order to
focus on the disaster at hand” all parties must be aware that opposing parties have incentives
to engage in activities that provide relief and reconstruction that are not entirely, if at all,
altruistic. Any one party’s decision-making calculus must include an awareness of the fact
that all parties face these incentives.
The United States-Iranian interaction in the midst of the 2003 Iranian earthquake in
the city of Bam is instructive on both of these policy points. Even though the earthquake
killed over 25,000 Iranian citizens and left over 75,000 homeless, the contentious nature of
the relationship between Iran and the United States created serious doubt that U.S. relief
assistance would be forthcoming. Both parties to the ongoing conflict seemed to be aware
the disaster presented particular opportunities to gain leverage (Enia 2006).
Iran, in particular, seemed to have this logic in mind when it rejected the offer for a
U.S. humanitarian mission led by senator and former head of the American Red Cross
Elizabeth Dole. The leadership in Iran argued “it was too preoccupied in dealing with the
aftermath of the devastating Bam earthquake and it would have been inappropriate at that
time to confuse it with another issue” (Iranian Daily Says Ignoring the U.S. Would Be
'Hugely Unwise' 2004). Despite the fact that U.S. rhetoric conveyed solely humanitarian
motives, the Iranians seemed to base their decision on the fact that there could be other

151
results to the interaction that might affect Iran’s relative bargaining power. Iranian President
Khatami stated, “Basically, you must not link the issue of the earthquake with chronic and
deeply-rooted political issues” (Mohammed Khatami 2003).
Finally, those outside the conflict should also be aware that any actions they take to
assist in the post-disaster relief and reconstruction may have important consequences for the
capabilities of one or more parties to the conflict. Any action taken to assist a disaster-
stricken region (be it a state or otherwise) may send powerful signals to the opposing side of
the preexisting conflict that may affect conflict. One state assisting another may, for
example, give the appearance of an alliance between the two that could extend beyond the
disaster relief issue. In other circumstances, a state or other external actor working with
separatist groups in a disaster-stricken region may send a signal that it considers the
separatist group’s claim to regional sovereignty somewhat valid. Many other examples seem
possible in this regard. In this sense, the goal of maintaining complete neutrality with
respect to the conflict while providing outside relief and reconstruction assistance seems to
be naïve.
In this regard the United States needs to think carefully about its disaster relief and
reconstruction policies. Much of the global scientific community is predicting that the
severity (and possibly quantity) of weather-related natural events will increase because of the
current pace of global climate change (Intergovernmental Panel on Climate Change 2007).
In addition, despite the fact that the global rate of earthquakes has been reasonably constant,
scientists are predicting that pace could increase as a result of the melting of the polar ice
caps (e.g., Ekstrom, Nettles, and Tsai 2006).

152
Some of the world’s major geographic fault lines run through countries like Pakistan,
which experienced an earthquake in 2005 that killed close to 80,000 people. Given
Pakistan’s current level of domestic political unrest, and the state of general lawlessness in
the northwest tribal areas, another major natural disaster could have severe consequences for
conflict in that country. In fact, the 2005 earthquake and the slow pace of organized relief
assistance generated a great deal of domestic hostility (Coll 2005).
The United States would be wise to think through the consequences that provision
of external disaster relief may have in similar situations. How would, for example, the
Pakistani or those in the tribal areas view international aid that is funneled through the
Pakistani government? In what ways would the processes of disaster be subject to
institution capture by groups like the Taliban? Could their ability to do so set a context for
greater violence in the region in a post-disaster context? Might any of these events lead to a
collapse of the Pakistani government? Or is there a way to provide external relief assistance
without contributing to further instability?
It seems imperative that diplomats, academics, and journalists be more careful with
the way they use the term disaster diplomacy. Too often, the hope has been expressed as if
peaceful interaction will simply materialize out of the sheer scope or intensity of the
destruction. This project suggests that the forces at work—although intimately related to the
specific context presented by the disaster—are indeed, much more complicated. Even so, if
there is any possibility that some peaceful diplomacy can arise from these dire situations, it is
critical that the policy-making community has a clear and well-informed understanding of
both the context and the mechanisms through which this process works. In this spirit, more
comparative research seems to be in order.

153
Bibliography
Abney, F. Glenn, and Larry B. Hill. 1966. Natural Disasters as a Political Variable: The
Effect of a Hurricane on an Urban Election. American Political Science Review 60 (4):
974-981.
Acemoglu, Daron, Simon Johnson, and James Robinson. 2001. The Colonial Origins of
Comparative Development: An Empirical Investigation. American Economic Review 91
(5): 1369-1401.
Aglionby, John. 2005. Legacy of Tsunami Brings Peace to Aceh. Guardian, August 15.
Agreement on Ceasefire. 2003. Agreement on a Ceasefire between the Government of the
Democratic Socialist Republic of Sri Lanka and the Liberation Tigers of Tamil
Eelam.
Aguswandi, and Judith Large. 2008. Aceh's New Era: An Interview with Irwandi Yusuf. In
Reconfiguring Politics: The Indonesia-Aceh Peace Process, edited by Aguswandi and J. Large.
London: Conciliation Resources.
Akcinaroglu, Seden, Jonathan M. DiCicco, and Elizabeth Radziszewski. 2008. Avalanches
and Olive Branches: Natural Disasters and Peacemaking between Interstate Rivals.
Paper presented at the Annual Meeting of the Peace Science Society, October 24-26,
Claremont, CA.
Albala-Bertrand, Jose-Miguel. 1993a. Natural Disaster Situations and Growth: A
Macroeconomic Model for Sudden Disaster Impacts. World Development 21 (9): 1417-
1434.
———. 1993b. The Political Economy of Large Natural Disasters: With Special Reference to
Developing Countries. New York: Oxford University Press.
Alexander, David. 1997. The Study of Natural Disasters, 1977-97: Some Reflections on a
Changing Field of Knowledge. Disasters 21 (4): 284-304.
Allen, Michael. 2008. Burma Cyclone a Catalyst for Change?
http://www.demdigest.net/blog/9/test.html (accessed 15 May, 2008).
Amnesty International. 1993. Shock Therapy: Restoring Order in Aceh, 1989-1993, Amnesty
International, London.
———. 2006. Sri Lanka: Amnesty International Condemns Killings of Civilians ASA
37/014/2006, Amnesty International.

154
Anand, Sudhir, and Martin Ravallion. 1993. Human Development in Poor Countries: On the
Role of Private Income and Public Services. Journal of Economic Perspectives 7 (1): 133-
150.
Andrienko, Yuri, and Louise Shelley. 2005. Crime, Violence, and Political Conflict in Russia.
In Understanding Civil War: Evidence and Analysis, edited by P. Collier and N. Sambanis.
Washington, DC: The World Bank.
ASNLF. 1976. Declaration of Independence of Acheh–Sumatra.
http://www.asnlf.net/asnlf_int/politics/declaration41276ing.htm (accessed 11
November, 2008).
Aspinall, Edward. 2000. Whither Aceh? Inside Indonesia 62.
———. 2007. The Construction of Grievance: Natural Resources and Identity in a
Separatist Conflict. Journal of Conflict Resolution 51 (6): 950-972.
Aspinall, Edward, and Mark T. Berger. 2001. The Break-up of Indonesia? Nationalisms after
Decolonisation and the Limits of the Nation-State in Post-Cold War Southeast Asia.
Third World Quarterly 22 (6): 1003-1024.
Aspinall, Edward, and Harold Crouch. 2003. The Aceh Peace Process: Why It Failed. Policy
Studies 1, East-West Center, Washington, DC.
Auvinen, Juha, and E. Wayne Nafziger. 1999. The Sources of Humanitarian Emergencies.
Journal of Conflict Resolution 43 (3): 267-290.
Awaluddin, Hamid. 2008. Why Is Peace in Aceh Successful? In Reconfiguring Politics: The
Indonesia-Aceh Peace Process, edited by J. Large and Aguswandi. London: Conciliation
Resources.
Azam, Jean-Paul. 2002. Looting and Conflict between Ethnoregional Groups: Lessons for
State Formation in Africa. The Journal of Conflict Resolution 46 (1): 131-153.
Babbie, Earl. 1995. The Practice of Social Research. 7th ed. Belmont, CA: Wadsworth Publishing
Company.
Badan Rehabilitasi dan Rekonstruksi (BRR) and International Partners. 2005. Aceh and Nias
One Year after the Tsunami: The Recovery Effort and Way Forward, Badan
Rehabilitasi dan Rekonstruksi/The World Bank/Asian Development Bank/Red
Cross/Red Crescent/United Nations Development Program, Banda Aceh.
Ballentine, Karen, and Jake Sherman. 2003. The Political Economy of Armed Conflict: Beyond Greed
and Grievance. Boulder, Colo.: Lynne Rienner Publishers.
Barber, Richard, ed. 2000. Aceh: The Untold Story. Bangkok: Asian Forum for Human Rights
& Development.

155
Barnhart, John D. 1925. Rainfall and the Populist Party in Nebraska. American Political Science
Review 19 (3): 527-540.
Barre, Anton, Patrick Barron, Samuel Clark, Arya Gaduh, and Matthew Zurstrassen. 2006.
GAM Reintegration Needs Assessment: Enhancing Peace through Community-
Level Development Programming, The World Bank, Banda Aceh.
Barron, Patrick. 2008. Managing the Resources for Peace: Reconstruction and Peacebuilding
in Aceh. In Reconfiguring Politics: The Indonesia-Aceh Peace Process, edited by Aguswandi
and J. Large. London: Conciliation Resources.
Barron, Patrick, and Samuel Clark. 2006. Decentralizing Inequality? Center-Periphery
Relations, Local Governance, and Conflict in Aceh, World Bank, Washington, DC.
Bartkus, Viva Ona. 1999. The Dynamics of Secession. Cambridge: Cambridge University Press.
Bedi, Rahul. 2005. Sri Lanka Aid Promises Peace with Tigers. Daily Telegraph, June 25.
Belgrad, Eric A., and Nitza Nachmias, eds. 1997. The Politics of International Humanitarian Aid
Operations. Westport, CT: Praeger Publishers.
Bennett, Jon, William Bertrand, Clare Harkin, Stanley Samarasinghe, and Hemantha
Wickramatillake. 2006. Evaluation Findings: Sri Lanka. TEC: Coordination of
International Humanitarian Assistance in Tsunami-Affected Countries, Active
Learning Network for Accountability and Performance in Humanitarian Action
(ALNAP), London.
Berdal, Mats R., and David M. Malone, eds. 2000. Greed & Grievance: Economic Agendas in Civil
Wars. Boulder: Lynne Rienner Publishers.
Bilham, Roger. 2005. A Flying Start, Then a Slow Slip. Science 308 (5725): 1126.
Birkland, Thomas A. 1996. Natural Disasters as Focusing Events: Policy Communities and
Political Response. International Journal of Mass Emergencies and Disasters 14 (2): 221-243.
———. 1998. After Disaster: Agenda Setting, Public Policy, and Focusing Events. Washington, DC:
Georgetown University Press.
Blocker, T. Jean, E. Burke Rochford, Jr., and Darren E. Sherkat. 1991. Political Responses to
Natural Hazards: Social Movement Participation Following a Flood Disaster.
International Journal of Mass Emergencies and Disasters 9 (3): 367-382.
Boone, Peter. 1996. Politics and the Effectiveness of Foreign Aid. European Economic Review
40 (2): 289-329.
Bowen, John R. 2002. The Myth of Global Ethnic Conflict. In Genocide: An Anthropological
Reader, edited by A. L. Hinton. Malden, MA: Blackwell.

156
Bowen, Merle. 1989. Peasant Agriculture in Mozambique: The Case of Chokwe, Gaza.
Canadian Journal of African Studies 23 (3): 355-379.
———. 2000. The State against the Peasantry: Rural Struggles in Colonial and Post-Colonial
Mozambique. Charlottesville: University of Virginia Press.
Brady, Henry E., and John E. McNulty. 2005. The Costs of Voting: Evidence from a Natural
Experiment. Presented at the Annual Meeting of the Western Political Science
Association. Oakland, CA.
Brambor, Thomas, William Roberts Clark, and Matt Golder. 2006. Understanding
Interaction Models: Improving Empirical Analyses. Political Analysis 14 (1): 63-82.
Brancati, Dawn. 2007. Political Aftershocks: The Impact of Earthquakes on Intrastate
Conflict. Journal of Conflict Resolution 51 (5): 715-743.
Braumoeller, Bear F. 2004. Hypothesis Testing and Multiplicative Interaction Terms.
International Organization 58 (4): 807-820.
Braungart, Richard G. 1984. Historical and Generational Patterns of Youth Movements: A
Global Perspective. Comparative Social Research 7: 3-62.
Bräutigam, Deborah A., and Stephen Knack. 2004. Foreign Aid, Institutions, and
Governance in Sub-Saharan Africa. Economic Development and Cultural Change 52 (2):
255-285.
Brecher, Michael, and Jonathan Wilkenfeld. 1982. Crises in World Politics. World Politics 34
(3): 380-417.
BRR and International Partners. 2005. Aceh and Nias One Year after the Tsunami: The
Recovery Effort and Way Forward, Badan Rehabilitasi dan Rekonstruksi/The World
Bank/Asian Development Bank/Red Cross/Red Crescent/United Nations
Development Program, Banda Aceh.
Buhaug, Halvard. 2007. README File for UCDP/PRIO Intrastate Conflict Onset Dataset,
1946-2006. Uppsala Conflict Data Program, Uppsala University & International
Peace Research Institute, Oslo. http://www.prio.no/CSCW/Datasets/Armed-
Conflict/Onset-and-Duration-of-Intrastate-Conflict/Onset-Data/ (accessed July 7,
2008).
Buhaug, Halvard, Lars-Erik Cederman, and Jan Ketil Rød. 2008. Disaggregating Ethno-
Nationalist Civil Wars: A Dyadic Test of Exclusion Theory. International Organization
62 (3): 531-551.
Buhaug, Halvard, and Scott Gates. 2002. The Geography of Civil War. Journal of Peace
Research 39 (4): 417-433.

157
Buhaug, Halvard, and Kristian Skrede Gleditsch. 2008. Contagion or Confusion? Why
Conflicts Cluster in Space. International Studies Quarterly 52 (2): 215-233.
Buhaug, Halvard, and Jan Ketil Rød. 2006. Local Determinants of African Civil Wars, 1970-
2001. Political Geography 25 (3): 315-335.
Burke, Adam, and Afnan. 2005. Aceh: Reconstruction in a Conflict Environment.
Indonesian Social Development Paper 8, Decentralization Support Facility, Banda
Aceh/Jakarta.
Burns, Jimmy, Shawn Donnan, David Ibison, and Hugh Williamson. 2005. Indonesia Begins
Talks with Aceh Separatists. Financial Times, 11 January.
Carey, Sabine C. 2007. Rebellion in Africa: Disaggregating the Effect of Political Regimes.
Journal of Peace Research 44 (1): 47-64.
Cessation of Hostilities: Framework Agreement between Government of the Republic of
Indonesia and the Free Aceh Movement. 2002. Geneva.
Chandrakanthan, A. J. V. 2000. Eelam Tamil Nationalism: An inside View. In Sri Lankan
Tamil Nationalism: Its Origins and Development in the Nineteenth and Twentieth Centuries,
edited by A. J. Wilson. Vancouver: University of British Columbia Press.
Charles, J. Ammon, Ji Chen, Thio Hong-Kie, Robinson David, and et al. 2005. Rupture
Process of the 2004 Sumatra-Andaman Earthquake. Science 308 (5725): 1133.
Chong, Alberto, and Luisa Zanforlin. 2000. Law, Tradition, and Institutional Quality: Some
Empirical Evidence. Journal of International Development 12 (8): 1057-1068.
Choucri, Nazli. 1974. Population Dynamics and International Violence: Propositions, Insights, and
Evidence. Lexington: Lexington Books.
Cincotta, Richard P., Robert Engelman, and Daniele Anastasion. 2003. The Security
Demographic: Population and Civil Conflict after the Cold War. Washington, DC:
Population Action International.
Clague, Christopher, Philip Keefer, Stephen Knack, and Mancur Olson. 1996. Property and
Contract Rights under Democracy and Dictatorship. Journal of Economic Growth 1 (2):
243-276.
Cohen, Charles. 2006. Essays in Political Economy and Corporate Finance. Ph.D.
Dissertation, Harvard University.
Coll, Steve. 2005. Fault Lines: After the Earthquake, Some Strange New Alliances. The New
Yorker, November 21.

158
Collier, Paul. 2000. Doing Well out of War: An Economic Perspective. In Greed and Grievance:
Economic Agendas in Civil Wars, edited by M. R. Berdal and D. M. Malone. Boulder,
CO: Lynne Rienner.
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 and Oxford University Press.
Collier, Paul, and Anke Hoeffler. 1998. On Economic Causes of Civil War. Oxford Economic
Papers 50 (4): 563-573.
———. 2001. Greed and Grievance in Civil War. Policy Research Working Paper 2355,
World Bank, Washington, DC.
———. 2002. On the Incidence of Civil War in Africa. The Journal of Conflict Resolution 46 (1):
13-28.
———. 2004. Greed and Grievance in Civil War. Oxford Economic Papers 56 (4): 563-595.
Collier, Paul, Anke Hoeffler, and Nicholas Sambanis. 2005. The Collier-Hoeffler Model of
Civil War Onset and the Case Study Project Research Design. In Understanding Civil
War: Evidence and Analysis, edited by P. Collier and N. Sambanis. Washington, DC:
The World Bank.
Collier, Paul, Anke Hoeffler, and Mans Soderbom. 2004. On the Duration of Civil War.
Journal of Peace Research 41 (3): 253-273.
Collier, Paul, and Nicholas Sambanis, eds. 2005a. Understanding Civil War: Evidence and
Analysis. Vol. 1: Africa. Washington, DC: World Bank.
———, eds. 2005b. Understanding Civil War: Evidence and Analysis. Vol. 2: Europe, Central
Asia, and Other Regions. Washington, DC: World Bank.
Comfort, Louise K. 2000. Disaster: Agent of Diplomacy or Change in International Affairs?
Cambridge Review of International Affairs 14 (1): 277-294.
Culbert, Vance. 2005. Civil Society Development Versus the Peace Dividend: International
Aid in the Wanni. Disasters 29 (1): 38-57.
Cuny, Frederick C. 1983. Disasters and Development. Oxford: Oxford University Press.
Dawood, Dayan, and Sjafrizal. 1989. Aceh: The Lng Boom and Enclave Development. In
Unity and Diversity: Regional Economic Development in Indonesia since 1970, edited by H.
Hill. Singapore: Oxford University Press.
De Soysa, Indra. 2002. Paradise Is a Bazaar? Greed, Creed, and Governance in Civil War,
1989-99. Journal of Peace Research 39 (4): 395-416.

159
Della Porta, Donatella, and Alessandro Pizzorno. 1996. The Business Politicians: Reflections
from a Study of Political Corruption. In The Corruption of Politics and the Politics of
Corruption, edited by M. Levi and D. Nelken. Oxford: Blackwell Publishers.
DeRouen, Karl R., and David Sobek. 2004. The Dynamics of Civil War Duration and
Outcome. Journal of Peace Research 41 (3): 303-320.
DeVotta, Neil. 2004. Sri Lanka in 2003: Seeking to Consolidate Peace. Asian Survey 44 (1):
49-55.
———. 2005. Sri Lanka in 2004: Enduring Political Decay and a Failing Peace Process.
Asian Survey 45 (1): 98-104.
Diamond, Larry, and Leonardo Morlino. 2004. The Quality of Democracy: An Overview.
Journal of Democracy 15 (4): 20.
Diehl, Paul F., ed. 1999. A Road Map to War: Territorial Dimensions of International Conflict.
Nashville, TN: Vanderbilt University Press.
Dietz, Simon, Eric Neumayer, and Indra De Soysa. 2007. Corruption, the Resource Curse,
and Genuine Saving. Environment and Development Economics 12 (1): 33-53.
Djuli, M. Nur, and Nurdin Abdul Rahman. 2008. The Helsinki Negotiations: A Perspective
from Free Aceh Movement Negotiators. In Reconfiguring Politics: The Indonesia-Aceh
Peace Process, edited by J. Large and Aguswandi. London: Conciliation Resources.
Doherty, Daniel, Donald Green, and Alan Gerber. 2006. Personal Income and Attitudes
toward Redistribution: A Study of Lottery Winners. Political Psychology 27 (3): 441-458.
Donnan, Shawn. 2005. Acehnese Rebels Freed in Jakarta Amnesty. Financial Times, 18
August.
Donnan, Shawn, and Rupini Bergstrom. 2005. Rebels Ease Stance on Sovereignty. Financial
Times, January 31.
Donnan, Shawn, and Taufan Hidayat. 2005. Aceh Rebels Start to Hand in Weapons.
Financial Times, 16 September.
Drury, A. Cooper, and Richard Stuart Olson. 1998. Disasters and Political Unrest: An
Empirical Investigation. Journal of Contingencies and Crisis Management 6 (3): 153-161.
Drury, A. Cooper, Richard Stuart Olson, and Douglas A. Van Belle. 2005. The Politics of
Humanitarian Aid: U.S. Foreign Disaster Assistance, 1964-1995. The Journal of Politics
67 (2): 454-473.
Dunning, Thad. 2008a. Improving Causal Inference: Strengths and Limitations of Natural
Experiments. Political Research Quarterly 61 (2): 282-293.

160
———. 2008b. Model Specification in Instrumental-Variables Regression. Political Analysis
16 (3): 290-302.
Easterly, William. 2001. Can Institutions Resolve Ethnic Conflict? Economic Development and
Cultural Change 49 (4): 687-706.
Eckstein, Harry. 1975. Case Studies and Theory in Political Science. In Handbook of Political
Science, Volume 7, edited by F. Greenstein and N. Polsby. Reading, MA: Addison-
Wesley.
———. 1980. Theoretical Approaches to Explaining Collective Political Violence. In
Handbook of Political Conflict: Theory and Research, edited by T. R. Gurr. New York: The
Free Press.
Eisinger, Peter K. 1973. The Conditions of Protest Behavior in American Cities. American
Political Science Review 67 (1): 11-28.
Ekstrom, Goran, Meredith Nettles, and Victor C. Tsai. 2006. Seasonality and Increasing
Frequency of Greenland Glacial Earthquakes. Science 24 (5768): 1756-1758.
Elbadawi, Ibrahim, and Nicholas Sambanis. 2000. Why Are There So Many Civil Wars in
Africa? Understanding and Preventing Violent Conflict. Journal of African Economies 9
(3): 244-269.
———. 2002. How Much War Will We See? Explaining the Prevalence of Civil War. The
Journal of Conflict Resolution 46 (3): 307-334.
Ellingsen, Tanja. 2000. Colorful Community or Ethnic Witches' Brew? Multiethnicity and
Domestic Conflict During and after the Cold War. Journal of Conflict Resolution 44 (2):
228-249.
Engdahl, E. Robert, and Antonio Villaseñor. 2002. Centennial Earthquake Catalog.
Washington, DC: U.S. Geological Survey.
Enia, Jason S. 2006. Between Promise and Delivery: Relief and Reconstruction after the
2003 Iranian Earthquake. Georgetown University Institute for the Study of
Diplomacy (GUISD): PEW Case Study Center.
———. 2008. Peace in Its Wake? The 2004 Tsunami and Internal Conflict in Indonesia and
Sri Lanka. Journal of Public and International Affairs 19 (1): 7-27.
Esteban, Joan, and Debraj Ray. 2008. Polarization, Fractionalization and Conflict. Journal of
Peace Research 45 (2): 163-182.
Evin, Ahmet. 2004. Changing Greek Perspectives on Turkey: An Assessment of the Post-
Earthquake Rapprochement. Turkish Studies 5 (1): 4-20.

161
Fearon, James D. 1994. Signaling Versus the Balance of Power and Interests: An Empirical
Test of a Crisis Bargaining Model. The Journal of Conflict Resolution 38 (2): 236-269.
———. 1995. Rationalist Explanations for War. International Organization 49 (3): 379-414.
———. 1997. Signaling Foreign Policy Interests: Tying Hands Versus Sinking Costs. The
Journal of Conflict Resolution 41 (1): 68-90.
———. 1998. Bargaining, Enforcement, and International Cooperation. International
Organization 52 (2): 269-305.
———. 2004. Why Do Some Civil Wars Last So Much Longer Than Others? Journal of Peace
Research 41 (3): 275-301.
———. 2005. Primary Commodity Exports and Civil War. Journal of Conflict Resolution 49 (4):
483-507.
Fearon, James D., and David D. Laitin. 2003. Ethnicity, Insurgency, and Civil War. American
Political Science Review 97 (1): 75-90.
Flanigan, Shawn Teresa. 2006. Charity as Resistance: Connections between Charity,
Contentious Politics, and Terror. Studies in Conflict & Terrorism 29 (7): 641-655.
———. 2008. Nonprofit Service Provision by Insurgent Organizations: The Cases of
Hizballah and the Tamil Tigers. Studies in Conflict & Terrorism 31 (6): 499-519.
Freedom House. 2004. Countries at the Crossroads 2004: A Survey of Democratic
Governance, Freedom House, New York.
Freudenheim, Ellen. 1979. Politics in International Disasters: Fact, Not Fiction. In Disaster
Assistance: Appraisal, Reform, and New Approaches, edited by L. H. Stephens and S. J.
Green. London: Macmillan.
Frieden, Jeffry A. 1991. Invested Interests: The Politics of National Economic Policies in a
World of Global Finance. International Organization 45 (4): 425-451.
Frödin, Lina. 2008. The Challenges of Reintegration in Aceh. In Reconfiguring Politics: The
Indonesia-Aceh Peace Process, edited by Aguswandi and J. Large. London: Conciliation
Resources.
Gaillard, Jean-Christophe, Elsa Clavé, Océane Vibert, Azarhi, Dedi, Jean-Charles Denain,
Yusuf Efendi, Delphine Grancher, Catherine C. Liamzon, Desy Rosnita Sari, and
Ryo Setiawan. 2008. Ethnic Groups' Response to the 26 December 2004 Earthquake
and Tsunami in Aceh, Indonesia. Natural Hazards 47: 17-38.
Gamson, William A. 1975. The Strategy of Social Protest. Homewood, IL: Dorsey.

162
Gates, Scott. 2002. Recruitment and Allegiance: The Microfoundations of Rebellion. The
Journal of Conflict Resolution 46 (1): 111-130.
Gates, Scott, Håvard Hegre, Mark P. Jones, and Håvard Strand. 2006. Institutional
Inconsistency and Political Instability: Polity Duration, 1800-2000. American Journal of
Political Science 50 (4): 893-908.
Gendron, Richard Arthur. 1998. Faultlines of Power: The Political Economy of
Redevelopment in a Progressive City after a Natural Disaster. Ph.D. Dissertation,
University of California, Santa Cruz, Santa Cruz.
George, Alexander L. 1979. Case Studies and Theory Development: The Method of
Structured, Focused Comparison. In Diplomacy: New Approaches in History, Theory, and
Policy, edited by P. G. Lauren. New York: Free Press.
George, Alexander L., and Andrew Bennett. 2005. Case Studies and Theory Development in the
Social Sciences. Cambridge: MIT Press.
Gerring, John. 2007. Case Study Research: Principles and Practices. Cambridge: Cambridge
University Press.
Glantz, Michael H. 2000. Climate-Related Disaster Diplomacy: A US-Cuban Case Study.
Cambridge Review of International Affairs 14 (1): 233-253.
Gleditsch, Kristian S., and Michael D. Ward. 1999. A Revised List of Independent States
since the Congress of Vienna. International Interactions 25 (4): 393-413.
Gleditsch, Nils Petter, Peter Wallensteen, Mikael Eriksson, Margareta Sollenberg, and
Håvard Strand. 2002. Armed Conflict 1946-2001: A New Dataset. Journal of Peace
Research 39 (5): 615-637.
Goldstone, Jack A. 1991. Revolution and Rebellion in the Early Modern World. Berkeley:
University of California Press.
———. 2001. Demography, Environment, and Security. In Environmental Conflict, edited by
P. F. Diehl and N. P. Gleditsch. Boulder: Westview Press.
Grossman, Herschel I. 1991. A General Equilibrium Model of Insurrections. American
Economic Review 81: 912-921.
———. 1995. Insurrections. In Handbook of Defense Economics, edited by K. Hartley and T.
Sandler. New York: Elsevier.
———. 1999. Kleptocracy and Revolutions. Oxford Economic Papers 51.

163
Guha-Sapir, Debarati, David Hargitt, and Philippe Hoyois. 2004. Thirty Years of Natural
Disasters 1974-2003: The Numbers, Department of Public Health, Université
Catholique de Louvain, Brussels, Belgium.
Gurr, Ted. 1970. Why Men Rebel. Princeton, NJ: Princeton University Press.
Gurr, Ted R. 2000. Peoples Versus States: Minorities at Risk in the New Century. Washington, DC:
U.S. Institute of Peace.
Gurr, Ted Robert. 1993. Why Minorities Rebel: A Global Analysis of Communal
Mobilization and Conflict since 1945. International Political Science Review 14 (2): 161-
201.
Hadi, Faisal. 2008. Human Rights and Injustice in Aceh: The Long and Winding Road. In
Reconfiguring Politics: The Indonesia-Aceh Peace Process, edited by Aguswandi and J. Large.
London: Conciliation Resources.
Haney, Patrick J. 1997. Organizing for Foreign Policy Crisis. Ann Arbor: University of Michigan
Press.
Hanks, Thomas C., and Hiroo Kanamori. 1979. Moment Magnitude Scale. Journal of
Geophysical Research 84 (B5): 2348-2350.
Harbom, Lotta, and Peter Wallensteen. 2007. Armed Conflict, 1989-2006. Journal of Peace
Research 44 (5): 623-634.
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. 2006. Sensitivity Analysis of Empirical Results on
Civil War Onset. Journal of Conflict Resolution 50 (4): 508-535.
Henderson, Errol A., and J. David Singer. 2000. Civil War in the Post-Colonial World, 1946-
92. Journal of Peace Research 37 (3): 275-299.
Herbst, Jeffrey. 2004. African Militaries and Rebellion: The Political Economy of Threat and
Combat Effectiveness. Journal of Peace Research 41 (3): 357-369.
Heston, Alan, Robert Summers, and Bettina Aten. 2008. Penn World Table, Version 6.2.
Center for International Comparisons at the University of Pennsylvania (CICUP).
http://datacentre2.chass.utoronto.ca/pwt/ (accessed June 15, 2008).
Hirshleifer, Jack. 1995. Theorizing About Conflict. In Handbook of Defense Economics, edited
by K. Hartley and T. Sandler. New York: Elsevier.

164
———. 2001. The Dark Side of the Force: Economic Foundations of Conflict Theory. Cambridge
New York: Cambridge University Press.
Hogg, Charu Lata. 2006. Sri Lanka: The Liberation Tigers of Tamil Eelam (LTTE) and Child
Recruitment, Coalition to Stop the Use of Child Soldiers, London.
Howden, Daniel. 2005. Tsunami Disaster: Indonesian Military Refuses Peace Talks with
Separatists. The Independent, 13 January.
Huggler, Justin. 2005. The Tamil Who Took on the Tigers. Independent, August 15.
Human Rights Watch. 2004. Living in Fear: Child Soldiers and the Tamil Tigers in Sri Lanka,
Human Rights Watch, New York.
Humphreys, Macartan. 2005. Natural Resources, Conflict, and Conflict Resolution:
Uncovering the Mechanisms. Journal of Conflict Resolution 49 (4): 508-537.
Humphreys, Macartan, and Habaye Ag Mohamed. 2005. Senegal and Mali. In Understanding
Civil War: Evidence and Analysis, edited by P. Collier and N. Sambanis. Washington,
DC: The World Bank.
Humphreys, Macartan, and Jeremy M. Weinstein. 2008. Who Fights? The Determinants of
Participation in Civil War. American Journal of Political Science 53 (2): 436-455.
Huntington, Samuel P. 1968. Political Order in Changing Societies. New Haven: Yale University
Press.
———. 1996. The Clash of Civilizations and the Remaking of World Order. New York: Simon &
Schuster.
Huther, Jeff, and Anwar Shah. 2005. A Simple Measure of Good Governance. In Public
Services Delivery, edited by A. Shah. Washington, DC: The World Bank.
Huxley, Tim. 2005. The Tsunami and Security: Asia’s 9/11? Survival 47 (1): 123-132.
Imai, Kosuke, Gary King, and Olivia Lau. 2007. Everyone's Statistical Software.
http://GKing.harvard.edu/zelig (accessed June 1, 2008).
Intergovernmental Panel on Climate Change. 2007. Climate Change 2007: Synthesis Report.
http://www.ipcc.ch.
International Crisis Group. 2001. Aceh: Can Autonomy Stem the Conflict? ICG Asia Report
18, International Crisis Group, Jakarta/Brussels.
———. 2002. Aceh: Slim Chance for Peace. ICG Asia Briefing 14, International Crisis
Group, Jakarta/Brussels.

165
———. 2003a. Aceh: How Not to Win Hearts and Minds. ICG Asia Briefing 27,
International Crisis Group, Jakarta/Brussels.
———. 2003b. Aceh: Why the Military Option Still Won't Work. ICG Asia Briefing 26,
International Crisis Group, Jakarta/Brussels.
———. 2005. Aceh: A New Chance for Peace. ICG Asia Briefing 40, International Crisis
Group, Jakarta/Brussels.
———. 2006a. Islamic Law and Criminal Justice in Aceh. ICG Asia Report, International
Crisis Group, Jakarta/Brussels.
———. 2006b. Sri Lanka: The Failure of the Peace Process. ICG Asia Report 124,
International Crisis Group, Jakarta/Brussels.
———. 2007a. Aceh: Post-Confict Complications. ICG Asia Report 139, International
Crisis Group, Jakarta/Brussels.
———. 2007b. Indonesia: How GAM Won in Aceh. ICG Asia Briefing 61, International
Crisis Group, Jakarta/Brussels.
Iranian Daily Says Ignoring the U.S. Would Be 'Hugely Unwise'. 2004. BBC Monitoring
International Reports, January 6.
Jagtiani, Sunil. 2005. Closing Door for Peace Talks in Sri Lanka. Christian Science Monitor,
December 28.
Jarstad, Anna K., and Desiree Nilsson. 2008. From Words to Deeds: The Implementation of
Power-Sharing Pacts in Peace Accords. Conflict Management and Peace Science 25 (3):
206-223.
Jones, Sidney. 2008. Keeping the Peace: Security in Aceh. In Reconfiguring Politics: The
Indonesia-Aceh Peace Process, edited by Aguswandi and J. Large. London: Conciliation
Resources.
Kahl, Colin H. 1998. Population Growth, Environmental Degradation, and State-Sponsored
Violence: The Case of Kenya, 1991-93. International Security 23 (2): 80-119.
Kaldor, Mary. 2005. The European Union and Sri Lanka: Tsunami Response and Long
Term Policy, The Centre for the Study of Global Governance, London School of
Economics and Political Science and Ministry for Foreign Affairs Sweden.
Kalyvas, Stathis N. 2001. 'New' and 'Old' Civil Wars: A Valid Distinction? World Politics 54
(1): 99-118.

166
Kalyvas, Stathis N., and Nicholas Sambanis. 2005. Bosnia's Civil War: Origins and Violence
Dynamics. In Understanding Civil War: Evidence and Analysis, edited by P. Collier and N.
Sambanis. Washington, DC: The World Bank.
Karl, Terry Lynn. 1997. The Paradox of Plenty: Oil Booms and Petro-States. Berkeley: University of
California Press.
Kearney, Marianne, and Richard Spencer. 2005. Indonesian Army Uses Tsunami Chaos to
Launch Attacks on Rebels. The Daily Telegraph.
Kearney, Robert N. 1985. Ethnic Conflict and the Tamil Separatist Movement in Sri Lanka.
Asian Survey 25 (9): 898-917.
Kell, Tim. 1995. The Roots of the Acehnese Rebellion, 1989-1992. Ithaca, NY: Cornell Modern
Indonesia Project.
Kelman, Ilan. 2005. Tsunami Diplomacy: Will the 26 December 2004 Bring Peace to the
Affected Countries? Sociological Research Online (1),
http://www.socresonline.org.uk/10/1/kelman.html.
———. 2006. Acting on Disaster Diplomacy. Journal of International Affairs 59 (2): 215-240.
Kelman, Ilan, and Theo Koukis. 2000. Disaster Diplomacy: Discord Disintegrated?
Cambridge Review of International Affairs 14 (1): 214.
Keohane, Robert O. 1984. After Hegemony: Cooperation and Discord in the World Political Economy.
Princeton, N.J.: Princeton University Press.
Ker-Lindsay, James. 2000. Greek-Turkish Rapprochement: The Impact of ‘Disaster
Diplomacy’? Cambridge Review of International Affairs 14 (1): 215-232.
Kimenyi, Mwangi S., and Njuguna S. Ndung'U. 2005. Sporadic Ethnic Violence: Why Has
Kenya Not Experienced a Full-Blown Civil War? In Understanding Civil War: Evidence
and Analysis, edited by P. Collier and N. Sambanis. Washington, DC: The World
Bank.
King, Gary. 1998. Unifying Political Methodology: The Likelihood Theory of Statistical Inference. Ann
Arbor: The University of Michigan Press.
Kingsbury, Damien. 2007. Indonesia in 2006: Cautious Reform. Asian Survey 47 (1): 155-161.
Kipp, Rita Smith. 2004. Indonesia in 2003: Terror's Aftermath. Asian Survey 44 (1): 62-69.
Knack, Stephen. 2001. Aid Dependence and the Quality of Governance: Cross-Country
Empirical Tests. Southern Economic Journal 68 (2): 310-329.

167
Knack, Stephen, and Philip Keefer. 1995. Institutions and Economic Performance: Cross-
Country Tests Using Alternative Institutional Measures. Economics and Politics 7 (3):
207-227.
Knowles, Stephen. 2005. Inequality and Economic Growth: The Empirical Relationship
Reconsidered in the Light of Comparable Data. Journal of Development Studies 41 (1):
135-159.
Korf, Benedikt. 2005. Rethinking the Greed-Grievance Nexus: Property Rights and the
Political Economy of War in Sri Lanka. Journal of Peace Research 42 (2): 201-217.
———. 2006. Functions of Violence Revisited: Greed, Pride and Grievance in Sri Lanka's
Civil War. Progress in Development Studies 6 (2): 109-122.
Kurlantzick, Joshua. 2008. Destructive Power: Natural Disasters Often Prompt Political
Reforms--and, Sometimes, Revolutionary Changes. The Boston Globe, 11 May.
Kydd, Andrew H. 2000. Trust, Reassurance, and Cooperation. International Organization 54
(2): 325-357.
———. 2005. Trust and Mistrust in International Relations. Princeton: Princeton University
Press.
Kydd, Andrew H., and Barbara F. Walter. 2006. The Strategies of Terrorism. International
Security 31 (1): 49-80.
La, John. 2004. Forced Remittances in Canada's Tamil Enclaves. Peace Review 16: 379-385.
La Porta, Rafael, Florencia Lopez-de-Silanes, Andrei Shleifer, and Robert W. Vishny. 1997.
Trust in Large Scale Organizations. American Economic Review 87 (2): 333-338.
Large, Judith. 2008. The Challenge of Hidden Economics and Predation for Profit. In
Reconfiguring Politics: The Indonesia-Aceh Peace Process, edited by Aguswandi and J. Large.
London: Conciliation Resources.
Lay, Thorne, Hiroo Kanamori, Charles J. Ammon, Meredith Nettles, Steven N. Ward,
Richard C. Aster, Susan L. Beck, Susan L. Bilek, Michael R. Brudzinski, Rhett Butler,
Heather R. DeShon, Göran Ekström, Kenji Satake, and Stuart Sipkin. 2005. The
Great Sumatra-Andaman Earthquake of 26 December 2004. Science 308 (5725): 1127-
1133.
Levi, Margaret. 1988. Of Rule and Revenue. Berkeley, CA: University of California Press.
Liberation Tigers of Tamil Eelam. 2003. The Proposal by the Liberation Tigers of Tamil
Eelam on Behalf of the Tamil People for an Agreement to Establish an Interim Self-
Governing Authority for the Northeast of the Island of Sri Lanka. Kilinochchi.

168
Lichbach, Mark Irving. 1989. An Evaluation of "Does Economic Inequality Breed Political
Conflict?" Studies. World Politics 41 (4): 431-470.
———. 1990. Will Rational People Rebel against Inequality? Samson's Choice. American
Journal of Political Science 34 (4): 1049-1076.
———. 1994. What Makes Rational Peasants Revolutionary? Dilemma, Paradox, and Irony
in Peasant Collective Action. World Politics 46 (3): 383-418.
Lowi, Miriam R. 2005. Algeria, 1992-2002: Anatomy of a Civil War. In Understanding Civil
War: Evidence and Analysis, edited by P. Collier and N. Sambanis. Washington, DC:
The World Bank.
Lund, Michael S. 2005. Greed and Grievance Diverted: How Macedonia Avoided Civil War,
1990-2001. In Understanding Civil War: Evidence and Analysis, edited by P. Collier and
N. Sambanis. Washington, DC: The World Bank.
Maddox, Bronwen. 2005. Quake Diplomacy at Work. Ottawa Citizen, October 11.
Makdisi, Samir, and Richard Sadaka. 2005. The Lebanese Civil War, 1975-90. In
Understanding Civil War: Evidence and Analysis, edited by P. Collier and N. Sambanis.
Washington, DC: The World Bank.
Manogaran, Chelvadurai. 1987. Ethnic Conflict and Reconciliation in Sri Lanka. Honolulu:
University of Hawaii Press.
Marshall, Andrew. 2007. The Escape Artist. Time, 15 February.
Marshall, Monty G., and Ted R. Gurr. 2003. Peace and Conflict, 2003: A Global Survey of Armed
Conflicts, Self-Determination Movements, and Democracy. Baltimore: Center for
International Development and Conflict Management, University of Maryland.
Mason, T. David, Joseph P. Weingarten, Jr., and Patrick J. Fett. 1999. Win, Lose, or Draw:
Predicting the Outcome of Civil Wars. Political Research Quarterly 52 (2): 239-268.
Matthews, Bruce. 1982. District Development Councils in Sri Lanka. Asian Survey 22 (11):
1117-1134.
Mauro, Paolo. 1995. Corruption and Growth. Quarterly Journal of Economics 110 (3): 681-711.
May, Bernhard. 2008. The Law on the Governing of Aceh: The Way Forward or a Source of
Conflicts? In Reconfiguring Politics: The Indonesia-Aceh Peace Process, edited by Aguswandi
and J. Large. London: Conciliation Resources.
McAdam, Doug, Sidney G. Tarrow, and Charles Tilly. 2001. Dynamics of Contention. New
York: Cambridge University Press.

169
McCarthy, John D., and Mayer N. Zald. 1977. Resource Mobilization and Social
Movements: A Partial Theory. The American Journal of Sociology 82 (6): 1212-1241.
McCarty, Nolan, and Adam Meirowitz. 2007. Political Game Theory. Princeton: Princeton
University Press.
McCawley, Tom. 2005. Indonesia Tightens Access to Aceh. The Christian Science Monitor, 14
January.
McCulloch, Lesley. 2005. Aceh: Then and Now, Minority Rights Group International,
London.
McDougall, Dan. 2005. Faith in Democracy Is Another Casualty of the Killer Wave. Observer,
October 30.
McNeil, Donald G., Jr. 2005. How Nature Changes History. New York Times, January 2.
Merkel, Wolfgang. 2005. Embedded and Defective Democracies. Democratization 11 (5): 33-
58.
Miguel, Edward. 2005. Poverty and Witch Killing. Review of Economic Studies 72: 1153-1172.
Miguel, Edward, Shanker Satyanath, and Ernest Sergenti. 2004. Economic Shocks and Civil
Conflict: An Instrumental Variables Approach. The Journal of Political Economy 112 (4):
725-753.
Miller, Michele Ann. 2008. The Conflict in Aceh: Context, Precursors and Catalysts. In
Reconfiguring Politics: The Indonesia-Aceh Peace Process, edited by Aguswandi and J. Large.
London: Conciliation Resources.
Mohammed Khatami. 2003. Interview by Vision of the Islamic Republic of Iran Network 1.
Unpublished Manuscript.
Morris, Eric Eugene. 1983. Islam and Politics in Aceh: A Study of Center-Periphery
Relations in Indonesia, Cornell University, Ithaca.
Morrow, James D. 1994. Game Theory for Political Scientists. Princeton: Princeton University
Press.
Murshed, S. Mansoob, and Scott Gates. 2005. Spatial-Horizontal Inequality and the Maoist
Insurgency in Nepal. Review of Development Economics 9 (1): 121-134.
Mydans, Seth. 2008. Rulers Keep Grip on Aid as Burmese Cast Votes. The New York Times,
May 11.
Nafziger, E. Wayne, and Juha Auvinen. 2002. Economic Development, Inequality, War, and
State Violence. World Development 30 (2): 153-163.

170
Nagel, Jack H. 1974. Inequality and Discontent: A Nonlinear Hypothesis. World Politics 26
(4): 453-472.
Nakashima, Ellen. 2005a. Aceh Rebels Describe Effort to Aid ‘Our Own’. Washington Post,
January 18.
———. 2005b. Indonesia Offering 'Olive Branch' to Rebels. Washington Post, January 28.
Ndikumana, Leonce, and Kisangani F. Emizet. 2005. The Economics of Civil War: The Case
of the Democratic Republic of Congo. In Understanding Civil War: Evidence and
Analysis, edited by P. Collier and N. Sambanis. Washington, DC: The World Bank.
Nel, Philip. 2006. Globalization and Violent Political Dissent in Developing Countries. In
Globalization and Conflict: National Security in a 'New' Strategic Era, edited by R. G.
Patman. London: Routledge.
Nel, Philip, and Marjolein Righarts. 2008. Natural Disasters and the Risk of Violent Civil
Conflict. International Studies Quarterly 52 (2): 159-185.
New York Times. 2005. From the Ruins. New York Times, January 8.
Ngaruko, Floribert, and Janvier D. Nkurunziza. 2005. Civil War and Its Duration in
Burundi. In Understanding Civil War: Evidence and Analysis, edited by P. Collier and N.
Sambanis. Washington, DC: The World Bank.
Noble, Gregory, and John Ravenhill. 2000. Causes and Consequences of the Asian Financial
Crisis. In The Asian Financial Crisis and the Architecture of Global Finance, edited by G.
Noble and J. Ravenhill. Cambridge: Cambridge University Press.
North, Douglas C. 1990. Institutions, Institutional Change and Economic Performance. Cambridge:
Cambridge University Press.
Oberschall, Anthony. 1973. Social Conflict and Social Movements. Englewood Cliffs, NJ:
Prentice-Hall.
Office of U.S. Foreign Disaster Assistance (OFDA) Centre for Research on the
Epidemiology of Disasters (CRED). 2008. EM-DAT: The OFDA/CRED
International Disaster Database. Université Catholique de Louvain. http://www.em-
dat.net (accessed June 15, 2008).
Oliver-Smith, Anthony. 1977. Disaster Rehabilitation and Social Change in Yungay, Peru.
Human Organization 36 (1): 5-13.
———. 1979. Post Disaster Consensus and Conflict in a Traditional Society: The 1970
Avalanche of Yungay, Peru. Mass Emergencies 4 (1): 39-52.

171
Olson, Mancur. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups.
Cambridge: Harvard University Press.
———. 1982. The Rise and Decline of Nations: Economic Growth, Stagflation, and Social Rigidities.
New Haven: Yale University Press.
Olson, Richard Stuart. 2000. Toward a Politics of Disaster: Losses, Values, Agendas, and
Blame. International Journal of Mass Emergencies and Disasters 18 (2): 265-287.
Olson, Richard Stuart, and A. Cooper Drury. 1997. Un-Therapeutic Communities: A Cross-
National Analysis of Post-Disaster Political Unrest. International Journal of Mass
Emergencies and Disasters 15 (2): 221-238.
Olson, Richard Stuart, and Vincent T. Gawronski. 2003. Disasters as Critical Junctures?
Managua, Nicaragua 1972 and Mexico City 1985. International Journal of Mass
Emergencies and Disasters 21 (1): 3-35.
Olsson, Ola, and Heather Congdon Fors. 2004. Congo: The Prize of Predation. Journal of
Peace Research 41 (3): 321-336.
Østby, Gudrun. 2008. Polarization, Horizontal Inequalities and Violent Civil Conflict. Journal
of Peace Research 45 (2): 143-162.
P-TOMS Agreement. 2005. Memorandum of Understanding ("Mou") for the Establishment
of a Post-Tsunami Operational Management Structure ("P-TOMS"). Colombo: Sri
Lanka.
Paxson, Christina H. 1992. Using Weather Variability to Estimate the Response of Savings
to Transitory Income in Thailand. The American Economic Review 82 (1): 15-33.
Pelling, Mark. 2003. The Vulnerability of Cities: Natural Disaster and Social Resilience. London:
Earthscan.
Philipson, Liz, ed. 1998. Demanding Sacrifice: War and Negotiation in Sri Lanka. Edited by J.
Armon, Accord. London: Conciliation Resources.
Philipson, Liz, and Yuvi Thangarajah. 2005. The Politics of the North-East: Part of the Sri
Lanka Strategic Conflict Assessment 2005, The Asia Foundation, Colombo.
Pitcher, M. Anne. 1998. Disruption without Transformation: Agrarian Relations and
Livelihoods in Nampula Province, 1975-1995. Journal of Southern African Studies 24 (1):
115-140.
Planning and Development Secretariat of the Liberation Tigers of Tamil Eelam. 2005. Post-
Tsunami Reconstruction: Needs Assessment for the Northeast, Kilinochchi.

172
Platt, Rutherford H. 1999. Disasters and Democracy: The Politics of Extreme Natural Events.
Washington, DC: Island Press.
Pocha, Jehangir S. 2005a. Amid Disaster, Hopes for Peace. Boston Globe, 15 January.
———. 2005b. Sri Lanka Rebels Take up Aid Effort. Boston Globe, January 5.
———. 2005c. Sri Lanka Struggles to Manage Aid Influx. Boston Globe, January 4.
Powell, Robert. 1999. In the Shadow of Power: States & Strategies in International Politics.
Princeton: Princeton University Press.
———. 2002. Bargaining Theory and International Conflict. Annual Review of Political Science
5 (1): 1-30.
———. 2004. Bargaining and Learning While Fighting. American Journal of Political Science 48
(2): 344-361.
Powell, Sian. 2005. Aceh’s Elusive Peace. Weekend Australian, January 29.
Quarantelli, Enrico L. 2001. Statistical and Conceptual Problems in the Study of Disasters.
Disaster Prevention and Management 10 (5): 325-338.
Quarantelli, Enrico L., and Russell R. Dynes. 1976. Community Conflict: Its Absence and Its
Presence in Natural Disasters. Mass Emergencies 1 (2): 139-152.
Raheem, M., and V. Gosselin. 2003. Informal Dispute Resolution in the Northeast and Puttalam.
Colombo: Centre for Policy Alternatives.
Raknerud, Arvid, and Håvard Hegre. 1997. The Hazard of War: Reassessing the Evidence
for the Democratic Peace. Journal of Peace Research 34 (4): 385-404.
Regan, Patrick M., and Daniel Norton. 2005. Greed, Grievance, and Mobilization in Civil
Wars. Journal of Conflict Resolution 49 (3): 319-336.
Reid, Anthony. 2004. War, Peace and the Burden of History in Aceh. Asian Ethnicity 5 (3):
301-314.
Reynal-Querol, Marta. 2002a. Ethnicity, Political Systems, and Civil Wars. The Journal of
Conflict Resolution 46 (1): 29-54.
———. 2002b. Political Systems, Stability, and Civil Wars. Defence and Peace Economics 13 (6):
465-483.
———. 2005. Does Democracy Preempt Civil Wars? European Journal of Political Economy 21
(2): 445-465.

173
Reynolds, Paul. 2003. The Politics of Earthquakes. BBC News Online, December 30.
http://news.bbc.co.uk/2/hi/middle_east/3351121.stm (accessed February 5, 2005).
Riley, John G. 2001. Silver Signals: Twenty-Five Years of Screening and Signaling. Journal of
Economic Literature 39 (2): 432-478.
Robinson, Geoffrey. 1998. Rawan Is as Rawan Does: The Origins of Disorder in New Order
Aceh. Indonesia (66): 127.
Rodrik, Dani. 1999. Where Did All the Growth Go? Journal of Economic Growth 4: 385-412.
Rofi, Abdur, Shannon Doocy, and Courtland Robinson. 2006. Tsunami Mortality and
Displacement in Aceh Province, Indonesia. Disasters 30 (3): 340-350.
Rogers, Steven. 2004. Philippine Politics and the Rule of Law. Journal of Democracy 15 (4): 111-
126.
Rogowski, Ronald. 1989. Commerce and Coalitions: How Trade Affects Domestic Political Alignments.
Princeton: Princeton University Press.
Ron, James. 2005. Paradigm in Distress? Primary Commodities and Civil War. Journal of
Conflict Resolution 49 (4): 443-450.
Rosenberg, Tina. 2002. Editorial Observer: To Prevent Conflicts, Look to Commodities
Like Diamonds. The New York Times, July 15.
Ross, Michael L. 1999. The Political Economy of the Resource Curse. World Politics 51 (2):
297-322.
———. 2004. What Do We Know About Natural Resources and Civil War? Journal of Peace
Research 41 (3): 337-356.
———. 2005. Resources and Rebellion in Aceh, Indonesia. In Understanding Civil War:
Evidence and Analysis, edited by P. Collier and N. Sambanis. Washington, DC: The
World Bank.
———. 2006. A Closer Look at Oil, Diamonds, and Civil War. Annual Review of Political
Science 9 (1): 265-300.
Sachs, Jeffrey D., and Andrew M. Warner. 2000. Natural Resource Abundance and
Economic Growth. In Leading Issues in Economic Development, edited by G. M. Meier
and J. E. Rauch. Oxford: Oxford University Press.
Sambanis, Nicholas. 2001. Do Ethnic and Nonethnic Civil Wars Have the Same Causes? A
Theoretical and Empirical Inquiry (Part 1). Journal of Conflict Resolution 45 (3): 259-282.

174
———. 2002. A Review of Recent Advances and Future Directions in the Quantitative
Literature on Civil War. Defence and Peace Economics 13 (3): 215-243.
———. 2004. What Is Civil War?: Conceptual and Empirical Complexities of an
Operational Definition. Journal of Conflict Resolution 48 (6): 814-858.
Sanchez, Fabio, Andres Solimano, and Michel Formisano. 2005. Conflict, Violence, and
Crime in Colombia. In Understanding Civil War: Evidence and Analysis, edited by P.
Collier and N. Sambanis. Washington, DC: The World Bank.
Sandholtz, Wayne, and William Koetzle. 2000. Accounting for Corruption: Economic
Structure, Democracy, and Trade. International Studies Quarterly 44: 31-50.
Sarkees, Meredith Reid. 2000. The Correlates of War Data on War: An Update to 1997.
Conflict Management and Peace Science 18 (1): 123-144.
Sarvananthan, Muttukrishna. 2007. In Pursuit of a Mythical State of Tamil Eelam: A
Rejoinder to Kristian Stokke. Third World Quarterly 28 (6): 1185-1195.
Schelling, Thomas C. 1960[1980]. The Strategy of Conflict. 2 ed. Cambridge, MA: Harvard
University Press. Original edition, 1960.
Scheuren, Jean-Michel, Olivier le Polain, Regina Below, Debarati Guha-Sapir, and Sylvain
Ponserre. 2008. Annual Disaster Statistical Review: The Numbers and Trends 2007,
Center for Research on the Epidemiology of Disasters (CRED), Brussels.
Schneider, Stephen H., Sergui Semenov, Anand Patwardhan, Ian Burton, Chris H.D.
Magadza, Michael Oppenheimer, A. Barrie Pittock, Atiq Rahman, Joel B. Smith,
Avelino Suarez, and Farhana Yamin. 2007. Assessing Key Vulnerabilities and the
Risk from Climate Change. In Climate Change 2007: Impacts, Adaptation and
Vulnerability. Contribution of Working Group Ii to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change, edited by M. L. Parry, O. F. Canziani, J. P.
Palutikof, P. J. van der Linden and C. E. Hanson. Cambridge: Cambridge University
Press.
Schock, Kurt. 1996. A Conjectural Model of Political Conflict: The Impact of Opportunities
on the Relationship between Economic Inequality and Violent Political Conflict.
Journal of Conflict Resolution 40 (1): 93-133.
Schulze, Kirsten E. 2003. Ceasefire or More? The World Today 59 (1): 24-25.
———. 2004. The Free Aceh Movement (GAM): Anatomoy of a Separatist Organization.
Policy Studies 2, East-West Center, Washington, DC.
———. 2005. Aceh Rebels Exploiting Tsunami Tragedy. The Straits Times (Singapore), 27
January.

175
———. 2008. A Sensitive Mission: Monitoring Aceh's Agreement. In Reconfiguring Politics:
The Indonesia-Aceh Peace Process, edited by Aguswandi and J. Large. London:
Conciliation Resources.
Seawright, Jason, and John Gerring. 2008. Case Selection Techniques in Case Study
Research: A Menu of Qualitative and Quantitative Options. Political Research Quarterly
61 (2): 294-308.
Sen, Amartya. 1998. Mortality as an Indicator of Economic Success and Failure. Economic
Journal 108 (446): 1-25.
Sengupta, Somini. 2005a. Nearly a Year after the Tsunami, Sri Lanka Strife Flares. New York
Times, November 2.
———. 2005b. Sri Lanka: Court Blocks Tsunami Aid Deal. New York Times, July 16.
Shastri, Amita. 2003. Sri Lanka in 2002: Turning the Corner? Asian Survey 43 (1): 215-221.
Shefner, Jon. 1999. Pre- and Post-Disaster Political Instability and Contentious Supporters:
A Case Study of Political Ferment. International Journal of Mass Emergencies and Disasters
17 (2): 137-160.
Shleifer, Andrei, and Robert W. Vishny. 1993. Corruption. Quarterly Journal of Economics 108:
599-617.
Sigelman, Lee, and Miles Simpson. 1977. A Cross-National Test of the Linkage between
Economic Inequality and Political Violence. Journal of Conflict Resolution 21 (1): 105-
128.
Simile, Catherine M. 1995. Disaster Settings and Mobilization for Contentious Collective
Action: Case Studies of Hurricane Hugo and the Loma Prieta Earthquake. Ph.D.
Dissertation, University of Delaware.
Snyder, Richard. 2006. Does Lootable Wealth Breed Disorder? A Political Economy of
Extraction Framework. Comparative Political Studies 39 (8): 943-968.
Snyder, Richard, and Ravi Bhavnani. 2005. Diamonds, Blood, and Taxes: A Revenue-
Centered Framework for Explaining Political Order. Journal of Conflict Resolution 49
(4): 563-597.
Soesastro, Hadi, and Raymond Atje. 2005. Survey of Recent Developments. Bulletin of
Indonesian Economic Studies 41 (1): 5-34.
Stallings, Robert A., ed. 2003. Methods of Disaster Research. Newark, DE: International
Research Committee on Disasters.

176
Steele, Jonathan. 2005a. After the Deluge, the Tides of War Are Flooding Back. Guardian,
December 30.
———. 2005b. Tsunami Disaster: Private Largesse, Public Lethargy. Guardian, January 3.
———. 2005c. Tsunami Disaster: Relief Effort Is Propaganda Trick, Say Tamil Tigers.
Guardian, January 5.
Stephan, Maria J., and Erica Chenoweth. 2008. Why Civil Resistance Works: The Strategic
Logic of Nonviolent Conflict. International Security 33 (1): 7-44.
Stewart, Frances. 2002. Horizontal Inequalities: A Neglected Dimension of Development.
QEH Working Paper 81, Queen Elizabeth House, University of Oxford.
Stokke, Kristian. 2006. Building the Tamil Eelam State: Emerging State Institutions and
Forms of Governance in LTTE-Controlled Areas in Sri Lanka. Third World Quarterly
27 (6): 1021-1040.
———. 2007. War by Other Means: The Ltte's Strategy of Institutionalising Power Sharing
in the Context of Transition from War to Peace--a Reponse to Muttukrishna
Sarvananthan. Third World Quarterly 28 (6): 1197-1201.
Stokke, Kristian, and Anne Kirsti Ryntveit. 2000. The Struggle for Tamil Eelam in Sri Lanka.
Growth and Change 31 (2): 285-304.
Strand, Håvard. 2006. Onset of Armed Conflict: A New List for the Period 1946-2004, with
Applications. Unpublished Manuscript.
Sukma, Rizal. 2004. Security Operations in Aceh: Goals, Consequences, and Lessons. Policy
Studies 3, East-West Center, Washington, DC.
Sulistiyanto, P. 2001. Whither Aceh? Third World Quarterly 22 (3): 437-452.
Svensson, Jakob. 2000. Foreign Aid and Rent-Seeking. Journal of International Economics 51 (2):
437-461.
Tamils Rehabilitation Organization (TRO). 2005a. Eighteen Month Tsunami Report:
Audited Financial Report as of 31 December 2005, Tamils Rehabilitation
Organization, Kilinochchi, Sri Lanka.
———. 2005b. Report on Emergency Relief and Rehabilitation for Tsunami Affected Sri
Lanka, Tamils Rehabilitation Organization, Kilinochchi, Sri Lanka.
Tanner, Marcus. 2005. After the Flood...Peace: Rebels Lay Down Their Arms. The
Independent, 21 December.

177
Tarling, Nicholas, ed. 1999. The Cambridge History of Southeast Asia, Vol. IV: From WWII to the
Present. Cambridge: Cambridge University Press.
Tavera-Fenollosa, Ligia. 1999. The Movimiento De Damnificados: Democratic
Transformation of Citizenry and Government in Mexico City. In Subnational Politics
and Democratization in Mexico, edited by W. A. Cornelius, T. A. Eisenstadt and J.
Hindley. La Jolla, CA: Center for U.S.-Mexican Studies, University of California San
Diego.
Taydas, Zeynep, Dursun Peksen, and Patrick James. 2008. Why Do Civil Wars Occur?
Understanding the Importance of IQ. Unpublished Manuscript.
Thakur, Ramesh. 2005. Natural Disaster as the Spur to Regional Cooperation. Daily Yomiuri,
January 6.
The Government of the Republic of Indonesia, and The Free Aceh Movement. 2005.
Memorandum of Understanding between the Government of the Republic of
Indonesia and the Free Aceh Movement. Helsinki.
Thorbecke, Erik, and Chutatong Charumilind. 2002. Economic Inequality and Its
Socioeconomic Impact. World Development 30 (9): 1477-1495.
Tilly, Charles. 1978. From Mobilization to Revolution. Reading, MA: Addison-Wesley.
———. 2003. The Politics of Collective Violence, Cambridge Studies in Contentious Politics.
Cambridge, U.K.; New York: Cambridge University Press.
Toft, Monica D. 2003. The Geography of Ethnic Conflict: Identity, Interests, and the Indivisibility of
Territory. Princeton, NJ: Princeton University Press.
Toset, Hans Petter Wollebæk, Nils Petter Gleditsch, and Håvard Hegre. 2000. Shared Rivers
and Interstate Conflict. Political Geography 19 (8): 971-996.
Trenbert, Kevin E., Philip D. Jones, Peter Ambenje, Roxana Bojariu, David Easterling,
Albert Klein Tank, David Parker, Fatemeh Rahimzadeh, James A. Renwick, Matilde
Rusticucci, Brian Soden, and Panmao Zhai. 2007. Observations: Surface and
Atmospheric Climate Change. In Climate Change 2007: The Physical Science Basis.
Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel
on Climate Change, edited by S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis,
K. B. Averyt, M. Tignor and H. L. Miller. Cambridge: Cambridge University Press.
Tschoegl, Liz, Regina Below, and Debarati Guha-Sapir. 2006. An Analytical Review of
Selected Data Sets on Natural Disasters and Impacts. Paper presented at
UNDP/CRED Workshop on Improving Compilation of Reliable Data on Disaster
Occurrence and Impact, 2-4 April, Bangkok, Thailand.

178
Tullock, Gordon G. 1974. The Social Dilemma: The Economics of War and Revolutions. Blacksburg,
VA: Center for the Study of Public Choice.
Urdal, Henrik. 2006. A Clash of Generations? Youth Bulges and Political Violence.
International Studies Quarterly 50 (3): 607-629.
———. 2007. The Demographics of Political Violence: Youth Bulges, Insecurity and
Conflict. In Too Poor for Peace? Global Poverty, Conflict and Security in the 21st Century,
edited by L. Brainard and D. Chollet. Washington, DC: Brookings Institution Press.
Uyangoda, Jayadeva. 2005. Ethnic Conflict, the State and the Tsunami Disaster in Sri Lanka.
Inter-Asia Cultural Studies 6 (3): 341-352.
Van Belle, Douglas A., Jean-Sébastien Rioux, and David M. Potter. 2004. Media, Bureaucracies
and Foreign Aid: A Comparative Analysis of the United States, the United Kingdom, Canada,
France and Japan. New York: Palgrave Macmillan.
Vogel, Richard Michael. 1996. Regional Growth, Structural Change, and Natural Disaster.
Dissertation/Thesis, Florida International University.
Wade, Robert. 1998. The Asian Debt-and-Development Crisis of 1997-? Causes and
Consequences. World Development 26 (8): 1535-1553.
Walker, Herman, Jr., and Peter L. Hansen. 1946. Rural Local Government: Local
Government and Rainfall: The Problem of Local Government in the Northern
Great Plains. American Political Science Review 40 (6): 1113-1123.
Walsh, James I. 2007. Do States Play Signaling Games? Cooperation and Conflict 42 (4): 441-
459.
Walter, Barbara F. 2006a. Building Reputation: Why Governments Fight Some Separatists
but Not Others. American Journal of Political Science 50 (2): 313-330.
———. 2006b. Information, Uncertainty, and the Decision to Secede. International
Organization 60 (1): 105-135.
Warren, Mark E. 2002. What Does Corruption Mean in a Democracy? American Journal of
Political Science 48 (2): 328-343.
Wayland, S. 2004. Ethnonationalist Networks and Transnational Opportunities: The Sri
Lankan Tamil Diaspora. Review of International Studies 30 (3): 405-426.
Weinstein, Jeremy M. 2003. Inside Rebellion: The Political Economy of Rebel Organization.
Ph.D., Harvard University, Cambridge, MA.
———. 2005. Resources and the Information Problem in Rebel Recruitment. Journal of
Conflict Resolution 49 (4): 598-624.

179
———. 2007. Inside Rebellion: The Politics of Insurgent Violence. Cambridge: Cambridge
University Press.
Weinstein, Jeremy M., and Laudemiro Francisco. 2005. The Civil War in Mozambique: The
Balance between Internal and External Influences. In Understanding Civil War: Evidence
and Analysis, edited by P. Collier and N. Sambanis. Washington, DC: The World
Bank.
Wickramasinghe, Anoja, and Kamalini Fernando. 2006. Local Response: Sri Lanka. TEC:
The International Community's Fudning of the Tsunami Emergency and Relief,
Active Learning Network for Accountability and Performance in Humanitarian
Action (ALNAP), London.
Wilson, A. Jeyaratnam, ed. 2000. Sri Lankan Tamil Nationalism: Its Origins and Development in the
Nineteenth and Twentieth Centuries. Vancouver: University of British Columbia Press.
Wolensky, Robert, and Edward J. Miller. 1983. The Politics of Disaster Recovery. In Sociology
toward the Year 2000: The Sociological Galaxy, edited by C. E. Babbitt. Harrisburg, PA:
Beacon Press.
Woodwell, Douglas. 2005. The "Troubles" Of Northern Ireland: Civil Conflict in an
Economically Well-Developed State. In Understanding Civil War: Evidence and Analysis,
edited by P. Collier and N. Sambanis. Washington, DC: The World Bank.
Zagare, Frank C., and D. Marc Kilgour. 2000. Perfect Deterrence. Cambridge: Cambridge
University Press.
Zimmermann, Ekkart. 1980. Political Violence, Crises, and Revolutions. Cambridge: Schenkman.
Zinn, Annalisa. 2005. Theory Versus Reality: Civil War Onset and Avoidance in Nigeria
since 1960. In Understanding Civil War: Evidence and Analysis, edited by P. Collier and
N. Sambanis. Washington, DC: The World Bank.

Abstract (if available)

Abstract What is the effect of natural disasters on political violence? Despite the widespread speculation about the possibilities for peace that typically occurs after a natural disaster strikes in a conflict zone, relatively little is known about the way disasters impact political mechanisms. In this project, I find that post-disaster strategic dynamics around control of disaster relief institutions are likely to impact levels of grievance and feasibility and thus have important implications for the likelihood of post-disaster violence. Using time-series, cross-sectional data, I find that post-disaster violence is more likely where pre-disaster institutional quality is relatively weak, as weak institutions create incentives for institution capture. A qualitative comparison of Indonesia and Sri Lanka following the 2004 tsunami suggests that post-disaster violence also depends on patterns of institutional control at the time of the disaster. The disaster results in a “lock-in” effect, accelerating and magnifying the material and non-material benefits of institutional control. These changes alter levels of grievance and the opportunities for rebellion.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The political economy of oil, terrorism and institutional development in Iran and its impact on the Middle East

PDF

The political economy of R&D collaboration: micro- and macro-level implications

PDF

The political economy of implementation: intellectual property rights protection across the world

PDF

Perspectives on state capacity and the political geography of conflict

PDF

Identity salience and political violence: centering social identity in the study of intrastate conflict

PDF

The impact of social capital on economic performance: lessons from small and medium size enterprises (SMEs)

PDF

Campus colorlines: the changing boundaries of race within institutions of higher education in the post-civil rights era

PDF

The murky risk of trade protectionism in an interconnected and uncertain global economy: a state, industrial, and regional analysis

PDF

Social policy and presidential ideology in Latin America: the political economy of social spending and anti-poverty programs

Asset Metadata

Creator Enia, Jason S. (author)

Core Title Shaking the foundations of violent civil conflict: institutions, disasters, and the political economies of state-rebel interaction

Contributor Electronically uploaded by the author (provenance)

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Politics

Publication Date 08/04/2009

Defense Date 05/11/2009

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Aceh,disaster diplomacy,Indonesia,institutions,natural disasters,OAI-PMH Harvest,signaling,Sri Lanka,Tamil,tsunami,violent civil conflict

Place Name Indonesia (countries), Sri Lanka (countries)

Language English

Advisor James, Patrick (committee chair), Gordon, Peter (committee member), Katada, Saori N. (committee member), Odell, John S. (committee member)

Creator Email enia@usc.edu,jason.enia@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m2471

Unique identifier UC197490

Identifier etd-Enia-2944 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-174696 (legacy record id),usctheses-m2471 (legacy record id)

Legacy Identifier etd-Enia-2944.pdf

Dmrecord 174696

Document Type Dissertation

Rights Enia, Jason S.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu