 

 

 

 

 

 

 

 

: r V
‘ .IT’.‘I< . $030“
.wﬁ...ru..tlt..ﬁ . . 3....

c ‘ . r ¢

42: ,
5...“qu .
a. 3n ~ :
umﬁx.) r.
{(31

.
z

.

I... ﬁr

3:: e.-

V 11.2.02. ,
'11
a;

II it. .o- v
.p s. :....!m....
a... 1.5:... .
€133 .
dddig.“ nu.
5.4.ch
3.1 :1...
33.251 ..

I

Jr

31.. I:
'53,!
‘9‘]

z .
1.2.3.2

53.. 3.3.:
3...": .
)9. .5."
u.
.4
.-
It
.5: .

. ‘1‘ V
‘ ..... b:
ﬁlési.
a ”wait...
..

rzu,

x
|« {vhf-1H.
.:
V. .2
a .19... (353.54»
.. a- .tﬁuhn.‘
A.

\
‘1.

«“1an .
.:3.?.l.x.

II

\6.‘ ex“!
. ~ it}: . n,
In I. . u

. .5.»

$53!... .
511.; ‘.«.a...4n .
c...70 IE! .4

s ..
ué. n... u:
9 ‘1.-

{ct

».r(.r..r...,
.535?
..‘

ﬂuvxa

 

OYTCOlO

2
2002/

 

LIBﬁARY
Mlchlgan State
- Unlversity

 

 

 

This is to certify that the

dissertation entitled

De-Autocratization And Violent Outcomes: The Role
Of Strategic Interaction And Information

presented by

Renee B. Agress

has been accepted towards fulﬁllment
of the requirements for

Doctorate degreein Political Science

 

 

Ag:

Major professor

 

Date N‘VJOI. ”0/

MS U is an Afﬁrmative Action/Equal Opportunity Institution 0-12771

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c-JCIRC/DateDuepes-ptts

 

DE-AUTOCRATIZATION AND VIOLENT OUTCOMES: THE ROLE OF
STRATEGIC INTERACTION AND INFORMATION

By

Renee B. Agress

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Political Science

2001

ABSTRACT

DE-AUTOCRATIZATION AND VIOLENT OUTCOMES: THE ROLE OF
STRATEGIC INTERACTION AND INFORMATION

By

Renee B. Agress

The question posed in this research is a relatively straight forward
question: when will de—autocratization lead to internal conﬂict. De-autocratization
refers to the process by which an authoritarian regime transitions from its current
regime type to a less restrictive regime type, whether the new regime is still
authoritarian or embraces democracy. As such, the term de-autocratization
encompasses the process of liberalization as well as democratization. Many
scholars have focused on the process of democratization in an attempt to explain
under what conditions democratization will succeed. In doing so, these scholars
choose not to explore why democratization may fail. Additionally, research of this
nature assumes that authoritarian regimes are interested in democratizing. In
reality, many authoritarian regimes are willing to allow some degree of change
but are not willing to democratize, at least not initially. The goal of this
dissertation is to explore the process of de-autocratization in such a manner as to
better understand when the process will lead to failure, speciﬁcally internal
conﬂict.

In order to accomplish this goal, a series of game theoretic models are

developed. These models focus on the strategic nature of the interaction

between various actors. A non-cooperative bargaining game is developed to
explore the interaction between regime members. An incomplete and imperfect
information model is developed in order to better understand how the regime
interacts with civil society. It is vitally important to understand how the different
actors interact with each other. It has long been argued that the regime is not a
homogeneous collection of individuals but that the regime is composed of
individuals with different goals and preferences. Yet the vast majority of the
literature makes an implicit assumption that the regime acts as a single unit. If
we are to understand how de-autocratization occurs, it is necessary to
understand why the regime may choose to liberalize and how that decision is
reached. It is equally important to develop a more complete understanding of
how the regime interacts with civil society. The model developed in this research
allows civil society to choose to demand liberalization even if the regime does not
initially choose to open. This is necessary addition to past game theoretic work
in the ﬁeld of democratization as it more accurately represents how the process
occurs.

The second half of the research develops an empirical model which
explores the different outcomes of de—autocratization and the variables that
inﬂuence these outcomes. A multi-nomial logit model is developed with the
dependent variable broken into ﬁve categories: no conﬂict, coup, political unrest,
minor conﬂict, and civil war. The outcome “no conﬂict” may include cases where

liberalization, or even democratization, is successfully achieved.

Copyright by
Renee B. Agress
2001

For my parents

ACKNOWLEDGMENTS

There are many people that I wish to thank for their support over the last
eight years.

I begin by thanking my parents. They instilled in me the desire to learn
and the love of education. I don’t know how far I would have gone if my mother
had not been willing to dedicate four hours a night to helping me with word
problems.

My brothers served as excellent role models for me. They would remind
me not to take myself too seriously yet always let me know they were proud of
my accomplishments. Their support was invaluable while completing my
educaﬁon.

I would like to thank Dr. Robert Benedetti and Dr. Brian Klunk at the
University of the Paciﬁc. Without their encouragement I would not have
contemplated graduate school.

I would like to thank Scott Gates for his honesty and listening when I
needed to talk to someone about school, research, or Babylon 5. In addition,
Scott introduced me to Thai, Korean, Indian, and Middle Eastern food. Strangely
enough, he never sought out a Norwegian restaurant when we were going out to
eat...

Dr. Brian Silver, Dr. Dennis Patterson, Dr. Kenneth Williams, and Dr.
Gretchen Hower all played an important role in my education. Gretchen
introduced me to the ﬁeld of International Relations and a wide variety of theories

focusing on ethnic conﬂict. After reading Ed Azar’s work in PLS 860, I knew that

vi

I wanted to study internal conﬂict. Brian, Dennis, and Ken provided numerous
comments on my dissertation. Without their help the research contained in this
document would not have been possible.

I could not have survived graduate school with out the support of Sara
McLaughlin Mitchell, Chris Butler, Chris Sprecher, Mark Souva, Dave Lektzian,
Brandon Prins, Misa Nishikawa, Jeff Reno, Damon Linker, Malia Reddick,
Michelle Kuenzi, Jon Monroe, and John U. Davis. Only Sara could have saved
me from Michelle's killer tomatoes and cheese bread. Chris Butler and Sprecher
spent many an hour listening to my discussion of ethnic conﬂict, civil wars, and
methods over coffee. Their comment on my work was invaluable, I hope that

they learned as much from me as I did them.

vii

TABLE OF CONTENTS

LIST OF TABLES ...................................................................................
LIST OF FIGURES .................................................................................

CHAPTER 1.
INTRODUCTION ...................................................................................
DEMOCRATIZATION VS. DE-AUTOCRATIZATION .................

CHAPTER 2.
SYNTHESIZING THREE LITERATURES .............................................
LITERATURE REVIEW ..............................................................
INTERNAL CONFLICT: PEASANT REVOLUTION .........
THE PATH-DEPENDENT APPROACH TO DE-
AUTOCRATIZATION .......................................................
AUTHORITARIAN TRANSITION THROUGH THE EYES
OF GAME THEORISTS ...................................................
WHAT IS GAINED FROM THIS RESEARCH ............................
THE ROLE OF STRATEGIC INTERACTION .............................
STRATEGIC INTERACTION WITHIN THE REGIME ......
STRATEGIC INTERACTION BETWEEN THE REGIME
AND CIVIL SOCIETY .......................................................

CHAPTER 3.
STRATEGIC INTERACTION AMONG REGIME MEMBERS ................
WHAT IS THE REGIME? ............................................................
THE POLITICAL AND ECONOMIC DIMENSIONS ....................
WHY OPEN? ....................................................................
POLITICAL DIMENSIONS ...............................................
ECONOMIC DIMENSIONS ..............................................
THE GAME THEORETIC MODEL ..............................................
DISCUSSION OF THE OUTSIDE OPTION .....................
THE FACTIONS ...............................................................
OUTCOMES AND PAYOFFS ..........................................
ORDER OF PLAY ............................................................
IMPLICATIONS OF THE MODEL ....................................

CHAPTER 4.

“—3

cocoon

12

17
20
21
22

27

29
29
31
34
34
35
38
40
43
46
47
48

STRATEGIC INTERACTION BETWEEN THR AUTHORITARIAN REGIME

AND CIVIL SOCIETY .............................................................................
THE ROLE OF CIVIL SOCIETY ..................................................

CIVIL SOCIETY AS DEPICTED IN THE LITERATURE...

CIVIL SOCIETIES TYPE ..................................................

viii

51
52
53
56

THE ROLE OF THE REGIME ..................................................... 57
THE REGIME AS DEPICTED IN THE LITEATURE ......... 58
THE REGIMES TYPE ....................................................... 60
INTERACTION BETWEEN THE REGIME AND CIVIL SOCIETY 61
POTENTIAL PITFALLS ON THE PATHWAY: EXPLORING THE
VARIOUS OUTCOMES OF REGIME TRANSITION .................... 63
THE GAME THEORETIC MODEL ............................................... 66
THE ACTORS .................................................................... 67
THE ORDER OF PLAY ...................................................... 68
OUTCOMES ....................................................................... 70
ASSUMPTIONS ................................................................. 75
COSTS ............................................................................... 77
PREFERENCE ORDERINGS ............................................ 79
SOLVE THE COMPLETE AND PERFECT INFORMATION
GAME ................................................................................. 83
THE INCOMPLETE AND IMPERFECT INFORMATION
GAME ................................................................................. 85
SOLVING THE INCOMPLETE AND IMPERFECT
INFORMATION GAME ........................................................ 88
CONCLUSION ................................................................................ 90
CHAPTER 5.
OPERATIONALIZATION ........................................................................... 92
UNIVERSE OF CASES .................................................................. 92
THE DATA SETS ........................................................................... 94
THE VARIABLES ........................................................................... 97
THE DEPENDENT VARIABLE: OUTCOME ....................... 97
THE INDEPENDENT VARIABLES ................................................ 101
REGIME TYPE .................................................................... 101
MEASURES OF TRANSITION ............................................ 101
ETHNIC FRAGMENTATION ........................... ' .................... 103
CODING RULES ............................................................................ 105
CONCLUSION ................................................................................ 108
CHAPTER 6.
EMPIRICAL FINDINGS ............................................................................. 109
HYPOTHESES ............................................................................... 108
THE EMPIRICAL MODEL .............................................................. 111
DATA ANALYSIS ................................................................. 1 11
CONCLUSION ................................................................................ 115
WHAT HAVE WE LEARNED? ........................................................ 116
BIBLIOGRAPHY ........................................................................................ 120

LIST OF TABLES

Table 4.1 List of Abbreviations for the Complete and Perfect

lnforrnation Game .................................................................................... 67
Table 4.2 A List of Cost Terms ........................................................... 77
Table 4.3 Establishing the Preference Orderings ............................... 80
Table 4.4 Cardinal Payoffs for Each Actor ......................................... 84

Table 6.1 Multi-Nomial Logit Results for all Cases Where
De-Autocratization is Attempted ............................................................... 1 14

LIST OF FIGURES

Figure 3.1 Non-Cooperative Bargaining Game ...................................... 44
Figure 3.2 A Spatial Representation of the Bargaining Area ................. 46
Figure 4.1 Complete and Perfect Information Game ............................. 66
Figure 4.2 Backwards Induction for the Complete and Perfect

lnforrnation Game ...................................................................................... 84
Figure 4.3 The Incomplete and Imperfect Information Game ............... 87

xi

xii

Chapter 1

INTRODUCTION

While watching the evening news in recent years one would have to note the
lack of militarized conﬂict between states. This lack of international conﬂict has
led to questions from students and family members as to whether there is a need
for scholars to continue to study war. The answer to this question is relatively
easy. The lack of international war today does not mean that there will not be a
war tomorrow.

Perhaps more importantly, while it may be the case that the international
system has been devoid of interstate conﬂict in recent years, there has not been
an absence of internal conﬂict in this time period. Rwanda, Bosnia, Kosovo, East
Timor, Sierra Leone, Russia, and Columbia are but a few examples of countries
that have struggled with internal conﬂict in recent years.

The research question that this dissertation seeks to answer is: How does
the process of democratization inﬂuence the likelihood of internal conﬂict in a
given country?

A great deal of existing literature in the ﬁelds of comparative politics and
international relations attempts to address the question of what causes internal
conﬂict. Many of these studies focus on one type of internal conﬂict, such as civil
war, coups, or everyday forms of resistance. It is rare that scholars explore how
a country may ﬁnd itself facing a civil war and not a military coup or political

unrest, such as riots and violent protests. One of the goals of this research is to

demonstrate whether different domestic factors will inﬂuence what type of
internal violence -a coup, political violence, or civil war— a country may ﬁnd itself
facing.

While the question of internal conﬂict is in itself interesting, this research
will focus on the outbreak of violence when a country attempts to democratize.
Why? Nigeria, Guatemala, Nicaragua, Columbia, El Salvador, Russia,
Yugoslavia, South Korea, Peru, Ecuador, and South Africa are but a few
examples of countries that have attempted democratization and, because of this,
have faced some type of internal conﬂict. It is not that the current literature fails
to recognize that violent outcomes are possible, but that most of the literature is
interested in explaining the cases of successful democratization, without
considering how democratization may lead to internal conﬂict. Much of the
literature dichotomizes democratization into cases where it succeeds and cases
where it failed. In reality, failure occurs for many different reasons and in many
different ways. If we are to understand why democratization fails, it is essential
that we understand the different types of failure and how a country might ﬁnd
itself facing a coup, an increase in authoritarianism, or a civil war.

Additionally, there is an implicit assumption by many scholars and American
decision-makers that democracy is the most desirable form of government.
O’Donnell and Schmitter state, “The first general and shared theme is normative,
namely, that the instaturation and eventual consolidation of political democracy
constitutes per se a desirable goal.” (1986, pg.3) This being the case, the

approach taken to the study of democratization has been biased toward

understanding how democratization succeeds, with less emphasis placed on the
question of how democratization fails.

While coups, increased repression, and civil war are all violent outcomes,
they are different outcomes. A coup occurs when one part of the ruling regime,
deﬁned as the decision-makers who determine policy for the country, attempts to
remove another part of the regime through the use of force. An increase of
repression, or an increase in authoritarianism, occurs when the regime chooses
to use force to silence or weaken some segment of civil society. Civil war occurs
when there is armed conﬂict between different factions within a country and there
are at least 1,000 battle deaths as a result of this conﬂict.

By paying attention to the difference in failures, we are forced to make more
of an effort to understand how a country ﬁnds itself facing a civil war and not a
coup. This focus forces us to develop a better understanding of why
democratization fails, the different types of failures, and how internal conﬂict is

one byproduct on this failure.

Democratization vs. De-autocratization

I contend that the manner in which democratization— in international
relations and comparative politics- is studied has led to a dichotomization of the
outcomes. The very fact that the concept being studied is referred to as
democratization leads to an emphasis on understanding how democracy
develops. The study then focuses on democratization or not democratization as

possible outcomes.

In his book, The Third Wave: Democratization in the Late Twentieth
Century (1991), Huntington discusses waves of democratization and reverse
waves of democratization. In his discussion of the reverse waves, Huntington
lists countries where the reversal came through in the from of a coup, but he
does not attempt to explain why some countries may ﬁnd themselves facing a
coup, others facing a civil war, and still others simply revert to an authoritarian
regime without violence. The discussion of reverse waves can be found in pages
17-21 and 290-294, 10 pages in a book that is 316 pages long. Clearly, then, the
emphasis is on explaining what, why, and how of democratization with a limited
discussion about the causes of its failures. While Huntington is not the only
scholar who writes on democratization, and this one book does not embody the
entirety of his work or the entire literature, it serves as an example of how
focused the democratization literature is on understanding success while
assuming that failure is the opposite of success.

I believe that a part of the problem is the use of the term democratization.
Huntington deﬁnes democratization as “the end of the nondemocratic regime, the
inauguration of the democratic regime, and the consolidation of the democratic
system.” (pg. 9) Huntington acknowledges that there is a difference between
democratization and liberalization, “...a partial opening of an authoritarian system
short of choosing governmental leaders through freely competitive elections.”
(pg. 9) He also believes that it is difﬁcult to determine if a leader or regime is
initially interested in liberalization and democratization and includes cases such

as South Africa in 1978 and the Soviet Union in 1985 in his research. (pg. 131-

134) Huntington is not the only scholar to make note of the difference between
liberalization and democratization. O'Donnell and Schmitter and Przeworski also
acknowledge this difference.

The debate between democratization and liberalization and how to tell when
a country is liberalizing or democratizing could be solved if scholars were to
focus on the overall process of de-autocratization. I deﬁne de-autocratization as
the process by which a country ﬁnds itself decreasing the control of an
authoritarian regime. The study of de-autocratization would allow for the including
of countries that were liberalizing as well as countries that are democratizing.

Focusing on the process of de-autocratization would prevent scholars

from falling into the trap of dichotomizing the outcome of the process. In much of
the democratization literature the emphasis is placed on whether or not
democratization succeeds, many times limiting the possible outcomes to
democratization fails or democratization succeeds. If scholars were to focus on
de-autocratization, the number and types of outcomes would change. The
development of a consolidated democracy would be one outcome. Additional
outcomes would include decreased authoritarian control and reversal of de-
autocratization. Perhaps most important, de-autocratization would allow scholars
to begin to unpack the many outcomes included in “failed democratization.”
Outcomes such as civil war, coups, political unrest, terrorism, and the like would
receive equal time with the outcome of successful democratization.

Why is this important? Huntington includes the Soviet Union in 1985 in his

discussion of the “Third Wave” of democratization. Since that time period, the

Soviet Union has ceased to exist with many of the Soviet republics becoming
independent countries in their own right, a move initially resisted by the ruling
regime headed by Gorbachev. More recently, violence has erupted in Chechnya,
where a portion of the population has pursued the goal of autonomy from the
Russian government. This violence can easily be attributed to the trials and
tribulations that are part of the de-autocratization process. But Huntington’s
research does not discuss how to prevent civil war, only the steps needed to
prevent the process of democratization from failing.

Does the re-occuring violence in Chechnya represent the failure of
democratization in Russia? Regular elections have been held and the results
have been honored. The elected legislature meets regularly and passes
legislation. New political parties have been formed, and free speech is permitted.
It is unclear to me how to code the case of Russia using Huntington’s criteria:
has it successfully democratized or not? The presence of internal conﬂict would
indicate that some portion of civil society does not see the democratically elected
ofﬁcials in Moscow as legitimate, even in the presence of regular elections. The
problem here is that the number of outcomes in the study of democratization has
been limited.

The very nature of the study of de—autocratization requires that scholars are
able to explain not only successful democratization or peaceful liberalization,
such as in Iran, but also the violent outcomes such as coups, riots, terrorism, and
civil war. Because of this research interest in violent outcomes, and wanting to

insure that the research can accommodate liberalization as well as

democratization, the research question addressed will be: How does de-

autocratization inﬂuence the likelihood of violent outcomes?

Chapter 2

SYNTHESIZING THREE LITERATURES

This research seeks to explain how the process of de-autocratization can
lead to violent outcomes. To answer this question, it is necessary to review
brieﬂy the existing literature on democratization and internal conﬂict. A
discussion of how this research differs from past research and what this research
adds to the existing literature will follow the literature review. Finally, a review of
the methods that will be employed while conducting this research and a chapter

guide will be provided.

Literature Review

The research question under consideration in this work asks how does de-
autocratization lead to violent outcomes? As such, the literature discussed below
addresses some part of this question. While there are many scholars and
volumes of literatures whose work could be included in this discussion, the
author has limited the discussion to the peasant revolution literature,
democratization, and game theoretic attempts to explain democratization.

Additional literature will be discussed in future chapters as needed.

lntemal Conﬂict: Peasant Revolution

One literature that focuses on the role of civil society in the face of political
and economic change is the Peasant Revolution literature. This literature
examines how peasants respond to change, what tools peasants have available,
and how conﬂict arises in these societies. This literature is important to the study
of de-autocratization as it focuses on the behavior of civil society in countries with
an authoritarian regime. Additionally, it explores how economic shocks and
changes can lead to internal conﬂict. The impact of economic change is one
variable that scholars in the democratization literature agree inﬂuence the
process of democratization. In reading the peasant revolution literature, we gain
insight into civil societies ability to respond to change, which is vitally important, if
we are to understand how de-autocratization can lead to internal conﬂict.

In The Moral Economy of the Peasant, James Scott explores why
peasants resort to violence and why the violence rarely leads to change. Scott
focuses on the role of the subsistence ethic in understanding peasant politics. He
argues that peasants are primarily interested in survival and will resort to
violence only when the village's survival is threatened. To survive, peasants
have developed a “subsistence ethic” that includes the notion of reciprocity,
forced generosity, communal land, and work sharing. (Scott, pg. 3) Landowners
are expected not to collect rent if the peasant has been unable to produce
encugh to survive, to provide snacks to workers in the ﬁelds, and to provide food

for struggling families.

Scott argues that the subsistence ethic clashes with capitalism and that
this clash has led to increased conﬂict in pre-capitalist societies. As landowners
begin to adopt capitalist practices, such forcing peasants as pay rent or
preventing them from using communal land, the subsistence ethic is endangered.
Yet these actions are not sufﬁcient for rebellion. Scott states that a substantial
shock is necessary for the peasant to rebel. (Scott, pg. 194)

Focusing on Southeast Asia, Scott points to three sources of income
variability that can lead to rebellion: natural yield ﬂuctuations, world market
ﬂuctuations, and mono-crop price ﬂuctuations. Scott’s research points to the role
of economic change in pre-capitalist societies. When conditions are right, the
regimes attempt to shift from a traditional economy to capitalist economy, which
can lead to conﬂict. With the subsistence ethic under attack due to the
developing capitalist economy, an economic shock can place peasants in a
position where they are unable to survive. If the landowner is not willing to help
the peasants, the peasants may choose to rebel. As such, it is the change in the
economic regime combined with economic pressures outside the peasants’
control of the peasants that can lead to internal conﬂict.

Benedict Kerkliet discusses the role of everyday forms of resistance in the
Philippines. Everyday forms of resistance are deﬁned as “considerable
resistance by subordinate villages against the claims on them by wealthier
people, capitalists, and the government and struggle for their own claims of what

should be theirs.” (Kerkliet, pg. 245) Kerkliet argues that this resistance is rarely

10

organized and attempts to remain hidden. Everyday resistance is important in

four ways. First

“It may supplement what poor people have to live on...Second, it says that
the absence of obvious discontent does not necessarily mean people are
content or that stability prevails... Ordinary resistance is central to political
discourse over the proper use and distribution of resources and how
people should treat each other... Fourth, peoples’ subdued, unobtrusive
discontent may also be the basis, in combination with other
circumstances, for organized confrontational resistance” (Kerkliet, pg. 246-

247)

Kerkliet looks at how the values of capitalism clash with the traditional
values of the village. Capitalism appears to ignore two important values found in
Philippine society- that people with more should help people with less and that
peoples basic needs must be satisﬁed. When these principles are ignored,
conﬂict is likely. Like Scott, Kerkliet ﬁnds that peasant revolution rarely

succeeds.

ln Peasants Politics and Revolution: Pressures Toward Political and

 

 

Social Change in the Third World. Joel Migdal states “the twentieth century has
been the century of peasant revolution” (Migdal, pg 226). Migdal focuses on how
revolutionary organizations develop and garner support from peasants. He
hypothesizes “that peasants participation in institutionalized revolutionary
movements is an attempt on their part to solve certain individual and local
problems through the immediate selective incentives offered by the revolutionary
organization.” (Migdal, pg 227) Peasant support is gained with the promise of
individual rewards. The peasants are faced with economic problems, normally
stemming from an attempt to adopt a new economic system. To solve these

problems, peasants may choose to join a revolutionary organization. If a political

11

leader is able to garner enough support from the peasants, then the leader may
be successful in overthrowing the existing political regime, according to Migdal.

The peasant revolution literature stresses the role of economic change in
order to explain why peasants might attempt to revolt. As a whole, the literature
ﬁnds that most revolutions are localized, poorly organized, and fail. Migdal is one
of the few scholars to attempt to discuss how peasants may become organized
and how a revolutionary leader can keep peasants involved in the struggle
against the current political regime. This literature is important if we are to
understand how de-autocratization can lead to lntemal conﬂict as it focuses on
the role of peasants, a large part of civil society in many non-democratic
countries, and what it takes to mobilized them.

The democratization literature is essential to the study of de-
autocratization. Democratization itself is one possible outcome of de-
autocratization. Additionally, the democratization literature provides many
valuable insights into what variables inﬂuence a regimes shift away from an
authoritarian regime. As such, the discussion below will focus on two
approaches to democratization. The ﬁrst is the path-dependent approach. The
second segment looks speciﬁcally at how game theory has been used in the

study of democratization.

The Path-dependent Approach to Democratization

To date, the majority of the research on this question has focused on de-

autocratization as a path—dependent process. This means that focus has been

12

placed on the major events and what variables affect these events. While there
has been a discussion of the actors involved in the process, it has been limited to '
interaction between civil society and the regime, leaving the interactions that
occur within the regime and civil society largely untouched.

Przeworski (1991), Agress and Gates (1999), Huntington (1991), Rustow
(1970), Dahl (1989), Bermeo (1992), and Casper and Taylor (1996), as well as
others, view democratization in this manner. There is an initial struggle within the
authoritarian regime; it can be violent or non-violent, which leads to the regime
allowing increased political, economic, and social liberties. This in turn leads to
civil society adjusting to the new openness of the regime and choosing to accept
what is given to them or to seek additional liberties. The regime can choose to
accept civil society’s, advances or it can chose to halt the process. Civil society
can then choose to accept this, or to challenge the regime.

It is not that the path described above is inappropriate or inaccurate. I
adopted this path while looking at the interaction between civil society and the
regime. By focusing on the events and the variables that inﬂuence these events,
we have limited our focus too dramatically. The body of literature to date has put
forward a minimum of 27 variables believed to inﬂuence democratization, with
many of the variables contradicting each other.1 If we are able to move away
from a path-dependent approach, then we can develop a more parsimonious

model. The path—dependent approach implicitly assumes that a variables

 

‘ Huntington (1991) has compiled a list of the most common variables in the literature, which he
presents in The Third Wave: Democratization in the Late Twentieth Century. A few of the
variables listed are the absence of feudalism in society, a feudal aristocracy at some point in the
history of society, communal homogeneity, and communal heterogeneity.

13

inﬂuence will last over the span of the entire process. In doing so, the impact of
a given variable is most likely weakened. How so?

Economic variables are included in most studies of democratization.
Huntington (1991) maintains that there are three ways that economic factors
inﬂuence the process of democratization: economic crises undermined the
authoritarian regimes, sufﬁciently high levels of economic development had been
achieved allowing democratization, and rapid economic growth destabilized the
authoritarian regime. Economic variables, then, behaved differently in many
cases. Economic crisis, a sudden down turn in a country’s economy,
undermined the authoritarian regime, a negative relationship. The slow or rapid
growth of the economy weakened the authoritarian regime, a positive
relationship.

More important, the variable is looked at over the entire course of
democratization. A sudden economic crisis or sudden economic growth serves
as the initial catalyst starting the process. It inﬂuences the interaction of regime
members among one another and it inﬂuences civil societies' internal interaction.
A sudden crisis could be cause for the reformers taking control of the regime and
beginning the process of de-autocratization as well as serving as the straw that
broke civil society’s backs and helped to unite some segments, at least
temporarily. What happens if there is another change in the economy during the
process? For example, there is a sudden improvement in the economy. Vlﬁll the
parts of civil society remain united or will some members choose to leave the

movement because they have lost their reason for participating? Will the regime

14

stop making reforms because it feels that the crisis is passed and it no longer
needs to placate civil society to remain in control? By focusing on a variable over
the entire process of democratization, a variable’s impact is weakened and not
properly acknowledged. If the focus is turned to understanding the interaction
between actors, we can allow the behavior of variables to alter, more accurately
reﬂecting what inﬂuence the variables have.

Finally, the path-dependent approach looks at democratization as one
continuous process, when in reality it is the culmination of a series of interactions
over time. The regime and civil society do not interact one time; they interact on
many occasions and over a variety of issues. Within the regime and civil society,
there are discussions and actions taken that can inﬂuence the outcome. The
outcomes from previous interactions will affect how each side behaves in later
interactions. Did the regime back down on some issue? Does that mean it is
weak and that civil society can exploit this? Did civil society push for some
concession that caused the hard-liners in the regime to attempt an lntemal power
struggle in order to halt the transition? By looking at de-autocratization as a
series of repeated events, we are able to understand better how and why
miscommunication between regime and civil society can occur, therefore
understanding why the violent outcomes, such as civil wars, military coups, and
increases in authoritarian control, can occur.

When we explore the process of authoritarian transition through the
different types of strategic interactions, it becomes clear that the process is not

path determined and that in looking at authoritarian transition in such a way, we

15

limit our understanding of the process and our ability to make predictions. By
focusing research on these series of interactions, we will be able to understand
better how the process of authoritarian transition works— while accounting for
past experiences and the unique impact some variables will have on the role of
various actors in the process.

If we assume that both the regime and civil society prefer to achieve their
goals without the use of force, we would expect that de—autocratization would
rarely be faced with a coup or civil war. The regime would only open when it
knew that civil society would not challenge it. Civil society would only mobilize
when it knew that the regime would not use force to stop the process. This
requires that both the regime and civil society know how one will react to the
other’s next move, requiring complete and perfect information. The repeated
interaction between members of the regime and civil society prevents this type of
knowledge from existing, allowing for both actors to misjudge how the other will
react to its decision. Civil society mobilizes believing that the regime will not
respond with force and ﬁnds itself involved in a civil war because the regime is
not willing to allow further liberalization.

The scholars who have studied democratization as a path-dependent
approach have described many of the scenarios that l have laid out above, but
none have attempted explicitly to model that interaction between the regime
members and within civil society. By focusing on these sets of strategic

interactions, as well as the actual interaction between civil society and the

16

regime, this research will help to develop a better understanding of how a country

ﬁnds itself facing a civil war.

Authoritan'an Transition Through the Eyes of Game Theorists

Scholars have approached the question of democratization by attempting
to identify the causes before understanding how the process works. Before we
can begin to focus on the variables that affect the process, it is necessary that we
develop an understanding of how the mechanism itself works. With that
understanding, we will be able to understand better who and what is being
inﬂuenced. Then we can address how they are being inﬂuenced. This research
is intended to develop a blueprint of authoritarian transition by focusing on the
strategic interactions within the regime, the collective action problem that civil
society faces, and the strategic interaction between civil society and the regime.
If successful, this will help scholars better understand why such divergent
outcomes are common and how these outcomes are reached.

Przeworski (1991), Gates and Humes (1997), and Agress and Gates
(1999) begin to formalize the process of de-autocratization, which allows us to
explore the more speciﬁc interactions between actors. The focus of the above
works is on the interaction between the authoritarian regime and civil society.
Each of these works allows for different regime types, although all assume in
their formal models that civil society is one large actor with aggregated
preferences.

Przeworski’s (1991) model is a ﬁrst attempt at better understanding how

authoritarian transitions proceed. Przeworski identiﬁes different regime types,

17

noting that liberalizers will open while hard-liners will remain closed, and different
types of civil society, differentiating between moderates and radicals. In his
discussion on civil society, Przeworski allows a dialogue to occur among the
different types within civil society. The regime will react based on what civil
society chooses to do. Przeworski acknowledges that civil society is not one
large entity. While he does not include different types of civil society in his game,
his inclusion of multiple actors in the discussion signals the need to acknowledge
civil society as a more complex actor. Przeworski’s model lays a solid
foundation for the use of game theoretic models in the study of authoritarian
transitions and provides both a theoretic and game theoretic model for future
work.

Przeworski’s model is an early attempt to demonstrate the dynamic nature
of the transition process. In allowing interaction between the regime and civil
society, he explicitly models the strategic nature of authoritarian transition in
order to derive a series of propositions about when the regime will choose to
open and how civil society should react to the regime's decision.

Gates and Humes (1997) build on Przeworski’s (1991) model by allowing
for incomplete and imperfect information.2 Using Przeworski’s model, they are
able to demonstrate that including incomplete and imperfect information in the
model would strengthen his ﬁndings. Not only do the results improve, but also the
inclusion of incomplete and imperfect information more accurately reﬂects the

real world. Civil society does not know how strong the authoritarian regime is.

18

Civil society may be able to read signals sent by the regime, but it does not know
with certainty how devoted the regime is to reform or the maintenance of the
status quo. In addition, Gates and Humes add a nature node at the start of the
game that determines if the state is more or less committed to liberalization.
These changes allow Gates and Humes to develop Przeworski’s model further
while more closely reﬂecting how the process of democratization works. Like
Przeworski, they do not model the interaction that occurs between members of
the regime and the different segment of civil society.

Agress and Gates (1999) further develop the Przeworski model by
allowing for nature to determine if the regime is weak or strong and then allowing
civil society the opportunity to choose to mobilize— even when the regime
chooses to remain closed. Additionally, Agress and Gates model the initial
decision to open or not open as a two sided incomplete information game,
increasing the opportunity for both sides to erroneously judge the type of
opponent they are facing. While this is the ﬁrst model to allow civil society the
opportunity to mobilize no matter what the regime does, civil society is still
treated as an aggregated actor.

The above research fails to take into consideration that the transition from
an authoritarian regime is not a path-dependent process, but that the interactions
between actors occur repeatedly. The interaction between the regime and civil

society is not a one-shot game but repeats itself as the two actors negotiate over

 

2 It is important to note that Gates and Humes are interested in demonstrating houw game
theoretic results are strengthened when including the appropriate types of information sets and
not to develop a new theory of authoritarian transition.

19

many issues.3 Additionally, these works do not take into consideration the
struggle within the authoritarian regime or the collective action problems faced by
civil society when modeling the process.

Additionally, many of the above works imply that, at various points along
the way, the process of democratization can stall. If the country begins the
process at a later date, it is treated as a separate event and the process is traced
from this new start date to its conclusion, whatever that may be. While
convenient for analytical purposes, the assumption that democratization is a
linear process is an inaccurate one. None of these scholars dismisses the role of
history in predicting whether democratization succeeds or fails. But the manner in
which the process is approached greatly discounts the effects that past attempts
at democratization have on more recent attempts. By viewing democratization
as a series of interactions among many different actors, we can allow for multiple
iteration of these interactions if needed which is a more accurate portrait of the

process.

What is Gained from this Research?

This research seeks to answer the question: How does de-autocratization
lead to violent outcomes? To answer this, it is necessary to approach the
question in a different manner then previous research. First, it is important to
understand that de-autocratization is a process that involves many actors. This is

not a point lost on any of the scholars whose work is discussed in the previous

 

3 Huntington (1991) identiﬁes many issues that a transitional regime must consider. He spends
considerable time on the problem of how to deal with former leaders who committed human rights
violations, the military, and other such issues of consolidation.

20

section; all include a discussion of the regime and civil society. What has been
missing from the literature is a discussion of the interaction among the different
actors. Before the regime interacts with civil society, regime members must
decide what they are willing to offer civil society. This process can be
contentious and lead to a faction in the regime choosing to attempt to remove
another faction through a coup. Yet the interaction between the regime members
is something that is not fully explored in much of the democratization literature.
Those who do, such as Casper and Taylor, fail to include the possibility of a
coup.

This research will explore authoritarian transition by looking at two distinct
sets of strategic interactions: the power struggle within the regime and bargaining
between the regime and civil society.4 Each of these strategic interactions will

inﬂuence the outcome of the process of de-autocratization.

The Role of Strategic Interaction

Why is this important? In treating past attempts at democratization as a
separate variable, the path-dependent approach weakens the impact these
experiences have on future attempts. If the initial attempt at authoritarian
transition ends with a successful coup, future reformers will keep that in mind

when contemplating transition or during a future transition attempt. There is a

 

‘ The collective action problem facing civil society is an additional set of interactions that must be
explored if we are to fully understand the nature of de-autocratization and how the process can
lead to violent outcomes. The complexity of this collective action problem forces the author to
address this particular set of interactions in future research.

21

direct impact on the reformer's perception of how the hard-liners will react to
certain changes in the regime structure.

By focusing on strategic interactions, we can develop a better idea of what
affect a coup will have on future attempts at democratization. Such an event will
affect how strongly regime members believe their compatriots are motivated in
maintaining an authoritarian regime or how far democratization can be allowed to
proceed before part of the authoritarian regime becomes skittish. The impact on
the regime members is accounted for directly in a model focusing on strategic
interactions and allows for a better understanding of how the process is affected.
By focusing on the types of strategic interactions that commonly occur during the
process of authoritarian transition, we develop a better understanding for what
variables will inﬂuence the different actors and in what way these variables will
inﬂuence that actors.

Finally, focusing on strategic interactions permits us to account for
repeated interaction between the actors. By taking into consideration multiple
iterations, the research more closely models how authoritarian transitions occur.
In reality, these different strategic interactions occur throughout an authoritarian
transition; civil society and the regime negotiate over many different issues. The
regime faces the possibility of internal power struggles throughout the process.
Different segments of civil society can splinter and form new groups, forcing civil
society to address the collective action problem again. While a case study

approach to democratization allows for these possibilities to be explored, path-

22

dependent models make it much more difﬁcult to understand what result multiple

iterations may have on an authoritarian transition.

Strategic Interaction Within The Regime

The ﬁrst series of interactions that will be addressed is the interaction
between members of the authoritarian regime. While it is common to think of the
regime as one uniﬁed actor, this is not the case.5 The regime is composed of
individual decision-makers who head the various branches of government.
Together these decision-makers determine and attempt to implement policy for
the country. The number of regime members varies in each country. As rational
actors, regime members have their own set of preferences, goals, and agendas,
which do not always agree with one another.

For the purpose of this research, I argue that regime members are either
reformers or hard-liners. A reformer is interested in liberalization. There can be
varying degrees of interest in liberalization, ranging from the desire to allow a
limited reform, such as granting language rights or allowing independent
newspapers, to someone interested in democratization. While it is possible for
reformers to have different interests, they share a desire to open the country to

some degree. Before suggesting that reform be attempted, they reach some

 

5 Huntington (1991 ), O’Donnell and Schmitter, Agress and Gates, and Przeworski are a few
examples of authors who assume that the regime is a uniﬁed actor. Huntington includes the
regimes actions in his discussion of democratization but does not elaborate on the possibility of
conﬂict within the regime. O’Donnell and Schmitter discuss the regimes decision to enter into a
pact with civil society but do not discuss how the regime reaches the decision. Agress and Gates
refer to the regime as effective and ineffective, based on their ability to maintain control in the
country but do not discuss how this is directly affected by internal conﬂict. Przeworski discusses
different regime types but he does not explore how the lntemal dynamics will affect the outcome

23

agreement on what type of reform they want. Reformers who are interested in
democratization will be willing to ally themselves with other reformers in order to
begin the process. As the regime continues to interact, reformers can become
hard-liners if they feel that liberalization has gone far enough, while others may
press for additional reforms.

Hard-liners are regime members who are not interested in liberalization.
As with reformers, there can be varying degrees of hard-liners, ranging from
those satisﬁed with the status quo to others interested in increasing their control
over the country. Hard-liners do not have to share the same goals. If the
reformers attempt to effect change, hard-liners will try to stop that change. Their
success or failure will depend on what ofﬁces they hold in the regime. For
example, military commanders will have a better chance of stopping change than
the minister of agriculture. A second factor is how much inﬂuence they have in
the regime. Hard-liners can become reformers. If the reformers are able to
liberalize and the hard-liners do not see a signiﬁcant decrease in power, then
hard-liners may experience an increase in legitimacy, some hard-liners may
choose to become reformers because they no longer feel threatened by
liberalization.

In the case of de-autocratization, the internal struggle occurs between
hard-liners and reformers, with the point of contention being relaxing the regimes’
control in some political, social, or economic area. The reformers suggest some

degree of liberalization, forcing the hard-liners to decide to accept or reject the

 

of the game. Crescenzi does include different regime types in his model but he does not discuss
how it is the reformers or hard-liners may come into power.

24

reformers’ proposal. How this struggle occurs and the result of the struggle play
a signiﬁcant role in explaining the ﬁnal outcome of the authoritarian transition. If
the reformers gain control by a slim margin, then de—autocratization may begin.
However, if this is the case, de—autocratization faces a greater probability of
stalling, reversion, or civil war than if the reformers are able to gain outright
control of the regime. The hard-liners may attempt a coup to remove the
reformers if they have gained control over the regime, as has been seen in
Russia and Nigeria. In Nigeria, the military coups replaced democratically
elected governments while in Russia the coup attempted to remove Gorbachev
from ofﬁce before holding a referendum. The inability of the regime members to
agree on a course of action is directly related to the use of military force against
the regime.

Not only are the interactions among the regime members crucial in the
initial stages of authoritarian transition, they can reappear at any point and can
inﬂuence the process. If the hard-liners are willing to allow some degree of
liberalization but are not willing to abide by a full transition to democracy, the
likelihood that another internal struggle may occur will increase, if the hard-liners
feel that the process has gotten out of hand. Additionally, reformers may push
for increased control of the regime if they feel that they will be able to gain full
control of the regime and force the transition to go further than the reformers
originally believed they could. Therefore, it is important that we develop a better
understanding of this struggle within the regime and how it affects the overall de-

autocratization process.

25

What do we gain from exploring the strategic interaction within the
regime? All the work on democratization and regime transition acknowledges the
importance of the regime. The regime is the ruling body of a given country; it
makes policy decisions and implements those decisions. It is an important actor
that must be understood if we are to understand the de-autocratization process.
By focusing on the interaction between regime members, we are able to gain
additional insight into how and why the regime makes the choices it does, and
how those choices directly inﬂuence the outcome of any attempt at de-
autocratization.

Interaction within the regime will determine what type of regime civil
society faces. Additionally, how the regime settles its differences on the issue of
liberalization will play an important role in the outcome. It is possible that the
reformers and hard-liners clash, creating a coup attempt that ends with the
removal of the reformers or the hard-liners. So understanding how regime
members interact becomes increasingly important in understanding how a
country reaches the point it does in the de-autocratization process.

Przeworski (1991), Crescenzi (1998), and Agress and Gates (1999) all make
references to the importance of regime type in their models.

Crescenzi and Agress and Gates allow nature to choose what type of
regime civil society faces. They do not take the next step: an attempt to model
the interaction between the regime members. While the use of a nature node to
determine if the regime is composed of hard-liners or reformers is a reasonable

proxy, it does not explore the actual interaction between the regime members.

26

By explicitly modeling the interaction between the regime members we are able
to develop a better understanding of how this interaction will inﬂuence the
outcome or de-autocratization. If reformers are able to take control of the regime
by only a slim majority, then they are less likely to pursue dramatic reforms that
may lead to democratization, than if they had greater control of the regime.

To explore the role of strategic interaction between regime members,
Chapter Three will present a non-cooperative bargaining model. This model will
demonstrate the variety of options available to hard-liners and reformers as well
as the importance of information and the role of strategic decision-making

throughout the process.

Strategic Interaction Between the Regime and Civil Society

The second series of interactions to be explored is that of the regime and
civil society. As was seen in the literature review on democratization, civil society
is allowed to move only after the regime has chosen to liberalize. This is
problematic for several reasons. First, it assumes that civil society cannot place
pressure on the regime to liberalize. Second, assuming that civil society cannot
mobilize until after the regime has chosen to liberalize assumes that civil society
has no political power and is at the mercy of the regime. Poland,
Czechoslovakia, and South Africa are three cases that clearly demonstrated that
civil society can mobilize prior to the regime choosing to liberalize. In all three
cases, civil society showed considerable political power, even after the leaders of

the various movements were jailed, and in the case of South Africa, killed.

27

This is why, it is important to understand how civil society and the regime
interact and the nature of that interaction. While it is clear that the regime holds a
great deal of power, this does not preclude the possibility that civil society will
challenge the regime. When will civil society choose to take this action? How will
the regime respond to civil society’s challenge?

Chapter Four seeks to address these questions. In order to do so, a
game of incomplete and imperfect information is developed. This model will
allow civil society to respond to the regime’s decision to open or remain closed,
which means that civil society can choose to mobilize no matter what the regime
chooses to do. This differs greatly from previous models of democratization
which allow civil society to move if the regime chooses to open. The model will
allow the reader to develop a better understanding of the options available to
both actors as well as demonstrate the importance of information during the
process of de-autocratization.

Chapters Five and Six will use empirical methods in order to develop a
better understanding of how speciﬁc variables will inﬂuence the outcome of the
de-autocratization process. Although the variables that will be empirically tested
are discussed in the theoretical chapters, they are not explicitly included in the
game theoretic models; as such, the empirical tests are not a direct test of the
game theoretic models. The variables under consideration: economic
development, the degree to which civil society is fragmented, regime stability,
and the degree of regime change attempted in the previous year, are used to

explain how the regime or civil society may respond to the others move.

28

Chapter Four will explain how the variables are operationalized, the data
sets employed in order to run the empirical tests, and a discussion of the overall
data set. Chapter Five will contain a discussion of the hypotheses, the multi-

nomial logit used to test the hypotheses, and the results.

29

Chapter 3

Strategic Interaction Among Regime Members

This research seeks to develop a better understanding of the mechanisms
driving the process of de-autocratization. I argue that more focus must be placed
on the interactions occurring omong the various actors involved in the process.
This chapter asks: When will the regime decide to open and how is that decision
made? To accomplish this, the chapter is broken into the following sections; a
discussion of what is meant by “regime”, a discussion of the economic and
political dimensions that regime members will bargain over, the development of
the model, and a discussion of implications drawn from the model.

Why does this chapter focus on the regime? The regime is composed of
many actors, from the designated leader to the various heads of governmental
branches. While some regime members will have more inﬂuence than others, all
are involved in the decision making process that affect the day-to-day activities of
the country and its population. As such, it is vitally important that we understand
how the members of the regime interact with one another and how these
interactions inﬂuence the process of de-autocratization.

Other scholars have discussed the importance of the regime, but few have
attempted to delve into the issue at this level. In most cases, scholars focus on
de-autocratization after the regime has made the decision to open. Przeworski
(1991) develops a model that focuses on what might happen if the regime

chooses to open. Huntington (1991) argues that the amount of control that the

30

regime has over transition will inﬂuence the outcome, but fails to explore the
interaction among the regime members and how these interactions will inﬂuence
the regime’s ability to control the process. O’Donnell and Schmitter (1986)
discuss the regime and why it enters into pacts with some segments of civil
society. They do not explore how the regime ﬁnds itself in such a position or how
the regime reaches that decision. In the previous work, the regime plays a large
role in the de-autocratization process. Yet how the regime members interact
internally and what impact that has on the regime’s ability to interact with civil
society is not discussed.

Liberalization can only occur when regime members choose to open or
agree to open following the mobilization of civil society‘. In order to understand
how the regime reaches this decision, it is necessary to understand how the
various factions in the regime interact with each other. The more divergent the
goals of the hard-liners and reformers, the greater the chance of conﬂict within
the regime. The different camps will be focused on attempting to advance their
conﬂicting agendas, which can destabilize the regime. How the interaction
among these camps occurs will be explored later in this chapter.

By developing an understanding of how and why the regime chooses to
begin liberalization, we can better understand the ﬁnal outcome of the transition
attempt as we approach the question with more information. The goal of this
chapter is to explore how the regime reaches the decision to liberalize and to
develop a better understanding of how this interaction can lead to a coup as well

as liberalization.

31

What is the Regime?

There are many actors involved in the process of de-autocratization. The
ideal research project would develop a model that takes each individual into
consideration, thereby providing accurate predictions. In reality, such a model
would be cumbersome and unparsimonious. As we shall see in the discussion
below, the regime is an aggregation of a larger number of individuals. If we are
to understand how it is that strategic interaction inﬂuences the outcome of a
country involved in de-autocratization, then we have to develop a better
understanding of who the actors are and why they behave in the manner they do.

By focusing on the interactions between regime members, we are forced
to begin to unpack the various actors and examine their choices more carefully.
In examining the interactions between regime members, we develop additional
insight into why the regime chooses to liberalize as well as how the forces that
drove that decision will inﬂuence the process’s ﬁnal outcome, whether that is
continued liberalization or a civil war or a coup. In order to understand how the
actors interact, it is ﬁrst necessary to deﬁne what regime means.

The regime is composed of a multitude of individuals, and the number will
vary from country to country. There is a designated leader, such as Castro in
Cuba or Gorbachev in the former Soviet Union, who is supported by the heads of

government institutions, such as the military commanders, the head of police, the

 

6 Chapter 4 will contain a discussion of Civil Societies role in the process of de-autocratization.

32

head of intelligence, agriculture, and ﬁnance. Each branch of government will
have varying degrees of inﬂuence over decisions reached by the regime. The
regime members will have their own goals and ideas about where the country is
heading and where they would like it to go. While there will be similarities in how
regime members view government, there can also be a great deal of variance. It
is this variance that can lead to a fragmented regime.

O’Donnell and Schmitter argue that “...there is no transition whose
beginning is not the consequence - direct or indirect — of important divisions
within the authoritarian regime itself, principally along the ﬂuctuating cleavages
between hard-liners and soft-liners”. (pg. 19, 1986) These divergent positions
cannot only lead to a peaceful transition from one regime type to another, but
also can lead to a coup attempt.

A recent example of conﬂicting ideologies can be found in Iran today. The
Council of Clerics remains a conservative body opposed to increasing political
liberties while the president of Iran, Khatami-Ardakani, is determined to increase
the political liberties available to the population. One step taken in this direction
was to allow a freer press. In 2000, newspapers and magazines were allowed to
criticize the government. The Council of Clerics felt that the papers had been
given too much freedom and insisted on a reversion to past policy. In this case,
President Khatami-Ardakani backed down, ordering the closure of newspapers
deemed too vocal and the arrest of the editors. Once President Khatami-
Ardakani was re-elected in 2001, he began once again to allow newspapers to

criticize the government. Clearly, the Council of Clerics and the president of Iran

33

are in disagreement over what type of political liberties should be granted to the
population. If the gap between the two factions grows, so will the possibility of
conﬂict.

As the distance between the two factions increases, the threat of a coup
can increase. How much power will the hard-liners be willing to relinquish? How
will the reformers know when they can no longer push for liberalization before the
hard-liners react with force? O’Donnell and Schmitter point to these very
questions.

“Once liberalization has been chosen - for whatever reason and under

whatever degree of control by incumbents - one factor emerges which

hangs like a sword of Damocles over the possible outcome. This is the

fear of a coup that would not only cut short the transition but impose a

regression to an even more restrictive and repressive mode of

governance.” (pg. 23)

To address this question, the research presented here will focus on
regime factions that are based on attitudes toward de-autocratization. Hard-
liners are regime members who are not interested in liberalization. They are
generally content with the status quo and in some cases are looking for
opportunities to increase their power, in effect decreasing the liberties available
to the population. Reformers are regime members who are interested in some
degree of liberalization. This does not mean that reformers are interested in

relinquishing all control, only that reformers are interested in allowing some

increase in political, social, or economic liberties.

34

The Political and Economic Dimensions

Why Open?

The authoritarian regime holds a great deal of economic and political
power, which the individual members can use for their own gain. The process of
de-autocratization requires that regime members relinquish some degree of
political control and the economic beneﬁts that come from holding ofﬁce. This
forces anyone attempting to understand the process of de-autocratization to ask
the question: Why reform? The model presented in this chapter will focus on the
economic and political reasons that drive de-autocratization.

Casper and Taylor (1996), Collier and Collier (1991), and O’Donnell and
Schmitter (1986) agree that de-autocratization begins when the regime comes to
a critical juncture. This critical juncture is a period of time where the authoritarian
regime is in some way vulnerable and liberalization may be possible. These
crises can take many faces: political protest, mass riots, and resignation of
government leaders, as well as international conﬂict. When the regime is faced
by such a crisis, the outcome is uncertain, allowing Reformers an opportunity to
suggest liberalization.

O’Donnell and Schmitter go as far as to argue that at the heart of the
political crisis are divisions between regime members and that the crisis itself

exasperates these divisions. (1986, 19) The political crisis at hand gives the

35

reformers an opening in that it exposes the inherent legitimacy problem within an

authoritarian regime.

“Opponents were stimulated to act because the failure was so obvious.
Ruling groups, including the armed forces, were less and less conﬁdent of
their own capacities, as well as deeply fragmented by recriminations over
who was responsible for the regime’s failures. Mediators were no longer
willing to arbitrate dissent and hold coalitions together. Faced with this, the
authoritarian rulers sought a rapid ‘political outlet’. This gave ample room
to the soft-liners, for whom it seemed less risky to launch the country into
liberalization, and even democratization, than to continue struggling
inﬂexibly and ineffectively against a rising tide of opposition, fed by
defection from the regime’s ranks.” (O’Donnell and Schmitter, 1986, 20)
Casper and Taylor argue that the critical juncture presents an opportunity
that reformers can choose to take advantage of. Huntington would agree, arguing
that one of the causal factors behind the third wave of democratization was the
declining legitimacy of the authoritarian regimes. (1991) According to Huntington,
the depth of the problem comes from the democratic experiences of many
authoritarian regimes that had been overthrown by authoritarian regimes during a
prior crisis. The authoritarian regime is able to temporarily solve the crisis, or give
the appearance of solving the crisis. Over a period of time, the crisis can re-
emerge or a new crisis arises that the authoritarian regime is unable to solve.

When this occurs, the legitimacy of the authoritarian regime is called into

question.
Political Dimension

When this critical juncture is reached, the reformers can choose to

propose a program of liberalization, or they can let the moment pass. If the

36

reformers choose to make an offer, it may include some type of political changes.
The degree of change offered by the reformers will be based on their
commitment to developing a new political regime, possibly a democracy, and
what the reformers feel they can receive without losing their place in the regime.
This can be thought of as a one-dimensional realm with a pure authoritarian
regime set at a score of zero and a pure democracy set as a score of 1. The
various points in between represent the different types of regimes that fall
between an authoritarian regime and a democratic regime. The reformers offer
will be something greater than the status quo position, but should not be so great
as to threaten the hard-liners enough that they fear for their position and are
willing to attempt a military coup to remove the reformers as regime members.

A wide variety of offers can be made along the political dimensions. Such
offers may include some of the following political reforms. Reformers can
demand the freedom of speech, the right of assembly, freedom of the press,
language rights for a speciﬁc group in a country, the right to a public and fair trial,

and suffrage for a speciﬁc group of people.

Economic Dimension

O’Donnell and Schmitter emphasize political unrest as the underlying
factor that causes de-autocratization. While political issues are at the root of
many attempts to de-autocratize, they are by no means the only cause of regime
transition. In addition to the political dimension, an equally important economic

dimension must be taken into consideration.

37

Huntington (1991), Lipset (1959), Dahl (1971), Bollen and Jackman
(1985), Morawetz (1977), and many others point to the importance of economic
development in understanding the development of democratic regimes. In
studying the countries involved in the Third Wave of democratization, Huntington
found that the urban middle class led many of the countries involved in
democratization. (pg.67, 1991) Huntington also points to the role of economic
crises. Huntington argues that rapid economic growth and economic crises can
undermine the authoritarian regime, in essence bringing about the critical
juncture that allows reformers the opportunity to begin the process of de-
autocratization. (pg 72, 1991)

Beyond the importance of economic development and economic crises in
setting the stage on which de-autocratization will occur, regime members have
an economic stake in maintaining control over the country. In many cases the
regime members are receiving economic beneﬁts from holding ofﬁce. These
beneﬁts can take the form of pay, bribes from industry, and the like. By entering
into de-autocratization, the regime members are relinquishing some of the
economic beneﬁts they have been receiving. As such, part of the negotiations
among regime members will include a discussion of economic changes allowed
in the country. As with the political dimension, this economic aspect of de-
autocratization must be taken into consideration if we are to understand how the
regime members interact with one another.

Like the political dimension, the economic dimension can be seen as a

one-dimensional line. In the case of the economic dimension, zero represents a

38

system where the state controls the economy, and one represents a free market.

The points in between represent a wide variety of hybrid systems.

The Game Theoretic Model

The modeling technique used in this chapter is a bargaining game with
outside options. The game pits two opposing factions, both factions coming from
within the regime, against one another. There are two dimensions that the
factions are bargaining over— the economic and political. The faction to move
ﬁrst will make an offer that the second faction can choose to accept, make a
counter offer, or attempt a coup. If the second faction chooses to make a
counter offer, the ﬁrst faction chooses between accepting the offer, rejecting the
offer, or staging a coup.

This model differs from a standard bargaining model because it adds an
additional element to the game that more accurately reﬂects the behavior of
regime members. The outside option gives both factions the opportunity to leave
the table and pursue a course of action that is not included in the set of outcomes
discussed at the bargaining table. In the standard bargaining model, the Hard-
liners and Reformers would have the option of making an offer, making a counter
offer, and then bargaining over how much reform, if any, to allow. The possibility
of a coup is not discussed at the bargaining table even though it is a choice
available to both parties. This choice is available only if one faction chooses to
leave the bargaining table. This will occur when one faction does not believe that

its needs will be met through the bargaining process, and that it will win if it

39

attempts a coup. In essence, the presence of the outside option alters the
bargaining space away from the good faith standard bargaining model.

This technique is an improvement over the standard bargaining model
because it allows both factions the opportunity of choosing an outside option,
military coup. The presence of the outside option of a coup changes the
bargaining structure such that each faction must consider the possibility of a
coup before making an offer. If a faction believes that it can gain more through a
coup than bargaining, the outside option becomes a credible threat. This will
inﬂuence the game’s outcome in the following manner. As one faction’s belief
that the opposing faction is likely to resort to a military coup, it will become more
willing to enter into negotiations and change its offer, in order to prevent a coup.

This is an improvement over previous modeling attempts by Przeworski
(1991) and Casper and Taylor (1996). First, the focus on the interaction between
regime members helps us to better understand how de-autocratization can lead
to violent outcomes. The outcome of the interaction between regime members
can derail the process of liberalization if the Hard-liners are able to remove the
Reformers with a coup. It is also the case that liberalization will occur more
quickly if the Reformers are able to remove the Hard-liners in a coup.
Przeworski's model does not take into consideration how the decision to
liberalize occurs, but instead assumes that liberalization has been decided and
proceeds from there. In doing so, Przeworski is unable to understand how de-

autocratization can be derailed or sped up prior to the regime opening. The

40

model developed in this chapter will demonstrate the importance of
understanding how regime members decide to liberalize.

Second, the inclusion of the outside option, coup, is an improvement
over the work of Casper and Taylor (1991). Their model begins with the
Reformers and Hard-liners negotiating over transition. While Casper and Taylor
discuss the possibility of a civil war, they make no attempt to model such an
option, instead arguing that such an action would redeﬁne the game. The model
developed in this chapter includes the possibility of a coup as it is a viable
outcome that has been seen in many countries attempting de-autocratization, for
example the Soviet Union in 1991 and Pakistan in 2000. If we are to understand
how the process of de-autocratization works, then it is necessary to develop
models that accurately reﬂect the variety of possible outcomes.

The model below is meant to focus only on the interactions between
regime members. Later chapters will explore how the regime interacts with civil
society. This is done in order to develop a better understanding of how the
decision made by the regime members can inﬂuence the overall process of de-

autocratization.

Discussion of the Outside Option.
The three outcomes that this research will focus on are the acceptance of
an offer, maintenance of the status quo, and a coup attempt. Neither Przeworski

nor Casper and Taylor allow for the possibility of a coup in their discussions of

41

democratization. Przeworski assumes that democratization starts with the
liberalizers controlling the regime and deciding to offer to reform or not.

Casper and Taylor do not assume that this is the case. The ﬁrst two
stages of their game focus on the regime's decision to liberalize. In the ﬁrst
stage, the critical juncture stage, there is an event that offers the opportunity to
liberalize. At this stage, the challenger must make the decision to challenge the
current regime structure. The second stage consists of three steps, known as
the sorting out stage. First, the challenger must identify him. Then the defenders
and the challengers make some initial proposal about the type of regime want
installed. Finally, the mass public is allowed to respond to the proposals. The
sorting out stage is followed by the deal cutting stage, where a ﬁnal outcome is
reached. By combining all three stages, Casper and Taylor's game
encompasses the entire process of democratization. In doing so, they negate the
importance of the bargaining process within the regime. Additionally, coups are
not included as a possible outcome of the game.

Casper and Taylor's model assumes implicitly that the challengers and the
defenders are entering into good faith negotiations. As such, the threat of a coup
is not possible. Yet in reality, the threat of a coup is something that every
authoritarian regime faces. The dilemma that scholars are faced with is how to
include the possibility of a coup into a model of de-autocratization. One answer
to this question is to view coups as an outside option.

An outside option requires that players involved in bargaining are willing to

walk away from the bargaining table so they can take advantage of an option that

42

is not found within the bargaining set. In the case of the model presented in this
chapter, players are negotiating over the degree of political and/or economic
reform. Each player can make an offer that can be accepted or rejected. Yet
there is an option that is found outside of the normal bargaining set, an attempt to
remove one part of the regime through a coup. According to Binmore, Shaked,
and Sutton (1985, 1989) and Scaramozzino (1991), the threat to resort to an
outside option would only be implemented if it is deemed a credible threat. A
threat is credible only when one player ﬁnds it is more proﬁtable to accept the
payoff from the outside option than to continue bargaining.

In the case of negotiations between hard-liners and reformers, both
players have the option of making an offer, accepting or rejecting the other
party’s offer and making a counter offer, as is presented in Casper and Taylor's
game. While these options are reasonable, recent events in Pakistan in 2000
demonstrate that there is another option that is available to members of the
regime, that of a coup.

If we are to truly understand how the members of the regime interact with
each other and how that can inﬂuence the outcome of de-autocratization, then
there must be the possibility of an outside option of a coup. Regime members
must know that they can leave the table and attempt a coup if they believe that
their side will emerge victorious. Thus, the outside option of a coup adds another
dimension to the model presented below that is not available in the models

presented by Przeworski or Casper and Taylor.

43

The Factions

Vlﬁthin an authoritarian regime, there are competing factions. I will refer to
these factions as hard-liners, H, and reformers, R. Przeworski and Casper and
Taylor assume that the reformers will always make the initial move in their
transition games. While this research is focused on the process of de-
autocratization, the game theoretic model in this chapter allows either faction to
move ﬁrst. This is an important departure from the previous models developed to
understand de-autocratization because this model can help in explaining why
countries that have begun the process ﬁnd their progress reversed and even in
understanding why some regimes become more autocratic. If, during the
process of de-autocratization, the hard-liners become uncomfortable with the rate
of liberalization or the direction that liberalization is taking, this model allows the
hard-liners to make a proposal to the reformers. In the Przeworski and Casper
and Taylor models, the hard-liners are forced to react to the reformers and are

not allowed to take the initiative.

A Spatial Presentation of the Policy Space
The factions are bargaining over desired policy positions in the political
and/or economic arena. The Status Quo is indicated as position q in the ﬁgure

below. Both factions have ideal points, represented as F; and Fj. If either

44

Offer

 

 

Fl
Accept Coup Attempt
Reject
()2 Coup
SQ

Figure 3.1 Non-Cooperative Bargaining Game

45

Ol=(p19 el)

02 = (P2 9 e2)

faction was to ﬁnd itself at its ideal point, there would be no need for that faction
to enter into bargaining, or continue bargaining, as there would be no way to
improve on its position.

In the example, Factionj would be comprised of Reformers as their ideal
point indicates an improvement on both the political and economic dimensions.
Faction i's ideal point indicates that the hard-liners would prefer to see the regime
increase its control over the country, as it is located below the status quo point.
The line connecting the ideal points of both factions is the contract curve. The
circles represent the limits within which each faction can bargain; any offer made
in this space can be accepted. The area where the two circles intersect is the
I bargaining space within which negotiation will occur. The ideal points for both
factions fall outside of the bargaining space. The dark line in the bargaining
space represents the possible location of the new status quo point.

The expected utility that either faction would receive from a policy position,
x = (px. ex), is represented in the utility function found below. X is the offer made
and i is the political and economic location of the faction.

Ui (x = px. ex) = - [(px - pa)2 + (ex - e: )2]

The utility function is a measure of distance between the policy point (x) on the
economic or political dimension and the ideal point for the faction in question.
The offers made will not be found at the factions ideal point, this is indicated in
the utility function by the negative sign. If the offer made is at the factions ideal
point, the value will be —0, indicating that there is no movement from the ideal

point.

46

 

Economic Dimension

 

 

 

- l
Pure Autocracy 0“ Pure Democracy
\

 

Political Dimension

Figure 3.2 A Spatial Representation of the Bargaining Area

Outcomes and Payoffs
There are three possible outcomes, a faction can accept the opposing

faction's offer, the status quo, and coup. There are no costs associated with the
bargaining process, as such, the payoff for moving the status quo points is found
using the utility function above. In the case of the status quo, the payoff is 0, as
there has been no change.

This is not the case when the outcome is a coup. The faction that is able
to defeat its opponent is able to move the status quo point to its ideal point. The

greater the cost of ﬁghting a coup, the smaller the payoff.

47

The payoff for a coup requires that the factions involved pay the cost of
ﬁghting, . This cost includes the loss of life, the cost of purchasing guns, and
the cost of paying people to ﬁght for your faction. Additionally, there is a
probability assigned to the coup outcome. P is the probability that F, wins; (1-P) is
the probability that F, wins. The faction that wins the coup is able to set policy at
its ideal point. Hence, the utilities for winning a coup are F; — i, for Fi, and F; - j,
for F,-. The greater the probability that the faction believes it has to win a coup,
the more credible the outside option of coup becomes. This will directly effect
what type of offer each faction will make and the likelihood that either faction will
exercise its outside option if it does not approve of the offer made by the
opposition. An offer outside of the bargaining set will be made only if the faction

making the offer believes that it can win if there is a coup.

Order of Play

The ﬁrst faction (Fi) opens the game by making offer, (91. e1) indicating its
offer on the political and economic dimensions. The second faction (F;) can
choose to accept the offer, make a counter offer, or act on its outside option and
attempt a coup. If Fj accepts the offer made by F: , then the game ends with the
outcome of accept and the payoffs associated with of p1, e1 for each faction If F;
chooses to exercise its outside option and attempt a coup, the payoff for the
winning faction is the ideal point minus the cost of ﬁghting. If F] chooses to make

a counter offer, (p2, e2), F. can choose to reject the offer, accept the offer, or

48

attempt a military coup of its own. If F; rejects the counter offer, then the game

ends with the status quo as the outcome.

Implications of the Model

As was discussed earlier, this model differs greatly from other models of
the process of de—autocratization. First, the model developed in this chapter
allows either faction to move ﬁrst. Secondly, this model allows for the use of
military coups.

These features offer us the ability to understand not only when de-
autocratization occurs but also the ability to understand why a country might
become more authoritarian while more accurately reﬂecting all possible
outcomes. Take for example a country that has begun the process of de-
autocratization and where the hard-liners are concerned that liberalization is
occurring too quickly or advancing beyond what they are comfortable with. In
this case, the hard-liners can make an initial offer that would move the status quo
point closer to an autocracy or controlled economy, effectively reversing what
liberalization had occurred. The reformers are left to ask themselves if the hard-
liners have the ability to win a coup, and if so, the reformer are more likely to
agree to slow down or reverse the process of liberalization. Why?

If the hard-liners threat of a coup is credible, then the reformers will be
accept the offer that the hard-liners put forth in order to maintain some presence
in the regime. This is done with the hope that in the future, the reformers will be

able to gain enough power that they can either withstand a coup or launch their

49

own coup and continue de—autocratization. This is not possible if the reformers
are forced out of the regime through a coup. If the hard-liners threat of a coup is
credible, then why negotiate? As the cost of ﬁghting increases, the payoff for the
coup option decreases, making negotiation a more attractive option. If the hard-
Iiners are able to force a halt to de-autocratization, or even reverse the process,
without ﬁghting, the hard-liners can receive a greater pay off. This means that a
coup will only be used as the last option by either the reformers or the hard-
liners.

This seems to accurately portray the process of de—autocratization as it
occurs in the real world. Of the one hundred and seventy-one cases of conﬂict in
the Wallensteen and Sollenburg data set, there were only eight cases were the
outcome was a coup7. While the Wallensteen and Sollenburg data set is not
dedicated to identifying coups, it does include all cases of conﬂict where there
were at least twenty-ﬁve battle dead. As such, countries were there was even a
mildly violent coup would be included in the data set.

The Wallensteen and Sollenburg data set seems to indicate what the
game theory model implies, that coups are a rarity. The model developed in this
chapter demonstrates why coups are so rare. Coups should only occur when
one faction in a regime believes that it will win and that the costs associated with
a coup are small enough that negotiations are no longer feasible. At this time,

the faction that believes that it can gain more through a coup will choose to coup.

 

7 I argue that political unrest and civil war require the participation of civil society. In order to
understand how these outcomes are reached, I develop a game theoretic model in chapter three
that explores the interaction between a regime and civil society. Chapter four explains how
measures of violent outcomes are developed.

50

While it is important to understand why a faction within the regime may
choose to coup, there are other forms of lntemal violence that can occur during
the process of de-autocratization. This chapter focused on the possibility of
lntemal violence based on the non-cooperative bargaining that occurs between
regime members. As such, the violent outcome that is most likely is a coup. In
order to understand all of the possible violent outcomes it is necessary to
examine how the regime interacts with civil society, this is the topic of discussion

in the next chapter.

51

Chapter 4

Strategic Interaction Between the Authoritarian Regime and Civil Society

The twentieth century has seen a large number of countries attempt some
form of authoritarian transition, some succeed while many have failed. Why? It
is the argument of this research that interaction between actors has a great deal
of inﬂuence on the process of de—autocratization. In chapter three we saw the
bargaining process used by the regime to determine if reform should be
attempted, what degree of reform should be allowed, and how the bargaining
process can lead to coups.

If we are to understand the interaction between the regime and civil
society, it is necessary that we understand how the regime members interact
amongst themselves. The decisions made by the regime in chapter three
determine what type of regime civil society is facing, what types of reform the
regime is comfortable allowing, and the strength of the different factions within
the regime. As such, the model developed in chapter three ties into the model
developed in this chapter.

It is the interaction between the regime and civil society that will guide a
country to one of many outcomes. Democratization represents perhaps the ideal
outcome of any transition process, as such it has received a great deal of
attention in recent years, unfortunately, it is but one possible outcome. On the
other end of the spectrum is civil war. Civil war brings with it the destruction of

lives, property, roads, and, in extreme cases, the loss of a unique group of

52

people. Knowing that the cost of civil war is so great, why do we see this
outcome occur throughout the nineteenth and twentieth century? This question
can be answered by focusing on how the interaction between the regime and civil
society occurs and the ramiﬁcation of the interaction within the regime and civil
society prior to entering into negotiation with one another.

I choose to use a game theoretic model in order to arrive at a better
understanding of this process due to the nature of the process being both
dynamic and strategic. The actors, the regime and civil society, have
preferences associated with the possible outcome of authoritarian transition and
will choose actions that they believe will allow them to achieve their goals.
Additionally, game theory allows us to explore the different strategies available to
the actors and the possible outcomes of those actions. The use of an
asymmetric incomplete information game allows us to model the process while

taking into consideration the actors lack of information about one another.

The Role of Civil Society
Civil society's ability to address the collective action problem will determine its

strength. Why is this important? The answer to this question is relatively simple
yet its impact on the process of authoritarian transition can be rather dramatic.
The stronger civil society is the more inﬂuence it will have when negotiating with
the authoritarian regime. This can be demonstrated by brieﬂy comparing the
experiences of Czechoslovakia and Nigeria. Both countries have a history of
authoritarian rule, Czechoslovakia was annexed and occupied by Germany

during WWII and then occupied by the Soviet Union until 1989 while Nigeria was

53

a British colony which has seen three separate attempts at developing a
democratic regime overthrown through military coups. Both countries have made
attempts to remove the authoritarian regime on several occasions with varying
degrees of success. Why the difference? I argue that civil societies role in the

process of transition is one part of the answer.

Civil Society as Depicted in the Literature
The idea of civil society playing a more active role in authoritarian transition is not
a new one. The work of scholars such as Casper (1995), Cohen (1994), Hegre et
al (1998), Huntington (1991), O’Donnell and Schmitter (1986), McAdam et al
(1996), and Sutter (1995) have included civil society in their explanations. In
many of these works, civil society is treated as a passive actor, able to react to
the regimes decisions but not able to initiate the process of regime transition.8

In O’Donnell and Schmitter’s 1986 work, Transitions from Authoritan'an
Rule: Tentative Conclusions About Uncertain Democracies, it is argued that civil
society can play an active role during the authoritarian transition through a
process called pacting. Pacting is deﬁned as:

“An explicit, but not always publicly explicated orjustiﬁed,
agreement among a select group of actors which seek to deﬁne (or better,
to redeﬁne) rules governing the exercise of powers on the basis of mutual
guarantees for the ‘vital interests’ of those entering into it” (1986, 37)

Pacting is used by the regime to control the degree of liberalization and occurs

between a small group of people. Pacts can be based on military, political, and

 

a Huntington and Przeworski acknowledge that civil society may initiate the attempt at transition.
Przeworski argues that it is a rare enough event that it is reasonable to leave such cases out of
his analysis of de-autocratization. Huntington does analyze cases where civil society is able to

54

economic issues.9 The regime signals that it is willing to allow new freedoms,
civil society begins to re-emerge, using the new liberties that the regime has
granted it. O’Donnell and Schmitter's do not allow civil society to act until the
regime chooses to open and the initial involvement is with a few segments of civil
society, which the regime chooses. While civil societies role can increase,
through what O’Donnell and Schmitter term a popular upsurge, civil society is still
initially forced to react to the regime.

Przeworski, Crescenzi, Casper and Taylor, and Gates and Humes all use
formal models to explain the initial stages of democratization and all include civil
society but limit civil societies ability to act by allowing it to move only if the
regime chooses to open. Agress and Gates make an attempt to address this
issue by developing an asymmetric, two-sided incomplete information game that
allows civil society to move even if the regime opts to remain closed. While these
models allow for different regime types, none allow for different types of civil
society. Przeworski acknowledges that there are different types of civil society,
which he labels moderates and radicals, in his theoretical discussion and that the
interaction between the different segments of civil society will inﬂuence how the
regime responds to civil societies demands but he does not include this in his
formal model.

Przeworski (1991) is the ﬁrst to develop a formal model of

democratization. In his model, civil society is relegated to moving only after the

 

initiate the process and argues that the majority of these cases lead to civil war or some form of
lntemal violence making this path a less acceptable path of transition.

9 This is analogous to Przeworski's (1991) description of the incorporation of elements of civil
society into a broad dictatorship.

55

regime has decided to open. This is intentional as Przeworski argues that civil
society may choose to mobilize even if the regime does not choose to open but
this rarely leads to the establishment of a democracy. While this may be the
case, Huntington also argues that replacements are rarely successful for a
number of reasons, it indicates one weakness in how the process of de-
autocratization has been studied. The emphasis on democratization has lead
scholars to approach the question in such a manner that making such simplifying
assumptions as restricting civil society’s opportunity to move has not been
challenged. Civil society does not have to force democratization in order to play
a role in the process. Civil society can place pressure on the regime through
demonstartaions, everyday forms of resistance, and armed conﬂict with the
regime which eventually may force the regime to consider some degree of
liberalization. How civil society inﬂuences the process of de-autocratization will
only be known once we allow civil society to participate equally in the process. It
is for this reason that the model presented in this chapter allows civil society to
move even if the regime does not choose to open.

While these works include civil society as an important actor, only a few
acknowledge that civil society can initiate the transition attempt and those that do
fall to acknowledge that civil society is not a uniﬁed actor, and vice versa. If we
are to develop a real understanding of how civil society interacts with the regime,
it is important that the model reﬂect civil society's ability to initiate transition and

the collective action problem civil society faces.

56

Civil Societies Type

As was demonstrated in chapter 3, civil society can not be treated as an
aggregated actor. In most countries civil society is composed of a multitude of
groups that may be centered on issues of ethnicity, religion, language, and
economic status. In treating civil society as a united actor, the above works
neglect the diversity within civil society and weaken the predictive power of their
models. The model presented in this chapter attempts to address this problem
by allowing for civil society to be weak or strong, dependent on how united civil
society is and allowing civil society to choose to mobilize even if the regime
chooses to remain closed. Why is this important?

Ultimately, any regime's longevity is based on civil society's compliance
with the rules by which the regime runs the country. Compliance does not require
or even imply acceptance of the rules, only that the rules are followed. As such,
compliance can be achieved through the use of force or the threat of force. In
countries were civil society has an opportunity to voice their opinion with the
belief that their opinion truly matters, compliance is easier to achieve and
maintain.

In the case of authoritarian transitions, civil society is not likely to have a
recognized voice or that voice will be limited and controlled by the regime. If civil
society is to inﬂuence the process, it must ﬁnd some way to be heard. The
louder the voice, the greater its possible impacts on the process of de-
autocratization. Achieving this voice is a difﬁcult task, as Collier and Hoefﬂer

demonstrate. The costs involved in rebelling are high, if the rebellion fails the

57

regime will punish the participants if the rebellion succeeds you have to trust that
the leaders of the rebellion will act in the best interests of civil society. (pg.8)
These two factors increase the incentive to free ride. The more free riders, the
more difﬁcult it is to rebel or mobilize any type of campaign directed at
challenging the regime.

Solving the collective action problem is essential if civil society is to play a
role in regime transition. The larger the number of people involved, the more
diverse the interests that are represented, the more seriously the regime have to
take civil society. This is not to say that if ninety percent of civil society comes
together to challenge the regime that their demands will be met. As we will see
when we explore the game theoretic model presenting the interaction between
civil society and the regime, even a tightly uniﬁed civil society can face the
possibility of civil war. What this does mean is that that the likelihood that their
voices are heard is greater and that their chance of affecting the process of de-
autocratization will be greater if civil society is able to unify behind some common

goal or cause.

The Role of the Regime

In any country, it is the responsibility of the regime to establish the rules and
norms by which the country is run. As such, the regime will play a large role in
any attempt at authoritarian transition even if it does not initiate the transition

attempt. The policies that the regime implements and how it reacts can alter the

58

manner in which transition occurs and the ﬁnal outcome of any attempt at de—
autocratization.

Once again turning to the example of Czechoslovakia in 1968 and in
1989, the type of regime plays an important role in the divergent outcomes. In
1968, the communist regime knew that it had the support of the Soviet Union.
When Dubcek took ofﬁce and began to make reforms, the hard-liners were
counting on the intervention of the Soviet Union to remove Dubcek and end his
experiment. In 1989, the Soviet Union has been weakened and is beginning to
make lntemal reforms of its own. The decline of the Soviet Union leaves the
communist regime in Czechoslovakia weakened which sucumbs to reform within
a year of the fall of the Berlin Wall. Solidarity’s victory in Poland signaled that the
Soviet Union would not interfere with the satellite states lntemal affairs, thus

weakening the position of the regime.

The Regime as Depicted in the Literature

The vast majority of the literature found in international relations and comparative
politics focuses on the regime and how it guides the process of de-
autocratization. Huntington (1991) describes three paths that democratization
can take and discusses what the regime’s role in the transition is. He ﬁnds that
democratic transitions which are driven by the regime, are more likely to be
successful and peaceful. Casper and Taylor also identify three paths that

democratization can take, although they differ from Huntington’s. Like

59

Huntington, an emphasis is on how the regime directs the process of
democratization.

Przeworski (1991) develops a game theoritic model which focuses on the
regime’s decision to open or not. The emphasis is placed on the regimes
decision, with civil society moving only after the regime has decided to liberalize.
He argues that there are two types of transitions, bottom up and top down, with
the topd down transition being more prevelant. For Przeworski, the question of
why the regime “cracked” is the most intersting question.

Like Przeworski, O’Donnell and Schmitter place a greater emphasis on the
role of the regime then they do on the role of civil society. The regime must send
a signal that they are willing to liberalize by actually allowing some degree of
opening. If civil society chooses to, they can attempt to negotiate with the regime.
Civil societ is included in the process but the emphasis remains on the role of the
reghne.

The model in this paper seeks to understand how the regime interacts with
civil society and how the decisions made by each actor inﬂuences the ﬁnal
outcome. In accordance with the views of Przeworski, O’Donnell and Schmitter,
and Huntington the regime is allowed to make the initial decision to open or not.
This is done because it is an accurate reﬂection of the process of de-
autocratization as the regime is responsible for the development of policy.

Taking no action, ie not opening, is also a decision that civil society can react to.
It is for this reason that I allow civil society to move even if the regime chooses to

not open.

60

The Regimes Type

In chapter three we saw that the regimes lntemal bargaining process will effect
the regimes types and impact the overall process of regime transition. Hard-liners
are less likely to open, and if they do the new liberties are likely to be very limited
and will not threaten the regimes control of the country. Reformers are more
likely to open and enter into some type of negotiation with civil society.

As with civil society, the type of regime matters. Is the regime composed
of hard-liners or reformers? If civil society had this information the interaction
between the two groups would be much lass complicated. Civil society will be
less likely to mobilize if the regime is composed of hard-liners then it is if the
regime is composed of reformers. Why? Reformers are more likely to support
some degree of increased liberties and are more likely to be willing to negotiate
on what it will give to civil society. Hard-liners are less likely to allow an increase
in civil liberties, if civil society is to mobilize against the hard-liners, civil society
will have to be uniﬁed if it is to stand a chance of overthrowing the regime.

The regime has the beneﬁt of controlling the military and police forces,
which allows it to dictate policy. Who controls these forces, how in favor they are
of reform, and how loyal the military and police forces are to the regime will play
an important role in determining the strength of the regime. If the regime cannot
count on the military or police forces to respond to commands to contain civil

society, than the regime is greatly weakened.

61

Interaction Between the Regime and Civil Society

Regime transition can occur with only one actor participating, but that is
rare. Whether the interaction between the regime and civil society occurs
through the use of force or repression against some segment of civil society or
through voting, the regime and civil society interact with one another on a day to
day basis. How this interaction occurs will dictate the ﬁnal outcome of any
attempt at de-autocratization.

While there are authors who have included civil society as an equal
partner with the regime in the process of democratization, most do not.
Przeworski (1991), O’Donnell and Schmitter (1986), and Crescenzi (1998) allow
civil society to move only if the regime chooses to reform. Bermeo focuses on
how the regime must experience some degree of political learning if
democratization is to succeed but does not include a discussion of how civil
society must also learn how to behave in a democratic society. Civil society is
relegated to the level of a second tier actor when in reality, its actions are equally
important as the decisions made by the regime.

Huntington (1991) uses a case study approach to explore the interaction
between the regime and civil society. Huntington demonstrates that the
multitude of ways that the interaction between the regime and civil society occurs
explains why there are different paths to democratization. If the regime is willing
to negotiate with some members of civil society, democratization can occur

peacefully. In cases where the regime is forcibly removed by some segment of

62

civil society, democratization tends to be less successful because of possible
conﬂicts within civil society and arguments over the development of new
governmental institutions and rules. While Huntington does not develop a formal
model of democratization or apply any econometric techniques to test his
hypotheses, he is able to demonstrate that democratization involves both the
regime and civil society.

As has been discussed earlier in this chapter, it is necessary to treat civil
society and the regime as equals in the proces of democratization. Even though
it may be rare, as Przeworski claims, that civil society rises in rebellion and
overthrows the regime it is still a possibility. More importnatly, if we are to
accuratly model de-autocratization it is essential that the model reﬂect the fact
that the process is an itterated one. It is rare that de-autocratization occurs over
night, especially if we are to focus on democracy as the ﬁnal outcome. Due to
the fact that de-autocratization is an itterated process, civil society must be
treated as an equal.

If we examine the case of South Africa we see what role civil society plays
in de-autocratization. Democratization came only after years of pressure had
been applied internally through consistent protests and challenges of the
governments authority by civil society. These very protests are what brought the
abuses of the Apartheid regime to the attention of people outside of South Africa
and lead to the external pressure placed on the regime. The ﬁnal stages of de-
autocratization where intiated by the Apartheid regime but only after the actions

of civil society forced the regime to the bargaining table.

63

Potential Pitfalls on the Pathway: Exploring the Various Outcomes of
Regime Transition

The ideal outcome in any regime transition is a peaceful outcome, one where
liberalization occurs with little ﬁghting or loss of life. While ideal, de-
autocratization rarely sees this outcome. Many times transition leads to coups,
repression, civil war, or a reversion to a more controlling authoritarian regime.
The question then, is how does a country ﬁnd itself facing on of these violent
outcomes?

We can assume that transition begins, for whatever reason, with a belief
that whatever goals are set by the actors can be reached peacefully. Any time
there is a violent outcome, both the regime and civil society will pay a high cost in
the loss of life, the money that goes into a violent conﬂict, loss of economic
production, and possibly the loss of power for the regime or eradication of some
segment of civil society. Knowing that the costs associated with violence can be
high, violence should only occur when the goals of the actors are too far apart to
peacefully negotiate a solution that both sides approve of. If the above
assumption is an accurate reﬂection of real world behavior, then the question that
must be answered is why do we see any of these possible violent outcomes?

The very process of de—autocratization can cause the outbreak of
violence. Mueller and Weede (1991) and Hegre et al (1998) agree that changes
in the level of repression can cause violence. In countries were repression is

high, rebellion is less likely to occur due to the high costs and slim chances of

success. In countries were repression is low, such as a democracy, rebellion is
less likely because the opposition has a voice and need not fear retaliation. Civil
society has peaceful avenues for voicing its disapproval making rebellion
unnecessary. It is when repression is at a moderate level that conﬂict is more
likely.

In countries were de—autocratization is occurring, repression is decreasing.
Some segment of civil society has been granted additional liberties. If
appropriate institutions have not been created in order to channel these new
liberties then members of civil society are faced with increasing liberties and no
formal method of exercising these new liberties. Additionally, Civil society may
desire additional liberalization once it has been granted its new rights. If civil
society chooses to press for additional liberalization the regime may see this
behavior as a threat to its continued existance and choose to use force to silence
civil society. In this case, de-autocratization leads to violence and not the
peaceful outcome that both actors would prefer.

Even if civil society chooses non-violent methods to demand additional
liberalization, regime members may see such demands as a challenge to its
authority and use force to end the call for additional liberties. By increasing
repression the regime could incite additional violence, leading to civil war. If
there are regime members who are interested in siding with civil society in calling
for increased liberalization, hard-liners in the regime may choose to remove the
reformers through a military coup. How civil society reacts to decreases in

repression and how the regime responds to civil society’s new demands will

65

dictate whether the outcome of de-autocratization is peaceful or violent and what
type of violence is seen.

Decreasing repression is not the only change in a country attempting de-
autocratization; possible changes in the economic structure of the society can
also inﬂuence the likelihood of a violent outcome. If the regime chooses to
relinquish some control over economic production there is an increase in
available resources, whether the resources are land or capital or both, for some
segment of civil society. Collier and Hoefﬂer ( 1999) ask what motivates civil
society to rebel: looting or justice seeking. Collier and Hoefﬂer develop two
models, one for looting and one forjustice seeking. In the looting model, Collier
and Hoefﬂer argue that an increase in resources will lead to an increase in
violence because there is more loot to be had. The regime's response, an
increase in attempted deterrence, can lead to an escalation of civil violence.
Additionally, the more proﬁtable the looting is the longer a civil war will last so the
more sweeping the economic changes the more likely de-autocratization will lead
to a violent outcome and the violence will last longer. Again, relinquishing some
control over civil society has the potential to end in violence.

The best intentions of the regime can serve as the impetuous for a violent
outcome. De-autocratization can lead civil society to believe that it can challenge
the regime, leading to an increase in repression or civil war, or through providing
an economic reason for lntemal violence by making more resources available to

civil society.

66

The Game Theoretic Model

In order to develop a better understanding of the interaction which occurs

between the regime and civil society, I develop a game theoretic model. By its

nature, game theory allows us to explore strategic interaction between actors in

order to understand how different outcomes are reached. Additionally, game
theory allows for different types of information, complete and incomplete, which
will help in developing a more accurate understanding of de-autocratization as
the model can be adjusted to more accurately represent the lack of information
that the actors face in the real world. As such, game theory provides us with a
powerful tool that can be used to answer the questions posed in this research.

The Game Tree

Nature

Reform Hard-Liner
Regime
Open ~ Open Open ~ Open
Civil Society
Mob ~Mob Mob ~Mob Mob ~Mob Mob ~M ob
Regime
SQ
LT 80
.. React ~React ..
React LT Rea ~React React React
iv "7 iv FT IV NT 'V FT

Figure 4.1 Complete and Perfect Information Game

67

 

 

 

 

 

 

 

Abbreviation Outcome

IV lntemal Violence

NT Negotiated Transition
LT Limited Transition

FT Forced Transition
SQ Status Quo

 

 

Table 4.1 List of Abbreviations for the Complete and Perfect lnforrnation Game

The Actors

The game tree above models the process of de-autocratization assuming
complete and perfect information. The assumption of complete and perfect
information is made here in order to better describe the game structure and will
be dropped further in the chapter.

There are two actors involved in this game, the regime and civil society.
The regime is composed of the primary decision makers within a given country.
The number of regime members will vary from country to country. The regime is
composed of reformers and hard-liners. Reformers are regime members who
favor some degree of liberalization. Hard-liners are regime members who prefer
to maintain the status quo and do not want to see liberalization occur. The
internal struggle faced by regime members is discussed in Chapter 2 of this

work.

68

 

Civil society is composed of various groups found within the country.
These groups can be based on religion, ethnicity, economic status, or other
factors that are deemed important within a country. Like the regime, there is no
set number of groups that can exist within a given country. As was discussed in
chapter 3, civil society must overcome the hurdles posed by a large scale
collective action problem. The degree to which civil society is able to overcome
the collective action problem indicates the strength of civil society. If civil society
is able to unite, for example 100 %, then civil society is considered strong. The
greater the level of disagreement within civil society, the more week civil society

and hence the more difﬁcult it is for civil society to challenge the regime.

The Order of Play

The game begins with nature determining if the regime is composed of
reformers or hard-liners. This is done in order to make the game more realistic.
Civil society does not know if it is facing a regime composed of hard-liners or
reformers. By allowing nature to determine if the regime is composed of hard-
Iiners or reformers, the model is able to capture the uncertainty that comes with
the lack of information and portray a more realistic model. Without a nature node
at the beginning of the game, there is an assumption that civil society knows the
regime’s type.

The regime moves next, choosing to open or remain closed. As
Przeworski (1991) acknowledges either actor can begin the process of de-
autocratization but in most cases the regime makes the ﬁrst move. Why is this

the case? The regime controls the institutions of governance and the military and

69

police forces. De-autocratization occurs when additional liberties are granted to
some segment of civil society, which only the regime is capable of allowing.
Additionally, the military and police forces can be used against civil society if civil
society attempts to force de-autocratization. For civil society to be able to force
change it must be able to overthrow the regime or convince the regime that is
strong enough to remove it from power. While there have been cases were civil
society has been able to remove the regime and begin the process on its own,
this is rare. Civil society can place pressure on the regime that eventually leads
to de-autocratization occurring, as we saw in Poland, Czechoslovakia, and South
Africa, even in these cases the regime eventually made the initial move to begin
to the process of regime transition. For this reason, the regime will make the
initial decision to open or remain closed.

Once the regime has moved, civil society must choose between mobilizing
or not mobilizing. Previous attempts at forrnalizing the process of de-
autocratization by Przeworski (1991), Crescenzi (1998), Casper and Taylor
(199?) have allowed civil society to move only if the regime chooses to open.
This does not accurately reﬂect how de-autocratization occurs. In the cases of
Poland, Czechoslovakia, and South Africa each country had a history of civil
society attempting to force de—autocratization, all of which were met with
resistance by the regime leading to the imprisoning or death of the movements
leadership. While the movements did not directly lead to de-autocratization
taking place, the cumulative effects of civil society’s efforts did play an important

role in pressuring the regime into eventually beginning the process of de-

70

autocratization. This must be accounted for if we are to understand how the
process of de-autocratization works and why these violent outcomes occur. The
only way to account for these instances is to allow civil society the option to
mobilize when the regime has decided to remain closed.

If the regime chooses to open and civil society chooses not to mobilize the
games end with a limited transition as the outcome. If civil society chooses to
mobilize, the regime is faced with a decision between reacting and not reacting.
The regime can choose to negotiate by not reacting or it can choose to attempt to
stop civil society's mobilization through the use of force. If the regime chooses to
react alter civil society chooses to mobilize, civil war is the outcome. Forced
transition occurs when the regime does not choose to open, civil society
mobilizes, and the regime does not react. In this case, the regime is removed

from ofﬁce by civil society.

The Outcomes

There are ﬁve possible outcomes from this game; lntemal violence, status
quo, forced liberalization, limited liberalization, and negotiated liberalization.
These outcomes occur once both the regime and civil society have moved,
allowing each an opportunity to react to the others decision. Since the research
is guided by a desire to understand how a country ﬁnds itself embroiled in a civil
war, I am not concerned with the ﬁnal outcome, for example who wins the civil
war. As such, there is not a nature node for probability of success for any of the

outcomes.

71

lntemal Violence occurs when the regime opens, civil society mobilizes,
and the regime reacts and when the regime remains closed, civil society
mobilizes, and the regime reacts. In both cases, conﬂict ensues clue to civil
societies choice to mobilize. The length and severity of conﬂict can vary from
case to case, but always involves some clash between civil society and the
regime.

It can be argued that Soviet Union began to de-autocratize in 1987 when
Gorbachev announced his new policies of glasnost and perostrokia. The Soviet
Union has ceased to exist, replaced by Russia and a series of independent
Republics. In Russia, de—autocratization has led to the development of a
democratic government that ﬁnds itself struggling with a wide variety of political
and economic challenges. One such challenge has been the civil war in the
Republic of Chechnya. In this case, the regime allowed democratization to
occur, the Chechnyans responded with a demand for greater autonomy, which
the regime denied. When the Chechnyan people began to use force in order to
gain autonomy, the Russian regime responded in kind. In this case, the regime
opened, civil society in Chechnya mobilized,a nd the regime responded with the
use of force.

In South Africa, the regime choose to remain closed, civil society
mobilized, and the regime responded with the use of force. While the regime
eventually choose to open, it was only after decades of armed conﬂict with the
Black population who had been demanding the end of Apartheid. Many of the

Black leaders, such as Mandela and Bike, where imprisoned and killed in an

72

attempt to silence the population. When the Black population organized marches
in protest of Apartheid the regime responded with deadly force. While the
violence never reached the levels used by scholars to code the case as a civil
war, there is no doubt that lntemal violence was occuring due to movement to
end Apartheid.

The transition which occurred in Romania in 1989 is another example of
internal violence, although an example which was a great deal more abrupt then
the transitions in Russia and South Africa. In 1989 many of the Eastern Bloc
countries where able to end the rule of the Communist regime because of the
actions taken by civil society. Most of these cases proved to be non-violent
transitions. In Romania the population began to call for the end of the communist
regime through protest. The protest quickly transformed to a mob action which
eventually lead to the capture and execution of Nicoli Ceausescu.

The Status Quo occurs when the regime chooses to remain closed, civil
society does not mobilize, and the regime does not react. This series of events
leads to no changes in the country. Two of the Soviet Unions former republics,
Latvia and Estonia, serve as examples of the status quo outcome. Between the
years of 1991-1996, Estonia and Latvia saw no change in its regime score in the
Polity 3d data set nor did either country experience any violence.

Forced Transition occurs when the regime remains closed, civil society
chooses to mobilize, and the regime chooses not to react. Liberalization, an

increase in some areas of political, social, or economic liberties, occurs but not

73

because the regime is interested in reforrn but because civil society forces it.
The regime may be replaced, if it is not it suffers a loss of power.

Czechoslovakia and Poland, serve as excellent examples of cases of a
negotiated transition. In the Case of Czechoslovakia and Poland, there was a
long history of struggle against the ruling Communist regime. The leaders of the
movements, Dubcek, Havel, and Welesa among others, where jailed in an
attempt to silence the movement. In the spring of 1968, an attempt to reorganize
the Communist regime in Chezcoslovakia was ended when Warsaw Pact tanks
arrived in Prague. In Poland, the labor movement organized strikes and protests
as a way of demanding de-autocratization. While the intial pressures placed on
the regime by civil society lead to arrests and no change, they set the stage for
the events of 1989. In 1989, the people of Poland and Czechoslovakia saw that
the regime was weakened and rose against the regime. The regimes in these
countries choose to accept civil societies demand for de-autocratization and the
change occurred in a peaceful manner.

Limited Transition occurs when the regime chooses to open, civil society
chooses to not mobilize, and the regime chooses not to react. In this scenario,
civil society is accepting the additional liberties that the regime is granting it and
choosing not to ask for more by mobilizing. Iran is a good example of a country
undergoing the process of de-autocratization that has led to a limited transition.
The last two elections have led to victories for candidates interested in reform,
including the President Khatami-Ardakani in 1987. Since his election, there has

been an increase in the number of non-govemment publications, political parties,

74

and student groups. While there has been some reform, it is believed that Iran
has not moved as quickly as the reformers would like to due to the presence of
the Council of Clerics. The Council of Clerics is a great deal more conservative
then the elected representatives and has such powers as approving political
candidates. The presence of the Council of Clerics has has led to a limited
transition in Iraq.

Negotiated Transition occurs when the regime chooses to open, civil
society chooses to mobilize, and the regime chooses not to react. In not
reacting, the regime indicates a willingness to work with civil society, or some
segment of civil society, and liberalization comes from negotiation between the
regime and civil society.

While it may seem that the different types of transition, limited, negotiated,
and forced, are really all the same thing, it is necessary to realize that the
differences in these types of transitions is important. A limited transition allows
for an increase in personal liberties for some segment of civil society that the
regime wanted. There is no challenge to the regimes new policy. A negotiated
transition occurs when the regime enters into some type of negotiation with civil
society. In doing this, the regime sends a signal to civil society that can be
interpreted in later iterations of the game. The regime is indicating that it is
composed of mostly reformers who are willing to move ahead with de-
autocratization or that the regime believes that civil war will be too costly or that
they cannot win a civil war. Finally, a forced transition means that the regime is

removed from ofﬁce by civil society. While all these outcomes indicate some

75

form of transition, how transition occurs is different and will affect future iterations
of the game.

Finally, it has been suggested that the above game does not allow for
democratization to occur. This is not the case. Transition reﬂects movement
towards a democracy even if democracy is not the ﬁnal outcome. Democracy
comes about through repeated interactions between the regime and civil society.
This can be represented as an outcome derived from iterating the game
presented in this chapter. This is a more accurate representation of how the
process of de-autocratization works. Steps are taken that may lead to
democracy but democracy is not a guarantee. By focusing on the steps, the
game acknowledges the gradual nature, even though it is sometimes conﬂictual,
of democratization.

Assumptions

In order to develop cost terms and preference orderings, it is necessary to
examine the assumptions that are used to build the model. Like all game
theoretic models, it is necessary to assume a unitary actor. On its face, this
assumption appears problematic for a model of strategic interaction between the
regime and civil society. It is difﬁcult to argue that civil society is a unitary actor.
As was seen in chapter three, civil society can be divided and the depth of those
divisions will inﬂuence civil society’s behavior when interacting with the regime.
In order to address this issue, chapter three develops a model which explores
how civil society is able, or unable, to solve the collective action problem it faces.

That information is used in this model in order to determine what actions civil

76

society will take when interacting with the regime. Civil society in this model
then, is deﬁned as a unitary actor in so far as it represents the ﬁnal outcome
reached when civil society attempts to solve its collective action problem as
described in chapter three.

A similar argument can be made about the regime. The regime is not
necessarily a homogeneous group. The model found in chapter two addresses
this concern. The model depicts the interaction between regime members and
feeds into the model found in this chapter. The regime knows it type based on
the outcome of the interaction between regime members. It is because of the
models developed earlier in this research that we can assume a unitary actor in
this model.

Beyond the unitary actor assumption, there are a seriesof assumptions in
regards to both actors. Let us ﬁrst explore the assumptions that surround the
regime.

1) Regime members prefer more control to less control.

1a) Hard-liners: U(SQ)=1, U(LT)=g, U(FT)=0 where 1>g>0
1b)Refomiers: U(LT)=1, U(SQ)=s, U(FT)=0 where 1>s>0

2) The value of transition for Reformers is enhanced by 2w>0 and 2 for

hard-liners where w (0,1) therefore UHL(LT)=g+2 and UR(FT)= 0+2w

3) lntemal Violence is conﬂict over government control. The value of

Internal Violence for the regime, Reformer or Hard-Liner, is based on the

Expected Value of conﬂict:

Hard-liner: {pU(SQ) + (1-p)U(FT)}

77

Reformer: {pU(LT) + (1-p)U(FT)}
minus the physical cost of the conﬂict, k>0, where p(Regime Wins)
EV(Civil WarlOpen)=pg-k and EV(Civil War|~Open)=p-k
4) Negotiated Transition is an outcome which the regime has shown a

willingness to open. As a bargaining problem, EV(NT) for either regime

type is {pU(LD+(1-p)U(CW)}. EVR(NT)=p1+2w > EVHL(ND=p9+2.

Costs

There are costs associated with any action taken by the regime and civil
society. Bothcivil society and the regime faces costs associated with opening,
mobilization, and ﬁghting. The table below summarizes the costs and the Greek

letters associated with each cost.

 

)3, Cost of Opening for the Regime

 

23wr Weighted Cost of Opening when Regime
is composed of Reformers

 

 

 

 

2c. Cost of Opening for Civil Society
:, Cost of Mobilizing for the Regime
:m Cost of Mobilizing for Civil Society

Mr Cost of Fighting for the Regime

 

Mwr Weighted Cost of Fighting when the
Regime is composed of Hard-liners

Mm Cost of Fighting for Civil Society

 

 

 

 

 

Table 4.2 A List of Cost Terms

78

If the regime chooses to open, there is a cost incurred by both the regime
and civil society. The regime incurs some loss of power but can gain additional
legitimacy. This additional legitimacy can come in the form of a more satisﬁed
civil society to an increase in funds received from other countries or international
organizations. As such, the regime can receive a positive beneﬁt for opening.
Opening can also lead to a negative payoff for the regime when it does not work
out as intended. If civil society attempts to force transition beyond were the
regime is willing to go, opening can lead to coups and civil war. Additionally,
Reformers pay a different cost for opening then hard-liners do. Reformers are
interested in some degree of transition, so they receive a weighted cost when
opening is selected. Civil society also incurs a positive beneﬁt. The opening
increases some political, social, or economic liberties. As the percentage of the
population which beneﬁts from the opening increases, so does the positive
beneﬁt incurred from opening.

If civil society chooses to mobilize both the regime and civil society incur a
cost. Civil society must pay for costs as mundane as photo copying ﬂiers,
brochures, pamphlets that it may use to educate others or to rally people to the
cause. Mobilization also brings with it the possible cost of imprisonment and
providing arms to those willing to use physical force. These costs can be offset
by the benefits received from successful mobilization, an increase in liberties or
costs can increase if mobilization fails and the regime cracks down and reduces
liberties. The regime also pays a cost if civil society mobilizes. It must provide

for additional police or military force, if it attempts to end the mobilization with a

79

show of force as well as a loss of power that comes from the mobilization. As
such, the regime never receives a positive payoff from mobilization while civil
society can receive a positive or negative return.

Finally, both the regime and civil society incur costs if ﬁghting occurs, if the
regime chooses to react to civil society attempt at mobilization. Both sides will
face casualties due to the ﬁghting. The longer and more violent the conﬂict
becomes, the greater the cost. Both actors must ﬁnd a way to pay for arming
those involved in the ﬁghting as well as a loss of economic beneﬁts for two
reasons, 1) the destruction of the infrastructure 2) the loss of economic gain
associated with production that could have occurred during the ﬁghting. The
costs increase for both actors if they lose. The regime loses power, control of the
country, and most likely the members of the regime will be killed or imprisoned.
Civil society faces the possibility of genocide, increased repression, and the loss
of even more political, social, and economic liberties. The only time that the
regime receives a positive payoff is when it uses force to increase its control over
the country. Hard-liners receive a weighted cost for ﬁghting. Since hard-liners
are not devoted to reform on any level, they will pay a different cost if ﬁghting is
necessary then reformers would pay.

Preference Orderings

One of game theories strongest assets is the ability to examine how
different actors, and different types of actors, behave in an environment requiring
strategic interaction and the resulting outcomes. The game presented in this

chapter identiﬁes two actors, civil society and the regime, with the regime having

80

 

 

 

 

 

 

 

 

 

 

 

 

 

Outcomes Regime Civil Society
lntemal Violence -Zw, -:r -MIr -2,,. -:cs -M,,.s
lntemal Violence‘ -2‘., -:r -Mwr -£,,. 4,, -M,,s
lntemal Violence w/ Regime .;r .Mr .;cs .Mcs
Close

lntemal Violencew/ Regime -:r -Mw, -:cs -Mcs
Close‘

Forced Transition -:r +:,,,

Forced Transition ‘ -:r +2cs

Limited Transition +ler +2,s

Limited Transition ‘ +2, +22cs

Negotiated Transition +22wr -:r +2,s +2cs

Negotiated Transition ‘ +2, -:r +21“ +:cs

Status Quo

 

 

 

Table 4.3 Establishing the Preference Orderings

two types; reformers, hard-liners, . The different types of regime is based on the
composition of each actor at the time the game begins. The regime is composed
of reformers when the members dedicated to some form of liberalization are able
to control the regimes actions. In the following paragraphs I deﬁne the preference
orderings for the actors. As in all cases, there can be different preference
orderings based on extreme cases, such as a regime which is composed of hard-

Iiners but realizes that the reformers are close to taking power and wish to

81

 

cement their factions control by allowing some liberalization. The preference
ordering developed for this research are meant to reﬂect what a “pure” hard-liner
or reformer prefers.

Vlﬁthin the regime the reformers preference orderings reﬂect the desire for
some type of liberalization. Reformers desire a change in the regimes policy, but
they prefer to control the process of de-autocratization, as such they prefer
Limited Transition over all outcomes. Negotiated Transition grants the regime
some degree of control even though the regime is required to relinquish more
control then it initially wanted to relinquish. Since negotiate transition allows
some liberalization and grants the regime some control over how the
liberalization is enacted it is prefered to the Status Quo. The Status Quo does
not allow for any type of change in the regimes control over the population, which
is contrary to the goals of a regime composed of reformers. Reformers prefer the
status quo to forced transition because the regime would lose all control if they
are forced from ofﬁce. Since reformers want to maintain some control over civil
society, even if they are willing to change the type of control that it has, removal
from ofﬁce is not a desired outcome. Yet forced transition is preferred to Civil
War. Civil war comes with high costs, both in terms of human life and economic
cost. Reformers are interested in allowing some type of change in the
government not the devastation that civil war brings.

If the regime is composed of hard-liners, the preference orderings take on
a entirely different shape. Hard-liners are interested in maintaining their position,

if not increasing their power. They are not interested in de-autocratization. Hard-

82

liners prefer the Status Quo to all other outcomes. The status quo maintains
their current level of control over the country, insuring continued beneﬁts that
regime members receive from holding ofﬁce. If some type of transition is
required, due to high levels of external pressure or the promise of rewards that
would overcome the amount of power that is relinquished, hard-liners prefer
Limited transition. Limited transition allows the hard—liners to control how much
power the regime relinquishes. Civil war is preferred to the remaining forms of
transition. The hard-liners are willing to relinquish some degree of control but are
not willing to move beyond the liberties that they propose. If civil society
challenges a hard line regime, then the hard-liners prefer civil war to any other
outcome. Civil war allows the regime some chance of regaining complete control
over civil society, and possibly increase its control if it wins. Negotiated
Transition is preferred to Forced Transition. Vlﬂiile the hard-line regime must
relinquish more control then it was initially was willing to, negotiated transition still
allows the regime to maintain ofﬁce and some degree of control over how the
transition occurs while still receiving some beneﬁts from remaining in ofﬁce.
Forced transition is the worst outcome for the hard-line regime as the regime is
forced out of ofﬁce and loses all control and any beneﬁts is had been receiving
while in ofﬁce.

Civil society Negotiated Transition to all other outcomes. Negotiated
transition allows civil society a voice in what type of liberties are granted and
guarantees that some of the changes that are made in the governments behavior

address issues that are important to civil society. Forced Transition is the next

83

most preferred outcome. Forced transition allows civil society to meet more of its
goals then any other type of transition but requires some use of force and is more
risky then negotiated transition. Preferring some degree of liberalization to none,
civil society prefers Limited Transition to the Status Quo or Civil War as it allows
some new liberties, even if those liberties are not the same as the ones that civil
society might want to see granted. Due to the cost of Civil War, civil society
prefers the Status Quo to Civil War. Civil war means that there is a possibility

that the regime is able to eradicate or greatly weaken segments of civil society.

Reformers: Limited Transition > Negotiated Transition > Status Quo > Civil War
>Forced Transition.

Hard-liners: Status Quo > Limited Transition > Civil War> Negotiated Transition
> Forced Transition

Civil Society: Negotiated Transition > Forced Transition> Limited Transition>

Status Quo > Civil War

Solve the Complete and Perfect Information game

Before exploring what the outcome of the game is using incomplete
information, it is necessary to identify equilibria assuming complete and perfect
information. By assigning ordinal payoffs to the possible outcomes we can
identify the equilibrium using backwards induction. Five represents the highest

rated outcome and 1 the lowest rated outcome.

84

 

Actor LT NT SQ FT IV

 

 

 

 

Reformers 5 4 3 2 l
Hard-liners 4 2 5 l 3
Civil Society 3 5 2 4 l

 

 

 

 

 

 

Table 4.4 Cardinal Payoffs for Each Actor. 5 is the best, 1 is the worst

Nature

Reform Hard-Liner
Regime
Open ~ Open Open ~ Open
Civil Society
Mob ~Mob Mob ~Mob Mob ~M ob Mob ~Mob
Regime
React LT :0 LT R ct Q
~ , ~ ea
Ru“ 5 React ~Rmt 2 React React 3 R g
3
IV 4m iv FT iv NT 'V fr
2 5 2 I 3 2 1 4
I l 4 I 5

Figure 4.2 Backwards Induction for the Complete and Perfect Information Game

85

 

The above outcomes are not surprising. If civil society knows that it is playing
against reformers, it will choose to mobilize knowing that the reformers prefer
some form of transition to none. The result is a Negotiated Transition which
beneﬁts both actors. When civil society knows that it is facing a hard-line regime,
Status Quo is the outcome. Civil society is aware that the regime will not tolerate
mobilization and the regime knows that if it opens, civil society will mobilize. As

such, the regime chooses not to open and civil society chooses not to mobilize.

Incomplete and Imperfect information Game

Is it realistic to assume that civil society has complete and perfect
information? Is it safe to assume that civil society knows the composition of the
regime? I argue that it is not. In authoritarian regimes, access to information is
restricted, how restricted will depend on the degree to which the regime is
authoritarian, which decreases civil society’s ability to determine what type of
regime it is facing. As a regime becomes more open, the amount of information
will increase, even this does not guarantee that civil society is fully informed as to
the type of regime it faces.

In order to increase the models accuracy, it is necessary to assume that
civil society does not have complete information. As mentioned above, the
amount of information that civil society has will differ based on the regime and
how strictly the regime controls access to information. This can be taken into

consideration by altering how strongly civil society believes it knows the regime’s

86

type. This is modeled using the game below, which adds an information set
where civil society is faced with the decision to mobilize or not mobilize. The
information set indicates that civil society is not sure what branch of the game
tree they are on, are they playing against reformers or hard-liners? In order to
solve the game, it is necessary to use Perfect Bayesian Equilibrium anaylsis.
Using the same ordinal preferences that were used in solving the complete and
perfect information, we can demonstrate why it the information sets are
necessary. Recall that in the complete and perfect information game, the
equilibrium outcomes were Negotiated Transition when playing against reformers
and Status Quo when playing against hard-liners. What happens when civil

society mistakes the regime’s identity?

87

Nature

Reform Hard-Liner
Regime
Open ~ Open Open ~ Open
.—.—.a—e—.—e—O .:;.-':—-'.’°.‘H: .............. CivilSociety
-- ~ ........
Mob ~Mob Mob ~Mob Mob ~Mob Mob ~Mob
Regime
R React L'l‘ :0 LT R t S
am 5 React ~React 2 React React g React eec ;
3
W L" [V FT iv NT 'V fr
2 5 2 l 3 2 l 4
l I 4 l 5

Figure 4.3 The Incomplete and Imperfect lnforrnation Game

If civil society believes that it is facing a hard-line regime when it is actually
facing a reform regime, the outcome becomes Limited Transition. The reform
regime opens but civil society does not mobilize fearing that if it does it will lead

to civil war, its least preferred outcome. Compare this to the original outcome of

Negotiated Transition in the complete and perfect information game. The reform
r eQime prefers Limited Transition to Negotiated Transition but civil society prefers
Negotiated Transition to Limited Transition. By misreading the regimes type, civil
society fails to press for additional reforms and loses an opportunity to improve

its position within the country.

88

If civil society believes that it is facing a reform regime when it is actually
facing a hard-line regime, the outcome is Civil War. The regime does not open
but civil society mobilizes. Hard-liners prefer Civil War to Forced Transition and
so initiate the use of force leading to a civil war. Compare this to the Status Quo
outcome found in the complete and perfect information game. Civil society would
have been better off choosing not to mobilize, Status Quo is preferred to Civil
War, by misreading the type of regime it is facing, civil society ﬁnds itself

embroiled in a civil war that it did not want.

In both cases, the outcome changes when a lack of information is taken
into consideration in the game theoretic model. This indicates that information is
essential to the outcome of the game. Knowing this, it gives the regime reason
to bluff. Reformers are interested in some degree of de—autocratization but prefer
complete control over the degree of reform. Knowing that information is
important, reformers have an incentive to pretend to be hard-liners as it could
lead to an outcome of limited transition and not negotiated transition.
Alternatively, hard-liners that are interested in increasing their control over the
country may bluff and present themselves as reformers. In doing so, civil society
may be baited into mobilizing giving the hard-liners reason to initiate conﬂict

allowing them the possibility of removing trouble makers or groups that hard-

"ners feel threaten their power base.

SOlving the Incomplete and Imperfect Information Game

89

While the use of backwards induction demonstrates the importance of
information in the cases where civil society and the regime interact strategically,
there are more sophisticated equilibrium options that will enhancing our
understanding of the interaction between the regime and civil society. Due to the
importance of information in this game, a Perfect Bayesian Equilibrium solution
set is appropriate. Once the regime, which knows its type, moves, civil society
must make a decision to mobilize or not mobilize. This decision is made, in part,
based on civil societies belief that the regime is composed of reformers or hard-
Iiners. The regime’s move sends a signal to civil society that it must interpret
when making its decision.

As was seen in the complete and perfect information game, reformers will
open when hard-liners will not. If the regime chooses to open then, civil society
may interpret this as a signal that the regime is composed of reformers. This is
not known with certainty. Hard-liners prefer the Status Quo to any form of
Transition but, as was seen in chapter two, there are reasons that will lead hard-
liners to allow some degree of opening. If civil society misreads the actor it is

playing against, then internal violence becomes a possibility.

p(Ail B) = p(Ai)p(B| Ai)/ p(A)p(Bl A) + p(~ A)p(B I~ A)
Using Bayes Theorum, we are able to solve the game.

For the game presented here, the formula is as follows:

p(ReformlOpen)= g(Refonner)Q(Open|Reformer)
p(Reformer)p(Open|Reforrner)+p(Hard-liner)p(Open|Hard-Iiner)

90

If the probability of Reformer is set to 112 then the probability of Hard-liner is also
1l2. Using the preference orderings found in prior discussion, we can assign
ordinal values to the payoffs and solve for the equilibrium.

p(ReformerlOpen)= “241) = 1l2 = 1l2

1/2(1/2)+1l2(1/2) 1l2 +1/2

What the above ﬁnding means is that civil society will choose to mobilize ﬁfty
percent of the time and not to mobilize ﬁfty percent of the time. This is due to the

fact that civil society does not have enough information to determine what the

regime type is.

Conclusion

The game presented in this chapter can only take place after the regime
determines its type, this is accomplished through the non-cooperative bargaining
game in Chapter Three. Once the regime determines what type it is, it then will
choose to open or not open. It is in the regimes best interest to hide its type from
civil society, as demonstrated in the above example, civil society will not mobilize
if they know the regime is comprised of Hard-liners and is more likely to mobilize
if it determines that the regime is made of Reformers.

In addition, we have Ieamed that internal violence is more likely to occur
when civil society believes that the regime is comprised of Reformers when in
reality the regime is made of Hard-liners. Based on their preference orderings,

Hard-liners wish to maintain the status quo. If a Hard-line regime does open, it

91

will not tolerate mobilization and lntemal violence will be the outcome. lntemal

violence will not occur when the regime is comprised of Reformers.

92

Chapter 5

OPERATIONALIZATION

In Chapters Three and Four the argument is made that the interaction
among regime members and between the regime and civil society during the
process of de-autocratization will inﬂuence the likelihood of lntemal conﬂict. In
the course of these chapters, it is argued that economic conditions, the degree to
which civil society Is fragmented, and the willingness of some regime members to
liberalize will directly impact the course of de-autocratization. The following two
chapters are dedicated to exploring the validity of the previous arguments using

empirical methods.

The Universe of Cases

In order to include the variables that inﬂuence the likelihood of internal
conﬂict in countries undergoing the process of de—autocratization, it is necessary
to draw data from four unique data sets. The data sets used in this research are
the Polity 3d data set, Wallensteen and Sollenbergs’ Data on Armed Conﬂict,
Ellingsens’ ‘Ethnic Wrtches’ Brew data set, and the World Bank Economic
Indicators data set.

The nature of the research question requires that the countries included in
the research were non-democracies in the years 1989-1996, the duration of that
data set, and where there was an attempt to de—autocratize. The polity 3d data

set was used to determine which cases which would be included in the data set.

93

The Polity 3d data set includes two variables, autocracy score and democracy
score. Both variables are based on a scale of 0-10, with 0 representing a low
degree of democracy or autocracy and 10 representing a high degree of
democracy or autocracy. An index variable, titled regime, was created by ﬁrst
changing the autocracy score to a negative value and then adding the autocracy
score and the democracy score. The regime variable was bounded between 10
and -10, with 10 representing a pure democracy and -10 representing a pure
autocracy. The decision rule to include a country in the empirical research
required that a country score a 6 or less in at least one year in 1989-1996. In
order to determine if a country had de-autocratized, its regime score had to
improve at least once between the years 1989-1996.

The Polity 3d data set was selected because of its inclusion of a variable
that measures both democracy and autocracy. In doing so, the Polity data
acknowledges that autocracy is not the mere absence of democracy and vice
versa. As such, the measures provide an accurate depiction of a country’s
regime type, which is essential for this research.

In addition to identifying which countries to include in the data set, the
Polity 3d data will be used to develop measure of regime change and regime
stability. The Polity data set is regularly updated which allows for the data to
reﬂect changes within a country's institutions of governance. How these

variables are constructed will be discussed later in the chapter.

94

The Data Sets

The Wallensteen and Sollenbergs' Data on Armed Conﬂict is used to
identify which the countries where conﬂict occurred and in what years. The
Wallensteen and Sollenberg data was selected for two reasons, the twenty-ﬁve
battle dead threshold and the possibility of multiple conﬂicts within a country.
The alternative to the Wallensteen and Sollenberg data set is Singer and Smalls'
Correlates of War Civil War data set. Singer and Small establish one-thousand
battle dead threshold for a case to be included in the data set. While this is a
reasonable threshold to identify a Civil War, indeed it is the threshold that
Wallensteen and Sollenberg use to identify a civil war within their own data set, it
is not reasonable to assume that the only type of internal conﬂict that can occur
during the process of de-autocratization is that of a civil war. Indeed, the models
developed in chapters three and four of this dissertation identiﬁes multiple
possible outcomes; political unrest, coups, and civil war. In order to develop an
empirical model that accurately represents the theory develop din the earlier
chapters of this dissertation it is necessary to a data set that establishes a lower
battle dead threshold for inclusion. The Wallensteen and Sollenberg data does
just that. In including cases where there is a minimum of twenty-ﬁve battle dead,
the Wallensteen and Sollenberg Data of Armed Conﬂict, allows for the inclusion
of cases where a bloody coup occurred, riots, terrorism, violent protests, and
other such outcomes.

Additionally the Wallensteen and Sollenberg data takes into consideration

the possibility of multiple conﬂicts within a country in any given year. For

95

example, in Indonesia it distinguishes between the conﬂicts in East Timor and
Aceh. Why is this important? Chapter four explores the inﬂuence that the
interaction between the regime and civil society has on the process of de-
autocratization and the possibility of internal conﬂict. It is argued that the degree
to which civil society is fragmented will inﬂuence its ability to effectively mobilize.
One possibility is that the more fragmented civil society is, the more likely that
there can be multiple challenges to the regime. While the game theoretic model
does not allow for this possibility, due to the intractability of the game once it
becomes an N person game, there is no reason to limit the empirical model in the
same manner. The Wallensteen and Sollenberg data allows for multiple conﬂicts
to exist in a given country. Using this data then allows the empirical model power
that is not found in the game theoretic model, which will strengthen the results of
the empirical model.

The Wallensteen and Sollenberg data set is not without its weaknesses.
Its greatest weakness is the time period for which data is available, 1989-1999.
This limits the empirical research to the Post Cold War period, preventing the
results from being generalized across time. While this is a weakness, the lack of
a data set with a similarly low threshold for inclusion requires the use of the
Wallensteen and Sollenberg data.

The third data set to be incorporated in this research is Ellingsens' 'Ethnic
Wrtches' Brew data set. In chapter four, the possibility of a fragmented civil
society is raised. In order to reﬂect accurately the theoretical concepts in the

empirical model it is necessary to include a measure of fragmentation within civil

96

society. The Ellingsen data provides these measures as she develops measures
of ethnic, religious, and linguistic fragmentation within a country. The Ellingsen
data set includes data from 1945-1994. This data is expanded to include the
years 1995-1999. It is assumed that the number of ethnic, religious, and linguistic
groups will remain constant overtime. In cases where there is a change in the
number of groups in a country, the years 1995-1999 are coded as missing data.
The ﬁnal data set used in this research is the World Bank Economic
Indicators 1998 data set. This data provides measures of economic development
for all countries in the international system through. the year 1996. As discussed
in chapters three and four, changes in a country's economic well being can be
the catalyst for liberalization and violence. As such, it is necessary to include
data that reﬂects the changes in a country's economic well being. The World
Bank data set provides a wide variety of variables that can be used to measure
economic change for all countries and years included in the data set.
Because the World Bank data was available through the year 1996, and the
Wallensteen and Sollenberg data does not begin until 1989, that data set is
limited to cases that include all authoritarian countries in the international system
from the years 1989-1996. While the limited time period decreases the ability to
generalize across time, it should not affect the models ability to demonstrate the
necessity of unpacking the possible outcomes of de-autocratization and not

simply focusing on democratize or not democratize as outcomes.

97

The Variables
The Dependent Variable: Outcome

The research question asked in this dissertation is when will the process
of de—autocratization lead to violence? As such, the dependent variable identiﬁed
in this research is the outcome of the process.

The preceding theoretical chapters argue that the study of de-
autocratization has been done a disservice by dichotomizing the possible
outcomes to "democratize" or "fail to democratize". The game theoretic models
developed in chapters three and four allow for the following outcomes, limited de-
autocratization with no violence, de-autocratization without violence, coups, and
violence. The violence outcome is left ambiguous in the game theoretic models
due to the difﬁculty of developing models that can clearly differentiate between
when riots, terrorism, or civil war may be possible outcomes.

This same difﬁculty does not exist when creating the dependent variable
for the empirical models. Wallensteen and Sollenbergs' data set provides the
basis for building the dependent variable titled Outcome. Wallensteen and
Sollenberg provide the years when conﬂict occurs for every country in the data
set. Additionally the variable "intensity level"10 indicated the number of battle
dead per year for each conﬂict. While the intensity level measure was a good

place to start, it did not accurately reﬂect the dependent variable.

 

1° Intensity Level: Found in the Wallensteen and Sollenburg data set.

0 - No conﬂict activity.

1 - Minor armed conﬂict: At least 25 battle-related deaths per year and fewer than 1,000 battle-
related deaths during the course of the conﬂict.

2 - Intermediate armed conﬂict: At least 25 battle-related deaths per year and an accumulated
total of at least 1,000 deaths, but fewer than 1,000 per year.

3 - War. At least 1 000 battle-related deaths per year.

98

The Outcome variable must differentiate between the type of conﬂict, not
simply the intensity of the conﬂict. In order to properly develop the Outcome
variable, the Wallensteen and Sollenberg data as an indication of when conﬂict
occurred and then conduct additional research in order to identify the nature of
the conﬂict. For the purpose of this research, the Outcome variable was broken
into ﬁve categories: no conﬂict, coups, political unrest, minor conﬂict, and civil
war.

The outcome variable was created using information found in the
Keesings Worldwide archives. The Keesings information was matched to each
country in order to determine if the conﬂict was a coup, political unrest, minor
conﬂict, or civil war.

A case was coded as a coup when the Keesings information indicated that
the only conﬂict in a country, in a given year, had been coup. The number of
battle deaths that occurred based on the coup was not taken into consideration.
Based on the reports found in the Keesings archives the following cases were
identiﬁed as coups: Paraguay in 1989, Azerbijan in 1993, Liberia in 1989,
Guatemala in 1989, Burundi in 1992, Liberia in 1989, and Sierra Leone in 1992
and 1 993.”

Cases where a segment of civil society had participated in protests against the
government and the government responded with the use of force where coded
as POIitical unrest. Additionally, for a case to be coded as political unrest, the

Conflict had to occur in one area and was not sustained. I deﬁned sustained

99

conﬂict based on the conﬂict continuing beyond one week. Forty-four cases of
political unrest were identiﬁed. According to Keesings, cases such as Nicaragua
(1989) and Niger (1990-1994) were coded as political unrest. In both cases,
there was conﬂict that did not last beyond a week and there were fewer then
1,000-battle deaths.

Minor conﬂict was deﬁned as conﬂict between the government and some
segment of civil society that occurs over an extended period of time and/or in
multiple locations. In addition, the death toll had to fall below 1,000 dead. There
were forty-ﬁve cases identiﬁed as minor conﬂict.

Civil war was coded as when conﬂict between the government and some
segment of civil society that occurs over an extended period of time and/or in
multiple locations and there were more then 1,000 battle deaths. Seventy-four
cases were identiﬁed as civil wars.

In many cases, the type of conﬂict changed year from year, depending on
the severity of the ﬁghting. Most frequently, cases ﬂuctuated between minor
conﬂict and civil war. One example of such a case is Rwanda. In 1990, Keesings
indicated that there was conﬂict through out the month of October with the death
toll ranging some where 100-1,000 dead. Due to the fact that there is not a ﬁrm
death toll reported, Rwanda in 1990 is coded as a minor conﬂict. From 1991-
1994, Rwanda is considered a civil based on the amount of conﬂict and the death

toll rising over 1,000 dead each year.

 

‘1 The only coups that are included in the data set are coups were there was a minimum of
twenty-ﬁve battles dead. This is due to the nature of the Wallensteen and Sollenberg data. In the
future, the data set will be expanded to include bloodless coups.

100

ln coding the outcome variable, several problems arose. First, there were
several instances were the Keesings archive did not provide evidence that
conﬂict had occurred in the year indicated in the Wallensteen and Sollenberg
data set. In these cases, I assumed that the Wallensteen and Sollenberg data
was correct and l coded the case as political unrest. This problem arose most
frequently in cases were there were conﬂicts in many different territories, for
example Myanmar. Keesings would report conﬂict but would not always identify
were the conﬂict was occurring. In all of these cases, Wallensteen and
Sollenberg coded the death toll as falling between 25-999 battle deaths. I choose
political unrest, as it was the lowest level of conﬂict that I allowed for. This
allowed me to keep the cases in question. in the data set but prevented me from
being more speciﬁc in the type of conﬂict that had occurred.

Secondly, there were several cases where the Keesings archive indicated
a death toll that was different from the death toll coded in the Wallensteen and
Sollenberg data. In the case of Rwanda, Keesings reported fewer then 1,000-
battle deaths, for many years Keesings indicated fewer then SOD-battle deaths,
while Wallensteen and Sollenberg indicated greater then 1,000 battle deaths. In
these cases, the death toll reported by Wallensteen and Sollenberg was used. I
choose to use the Wallensteen and Sollenberg data due to Keesings lack of a

death toll in many of the cases.

101

The Independent Variables

Regime Type

As was discussed earlier, a country's regime type was determined through the
use of the Polity 3 data set. This data set provides a measure of democracy as
well as autocracy, both ranging between 0-10 with 0 being the least democratic
or autocratic and 10 being the most democratic or autocratic. Using these two
variables I was able to develop an index variable of regime type. For the
purpose of this research, any case scoring greater then a 6 was declared a
democracy and removed from the universe of cases.

The cases that were left range for a score of 6 to -10, developing an
ordinal measure for regime type that allows me to distinguish between countries
that are on the verge of becoming a democracy and those that are highly
autocratic. It should come as no surprise that there was a spike in the number of
countries that transitioned from an authoritarian regime to a democratic regime in
1989 following the fall of the Berlin Wall and the Soviet Unions decision to allow

its satellite countries determine what type of regime they wanted.

Measures of Transition

The goal of this research is to better understand when de-autocratization
leads to violence. In chapters three and four, game theoretic models are
developed that allow us to better understand how the process of de-
autocratization occurs and how the process can lead to a violent outcome. If this

research is to accurately develop an empirical model that enhances the ﬁndings

102

of the game theoretic models it is necessary to measure when how quickly de-
autocratization occurs. The measures used in this research are titled regime
stability and regime change. Both measures are developed based on the regime
index variable developed from the Polity 3d data.

regime stability and regimes change are similar concepts yet the
differences are very important. Regime change represents the degree to which
the regime is willing to change its policy. Regime change, then, is an attempt to
better identify the degree to which the regime is willing to open. The game
theoretic model in chapter three demonstrates that the large the proposed regime
change, the less likely the opposing faction will be to attempt a coup. As such, it
is imperative that the empirical model includes a measure that indicates how
much change occurred in a given year if we are to understand when coups may
OCCUR

The Polity 3d regime score was used to develop the regime change
variable. Regime change measures the amount of de-autocratization that takes
place in a given country. Subtracting the score from one year to the next
accomplishes this task. This variable measures the degree of transition in order
to see what impact the amount of change has on transition outcome.

Regime stability refers to the amount of time between measurable
changes in policy. Stability then, refers to how long a given regime, and that
regimes policy initiatives, have controlled a given country. Regime stability is an
important concept derived from the game theoretic model in chapter four. The

model demonstrates that in a complete and perfect information setting, a violent

103

outcome is not possible even when the regime chooses to open. This is due to
the fact that civil society is aware of how much change the regime will allow and
will not challenge the regime. In the incomplete and imperfect information game,
a violent outcome is possible. Why the difference? WlthOUt complete and perfect
information, civil society may see the regimes attempt to open, i.e. change its
regime score, as a sign that the regime is weakened and hence vulnerable.
Regime stability, the length of time between policy changes, then can be seen as
a signal to civil society as to how strong or weak the regime is.

Regime stability measures the number of years between transition attempt
by counting the years between attempts at transition. This measure is developed
using the Polity 3d data. The regime stability score was derived from the regime
change score. Each year that the regime change score remains the stable, adds

another year to the regime stability score.

Ethnic Fragmentation

In Chapter Four I argue that civil society can play a very important role in
the process of de-autocratization if it is able to unite and demand liberalization
from the regime. One factor that must be taken into consideration is how uniﬁed
is civil society? The more uniﬁed civil society is, the more pressure it can place
on the regime. The more fragmented civil society the less pressure it can bring to
bear. As such, the empirical model can ill afford to ignore the role of civil society

in the process of de—autocratization. Yet measuring civil society is a task that

104

poses many challenges and concerns, some of which cannot be addressed by
empirical measures that are available today.

Civil society is a difﬁcult concept to measure for many reasons. First, the
exact deﬁnition of civil society is elusive. Civil society can be thought of in many
ways; as ethnic groups, economic classes, religious differences, or any
combination of the above. To date no measure has been developed that
addresses all of these components or the cross cutting nature of civil society.
This makes measuring civil society challenging and means that any measure of
civil society will be imperfect.

Additionally the degree to which one segment of civil society is willing to
go in order to achieve its goals would be a more accurate measure of the depth
of fragmentation. The fact that different groups exist within civil society is not as
important has to how strongly those groups disagree and what actions they are
willing to take in order to affect change. For example, the willingness of some
white Americans to use force in order to halt civil rights changes in 2001 is very
different then the willingness of similar groups in 1951. That is not to say that
white supremacist groups in the United States do not exist in 2001 or that the
depth of their feelings towards minorities have changed, only that their
willingness to use violence to intimidate and halt the civil rights movement has
changed. While the most appropriate measure of how fragmented civil society is
should include a measure of willingness, this is not something easily
conceptualized or measured. Due to these difﬁculties the measure used in this

research does not include a measure of willingness.

105

For the sake of this research, I will use a measure of ethnic fragmentation
developed by Tanja Ellingsen in the 'Ethnic Wrtches' Brew Data Set. I choose a
measure that relies on ethnic fragmentation because such a measure
incorporates what appears to be the most divisive component of civil society,
nationalism.

Ellingsen develops a measure of ethnic fractionalization that is a count of
the number of ethnic groups found in a given country between the years 1945-
1994. She uses data found in the Handbook of the Nations 1979, 1983, 1986,
1993, Britannica Book of the Year 1969, 1974, 1979, 1986, 1987, 1989, 1992,
1993, and the Demographic Yearbook 1964, 1971, 1973, 1979,1983, 1988. In
order to insure that the various sources were compatible, and that the measure
was reliable and valid, Ellingsen checked the correlation of the different
measures against each other and found that the correlation was high allowing the
creation of one composite variable for each of the different measures in her data
set.12

As a test of the validity of my chosen measure of civil society, ethnic
fragmentation, I will use the variable number of linguistic groups and number of

religious groups also found in the Ellingsen data set.

Economic Indicators
Economic development is one factor of de-autocratization that is

discussed in chapters three and four. In chapter three, the game theoretic model

 

'2 Ellingsen does develop a series of decision rules for when the correlation appeared to be
weaker then she was comfortable with.

106

discusses the role of economics in inﬂuencing the regimes decision to open or
not. One example of the importance of the economic dimension can be found in
the Soviet Union in 1987 when Premier Gorbachev announces a series of
economic policy changes under the umbrella of perestroika. In chapter four, the
role of economic development is discussed as one variable that can motivate civil
society to mobilize and attempt to pressure the regime to open.

Economic data was collected from the World Bank data set. The economic
indicator chosen for this research is Gross Domestic Product Per Capita (1987).
GDP will measure changes in the economic performance for a given country.

This data is gathered annually and is available for most countries.

Coding Rules

While building the data set used for this dissertation, a multitude of
decisions had to be made on how to code the data. What follows below is a
discussion of the coding rules used to develop my data set.

When selecting the cases to be included in the data set, cases where the
regime score was greater then six were not included. Adding the autocracy
developed the regime score and democracy scores found in the Polity 3 data set.
Both scores shared a range of 0-10, with 0 being the lowest score and 10 being
the highest score. The Autocracy scores were translated into negative numbers
and then added to the democracy scores producing an index variable with a

range of -10 to 10.

107

There were cases in the Polity data set that did not have a regime score
for each year. If a case did not have regime data available and there was no
recorded conﬂict, the case was removed from the data set. If a case had no
regime data available and there was a record of conﬂict the case a dummy
variable was created to indicate that this was a new country. These cases are
Croatia, Bosnia, Georgia, Azerbijan, Eritrea, Suriname, and Tajikistan. In the
years that there was no regime data the cases were coded with a dummy
variable where 1 indicated that it was a new country. The year were regime data
became available the cases were coded as a 0, indicating that the regime data
was now available. In order to include the cases in the data, the regime scores
from the ﬁrst year that the case had data available was copied for the years were
no data was available.

Additionally, there were countries that had been involved in multiple
conﬂicts between 1989-1994. In these cases the regime data was copied for
each year that there contained multiple conﬂicts so to include all the conﬂicts.
Finally, Ellingsens' ethnic, religious, and linguistic variables all allow for the
possibility of fractions. This occurs because she creates an index variable for
each, which combines the data from her three sources which leads tot he
creation of fractions. In order to remove the fractions from the data set, the
variables were re-coded such that any case where a fraction occurred was
rounded down. I believe that whole numbers serve as a more accurate count of
the number of religious, ethnic, or linguistic based groups there are in a country.

By rounding down, the data has been altered in such a way that it is biased

108

against my position that the number of groups will increase the likelihood of

conﬂict.

Conclusion

This chapter has served to introduce the reader to the data sets utilized to
build the overall data that will be used to empirically test a series of hypotheses,
found in chapters three and four. The reader has been introduced to the
dependent and independent variables as well as how those variables are
operationalized. Chapter six will use the measures developed in this chapter to

empirically test the hypotheses derived in chapters three and four.

109

Chapter 6

EMPIRICAL FINDINGS

The focus of this research is when will internal conﬂict be the result of the
process of de-autocratization. The proceeding chapter established the universe
of cases and discussed the operationalization of variables that will be used in this
chapter. The game theoretic models in chapters three and four demonstrate the
importance of information and strategic interaction in determining whether or not
de-autocratization will occur peacefully, end with no change in a country’s regime
type, or end in violence. This chapter will develop a series of hypotheses, and
empirically test said hypotheses, that will help us to better understand what
factors will inﬂuence the type of violent outcome that countries under going the
process of de-autocratization may fall prey to.

It is important to understand that the empirical model developed in this
chapter are not explicit tests of the game theoretic models developed in chapters
three and four. The game theoretic models demonstrate the importance of
information as well as the wide variety of outcomes that are possible if de-
autocratization is attempted. It is argued that economic growth, ethnic
fragmentation, regime type, regime stability, and the degree of regime change
will inﬂuence the decisions made by the actors, the regime members as well as
civil society, and hence the outcome of the game. These variables are not

explicitly modeled; hence the game theoretic models cannot be explicitly tested.

110

The empirical model presented in this chapter is meant to more closely
exam the various types of internal conﬂict and which of the variables mentioned
above inﬂuence the likelihood of each type of conﬂict. Instead of creating a
dichotomous variable, no conﬂict or conﬂict, the outcome variable includes ﬁve
categories; no conﬂict, coup, political unrest, minor conﬂict, and civil war. This is
done in an effort to distinguish between the many types of internal conﬂict that a
country may face and with the understanding that there are unique conditions
that may cause a country to confront a wide variety of violent outcomes, each
with its own unique costs. As such, this chapter seeks to unpack the different
types of violent outcomes that may be associated with the process of de-

autocratization.

Hypotheses

The variables that are the focus of the empirical model presented in this
chapter are outcome, economic development, regime type, regime change,
regime stability, and ethnic fragmentation. The hypotheses derived below are
based on the literature found in the ﬁelds of comparative politics and international
relations. Each of the variables are discussed in chapters three and four as
being important in inﬂuencing the decisions made by regime members, the
regime as a whole, and civil society in the game theoretic models. In order to
examine the inﬂuence that each of these variables has on the likelihood of each
of the possible violent outcomes, it is necessary to develop hypotheses for each

of the variables.

111

Hypothesis 1: As economic development increases, the likelihood of
lntemal conﬂict decreases

Hypothesis 2: The more democratic a regime is, the less likely it is to face
lntemal conﬂict.

Hypothesis 3: The more quickly regime change occurs the less likely there
will be internal conﬂict.

Hypothesis 4: The more stable a regime is the less likely there will be
internal conﬂict.

Hypothesis 5: The greater the degree of ethnic fragmentation, the greater

the chance of internal conﬂict.

The Empirical Model

The multi-nomial logit allows the development of a dependent variable
with multiple categories, such as the dependent variable developed in this
dissertation. As such, it is the most appropriate technique for testing the above

hypotheses

Data Analysis

The universe of cases was all countries coded as an authoritarian regime
in the Polity3d data and where the regime score indicates a decrease in the
amount of authoritarian control. The results indicate that not all lntemal conﬂicts

are the same; different variables will inﬂuence what outcome is reached

112

dependent on the internal conditions of a country undergoing the process of de-
autocratization.

In chapter three, it is argued that a coup is possible only when one of the
factions within the regime believes that it can win if it attempts a coup and if the
offer made by the opposing faction is too far away from its ideal point. While the
empirical results discussed here are not an empirical test of the game theoretic
model, they do indicate that the variables that inﬂuence the likelihood of a coup
occurring are related to regime stability and the degree of regime change. The
two variables that are statistically signiﬁcant when discussing the coup outcome
are change and stability. A coup is more likely to occur if the regime attempts
change at a more leisurely rate then if the regime attempts change more quickly.
Additionally, the more stable a regime is the less likely a coup will occur.

Initially, the change results appear to be troubling. One would expect that
the greater the amount of change that is attempted, the higher the probability of a
coup. Why? The Hard-line faction is not interested in relinquishing a signiﬁcant
amount of power to the Liberalizers or civil society. The game theoretic model in
chapter three argues the opposite. If the Liberalizers believe that they are strong
enough to win if they attempt a coup, they will make the Hard-liners an offer that
demonstrates that they believe that they will win. Such an offer would be close to
the Liberalizers ideal point, meaning that the degree of change would most likely
be large. As such, the ﬁndings of the empirical model match the expected results
from the game theoretic model. As the degree of change increases, the

likelihood of a coup decreases.

113

The stability variable is signiﬁcant in the hypothesized direction for the
case of a coup. The longer the regime maintains control over the country without
attempting change, the less likely the regime will experience a coup.

Chapter four explores the interaction between regime members and civil
society, emphasizing the importance of information. It is argued that economic
development, or the lack there of, can cause civil society to demand that the
regime makes some type of change. The only outcome that the GDP variable is
statistically signiﬁcant is that of political unrest. Political unrest is deﬁned as
conﬂict with fewer then 1,000 deaths and that lasts for fewer then 7 days. For
example, one time riots and violent protests are categorized as political unrest. It
is not surprising that GDP is signiﬁcant for this outcome. Economic shortfalls,
which would lead to an increased likelihood of violence, will place a strain on the
populations' ability to survive, angering the population.

The relationship between economic development and political unrest is
one that has been discussed through out the comparative politics literature.
Scott and Kerkliet argue that changes in economic policies will lead to what he
refers to as everyday forms of resistance in areas where the population is
struggling to survive. In countries were the majority of the population is
struggling to survive, any economic change that makes survival more difﬁcult is
going to be met with resistance.

The third category is that of minor conﬂict. Minor conﬂict was deﬁned as
conﬂict with fewer then 1,000 deaths but where the ﬁghting was ongoing through

out the year. In this case, the variables that are statistically signiﬁcant are regime

114

Political
Unrest

 

group

Table 6.1. Multi Nomial Logit Results for all cases were De—autocratization is
attempted

115

type and religion. The more autocratic a regime, the greater the probability the
country is to face conﬂict throughout the year.

Religion is signiﬁcant, although not in the hypothesized direction. The
results indicate that as the number of religious groups found in a country
increases there is a decline in the probability of minor conﬂict. Of all the
measures provided in this research, the measures of ethnic fragmentation,
including religion, are the most problematic. Religion is measured as the number
of religious groups in a given country. This does not take into considerations
important factors such as the presence of cross cutting cleavages within a
country.

Finally, regime stability was the only variable statistically signiﬁcant for the
civil war outcome. These were cases where there was greater then 1,000 dead
and the ﬁghting occurred throughout the year. In this case, the more stable the

regime, the less likely there was to be a civil war.

Conclusion

What do we learn from the above results? It is clear that different variables
inﬂuence what type of internal conﬂict a country may face when under going the
process of de—autocratization. If we know that a country is struggling with
economic development then there is an increased probability that said country
will ﬁnd itself having to deal with political unrest while a country that has seen
many regime changes in the last few years is more likely to ﬁnd itself involved in

a civil war or facing a coup.

116

What Have We Learned?

This research began with a deceptively simple question: when will de-
autocratization lead to violence? VVItl'l the focus of numerous scholars and policy
makers on the study and promotion of democratization throughout the world, it
was clear that this question must be addressed. Is it possible that
democratization itself leads to violence? If so, is it not possible then that the
promotion of democratization is something that policy makers should examine
more carefully? It was clear that this question must be addressed.

Traditional research approached the subject of democratization with
several assumptions. First, it was assumed that the goal of any regime transition
attempted by an authoritarian regime was to develop a consolidated democratic
regime. The very title “democratization” points to this assumption. Yet it has not
been clearly demonstrated that this is the case. The number of “stalled”
democratic transitions should clearly demonstrate that a democratic regime is not
necessarily the goal of the regime. Perhaps those countries are not stalled at all
but instead the regime was content with the degree of liberalization and halted
the process.

Second, the process of democratization was treated as a linear, path-
dependent process dictated by the whims of the regime. Many of the scholars
whose work is discussed earlier in this document, state that democratization
cannot begin unless the regime chooses to open.

The research presented in this paper approached the question only after

disregarding these two assumptions. The term “de—autocratization” was adopted

117

in order to try and avoid the problems associated with the ﬁrst assumption. De-
autocratization includes all countries making some degree of transition away
from a pure authoritarian regime, no matter how small the step may be.
Additionally, the term de-autocratization removes the implicit dichotomization of
the ﬁnal outcome. No longer is the ﬁnal outcome successful or unsuccessful
democratization. Instead the ﬁnal outcomes include coups, political unrest, and
civil war as well as democratization and limited liberalization.

The models developed in this research are a ﬁrst step towards exploring
de—autocratization as a non—linear process. The regime may return to the
bargaining table on any number of occasions through out the course of the
process in order to decide what should occur next. If the regime chooses to open
and does so successfully, the regime will have to meet again to decide if it should
allow further opening or not. Additionally, by moving from one model of de-
autocratization, such as Casper and Taylor developed, and acknowledging the
presence of multiple models in the overall process, the ﬁrst step has been taken
to understanding to non-linear nature of the process. This aspect of the process
has not been fully developed in this research, it is a research project for the
future.

The non-cooperative bargaining game, found in Chapter Three, explores
how regime members interact with each other and how this interaction can lead
to coups. The model demonstrates that the greater the demand made, the less
likely a coup will occur. This is due to the belief that the faction making the

demand has the ability to force the opposing faction from the regime. Rather

118

then being forced from the regime, the opposing faction chooses to accept the
offer in order to maintain the beneﬁts of ofﬁce. This interaction demonstrates the
importance of information. If either faction is incorrect in assessing its strength it
can lead to a coup or to acquiescing to an offer when it was not necessary.

The incomplete and imperfect information game in Chapter Four
reinforces the importance of information. Civil society will not choose to mobilize
if it knows that the regime is comprised predominantly of Hard-Liners because
civil society knows that Hard-Liners will not allow additional liberalization. As
such, internal violence is avoided. If civil society believes that it is facing a regime
lead by Reformers, when in reality the regime is lead by Hard-Liners, conﬂict will
occur.

In addition, the model developed in Chapter Four allows civil society to
mobilize even if the regime chooses not to open. This is an important addition to
the research in this ﬁeld as it more accurately reﬂects the real world.
Czechoslovakia, Poland, and South Africa are but a few examples of countries
where civil society mobilized and demanded liberalization when the regime was
not [willing to liberalize. In all three cases, members of civil society were willing to
face imprisonment, violence, and even death in order to try and force a change in
the type of regime. The model in Chapter Four allows for this possibility where
past models did not.

Finally, the empirical model, found in this chapter, clearly demonstrates
that different independent variables will inﬂuence not only the likelihood of

violence but the type of violence we should expect to see. This is an important

119

ﬁnding as it reminds us that the factors that we believe necessary for successful
democratization, such as economic development, can also lead to violence.

The research developed in the last 120 pages has done a great deal to advance
our knowledge of how the process of de-autocratization works, how violent
outcomes can come from the process, and the role of information in the process.
This is not the end of the research. Future research will focus attention on civil
society and how civil society overcomes the collective action problem in order to
mobilize as well as explore the role of multiple iterations of the games developed
here in order to better understand how the long term process may lead to

conﬂict.

120

BIBLIOGRAPHY

121

ll

 

Bibliography

Bates, Robert H., and Barry R. Weingast. 1997. “Rationality and Interpretation:
The Politics of Transition.” , Hoover Institution, Stanford University.

Binmore, Ken. Avner Shaked, and John Sutton. 1985. "Testing Noncooperative
Bargaining Theory: A Preliminary Study." The American Economic Review 75:
1 178-1 180.

Binmore, Ken, Avner Shaked, and John Sutton. 1989. "An Outside Option
Experiment." Quarterly Journal of Economics 1989: 753-770.

Bratton, Michael, and Nicholas van de Walle. 1996. Democratic Experiments in
Africa: Regime Transitions in Comparative Perspective. Cambridge: Cambridge
University Press.

Bremeo, Nancy. 1992. "Democracy and the Lessons of Dictatorship."
Comparative Politics. April, 1992: 273-289.

Casper, Gretchen, and Michelle M. Taylor. 1996. Negotiating Democracy:
Transitions from Authoritarian Rule. Pittsburgh: University of Pittsburgh Press.

Collier, Paul and Anke Hoefﬂer. 2000. "On Incidence of Civil War in Africa."
World Bank. (http:/lwww.worldbank.org/research/conﬂict/papersﬁncidence.htm)

Collier, Paul and Anke Hoefﬂer. 2001. "Greed and Grievence in Civil War." World
Bank.

(http:l/www.worldbank.org/research/conﬂictlpaperslgreedandgrievance.htm)

Crescenzi, Mark JC. 1998. "Uncertainty in Transitions from Authoritarian
Regimes: Explaining Violence and Mass Opportunities." University of Illinois at
Urbana Champaign and Merriam Laboratory for Analytic Political Research.
Work in Progress. '

Dahl, Robert A. 1989. Democracy And Its Critics. New Haven: Yale University
Press.

Davies, James C. 1962. “Towards a Theory of Revolution.” American
Sociological Review 39: 5-19.

de Figuiredo, Rui JP, and Barry R. Weingast. 1997. “The Rationality of Fear:

Political Opportunism and Ethnic Conﬂict.” , Hoover Institution, Stanford
University.

122

Fearon, James D., and David D. Laitin. 1996. "Explaining Interethnic
Cooperation." The American Political Science Review 1996: 715-735.

Gates, Scott, and Brian D. Humes. 1997. Games, lnfonnation and Politics:
Applying Game Theoretic Models to Political Science. Ann Arbor: University of
Michigan Press.

Gates, Scott, Havard Hegre, Mark P. Jones, and Havard Strand. 2000.
"Institutional Inconsistencey and Political Instability: Persistence and Change in
Political SystemsRevisted, 1800-1998." Presented at the Anuual Metting of the
American Political Science Association, Washington DC.

Geddes, Barbara. 1991. “A Game Theoretic Model of Reform in Latin American
Democracies.” American Political Science Review 85: 371 -392.

Gleditsch, Nils Petter, and Havard Hegre. 1997. “Peace and Democracy: Three
Levels of Analysis.” Journal of Conﬂict Resolution 41: 283-310.

Gurr, Ted Robert. 1970. Why Men Rebel. Princeton, NJ: Princeton University
Press.

Gurr, Ted Robert. 1993. “Why Minorities Rebel: A Global Analysis of Communal
Mobilization and Conﬂict Since 1945.” lntemational Political Science Review 14:
161-201.

Gurr, Ted Robert. 1994. Minorities at Risk. Washington, DC: United States
Institute of Peace.

Gurr, Ted Robert, and Roger Duval. 1973. “Civil Conﬂict in the 1960s: A
Reciprocal Theoretical System with Parameter Estimates.” Comparative Political
Studies 6: 135-170.

Gurr, Ted Robert, and Mark Irving Lichbach. 1981. “The Conﬂict Process: A
Formal Model.” Journal of Conﬂict Resolution 25: 3-29.

Huntington, Samuel P. 1991. The Third Wave: Democratization in the Late
Twentieth Century. Norman, OK & London: University of Oklahoma Press.

Jackman, Robert W., Rosemary H.T. O'Kane, Thomas H. Johnson, Pat
McGowan, and Robert O. Slater. 1986. The American Political Science Review
80: 225-250.

Johnson, Thomas H., Robert O. Slater, Pat McGowan. 1984. American Political
Science Review 78: 622-640.

123

II

 

Krain, Matthew. 1998. "Responding to Revolution: Repression, Accomodation,
and Opposition in Post-lntemal War States." Presented at the Annual Meeting of
the Midwest Political Science Association, Chicago, Il.

Krain, Matthew, and Marissa Edson Myers. 1997. “Democracy and Civil War: A
Note on the Democratic Peace Proposition.” lntemational Interactions 23: -.

Licklider, Roy. 1995. "The Consequences of Negotiated Settlements in Civil
Wars, 1945-1993." The American Political Science Review 1995: 681-690.

Londergan, John B. and Keith T. Poole. 1990. "Poverty, The Coup Trap, and
the Seizure of Executive Power. Worid Politics 42: 151-183.

Ma, Ching-to Albert, and Michael Manove. 1993. "Bargaining with Deadlines
and Imperfect player Control." Econometrica 61: 1313-1339.

McLaughlin, Sara, Scott Gates, Havard Hegre, Ranveig Gissinger, and Nils
Petter Gleditsch. 1998. "Timing the Changes in Political Structures: A New Polity
Database." Journal of Conﬂict Resolution 42: 231 -242.

Migdal, Joel. 1974. Peasants, Politics, and Revolution: Pressures Toward
Political and Social Change in the Third World. Princeton University Press.

Moore, Will H. 1998. “Repression and Dissent: Substitution, Context and Timing.”
American Joumal of Political Science 42: 851 -873.

Mueller, John, and Erich Weede. 1990. “Cross-National Variations in Political
Violence: A Rational Action Approach.” Joumal of Conﬂict Resolution 34: 624-
651.

Muller, Edward N. 1985. “Income Inequality, Regime Repressiveness, and
Political Violence.” American Sociological Review 50: 47-61.

Muller, Edward N., and Karl-Dieter Opp. 1986. “Rational Choice and Rebellious
Collective Action.” American Political Science Review 80: 471-487.

Muller, Edward N., and Mitchell A. Seligson. 1987. “Inequality and Insurgency.”
American Political Science Review 81: 425-451.

O'Donnell, Guillermo, and Phillipe C. Schmitter. 1986. Transitions from
Authoritarian Rule: Tentative Conclusions About Uncertain Democracy.
Baltimore: Johns Hopkins University Press.

O'Donnell, Guillermo , Philippe C. Schmitter, and Laurence Whitehead. 1986.

Transitions from Authoritarian Rule. Prospects for Democracy. Baltimore, MD:
Johns Hopkins University Press.

124

Posusney, Marsha Pripstein. 1993. "Irrational Workers: The Moral Economy of
Labour Protest in Egypt." World Politics 46: 83-120.

Przeworski, Adam. 1991. Democracy and the Market: Political and Economic
Reforms in Eastern Europe and Latin America. Cambridge: Cambridge University
Press.

Rule, James B. 1988. Theories of Civil Violence. Berkeley, CA: University of
California Press.

Rustow, Dunkart A. 1970. "Transitions to Democracy: Toward a Dynamic
Model." Comparative Politcs April 1970: 337-363.

Scaramozzino, Pasquale. 1991. "Bargaining with Outside Options: Wage and
Employment in UK Manufacturing 1974-82." The Economic Joumal 101: 331-
342.

Scott, James C.. 1977. "Moral Economy of the Peasant: Rebellion and
Subsistence in Southeast Asia." Yale University Press, New Haven, Ct.

Sutton, John. 1986. "Non-Cooperative Bargaining Theory: An Introduction."
The Review of Economic Studies 53: 709-724.

Sweeny, Kevin J. 2000. "An Investigation of the Causes of Civil War. 1960-
1994." Presented at the 2000 Annual Meeting of the American Political Science
Association, Washington DC.

Taylor, Michael. 1988. “Rationality and Revolution.” , Cambridge University
Press, Reading.

Tilly, Charles. 1978. From Mobilization to Revolution. Reading: Addison-Wesley.
Wayman, Frank Whelon, David J. Singer, and Meredith Sarkees. 1996. “Inter-
State, lntra-State, and Extra-Systemic Wars 1816-1995.” Presented at the
Annual Meeting of the lntemational Studies Association, San Diego, CA.
Weingast, Barry R. 1997. “The Political Foundations of Democracy and the Rule
of Law.” American Political Science Review 91: 245-263.

125