Dilemmas of Democratic Consolidation: A Game-Theory Approach 9781626373341

Citation preview

DILEMMAS OF DEMOCRATIC CONSOLIDATION

DILEMMAS OF DEMOCRATIC CONSOLIDATION A Game-Theory Approach

Jay Ulfelder

Published in the United States of America in 2010 by FirstForumPress A division of Lynne Rienner Publishers, Inc. 1800 30th Street, Boulder, Colorado 80301 www.firstforumpress.com and in the United Kingdom by FirstForumPress A division of Lynne Rienner Publishers, Inc. 3 Henrietta Street, Covent Garden, London WC2E 8LU © 2010 by FirstForumPress. All rights reserved Library of Congress Cataloging-in-Publication Data A Cataloging-in-Publication record for this book is available from the Library of Congress. ISBN: 978-1-935049-18-0 British Cataloguing in Publication Data A Cataloguing in Publication record for this book is available from the British Library. This book was produced from digital files prepared by the author using the FirstForumComposer. Printed and bound in the United States of America The paper used in this publication meets the requirements of the American National Standard for Permanence of Paper for Printed Library Materials Z39.48-1992. 5 4 3 2 1

Contents

List of Figures Preface

vii ix

1

Introduction

2

The Democracy Game

25

3

How Fragile Is Democracy? Patterns in Cross-National Data

55

How Does Democracy Break Down? Evidence from Random Narratives

73

4

1

5

Exploring Confounding Cases

113

6

Implications for Democracy Promotion

125

7

Conclusion

151

Appendix: Episodes of Democracy, 1955–2007

159

References Index

165 173

v

Figures

2.1 The Democracy Game

26

3.1 Global Prevalence of Democracy, 1955–2007

59

3.2 Democratic Breakdowns by Type, 1955–2007

60

3.3 Annual Counts of Breakdowns by Type, 1955–2007

61

3.4 Survival Rates of Democracies Worldwide, 1955–2007

62

3.5 Survival Rates of Democracies by Region, 1955–2007

63

3.6 Survival of Democracies by Income at Founding Elections

64

3.7 Survival of Democracies During and After the Cold War

68

3.8 Annual Share of Elections with International Observer Missions, 1960–2005

69

vii

Preface

This book offers a theory of politics in modern democracies that is meant to help explain why so many attempts at this form of government end badly. Democracy is arguably humankind’s greatest political invention, but it is also a construct that has proven difficult to sustain in practice. The theory I develop in the pages that follow portrays the usurpations of power that terminate most attempts at democracy— military coups, executive coups, and rebellions—as tragedies in the classical sense. The institutions of democratic government help put into practice at the scale of whole societies two of the noblest human impulses: empathy and a keen sense of justice. We construct democracy because we aspire to make life richer and fairer, not only for ourselves but also for our families and neighbors. Once we have established democratic government, however, the venality and fear that course alongside those nobler impulses can impel us to subvert it, sometimes even in cases where everyone involved would rather see democracy survive. In a sense, those baser instincts become the tragic flaw that carries us back to the harsher forms of rule that democracy was supposed to replace. This book grew out of work I did in the 2000s as research director for the Political Instability Task Force (PITF), a social-science research program funded by the US government’s Central Intelligence Agency (CIA) and managed under contract with my employer, Science Applications International Corporation (SAIC). All statements of fact, opinion, or analysis expressed are those of the author and do not reflect the official positions or views of the CIA or any other US government agency. Nothing in the contents should be construed as asserting or implying US government authentication of information or Agency endorsement of the author’s views. This material has been reviewed by the CIA to prevent the disclosure of classified information. Part of my work for the Task Force has involved developing statistical models to help forecast and explain transitions to and from democracy. Not long after I had developed a pair of models that seemed to fit the data pretty well, Thailand was struck by a military coup. According to my statistical model and the various theories of democratic consolidation on which it drew, the Thai coup of 2006 should not have happened.

ix

x

Dilemmas of Democratic Consolidation

The statistical model’s failure to anticipate this important event led me to take a closer look at the politics that preceded it. Looking back at the press coverage prior to the military coup, it sometimes seemed like everyone except theorists of democratic consolidation (and me with my statistical model) had seen that coup coming. The game-theoretic model developed and applied in the pages that follow is my attempt to construct a theory that could account for the strategic concerns those contemporary observers were seeing without ignoring the broader regularities identified in statistical research. Many people played important roles in producing this book. I would like to start by thanking the people at the CIA with whom I have interacted in my role as PITF research director for offering constructive comments on this particular project and for seeing value in its eventual publication. I am also grateful to Irv Jacobs and Andrew Stifel of SAIC, my program managers during the period I was working on this book, and to Jessica Gribble at Lynne Rienner Publishers, for giving me the opportunity to publish. Earlier versions of some portions of this book were presented at quarterly meetings of the Political Instability Task Force, at the National Meeting of the Midwest Political Science Association in April 2008, and at the Annual Meeting of the American Political Science Association in August 2008. For thoughtful comments along the way, I would like to thank Robert Bates, Emily Beaulieu, Michael Coppedge, Michael Cutrone, Jack Goldstone, Christian Houle, Bruce Kay, David Laitin, Mike Lustik, Monty Marshall, Brian McCotter, Mark Mullen, Pat Regan, Milan Svolik, and three anonymous reviewers. I am also very grateful to my mother, Linda Potter, for her help in editing the penultimate version. Any errors that remain are, of course, my own. Last and most of all, I would like to thank my wife, Francesca, and my sons, Parker and August, for their encouragement.

1 Introduction

“The Silk Revolution,” one observer dubbed it. In September 2006, after months of political turmoil that saw citizens repeatedly throng the streets of Bangkok for dueling protests, Thailand’s military leaders carried off a bloodless coup d’etat against the democratically elected government of Prime Minister Thaksin Shinawatra. While the prime minister attended a meeting of the UN General Assembly in New York, top officers declared martial law, announcing on television that they had suspended the constitution and dismissed the government, both houses of parliament, and the Constitutional Court. The officers said they were temporarily replacing those bodies with a Council of Administrative Reform on behalf of the country’s king, and military spokesman Gen. Prapas apologized to the Thai people for any “inconvenience.”1 The coup punctuated a months-long political crisis during which a snap election was boycotted by the leading opposition party and then voided by the Constitutional Court and Prime Minister Thaksin had taken an unusual seven-week “break” from politics. Democracy had been tried in Thailand before, and each previous episode has also ended with a military coup, so from a local perspective the turn of events in 2006 was not entirely exceptional. From a global perspective, however, the coup against the Thaksin government was noteworthy for several reasons. First, the 2006 coup terminated the longest episode of democratic government in Thailand’s history—fifteen years, an age at which some theories of democratization assert that democratic norms and habits ought to have taken hold in a way that would prevent usurpations of power. Second, the coup came after the country had experienced a peaceful transfer of power from one party to another via fair elections, another widely used marker of democratic consolidation. Third, the coup made Thailand one of the richest countries ever to suffer a breakdown of democracy in a world where economic development and democracy are often assumed to go hand in hand.2 Finally, the coup came after several years of solid economic

1

2

Dilemmas of Democratic Consolidation

growth under Thaskin’s leadership, a performance that some theories would identify as a crucial source of legitimacy and thus inoculation against coups. In short, the usurpation of power that occurred in 2006 contradicted much of the conventional wisdom about the structural conditions under which democracy should survive or fail. Although rumors of an impending coup had circulated for weeks before the event finally happened and many citizens apparently welcomed the military’s attempt to break the stalemate between rival political camps, prevailing theories of democratic consolidation would not have led us to anticipate this outcome.3 In December 2007, little more than a year after Thailand’s coup, democracy also came undone in Kenya. In contrast to Thailand, however, the Kenyan military did not play a direct role in this seizure of power. Instead, soldiers only deployed to try to restore order after a flawed election sparked widespread violence between political rivals. In the Kenyan case, it was the elected government that dismantled democracy, apparently by rigging the vote to prevent opposition candidate Raila Odinga from unseating incumbent President Mwai Kibaki. The bloodshed triggered by that fraud claimed hundreds of lives and shut down economic activity in many parts of the country for days, but that violence was a symptom, not a cause, of democracy’s destruction. Kibaki’s controversial re-election and the violence that ensued occurred in a country generally regarded as one of sub-Saharan Africa’s most stable and most prosperous. Although democracy was still shortlived in Kenya at the time—Kibaki’s election in 2000 had marked the end of a period of authoritarian rule that began in the 1960s and bent but did not break in the 1990s—the country’s vigorous press, active and growing civil society, and vibrant economy were supposed to have guarded it against the troubles that have characterized many of the attempts at democratic government in African countries since independence. In simple terms, Kenya was thought by some prominent observers to have “too much to lose” to succumb to the temptations of power politics. These breakdowns of democracy in Thailand and Kenya were notable because they contradicted conventional ideas about the circumstances under which democratic regimes become consolidated. 4 Unfortunately, those breakdowns are not isolated incidents. One piece of good news from global politics in the early part of the twenty-first century is that democratic regimes have become more prevalent than ever. The less happy corollary to that pattern is that many attempts at democracy continue to fail, even under conditions traditionally

Introduction

3

considered auspicious for the establishment of elected government. In fact, most attempts at democratic government in the past half-century have ended with a return to authoritarian rule, often not so many years after their start.5 These failures matter tremendously in their own right, because they sharply diminish the political rights of the citizens who suffer them. In a highly interconnected world, those failures can also have deleterious consequences for governments and citizens elsewhere, a fact that has led many of the world’s most powerful countries to make the promotion of democratic government a pillar of their foreign policies. Why do so many attempts at democracy fail? To provide novel answers to that question, this book pays less attention to the structural conditions that dominate prior theorizing on this subject and concentrates instead on the process of democratic breakdown. In the pages that follow, I develop and apply a game theoretic model of democratic politics to explore how and why the institutions of elected government might survive or fail under a variety of conditions. This model allows us to connect different forms of breakdown to specific actors and the strategic incentives they confront under different circumstances. Conventional accounts of democratization claim that democracy is likely to fail when structural conditions do not favor its survival. Although these arguments generally accord with robust patterns in cross-national data, they fail to explain exceptional cases and do not provide clear insights into the timing and mode of democracy’s failure. Some scholars have previously used game theory to explore the strategic aspects of democratic breakdown, but those works have generally failed to consider the breadth of the ways in which democracy can break down and the full variety of incentives at work. Empirical observation of democratic breakdowns in the latter half of the twentieth century shows that these events comes in three basic forms: 1) the elected government rigs the electoral process in its favor or dismantles that process entirely; 2) the military steals power from that elected government; or 3) a popular rebellion topples the elected government. The game-theoretic model developed in this book connects these events to the incentives confronted by the organizations that perpetrate them. As is widely recognized, partisan rivals and military leaders may be tempted to try a coup or rebellion by the desire to control the spoils that come with state power, strategic considerations that might lead political actors in democracies to try to seize power illegally. What most theories of democratic consolidation fail to recognize, however, is that uncertainty about rivals’ interests, capabilities, and intentions may also compel those organizations to attempt a coup or rebellion by

4

Dilemmas of Democratic Consolidation

amplifying their fears of the exclusion and exploitation they could suffer if their political rivals take power—perhaps through the ballot, but also possibly through a coup of their own. Because of this strategic uncertainty, coups and rebellions can occur even in situations where all of the relevant actors most prefer that democracy survive. Taken together, these temptations and fears means that political parties and militaries in democratic regimes often have substantial incentives to try to usurp power. Those strong incentives, in turn, help to explain why democracy so often fails, even in conditions that structural theories might regard as auspicious. Defining Democracy

While democracy is surely one of the most familiar concepts in political science, scholars often disagree on its meaning. Most contemporary researchers follow Schumpeter (1945) by defining democracy with reference to processes rather than outcomes, but not all do, and even the procedural definitions used by many scholars vary significantly in their content.6 The persistence of a debate over definitions of democracy does not mean that there are no points of agreement. In 1830—thirty-three years before Abraham Lincoln would echo the line in his Gettysburg Address—Daniel Webster delivered a speech in which he spoke of a “people’s government, made for the people, made by the people, and answerable to the people” (Lijphart 1999: 1). This notion of a government of, by, and for the people it governs—of popular sovereignty—has emerged in many societies over the course of human history (Dahl 1998: 7–25). The notion of popular sovereignty also underpins most contemporary efforts to identify and measure democracy. In this book, I rely on a definition of democracy that leans heavily on the work of Robert Dahl (1971, 1998), with additions or qualifications proposed by other scholars who have also struggled to measure democracy in real countries over time for purposes of comparative analysis. Specifically, I define democracy as a form of government in which citizens freely and fairly choose and routinely hold accountable their rulers. In practice, this form of government occurs when four conditions hold. 1. Elected officials rule (representation). Representatives chosen by citizens make policy by law and in fact, and unelected entities

Introduction

5

cannot unilaterally block those representatives’ collective decisions or impose policy changes of their own. 2. Elections are fair and competitive (contestation). The process by which citizens select their rulers provides voters with distinct choices and is generally free from deliberate fraud or abuse. 3. Basic civil liberties are respected (freedom). Freedoms of speech, association, and assembly routinely afford citizens opportunities to deliberate on their interests, to organize in pursuit of those interests, and to monitor the performance of their elected representatives and the agencies and organizations on which those officials depend. 4. Politics is inclusive (inclusion). Adult citizens have equal rights to vote and participate in government, and they enjoy fair opportunities to exercise those rights. The fundamental procedural element of democracy defined this way is fair, competitive, and multiparty elections in which virtually all adult citizens may participate and vote for ruling officials. As to which kinds of officials to consider, I focus on the individuals who actually perform the state’s legislative and executive functions at the national level: the head of government and the members of the legislature. For elections to these offices to fulfill the promise of participation and representation, however, democracies must routinely secure certain civil liberties, including freedoms of speech, association, and assembly. These freedoms are essential to citizens’ opportunities to deliberate on their own interests, to make informed choices about candidates for office, to appeal for votes when running for office, to advocate for or against specific policies, and to monitor their government’s actions. For the principles of representation and accountability to be realized, democracies must also protect the policy-making process from interference by unelected entities (Dahl 1971; Karl 1995; Linz 1978). Drawing on Tsebelis’ (2002) notion of a veto player, I consider unelected entities to wield undue power when they can unilaterally block or produce change in major issue areas such as national security, taxation, or property rights. The unelected entities in question can be any one of a variety of individuals or organizations, including monarchs, military leaders, religious or tribal elders, or even foreign governments. In bicameral systems, the upper house of the legislature sometimes holds veto power; these upper houses are considered unelected when a

6

Dilemmas of Democratic Consolidation

substantial proportion of their members either inherit their seats or are chosen by unelected bodies or officials.7 Defined this way, democracy breaks down when routine opportunities for representation or accountability are substantially diminished or eliminated. This can happen abruptly, as it did in both Thailand and Fiji in 2006, when an unelected individual or group announces that they have replaced elected officials at the head of government and those elected officials actually step aside or are exiled, jailed, or killed. Usurpations of power by unelected officials are usually plain to see—military officers announce their coup on television or radio, antigovernment protesters swarm the legislature, and so on. Democratic accountability can also erode more gradually, as the institutions and practices required to produce transparency, access, and competitiveness are incrementally subverted or dismantled, as they were in the 2000s in Russia and Venezuela. These “creeping coups” are more difficult to observe than the abrupt ones, partly because the incumbents undertaking them have motive and opportunity to conceal their actions. In fact, the repertoire of techniques for subverting democracy may be bounded only by the limits of incumbents’ creativity. As international election-monitoring has become more common and more sophisticated since the end of the Cold War (Hyde 2007), officials attempting this kind of subversion seem increasingly to be focusing their efforts on elements of the process that are further and further removed from the balloting itself, including obstructing challengers’ candidacies, manipulations of the media environment, tampering with voterregistration procedures, clever redistricting, and even the adoption of rules or movement of troops to distribute votes from rank-and-file soldiers in more favorable ways (Bratton 1998). Prior Theory and Research

Many of the democracies in existence as I write this book have not experienced attempts to usurp power from their elected governments for decades. During that same time, however, many other democratic regimes have come and gone, sometimes more than once in the same country. What explains this variation in outcomes? Why does democracy survive in some cases and break down in others? And, in the cases where democracy fails, what explains the form and timing of those events? In other words, why do those breakdowns happen when and how they do? Theories of comparative democratization that might shed light on these questions fall into three broad groups. One set of explanations

Introduction

7

emphasizes the influence of structural conditions—the political, economic, and social environment in which governments are situated. A second set emphasizes process—the dynamic ways in which democracy arises and sometimes fails. A third set uses game-theoretic models to explain the survival or failure of democracy as a consequence of strategic interactions among rational actors. The weakness of structural theories as explanations for democratic breakdown is that they tell us little about the timing and course of that process. In other words, even if they offer a compelling general story about why democracies die, they usually say little about who kills them and even less about how and when they go. Meanwhile, theories of democratization emphasizing process deal explicitly with questions of who, how, and when, but they often do so descriptively in ways that do not readily accumulate into more general explanations and predictions. The best rational-choice theories combine aspects of structure and process in formal models that help to clarify existing issues and, sometimes, to illuminate new ones. Still, none of the existing rationalchoice theories of democratic breakdown has managed to capture all of the crucial actors and pathways and then link those elements to observed structural patterns. Structural Theories

Structural theories of democratization identify elements of context that are thought to shape the chances for democracy’s emergence and survival. For several decades, modernization theory has dominated discourse among political scientists in the United States about the process of democratization. Modernization theory understands democracy in teleological terms, as an outgrowth of certain socialstructural changes that occur as societies experience certain kinds of economic growth and development. According to this view, economic development in the industrial age transforms societies through mutually reinforcing processes of urbanization, education, increased communication, and the accumulation of wealth by ordinary citizens. These processes change the way that society is organized and give rise to new values and interests that are conducive to democracy. Modernization theory is perhaps most strongly identified with the work of Seymour Martin Lipset (1959), who saw the emergence of an educated middle class with “moderate” values as the crucial sociological foundation for democratic stability.8 In a recent restatement of modernization theory, Ron Inglehart and Christian Welzel (2005: 134) locate the causal mechanism of this process in the emergence of what

8

Dilemmas of Democratic Consolidation

they call self-expression values, which are said to arise in conjunction with the increases in education, knowledge-intensive work, and social complexity that generally accompany economic development: Modernization tends to bring both cognitive mobilization and growing emphasis on self-expression values. This is turn motivates ever more people to demand democratic institutions and enables them to be effective in doing so as elites watch the costs of repression mount. Finally, with intergenerational replacement, the elites themselves may become less authoritarian and repressive if their younger cohorts are raised in societies that value self-expression. Social change is not deterministic, but modernization increases the probability that democratic institutions will emerge.

Occasionally, societies might sputter along under authoritarian rule in spite of modernization or might attempt democracy when these structural prerequisites are still lacking, but the basic trajectory is thought to be universal. Consequently, these aberrations are expected eventually to correct themselves, either by reverting to autocracy or restoring democracy. While modernization theory has dominated the field of comparative democratization, scholars have tabled many other important ideas about the influence of structural forces on the prospects for democracy. Some scholars argue that enduring aspects of a society’s “political culture”— “the beliefs and values concerning politics that prevail within both the elite and the mass” (Diamond, Linz, and Lipset 1990: 16)—have lasting effects on the prospects for democracy’s success. Juan Linz and Alfred Stepan (1996) draw attention to the effects of prior regime type, positing a variety of ways in which structural features of the authoritarian regime that immediately precedes democratization can shape the processes of democratic transition and the prospects for consolidation. Other scholars have made claims about the impact of a democracy’s institutional design on the odds that it will survive. The most prominent of these arguments comes from Juan Linz (1978, 1990a, 1990b), who initiated what might be regarded as a field within the field of comparative democratization with his assertion that presidential systems are more susceptible to deadlock and therefore breakdown than parliamentary ones.9 Still other researchers have focused on the relationships between electoral systems and the dynamics of politics in democracies, with the conventional wisdom claiming that proportional representation (PR) is more conducive to democratic survival than majoritarian systems because of PR’s greater inclusivity.10

Introduction

9

In spite of this variety of ideas about relationships between structural forces and democratization, modernization theory remains the starting point for most contemporary explanations of democracy’s establishment and survival. Probably the most powerful source of this theory’s staying power is the widely acknowledged empirical fact that wealth and democracy hang together. In an attempt to confirm the causal pathway posited by modernization theory, Inglehart and Welzel (2005) have demonstrated a similar cross-national association between what they call self-expression values and levels of effective democracy. As Geddes (1999) surmises, this empirical regularity—or “stylized fact,” as she puts it—does seem to confirm modernization theory’s fundamental assertions about affinities between certain forms of economic development, human values, and political institutions. This hypothesis has been confirmed again most forcefully by Przeworski et al. (2000: 101), who conclude on the basis of their rigorous statistical analysis that “Lipset was right in thinking that the richer the country, the more likely it is to sustain democracy.”11 And yet, in spite of this broad empirical regularity, modernization theory still leaves us scratching our heads about specific failures of democracy in specific countries. Although modernization theory tells a compelling story about how certain macro-structural processes are broadly conducive to certain kinds of political change, it is essentially silent on the question of how and when those changes occur. So, while it may provide important insights into long-term trends in the population of states, it offers little guidance on the proximate forces behind failures of democracy in specific real-world cases. Many of the other ideas about structural causes mentioned here were apparently intended to help fill in that blank, but they ultimately share the same fundamental weakness: any theory that relies primarily on structure will be hard pressed to explain change. Even when the structural element in question is a dynamic process rather than a static characteristic—as is the case with modernization theory—these theories do not generate clear hypotheses about how and when specific democracies might fail.12 Process-Centered Theories

Sensibly enough, the major body of work that developed as a counterpoint to modernization theory shifted the focus from structure to process in an attempt to glean new ideas about the causes of democratic transition and breakdown. This literature generally traces its origins to work by Dankwart Rustow (1970), who sought to move the conversation from the search for preconditions to the dynamic process by which

10

Dilemmas of Democratic Consolidation

democratic regimes come into existence—what he called a “genetic” theory of democracy. Ironically, Rustow’s critique of structural theory’s inflexibility led him to a model in which the genesis of stable democracy was said to depend not only on the presence of a few “indispensable” preconditions but also on the sequence in which those preconditions were assembled. Although this particular model has not withstood empirical scrutiny, Rustow’s call to consider process apart from precondition has.13 Much of the work in this vein examines problems with the establishment of democracy rather than its consolidation, which is the subject of this enquiry.14 Not all of it does, though. In contrast to modernization theory’s emphasis on structural scaffolding, Linz (1978) sought to examine the process of democratic collapse in several prominent historical cases. In so doing, he sought to identify recurring themes and mechanisms that would hint at the underlying causal dynamics. Perhaps inevitably, the model that emerged from this project was primarily descriptive and emphasized the role of agency. According to Linz, democracies break down when elected leaders fail to respond effectively to crises. This failure of leadership makes the problems producing those crises appear unsolvable, and that sense of insolubility creates an opportunity for semi-loyal or disloyal oppositions to usurp power. Linz (1978: 50) writes: In the last analysis, breakdown is a result of processes initiated by the government’s incapacity to solve problems for which disloyal oppositions offer themselves as a solution. That incapacity occurs when the parties supporting the regime cannot compromise on an issue and one or the other of them attempts a solution with the support of forces that the opposition within the system perceives as disloyal. This instigates polarization within the society that creates distrust among those who in other circumstances would have supported the regime.

At root, then, this theory posits that democracy fails because of poor leadership. According to Linz (1978: 51), “Oversimplifying somewhat, we can say that a regime’s unsolvable problems are often the work of its elites.” Linz argues that elected leaders often set themselves up for this problem when they adopt agendas more ambitious than their means allow, but he also claims they retain some ability to salvage or sabotage democracy right up until a regime’s final moments. The presence and growth of disloyal or semi-loyal opposition is a crucial element in this story, and it often has its origins in conditions and decisions that long precede the crisis, such as the mode of transition to democracy, the extent of power-sharing in the new democracy, the toleration of militia

Introduction

11

groups, and the design of democratic institutions. This whole process is said to be mediated by the depth of legitimacy and the regime’s efficacy and effectiveness prior to the crisis, and those conditions are shaped by structural characteristics, but none of these relationships is deterministic. The heart of Linz’s explanation is the shift from centripetal to centrifugal politics, but the unanswered question is why this dynamic tips one way or the other. His descriptive model identifies leadership, statesmanship, flexibility, and timing as key variables, but that seems like an incomplete answer at best. His comparative analysis begins to get at the why, but—perhaps because of his initial decision to restrict his sample to a fairly narrow set of cases—it ignores the possibility of breakdown outside of a prior political crisis. In other words, there is a substantial range of strategic possibilities that his model overlooks. A newer but now substantial body of work frames the problem of democracy’s survival as the result of a dynamic and uncertain process of regime consolidation. What exactly consolidation entails remains the subject of much discussion and debate (Schedler 1998). Some scholars understand consolidation in probabilistic terms, arguing that it refers simply to the expectation that a particular democracy is almost certain to survive indefinitely (Acemoglu and Robinson 2006, Przeworski et al. 2000, Svolik 2008). Others, however, seek to explain how that expectation emerges, usually by examining changes over time in attitudes and institutions. One of the most elaborate statements of the latter view comes from Larry Diamond (1999), who defines consolidation (p. 65) as “the process of achieving broad and deep legitimation, such that all significant political actors, at both the elite and mass levels, believe that the democratic regime is the most right and appropriate for their society, better than any other realistic alternative they can imagine.” In his view, this process depends on developments in three areas: the regime’s economic and political performance, the strengthening of political and civil institutions, and the deepening of democracy through expanded participation and accountability. Where these trends occur, democracy is likely to survive; where they fail to start or stall, democracy will remain tenuous. While rich and provocative, Diamond’s claims are illustrative of what I consider to be the chief shortcoming of work on consolidation: these theories are primarily descriptive or normative, not explanatory. They suggest one way to observe whether or not the prospects of democracy’s survival are improving or deteriorating, but they generally say little about the forces driving those trends. To the extent that they do, they usually return to transitology’s emphasis on the “will and skill”

12

Dilemmas of Democratic Consolidation

of political elites. For theories that see changes in norms and values as the outcome of interest, this explanation flirts with tautology. Even more problematic in my view, these explanations also fail to make explicit how those elites’ choices are shaped by strategic considerations. Scholars associated with this school often acknowledge that elites do not simply pursue their own interests in a linear fashion, but they generally do not spell out what other issues those actors might consider and how those considerations might affect their behavior. Game-Theoretic Approaches

Another and generally newer literature uses game theory to explore in a more rigorous way the strategic interactions among key actors. This body of work tries to combine insights about the influence of structural forces with careful consideration of the kinds of choices that individuals and organizations must actually make, and then to link those individual choices to social outcomes. As Barry Weingast and Rui de Figueiredo (1999: 263) summarize, “The hallmark of rational choice theory for explaining macrosocial failure is its approach to social dilemmas of cooperation. In a variety of circumstances individually rational actions produce socially irrational outcomes.” In one sense, democracy is an attempt to institutionalize macrosocial cooperation, and its failure can be studied by searching for ways in which the choices of specific individuals and organizations produce what are often socially irrational results.15 In his pathbreaking work on democracy, Adam Przeworski (1991) used the logic of game theory to turn our gaze away from the agent’s skills onto the incentives to which those agents respond. His model focuses on the actions of election losers as the determinants of democracy’s survival, and it implies that democracy will endure when the losers in any particular election have sufficient prospects for winning in future elections that it is better to concede defeat and wait to fight again at the ballot box instead than to rebel in response to the latest loss. In an important extension of that model, Przeworski (2006) incorporates the prospect that election winners may also “rebel,” where rebellion by either party is understood as an attempt to impose a dictatorship in order to redistribute income to their supporters. In this version of the model, if either party chooses to rebel, a violent confrontation ensues, and the outcome of that confrontation is determined by the balance of military force. Nevertheless, the key insight from the earlier version is essentially unchanged: actors comply because they believe future elections afford them a better chance to advance their interests than subversion would.

Introduction

13

Consistent with Przeworski’s general logic but with a sharper focus on economic inequality as the main engine of politics, Carles Boix (2003) and Daron Acemoglu and James Robinson (2006) posit that we may usefully understand politics as a struggle between rich and poor over the distribution of wealth. Broadly speaking, Acemoglu and Robinson argue that democracy is established when the threat of rebellion by the poor against a dictatorship of the rich is credible and formidable, and democracy persists as long as the resulting redistribution is not costlier to the rich than the coercion and repression they would have to supply to mount a coup and sustain the ensuing dictatorship. In particular, these authors emphasize the way that democracy links policy outcomes to the preferences of the median voter and thereby offers a credible commitment by the rich to redistribute enough wealth to the poor to constrain the threat of a popular rebellion. Boix’s theory follows a similar logic but associates the commitment problem with both of the actors, thus adding the possibility that democracy can fail by revolution as well as coup. All of these authors model democracy as a strategic interaction among actors seeking to advance their own interests, often at the expense of others. Where Przeworski’s model emphasizes those actors’ future electoral prospects as an incentive to, or constraint against, the subversion of democracy, Boix and Acemoglu and Robinson emphasize the policies that democracy is expected to produce, and thus the relative costs to the wealthy of accepting some amount of redistribution or reneging on that commitment and sustaining or re-imposing dictatorship. What Przeworski’s model seems to lack is a way to address the kinds of commitment problems Acemoglu and Robinson and Boix spotlight. The contestants in democratic politics have to worry not just about how the next election might turn out, but whether it will occur at all, and if it might not, what they ought to do about it. Meanwhile, Acemoglu and Robinson’s model seems to obscure the point that democracy produces not just policies but also electoral winners and losers, and both of these groups may find reason to prefer a change in the political order. This can happen, in part, because in an immediate sense those electoral winners and losers are not social groups or economic classes but political parties—in other words, a specific kind of organization that does not always act neatly on behalf of the citizens it claims to represent. At the same time, none of these models deals directly with the potentially autonomous role of the military in this process.

14

Dilemmas of Democratic Consolidation

Democracy’s Dilemmas

To develop a theory of democratic breakdown and consolidation, I start from the observation that different actors can usurp power in different ways, depending on where they sit (figuratively, in political space) and what skills they possess. As Przeworski (1991: 12) argues, “Democratic societies are populated not by freely acting individuals but by collective organizations.” These organizations are composed of aggregations of citizens, and they typically make claims on behalf of even larger groups defined by some commonality of interest. The fact that they make collective claims, however, does not mean that these organizations simply channel their constituents’ interests. Instead, organizations bring specific skill sets, structures, and even interests of their own to their interactions with other actors, and these attributes shape their behavior, just as external incentives do (Milgrom and Roberts 1992, Simon 1976). What’s more, organizations often develop interests that diverge from the interests of the collections of individuals on whose behalf they are supposed to act. As a result, those constituents often find it difficult or costly to monitor the organization’s behavior and to punish it for wrongdoing. Economists refer to this as the principal-agent problem, and it pervades collective action (Milgrom and Roberts 1992; Niskanen 1971). Organizations also tend to persist. Because coordination problems must be overcome to get them off the ground, the costs of starting them often exceed the costs of maintain them, and once they exist they usually fight for their own survival. As a result, organizations usually take on lives of their own, even as they serve to facilitate exchanges among the individuals they purportedly represent. Organizations do emerge, change, and die, and whole fields of study in economics and sociology are devoted to understanding these dynamics.16 Nevertheless, the tendency for organizations to develop interests of their own and to outlive the impulses that led to their creation makes it reasonable to treat them as political actors in their own right at particular moments in time. Following this reasoning, the first of the simplifying assumptions I make in trying to understand the process of democratic breakdown is to focus on specific types of organizations, not competing social classes or specific individuals, as the most relevant actors. Observation of the ways that democracies fail in the real world identifies three organizations as the crucial ones: 1) election winners, a.k.a. the incumbents; 2) election losers, a.k.a. the opposition; and 3) state security forces, a.k.a. the military. The differences in these organizations’ roles and capabilities

Introduction

15

mean that they respond to different sets of incentives and usurp power in different ways. As a second simplification, I assume that the organizations most relevant to democracy’s survival seek to maximize their material welfare. Importantly, this assumption implies that those organizations do not automatically seek to gain or retain political office for its own sake (Geddes 1999). State authority is understood here as an instrument used to pursue other ends, not an end in and of itself, and the value associated with controlling that instrument varies according to the outcomes it can help produce. If this were not true, democracy would never survive for long, because incumbents would routinely engage in extreme behavior to retain their positions no matter how long the democracy had existed. In this framework, the incumbent and opposition are political parties or coalitions of parties, meaning that they are organizations composed of citizens with political skills. These organizations exist for the purpose of mobilizing voters and producing policy. Those functions can be performed in many ways by many different kinds of individuals, from village elders to precinct captains, from thugs to marketing professionals. The skill and success of those organizations at mobilizing voters on their behalf determines whether or not they win elections, and the winning or losing of elections determines whether or not their preferred policies are pursued. Generally speaking, there are no functional differences between the incumbent and opposition parties; the distinction between the two is strictly the result of the preceding election, and those roles are interchangeable. The military is a very different kind of organization. It is composed of “specialists in violence” (Bates 2001) who have explicitly or implicitly entered into a contract with the state to act as its agents of legitimate coercion—that is, its “muscle.” What is essential to this book’s theory is that, in a democracy, this organization is subordinated only to the state, not to a particular political party, which means it really isn’t subordinated to anyone at all, because in this relationship the state is just an abstraction.17 The individuals who exercise the associated authority are interchangeable, and the idea of what constitutes the state is subject to interpretation.18 Put another way, Weber’s legitimacy is an idea, but the guns are real. The military’s coercive skills and capacity give it an inherent capability for independent political action. In other words, the statesanctioned military is not just an element in the calculus determining the balance of coercive power among political parties. Instead, it is a distinct organization with an inherent potential for autonomous action, and it may choose at any time whether or not to ally itself with any particular

16

Dilemmas of Democratic Consolidation

party or to act on its own behalf in a manner that might not reflect other actors’ ideas about loyalty to the state. As much as those of us who live in countries that have not recently suffered coups would like not to think about it, this description applies to militaries in rich countries with long democratic traditions just as well as it does to militaries in poorer countries with a recent history of praetorian rule. As Rapoport (1968: 552) argues, “As long as we can distinguish between a government on the one hand, and the armed and unarmed portions of its public on the other, military usurpation is always conceivable; and, in fact, history gives little support to the supposition that an unmistakable movement from military insubordination to subordination exists.” Rapoport’s reference to the military as the armed portion of the public hints at an important point about the origins of the military’s political interests. Many theories of political rule do not treat the military as an independent actor, but ones that do often see its interests arising from the nature of the military as an organization. As an actor, the military is often assumed to be interested primarily in maximizing its budget and benefits, minimizing its costs, and protecting its reputation. While these issues will often be important, we should also keep in mind that the military is composed of armed citizens who, presumably, also have interests as citizens. When citizens in a democracy suffer, soldiers and their families often suffer, too, and this suffering can impel those soldiers to react in their capacity as soldiers. In other words, the forces shaping the incentives for soldiers to act politically in a democracy are not limited to the ones that directly concern the military as an organization.19 The reference above to affirmation of democratic government hints at the point that, in order to persist, democracy must constantly be produced through the behavior of individuals and groups engaged in it. This is true of any set of political institutions, and democracy is no exception. There is no magical set of conditions under which democracy becomes permanent and the risk of failure is therefore zero. As other scholars have observed (Acemoglu and Robinson 2006; Boix 2003; Przeworski 1991, 2006), there do appear to be conditions under which the persistence of democracy becomes an equilibrium from which actors are highly unlikely to deviate. That persistence should not be confused with permanence, however, and even in the world’s oldest democracies, there exists at least the possibility that unexpected shocks to the system or an accumulation of other processes could lead to authoritarian rule.20 Because it controls state authority, the party in power—the incumbent—uniquely possesses the opportunity to abrogate democracy from the inside. In other words, the incumbent can usurp power without

Introduction

17

directly employing coercion, by simply changing the formal rules or impinging upon the requisite supporting practices in ways that favor its continuation in office. In some instances, these alterations take the form of direct electoral fraud, declared annulments of constitutional procedures, or unilateral and extra-constitutional dismissals of other elected officials. In others, the party in power infringes more gradually and perhaps more subtly on important enabling conditions for representation and transparency, such as rights of free speech, procedures for voter registration, or rules governing the formation and operation of political parties. In all cases, however, the essential dynamic is the same: incumbent officials use their authority and influence to alter electoral procedures or to diminish civil liberties in ways that ensure their continuation in office. For analytical purposes, we can treat all of these actions as a single mode of democratic failure, the executive coup. A critical aspect of executive coups is that they can be carried out without acquiring new skills or mobilizing new supporters. The incumbent party simply uses its de jure power to rewrite rules, alter institutions, or influence key officials—exactly the kinds of tasks it was organized and elected to do. This is important to the prospects for democracy’s survival because it means that executive coups are relatively cheap, at least in their execution. They certainly require collective action, but the number of individuals who must act collectively and the immediate costs of their actions are usually small, and those individuals are already bound together by organizational ties which at least partially align their interests and help them to overcome the kinds of coordination problems that generally inhibit more spontaneous forms of collective action. Many of the tactical maneuvers involved in carrying out an executive coup are also inherently ambiguous in their intent. Political parties are expected to seek an edge wherever they can find one, and elected governments sometimes have legitimate reasons for imposing restrictions on civil liberties or political action. For the opposition and the military, this ambiguity of intention translates into uncertainty about the strategy the incumbent party is pursuing. As events unfold, it will often seem unclear whether the incumbent is seeking partisan advantage within the “normal” parameters of democratic politics or is instead building incrementally toward a decisive break with democracy. As revealed by the formal model developed later in this book, under the right (or wrong) circumstances, that uncertainty can significantly affect the prospects that democracy will live or die.

18

Dilemmas of Democratic Consolidation

Opposition parties see two paths to power. They can try to win the next election (assuming that it occurs), or they can try to overthrow the party in power by extralegal means—in a word, through rebellion. In contrast to executive coups, rebellions require a type of organization and skills that differ significantly from the ones a political party uses to contest elections. Rebellions are also illegal, which means that organizations preparing for them must take measures to hide their actions or risk punishment or defeat if they do not. Finally, because rebel movements usually lack a dominance of force, the formal trappings of state authority, or both, rebellions are just plain hard to finish. These aspects of rebellions make them a costly and difficult undertaking compared with executive coups.21 Even when opposition groups manage to mobilize a large and imposing force, the “end game” of actually seizing power usually affords other actors, domestic and foreign, numerous opportunities to redirect the outcome in their favor. Events in Ecuador in February 1997 provide a case in point. Amidst a deepening economic crisis, opponents of President Abdalá Bucaram managed to bring literally millions of people into the streets as part of a general strike aimed at compelling Bucaram to resign. The National Assembly responded to the show of popular force by voting to remove the eccentric president on grounds of “mental incapacity” and then installed its leader, Fabián Alarcón, as interim president. Thus, in spite of their demonstration of tremendous de facto power, the citizens who effectively forced Bucaram’s ouster played no direct role in determining what happened next. It is also important to clarify that rebellions need not involve the direct use of violence in order to pose a threat to democracy. Nonviolent uprisings that aim to topple elected governments also represent a form of rebellion that can lead to a break from democratic governance. Philosophically, the democratic credentials of these nonviolent popular uprisings is sometimes ambiguous. Democracy is supposed to entail government of, by, and for the people, and in cases where an elected government has become broadly unpopular, mass uprisings that lead to the installation of a new set of rulers can seem like an expression of the “will of the people” every bit as legitimate as an election. The key point, though, is that democracy as a system of government depends fundamentally on adherence to rules and procedures—constitutionalism, as Diamond (2008) puts it—and the procedures at the core of this arrangement are the ones that describe who votes, when, and how those votes are translated into seats in office. Thus, efforts to circumvent those procedures can pose an inherent threat to democracy as a system of government, whether or not they directly use violence.22

Introduction

19

All national governments, whether democratic or authoritarian, depend constantly on the support of an armed force, an organized agent of coercion, to promote or protect their claim to a monopoly on state authority within their territory. For democracy to exist, however, this unelected force must limit itself to that negative role of protection; it must not engage in positive political action. Soldiers are citizens too, but those roles must be separated, and their actions as soldiers must come at the discretion of their elected rulers. When military officers abrogate democracy by claiming control over central state authority, it is a military coup.23 This mode of democratic failure involves military leaders acting as representatives of the military as a state-sanctioned organization. When soldiers act against their government as participants in opposition parties or armed insurrections, any resulting abrogation of democracy is not considered a military coup. In spite of the military’s preponderance of coercive strength, military coups are still difficult to execute. The degree of difficulty depends, in part, on the internal organization of the armed forces. In cases where officers and soldiers are loyal to its perpetrators, a coup may be bloodless and relatively cheap; in cases where the military is internally divided, the execution of a coup may exact a much higher price, or it may fail as a result of the ensuing infighting. Even when the military acts coherently, however, the process of converting the coercive power of guns into the persuasive power required to sustain political authority is not a straightforward one. Thus, military coups can founder in spite of initial tactical successes as long as enough of the rest of the relevant actors—including ordinary citizens—refuse to play along. The need to inspire quasi-voluntary compliance in order to obtain their strategic objective poses an additional obstacle to political aspirations of military leaders, and this need helps explain why coups are relatively rare in spite of the military’s inherent combination of coercive power and political interests. In deciding whether or not to abide by the democratic rules of the game, the ruling party and its political rivals both must consider more than their prospects for victory or defeat in the upcoming election and their expected welfare in an authoritarian regime of its own making. Likewise, the military cannot focus myopically on its own preferences. Instead, each of these actors must take into account the risk of a usurpation of democracy by the others. The multiplicity of threats and simultaneity of action in a context of incomplete and imperfect information raises the possibility that any of these actors might act strategically by staging what amounts to a preemptive strike, attempting to establish an authoritarian regime that it controls before another actor

20

Dilemmas of Democratic Consolidation

does the same. In other words, this representation of the problem suggests that the establishment of an autocracy does not have to be any of the actors’ most preferred outcome in order for a coup to occur. Any one of them might also attempt to abrogate democracy as a means to reduce the risk of ending up with a least-preferred outcome in which another party usurps power and shuts them out of the ensuing authoritarian regime. Methodology

My aim in this book is to develop a new, mid-range theory of democratic breakdown. If this theory is to prove useful, it must focus our attention on certain actors, their interests and beliefs, and, ultimately, the strategic choices they make in ways that are both logically coherent and accord with observed outcomes in the real world. In other words, the situations and processes the theory describes in a stylized way must really occur, and viewing those occurrences through the lens of this theory should shed some light on why they happened in ways that prior theory does not. A formal model, developed in Chapter 2, serves as the theory’s foundation. Game theory is a powerful tool for theory building because it forces us to clarify our logic. At the same time, as Bates et al. (1998) argue, game theoretic models are most useful when they are based on assumptions that accord with real-world patterns. The conventions of formal modeling—the symbols, equations, diagrams, and such—are not the theory. They are only a language and grammar used to help tease out insights from a set of prior ideas about the nature of politics in democracies. The assumptions made here about the actors involved, their interests, and their decision-making processes are, of course, gross simplifications. Still, they are not chosen solely on the basis of their ability to produce novel hyoptheses and accurate predictions. They are also chosen because they are thought to reflect important elements of observed human behavior. To explore the theory’s relevance to the real world, I use “analytic narratives”—retellings of historical events that trace the actors and concerns spotlighted by the theoretical model to see if they offer fresh insight into real-world cases. These narratives are not sufficiently precise or detailed to falsify the theory, a goal that would probably be inappropriate at this stage anyway.24 Instead, these sketches are meant to serve the more modest goals of illustrating the elements of the theory in action and, in so doing, of providing some validation for the choices made in its construction. As Bates et al. (1998: 234) argue,

Introduction

21

Rational choice theory (and in particular the theory of games) offers a theory of structure: it suggests a way in which structures create incentives that shape individual choices and thereby collective outcomes. Insofar as a game yields multiple equilibria, it may be difficult to test its explanatory power. But the ‘force’ of a game may lie in the properties of structure that it highlights and in the strategic problems to which it gives rise.

The formal model developed in this book is dynamic and complex, so it is not surprising that it yields what game theorists call “multiple equilibria”—in other words, that it does not produce unique predictions about choices and outcomes in all situations. Given this complexity, my chief aspirations for the theory are that it identifies significant strategic dilemmas common to all democracies and helps to clarify how certain structural conditions and external shocks or interventions can exacerbate or diminish those dilemmas. If it does those things, it should also guide us toward more insightful descriptions of real-world cases and, perhaps, to more accurate predictions about their future survival or failure. The theory I develop is meant to be mid-range; in other words, it is meant to produce general insights, but only across a narrow class of cases. My game-theoretic model only applies to situations where the formal and informal institutions required to produce democracy, understood here as a system of government in which rulers are routinely held accountable to citizens (Schmitter and Karl 1991), have already come together, even if only briefly. My hope is that the theory sheds light on the behavior of certain sets of individuals in democracies across history and well into the future. That said, the model is most certainly not intended to be a general theory of politics. It is not designed to help us understand politics in authoritarian regimes or how democracy arose in the first place. If it is useful, the theory will shed light on elements of the interplay among actors and between actors and structure in democracies that help explain outcomes in specific cases in new ways, that accord with general trends in empirical data, and that offer specific and testable predictions about outcomes in cases that occur in the future. Organization of the Book

The book proceeds as follows: Chapter 2 takes the assumptions and intuitions about politics in democracies discussed in the Introduction and translates them into a formal game-theoretic model. Comparative statics are used to see what the model can tell us about the prospects for democratic breakdown or consolidation in a few archetypal scenarios,

22

Dilemmas of Democratic Consolidation

and anecdotes are used to illustrate some of those scenarios in action. Chapter 3 applies descriptive event-history techniques to cross-national data, summarizing broad patterns in the survival of democracy over the past half-century and discussing what my model suggests about those patterns’ origins. Chapters 4 and 5 use process-tracing narratives to demonstrate that this book’s game-theoretic model sheds new light on how the democracies represented in those cross-national data actually survived or failed. The first of those two chapters explores four episodes of democracy, one selected at random from each of four sets according to their outcomes: breakdown by executive coup (Ukraine in the 1990s), breakdown by military coup (Fiji in the 2000s), breakdown by rebellion (Cyprus in the 1960s), and consolidation (Spain after Franco). The second of those chapters turns to a pair of recent breakdowns that most theories of democratic consolidation would have failed to anticipate— Venezuela and Thailand—to show that this book’s theory can help to explain those surprises in ways that conventional theories cannot. Chapter 6 spells out some implications of the model for international democracy-promotion efforts; broadly speaking, it shows how the technical view that motivates most current work in the field of democratic development overlooks the political implications of specific interventions and may therefore produce unintended and undesirable consequences, and it uses the model to generate a few concrete recommendations for future efforts at democracy promotion. The book concludes by revisiting the theory’s major implications, identifying areas for further research, and speculating about future trends in the global spread and consolidation of democracy.

Notes 1 “With Premier at U.N., Thai Military Stages Coup,” The New York Times (September 20, 2006). 2 According to the World Bank’s World Development Indicators, in Thailand in 2004, per capita income measured in constant 2000 $US was $2,356, on par at the time with Peru and Russia. 3 See Ismael Wolff, “The Silk Revolution,” The New York Times (September 26, 2006). 4 Throughout this book, the term “consolidation” is used to refer narrowly to expectations about regime survival. Following Schedler (1998), a democracy is considered consolidated when most observers, including the key actors themselves, expect it to survive indefinitely. Of course, these expectations do not guarantee that democracy will survive, and one of the points of this book is to illuminate strategic concerns that can affect those beliefs and the behavior of the actors to whom they adhere.

Introduction

5

23

For a detailed discussion of the numbers, see Chapter 3. Some of the definitions of democracy that have most powerfully influenced contemporary research on democratization are found in Schumpeter (1945), who identified the process of election as the sine qua non of democratic government; Dahl (1971), who proposed to measure regimes by their inclusiveness and contestation for office and emphasized the supporting role of civil liberties; Schmitter and Karl (1991), who linked democracy to the concepts of citizenship and accountability; and Przeworski et al. (2000), who assert that democracy is inherently a yes/no concept and understood it as a system in which governments lose elections. 7 I do not consider judicial review by a supreme or constitutional court whose members are appointed by elected officials to be veto power by an unelected entity. Likewise, I do not consider routine consultation or advisement of elected officials by unelected entities to be inherently undemocratic, either. 8 For an insightful intellectual history of modernization theory, see Gilman (2003). 9 Cheibub (2007) offers the most thorough and compelling rebuttal of this argument, showing on both theoretical and empirical grounds why this claim is flawed. 10 See Norris (2004: 3–22) for a careful review of arguments in this vein and counterarguments concerning the enduring effects of culture and values. 11 See also Epstein et al. (2005) and Ulfelder and Lustik (2007). 12 See Doorenspleet (2005) for a thoughtful effort to integrate and test ideas from several structural theories of democratization. 13 Also worth noting, the idea that a special sequence of changes holds the key to the establishment of stable democracy has also lingered in spite of various attempts to declare it dead. 14 The most influential expression of this “transitology” school comes from O’Donnell and Schmitter’s (1986) comparative analysis of democratization in Latin America and Southern Europe. From those cases, these authors concluded that the “will and skill” of political elites—along with significant doses of luck and chance—represent the driving forces behind the outcome of transitions from authoritarian rule. 15 Some scholars have used game theoretic models to examine other aspects of the democratization process. Colomer (1991) and Przeworksi (1991), for example, develop models of transitions from autocracy to democracy, while Geddes (1999) sketches games that help explain how different forms of authoritarian rule are sustained. To the best of my knowledge, however, little of the prior work in this vein applies directly to the subject of this book, namely, the survival and breakdown of democratic regimes. 16 In sociology, see especially Hannan and Freeman (1989). 17 As Hobbes writes in Behemoth, “For if men know not their duty, what is there that can force them to obey the laws: An army, you will say. But what shall force the army?” 18 I do not say that it is loyal to the constitution because the notion that the constitution embodies the state is just one (particularly American) version of the kind of idea I’m talking about. In Turkey, the military has long considered itself loyal to an ideology, Ataturkism, which is only partially expressed in that country’s constitution. In Thailand, the military appears to emphasize loyalty to 6

24

Dilemmas of Democratic Consolidation

the king over the constitution. The common thread here is some idea about the proper foundation of state authority, which, ironically, cannot exist without some coercive means of enforcement. 19 Coup attempts in Venezuela in 1992 and Ecuador in 2000 illustrate this idea. In both situations, economic crisis led to widespread popular frustration with elected governments. That frustration produced popular unrest that culminated in attempted coups involving middle-ranking officers and rank-andfile soldiers from the communities that were suffering. Thus, although they both involved segments of the armed forces, these grabs for power do not seem to have had much to do with the interests of the military as an organization. 20 For a thoughtful discussion of this point, see Armony and Schamis (2005). The notion that the risk of failure is never zero also has important implications for qualitative and quantitative analysis, because it implies that all democracies can be compared to one another in the search for correlates of survival and termination. 21 In his study of democratic breakdown, Linz (1978: 15) observes that, “The twentieth century has seen fewer revolutions started by the populace than the nineteenth, and their fate in modern states has generally been defeat. The Communists and Nazis learned that lesson. Mussolini’s combination of illegal action and legal takeover became the new model for overthrow of democracies. Only the direct intervention of the military seems to be able to topple regimes in modern stabilized states.” 22 That said, it is also worth noting that subversions of an existing elected government may sometimes be required to establish a democracy that will be more durable in the long run. The principle here is the one described by Thoreau in his essay on civil disobedience: “Those who, while they disapprove of the character and measures of a government, yield to it their allegiance and support are undoubtedly its most conscientious supporters, and so frequently the most serious obstacles to reform” (Bode 1947: 118). 23 When the military routinely makes policy decisions, we don’t call it a democracy in the first place. Coups also differ from situations in which the military is called on by the elected government to provide internal security, perhaps during a period of emergency rule. 24 On the futility of seeking absolutely to falsify social-science theory, see Bates et al. (1998: 14–18).

2 The Democracy Game

To see what the ideas developed in the Introduction suggest about how and why democratic regimes consolidate or fail, this chapter translates those ideas into a formal game-theoretic model and then explores what that model predicts under different sets of conditions. To start, we can think of politics in a democracy as a dynamic, three-player game involving two rival political parties and the military. 1 An election cycle—that is, the period from one national election to the next—comprises a single iteration (or round) of the game, which is assumed to repeat for an indefinite number of iterations. At the start of the game, the institutional design of the democracy has been set and elections have been held to determine which of the rival political parties holds power. The game then unfolds as follows, where “defection” simply means a player attempts to usurp power and “cooperation” means he does not. Figure 2.1 on the following page shows the game as a tree, with nodes representing decision points and branches representing the decision-makers’ available courses of action. 1. The opposition party (O) chooses to cooperate or to defect. 2. The incumbent party (I) chooses to cooperate or to defect. 3. The military (M) chooses to cooperate or to defect. 4. Nature determines the outcome, contingent on the players’ moves. a. If all players cooperate, democracy survives until the next election. Nature determines the outcome of that election, and players receive the associated payoffs. b. If one or more players defect, Nature determines whether any of those defections succeeds. Only one can succeed. i. If any defection succeeds, the game ends with the establishment of a dictatorship led by the player who attempted it, and all players receive the payoffs associated with that outcome.

25

26

Dilemmas of Democratic Consolidation

ii. If no defections succeed, democracy survives until the next election, when Nature determines the outcome of that election. All players receive the payoffs associated with that outcome, and any player who defected suffers a penalty. Figure 2.1. The Democracy Game

O

D

C

I

C

I

D

M

C

{CCC}

C

M

D

C

{CCD} {CDC}

D

M

D

{CDD}

C

{DCC}

M

D

{DCD}

D

C

{DDC}

{DDD}

This game has five possible outcomes, denoted here as follows. Di = incumbent-led democracy Do = opposition-led democracy Ai = incumbent-led autocracy Ao = opposition-led autocracy Am = military autocracy The ordering of the players’ moves—opposition first, then incumbent, then the military—is meant to capture typical differences in the information available to the players about each others’ strategies and the specific forms of usurpation each one may attempt, based on the nature of their organizations. Rather than serving as an advantage as it often does, the opposition’s status as first mover in this game reflects its

The Democracy Game

27

incomplete information about the actions the other two players will take and the cumbersome and public nature of the form of defection available to it (rebellion). The opposition is uncertain about the incumbent’s strategy, in part, because the incumbent will often be able to pursue an executive coup in secret, at least initially. Meanwhile, the military’s capacity for rapid and forceful action positions it as a kind of veto player, giving it the advantage of waiting and then responding to the parties’ actions as they become clearer. The military’s position as final mover also reflects uncertainty among the political parties over the nature of the military’s loyalties, and therefore its preferences, due in part to the secretive nature of militaries as agents of national security. To put the game to work, I assume that all of the players are instrumentally rational, meaning that they will choose the course of action which they expect to yield their most preferred outcome (Zagare and Kilgour, 2000: 39). As for why those actors prefer some outcomes over others, I assume for purposes of simplification that all actors seek to maximize their material welfare. Cultural and organizational forms can affect the actors’ ideas about material well-being and the appropriate and possible paths to it, but the key point is that none seeks political office for its own sake. This means that outcomes involving control of state authority are not necessarily valued over others, although for the political parties, at least, they often are. The solutions to the game discussed below involve several additional parameters and simplifying assumptions. First of all, if democracy survives, each player’s expected payoffs (πi) can be represented as πi = (p)(Di) + (1 − p)(Do) − κi

where p is the probability that the incumbent party wins the next election and κi is the expected cost of punishment for having attempted a coup and failed (where κi always equals zero for players who chose cooperation). As the addition of the punishment parameter suggests, I assume that actors can expect to be punished for attempting to usurp power and failing, if democracy survives. Any punishment likely to be meted out by an authoritarian regime led by a rival organization is factored into the expected payoff associated with that outcome. Second, I assume that coup attempts and rebellion require collective action within the relevant organizations, and that collective action is costly. This cost is represented here as µ, where µi = 0 if the ith player chooses to cooperate and µi > 0 if the ith player chooses to defect. These costs vary across actors and situations. In general, I expect that

28

Dilemmas of Democratic Consolidation

mobilization is usually but not always less costly for the military, because of its hierarchical organization. I make a similar assumption about the costs of collective action by the incumbent party, because executive coups can be executed by a relatively small number of individuals doing their regular jobs. By contrast, I assume that rebellions are generally much costlier to attempt, because they usually require the mobilization of a larger set of participants for sustained and extraordinary action. In general, though, the cost of collective action is represented in the game as a continuous variable that is always greater than zero. Third, I assume that the probability any actor’s attempted usurpation of power will succeed is a function of his capabilities, which are indicated here as σ, where σi represents the probability that a coup attempt by ith player will succeed in the absence of coup attempts by either or the other players (so 0 < σi < 1). To simplify the math, I assume that these capabilities are independent, and that the odds that any one player’s attempt will succeed relative to the odds of the others’ attempts succeeding are the same, no matter how many players defect. This idea is captured mathematically in the following equation, where σi = 0 if the ith player chooses to cooperate (i.e., does not attempt a coup).2 σi′ = [σi / (σi + σj + σk)] × {1 − [(1 − σi) × (1−σj) × (1 − σk)]}

There are eight possible combinations of strategies in this game. Based on the assumptions described and notation used above, the ith players’ expected payoffs under those eight combinations of strategies can be represented as follows, where C indicates cooperation and D indicates defection. πi(Ci,Cj,Ck) = (p)(Di) + (1 – p)(Do) πi(Ci,Dj,Ck) = (σj)(Aj) + {(1 – σj) × [(p)(Di) + (1 – p)(Do)]} πi(Ci,Cj,Dk) = (σk)(Ak) + {(1 – σk) × [(p)(Di) + (1 – p)(Do)]} πi(Ci,Dj,Dk) = (σj′)(Aj) + (σk′)(Ak) + {(1 – σj′ – σk′) × [(p)(Di) + (1 – p)(Do)]} πi(Di,Cj,Ck) = (σi)(Ai) + {(1 – σi) × [(p)(Di) + (1 – p)(Do) – κi]} – µi πi(Di,Dj,Ck) = (σi′)(Ai) + (σj′)(Aj) + {(1 –σi′ – σj′) × [(p)(Di) + (1 – p)(Do) – κi]} – µi

The Democracy Game

29

πi(Di,Cj,Dk) = (σi′)(Ai) + (σk′)(Ak) + {(1 – σi′ – σk′) × [(p)(Di) + (1 – p)(Do) – κi]} – µi πi(Di,Dj,Dk) = (σi″)(Ai) + (σj″)(Aj) + (σk″)(Ak) + {(1 –σi″ – σj″ – σk″) × [(p)(Di) + (1 – p)(Do) – κi]} – µi

We can explore the game’s implications by assuming different preference orderings for the players under conditions of complete and perfect information and then relaxing some of those assumptions to the real-world uncertainty about the other actors’ interests, capabilities, and intentions. This rest of this chapter derives and then discusses results under a few sets of assumptions that are intended to represent archetypal situations often described in other theories of democratization. Scenario 1: Democracy Preferred

As a starting point, let’s assume that all of the players prefer democracy to all forms of authoritarian rule. They probably have preferences over which party wins elections, and they might like some kinds of autocracy even less than others, but as a rule they always favor democracy over autocracy. For each of the political parties, we can safely assume that they would like to be and stay the incumbent—in other words, to win elections. If their rival party is to hold power, they would prefer that it does so in a democracy, where institutional arrangements give them greater influence over the policies that rival adopts and the odds of returning to power are greater. If they must live in an autocracy, they would prefer to lead that regime over being ruled by someone else. If they must live under someone else’s authoritarian rule, they would prefer to see the military govern instead of their political rivals, because the military may seek a hasty return to civilian rule or establish an inclusive government, while their rivals can be expected to shut them out of power and pursue policies more detrimental to their interests. This means that an authoritarian regime led by their partisan rival is their least-preferred outcome. The military is assumed to share this general preference ordering; the only difference is that it is indifferent to which of the political parties rules. These assumptions produce the following preference orderings: Incumbent: Di > Do > Ai > Am > Ao Opposition: Do > Di > Ao > Am > Ai

30

Dilemmas of Democratic Consolidation

Military: Di = Do > Am > Ai = Ao

With complete information, this scenario produces a unique subgame-perfect Nash equilibrium in which all players cooperate and democracy survives. To solve the game through backwards induction, we start by examining the military’s decision problem in the third stage. The fact that the military prefers either version of democracy to all types of autocracy means that the military will only choose to defect if one or both of the political parties has already defected as well, because a military coup in isolation would only make the military worse off, even if the costs of its collective action and expected punishment were zero. This is plain to see if we use the following equation to consider this decision problem in a general way. If player i expects players j and k to cooperate, he will only prefer to defect when the following holds true: (σi)(Ai) + {(1 − σi) × [(p)(Di) + (1 − p)(Do) − κi]} − µ > (p)(Di) + (1 − p)(Do)

If we represent the expected payoff from the survival of democracy—the right-hand side of that equation—simply as D, we can simplify the equation as follows: (σi)(Ai) + (1 – σi)(D) – (1 – σi)(κi) – µ > D

This further simplifies to: [(σi)(Ai – D)] – [(1 – σi)(κi)] – µ > 0

If D ≥ Ai, then Ai − D ≤ 0, so this inequality can never be satisfied, no matter what value the other parameters take. Consequently, the military that prefers democracy under either political party to autocracy in any form will never choose to defect when both of the political parties have chosen to cooperate. Recognizing this, the incumbent party will only choose to defect in the second stage if the opposition has already defected, because it also prefers either version of democracy to all versions of autocracy for the same reason. Recognizing this at the start of the game, the opposition party will always choose to cooperate, because defection can only lead to one of its less-preferred outcomes. This equilibrium happens to be efficient as well, in that all three actors achieve their best payoff through the mutual cooperation this situation produces. This result is not sensitive to variation in the actors’ expectations about the outcome of the next election; even a certain

The Democracy Game

31

victory for their rival doesn’t alter the party’s choices. Neither is this result sensitive to variations in the size of the payoffs (assuming the order remains the same), the costs of collective action, or the expected penalties for transgression of the democratic rules. In other words, with complete information, a collection of actors who all prefer democracy to any kind of authoritarian rule can be expected to forego attempts at usurping power, and so democracy can be expected to survive. This is probably the situation many of us have in mind when we talk about the “consolidation” of democracy, and the game elaborated here supports the intuition that this situation produces a stable and efficient outcome. Importantly, though, this same set of assumptions also shows how democracy can be vulnerable to coups or rebellion even when all players prefer the survival of democracy to all versions of authoritarian rule. Specifically, the game reveals that there are situations in which the incumbent, the military, or both would try a coup if either of the other players were expected to defect. This expectation is not possible under the assumption of complete information, but it emerges as soon as we relax that assumption and allow uncertainty about the other actors’ expected payoffs or choice of strategies. These patterns are sensitive to variations in the costs of mobilization, the expected penalties for failed coups or rebellions, and the actors’ capabilities. Nevertheless, the point remains that under certain plausible assumptions the incumbent’s and the military’s strategies (in the game theoretic sense of a complete plan of action) would include coup attempts, even if those actors prefer democracy with either party in office to all versions of autocracy. To see how this is so, consider the ith player’s decision problem when confronted with the belief that player j will defect. Under these circumstances, player i will choose to defect in spite of his preference for democracy if: (σi′)(Ai) + (σj′)(Aj) + {(1 – σi′ – σj′) × [(p)(Di) + (1 – p)(Do) – κi]} – µi > (σj)(Aj) + {(1 − σj) × [(p)(Di) + (1 – p)(Do)]}

This simplifies to: (σi′)(Ai) − [(1 − σi′ − σj′)(κi)] − µi > (σj – σj′)(Aj)

If we temporarily assume that there are no mobilization or punishment costs (so κi = 0 and µi = 0), this simplifies further to: (σi′)(Ai) > (σj – σj′)(Aj)

32

Dilemmas of Democratic Consolidation

In prose, this inequality implies that player i would rather defect if his expected payoff from ruling as an autocrat (Ai), weighted by the conditional probability that his coup attempt will succeed (σi′), is greater than his expected payoff under an autocracy led by player j (Aj), weighted by the marginal difference in the chances that j’s coup attempt will succeed even though i has also defected. As long as Ai > Aj (i.e., player i would rather rule in an autocracy than be ruled in one), this inequality implies that there can be values of the other parameters (κi, µi, σi, and σj) such that the initial inequality will be satisfied, leading player i to defect. The specific value of any one of those parameters, including the relative size of Ai and Aj, required to induce i to defect will depend on the values of all the others, but this inequality generally implies that the incentive for player i to defect strengthens as: 1) i’s distaste for an autocracy led by j rises (i.e., as Ai - Aj grows larger); 2) i’s capability in executing a coup (σi) increases; 3) i’s mobilizational costs decline; and 4) the expected costs of punishment to i for a failed coup decline. The effects of variation in the players’ capabilities are more complex because, as the preceding inequality suggests, they depend not only on the relative size of those parameters but also their conditional effects on each other.3 As a result, any attempt to derive specific solutions would require a series of simplifying assumptions that may not be terribly realistic, so I do not undertake that here. If an actor believes it will do better under any version of democracy than it will under any form of autocracy and believes the other players prefer democracy as well, how could it ever be tempted to try to usurp power? The logic is the same as that revealed in game theory’s archetypal Stag Hunt: coup attempts sometimes form part of a players’ strategy under these circumstances because they offer a means for that player to reduce her risk of exploitation. In my model, even when a player prefers the cooperative outcome in which all players abide by the democratic rules of the game, uncertainty about the other players’ commitment to that same goal introduces the risk of being “suckered,” and a preemptive grab for power is one way to reduce that risk. The presence of coup attempts in the strategies of the incumbent party and the military has important substantive implications if we consider the realistic possibility of uncertainty about the opposition’s actions (or, in the jargon of game theory, imperfect information). If the incumbent or the military has reason to believe the opposition is attempting to rebel, whatever their beliefs about that opposition’s preferences, they may be tempted to try a coup of their own. One of the troubles for democracy is that mobilization for electoral competition and mobilization for rebellion can look a lot alike. As a result, “normal”

The Democracy Game

33

party-building and campaigning efforts by opposition parties may be misunderstood as efforts to develop the capability to seize power by extraordinary means, and this misunderstanding can compel governments and militaries otherwise favorable to democracy to strike preemptively. This problem is likely to be especially acute early in the life of a democratic regime, where parties have not yet had the opportunity to develop reputations for cooperation that would help to overcome this dilemma. It is also likely to be more acute when elements of the opposition have a specific reputation or capacity for rebellion, as might be the case in societies attempting to recover from civil war, where the lines between militias and parties are sometimes blurred. The kind of defensive coup envisioned in this scenario appears to have occurred in Bangladesh in early 2007, when the military declared a state of emergency and installed its own “caretaker” government in response to a protracted showdown between the country’s archrival political parties, the Awami League and the Bangladesh Nationalist Party (BNP). In accordance with Bangladesh’s constitution at the time, a supposedly nonpartisan caretaker government was installed by the ruling BNP in October 2006 ahead of scheduled parliamentary elections. The opposition Awami League, however, claimed that caretaker government was not impartial, and it objected to the composition of the country’s electoral commission, which it claimed was preparing to rig the elections in favor of the BNP. Violent protests ensued, including a blockade aimed at delaying the elections. The military intervened in January when it apparently coaxed the caretaker government to declare a state of emergency and suspend the elections and then replaced the interim president with an official of its choosing. Before the January 2007 coup, the Awami League and BNP had a long history of intense and sometimes violent confrontation around elections, but the marches and clashes that preceded the declaration of a state of emergency were substantial even by Bangladesh’s standards. Put another way, each side’s accusations that its rival was attempting to usurp power—the BNP by stacking the electoral deck in its favor, the Awami League by mass uprising—seemed credible at the time. Arguably, then, the country had already reached the point where the opposition and incumbent appeared to be defecting, leaving it up to the military to decide how to respond. In the years prior to this showdown, the Bangladeshi army had become a substantial contributor to UN peacekeeping missions, and those contributions had become an important source of financing for the military’s operations and prestige for its leaders. As the election crisis worsened in early 2007, the UN suspended technical support for the upcoming elections and threatened

34

Dilemmas of Democratic Consolidation

to make future participation by the Bangladeshi army in UN peacekeeping missions conditional on the persistence of democracy in Bangladesh.4 This conditionality gave the country’s military a vested interest in blocking a switch to a civilian dictatorship—even if it had to launch a coup to do so. Argentina’s military coup of 1966 offers another example of military intervention aimed at “saving” democracy. According to O’Donnell (1973), the perpetrators of that coup saw their actions as a means to rescue democracy, not destroy it, while also protecting their own career trajectories and their organization’s interests. Where previous coups by gorila officers had a clear partisan tinge, the 1966 coup was based on the premise that the system as a whole was broken, and only the military could save it. Ironically, as O’Donnell (1973: 162) notes, the system-saving coup occurred after several years of intensive, U.S.– and French-backed military training and reorganization intended to instill a deeper professionalism within an organization that had been rent apart by politicization and fractionalization in the 1950s. Apparently, that retraining succeeded at enhancing the military’s capacity to act cohesively and may even have altered the organization’s culture in some important ways. What it clearly did not do, however, was resolve the underlying strategic tensions that were driving the cycle of putsches, planteos, and rebellions that had characterized Argentina’s politics since World War II. And, as anticipated by the theory developed here, those strategic considerations apparently trumped any short-term changes in culture and values. Scenario 2: Power Preferred

Unsurprisingly, the game-theoretic model outlined above identifies many pathways to attempted coups and rebellions when we alter the expected payoffs to indicate that the players prefer political power to democracy as such. For this scenario, we start by assuming that both of the political parties’ first preference is to lead an authoritarian regime, allowing them to set state policy as they see fit. Their second preference is to form the government in a democracy, where they still hold power, but that power is constrained by constitutional and institutional checks on state authority. If they cannot rule, they would rather live in a democracy led by their political rival, because that arrangement will constrain their rival’s use of power and give them institutional channels to shape policy outcomes. If they must live in an authoritarian regime led by someone else, they would prefer that it be the military, leaving an autocracy led by their political rival as the least-preferred outcome for

The Democracy Game

35

both parties. For its part, the military is also assumed to prefer to rule directly through an authoritarian regime. If it cannot rule directly, the military prefers a civilian-led authoritarian regime to democracy, because the civilian dictator’s dependence on the military’s coercive power for internal repression is expected to provide the military with more influence over state policy than it would have in a democracy. In sum, we assume the following preference orderings: Incumbent: Ai > Di > Do > Am > Ao Opposition: Ao > Do > Di > Am > Ai Military: Am > (Ai ≈ Ao) > (Di ≈ Do)

To see what this means for the risk of coups and rebellions, we can start by considering the simplest situation, in which player i believes both of the other two players will choose to cooperate. Under these circumstances and recalling the simplifications shown earlier, we can say that i will defect if the marginal gains associated with a successful coup, discounted by the probability that the coup will succeed, are greater than the expected costs of collective action and penalty for trying and failing. Algebraically, this means player i will defect when: [(σi)(Ai – D)] – [(1 – σi)(κi)] – µi > 0

With this preference ordering, (Ai – D) will always be a positive number, so i’s best course of action will be determined by the specific values of the other parameters. In general, though, we can see that the incentive to defect strengthens with increases in the size of i’s marginal gains from ruling as an autocrat compared with continued democracy (i.e., as Ai – D grows larger), and with increases in the odds that its coup attempt would succeed (σi). At the same time, we can also see that the incentive for i to defect weakens with increases in its mobilizational costs (µi) and the expected cost of punishment for a failed coup (κi). Because D depends, in part, on the outcome of the next election, we can see that this parameter (p) will also factor into i’s calculations, as will the relative value of the two democracy outcomes (Di and Do). The larger the difference between Di and Do, the stronger the effect p will have on i’s expected payoffs. For either party, though, the incentive to defect strengthens as its prospects in the next election decline, other things being equal.

36

Dilemmas of Democratic Consolidation

If player i expects one of the other two players to defect, we can again lean on the simplification derived above to see that player i will choose to defect if: (σi′)(Ai) – [(1 – σi′ – σj′)(κi)] – µi > (σj – σj′)(Aj)

In a “power hungry” scenario, this inequality pits i’s most preferred outcome (an autocracy it leads) against one of its two least preferred outcomes (an autocracy led either by the military or by its political rival), so the temptation to try a coup as well will presumably be strong. That temptation can be offset, however, by high mobilizational costs, a high cost of punishment, or a disparity in i’s coup capabilities relative to j’s. Of these several constraints, the costs of collective action will have the most direct impact on i’s expected payoffs, because those costs will be incurred whenever i defects, no matter what the outcome. When player i believes both of the other players are going to defect, it will choose to attempt coup of its own if: (σi″)(Ai) + (σj″)(Aj) + (σk″)(Ak) + {(1 – σi″ – σj″ – σk″) × [(p)(Di) + (1 – p)(Do) – κi]} – µi > (σj′)(Aj) + (σk′)(Ak) + {(1 – σj′ – σk′) × [(p)(Di) + (1 – p)(Do)]}

By grouping some of the terms, we arrive at: (σi″)(Ai) – [(1 – σi″ – σj″ – σk″)(κi)] – µi > (σj′ – σj″)(Aj) + (σk′ – σk″)(Ak) + {[(1 – σj′ – σk′) – (1 – σi″ – σj″ – σk″)] × [( p)(Di) + (1 – p)(Do)]}

The inequality now includes more terms than it did when only one other player was expected to defect, but the basic issues are generally the same: player i must weigh the expected benefits of ruling as a dictator (discounted by the probability of his success in the face of coup attempts by both of the other players) against the expected costs of collective action and punishment and the payoffs he would receive under each of the other four outcomes (discounted by the marginal probabilities that each of these outcomes will come to pass). The added twist is that player i now must also consider how he would fare if democracy survives, a calculation that depends both on the size of the expected payoffs associated with victories by either party (Di and Do) and on expectations about the outcome of the next election. For either of the political parties, the incentive to defect strengthens as its electoral prospects decline or its expected payoff under an elected

The Democracy Game

37

government led by its rival shrinks, other things being equal. For a military indifferent to which party governs, the chief question will be the degree to which it prefers power to democracy, and the outcome of the next election will be irrelevant. If the military has a political favorite, though, it will also need to consider the likely outcome of the next election. In contrast to the first scenario, which produced a unique subgameperfect Nash equilibrium favoring the survival of democracy under conditions of complete and perfect information, all combinations of actions are possible in this scenario, depending on the values assigned to the model’s other parameters. This result confirms the intuition that democracy will be inherently susceptible to breakdown when the relevant actors see control of state power as the means to their best payoffs and believe their rivals might have similar preferences. The logic here is essentially the same as that found in game theory’s familiar Prisoner’s Dilemma. As long as players believe they will get their best payoffs from ruling as autocrats, they will be tempted to try a coup. They might prefer losing elections in democracy to living an autocracy led by their rival or the military, but the fear that other players might attempt usurpations of their own actually strengthens the incentive to defect for actors who believe their own attempts can diminish the chances that their rivals’ attempts will succeed. The only real constraints on defection under these conditions are the expected costs of collective action and punishment, with the former usually looming larger than the latter because of its direct impact on a player’s welfare. Credible Commitments

When all of the relevant actors do not obviously prefer that democracy survives, those actors’ beliefs about the other actors’ preferences and strategies will often be malleable. In these circumstances, democracy’s survival will often depend on the deliberate development of credible commitments to forego grabs for power. In order to be credible, these commitments must be costly to undo or ignore. In formal terms, this means that, for each actor involved, the expected gains from adhering to such a commitment minus the costs of ignoring or dismantling it must exceed the expected gains from pursuing the strategy those commitments were meant to discourage. In order to affect other players’ strategies, those commitments also must be communicated clearly. If the other actors are not aware of a commitment, it cannot affect their beliefs, and their beliefs are what determine their choice of strategies. Finally, non-compliance with the terms of the commitment must be readily

38

Dilemmas of Democratic Consolidation

observable. In other words, the players must be able to monitor each others’ adherence to the terms of the commitments they have made. Commitments that are not costly to break, are not communicated, or are difficult to monitor will generally be ignored. Any discussion of ways in which credible commitments might support the survival of democracy raises the subject of pacts. In the literature on comparative democratization, the term “pact” is generally used to describe a bargain struck among rival elites during a transition to democracy that delimits the legitimate uses of political authority in exchange for the distribution of specific benefits to the actors who participate.5 The benefits in question are usually understood to be proscriptive commitments not to threaten the vital interests of the parties involved. Put another way, the key benefits are usually freedoms from harm, not specific material incentives or permissions to do new things. In this sense, a pact is seen as a set of promises meant to reduce the relevant actors’ uncertainty about the harms that democracy might cause them. As O’Donnell and Schmitter (1986: 38) put it, “At the core of a pact lies a negotiated compromise under which actors agree to forego or underutilize their capacity to harm each other by extending guarantees not to threaten each others’ corporate autonomies or vital interests.” So, why do some pacts succeed while others fall apart? Much of the work on pacts in the field of comparative democratization has been vague on this point. The core theorizing on the subject (O’Donnell and Schmitter 1986, Karl 1986) implies that pacts work because they promise key players protection from harm, but it says little about what gives those promises the force they need to induce cooperation. Stepan (as quoted in Linz and Stepan 1996: 61) frames the answer to that question as a two-level collective action problem, positing that successful pacts “have two requirements: first, leaders with the organizational and ideological capacity to negotiate a grand coalition among themselves; second, the allegiance of their political followers to the terms of the pact.” These requirements are ultimately more descriptive than they are explanatory, however, in that they beg questions about how leaders acquire that capacity to coalesce and why political followers would give their allegiance. Weingast (1997: 258) makes more headway on the matter by using game theory to reframe the problem as a matter of incentives. He shows that pacts succeed when they are self-enforcing, and to be self-enforcing, pacts must do two things. First and most obvious, parties to the bargain must believe they are better off under it than not. Second, “elites and their followers must be willing to punish those who seek unilateral defections from the pact.” These two conditions are sometimes difficult

The Democracy Game

39

to distinguish in practice—after all, why would a party to a pact be willing to suffer the cost of punishing a defector unless he thought he would realize a net gain as a result?—but the distinction does highlight the importance of gains from cooperation and the need for credible threats of punishment for actors who violate the terms of the pact. My model contributes to this literature by further clarifying what pacts must accomplish in order to facilitate the survival of democracy. Weingast (1997) emphasizes the need for pacts to be self-enforcing, and this book’s model supports that claim. Pacts are basically promises, and game theory shows us that promises are only credible insofar as others believe it is in the self-interest of the promise-maker to abide by that commitment. At the same time, the concept of self-enforcement can be a bit vague, and the model developed here helps make the idea of selfenforcement more specific for this particular problem by spotlighting more factors the relevant actors might consider when weighing the benefits and costs of violating a pact’s terms. When compliance with the terms of a commitment is not obviously aligned with an actor’s self-interest—in other words, when the commitment is not self-enforcing—the commitment can still be made credible by vesting monitoring and enforcement capacity in some third party not susceptible to the same temptations. For this arrangement to work, that third party must itself be regarded as credible and capable. If that third party is a domestic organization or agency, it must be reliably rewarded for behaviors that impede transgressions, such as running a clean election or ruling impartially on legal matters, and must be insulated from partisan manipulation. If it is an outside actor, such as a foreign government or international organization, the players must believe that it is in that outside actor’s self-interest to enforce the terms of the deal. As Weingast (1997: 258-59) observes, however, these conditions are not easy to meet, and as a result self-enforcing pacts may sometimes be impossible to devise. In contrast to the traditional view of pacts as proscriptive arrangements, my model also suggests that these bargains need not center on promises to forego action. Instead, the focus on relative gains implies that successful pacts can also center on promises to collaborate. In other words, pacts could be about realizing gains from cooperation as well as reducing threats.6 O’Donnell and Schmitter implicitly acknowledge this point when they argue (1986: 38) that conditions are ripest for pacts when “conflicting or competing groups are interdependent, in that they can neither do without each other nor unilaterally impose their preferred solution on each other if they are to satisfy their respective divergent interests.” Still, my model helps to

40

Dilemmas of Democratic Consolidation

clarify why that is true. Interdependence implies that there are gains to be had from cooperation, and an inability to impose outcomes implies that there is a rough balance of coercive strength. Under these conditions, a forceful grab for power is unlikely to succeed, and if it does, it is unlikely to produce significant gains in welfare. As a result, actors have an incentive to reassure rivals of their desire to cooperate in order to realize gains that are only available under democracy. Of course, opportunities for gains from cooperation will not always exist. When they do exist, expected gains from cooperation may be too small to outweigh the temptation to chase the spoils of authoritarian rule. Attempts to commit to sharing those gains will also be subject to concerns about cheating and the need for a reliable and efficient means to distribute those gains.7 Nevertheless, my model clearly shows that the promise of mutual gains from cooperation can serve as the foundation for pact-making in democracies. Finally and perhaps most novel, my model also shows how these commitments need not be “pacts” in the traditional sense in order to improve the prospects for democracy’s survival. As discussed above, scholars have traditionally thought of pacts as bargains painstakingly and explicitly negotiated by elites representing rival interests. To some observers, the process of negotiation is the real reason pacts have any effect, in that the act of negotiating can kick-start a virtuous cycle of evolution in norms and habits towards collaboration and compromise (Levine 1978). Intriguingly, though, there might also be situations in which the prospects for democratic consolidation are improved because a political party or the military ties its own hands against a usurpation of power without ever negotiating at all. Why would any self-interested organization do this? Schelling (1980: 22) answers this question as follows: The essence of [one class of bargaining] tactics is some voluntary but irreversible sacrifice of freedom of choice. They rest on the paradox that the power to constrain an adversary may depend on the power to bind oneself; that, in bargaining, weakness is often strength, freedom may be freedom to capitulate, and to burn bridges behind one may suffice to undo an opponent.

My model suggests that players in democratic politics might engage in this kind of hand-tying as a way to affect rivals’ beliefs about their preferences and strategies in order to entice their cooperation when there are gains to be realized from it. Those gains might depend on the

The Democracy Game

41

ensuing cooperation, but they might also come because the organization that ties its own hands thereby frees itself from a mobilizational “arms race” with its rivals. What might those commitments look like in the real world? Generally speaking, we would expect to see actors investing in institutions or actions that limit their own ability to carry out a coup or rebellion or increase the expected cost of punishment. For incumbents, these commitments might take the form of investment in a more effective judiciary, the establishment of a more powerful and more independent electoral management body, and the adoption of laws imposing tougher penalties for elections-related transgressions. For opposition parties in post-conflict situations, they might include participation in disarmament or demobilization programs that reduce their own capacity to rebel. El Salvador offers another interesting case in point. In her work on the transition to democracy and end of civil war in El Salvador in the early 1990s, Elisabeth Wood (1995) spotlights a pact committing all three sides to reducing their capability to execute a coup and establishing an independent body of international and domestic observers to monitor compliance. The decision to negotiate this pact depended on the experience of civil war and the parity of political forces on the two warring sides. With that as the status quo, all sides believed they could be better off under some kind of power-sharing arrangement, which would not obviously entail a tyranny of the majority and would improve material gains for all. The impartiality of the body tasked with monitoring compliance ensured that the parties to the pact would have good information about their rivals’ strategies, reducing the risk that those rivals would be able to renege by surprise. So far, the pact has stuck, and El Salvador remains a democracy today. Elections have produced transfers of power without violence, and elected governments have purged and overhauled the country’s military without triggering a coup d’etat. How has this happened? Contrary to conventional thinking on the sources of democratic stability, the survival of democracy in El Salvador cannot be attributed to popular legitimacy and regime performance; after all, this is a country where poverty remains endemic, violent crime is a significant and growing concern, and voter turnout in national elections is consistently low (Wood 2005). Remarking on El Salvador, Mozambique, and several other cases from the second half of the twentieth century, Wantchekon (2004) argues that stalemated civil wars make conditions propitious for democracy precisely because they reveal the high costs of violent conflict and demonstrate the inability of the warring parties to impose their preferred

42

Dilemmas of Democratic Consolidation

outcomes on each other. “Post–civil war democratization is primarily motivated by the need for political order,” he writes (p. 18). Citizens like it because it provides protection from predation and allows them to change governments peacefully, and warring factions (akin to the rival parties in my model) like it because it induces citizens to invest in productive economic activity—activity that produces a positive externality on which their own well-being depends. Consistent with this book’s model, though, this logic only holds when wealth depends on citizens’ investment. In situations where wealth is derived from things like natural resources or foreign aid, the gains are more likely to be zerosum, so the incentive to keep fighting in pursuit of outright control over those resources will be stronger. In sum, pacts represent one means by which the players in a democracy game might credibly commit to forego coups and rebellion in an attempt to realize mutual gains that depend on cooperation. Of course, the flip side of a story in which successful pacts are understood as credible promises made in pursuit of gains from cooperation is that pacts (or successful ones, anyway) will not always be possible. This is true because the expected payoffs from cooperation will not always be larger than the ones from defection, and even when they are, the actors involved will not always succeed in finding ways to make their agreements self-enforcing (Weingast 1997: 258–259). Horizontal Accountability

The inequalities sketched above suggest that fear of costly penalties for failed coups or rebellions can deter those attempts, even when players prefer to rule as autocrats. This constraint can bind if the marginal benefits of holding power are sufficiently small, if the expected penalty itself is sufficiently large, or—and this is probably the most interesting possibility—if that penalty is expected to be incurred as long as democracy survives, regardless of which party wins the next election. When punishment is contingent on the outcome of the next election, it imposes a weaker constraint on the actors’ behavior because it is discounted by both the probability of winning the next election and the probability that the coup attempt will succeed. When punishment is likely to be incurred under any democratic government, it is only discounted by the probability that the coup will succeed, so it weighs more heavily against trying to usurp power. This insight has important real-world implications if we think of the expected costs of punishment as a function of the strength of mechanisms for horizontal accountability—that is, routine checks on the

The Democracy Game

43

exercise of state power. According to O’Donnell (1998: 117), horizontal accountability “depends on the existence of state agencies that are legally empowered—and factually willing and able—to take actions ranging from routine oversight to criminal sanctions or impeachment in relation to possibly unlawful actions or omissions by other agents or agencies of the state.” In contemporary democracies, the most relevant agent of horizontal accountability is the judiciary. Horizontal accountability can also flow from the work of election management bodies and of non-governmental organizations (NGOs) who cannot directly punish transgressions but often play a critical role in monitoring compliance. Most important in this regard are the press and other issuespecific “watchdog” groups. When the agents of oversight and enforcement are in the pocket of the party in power, punishment for transgressions of the democratic rules of the game by the incumbents will usually hinge on a loss of office. Under these conditions, a vicious cycle can emerge: election winners can expect to get away with rule-breaking as long as they keep winning, and they can improve the odds that they will keep winning by breaking the rules. Thus, incumbents will generally be less constrained by the threat of that punishment, even when the threatened penalties are substantial. The result is similar when the agents of oversight and enforcement are weak or ineffective, whatever their political loyalties. Where the judiciary and other agents of oversight are effective and independent of the party in power, however, punishment for such transgressions will not depend on the outcome of the ensuing election. Under these circumstances, the punishment constraint will bind at lower levels. Of course, the strength of this constraint will also depend on the size of the sanctions involved. When the penalties for transgressions are small, they are unlikely to have any deterrent effect, no matter how autonomous or effective the agencies that mete them out. Only large sanctions are likely to tip actors away from strategies they were otherwise determined to pursue. The capability and independence of India’s judiciary, and in recent years of its elections management body, arguably help to explain the persistence of democracy in that country, a case often regarded as an anomaly for structural theories of democratization. Mehta (2007: 70) describes India’s Supreme Court as “one of the world’s most powerful judicial bodies” and posits that this court “has by and large played a significant and even pivotal role in sustaining India’s liberal-democratic institutions and upholding the rule of law” by consistently protecting civil liberties and safeguarding the electoral process. In recent years— notably, since India’s sole reversal of democracy during The Emergency

44

Dilemmas of Democratic Consolidation

of 1975–1977—the country’s election management body appears to have developed similar strengths. According to Ganguly (2007: 34–35), “The National Election Commission (NEC), once a glaring failure at its mission of conducting elections free of intimidation and fraud, has over the past two decades become a robust and highly effective body. No government or politician dares publicly to challenge its prerogatives. Regardless of which individual holds the office of chief election commissioner, the writ of the NEC is now mostly beyond question or reproach.” Surely any full accounting of democracy’s persistence in India would have to address a host of other forces, including ethnic diversity, economic change, the role of the military, and the effects of federalism. Nevertheless, it is at least encouraging for this book’s theory that other authors have identified the efficacy and independence of the judiciary as a vital source of democratic stability in a “crucial” case such as India.8 Scenario 3: A Partisan Military

In the first two scenarios, the military was assumed to be indifferent to a choice between the two political parties; either it favored democracy or it sought political power, but it had no preference about which party ruled. For this next scenario, we will assume instead that the military does not seek political power for itself but is still “politicized” in the sense that it has a strong affinity to only one of the two rival parties. This preference could arise because the military and that party’s interests are aligned, as was sometimes the case during the Cold War, for example, when militaries supported by one superpower or the other had natural affinities with parties at their patron’s end of the political spectrum. It could also arise because one party has promised side payments to the military in exchange for its loyalty—say, a larger share of the state budget, access to lucrative rents, or immunity from prosecution for past misdeeds—that can only be delivered as long as that party controls central state authority. In any case, for purposes of this scenario we will assume that the military’s most-preferred outcome is either a democracy or an autocracy led by its ally. The military is assumed to be indifferent to the type of regime because each has its pros and cons. An autocracy gives its ally more power, but it may also tax the military more heavily as an agent of internal repression. A democratic government led by its ally will usually demand less of the military, but it will usually offer less as well. If its favored party cannot rule, the military prefers that it rule instead, through a military dictatorship. If its ally’s political rival is to rule, the

The Democracy Game

45

military prefers that it does so in a democracy, in which that rival’s power will be constrained. Finally, we assume that the military’s leastpreferred outcome is a dictatorship led by the party it does not favor. This set of assumptions producing the following preference ordering, where f is the party the military likes (friend) and e is the party it dislikes (enemy): Af ≈ Df > Am > De > Ae

The military’s position as the last mover in this game plays an important role in the results that follow. To examine the effects of a partisan military, we can start by considering its decision problem when only one of the other actors has chosen to defect: [(σm′)(Am)] – [(1 – σm′ – σj′)(κm)] – µm > (σj – σj′)(Aj)

As long as player j—the party already attempting a coup or rebellion—is the party the military favors, the military will generally prefer to cooperate in the game’s third stage, because (by assumption) Aj >> Am. It is possible that differences in the two parties’ capabilities will still give the military an incentive to defect (i.e., if σj is small and σi′ is large), but the costs of mobilization and punishment weigh against this possibility, as does the impact the military’s attempt will have on the chances that its most-preferred outcome (Aj) will be realized. If the party attempting a coup or rebellion is the one the military does not favor, however, the temptation for the military to try a coup of its own will generally be much stronger. The model’s specific prediction will depend on the expected costs of punishment for a failed military coup (k); the military’s confidence in its ability to execute a successful coup in the face of j’s attempt (σi); the marginal effect of the military coup attempt on the likelihood that j’s attempt will succeed (σj – σj′); and the expected costs to the military of mobilizing for its coup attempt. As long as the expected costs of mobilization and punishment are not especially large, however, a partisan military will have an incentive to try a coup when it believes its ally’s rival is already trying to usurp power and to stay in the barracks when it is not. If both parties have chosen to defect, we can summarize a partisan military’s decision problem by the following inequality: [(σi″)(Ai)] – [(1 – σi″ – σj″ – σk″)(κi)] – µi > [(σj′ – σj″)(Aj)] + [(σk′ – σk″)(Ak)] + {[(1 – σj′ – σk′) – (1 – σi″ – σj″ – σk″)] × [( p)(Di) + (1 – p)(Do)]}

46

Dilemmas of Democratic Consolidation

Now, the military must also consider the two parties’ relative capabilities and the outcome of the next election. The fact that (by assumption) Aj >> Ai, however, suggests the military’s incentive to attempt a coup declines as its ally’s capabilities increase. The final inequality shown above also implies that this is one scenario in which the parties’ prospects in the next election can make a significant difference in the players’ temptation to defect. Interestingly, though, those prospects still have little effect on the strategies adopted by the parties themselves. Instead, they operate here by shaping the military’s expectations about the risk that the survival of democracy will lead to one of its least-preferred outcomes—a government led by its ally’s rival—and thus the strength of its incentive to attempt a coup. The military’s incentive to try a coup is strongest when its ally is expected to abide by the democratic rules but its political rival is expected to make a grab for power. Here, the military’s position as third mover in the game becomes a distinct advantage, giving it a chance to launch a countercoup and thereby reduce the risk that it and its ally will be stuck with their least-preferred outcome. At the same time, the game also implies that a partisan military can be tempted to try a coup even when neither of the political parties defects, if its distaste for the rival party is sufficiently strong or its ally’s electoral prospects are sufficiently weak. Recognizing this temptation, the rival party may then be compelled to try a coup of its own. Once again, then, we find that a situation in which all of the actors’ most-preferred outcomes involves a democratic government can still lead to one or more attempts to usurp power. The discussion of the first scenario explored the implications of uncertainty about the opposition’s actions for the decisions of the incumbent party and the military. This scenario illuminates the implications of uncertainty about the military’s interests for the actions of the political parties. The comparative statics show that the parties might be tempted to try an executive coup or rebellion even when they are sure where the military’s loyalties lie, either because they see it as a way to reduce the risk of a less-preferred outcome or because they expect the military will not obstruct them. In a world of incomplete information, however, even parties that would prefer to see democracy survive and believe they are confronting weak rivals might be prodded to attempt a coup because they cannot rule out the possibility that the military is loyal to their rival and is therefore more likely to make its own grab for power. Of course, the model also suggests that the military is subject to similar fears and temptations and may therefore be compelled to attempt a coup of its own. The so-called Colonels’ Coup in Greece in 1967

The Democracy Game

47

offers one example of a partisan military striking to preempt an election that was expected to prove unfavorable to its political ally, even when military leaders apparently had no strong political ambitions of their own (at least initially). After Greece’s civil war ended in 1949 with the military defeat of leftist forces, a series of conservative governments held power through parliamentary elections. During this time, the threat of renewed rebellion was managed by banning the Communist party and conducting intelligence activity to guard against a leftist coup. In 1965, however, a scandal involving the prime minister’s son triggered a period of political instability that led the government to call elections for May 1967. Ahead of those elections, expectations developed that the incumbent Center Union party, whose more conservative wing had recently split from it, would win a plurality but would have to strike an alliance with leftists in order to form a new government. The threat of a political turn to the left prompted a group of vehemently anti-communist, middle-ranking army officers to seize power in April 1967, just weeks before the scheduled vote. These officers apparently had no objection to democracy as such and no essential thirst for political power. Instead, they justified their usurpation of power as a “revolution to save the nation” from a “communist conspiracy” that purportedly aimed to take control of national government by any means necessary. In terms of the formal model, the colonels apparently saw the prospect of a leftist government as a tremendous threat to their country, to the military as an organization, and thus to themselves. By early 1967, they believed the left was already attempting to usurp power, and they believed it also stood a good chance of legally taking power through the scheduled elections. Under these circumstances, the officers probably felt they had little to lose and much to gain by subverting democracy in the name of saving it. A similar process occurred in Honduras in 2009. On the morning of June 28, soldiers entered the presidential palace, disarmed the presidential guard, roused President Manuel Zelaya Rosales from his bed, and forced him to board a plane to Costa Rica, effectively sending him into exile in his pajamas. With Zelaya removed, the country’s legislature voted to install President of the Congress Roberto Micheletti as interim chief executive ahead of elections scheduled for the fall. The Organization of American States (OAS) quickly condemned these actions and insisted that President Zelaya be allowed to return to his country and his office. As of this writing in November 2009, however, negotiations over Zelaya’s return had failed to restore him to the presidency, and the interim government retained power.

48

Dilemmas of Democratic Consolidation

Zelaya’s ouster marked the first military coup in Latin America since the early 1990s, when Haitian president Bertrand Aristide was toppled soon after his election. What’s more, the Honduran coup occurred in spite of procedures adopted years earlier by the OAS and some of its member countries, including the United States, to discourage coups by credibly committing to punish countries that suffer them with diplomatic isolation and reductions in foreign aid. Predictably, the coup also imposed significant costs on the Honduran economy, even for business owners who supported it in principle.9 Why would the Honduran military choose to intervene in politics under these circumstances? In fact, the military was not the first mover in this case, and the events that preceded its intervention—along with military leaders’ own statements on the matter—help to clarify why it ocurred. Days before he was forced into exile, President Zelaya had declared that he would ignore a Constitutional Court ruling and proceed with a planned referendum to scrap the country’s term limits, allowing him to stand for office again in elections scheduled for November. The military and Zelaya’s political opponents appear to have concluded from this action that Zelaya was attempting an executive coup, and that the only way to prevent that executive coup from succeeding was to remove Zelaya from office. As one journalist describes, “It was a ‘golpe profiláctico,’ its supporters said—a coup meant to prevent whatever was coming next.” 10 The prospect of an “autogolpe” was probably not sufficient to draw the Honduran military back into politics, however. What military leaders found particularly threatening was the spectre of leftist radicalism that they saw standing behind the president’s actions. Appearing on television a month after the coup, the five generals who lead the Honduran armed forces argued that they only acted to defend democracy, not to damage it, and they specifically identified the government’s leftist bent and affinity with Venezuelan president Hugo Chávez as a crucial source of their concern. “Central America was not the objective of this communism disguised as democracy,” Gen. Miguel Ángel Garcia Padget said. “This socialism, communism, Chávismo, we could call it, was headed to the heart of the United States.”11 For an organization with deep roots in the United States’ efforts to contain Communism during the Cold War, the threat of an executive coup by a president allied with a deep-pocketed Venezuelan regime determined to export its Bolivarian revolution throughout the region was too much to stomach. Although the generals had no evident desire to rule and apparently saw democracy as their country’s best form of government, they were nonetheless compelled to perpetrate a coup by the strength of

The Democracy Game

49

their distaste for the prospect of a semi-permanent leftist regime in their own country.12 Intriguingly, the solutions explored above imply that a partisan military does not automatically diminish the chances that democracy will survive. In situations where a political party may be tempted to try to usurp power, the game implies that the military’s loyalty to the other party, coupled with its last-mover advantage, could have a deterrent effect. For the party the military dislikes, the expectation that its own coup attempt will be met with a countercoup by the military will reduce its expected probability of success and add the possibility that one of its least-preferred outcomes, a military dictatorship, will come to pass. In this situation, a partisan military can discourage that disfavored party from attempting a coup in the first place. Of course, this deterrent effect will only hold if the rival party does not already expect the military to attempt a coup of its own; otherwise, the situation described above will prevail, and the military’s partisanship might entice the opposing party into a grab for power. Thus, the effect of military partisanship on the risk of executive coup or rebellion should depend not just on how intense is the military’s preference for the one party over the other, but also on the parties’ beliefs about whether or not the military covets power for itself. Turkey’s military has arguably served as this kind of deterrent in recent years vis-à-vis the country’s strengthened Islamist movement. The Turkish armed forces have long played the role of guarantor of the secularist ideals espoused by Kemal Ataturk, founder of the Republic of Turkey, and codified in each of the country’s constitutions. Those armed forces have intervened in politics on numerous occasions, including three coups since 1950, to defend what its leaders saw at the time as the country’s national interests, including its secular foundations. At the same time, Islamists have always played a powerful role in Turkish politics, and over the decades since the last military coup in 1980 its governments and military leaders have gradually and fitfully loosened restrictions on social and political activity with an Islamist bent. The movement that both spurred and benefited from this shift eventually succeeded in taking power through the Justice and Development Party led by Recep Tayyip Erdoğan, former mayor of Istanbul and later prime minister of the Republic. As recently as the 1990s, it would have been difficult to imagine an Islamist party winning control of national government in Turkey without witnessing a more forceful response by the country’s military. Indeed, Erdoğan’s mentor and predecessor as head of the country’s leading Islamist party, Necmettin Erbakan, became prime minister as head of a

50

Dilemmas of Democratic Consolidation

coalition government in 1996, but he was ousted the next year under pressure from military leaders after he tried to carry the country into closer ties with other Muslim nations. Erdoğan himself experienced this pressure personally when he was imprisoned for five months and, at the time, banned from politics for life after appearing at a political rally in 1997 where he declared Islam to be his compass and read a line from a famous poem—“The mosques are our barracks, the domes our helmets, the minarets our bayonets and the faithful our soldiers”—that many secularists considered to be provocative. As already noted, the prospect of eventual membership in the European Union, and the economic benefits it would convey, have surely encouraged cooperation by all of the relevant actors, including not only Turkey’s military leaders but also its mainstream Islamists, by promising substantial long-term mutual gains from sustaining and deepening democracy. At the same time, the theory developed here spotlights how the military’s obvious and powerful preference for secular parties has also encouraged the Islamists to invest heavily in developing a reputation as cooperative moderates. The Islamists’ political strength—the Justice and Development Party won a sweeping victory in 2002 parliamentary elections and has only strengthened its hold on local and national government since then—clearly gives them the de jure power to rewrite the rules in their own favor. Still, recognition that aggressive grabs for power would almost certainly provoke a military crackdown or coup has given the Islamists a strong incentive to “play nice.” In the wake of the harsh rebukes of their leaders, the Islamists have successfully regrouped as a populist movement that portrays itself as more democratic than the supposedly liberal secularists. As Sontag (2003) writes, For the modernists in the Welfare Party, Erbakan’s ouster followed by Erdoğan’s conviction undeniably demonstrated that confrontation with the establishment wasn’t getting them anywhere. Fehmi Koru, columnist for an Islamic-oriented newspaper, told me: ‘When I first started writing about democracy, some members of the community criticized me openly, saying Islam and democracy were incompatible. But they grew ready for a change.’ They decided to start a new party that would aim for a broader political base. They would stop conducting politics with religious symbols and demonstrate instead how true belief informs politics wisely.

This strategy appears to have worked. So far, Justice and Development has managed to strengthen its electoral mandate without provoking a military coup or rebellion. As one might expect from an opposition

The Democracy Game

51

movement incapable for the moment of winning an election or mounting a rebellion, secularist forces have at times seemed to invite military intervention to knock Justice and Development from power. The generals have not obliged, however, probably because they know that doing so would cost them deeply through diminished support from Western allies and a setback to the country’s European ambitions. In the thick of a mass showdown in 2007 between secularists and Islamists over the election of a president, the generals made veiled threats to intervene, but those remarks were immediately derided by other Western governments. The passage (if not resolution) of that confrontation without a military intervention was widely seen as evidence that Turkey had finally turned a corner on the path away from military rule toward democratic stability, and the Islamists’ strategic embrace of a proWestern, traditionally liberal platform has played a critical role in enabling this trajectory.13 Summary

This chapter developed and then explored the implications of a gametheoretic model meant to cature the dilemmas of democratic politics described in the Introduction. The next three chapters turn to empirical evidence—first cross-national data on the survival and failure of democracies in the past half-century, then process-tracing case studies of several of those episodes—to see if that model helps us understand observed patterns and outcomes. Before turning to the evidence, however, it is worth recapitulating the model’s most important implications. •

Under conditions of complete information, if every actor expects to fare better in a democracy than in any kind of authoritarian regime (including one they lead), all of the actors will choose to forego attempts at usurping power, and democracy can be expected to survive indefinitely.

•

Under more realistic conditions of incomplete information, the model suggests that players might sometimes choose to try to usurp power even when every player prefers democracy to all forms of dictatorship, including one they lead. These patterns are sensitive to variations in the costs of mobilization, the expected costs of punishment, and the actors’ capabilities, but the point remains that, under certain (plausible) assumptions, the incumbent’s and the military’s strategies (in the game theoretic sense of a complete plan

52

Dilemmas of Democratic Consolidation

of action) could include coup attempts. •

Contrary to the view that elections are the crucial source of political uncertainty in democratic regimes, this book’s model suggests that fears of other actors’ coups and rebellions can overshadow players’ beliefs about the outcome of future elections in shaping their own decisions about whether or not to try to usurp power. Concerns about election outcomes only loom so large when fears of these other events have receded.

•

When players believe they would be best off ruling as autocrats, they will often be tempted to try a coup or rebellion. The strongest constraint against defection under these conditions is the expected cost of collective action—that is, the direct costs of mounting a coup attempt or rebellion. Players may also be deterred by the prospect of punishment if their grab for power should fail, but that constraint weighs less heavily because it is discounted by the probability of success. In situations where punishment is contingent on the outcome of the next election, this constraint will be even weaker.

•

Uncertainty about partisan loyalties in the military also affects the likelihood of attempted coups and rebellion. In a world of incomplete information, even parties that would prefer to see democracy survive and are confronting a relatively weak rival might be prodded to try to usurp power because they cannot rule out the possibility that the military is loyal to their rival and is therefore likely to attempt a coup of its own.

•

A partisan military does not automatically increase the risk of democratic breakdown, however. In situations where a political party is tempted to try to usurp power, the model implies that the military’s loyalty to the other party, coupled with its last-mover advantage, can have a deterrent effect. For the party disfavored by the military, the expectation that a grab for power will be met with a countercoup by the military reduces its expected probability of success and increases the risk that one of its least-preferred outcomes, a military dictatorship, will come to pass. These concerns may be strong enough to discourage that rival party from trying to usurp power in the first place.

•

Credible commitments to follow the democratic rules of the game can help actors overcome the fears of exploitation in the face of

The Democracy Game

53

uncertainty about other players’ choice of strategies. These commitments need not depend on disincentives for defection; they can also depend on credible promises to share gains from cooperation. Nor must these commitments be codified in a formal “pact,” although it is helpful if they are communicated clearly so they can affect the other players’ beliefs. Opportunities for gains from cooperation will not always exist, however, and even when they do, they may be too small to outweigh the temptation to try to rule as a dictator.

Notes 1 In game-theory jargon, dynamic just means that the game involves more than one move in sequence within each round of play. 2 The game’s implications do not depend on this assumption about the relationships among the players’ capabilities and the probabilities of different outcomes. This assumption is just one way to simplify the game for purposes of generating comparative statics. 3 For example, is player j’s coup attempt less likely or more likely to succeed when player i also attempts a coup, and if so, by how much? In the real world, we can imagine that the answer to this question would depend on the specific actors involved and the context in which their confrontation occurred. 4 “In Bangladesh, State of Emergency and Election Delay,” The New York Times (January 12, 2007). 5 In O’Donnell and Schmitter’s (1986: 37) terms, “A pact can be defined as an explicit, but not always publicly explicated or justified, agreement among a select set of actors which seeks to define (or, better, to redefine) rules governing the exercise of power on the basis of mutual guarantees for the ‘vital interests’ of those entering into it.” 6 This point echoes the arguments made by neoliberal institutionalists in international relations theory. See especially Keohane (1984). In essence, I am arguing that pacts are to democracy what international regimes are to international relations, in the sense understood by Keohane. When there is not a Leviathan and incentives don’t obviously favor power-sharing, the survival of democracy depends on credible commitments to forego coups and rebellion. What my model shows is that democratic procedures are not sufficient in and of themselves. What are also required are commitments not to usurp power, and these often depend on additional promises or hand-tying measures. Once such commitments are in place, exogenous change continues. Those changes can support the survival of democracy by tilting expected payoffs in favor of cooperation, but they can also undercut it. 7 Acemoglu and Robinson (2006) consider democracy itself to serve exactly this purpose—committing credibly to the redistribution of wealth from the rich to the poor by institutionalizing a policy process that favors the latter. What I have in mind here are the more mundane aspects of the problem of

54

Dilemmas of Democratic Consolidation

actually accounting for those gains and then mechanically dividing and distributing them to their agreed-upon beneficiaries. In other words, I am concerned here with the capacity to execute a bargain, not the process by which that bargain is reached. 8 I am using the term “crucial case” here as Eckstein (1975: 118) does, referring to a case “that must closely fit a theory if one is to have confidence in the theory’s validity, or conversely, must not fit equally well with any rule contrary to that proposed.” Because India is widely regarded as an outlier for structural theories of democratization, theories aiming to complement or supplant those structural approaches ought to have something useful to say about how democracy has survived there in spite of profound poverty and inequality and a highly fractious society. 9 “Honduran Businesses Still Wait to Heal,” The New York Times (November 1, 2009). 10 William Finnegan, “Letter from Honduras: An Old-Fashioned Coup,” The New Yorker (November 30, 2009), p. 38. 11 “On TV, Honduran Generals Explain Their Role in Coup,” The New York Times (August 4, 2009). 12 According to my model, credible threats from the United States government of significant sanctions in response to coups should have helped prevent President Zelaya’s forcible ouster by shrinking coup-plotters’ expected payoffs under a post-coup government. Why didn’t that threat work here? Apparently, the coup-plotters simply miscalculated, believing that threat had been tempered by mutual distaste for Venezuela’s President Chávez. According to Finnegan (op. cit., p. 39) “The golpistas were privately stunned, I was told, by the firmness of the U.S. reaction—this would never have happened if the Republicans had still been in power.” 13 In describing this approach as strategic, I do not mean to imply that the Justice and Development Party’s deepened liberalism is some kind of Trojan horse for a deeper religious agenda. I have no idea what the party leadership is “truly” aiming to accomplish.

3 How Fragile Is Democracy? Patterns in Cross-National Data

The game-theoretic model developed in the preceding chapter suggests that powerful organizations in democratic regimes often face strong incentives to try to usurp power, sometimes even when they would prefer that democracy survives. These incentives arise, in part, from strategic dilemmas that inhere in democratic politics, and recognition of those dilemmas help to explain why democratic consolidation is rare. If this view is correct, then attempts at democracy will often fail. More specifically, my model suggests that these failures are more likely to occur by military coup or executive coup than rebellion because of differences in the costs of collective action associated with the different paths to power. At odds with theories that view democratic consolidation primarily as a learning process, my model also suggests that democracies will sometimes fail even after they have survived for several election cycles and perhaps even produced a peaceful transfer of power between rival parties. Furthermore, my model implies that democracy can sometimes survive in cases where most citizens consider it to be ineffective or illegitimate, as long as the incentives of key organizations align in that direction or the expected costs of collective action and punishment for failed coups are sufficiently steep. This chapter uses an original cross-national time-series data set to see whether or not real-world patterns bear out these expectations. These data identify the occurrence and termination of episodes of democracy since 1955 in all countries with populations of 500,000 or larger. In cases where episodes of democracy ended, the data identify whether the failure occurred by military coup, executive coup, rebellion, or some other mode. After describing how the data were constructed, I use descriptive techniques from a statistical approach called event history analysis to identify empirical patterns in the survival and failure of democracy over the past half-century and then discuss how those patterns accord with the model’s general predictions.

55

56

Dilemmas of Democratic Consolidation

Data

As stated in the Introduction, I understand democracy to be a form of government in which citizens routinely choose their rulers through free, fair, and competitive elections. To determine whether or not the national political regime that exists in a particular country at a particular time meets this definition, I applied the following coding criteria to information from various secondary and tertiary sources on all countries of the world during the period 1955–2007 with populations larger than 500,000.1 To be considered a democracy, a regime had to satisfy all of the listed conditions. In other words, each of these conditions is considered necessary for a regime to be called democratic. 1. Are the officials who actually rule chosen through elections? • The head of government is chosen directly or indirectly by popular election, or he/she is the constitutionally designated successor to an elected head of government who has resigned, died, or become incapacitated while in office. • The members of the legislature are chosen by popular election. • No unelected individual, body, or organization—domestic or foreign—wields veto power across a range of national policy issue areas. 2. Are those elections competitive? • At least two independent political parties field candidates for most or all national offices, including the head of government in cases where that office is filled directly by election. • Independent news media exist and are accessible to most citizens. • Processes of voter registration and identification and lists of registered voters are not manipulated, restricted, or impeded on a large scale to partisan advantage. • State resources are not used directly and extensively in political campaigns to the advantage of incumbent officeholders. • The vote-tallying process is not subject to abuse or fraud that is widespread or sufficient to change either the balance of power in the legislature or the outcome of a direct election for head of government. 3. Is the political process broadly inclusive?

How Fragile Is Democracy?

•

• •

57

Citizens may form independent political parties or associations without substantial interference or impediment by the state. Nearly all adult citizens may stand as candidates for office. Elections are based on the principals of universal and equal suffrage.

Democracies may choose their rulers directly or indirectly, and they may use a variety of electoral systems to convert votes into outcomes. Variations in these elements establish different types of democracy, but they do not determine whether or not democracy exists. Similarly, democracies may locate the boundaries of public authority in different places and by different means. For example, some might consider private property an inviolable right, while others might emphasize personal welfare. Again, these variations might give rise to different forms of democracy, but they do not determine if it exists in the first place and therefore were not considered in the construction of these data. For purposes of my model, it was important to determine not only that an episode of democracy had occurred but also how it ended if it did. The following rules were used to classsify democratic breakdowns according to their type. Military Coup. A military coup occurs when the state’s armed forces or some faction thereof, acting on their own behalf or purportedly on behalf of the “nation,” seize control of national government. When soldiers act against their government as participants in an uprising led by the opposition or refuse orders to suppress that kind of uprising, any resulting usurpation of power is considered a rebellion (see below), not a military coup. Executive Coup. An executive coup occurs when the incumbent government terminates democratic competition by fiat or otherwise changes the formal rules or supporting practices of democracy in ways that virtually ensure that government’s continuation in office. In presidential democracies, the president and his or her party are considered the incumbent government. In parliamentary or mixed democracies, the party or coalition that selects the prime minister play this role. In some cases, actions comprising an executive coup will take the form of direct electoral fraud, declared annulments of constitutional procedures, or unilateral and extra-constitutional dismissals of an

58

Dilemmas of Democratic Consolidation

elected legislature. In others, the party in power infringes more gradually and perhaps more subtly on important enabling conditions for representation and transparency, such as rights of free speech, procedures for voter registration, or rules governing the formation and operation of political parties. Rebellion. A rebellion occurs when representatives of a group or groups that contested but lost the preceding election use extralegal means to seize control of national government or to produce a collapse of central state authority. The actions that produce a revolution can be violent, such as riots or insurgency; non-violent, such as demonstrations and strikes; or both. Situations in which the national government is toppled or supplanted by some other organization—such as a foreign army or a militia that does not claim to represent a group which contested the preceding election—are not considered rebellions and are placed instead in the “other” category (below).2 Other. Any termination of democracy that does not clearly fit into one of the aforementioned categories. Patterns

Figure 3.1 charts the prevalence of democracy worldwide over the past several decades. As the chart shows, the ranks of democracies expanded in the 1960s and again in the 1990s, two periods of prolific state creation. The expansion of democracy that accompanied decolonization in Africa around 1960 was followed by a decline in the stock of democracies that reached its nadir in the mid-1970s. That trend was then reversed, however, and the democratic share of the world’s regimes grew steadily throughout the 1980s, thanks mostly to transitions that occurred first in Latin America and then in Africa and Eurasia. By the end of 1992, more than half of all countries worldwide had some form of democratic government. The global prevalence of democratic regimes peaked in 1995 at 57 percent, but in contrast to the pattern from the 1970s it has not diminished significantly since then. At the end of 2007, 91 countries had democratic regimes according to this book’s measure, representing 55 percent of all regimes worldwide.

60

Dilemmas of Democratic Consolidation

Figure 3.2. Democratic Breakdowns by Type, 1955–2007 50

45

40

35

30

Event Count

25

20

15

10

5

ARG 62 ARG 66 ARG 76 BFO 80 BNG 07 BRA 64 CEN 03 CHL 73 COM 75 COM 99 CON 63 CON 97 CYP 74 DOM 63 ECU 61 FJI 87 FJI 06 GAM 94 GHA 72 GHA 81 GNB 03 GRC 67 HAI 91 HON 63 HON 72 LAO 60 MYA 58 MYA 62 NIG 66 NIG 83 NIR 96 PAK 58 PAK 99 PAN 68 PER 62 PER 68 ROK 61 SIE 67 SIE 97 SOM 69 SUD 58 SUD 69 SUD 89 SYR 62 THI 76 THI 91 THI 06 TUR 60 TUR 71 TUR 80 Mil. Coup

ALB 96 ARM 96 BNG 74 AZE 93 BEN 63 BOL 64 CAM 97 CRO 95 ECU 70 EQG 69 GHA 58 GRG 00 GUY 68 HAI 99 IND 75 INS 57 KEN 66 KEN 07 MAG 65 MAL 69 MLI 97 MZM 04 PAK 77 PER 92 PHI 72 RUS 93 RUS 03 SEN 63 SEN 07 SIE 71 SRI 82 UGA 66 UKR 98 URU 72 VEN 05 ZAM 68 ZAM 96 ZIM 87 Exec. Coup

BUI 93 CYP 63 ECU 00 FJI 00 GNB 98 GRG 91 IVO 02 LBR 03 LEB 75 LES 98 SOL 00 Rebellion

CZE 92 GFR 89 LES 70 NEP 60 NEP 02 PNG 02 SWA 73 SYR 58 YGS 06 Other

How Fragile Is Democracy?

65

from organizations rooted in contracts rather than coercion; the existence of a state that possesses a monopoly on the legitimate use of violence; and the ability of competing organizations to sanction members of the government through well-specified, well-understood, and non-violent mechanisms. Crucially, in an open-access order, the generation of new wealth requires cooperation and free exchange among many individuals who choose to invest their resources—time, skill, land, money—in productive activities. As Bates (2001: 26) describes, “For power to be used to produce wealth, coercion must therefore be used in new ways. Those who specialize in the use of force must refrain from violence and delegate their authority to those who will employ it productively.” These incentives encourage individuals to construct political institutions that routinize cooperation and resolve conflict by non-coercive means. By contrast, economies organized around the control of rents encourage actors to focus their efforts on domination rather than cooperation. In the modern context, this usually means trying to control the state, because state authority typically determines the distribution of those rents. Countries with economies organized around rent-seeking are less likely to democratize in the first place (Ross 2001; Ulfelder 2007), but if they do—or in cases where rents expand significantly after democracy is already in place—the pursuit of those rents encourages actors to usurp power in an attempt to monopolize them. Without the economic incentive to pursue cooperation and encourage investment in production, actors can profitably treat politics as a zero-sum game. In the language of my game-theoretic model, increases in the value of rents tied to state power can be said to increase the relative value actors assign to ruling as a dictator, and thus the temptation to try a coup or rebellion. The rents are spoils, and in authoritarian regimes those spoils go exclusively to the victor. The spoils of authoritarian rule are smaller in economies organized around profit-seeking because they depend on the actions of the individuals engaged in productive activity. Those individuals will generally produce less if their government increases taxes beyond a certain point because the marginal returns on their additional investments will have turned negative. This elasticity of production, as economists call it, limits the payoffs autocrats can expect to reap in profit-seeking economies in a way that is not true in rentseeking economies. The process of usurping power can also carry economic costs in profit-seeking economies that will further reduce the expected payoff from a coup or rebellion. Economic growth depends on investment in production, but individuals will only make those investments when they

66

Dilemmas of Democratic Consolidation

believe they will reap some benefit from it. The more concerned they are that someone will snatch those future gains from them, the weaker the incentive to invest. The disruptive institutional changes associated with coups and rebellions produce uncertainty that may discourage individuals from investing, and specific concerns about property rights and the validity of contracts under the new regime may have a similar effect.9 In other words, political volatility can impose its own costs on actors in economies where the production of wealth depends primarily on continuing investment in production.10 The effects of this kind of volatility on the potential payoffs to new dictators may have become especially deleterious in the past couple of decades, when the increased mobility of capital allows what Thomas Friedman (2000: 109) calls the Electronic Herd—the vast array of stock, bond, and currency traders that now represent one of the leading sources of capital for companies and countries alike—so quickly and so ruthlessly to punish markets that fail to provide reliable or large returns on their investments. To attract capital in this environment, most governments must adhere to an array of liberal economic and political policies Friedman terms the Golden Straightjacket. As Freidman notes (2000: 105), “As your country puts on the Golden Straightjacket, two things tend to happen: your economy grows and your politics shrinks.” Bates (2001: 107) sharpens this argument when he writes that, “The creation of limited government may not be sufficient to secure high levels of investment, much less the growth of national economies. But assurances to investors surely are necessary to secure the formation of capital,” and thus to allow economic growth to occur. These forces have weaker effects on economies organized around the exploitation of oil, metals, and other natural resources, because growth in those economies depends less on fresh investments of private capital. For most countries, however, anticipation of punishment by the market for uncertainty about property rights and policy choices now represents an important source in the political economy of regime change of the kind of positive feedback that Pierson (2004) sees as a crucial source of path dependency. In short, my model suggests that it is not only the level of national wealth but also the ways in which that wealth is produced that can affect the prospects for democratic survival. As Hardin (1997: 252) surmises, “Democracy works especially well when there is a large capital stock that is at risk if order breaks down, especially if the capital stock is spread fairly broadly through the society.” If these kinds of concerns can be expected to limit productive exchange or threaten property rights under the new regime, they will reduce the benefits of dictatorship to the ruler and may even tip the payoffs far enough to cause those actors to

How Fragile Is Democracy?

67

prefer governing in a democracy to ruling as a dictator. This connection helps to explain the often-observed correlation between economic wealth and durable democracy. Unfortunately, as North, Wallis, and Weingast observe, it does not clarify exactly how that conjunction arises, as it rarely has in human history. Nevertheless, the diverse experience of countries such as the United Kingdom, France, Germany, and the United States suggests that, to a large extent, it arises in fits and starts as individuals relentlessly pursue new gains from trade (McMillan 2002: 26). I do not mean to imply that there are only two kinds of economies, or that the possibility of gains from cooperation leads automatically to the establishment and survival of democracy. Real-world cases inevitably entail some combination of the two stylized situations described by North, Wallis, and Weingast, and democratic institutions do not arise spontaneously in situations where they might be expected to produce larger gains. Rather, the point is that the specific mix of opportunities for economic gains available in a particular democracy at a particular moment powerfully affects actors’ evaluations of the payoffs associated with different regime outcomes, and thus their incentives to pursue cooperation or revert to exploitation. In situations where markets strongly favor investment and free exchange over rent-seeking, democracy can become self-enforcing, and economic growth can become sustained. In situations where riches flow from exclusionary rather than cooperative behavior, democracy will struggle to survive, if it is attempted at all. Most often, actors will encounter a mix of the two patterns, and in these cases, the organization of politics and the rate of economic growth will usually be more volatile. Survival and International Context

National borders do not neatly contain national politics. As many scholars have observed, regional and global forces can also affect the odds that democracy will become consolidated (Levitsky and Way 2007, Schmitter 1993, Ulfelder 2008, Whitehead 1996). Circumstantial evidence of this kind of “outside-in” effect can be found in a simple comparison of survival rates for democratic regimes during and after the Cold War. As Figure 3.7 shows, democracies appear to be surviving longer since the end of that superpower rivalry.

70

Dilemmas of Democratic Consolidation

and politics of election observation missions varies widely, it seems fair to say that their near-ubiquity in recent years has contributed to the extension of some democracy’s life spans since the end of the Cold War. As discussed so far, these causal mechanisms linking foreign assistance and election observation missions to the risk of democratic breakdown would seem primarily to affect the incentives to executive coups. Why, then, did the risk of military coups decline after 1989? This pattern seems to stem, at least in part, from changes in the incentives provided by key players in the international system. During the Cold War, U.S. and Soviet governments generally put their own rivalry ahead of other foreign-policy concerns, and both considered friendly relations with developing countries to be a crucial asset in that struggle. Under these circumstances, coup plotters knew that if they managed to seize power, they could probably gain diplomatic and even material support from one or the other of the superpower rivals. In some cases, militaries in developing countries were tacitly or even actively encouraged by superpower patrons to seize power from elected governments.14 Viewed through the lens of my game-theoretic model, those policies would have increased the expected payoff to restive militaries from seizing power (in the form of material support from those external actors) and possibly also reduced the expected punishment for trying a coup but failing (due to protections extended by those external actors). The end of the US-Soviet rivalry effectively ended those policies. In so doing, it sharply altered the expected payoffs from usurpations of power for militaries in many developing countries, thereby reducing the incidence of military coups.15 Summary

This chapter has shown that most attempts at democracy end in failure, and those failures often come fairly quickly, especially in poorer parts of the world. These patterns seem all the more stark if we consider that they only represent the cases where coups or rebellions succeeded, not all of the other cases where actors tried but failed to usurp power. My data do not identify failed military coups, executive coups, and rebellions in democracies, so it is impossible for me to quantify the prevalence of those events.16 Still, cross-tabulation of the regime data used here with one independent record of military coups d’etat identifies 57 failed coup attempts in democracies during this same fifty–year period. A cursory review of cases such as Pakistan in the 1980s or Venezuela in the 1990s suggests that successful usurpations are often preceded by a series of unsuccessful attempts or various thrusts and

How Fragile Is Democracy?

71

parries that heighten the strategic uncertainty at the heart of this book’s theory. Even in well-established and apparently resilient democracies, incumbents and oppositions continue to engage in behaviors—from gerrymandering and questionable campaign finance practices to outright fraud—that represent variations on the same themes (Armony and Schamis 2005). Sadly, then, the grabs for power anticipated by my theoretical model seem to be quite common, and democratic consolidation appears to be even more difficult to attain than conventional theories recognize.

Notes 1 These data were developed by the author for the Political Instability Task Force (PITF), a research program funded by the Central Intelligence Agency (CIA). A complete list of episodes of democracy identified by these data and the ways in which those episodes ended is provided as an appendix to the book. Because these data were created for CIA, they do not include the United States. For more on PITF, see http://globalpolicy.gmu.edu/pitf/index.htm. 2 In the jargon of event history analysis, episodes that end with other forms of breakdown are treated as right-censored cases, which means they exit the risk set but do not count as a failure event. 3 Linz (1978: 15) observes the same pattern but explains it in a different way. “In modern societies,” he writes, “governments faced with such threats can generally count on the compliance of many citizens, their staff, bureaucrats, policemen, and the military if they decide to activate their commitments to legitimate authority. Therefore, disloyal oppositions have tended increasingly to avoid direct confrontation with governments and their agents and have aimed instead at combining their illegal actions with a formally legal process of transfer of power. In that process the neutrality, if not the cooperation, of the armed forces or a sector of them has become decisive. The twentieth century has seen fewer revolutions started by the populace than the nineteenth, and their fate in modern states has generally been defeat.” 4 Countries with episodes of democracy that began before 1955 or that were continuing when observation ended in 2007 are included in the calculations. The data set includes the start time for the episodes that were ongoing in 1955, and the estimation technique differentiates between cases that ended because observation ended and ones that experienced the terminal events of interest. 5 The point at which the survivor function crosses the 0.5 mark on the yaxis shows the average survival time for the observed group. 6 The studies confirming this relationship are too numerous to list, but a few recent examples include Epstein et al. (2006), Przeworski et al. (2000), Svolik (2008), and Ulfelder and Lustik (2007). 7 Per capita income is measured here in constant 1990 US dollars, as estimated by Angus Maddison. The groups represent quartiles based on global data for the period 1955–2007.

72

Dilemmas of Democratic Consolidation

8 “Rent” in this context is a technical term referring to the difference between what a factor of production is paid and what it would have to be paid to stay in its current use. It is used as a measure of market power. 9 On the economic effects of military interventions in democratic politics, the words of a mother of three in Fiji in the wake of that country’s most recent military coup provide an anecdotal illustration of this point. She told the BBC, “‘I know what we went through after the last coup. The children didn’t go to school, the prices of goods went up, our husbands’ pay was cut off. We don’t want to go through that again,’ she recalled gloomily.” See BBC Online, “Fijians Divided Over Coup Outlook” (December 5, 2006). 10 Using a formal model similar to the one developed here but with only two players instead of three, Przeworski reaches a similar conclusion. He writes (2005: 265), “As income increases, the gap between the well-being of electoral losers and of people oppressed by a dictatorship becomes large. The stakes are too high to risk losing the income guaranteed under democracy.” 11 For the OECD data, see their International Development Statistics (IDS) online database at http://www.oecd.org/dataoecd/50/17/5037721.htm. 12 On the point about the ambiguity of causal mechanisms, see also Improving Democracy Assistance: Building Knowledge through Evaluations and Research, a 2008 study funded by USAID and undertaken by the National Academy of Sciences’ National Research Council. The report is available online at http://www.nap.edu/catalog.php?record_id=12164. 13 Ahead of parliamentary elections in 2007, the Russian government effectively uninvited a mission from the Organization for Security and Cooperation in Europe (OSCE) that was slated to observe that balloting. 14 Support from the U.S. government for coup plotters in Guatemala in 1954 and Chile in 1973 is a matter of public record. See The Cold War Museum at www.coldwar.org. 15 In research conducted independent of the work reported here, Morrison (2008) used similar categories and observed the same general pattern and the same decline in military coups since the end of the Cold War. On the latter trend in particular, see also Marinov and Goemans (2008). 16 Existing data on failed military coups, insurgencies, and popular protest are probably sufficient to generate at least a preliminary read on the prevalence of attempted military coups and rebellions. What’s lacking at the moment are data on attempted but failed executive coups.

4 How Does Democracy Break Down? Evidence from Random Narratives

The patterns identified in the preceding chapter appear at least circumstantially to confirm a few broader hypotheses arising from this book’s theoretical model. We saw, for example, that most attempts at democracy are short lived, as we would expect in a world where incentives to usurp power abound. Contrary to the view that popular legitimacy is the source of regime survival, we also saw that democracy is far more likely to fall as a result of coups by elected incumbents or the military than it is to fall to rebellious citizens. The analysis also confirmed once more that democracy is more likely to survive when it arises in wealthier countries, and it showed that democracies are surviving longer on average since the end of the Cold War—both patterns that can be explained by examining the actors and incentives emphasized in this book’s theoretical model. Ultimately, though, the theory developed in this book is really meant to help us understand the process by which democracies survive or fail. More specifically, this theory is meant to help us see which organizations may be tempted to try to usurp power under what conditions and why. To properly test a process-oriented theory like this one, we need to examine real-world attempts at democracy to see whether the actors, incentives, and beliefs spotlighted here actually have the kinds of effects the model suggests they should. In other words, we need to examine the process of democratic survival or failure in specific cases to see if it resembles the “game” posited here and produces outcomes that are consistent with the model’s implications (George and Bennett 2004). This chapter and the one that follow take up that task, using two different logics to select the cases for closer scrutiny.

73

74

Dilemmas of Democratic Consolidation

Methodology

To test theories founded on formal models, Bates et al. (1998) advocate that scholars use those models to try to explain real-world events by constructing “analytic narratives.” As the authors use the term, analytic narratives are not just rich descriptions of particular historical events or processes. Instead, they are explanatory narratives focused specifically on the actors, interests, and incentives identified in the theoretical model, and the mechanisms by which those interests and incentives are translated into real-world choices. The point of an analytic narrative is to see whether or not the elements identified the model can lead us to a coherent and compelling explanation of an observed outcome. More specifically, Bates et al. (1998: 14–16) propose several questions we can apply to an analytic narrative in order to test the theory on which it is based. The studies that follow emphasize three of those questions: 1) Do the assumptions underpinning the formal model fit the observed facts of the case? 2) Do the model’s core implications find confirmation in the historical record? And 3) How well does the model do by comparison with explanations rooted in other theories? To select a manageable number cases for closer study while preserving some ability to generalize from those cases, I employed the “random narrative” strategy proposed by Fearon and Laitin (2008). As its name suggests, this strategy requires the researcher to select cases at random from the larger sample to which the theory is meant to apply. According to Fearon and Laitin, randomizing the selection of cases for closer scrutiny guards against bias in the inferences that ensue. When researchers choose cases because those cases are familiar or appear to be more relevant, they risk introducing known or unknown selection biases that can skew the ensuing findings. Randomization avoids this problem, and in so doing it provides a sounder basis for making general inferences. It also increases the likelihood of discovering new variables or mechanisms that the original theorizing failed to consider—in a word, of surprises. To select cases for this study, I applied the random-narrative approach to a list, stratified by observed outcome, of all episodes of democracy in the past half-century identified in my data set. The theoretical model envisions four possible outcomes for any democratic episode: democracy survives, or it breaks down by military coup, executive coup, or rebellion. By stratifying the selection according to those outcomes, we can ensure that each of the major pathways to democratic failure is considered, as is democratic survival.1 At the same

How Does Democracy Break Down?

75

time, by treating the entire episode of democracy as the unit of analysis, we also allow for variation in outcomes over time within each case, thereby expanding the range of comparisons we can make. The four cases selected for this study using this approch were Fiji, 2001–2006 (military coup); Ukraine, 1994–1998 (executive coup); Cyprus, 1960–1963 (rebellion); and Spain since 1977 (regime survival). Each of the analytic narratives that follow briefly describes the events marking the onset and (if relevant) termination of democracy for the selected episode; establishes initial conditions, including the identity and apparent interests of the relevant actors; describes the politics that ensued with special attention to the capabilities and beliefs emphasized in the model; and, finally, discusses how this interpretation of events compares with ones suggested by other theories of democratization. After all of the cases have been examined, the chapter concludes with an assessment of the theory’s plausibility and relevance in light of the accumulated evidence. Breakdown by Military Coup: Fiji 2001–2006

To examine the strategic considerations behind the 2006 military coup that terminated Fiji’s third episode of democracy, it helps to walk briefly through the politics of the country’s previous two democratic breakdowns. Although they are not directly part of the case under consideration, those prior breakdowns help to establish the players and provide insights into their interests and strategic beliefs. Along they way, they also show that the dilemmas and behaviors emphasized by this book’s theoretical model have been a regular feature of politics in this small Pacific island nation for decades.2 Fiji’s first attempt at democracy lasted 17 years: from 1970, when the country gained its independence from the United Kingdom, until shortly after parliamentary elections in April 1987. Political organization around a pair of communal identities—indigenous Fijians on the one hand, descendants of Indian laborers called Indo-Fijians on the other— has played an important role in each of Fiji’s democratic breakdowns, including this first one and the third one on which this narrative will eventually focus. In the 1987 elections, the Fiji Labor Party—widely considered a vehicle for Indo-Fijian interests—outpolled the incumbent Alliance Party and formed a coalition government with a smaller third party. That defeat marked the first time since independence that the Alliance Party, associated with the indigenous Fijian community, had failed to win a national election. Fiji’s army overthrew the new government the next month, allegedly with tacit approval from Alliance

76

Dilemmas of Democratic Consolidation

Party leader Ratu Sir Kamisese Mara. In this book’s language, Fiji’s first episode of democracy ended with a military coup, and that coup was sparked by the election victory of a party that was the enemy of that military’s friend. Following negotiations to secure constitutional changes that would bolster the political power of indigenous Fijians, Brigadier Sitiveni Rabuka on 25 September tossed out the Advisory Council that was supposed to guide the country back to democracy and replaced it with a military council under his own leadership. The country returned to civilian rule that December, when Brigadier Rabuka stepped aside as head of state and became minister of Home Affairs in an interim government in which Alliance Party leader Mara was reinstated as prime minister. A new constitution promulgated in July 1990 established Fiji as a republic with a bicameral legislature. The new constitution also guaranteed indigenous Fijians a majority of seats in both houses and included a controversial clause that was criticized by some as giving the army authority to stage further coups. Competitive elections were finally held again in 1992 and won, unsurprisingly, by the indigenous-led party, by-then reconstituted as the Fijian Political Party and better known as SVT for the initials of its Fijian name. In 1997, still smarting from economic decline and capital flight in the wake of the 1987 coup, Fiji’s SVT-led government adopted a new constitution that was intended to promote ethnic inclusion and respect for human rights (Lal 2000). After the 1987 coup and the rioting and looting that accompanied it, Indo-Fijians had begun emigrating from the island at a rapid clip, leading to a demographic shift in which indigenous Fijians came to comprise a majority of the national population by the mid-1990s. In spite of its dwindling voter base, however, the Indo-Fijian led Fiji Labor Party still managed to win thirty-seven of seventy-one seats in parliamentary elections in 1999, thanks in part to a change in the voting system mandated by the 1997 constitution (Fraenkel 2001). On 19 May 2000, a band of armed indigenous Fijian civilians led by George Speight responded to the SVT’s election loss by storming parliament and seizing several hostages, including Indo-Fijian Prime Minister Mahendra Chaudhry. Saying he had seized power on behalf of all indigenous Fijians, Speight declared himself president, appointed an opposition lawmaker as prime minister, and suspended the constitution. Ten days later, the country’s armed forces, led by Commodore Frank Bainimarama, declared martial law and assumed control of the country. The hostage crisis ended in July when the gunmen finally surrendered. Instead of restoring the government elected in 1999, the military decided to dissolve parliament and set up an interim government under ethnic

How Does Democracy Break Down?

77

Fijian Laisenia Qarase. This government promptly established a committee tasked with revising the 1997 Constitution in order to guarantee indigenous Fijians’ control of government. The installation of the Qarase government marked the end of Fiji’s second democratic episode. Once again, the breakdown was triggered by the electoral loss of an indigenous Fijian party to the country’s predominantly Indo-Fijian opposition. This time, though, the election losers first attempted a rebellion on their own, and the military only stepped into an openly political role after that attempted rebellion led to a protracted standoff that paralyzed and embarrassed the country. Nevertheless, the outcome was largely the same: a coerced return to government led by indigenous Fijians. The move back toward elected government began in November 2000, when Fiji’s top court deemed the military-installed interim government unconstitutional. The following March, the High Court upheld its decision and the key players chose not to resist. That ruling left the government in the hands of the president and the Great Council of Chiefs, a traditional indigenous Fijian body, and Laisenia Qarase was named head of an interim government soon thereafter. Fresh elections were held in August and September under the terms of the 1997 Constitution, and Qarase’s United Fiji Party (SDL) won a plurality of seats. The SDL subsequently formed a ruling coalition with the Conservative Alliance Party (CAP), and Qarase was again named prime minister. Although Chaudhry’s Labour Party accused the victors of fraud, election monitors from the United Nations, the Commonwealth, and the European Union deemed the balloting free and fair. The newly elected parliament was seated in October 2001, marking the start of the country’s third episode of democracy and the case selected at random for this analysis. That episode terminated in a bloodless military coup five years later, several months after parliamentary elections had extended the incumbent, SDL-led government’s mandate. After threatening for more than a year to oust that government if did not accede to a variety of demands, Commodore Bainimarama—the same commodore who had declared martial law and temporarily assumed control of the country in the midst of the parliamentary hostage crisis several years earlier—finally did so on 5 December. Bainimarama said he would eventually allow a return to elected civilian rule, but he did not immediately set a timetable for elections. As of late 2009, after several false statrts toward fresh elections, Bainimarama remained in power as Fiji’s self-appointed leader.

78

Dilemmas of Democratic Consolidation

What explains why Bainimarama led Fiji’s military to its third coup in twenty years, and why did he act when he did? Relations between indigenous Fijians and Indo-Fijians have long served as an important axis of national politics in Fiji, and accounts of the country’s several coups and coup attempts usually point to this communal divide—and the conflict of socio-economic interests associated with it—as a crucial structural problem inhibiting the consolidation of democracy in Fiji. Most of the country’s Indo-Fijians are descendants of Indian laborers brought to the country during British rule to work on its sugar plantations, then and still an important source of income. Although Indo-Fijians comprised a majority of the country’s population at independence,3 indigenous Fijians have retained ownership of most of the country’s profitable land and have traditionally dominated national politics. Tensions between these two communities over political power and economic distribution clearly contributed to previous breakdowns of democracy. Consistent with Acemoglu and Robinson’s (2006) model, in which coups occur when wealthier citizens come to view democracy as too costly to themselves, both of Fiji’s previous failures of democracy occurred shortly after the country’s traditional indigenous-Fijian elite lost power to coalitions led by leftist parties advocating what might be regarded as Indo-Fijian interests. What is vexing about the 2006 coup for Acemoglu and Robinson’s model is that the military in this instance acted against the country’s entrenched elite, not for it. Boix’s (2003) model, which also portrays conflicts over the redistribution of wealth as the engine of regime change, seems to leave room for this kind of coup by including the possibility of democratic breakdown by leftist revolution. Here again, though, the model does not seem to fit the case of Fiji’s 2006 coup, where the military apparently neither coordinated its actions with nor acted on behalf of the leftist parties that sought redistributive policies. Instead, Bainimarama’s demands ahead of the coup centered on allegations of official corruption and concerns about lenient treatment for soldiers who mutinied in 2000 during the two-month standoff with George Speight and his collaborators.4 The disconnect between Bainimarama’s apparent motives in perpetrating the 2006 military coup and the interests of the country’s Indo-Fijian community is just as problematic for theories that seek to explain Fiji’s failure to consolidate democracy as a consequence of entrenched ethnic rivalry. By 2001, demographic changes associated with the 1987 coup and its aftermath had effectively quashed the threat of Indo-Fijian domination, and changes to the constitution and electoral system were encouraging the development of new parties, some of

How Does Democracy Break Down?

79

which were forging multiethnic coalitions. According to Lal (2000: 2324), the campaign surrounding the 1999 elections “was the most relaxed in living memory…For once, race was relegated to the background because both coalitions were multiracial.” Instead, Fijian politics had shifted from competition between indigenous and Indo-Fijians to competition among rival factions within the indigenous community. “The notion of indigenous Fijian political unity was a social construction which emerged after the establishment of the colonial state in 1874,” Durutalo (2000: 73) asserts—a construction that was reified through policies of indirect rule and institutionalized in various ways in the 1970 constitution, included an electoral system based on communal organization. At the same time, this construct of unity masked a historical rivalry within the indigenous population between eastern and western Fijians and, later, emerging conflicts based on divergent socio-economic interests. In the regional rivalry, eastern chiefs traditionally dominated national political authority, including favoritism for their clans in employment in local administration and civil service, and western Fiji was more independent-minded, having only been conquered at the time of colonization. This independent streak was evident in the formation of the Western United Front party in the late 1970s, when some western Fijian landowners clashed with the ruling Alliance Party government over payment of a forest royalty that led the westerners to claim their resources were being taken without adequate compensation. According to Durutalo (2000: 76), “The western Fijian chiefs and their people saw their politico-economic marginalization from the colonial to the post-colonial era, as similar to those [sic] of the descendants of the indentured laborers.” Meanwhile, the emergence of the Fiji Labor Party in the mid1980s—a development triggered in part by austerity measures mandated by the International Monetary Fund (IMF)—marked a sharper turn toward politics based on economic rather than purported racial interests. Although it is sometimes described as a vehicle for the interests of the Indo-Fijian community, the Labor Party has been multiethnic and has promoted a leftist rather than racial agenda from the start. This shift away from racial politics was accelerated by the 1987 and 2000 coups. Of these events, Durutalo (2000: 78) writes that, “The coups themselves have been a catalyst for rapid social change amongst indigenous Fijians, unleashing many forms of intra-societal rivalry.” Commodore Bainimarama apparently shared this perception; a 1 January 2001 article from the Suva Post quotes him as saying, “There is a lot of animosity towards one another among the indigenous people. After 19 May [2000, the start of the failed coup attempt led by George Speight], all of a

80

Dilemmas of Democratic Consolidation

sudden the Fijians started fighting each other. The question is what were they fighting for and who was fighting? People were fighting for power and positions.” To account for the path democracy took in Fiji after 2001, we need to understand not only the occurrence and timing of the 2006 military coup, but also the absence of an executive coup or rebellion in the five years before that event. To do that, this book’s model prods us to consider the expectations of Fiji’s partisan rivals and its military about their well-being under different political futures, their ability to usurp power should they try to do so, and the fear that others might seize power first. Let us start with the opposition. Whatever the Indo-Fijians, urbanites, and other segments of the constituency that lost the 2001 and 2006 elections might have expected regarding their prospects under future versions of democracy or autocracy, there was little prospect of a rebellion during that time because this constituency was so badly outgunned. The reactionary violence of 1987 and 2000 and the demographic changes that ensued made clear that the odds a rebellion would succeed were slim at best, and occasional threats of violence from indigenous militias gave would-be rebels every reason to expect that any attempted uprising would be harshly punished.5 The weakness of the revolutionary threat from the country’s partisan opposition rendered their preference ordering moot to the problem of democracy’s survival following the restoration of democracy in 2001. What moved front and center instead was a budding rivalry between the traditional ethnic-Fijian elite and the military under Bainimarama. The major interest of the traditional elite who held power during this episode of democracy appears to have been the use of state power to protect and even expand their wealth. During this time, the government sought to expand the property holdings of its key constituents through legislation that would reserve ownership of all coastal land for ethnic Fijians, mostly traditional chiefs.6 In 2004, the government endorsed the creation of a state-funded entity, the Vanua Development Corporation, to invest in projects that would generate income for those landowners.7 The incumbent party also succeeded at blocking further investigation in the failure of the National Bank of Fiji, which collapsed under a load of bad debt in 1997. That failure came after a decade during which the bank expanded advances to indigenous Fijians and made increasingly risky loans under lending criteria that apparently favored traditional elites. When the bank failed, the state took it over, costing Fijian taxpayers hundreds of millions of dollars, roughly 8 percent of the country’s gross domestic product.8 Ultimately, the government

How Does Democracy Break Down?

81

successfully stymied efforts to prosecute individuals responsible for the collapse and to investigate the lending practices that precipitated it. During this period, Commodore Bainimarama also complained about the opacity of spending by communal councils led by traditional chiefs and the government’s failure to audit them. Based on these exclusionary interests and as demonstrated by their sympathy for the failed coup of 2000, Fiji’s traditional elite, represented by the Great Council of Chiefs, fit the mold of an organization that saw political power as a crucial tool for improving its material well-being and saw its interests as gravely threatened by the prospect of an opposition-led government. Its behavior in office suggests that, in game theoretic terms, the incumbent party preferred ruling, whether in a democracy or an autocracy, to all other outcomes. Meanwhile, election losses in 1987 and 1999 had given the incumbents reason to fear that their hold on power was not permanent. At the same time, Commodore Bainimarama’s evident willingness to confront them after the 2000 putsch gave them reason to fear that the military would intervene if they tried to cement their authority through an executive coup or if they rebelled again in response to an election victory by the opposition. The military gave specific assurances ahead of the 2001 election that it would respect the outcome no matter which party won and would protect voters against intimidation; according to a radio broadcast from that time, “The military has expressed similar views, saying eligible voters should feel secure in casting their votes for the party of their choice ‘in the secure knowledge that they have the full protection of the RFMF [Republic of Fiji Military Forces].’”9 In other words, the incumbent party faced significant uncertainty about its hold on power. The model developed here implies that this combination of preferences and beliefs will often impel the ruling party to attempt an executive coup. In fact, evidence from the May 2006 elections, in which the incumbents won a very narrow majority, shows this was the case. After those elections, Fiji’s Human Rights Commission empanelled a Commission of Inquiry to investigate allegations of fraud and abuse by the election’s winners. According to a 2007 editorial in The New Zealand Herald by commission member Dr. David Neilson, the evidence the panel found “strongly indicates a pattern of bias that clearly disadvantaged Indian voters and the Fiji Labour Party, while clearly advantaging the [ruling] SDL party…. The registration process was both inadequate and biased and submissions strongly indicate campaigning involved deliberate and explicit vote-buying near polling day by the SDL party in league with the broader state.” According to the commission’s report, these irregularities were especially acute in the

82

Dilemmas of Democratic Consolidation

“swing” urban constituencies in which Fiji’s elections are won or lost under the electoral system instituted in 1997.10 International observers reported technical irregularities but did not allege partisan bias and declared the elections to be “credible.”11 According to Neilson, however, the observer missions did not find bias in the irregularities because they did not look for it. The commission’s findings suggest that democracy may already have ended in Fiji before the 2006 military coup. In any case, the irregularities observed in the 2006 election certainly gave the opposition and the military reason to believe that the incumbent preferred political power to democracy as such and was attempting to retain office by any means necessary—an attempt the military eventually blocked with its own usurpation of power. Why would the military choose to intervene? The model developed here suggests that it would do so for one of two reasons: either it preferred to rule and believed it could seize power without paying too high a price, or it feared the consequences of a successful executive coup and believed those consequences would outweigh the costs of its own seizure of power. There is little reason to believe military leaders sought power for its own sake. None of the preceding breakdowns of democracy was instigated by a power-hungry military, and the fallout from those previous failures gave officers reason to believe that another military coup would produce negative consequences for the country’s economy and international standing, the functioning of their organization, and perhaps even their personal safety. Instead, the 2006 coup appears to have been primarily a defensive one, provoked by its fear of the slide toward an incumbent-led autocracy. After the mutinies of 2000, the military had serious concerns about fractiousness within its own ranks provoked by the political fighting. Military leaders also apparently dreaded the prospect that they would be forced to referee fighting between indigenous and Indo-Fijians or to confront another armed threat from supporters of the traditional elites if and when the ruling party lost elections again. For Bainimarama, that fear was personal; he was nearly assassinated by rebels in one of the 2000 mutinies, and he believed some members of the indigenous Fijian SDL party had instigated that failed attack.12 Bainimarama’s concerns about the incumbent party’s intentions led him to show its distrust in public and voice its displeasure with some of the government’s policy initiatives. These expressions gave the incumbent party reason to fear that the military would either seize power directly or block its efforts to retain office. The resulting mutual mistrust

How Does Democracy Break Down?

83

spurred a tit-for-tat spiral of threats between the military and the incumbent party that ended—ironically for the government—with the December 2006 coup. Apparently at the top of the Bainimarama’s list of concerns was the government’s efforts to establish a Reconciliation and Unity Commission with the power to pardon people involved in the 2000 events whose motives were deemed to be political rather than criminal; Bainimarama specifically complained that military discipline and integrity would be undermined if the 2000 mutineers were released early. In 2004, the government and military wound up in a standoff over the decision to renew Bainimarama’s contract, and the government made noises about trying Bainimarama for attempting to organize what they were calling a coup. In 2004 and 2005, the government continued to pursue reconciliation with the 2000 mutineers and release prisoners early, and Bainimarama continued to object. The government escalated the conflict in October 2005 when it tried to shift control of the budget for the police and military to the Ministry of Home Affairs, but Bainimarama and the police chief successfully protested, saying that change would effectively politicize their organizations. Soldiers’ cell phones were cut off soon thereafter, allegedly as a money-saving measure. Last but not least, a protracted and public legal dispute had also developed over other budget issues that included financial penalties imposed on Bainimarama personally. By the time the SDL government tried to relieve Bainimarama of his command, military leaders apparently saw themselves as being pinned in a corner. If they acquiesced to the government’s efforts, the military would be gravely weakened and increasingly beholden to a faction whose interests they did not share. The main alternative course of action at this point was to remove that threat by overthrowing the government, which it finally did on 5 December 2006. Contrary to the explanations suggested by prominent game theoretic models of regime change, Fiji’s military did not act at the behest of the country’s wealthy elite, nor were they the coercive arm of a rebellious and poor opposition. More consistent with this book’s model, it seems that Fiji’s military acted autonomously when it seized power in December 2006. Moreover, that action apparently occurred in response to a strategic dilemma rather than a hunger for power in its own right. Under Commodore Bainimarama’s leadership, the Fijian military was driven toward a coup by the hostile actions of an incumbent government that appeared to be committed to retaining power by any means, including electoral malfeasance that may have amounted to an executive coup. When that government turned directly on its own security forces and their

84

Dilemmas of Democratic Consolidation

leadership in an effort to consolidate its gains, Commodore Bainimarama appears to have judged that the costs of continuing to cooperate had finally come to outweigh the expected costs of a coup, and he acted accordingly. Breakdown by Executive Coup: Ukraine 1994–1999

Ukraine gained its independence from the Soviet Union in 1991, but by this book’s criteria it did not have a democratic regime until a new legislature was elected and seated in 1994. In March 1990 elections to Ukraine’s legislative body, the Supreme Soviet, an opposition bloc led by the Ukrainian nationalist movement Rukh won one-quarter of the seats in the face of widespread abuse of the system by the incumbent Communist Party. Under pressure from large popular demonstrations, the Communist-dominated legislature in July 1990 passed a resolution on sovereignty, but conservative forces soon reasserted their control over the republican government and blocked further moves toward independence from Moscow. Stduent protesters pushed back in October 1990, and workers from Kiev's largest factory soon joined them in the streets. Faced with this growing movement, the government agreed to meet the students’ demands for more aggressive movement toward Ukrainian sovereignty. In July 1991, the legislature passed a law establishing the post of president of the Ukrainian SSR. Leonid Kravchuk was elected president in elections held in December 1991, and Ukraine was recognized internationally as an independent state later that month, after the USSR formally dissolved. As part of a political deal concluded in September 1993 in the wake of a miners’ strike, early parliamentary and presidential elections were scheduled for 1994. The parliamentary elections held that March and April produced a fractious legislature. The Communist Party won a plurality with 83 of 450 seats, and the allied Socialist Party and Interregional Bloc won another 51 seats, but Rukh and its allies won 117 seats of their own, (Bojcun 1995). These results left the two blocs fairly evenly matched and gave neither one a legislative majority. Shortly after parliamentary elections had concluded, Ukrainian went to the polls again to elect a president. Contrary to most observers’ expectations, incumbent president Kravhcuk was defeated by Leonid Kuchma, who was supported by the Communist Party, its allies, and ethnic Russians in the country’s eastern regions. After these technically messy but substantively competitive elections, Ukraine had finally crossed the threshold to democratic government, with both a chief executive and a

How Does Democracy Break Down?

85

legislature chosen in competitive, inclusive, and largely free and fair balloting. Two years later, after protracted and intense political wrangling, the country’s legislature in June 1996 adopted a new constitution. This constitution replaced the bicameral legislature with a unicameral one to which the prime minister and government were responsible, and it called for a directly elected president to serve as head of state in five-year terms. The first national elections under the new constitution were held in 1998 and 1999. By the time those elections were over, Ukraine was no longer a democracy. According to an election-observation mission affiliated with the Organization for Security and Cooperation in Europe (OSCE)’s Office for Democratic Institutions and Human Rights (ODIHR), significant abuses marred the 1998 legislative balloting, the first national elections under the new constitution. While their report describes a legal and administrative framework that was “adequate” for the task, it also decries (p. 3) a campaign characterized by “incidents of violence, arrests and actions against candidates and abuse of public office” that “raises questions about the neutrality of the state apparatus in the election.” The observer team seems to have been particularly concerned about erosion of press freedom in ways that favored incumbent officials. Its report states (p. 17) that, “Television was the most influential means of communication during the election campaign in Ukraine,” and then concludes that, “State television was clearly under the control of the government, which used both state channels for the promotion of the party of power.” OSCE observer reports suggest that the 1999 presidential poll was similarly flawed, only more so. According to the observer mission’s final report (p. 1), the presidential contest “failed to meet a significant number of the OSCE election related commitments.” In language that is unusually strong and frank for international election observation teams, the mission’s report (p. 1) further claims that, “State administration and public officials were observed campaigning for the incumbent President and against his challengers. The OSCE/ODIHR Election Observation Mission uncovered clear evidence that this campaign by State institutions was widespread, systematic and co-ordinated across the country, and provided the incumbent with a substantial and sustained advantage over his competitors.” As the OSCE reports suggest, the failure of democracy in Ukraine occurred by executive coup. In this instance, the incumbent party used its position in government to advantage itself and impede its competitors to such an extent that elections were no longer sufficiently competitive

86

Dilemmas of Democratic Consolidation

to satisfy the basic conditions for democratic government. Here, the incumbent government apparently did not employ the boldest means, such as rigging the vote count or restricting registration of voters or candidates, to cement its hold on power. Instead, the incumbent government employed its influence over print and broadcast media and its control of other state resources to sharply tilt the playing field in its favor.13 What did not occur in Ukraine during this five-year period were attempts to usurp power by the military or the government’s political opposition. Thus, the outcomes to be explained in this case are the occurrence of a (successful) executive coup attempt and the absence of attempted military coups or rebellions. Why would the incumbent in this case be tempted to dismantle the mechanisms of democratic accountability? This book’s model suggests that it was the material benefits of retaining power and the risks associated with losing office that tempted the “party of power” to rig the system in its favor. Consistent with this explanation, observers of the country’s political economy at the time describe Ukraine as a closed-order society in which wealth flowed not from free exchange but instead from access to state authority. In her analysis of politics in Ukraine between the 1994 and 1998 elections, Birch (2000: 102) writes that, "Various ‘clans,’ tied to different business interests, operated increasingly as patronage networks, and much of politics revolved around the distribution of state wealth among competing groups.” Similarly, van Zon in his 2000 study of the political economy of independent Ukraine describes a ruling elite that viewed the state as a “feeding ground” (p. 4) and observes (p. 39) that, “The state is rather used as a tool by ruling clans to promote their interests, that is, selling off state property.” This kind of rent-seeking behavior often arises in countries where the state controls substantial natural-resource wealth. In post-Soviet Ukraine, however, it was the effort to transform the Soviet-era command economy into a market-based one, and the process of privatization associated with that transition, that produced tremendous opportunities for the exercise of state power for personal gain. Van Zon (2000: 36) specifically links the privatization process to personal financial benefits for the incumbent government under President Kuchma: The office of the president and government organized numerous selloffs of state property in which members of government and presidential staff profited greatly. The dealings with gas companies are widely published. A telling example is the selling off of the Black Sea fleet. Although parliament prohibited privatization of the fleet, with an

How Does Democracy Break Down?

87

estimated value of US$1 billion, approximately 100 ships have been sold or leased.

The nature of Ukraine’s lucrative post-Soviet rents meant that the timing of control of state authority also mattered greatly. These opportunities would not last forever; once the sell-off of state assets was completed, the size of the gains associated with state authority would be significantly reduced. The expectation of diminishing returns arguably strengthened the incentive to hold onto office by any means necessary during this transitional period. Ukraine’s executive coup might have been prevented in spite of these strong incentives if the government had sufficient reason to fear the consequences of “cheating” to sustain its hold on state power. At the time, however, there were few constraints on government efforts to rig the system in its favor. In the mid-1990s, the judiciary and the country’s election management bodies were effectively captive to the government. On Ukraine’s judiciary, Van Zon (2000: 33) reports that, “There is no due regard to law because the law enforcement agencies are very weak and often closely linked to the state bureaucracy. There is no independent judicial power.” Meanwhile, the OSCE observer missions describe an apparatus of election administration that had the potential on paper to serve as an effective and independent agent but in practice was effectively captured by the party of power. For example, in its final report (pp. 5–6) the mission that observed the 1999 presidential elections note a pattern in which associates of President Kuchma were overwhelmingly selected to serve as chairs of regional and local polling commissions in districts where the president’s opponents had performed well in parliamentary balloting the year before. Outside powers also failed to impose any significant constraints on the Kuchma government’s efforts to consolidate its hold on power at the time. Rhetorically, the West was certainly pressuring Ukraine to democratize, but it was also acutely worried about the success of Ukraine’s economic transition, the fate of the country’s nuclear weapons, and the threat of Russian expansion. In the spring of 1995, U.S. President Bill Clinton visited Ukraine at a time when the economy was showing signs of improvement and lauded Kuchma for the steps he had taken toward economic reform. Around the same time, the U.S. Government announced a large increase in foreign aid to Ukraine. This increase made Ukraine the fourth-largest recipient of U.S. assistance at the time and was at least implicitly tied to the country’s decision to dismantle its nuclear arsenal. Meanwhile, the International Monetary Fund (IMF) in the mid-1990s began making large loans to Ukraine that

88

Dilemmas of Democratic Consolidation

were conditional on the adoption of institutional and policy reforms the IMF favored. Shortly before President Clinton’s visit, the Kuchma government agreed to terms with the IMF for a $1.8-billion loan, a marked increase over the $400-million package the IMF had announced the previous fall. While the expected costs of punishment for attempting to rig the system and failing appeared low, the anticipated penalties for losing office to a rival probably loomed much larger. Specifically, Kuchma’s “clan” had to fear that it would be prosecuted for the very corruption that made state power so attractive, if a rival group won control of government. Official corruption was widely recognized as a major problem for Ukraine at the time and was profoundly frustrating to most voters, who were struggling to survive the economic collapse that followed independence (Van Zon 2000). The legislature responded to this frustration, and to the sense that corruption was retarding economic development, by passing numerous laws intended to rein in official corruption. It also established an Organized Crime and Corruption Committee that investigated and successfully documented corrupt practices among sitting officials in the Kuchma government. These conditions suggested that a loss of power might be followed by tougher investigations that could lead to serious legal penalties, the redistribution of wealth, or both. In the language of this book’s formal model, this looming threat would have sharply lowered the expected payoffs to the Kuchma “clan” in an opposition-led democracy, thereby strengthening its preference for retaining power. Other things being equal, the smaller the payoff associated with an opposition-led democratic government, the larger the relative gains from retaining power and thus the stronger the incentive to attempt an executive coup. Where were Ukrainian voters in this story? The central cleavage in Ukrainian politics after independence pitted ethnic Ukrainians concentrated in the country’s western half against ethnic Russians concentrated in its eastern regions. This ethnic and regional divide often coincided with different attitudes toward domestic and foreign policy. As Birch (2000) documents, ethnic Russians in the east tended to support leftist parties and to press for closer ties to Russia, while ethnic Ukrainians in the West were more inclined to favor market reforms and resist Russian influence, but this apparent polarization was only weakly reflected in the dynamics of government at the time. Although Kuchma won heavy support from ethnic Russians in his 1994 and 1999 election victories, his government’s actions in office can hardly be viewed as pandering to that constituency. The Popular Democratic Party formed by Kuchma in 1995 as the “de facto party of government” was essentially

How Does Democracy Break Down?

89

centrist in its policy orientation (Birch 2000: 103), and once in power Kuchma actually pushed to cut state subsidies to industry and agriculture that were avidly supported by the Communists and other parties of the left. As it happens, while popular politics in post-Soviet Ukraine have consistently revolved around the broad ethnic and regional split that has repeatedly produced a divided legislature, elite politics has centered more often on “clans” led by specific individuals with strong local bases. As Birch (2000: 102) describes, The emergence of political fiefdoms in Ukrainian politics was reflected in the changing structure of parliament, where factions formed, dissolved, and reformed to such an extent that by 1998 the political structure of the assembly bore only a vague resemblance to the party affiliations of the deputies elected four years earlier. Communist and Rukh strongholds on the left and right still marked out the main ideological poles of the political field, but the region between these two factions was one of continued flux as aspirant leaders strove to attract followings from among the weakly-aligned centrist mass of the parliament.

Thus, while the Kuchma government certainly enjoyed significant popular support from one of the factions that characterized mass politics in post-independence Ukraine, neither this support nor the broader fact of a polarized electorate appear to have played a significant role in the creeping executive coup of 1998–1999. If anything, the presence of this broad regional, ethnic, and ideological divide may have helped to empower these clans by effectively ceding political power to the “swing” groups in the middle. Given this disconnect between elite politics and mass concerns and the dismal state of the Ukrainian economy throughout much of the 1990s, why was there no attempt by the opposition to seize power from the incumbent before this executive coup or in direct response to it? It is difficult to state with much confidence why Kuchma’s opponents did not mobilize more aggressively at the time, when they clearly had much to lose by their exclusion from power. That said, a look ahead to the Orange Revolution that occurred in response to fraudulent elections in 2004—an uprising that eventually reversed the previous executive coup and restored democracy in Ukraine—suggests that collective action problems may have hobbled the opposition’s response to Kuchma’s consolidation of power several years earlier. In an analysis of the origins of the 2004 uprising, McFaul (2005: 9) states that:

90

Dilemmas of Democratic Consolidation

For much of the previous decade, Ukraine’s democratic forces had remained divided and disorganized. The crafting of an opposition party was complicated by the presence of [a] strong and legitimate Socialist Party, which made cooperation with liberals difficult. Nor, for many years, was there a single, charismatic leader of the opposition who stood out as an obvious first among equals.

The mass mobilization that occurred in Kiev after the 2004 elections was substantially facilitated by a specific organization, Pora, that had not existed in 1998. As McFaul and others have noted, Pora’s emergence in the intervening years was helped tremendously by financial support and organizational training from Western governments and non-governmental organizations. The form and direction of this external support was explicitly modeled on the success of the uprising that overthrew Slobodan Milosevic in Yugoslavia in 2000, and therefore was not available prior to the Kuchma government’s 1998–1999 executive coup. The difficulties involved in mobilizing the kind of collective action required to block the executive coup or seize power for the opposition help to explain why there was no rebellion before 1999 in spite of the obvious temptation to control state power and opportunity costs of outsider status. Fears of civil war and Russian influence probably hobbled opposition mobilization as well. After Ukraine became independent in 1991, many Ukrainians worried that the ethnic and regional split in their country would lead to the kind of violent fights that struck other post-Communist countries with similar demographic and historical situations. A New York Times article from 27 June 1994, on the eve of parliamentary elections, quotes one woman as saying, “People are really afraid of civil war, of Nagorno-Karabakh and Georgia.” Domestic politics in Russia fanned this fear of revived imperialism, heightening concerns about the fate of the country.14 In light of these concerns, opposition activists likely feared that a popular uprising led by ethnic Ukrainians against a president supported by ethnic Russians would have provoked violent conflict and eventually drawn Russia into the fight, thus leading to their least-preferred outcomes of civil war and a possible loss of independence. Where was Ukraine’s military in this story? At independence, Ukraine inherited a massive fighting force from the USSR— approximately 750,000 troops—that was organized for a war against NATO forces positioned to the west. For the government of newly independent Ukraine, however, the two major security threats were thought to be civil conflict and Russia. As a result, the key task for the

How Does Democracy Break Down?

91

military in newly independent Ukraine was its own reorganization, and that reorganization was expected to hurt (D’Anieri et al. 1999: 234). In the early 1990s, the Ukrainian military faced budget and staffing cuts and was subject to a strategic reorientation that was clearly eroding its political prerogatives. In the face of these challenges, military rule would presumably look more appealing. What is more, D’Anieri et al. note that military reform really began in earnest after Kuchma’s election in 1994, so soldiers would have had good reason to associate those challenges specifically with the Kuchma government. Those same soldiers may also have been aware of the financial benefits to be gained by guiding the privatization process, an awarenes that would only have added to the temptation to seize power. Under these circumstances, why were there no attempted military coups during this period? The prospect that these incentives would translate into action seems once again to have been mitigated by the challenges involved in undertaking collective action. D’Anieri et al. explain the absence of forceful intervention by the military in politics as the result of a vacuum in military leadership; as they describe it (1999: 253), In separating the Ukrainian military from the Soviet military, Ukraine created a military without a head, that is, without the strong leadership that would facilitate a role in politics. Prior to 1991, there was no Ukrainian general staff, no independent organizational capacity, no Ukrainian command structure. These were created by the newly independent state and for its purposes.

Betz (2004: 43) makes essentially the same point, remarking that “what Ukraine inherited was not an army as such; rather, it simply laid claim to the parts of the dismembered Soviet army that happened to have been based on its territory, minus the bureaucratic structures needed for its management and control which existed only in Moscow.” Thus, the peculiar circumstances of USSR’s collapse left Ukraine with a large army, but one that lacked sitting leaders with a vested interest in the organization’s status or their own status within it. In the absence of such leaders and a coherent command structure linking them to their soldiers, there were arguably fewer individuals who could organize a serious coup attempt, and the prospects that a coup attempt would succeed probably seemed slim. The shoddy state of the military inherited by Ukraine probably also served to raise the costs of collective action and to lower soldiers’ expectations about the payoff from a military government. Rather than

92

Dilemmas of Democratic Consolidation

being committed to enhancing the status of their organization or advancing their own careers within it, many officers and soldiers in Ukraine’s military at the time were looking for a way out of service. According to D’Anieri et al. (1999: 248), “Surveys show also that about one-half of all young officers would like to quit the armed forces. They stay in the military largely because of their family situation or the fear of unemployment rather than out of any genuine desire to be officers.” Morale may have been even lower among rank-and-file soldiers, who were conscripts rather than volunteers. As in the Soviet army from which it derived, suicide was a serious problem in Ukraine’s armed forces, and soldiers and officers alike fled the services in droves. As Betz (2004: 55) summarizes, “What was happening to the armed forces was not reform but collapse.” The deteriorating state of Ukraine’s military and the weakness of the command structure may have mitigated the risk of a coup, but they did not prevent military officers from pushing back against some of the elected government’s efforts to reshape their organization. D’Anieri et al. cite two instances in which the military intervened in national politics in the mid-1990s. The first occurred in September 1993, when Gen. Morozov, then the country’s defense minister, publicly denounced agreements reached between Ukraine and Russia over the fate of the Black Sea fleet, to which both countries had laid claim after the USSR’s demise. Even though the agreements were never ratified, Morozov was pushed out of his job after this incident, suggesting that the civilian government would not tolerate that kind of public dissent. In the second and more significant instance, a group of officers led by Colonel General Anatoliy Lopata publicly clashed with Kuchma’s civilian defense minister Shmarov over Shmarov’s plan for military reform. According to D’Anieri et al. (1999: 244), Conflict over the reform program was intense, such that Shmarov constructed a completely separate working group under Bizhan, which some characterized as an ‘underground general staff,’ to complete the plan beyond the reach of Lopata and the general staff. Lopata was popular within the officer corps, and the dispute spurred resentment that officers faced political pressure from the Ministry of Defense to side against their leader.

Kuchma responded to the conflict by eventually firing both of the protagonists and tabling a new reform plan that represented a compromise solution. These concessions may have further weakened the incentives to attempt a coup by signaling to the military that its interests

How Does Democracy Break Down?

93

would not be neglected under democratic government, a fact that would narrow or eliminate any relative gains to be won by establishing direct military rule. Summing up, this book’s model focuses our attention on aspects of Ukrainian politics in the 1990s that provide reasonable explanations for the occurrence of an executive coup and absence of attempted military coups and rebellions during this period. Once elected, the Kuchma government enjoyed substantial material benefits from its position of power and had every reason to believe those benefits would be erased and perhaps even reversed if it were turned out of office. These calculations gave the incumbent a powerful incentive to rig the system in its own favor. Meanwhile, the Kuchma government’s political rivals and its own military were hamstrung by collective action problems that helped to prevent either of them from attempting their own usurpations. Under these conditions, the path to an executive coup was fairly clear. Less than a decade after the USSR’s disintegration, one of the Soviet successor states considered most receptive to democracy had returned to authoritarian rule, and a careful look at the strategic considerations of the crucial actors involved helps us understand why. Breakdown by Rebellion: Cyprus 1960–1963

Under British rule after 1878 and Ottoman rule for three centuries before that, Cyprus gained its independence in 1960. At the time, the Mediterranean island’s population numbered less than 1 million and was roughly four-fifths Greek Cypriot and one-fifth Turkish Cypriot. Both of those communities had strong ties to their respective “motherlands,” which viewed Cyprus as an extension of their own hostile relations, and were only loosely unified politically and socially with each other. As early as the 1820s, when Greece regained its independence from the Ottomans, Greek Cypriots had violently resisted Turkish rule and sought unification with Greece, or enosis. In the mid-1950s, Greek Cypriots amplified their calls for enosis again and launched a violent campaign to push for that goal under the banner of a militant group known as the National Organization of Cypriot Fighters (EOKA by its Greek initials). Keen on protecting its military base on the island, Great Britain responded with a massive infusion of troops and sought to repress the Greek uprising. Some Turkish Cypriots collaborated with the British in trying to quell a nationalist campaign they apparently perceived as an existential threat. Turkish Cypriots also formed their own militia, known as Vulkan, and sought to institutionalize a split from Greek Cypriots under the principle of partition, or taksim.

94

Dilemmas of Democratic Consolidation

Faced with this draining challenge at a time of decolonization, Britain began to talk with Greece and Turkey and Cypriot leaders from both communities about exit options that would cede political control of the island while preserving access to its military bases. In 1959, those five parties hammered out the Zurich-London Agreement which, among other things: called for a constitution that would divide power between the two communities at both the national and municipal levels; prohibited changes to the basic structure of the ensuing republic and permanently barred both enosis and taksim; preserved Britain’s sovereignty over territory housing its military bases on the island; and guaranteed the right of each of the three interested outside powers to intervene unilaterally and forcefully if the terms of the deal were violated. The agreement also allowed Greece and Turkey to garrison soldiers on the island—up to 950 for Greece and 650 for Turkey. The constitution that emerged from those negotiations explicitly divided political authority at both the national and local levels between the rival Cypriot communities in various ways. First, it mandated that Greek Cypriots would perpetually choose the country’s president and Turkish Cypriots its vice president and gave significant veto powers to both offices. The constitution also vested executive authority in a Council of Ministers to be composed of seven Greek Cypriots and three Turkish Cypriots. In the fifty-seat unicameral House of Representatives, thirty-five seats were reserved for election by the Greek populace and the other fifteen for election by the Turkish populace, and the constitution gave substantial veto powers to both factions in the form of requirements for joint majorities on legislation on key national issues. 15 Following the terms of the Zurich-London Agreement, the constitution also provided for the establishment of parallel communal chambers, each with a separate right to impose taxes on its own constituents for the needs of its community and each with final authority in the areas of religion, cultural affairs, education, and personal status. At the local level, each of the five largest towns on the island was divided administratively into Greek and Turkish municipalities, each with its own elected council. Across the island, civil-service jobs were apportioned according to the same 7:3 ratio used to distribute seats on the Council of Ministers and in the House of Representatives. Finally, the constitution also called for the establishment of a unified national army drawn from both communities according to a slightly more even 3:2 ratio. Archbishop Makarios III, a leader of the Greek Cypriot nationalist movement during the preceding decade, defeated fellow Greek Cypriot John Clerides for the presidency in elections held in December 1959,

How Does Democracy Break Down?

95

and Turkish Cypriot Dr. Fazil Küçük became vice president after running unopposed. Elections to the unicameral House of Representatives were held in July 1960. All but five of the thirty-five seats allotted to Greek Cypriots went to Makarios’ Patriotic Front and its allies—the Greek Cypriot communist party won the other five—and all fifteen of the Turkish Cypriots elected to the legislature were supporters of Vice President Küçük. The final version of the multilateral agreement was signed in April 1960, and Cyprus officially became independent on August 16 of that year. The democratic regime painstakingly constructed under the powersharing constitution of 1960 crumbled in December 1963 when civil war broke out, effectively partitioning the island into Greek and Turkish enclaves. Fighting flared after an incident in which a Greek Cypriot policeman killed two Turkish Cypriots in the capital, Nicosia, but there had been other violent incidents earlier in the year. The war drove tens of thousands of Cypriots to flee to parts of the island inhabited by coethnics and dragged on into 1964 in spite of threats by both Greece and Turkey to intervene on behalf of their favored sides. In mid-1964, the U.N. Security Council voted to establish an international peacekeeping force in Cyprus and began to try to broker a resolution to the dispute. Some fifty years and billions of dollars later, the island remained divided, and that peacekeeping force remained in place. This case was chosen for closer study here as one that ended with the breakdown of democracy by rebellion—in this instance, through the collapse of central state authority rather than the outright seizure of power. In fact, closer study shows the nature of the democratic breakdown was more ambiguous than the “rebellion” label suggests. Most significant, it is not clear which side started the fighting in December 1963 that precipitated the collapse of the national government. Before that moment, both sides had been strengthening the private armies they had first built under colonial rule, and each has claimed the other was responsible for the outbreak of violence that winter. Also different from the archetypal rebellion, the Turkish Cypriot faction that fought this war never seemed to seek control of national government. Their actions were at least partly defensive, and it is not clear whether they wished to rule Cyprus as a whole or were just seeking autonomy or secession. In spite of these ambiguities about the precise nature of the breakdown, the Cyprus case clearly exemplifies the kind of mutual insecurity emphasized in this book’s theoretical model. This is a case in which democracy stopped functioning as constitutionally prescribed because mutual mistrust and strategic uncertainty put the rival factions

96

Dilemmas of Democratic Consolidation

on a downward spiral toward confrontation rather than cooperation. In the end, what mattered most was not the “true” nature of the interests pursued by the two sides, but the beliefs—or fears, really—that each side held about its rivals’ goals. Whether or not the Greek Cypriots actually sought to establish a tyranny of the majority or the Turkish Cypriots actually wanted to return the island to Turkish rule is less important to understanding the breakdown of national democracy in 1963 than the two sides’ perceptions that their rivals harbored these kinds of goals. To see how the formal model developed in this book applies to this case, we can start by making some deductions about the preferences of the organizations leading the Greek and Turkish Cypriot communities at independence. For starters, the absence of a national army removed the possibility of one of the model’s five outcomes: a dictatorship led by the military as an autonomous actor. Although the 1960 constitution called for a unified national military, that organization had not come into existence before the breakdown in 1963. According to Salih (1978: 21), “Even though the nucleus of an army was formed, a Cypriot army never materialized.” Meanwhile, the demographic structure of the island’s population—Greek Cypriots outnumbered Turkish Cypriots by a ratio of four to one—meant that Turkish Cypriots could never hope to win control of government through democratic elections, thereby removing another of the five usual outcomes: democracy governed by the current opposition. Thus, the peculiarities of the situation defined the interaction in Cyprus as a two-player game with three possible outcomes: the persistence of democracy with a government led in perpetuity by Greek Cypriots, or an authoritarian regime led by one or the other of the two communities’ dominant political organizations. Given their recent history of hostilities—more than 500 Cypriots were killed in the fighting and repression that immediately preceded independence—both sides presumably saw an autocracy led by their communal rivals as their worst-case scenario. On fears of the subjugation that outcome would entail, Borowiec (2000: 33) quotes Turkish Cypriot Necati Ertegün, a Turkish judge on the country’s first Supreme Court, as saying, “What was ‘freedom’ for the Greek Cypriots was ‘enslavement’ for the Turkish Cypriots,” and the Greek Cypriots undoubtedly saw the prospect of Turkish rule in similar terms. Meanwhile, the power-sharing provisions of the 1960 constitution narrowed the range of policy outcomes in ways that limited the expected payoffs from democracy to the Greek Cypriots.16 This arrangement presumably increased the size of the relative gains for the Greek Cypriots from dictatorship compared with democracy and thus

How Does Democracy Break Down?

97

strengthened their temptation to attempt an executive coup. Permanently relegated to opposition status in the democratic regime, Turkish Cypriots presumably would have most preferred ruling as a dictator as well, because only then could they hope to achieve their favored policies. In short, the “game” in this instance simplified to a rivalry between two players who wanted most of all to break from the constitution and claim power exclusively for themselves; who wanted least of all to see their rival successfully seize power; and who expected their payoffs from the survival of the constitutional regime to fall somewhere in the middle. Under these circumstances, the formal model suggests that the two parties’ choices of strategies would have been heavily influenced by three concerns: the expected costs of collective action associated with attempting an executive coup or rebellion, their beliefs about the chances that such an attempt would succeed, and their beliefs about the rival’s choice of strategies. The legacy of armed confrontation between the two communities that predated independence would have affected all three of these parameters. It would have lowered the cost of collective action by giving both parties pre-formed militias—Greek Cypriots’ EOKA and Turkish Cypriots’ Vulkan, later recast as the Turkish Defense Organization (TMT)—that could provide the muscle to support a defection from democracy. At the same time, the knowledge that their rivals were also armed and organized would have increased both parties’ uncertainty about the prospects that an attempted usurpation of power would succeed and lowered the expected gains from authoritarian rule by the costliness of the fight required to produce it. Finally, as suggested by theories on how security dilemmas can lead to war (Fearon 1998, Jervis 1976, Posen 1993, Walter and Snyder 1999), the very existence of their rival’s militia could have goaded the parties toward confrontation by increasing each side’s uncertainty about their rival’s intentions. Importantly, the existence of these militias was common knowledge; according to Salih (1978: 21-22), “The leaders of both communities on the island openly admit that they had men going through military training in Greece and Turkey, and that weapons were smuggled and hidden in planned areas all over the island.” Each side claimed that its militia was intended as protection against aggression by the other, but rumors abounded on both sides that their rival was preparing some kind of surprise attack. According to Borowiec (2000), Turkish Cypriots believed at the time that their Greek counterparts had a specific plan in place (the Akritas Plan) by which the Greek militias would eradicate the Turks, while Greek Cypriots reportedly believed

98

Dilemmas of Democratic Consolidation

their Turkish counterparts were looking to provoke conflict in order to induce an intervention by Turkey that would turn the fight in their favor. The threatening signals sent by the strengthening of these militias were only reinforced by the actions of the two sides’ political leaders. One of Makarios’ first actions as president of the new republic was to appoint several leaders of the violent struggle for independence to top government posts, including Minister of Interior. In 1962, Greek Cypriot politician John Clerides tabled a proposal that would have unified municipal administration in ethnically mixed cities under governments divided according to the usual 7:3 ratio and would also have ceded to Turkish Cypriots greater control over municipal budgets in cities with Turkish majorities. Although the proposal offered something to both sides, Turkish Cypriot leaders rejected it, apparently out of fear that it would erode the standard of partition underlying taksim (Borowiec 2000). Makarios then responded by forcing a vote on the Executive Council in favor of unified administration with proportional representation, but the Constitutional Court—composed of one Greek Cypriot, one Turkish Cypriot, and one “neutral” international representative, all three of whom had to be approved by all parties to the London-Zurich Agreements—rejected that move in the spring of 1963. The move widely regarded as the trigger to the eventual escalation of violence came in November 1963, when Makarios offered his socalled 13-Point Plan to unify national government in Cyprus under a revised constitution.17 Among other things, this plan sought to eliminate the president and vice president’s veto powers, to remove the requirement for separate majorities within the two legislative constituencies, and to revise the quotas for representation in the civil service and the national army to match the 4:1 proportions of the population as a whole instead of the 7:3 and 3:2 ratios prescribed in the 1960 constitution. Although the plan incldued some concessions to the Turkish side—for example, it would have allowed the Turkish vice president to fill in for the Greek president in case of the latter’s incapacitation, and it would have eliminated the Greek Communal Chamber—this proposal seems to have been received by Turkish Cypriots as an attempt to establish a Greek Cypriot dictatorship. If the numerous veto points guaranteed by the 1960 constitution were removed, the island’s demographics would have relegated the Turkish Cypriot community to the status of a permanent and weak opposition. Thus, by late 1963, both sides had reason to believe the other was already defecting from democracy. The Turks feared the Akritas Plan and Makario’s proposed shift to unified government, while the Greeks feared threats of Turkish military intervention. In game theoretic terms,

How Does Democracy Break Down?

99

this belief would have put both parties on a pathway in which the persistence of the status quo was seen as an unlikely outcome. Instead of choosing between democracy and dictatorship, they were choosing between passively risking that their rival’s attempt at usurpation would succeed and actively countering that threat with an attempt of their own. As de Figueiredo and Weingast (1999) show, in light of the abysmal payoffs anticipated from subjugation by the rival community—payoffs based partly on the possibility of unification with the rival’s motherland and the associated losses of life, property, and political rights—the expected probability that such an attack would succeed need not have been large to tempt both sides toward defection. Fear of statelessness, eviction, and even annihilation are powerful motivators. It is tempting to blame the civil war that ensued on the convoluted provisions of the power-sharing constitution, an explanation that would suggest the conflict could have been prevented by better institutional design. In retrospect, however, it seems clear that the arrangement under which independent Cyprus was born was an outgrowth of those troubled relations, not the source of them. Long before the British finally ceded their authority over the island, the two communities had mobilized for military confrontation and engaged in fighting that claimed hundreds of lives. The prevailing feeling at the moment of Cyprus’ independence seems to have been one of profound mutual mistrust. The ZurichLondon Agreement and associated constitution certainly did not resolve the insecurities that led to that fighting; in fact, their provisions for multiple veto points for both communities seemed only to guarantee a perpetual political stalemate, thereby lowering the expected payoffs from democracy for both sides. Even so, it is difficult to imagine any other arrangement that would have been acceptable to the five parties to the negotiations that produced that constitution. Without those kinds of provisions, it is unlikely there would have been an attempt at democracy under a unified government in the first place. More than anything else, the involvement of foreign powers seems to have heavily conditioned the outcome in this case. Had it merely been a political struggle between the two communities on the island, this confrontation might have been more manageable. The deep involvement of external powers helped spur those communities toward conflict, however, by simultaneously increasing their co-ethnic’s coercive capacity and heightening the other side’s sense of insecurity. Both countries did this directly, with military training and materiel channeled through the two communities’ private militias, and indirectly, through a guarantee that they would intervene to defend their “brethren” in the face of a security threat.18 These external interventions gave both sides

100

Dilemmas of Democratic Consolidation

reason to believe they were more capable of achieving their political goals by force than they would have been without such support. Perversely, they may also have limited the expected costs of a military confrontation to the Cypriot groups by implying that the external powers would bear a heavy burden in any war. That both of those powers were members of the same security alliance, the North Atlantic Treaty Organization (NATO), may have reinforced a sense that war would be less costly by producing an expectation that a broader international intervention would occur to stop fighting if it ever did break out. In fact, that expectation was proven correct when the United Nations moved quickly to intervene after war erupted in 1963. As it happens, the dynamics of domestic politics in those external powers may also help to explain the occurrence and timing of the breakdown of the national democratic regime in Cyprus. In parliamentary elections held in Greece in the summer of 1963, challenger George Papandreou of the Center Union party encouraged Makarios to push forward with his plan to amend the Cyprus constitution and promised to support enosis if elected (Salih 1978: 22). The Center Union party won those elections, but only with a plurality, and Papandreou asked Greece’s king to call fresh elections, which happened in February 1964. Meanwhile, the Center Union’s victory appears to have emboldened Makarios by suggesting that support from Greece was at its strongest. According to Salih (1978: 23), Makarios also believed political instability in Ankara at the time would prevent Turkey from intervening. Taken together, these beliefs could have convinced Greek Cypriot leaders that it was an unusually opportune time to push for their goal of unified national government and eventual enosis. The timing of the 13-Point Plan seems to have been driven by this logic, and that plan, in turn, contributed to Turkish Cypriots’ decision to respond forcefully to the perceived provocation of 21 December 1963. In its edition the next day, The New York Times reported the incidents widely regarded as the start of the civil war as follows:19 The violence began with an hour-long battle with machine guns, pistols and rifles that erupted shortly after midnight between a Turkish crowd and a police patrol on narrow streets dividing the Greek and Turkish sectors of the capital…The police said the crowd had started the clash by firing on the patrol. The Turkish leadership charged that the incident was ‘a consequence of police methods used against the Turkish community…Later, policemen and Turkish Cypriote [sic] students clashed in the main square of the Turkish sector. Seeking to quell a riotous demonstration, the police opened fire. Two Turkish

How Does Democracy Break Down?

101

youths were wounded. The officers said demonstrators had stoned or shot at Greek cars and buses, though none of the occupants was hit, and stormed the sector’s police station in an attempt to raise the Turkish flag.

Prevailing theories of democratic consolidation and breakdown do not shed much light on why the republican regime established in Cyprus under the constitution of 1960 collapsed in the face of civil war just a few years later. Upon independence, Cyprus was not an especially poor country—its infant mortality rate at the time was on par with Australia’s or France’s in the mid-1950s—and there was no economic crisis occurring at the time of breakdown. Consistent with Acemoglu and Robinson’s (2006) emphasis on economic inequality as a cause of regime change, there is evidence that Greek Cypriots resented redistribution through subsidies and the allocation of civil-service jobs that they perceived as favoring their Turkish rivals.20 The breakdown that eventually occurred, however, was not the result of a reactionary coup on behalf of the wealthier community, as Acemoglu and Robinson’s (2006) theory suggests. Instead, it was the result of an armed confrontation born of mutual mistrust and the associated fears of exploitation. The outcome in Cyprus is also broadly consistent with Rustow’s (1973) claim that democracy is inherently susceptible to failure in countries that lack a shared sense of national identity, but that general assertion is also contradicted by numerous cases in which democracy survives in spite of national disunity,21 and it tells us virtually nothing about how or under what conditions the lack of a shared identity does or does not lead to breakdown. Even though the outcome in Cyprus does not fit neatly into one of the categories used in this book’s formal model, the theory developed here arguably does a better job capturing the causal mechanisms at work in the spiral toward violent confrontation and regime breakdown in Cyprus after 1960 than do other theories of democratic breakdown. The constitutional order disintegrated because both parties felt they had much to fear from domination by their rival, and the behavior of those parties and their external patrons before, during, and after the establishment of democracy gave both communities reason to fear that a defection by their rival was imminent and would expose them to grave risk of exploitation. Under these difficult conditions, it is hard to imagine how a unified, national, democratic government might ever have survived in Cyprus.

102

Dilemmas of Democratic Consolidation

Survival of Democracy: Spain Since 1977

In November 1975, nearly four decades after he came to power through the military coup that triggered Spain’s bloody civil war, Generalisimo Francisco Franco died. Concerned about the instability of succession, Franco in 1969 had named as his successor then-Prince Juan Carlos, the grandson of King Alfonso XIII and heir to the Spanish throne. On 30 October 1975, Juan Carlos provisionally assumed the powers of chief of state and head of government. In 1976, amidst a wave of strikes, demonstrations, and Basque separatist violence, Juan Carlos and his newly appointed prime minister, Adolfo Suárez González, ushered in a program for political reform that won begrudging support from key factions within the regime but was overwhelmingly endorsed by voters in a December 1976 referendum.22 Political parties were legalized, elections were held, and in 1977 the new legislature was seated, marking the restoration of democratic government nearly half a century after the fall of the Republic in the 1936. Led by Prime Minister Suárez, the right-leaning Democratic Center Union (UCD) won a near-majority of seats in the 1977 elections and expanded its plurality in March 1979 balloting. In 1978, a new constitution was adopted by that legislature and then overwhelmingly approved by popular referendum. Suárez’s UCD-led coalition government collapsed in 1981, however, and military officers attempted a coup on 23 February as a new prime minister was preparing to be sworn into office. The so-called F-23 coup fizzled when the king failed to endorse it and mustered support from other officers for the elected government. Elections in 1982 transferred power from the UCD to the leftist Spanish Socialist Workers’ Party (PSOE), which retained control of government through elections in 1996. As of 2009, democracy had survived without interruption in Spain since its restoration in 1977. The process by which democracy was established and then persisted in Spain after Franco’s death is probably one of the most closely studied cases in the literature on comparative democratization,23 and virtually all observers of those processes agree that the risk of a democratic breakdown in Spain has been minimal since the failure of the F-23 coup attempt in 1981. The brief study sketched below does not attempt to explain the occurrence of Spain’s transition to democracy after Franco’s death—a topic that lies beyond the purview of the theory developed in this book—nor does it try to detail the events of the several years during which Spain’s democratic “consolidation” is said to have taken place. Instead, its much narrower goal is to see if evidence about the

How Does Democracy Break Down?

103

preferences and beliefs of key political actors during and after that initial period can be distilled to explain four outcomes that are within the scope of the current theory. The outcomes to be explained are: the absence of attempted executive coups; the absence of attempted rebellions; the occurrence of attempted military coups in the years immediately after democracy’s restoration; and the absence of attempted military coups since then.24 In assessments of prior research on the subject, some authors have described a prevailing view that the survival of democracy in Spain after 1977 was “overdetermined”—that is to say, more or less inevitable—by favorable structural conditions. Researchers have cited Spain’s level of economic development, the lure of closer integration with Western Europe, painful memories of its civil war, and other structural factors as conducive to the establishment and persistence of democratic government after Franco’s death. In retrospect, though, the survival of the new regime hardly seems to have been a foregone conclusion. As Linz and Stepan (1996: 89) suggest, “It is well to remember that even the easiest and most successful transition was lived as a precarious process.” Just because Spain’s successful democratization fits the predictions of structural theories does not mean that the variables identified by those theories tell us much about how or why that outcome occurred. In fact, the economic malaise that afflicted Spain in the late 1970s, the escalation of separatist violence, and the eventual collapse of the UCD government in the face of these challenges suggest the situation was more perilous than the “overdetermined” argument implies. Given the myriad challenges and profound uncertainty of politics in Spain after the restoration of democracy, why did neither of the leading political parties—the UCD and its leading leftist rival, the PSOE— attempt to usurp power? Consistent with this book’s model, work by Alexander (2002) suggests that major right-wing forces supported democratization after Franco’s death because they perceived the revolutionary threat from the left to be weak and expected better payoffs from democracy than autocracy. On the initial decision to democratize, he cites survey data and extensive interviews with transitional elites and conservative small-business owners to support his contention (2002: 141) that: Conservatives became convinced that the prospective democratic electorate was highly unlikely to generate property-threatening policies or extensive civil violence. The right believed left leaders would be hemmed in by the fact that the social bases of the radical left

104

Dilemmas of Democratic Consolidation

had been powerfully eroded by changes in Spain’s social structure, income and educational levels, and to a lesser extent the effects of the civil war. This meant that the right could expect to secure its basic preferences over both policy issues and safety in democracy—even in the event of electoral alternation—without having to incur the risks of exposure to repression inherent even under a sympathetic authoritarian regime.

Importantly, Alexander’s interviews suggest that the right believed authoritarian rule could be maintained after Franco but did not believe it was worth saving. The key risk in the 1930s was the threat to security and property posed by a mass of hungry and landless people. By the 1970s, however, that impoverished populace had been transformed by economic and social development into an organized working class and an expanding middle class supported by a modern welfare state. This transformation was reflected in survey data indicating that four out of five Spaniards in 1977 identified themselves as centrist (Preston 1990: 122). Under these circumstances, leftist organizations no longer had a strong incentive to pursue forcible redistribution through radical action. As a result, “Exclusion [via autocracy] was less costly than forty years earlier but was a solution to a problem that no longer existed” (Alexander 2002: 161). As Alexander describes, the beliefs of many right-wing actors that the left had moderated were only reinforced by events after 1977. Following the restoration of democracy, the Communist Party of Spain (PCE) performed poorly in electoral competition, membership in labor organizations declined, the radical left’s occasional attempts at direct action fizzled, and the PSOE advocated policies in many areas that were not radically different from those of the UCD. The PSOE’s failure to make significant gains in 1979 elections drove the party leadership even further toward the center, a shift crystallized in an extraordinary congress later that year in which the moderate Felipe González emerged as the party’s leader. Even if it preferred democracy to most arrangements, the UCD still might have been tempted to try an executive coup as a way to ensure adoption of its preferred policies. Any inclination among UCD leaders to try to cement their hold on office before or immediately after their party’s election loss in 1982, however, was likely mitigated by the anticipated costs of a coup attempt and uncertainty about the payoffs associated with a return to authoritarian rule. Although the left had moderated its demands since the 1930s, it had also shown repeatedly— most recently, in the period following Franco’s death—that it could

How Does Democracy Break Down?

105

mobilize substantial popular support in defense of those more moderate demands, including access to power by elections. Indicative of the challenges right-wing politicians would have faced in re-imposing authoritarian rule, one transition leader told Alexander (2002: 181) that he expected any coup would be faced within days by “millions” of protesters in the streets. What’s more, strikes and other work stoppages during the transition had imposed real costs on the Spanish economy, and anyone contemplating a return to autocracy had to bear these economic costs in mind. The extent to which their income now depended on the labor and consumption of the working and middle classes would have prodded business-owners and landholders to devalue any attempt to re-establish authoritarian rule. Last but not least, the UCD was hardly a monolithic organization that could efficiently impose the kinds of procedural changes usually used to carry out an executive coup. Instead, its power in the legislature depended on a loose coalition of parties with divergent interests. This dependency on a loose-knit coalition implied that the costs of the collective action required to execute an executive coup would have been steep. These beliefs and conditions help to explain why the incumbent UCD government did not attempt to dismantle democracy once it was established, even when their electoral fortunes turned in 1982. What about the leftist opposition? Unfortunately, Alexander’s research does not include interviews with leftists that would shine direct light on their beliefs during this period as it does on the beliefs of conservatives. Still, we can deduce some tentative conclusions about those beliefs from their statements and behavior at the time. Election results, survey data, and a decline in the ranks of organized labor after 1977 suggested an electorate that was more centrist than split, and, as previously noted, the PSOE modified its behavior under democracy to fit this tendency. Importantly, the PSOE won enough seats in the 1977 elections to bargain with the UCD over the terms of a new constitution from a position of strength. As Arango (1995: 115) observes, “Some form of consensus…was necessary if the constitution writers were to succeed,” and the congressional committee established to draft this constitution reflected this need for consensus by drawing on several parties, including the PSOE and PCE as well, of course, as the UCD. Against a backdrop of moderation in popular attitudes, this balance of power helped to produce a constitution that did not specifically favor one of the major party’s interests over the other. Meanwhile, the military was clearly sympathetic to the right, suggesting that any attempt by the PSOE at rebellion—or, once in power, an executive coup—would be resisted with powerful force.

106

Dilemmas of Democratic Consolidation

Absent a reactionary coup attempt that directly threatened the gains they had made under the new regime, the left’s ability to mount a rebellion seemed weak. On top of all these domestic factors, the Cold War context meant that a leftist authoritarian regime in Western Europe was almost certain to provoke a strong, negative response from Europe and the United States. Thus, even if leftist forces had hoped to rule by authoritarian means—and circumstantial evidence suggests that they did not—the profound challenges and steep costs of achieving that goal seem to have been sufficient to deter them from trying. Some observers credit the Moncloa pacts—bargains on economic policy negotiated by leaders of several parliamentary factions in October 1977, before a new constitution was established—with helping to protect Spain’s fledgling democracy through its initial years. As Alexander (2002: 179) describes them, the deals struck at Moncloa “were a set of essentially economic agreements: the left agreed to limits on wages, public spending, and credit, while the UCD government agreed to implement income tax reform, reorganize welfare state programs, and initiate administrative reforms.” The argument developed here implies that these pacts were neither necessary nor sufficient for the survival of democracy in Spain at the time. The process of pacting may have reinforced the players’ beliefs that their rivals would not seek to exploit them by attempting to usurp power, but the aforementioned survey and interview data indicate that those beliefs preceded the pacts and therefore could not have been produced by them. In this case, the formal bargains appear to have been less important to the survival of democracy than the rival parties’ beliefs that the other had little stomach for the anticipated consequences of any attempt to impose authoritarian rule. What’s more, neither the spirit of compromise and consensus supposedly embodied in the pacts nor the specific bargains codified in them survived for long (Alexander 2002: 179–180). The rapid disintegration of these deals makes it difficult to imagine how the pacts themselves could have played a pivotal role in the longer-term survival of the Spanish regime. In short, both the incumbent and opposition parties in Spain after 1977 eschewed attempts to usurp power because they expected that continued participation in the democratic regime would prove more rewarding than any attempt to rule as autocrats, even when the economy foundered as it did in the late 1970s. Neither party’s commitment to democracy was contingent on winning elections, in part because both sides believed the left no longer posed a credible revolutionary threat and the right had little to gain by attempting to re-establish a conservative autocracy. Under these favorable conditions, adherence to

How Does Democracy Break Down?

107

the democratic bargain was self-enforcing; both sides believed they were better off in a democratic regime, and both had reason to believe their rival preferred to continue to cooperate as well. The ultimate outcome in this case—the indefinite survival of democracy—and the class-based explanation for that outcome offered by Alexander are consistent with Acemoglu and Robinson’s (2006) theory, which emphasize elites’ concerns about the redistribution of their wealth as the causes of coups in democracies.25 What those classcentered theories cannot offer is an explanation for the military coup plots that emerged in spite of the incumbent right-of-center party’s evident commitment to democracy. Why, then, were some elements in the Spanish military disloyal to the government during this period? Many historians point to the decision during the transition to legalize the Communist Party of Spain (PCE)— the bête noire of the right in the 1930s and a force the military had been indoctrinated to view as its enemy—as the trigger to military conspiracies against the new regime. What that explanation overlooks is that military conspiring continued even as the ensuing several years revealed the weakness of the radical left. The PCE won just 10 percent of the vote in 1977 and 1979 elections, the PSOE shifted toward the center, and the radical left’s few attempts at direct action garnered little support. Under these circumstances, it is hard to see how military leaders would not have updated their assessment of the Communist “threat” and adjusted their strategies accordingly. A more compelling explanation can be found in the persistence and escalation of separatist violence in the late 1970s, including the successful and attempted assassination of numerous military officers, as a crucial factor in the military’s thinking at the time. According to Linz and Stepan (1996: 99), “This armed violence created the very real potential of military opposition to the democratic transition and consolidation because, while not one army officer was killed during the Basque insurgency in 1968-75 under Franco, or in the 1975-77 transition period, in the post-electoral period of democratic rule between 1978 and 1983, thirty-seven army officers died due to Basque nationalist violence.” Preston (1990: 99) describes the military’s reaction to some of these killings, noting for example that, after the murder of two officers in May 1979, “Tension in the barracks could hardly have been greater,” and there were shouts at the funeral of “Ejercito al Poder!” As the violence was escalating, the civilian leadership was pursuing a policy of accommodation that led in 1979 to successful referenda on formal autonomy for the Basque region and Catalonia. Unfortunately, these concessions failed to slow the rate of separatist violence, as

108

Dilemmas of Democratic Consolidation

civilian politicians had suggested they would. Meanwhile, reactionary leaders decried these concessions as stepping stones on a path to “chaos” and insecurity. The military’s concerns about rising insecurity appear to have been exacerbated by the government’s efforts to tighten the reins on an officer corps it (correctly) perceived as a political menace. Some influential officers apparently saw these efforts as direct threats to their own careers. According to Preston (1990: 130), “There were rumors throughout the summer of 1977 that the cabinet intended to purge older officers and thereby prevent them fulfilling their career and pension prospects.” In 1979, around the same time as the murders of two prominent officers in Basque Country, Prime Minister Suárez bucked traditional promotion policy in his appointment of liberal Gen. José Gabeiras Montero as Chief of General Staff, passing over two more senior officers who were widely viewed as reactionaries. (One of those two officers, Captain-General Jaime Miláns del Bosch, would later help to lead the F-23 putsch.) In 1980, the legislature took up a draft law on military amnesty that was meant to reintegrate officers who fought with the Republic during the Spanish civil war. According to Preston (p. 179), “Anger in army quarters was reported to be even more intense than it had been after the legalization of the PCE.” Thus, some Spanish officers in the early years of the new democracy seemed to face significant threats from two directions: threats to their lives from the Basque insurgency, and threats to their careers from the government’s efforts to reorient the armed forces in a more liberal direction. In this environment, it is not entirely surprising that coup-plotting began early. According to Arango (1995: 144), “The coup d’etat [of 23 February] had been building since events during the early days of the transition convinced the diehards in the military that the demolition of the Franco regime and the transition to democracy were in earnest.” In September 1977, just months after founding elections, the government uncovered an alleged conspiracy by a few top generals and responded by reassigning them. In late 1978, the government foiled Operation Galaxia, so named for the café in which it was supposedly hatched. When the plotters—including one of the leaders of the F-23 attempt—were finally tried in May 1980, however, they were given light sentences amounting to time served and immediately returned to duty. By late 1980, after a series of assassination attempts on top generals by radical Basque nationalists and the leftist GRAPO, talk of an impending coup had become widespread, especially in the right-wing press (Preston 1990).

How Does Democracy Break Down?

109

In game-theoretic terms, the escalation in violence and the officers’ responses to it suggest that those officers saw their expected gains under democracy as low and declining, presumably because of the serious threats to their personal and organizational security and the apparent inability of the civilian leadership to stem those threats. Officers may have been uncertain about the chances that a military coup would succeed, but the formal model shows that those chances did not have to be especially great to induce an attempt, as long as the expected gains from a military-led government were much larger than the expected gains under continued democracy. Moreover, the government’s flaccid response to previously uncovered coup plots gave would-be conspirators reason to believe that there was little to lose by trying and failing. Of the light sentences given to the Galaxia plotters, Preston (1990: 168) argues that, “A greater encouragement for plotters could hardly have been imagined.” Under these circumstances, it is not surprising that some officers decided to gamble on a coup attempt in spite of uncertainty about their chances for success and the political and economic fallout the attempt might produce. Of course, the F-23 coup did not succeed, and the decades since that putsch have apparently been free from serious coup-plotting. What changed? The absence of military conspiracies since 1981 is probably due, at least in part, to revisions in the beliefs of military officers about the probability that a coup would succeed and the penalties for trying and failing. Many observers credit King Juan Carlos with “saving” Spain’s democracy because he declined to endorse the officers’ plan, as they had apparently expected him to do, and instead cajoled other officers and appealed to the public to resist the coup. This response clarified that one powerful political actor was an obstacle rather than an ally. Meanwhile, the severe punishment meted out to leading plotters Tejero and Miláns del Bosch—both were sentenced to decades in prison for the F-23 coup—helped to discourage later attempts, just as the weak response to the 1979 plot seems to have encouraged them, by giving conspirators reason to expect to pay a steep price for trying and failing. “The trials helped to consolidate democracy,” Linz and Stepan write (1996: 110), “because they showed how divided and without an agenda the military ‘alternative’ really was. The most important hard-liners were defeated, disgraced, and jailed. After the trials there was a ‘steady realization among large numbers of officers that democracy was there to stay and that the military ought to accommodate itself within it.’” Last but not least, the Basque insurgency ebbed after 1981 as well. The years 1978–1980 would ultimately prove to have been the campaign’s deadliest. One factor behind this diminution of violence may

110

Dilemmas of Democratic Consolidation

have been the establishment—with the endorsement of some officials in the PSOE government—of clandestine paramilitary groups aimed at combating ETA in the Basque Country, a scandalous program exposed years later by the Spanish newspaper El Mundo. Another possible factor was the lack of public support, even among Basque nationalists, for the kind of lethal terrorism ETA’s militant wing had practiced in the late 1970s. Whatever the causes, though, it seems telling that the end of a period of military conspiracy and open coup talk coincided with the public defeat of one major coup attempt and a decline in the separatist violence that evidently had helped to motivate that putsch. In sum, three things came together to allow Spain to succeed in sustaining democracy since 1978. First, as Alexander (2002) convincingly shows, socioeconomic changes that occurred under Franco led the country’s leftist and rightist rivals to believe that democracy was their most preferred outcome, regardless of which party held power at the moment. Second, neither of those saw the other as particularly interested in carrying off an executive coup or rebellion, leaving little incentive for the sort of “preemptive strike” that might have ended democracy in spite of those preferences. Last, although elements in the military clearly wished to usurp power in the years immediately following the transition, they ultimately proved incapable of doing so. Fortunate for Spain, the lessons from that failure aligned with exogenous changes in the conditions that had helped provoke the F-23 attempt to discourage future military interventions and thus solidify the country’s democratic regime. Summary

The paths followed by the four cases examined in this chapter are essentially consistent with the theory developed in the preceding ones. In each of the cases, the conditions and mechanisms emphasized in the theoretical model could be connected to the observed outcomes through a coherent chain of causal logic. In some of the cases, the temptations and dilemmas identified by the model were actually cited by the actors themselves as important motivators of their behavior. In others, there was no “smoking gun” of that sort, but the expected patterns still held in a circumstantial way. Perhaps most significant, none of the cases examined in this chapter defied explanation in the model’s terms. Given that those cases were selected at random from scores of episodes worldwide over half a century, this is no small feat. Of course, such a feat hardly means that my model perfectly explains the survival of democracy everywhere at

How Does Democracy Break Down?

111

all times. As critics of the “analytic narrative” approach have noted, this kind of interpretive analysis is susceptible to selective identification and treatment of the evidence. It is possible that actors in the selected cases were actually motivated by considerations different from the ones emphasized here, or that factors not discussed in these brief narratives also played important roles in those cases. Nevertheless, the fact that we could use the theoretical model to explain cross-sectional and longitudinal variation in outcomes across four diverse cases suggests that the model offers a useful tool for understanding the sources of democratic breakdown and consolidation.

Notes 1 The episodes of democracy were grouped according to their observed outcomes; a random-number generator was used to order the cases within each of the four groups; and then the case at the top of each of the four ordered lists was selected for study. 2 Fiji’s population numbers less than 1 million. 3 The Economist (September 10, 1977: 76). 4 On 7 July 2000, soldiers supporting Speight mutinied in Labasa, allegedly looting and raping Indo-Fijians. A second mutiny happened on 2 November 2000 when soldiers in the Queen Elizabeth Barracks tried to depose Bainimarama. It failed, four of the mutineers were beaten to death by loyalists, and 44 soldiers from the Counter Revolutionary Warfare Unit were later convicted of involvement in the mutiny. 5 Evidence of this threat comes from a 12 January 2001 broadcast of Radio New Zealand International as transcribed by the US Government’s Open Source Center: “A report from Fiji says there are threats of massive violence in the country’s west should the court of appeal rule in favour of the 1997 constitution, putting the interim administration out of office…The Sun says a chief who called their office warned that a major fight could break out in Lautoka if the interim administration was declared illegal and asked to step down. It quotes the chief, who is not named, as saying groups of people were planning pockets of violence, mainly against ethnic Indians, should the decision be made in favour of the deposed government, or for a government of national unity.” 6 According to the Report of the Commonwealth Observer Group on Fiji’s May 2006 elections (p. 28), 86 percent of the country’s land is controlled by the country’s traditional Fijian chiefs, who lease much of that land to the IndoFijians who farm it. 7 Government of Fiji press release of 21 April 2004 at http://www.fiji.gov.fj/publish/page_2311.shtml. 8 “Crisis: Collapse of the National Bank of Fiji,” by R. Grynberg, D. Munro and M. White, Listener (26 April 2003). 9 “President’s Spokesman Says Military to Respect Poll Results,” Fiji Village News (23 Jul 2001).

112

Dilemmas of Democratic Consolidation

10 “What if Fiji Ballot Was Less than ‘Free and Fair?’” New Zealand Herald (15 October 2007). 11 See for example the Report of the Commonwealth Observer Group (21 May 2006) and the Preliminary Statement issued by the European Union’s Election Observation Mission (18 May 2006). 12 Report of the Commonwealth Observer Group (21 May 2006), p. 15. 13 It is interesting to note that elected governments in Armenia, Georgia, and Russia apparently employed similar methods to sustain their power. 14 In 1996, for example, Russian parliamentarians passed a resolution supporting independence from Ukraine for the territory of Crimea, which had been pushing for that goal since 1990. 15 The constitution required majorities within each community’s set of representatives for the passage of legislation in some areas, including electoral laws, municipal authority, and taxes. To alter the membership arrangement, separate supermajorities of two-thirds were required. 16 Exemplifying this situation, Greek Cypriots in the early 1960s complained that tax and subsidy policies established in the 1960 constitution unfairly favored Turkish Cypriot interests, but Turkish Cypriot vetoes prevented them from adjusting those policies in spite of their putative control of national government. 17 In point of fact, some violent incidents occurred before December 1963. Salih (1978) specifically mentions clashes in the Paphos District in May 1963 and a bombing earlier in December that damaged a statue of an EOKA “hero.” 18 Exactly why Turkey and Greece were so deeply entangled in Cyprus is a matter of enduring debate that I will not attempt to resolve. On this question, see Bahcheli (1990) and Hitchens (1997). 19 “2 Die, 9 Hurt in Cyprus Rioting; Greek-Turkish Strife Revived,” The New York Times (22 December 1963), p.1. 20 Salih (1978) cites these issues as important drivers of the confrontation that ensued. 21 Recent counterexamples include Estonia and Latvia. 22 On the politics preceding the transition, see Arango (1995) and Preston (1990) for rich descriptions and Colomer (2000) for game theoretic analysis. 23 Linz and Stepan (1996: 88, fn. 1) note that “the bibliography on the Spanish transition is the most extensive of any of the cases we consider in this book,” which covers more than a dozen countries. 24 The word “apparent” is used here because it is virtually impossible to establish with certainty that no serious efforts were made to cement incumbentparty rule or put military officers in charge of government during this thirty-year period. 25 Acemoglu and Robinson (2006: 224) write, “Our approach presumes that for various reasons, the military represents the interests of the elites more than those of the citizens.”

5 Exploring Confounding Cases

In the preceding chapter, I tested the explanatory power of the theory of democratic breakdown developed in this book with analytic narratives on a set of divergent cases drawn at random from all episodes of democracy in the past half-century. In this chapter, I further probe the model’s explanatory power by turning it on a pair of recent cases of democratic breakdown—Venezuela and Thailand—that prominent prior theories of democratic survival cannot adequately explain. Previously considered an important and even inspiring example of democratic consolidation under difficult conditions, Venezuela saw its democractic regime dismantled by executive coup in the 2000s, several decades after it was established. This breakdown of democracy confounds process-oriented theories which assert that the accumulation over time of experience in cooperation and accommodation under elected government makes democracy permanent, and it undercuts claims that explicit power-sharing arrangements called “pacts” contribute permanently to the consolidation of democracy. As described in this book’s Introduction, Thailand also suffered a democratic breakdown in the mid-2000s. The military coup that ended that episode of democracy confounds structural theories which assert that democratic regimes are almost certain to survive in countries where per capita income has risen above certain levels. The narratives that follow are meant to see whether a theory centered on democracy’s strategic dilemmas allows us to see how and why democracy broke down in these two cases in ways that other theories cannot. Venezuela 1958–2005

Long considered one of the iconic and most successful cases of a “pacted” transition to democracy, Venezuela shows how explicit powersharing agreements can help to produce credible commitments that can help to sustain democracy through less-than-propitious initial

113

114

Dilemmas of Democratic Consolidation

conditions. At the same time, the course the country has taken since that transition also provides a cautionary example of how such commitments can eventually unravel when the conditions under which they were created no longer hold. Venezuela’s latest and longest episode democracy began in 1958, a decade after a short-lived initial try at popular government known as the trienio. The restoration of democratic rule in 1958 occurred only after the country’s leading political parties hammered out a deal—the Pact of Punto Fijo—that “expressly committed party leaders to use their organizations to moderate political conflict” as a means to allow both sides to realize larger gains from economic growth and, apparently, to share the spoils of corruption (Coppedge 1994). According to Karl (1987: 83), the pact guaranteed each of the leading parties “some share of the political and economic pie through access to state jobs and contracts, a partitioning of the ministries, and a complicated spoils system ensuring the political survival of all signatories.” The military, which had played a leading political role in the preceding authoritarian regime, was rewarded for its loyalty to the new system with promises of amnesty from prosecution for abuses committed under its leadership, along with generous salaries and other perquisites (Haggerty 1993; Karl 1987). The model developed in this book suggests that the Venezuelan pact stuck because the parties to it gained substantially from their adherence to it. The oil boom of the 1970s helped the Venezuelan government turn its natural-resource wealth into a growing stream of revenue. In contrast to cases of “pacted” democracy such as Spain, however—where labor and capital-holders found themselves in a situation of mutual dependency by the 1970s—production of oil revenues in Venezuela did not depend so much on continuing collaboration.1 This meant that any actor who successfully seized control of Venezuela could anticipate larger financial gains from holding power alone than from continuing to share it with rivals. In that sense, the expected payoffs associated with usurping power in Venezuela were larger than they were in Spain, so the actors’ economic incentives to sustain cooperation were not as clear. What helped cement actors’ commitments to democracy in Venezuela after their return to democracy in 1958 was shared fear of a return to the costly conflict that preceded it. Similar to El Salvador after its brutal civil war—another iconic case of “pacted” democracy— Venezuela had learned from the violence under military rule that none of the rivals was strong enough dominate the others. Without a significant advantage in de facto power, those rivals could expect to gain more from sharing the spoils of the country’s natural-resource wealth and the

Exploring Confounding Cases

115

economic growth it fueled than they could from renewing the fight for sole control. As Karl (1987: 85) summarizes, the Venezuelan pacts “represented a classic exchange between ‘the right to rule for the right to make money.’” This cooperative approach was embodied in the policies the government pursued under President Romulo Betancourt, who took office in January 1959. On the first anniversary of his administration, President Betancourt announced to a crowd of tens of thousands that, after four months of negotiations, the country’s leading oil companies had reached a settlement with the industry’s 42,000 workers which would provide those workers with substantial new benefits, including annual wage increases of more than 10 percent and totaling $100 million, nearly half of which would be paid by the government; limits on the length of the work week; and fringe benefits such as paid leave. 2 This contract, described by The New York Times at the time as “one of the most generous in the oil industry,” explicitly committed to sharing the spoils of Venezuela’s resource wealth with the workers whose cooperation was required to produce them. Equally important to actors’ expectations about future payoffs from continued cooperation, the agreement was widely expected to help boost growth in the wider economy, in part by encouraging new investment.3 In spite of the apparent success of the agreements forged at Punto Fijo, democracy did not go unchallenged in Venezuela during this period. In January 1960, an alleged coup plot involving military officers loyal to ousted ruler Brig. Gen. Marcos Pérez Jiménez and apparently supported by the government of Dominican dictator Rafael Trujillo was disrupted by the arrest of one of its apparent leaders, a civilian lawyer who had headed the country’s Chamber of Agriculture.4 Riots and purported coup plots by leftist elements—some with alleged support from the Dominican Republic and inspiration from Cuba—led President Betancourt and Venezuela’s Congress to suspend freedoms of speech and assembly and protections against arbitrary arrest in late 1960 and again in early 1961. Those measures did not prevent a failed rebellion in February 1961 in which bands of armed men, some of them from the ranks of the county’s military, briefly seized police barracks and a radio station.5 Foiled coup plots and attacks on the Communist Party continued for a few more years, at which point the erstwhile rebels effectively sued for peace in an effort to salvage some political gains in spite of their apparent inability to seize power outright.6 In the meantime, however, they demonstrated that even agreements as comprehensive as the ones struck at Punto Fijo could guarantee the survival of democracy.

116

Dilemmas of Democratic Consolidation

If pacting leads to the consolidation of democracy by institutionalizing norms of cooperation and compromise, then Venezuela should still be a democracy today. Likewise, if consolidation depends primarily on the passage of time, the observed alternation in power, and the associated habituation to democratic procedures, the survival of democracy in Venezuela should have been virtually guaranteed after the first couple of decades. Unfortunately, Venezuela is not a democracy anymore, and that outcome tells us that neither the instrument of a pact nor habits of cooperation have as clear an effect on the survival of democracy as many theories of consolidation would suggest. In Venezuela, the commitment to cooperation that appeared to solidify in the early 1960s eroded when the spoils from the oil boom began to run out. The decline in oil prices that began in 1981 led to a decline in living standards to the point that many Venezuelans believed they were experiencing “the worst economic crisis in a generation,” according to The New York Times. The sharp drop in state revenues led to a debt crisis to which the government responded with a structural adjustment program that proved broadly and deeply unpopular, sparking major riots in 1989 and 1990. Describing the impact of the fiscal crisis, Coppedge (1994) writes that, “Fewer resources were available for patronage or even for meeting the routine obligations of the state; public services declined, and infrastructure was allowed to deteriorate. The parties lost some of their ability to fulfill their promises, to co-opt new organizations…and to provide government jobs for friends and (former) enemies.” Democracy survived in Venezuela for nearly three decades under the politics of concertación orchestrated by the Pact of Punto Fijo. Once the mutual gains on which that arrangement was predicated began to disappear, however, the commitment to cooperation crumbled, and Venezuela suffered a series of attempted coups and rebellions that led eventually to a return to authoritarian rule under President Hugo Chávez. The first blows were struck in the early 1990s by junior officers in the country’s armed forces whose families were suffering from the economic malaise, even as senior officers continued to receive lucrative salaries and benefit from corruption. A leftist putsch led by then-Colonel Chávez fizzled in December 1991, and a second grab for power by a group of more centrist officers failed the following November. Chávez was elected president in 1998 and then reelected in 2000 following the passage by referendum of a new constitution the year before. When the Chávez government tried in 2002 to reform the state oil company PDVSA, his political opponents launched mass protests and a general strike. This resistance culminated in a coup on April 11

Exploring Confounding Cases

117

that involved some senior military commanders and briefly landed Chávez in jail but lasted less than two days before the Presidential Guard backed by throngs of Chávez loyalists restored the president to office. Over the next few years, Chávez’s government restricted civil liberties and stacked key state bodies with loyalists in a process that amounted to a rolling executive coup. After surviving a recall referendum of questionable quality in 2004, Chávez stacked the country’s electoral commission and Supreme Court with political supporters, and questions of fraud in voter registration and votecounting dogged the country’s December 2005 legislative elections in which Chávez supporters won virtually all the contested seats. Had either one of the military coups of the early 1990s stuck, or had the uprising of 2002 succeeded in ousting Chávez and claiming power for the opposition—all plausible counterfactuals—the mode of democratic breakdown would have been different, but the basic outcome would have been the same. Consequently, the crucial puzzle in this case is the timing of the switch from purportedly consolidated democracy to vulnerable regime. What my theory seeks to explain is not precisely when democracy ended under Chávez’s rule, but why coup attempts began occurring in the early 1990s, after so many years without them. Once usurpations of power were back within the realm of the possible, the credibility of the parties’ and the military’s commitment to uphold the democratic rules of the game was severely eroded, and the basic social dilemmas highlighted by my model were back in play. The resulting mutual insecurity led to a series of attempts to usurp power by each of the players: first the conservative wing of the military in 1992, then the opposition under Chávez with the failed rebellion of 2002, then Chávez himself after that. Thailand 1992–2006

On 19 September 2006, military leaders in Thailand seized power in a bloodless coup d’etat, ending an episode of democratic government that began fourteen years earlier, in September 1992. With Prime Minister Thaksin Shinawatra traveling overseas, top officers declared martial law, suspended the constitution, and dismissed the government, both houses of parliament, and the Constitutional Court. The military’s seizure of power broke a months-long impasse between the country’s leading political parties that saw near-constant rallies in Bangkok by supporters of both sides. Prime Minister Thaksin had tried to break the impasse by calling snap elections in April 2006, but the leading opposition party boycotted the vote, which was then

118

Dilemmas of Democratic Consolidation

declared invalid by the Constitutional Court on grounds that the positioning of voting booths in polling stations had failed to ensure the secrecy of the ballot.7 In the run-up to those elections, Thaksin had announced that he would not serve as prime minister when parliament reconvened, delegated his powers to his deputy, and took what amounted to a seven-week vacation. The invalidation of the April balloting pushed the timetable for fresh elections into the fall, but the military acted before those elections were held. The officers who led the coup said they were temporarily replacing parliament with a “council of administrative reform” on behalf of the country’s king, and military spokesman Gen. Prapas Sakultanak apologized to the public for any “inconvenience” caused by the coup. As noted in the Introduction to this book, the breakdown of democracy in Thailand in 2006 was noteworthy because it contradicted several widely held beliefs about the conditions under which democracy should survive. For one thing, the September coup terminated the longest episode of democratic government in Thailand’s history, fourteen years. According to some theories of consolidation centered on the idea of learning and habituation, fourteen years is a long time for the development of habits that allow rivals to solve political problems through cooperation rather than confrontation. Thailand’s coup also occurred after several years of rapid economic growth, apparently contradicting the belief that regime performance is the key to democratic survival. Last and probably most significant, the 2006 coup put Thailand into a small group of countries that have suffered a breakdown of democracy after its per capita income had risen to middle-income levels. According to the World Bank’s World Development Indicators, in Thailand in 2006, per capita income measured in constant 2000 $US was $2,600, on par with Chile in the early 1980s and Turkey in the early 1990s. As Przeworski et al. (2000) demonstrate, at those income levels, democracy very rarely breaks down. Thus, the 2006 Thai coup contradicted several widely held beliefs about the conditions under which democracy is expected to survive. Although none of the theories on which these beliefs are based predicts with certainty that democracy will survive in situations like Thailand’s, the Thai case is striking for failing despite having so many elements seemingly working in its favor. What was it, then, that propelled politics in Thailand from the apparent stabilization of democratic rule in the 1990s toward political polarization, confrontation, and, eventually, military intervention? Viewing the case through the lens of the theory developed here, the change in the tenor of Thai politics in the 2000s was probably caused by

Exploring Confounding Cases

119

the rapid rise of Prime Minister Thaksin Shinawatra’s Thai Rak Thai (TRT) party. The TRT’s rise, in turn, resulted from the coincidence of two exogenous “shocks” to Thai politics: the Asian financial crisis of 1997, and the strategic leadership and financial wealth combined (serendipitously or unfortunately, depending on whom you ask) in the person of Mr. Thaksin. Once TRT had taken power, its apparent lock on Thai elections for the foreseeable future altered its rivals’ expectations about their welfare under a democratic regime, leading them to prefer a return to authoritarian rule in spite of the associated risks and costs. Thaksin Shinawatra, one of Thailand’s wealthiest businessmen, founded the Thai Rak Thai (Thais Love Thais) party in July 1998, one year after his government’s decision to float the country’s currency helped to trigger a financial crisis that swept Asia and unsettled markets around the globe. That financial crisis spurred significant protests among Thai farmers and people engaged in the country’s urban informal sector, two proportionately large but politically underrepresented segments of the country’s population (Phatharathananunth 2008). From the start, Thaksin and TRT adopted a platform aimed at appealing to these constituencies by promising cheap and universal health care, agrarian debt relief, and village development funds. The party also used its financial resources to lure members of parliament away from other parties under its umbrella, eventually assembling the coalition that nearly won an outright majority in its first general election and swept to a substantial and unprecedented majority in its second one, in 2006. The financial crisis energized and aligned the interests of large constituencies in the Thai electorate that had held little sway in the past. In May 1998, Thai farmers engaged in some of their largest protests in years. The energy evident in those protests, however, was only converted into election victories with deliberate direction from Thaksin, who possessed an unusual combination of strategic leadership, political ambition, and personal wealth. Absent his very deliberate and clever efforts to capitalize on the swelling discontent and his possession of the financial resources required to do so successfully, it is unlikely that Thai politics would have been as thoroughly realigned as they were. The meteoric rise of the TRT party effectively marginalized the groups that had held power in Thailand since the restoration of democracy in 1992. What’s more, those groups perceived TRT’s policies as profoundly threatening to their interests, and they believed TRT was attempting to cement its hold on power through a slowburning executive coup. Describing the TRT’s first four year in office, Brown (2007) writes that Thaksin sought “to restructure Thailand’s political economy” through actions that “entrenched his party's control

120

Dilemmas of Democratic Consolidation

over parliament, bureaucrats and the state apparatus, limited the political space of civil society, imposed greater discipline on society, developed new strategies for capital accumulation and negotiated a new social contract with subordinate classes.” This project depended on the construction of a “formidable political and economic power network” that spanned several large business conglomerates and political parties and reached into the police and military; as it happened, this network also displaced the previous one that tied together “prominent bureaucratic and aristocratic families and the Democrat party” and centered on Thailand’s constitutional monarch. The concerns of this “old guard” faction aligned with the fears of many members of Thailand’s urban middle class. According to Phongpaichit and Baker (2008), “The evolution of Thaksin’s populism and the growth of middle-class opposition were interrelated in a kind of dialectic: as Thaksin lost urban middle-class support, he intensified his populist appeal to the informal mass; as Thaksin’s populism became more strident, the middle class felt more alienated. Thaksin’s threat to bring millions of rural supporters into the capital was the logical conclusion to this spiral, and a key trigger for staging the coup. Beneath this interplay lay the massive gap in incomes between city and village, and a long-standing middle-class fear of empowerment of the rural mass. The threat which the middle class perceived in Thaksin’s populism was partly fear that they would be obliged to pay for his redistributive schemes, but more fear that they would no longer have a privileged position to influence the state agenda.” Whatever the merits of the claims that Thaksin was unfairly cementing his party’s hold on power, the TRT clearly owed much of its success to its real popularity among large swaths of the Thai electorate, and in the mid-2000s it seemed clear that TRT would keep winning elections for the foreseeable future.8 At the same time, Thaksin did little to assuage fears that his government was bent on cementing its hold on power by undemocratic means. If anything, he explicitly fueled those concerns. As Phongpaichit and Baker (2008) describe, “Thaksin devalued the importance of parliament, neutralised the check-andbalance bodies of the 1997 constitution, micro-managed the electronic media and said in public that law, the rule-of-law, democracy and human rights were not important because they often got in the way of ‘working for the people.’” These fears also gained credence through court rulings regarding electoral malfeasance on behalf of some TRT deputies in the 2001 and 2005 contests. In short, the political parties that held sway in Thailand during the 1990s found themselves in opposition in the 2000s, facing a government

Exploring Confounding Cases

121

that was, in their eyes, inimical to their interests and attempting an executive coup. In game theoretic terms, we might say that the former ruling party’s expected payoff from democracy had been greatly diminished, and the group had reason to fear that its least-preferred outcome, a TRT dictatorship, would come to pass if they did not move more forcefully to prevent it. While this party might still have anticipated a large payoff from democracy were it to govern, the odds that it would win the next election were obviously long, so its expected payoff from democracy was determined primarily by the (small) value it assigned to living in a democracy led by the TRT. Recognizing that they could not defeat the TRT at the ballot box—even in a fair fight—this group began to pursue nonviolent rebellion as a means to improve its supposedly dire situation. The mass protests mobilized under the banner of the People’s Alliance for Democracy (PAD) in 2006 paralyzed politics in Thailand and induced a constitutional crisis. The country even found itself without a valid government for a brief period after the Constitutional Court invalidated the April elections the TRT had called in response to the protests. The rebellion never quite succeeded in removing Thaksin or the TRT from office or remaking the Thai regime, however. That “honor” fell instead to the country’s armed forces, which executed a successful and bloodless coup in September 2006, while Thaskin was traveling overseas. The rationale for the PAD-led uprising seems clear enough. Why, though, did the Thai military choose to try to usurp power? Two developments under the TRT government seem to have reordered the military’s preferences, aligning them with the country’s traditional political establishment: 1) the resurgence of a separatist insurgency in the country’s southern provinces, and, ironically, 2) Thaksin’s attempts to overhaul the leadership of the armed forces in order to ensure its loyalty to him and TRT. Muslims comprise a majority of the population in Thailand’s southern provinces near the border with Malaysia, and this region has suffered separatist violence for decades. That conflict ebbed in 1998 after a successful effort in coordination with the Malaysian government to enhance border security, but it flared anew in late 2001 and then surged in 2004. Between January 2004 and August 2007, more than 2,600 people were killed—roughly two each day—as a result of the war over southern autonomy (Chalk 2008). Defeating the southern insurgency has been a top priority for Thailand’s military leaders for decades, but the TRT government elected in 2001 pursued policies that many officers saw as inimical to that goal.

122

Dilemmas of Democratic Consolidation

A noteworthy example involved the new government’s decision in 2001 to dismantle a multi-agency coordination group, the SBPAC, that the military believed was an important instrument in its counterinsurgency efforts—a belief that was demonstrated by the junta’s prompt restoration of that group after the 2006 coup (Chalk 2008). The Thaksin government’s decision to subordinate the armed forces to the police in the fight against the southern insurgents was also connected to the other policy direction that military leaders saw as a threat. According to Chalk (2008), Thaksin’s decision to scrap the SBPAC “reportedly reflected his concern that the center was staffed by officials whose loyalties flowed primarily to the opposition Democrats rather than to him and his government.” These political concerns apparently motivated other appointments, promotions, and replacements under the TRT government as well, aggravating relations with between the civilian government and other military officers in the process. Taken together, the TRT government’s redirection of counterinsurgency efforts in the south and politicization of military appointments may have reduced many military leaders’ expected welfare under the extant democratic regime, a trend that would make military-led authoritarian rule look more appealing by comparison. Concerns among military officers about the costs of a TRT-led democracy were undoubtedly exacerbated by the mass political showdown in 2006. When the junta leaders explained why they intervened in September, one of the four reasons they gave was that Thaksin had “caused an unprecedented rift in society.” This rift threatened to explode into open conflict, an eventuality that could have proven quite costly to the military in both personal and organizational terms. Meanwhile, military leaders probably anticipated that they could execute a coup fairly cheaply. In spite of its electoral appeal, the TRT had never done much grass-roots organizing (Phongpaichit and Baker 2008). As a result, the party lacked the kind of organizational base that would have allowed it to easily rebuff the PAD-led rebellion or to engage in a direct confrontation with the country’s armed forces. At the same time, the coup leaders probably expected that many participants in the ongoing rebellion would support a military government as a secondbest outcome—still vastly preferable to the status quo and without the direct costs and uncertainty of continued rebellion—and demobilize in response. In sum, military leaders who were not personally loyal to Thaksin in mid-2006 found themselves living under an elected government that was increasingly costly to their interests and that they probably believed they

Exploring Confounding Cases

123

could successfully displace at little direct cost. Using a tactic common to many military coup attempts, they struck when Thaksin was traveling overseas, thus preventing him from rallying supporters in person and avoiding the costs of capturing and keeping him. As was probably anticipated, the PAD-led protests quickly evaporated in response, and Thaksin’s supporters failed to mobilize any significant challenge to the junta. Thai politics has taken many strange turns since the September 2006 coup. Democracy was restored under what amounted to a new TRT government in early 2009—the TRT itself was banned, and Thaksin was living in exile in London at the time—only to fall later that year when courts sympathetic to the TRT’s rivals effectively handed power to the last election’s losers in a December 2009 ruling with dubious legal merits. That ruling came after a surge in PAD-led street protests led to fresh violence in the capital, shut down Bangkok’s international airport, and crippled the country’s trade-dependent economy. Some observers speculated that the military quietly facilitated the rebellion by pressuring selected lawmakers to cross the aisle in parliament and support the new government. In other words, in a single year Thailand managed to repeat the cycle of democratization, polarization, and breakdown it had experienced in the fourteen-year episode that ended in 2006. For current purposes, however, the essential point is this: the theoretical model developed in this book can produce a narrative that credibly explains why democracy survived in Thailand in the 1990s only to fail in the mid-2000s, while most conventional theories of democratic transition and consolidation cannot. By focusing our attention on the interests of the actors involved and the dilemmas they confronted, we can see how a regime that ought to have been settling into “normal” democratic politics if theories rooted in modernization or habituation are to be believed instead came unglued.

Notes 1

See the preceding chapter for a discussion of the Spanish case. “Betancourt Ends a Year in Office,” The New York Times (February 13, 1960). 3 “Venezuela Hails Oil Industry Pact,” The New York Times (February 15, 1960). 4 “Caracas Reports Smashing a Plot,” The New York Times (January 21, 1960). 5 “Venezuela Quells Rebel Outbreaks,” The New York Times (February 20, 1961). 2

124

Dilemmas of Democratic Consolidation

6 “Venezuelan President Imposes Rights Curbs,” The New York Times (January 23, 1961); and “Venezuela Renews Curbs on Liberties,” The New York Times (February 3, 1961). 7 Booths were positioned such that voters had their backs turned to the queue of waiting voters while casting their ballots, an arrangement that arguably made it difficult to protect the secrecy of their votes. 8 Phongpaichit and Baker (2008) estimate that the agricultural and urban informal sectors which generally supported TRT comprise about two-thirds of the national electorate.

6 Implications for Democracy Promotion

What does this book add to the conversation among policymakers and democratization practitioners about how to help countries improve the odds that their democratic regimes will survive? First and foremost, I hope my theory encourages or reminds individuals working in this field to think carefully about the ways that their words and deeds might shape the strategic considerations of individuals and groups in the situations they are seeking to affect. Democratization is not just a process of institutional and organizational engineering; it is also, if not more so, a high-stakes political fight. A theory that underscores the strategic aspects of democratic consolidation encourages us to consider how even highly technical interventions can shape incentive structures and influence the beliefs of actors in ways that can have unintended as well as intended consequences.1 In short, a game-theoretic approach to democratic consolidation reminds us to think politically as well as programmatically. O’Donnell and Schmitter (1986) were right when they likened the strategic complexity of the transition from authoritarian to democratic rule to a multilevel game of chess. Somehow, though, the wrong lesson has been drawn from that analogy by scholars and practitioners who seem to have concluded that the preeminence of strategy over structure, of decisions over predestination, means that anything is possible. Just like a game of chess, the strategic possibilities in any given situation are heavily constrained by the prior arrangement and abilities of the pieces and the player’s expectations about their rival’s future actions. Players may have many courses of action open to them, but the rules and trajectory of the game will usually mean that only one or a few of those options look promising at any given moment. My models shows that even individuals who are firmly committed to democracy in principle can be tempted to usurp power by uncertainty about the capabilities and strategic intentions of others. This finding implies that attempts to sustain democracy by supporting elites who

125

126

Dilemmas of Democratic Consolidation

seem to want its survival, or by convincing citizens that democracy is the best form of government, will not always be effective. We should be able to do better by helping establish incentive structures that specifically discourage coups and rebellions and provide material rewards for continued adherence to democratic practices. In a broader sense, this finding also helps to turn our focus away from the fuzzy concepts of “legitimacy” and “good governance” as the foundation for democracy’s survival toward sharper assessments of the interests, organization, and expectations of specific, powerful actors. If incentive structures are so important, then different interventions will be appropriate in different cases, depending not only on structural factors such as the level and character of economic development but also on the capabilities and vested interests of the specific organizations involved in the democratic process. It is trite to say that democracypromotion efforts should be tailored to local conditions. I hope my theory gets beyond that platitude by providing another lens through which policymakers and practitioners can examine those local conditions, deduce key concerns of powerful actors, and derive strategies for diplomacy or assistance that are sensitive to those concerns and the ways in which the act of intervention itself will affect them. Determining which kinds of interventions are likely to be constructive and which might be counterproductive in a particular case requires careful assessment of the interests, capabilities, and beliefs of the relevant actors, and of the ways those elements combine to shape the actors’ strategic concerns. This book’s model does not and is not meant to produce generic prescriptions for democracy promotion worldwide. What I hope this model does instead is contribute a new tool to that diagnostic process—a tool that emphasizes the social dilemmas as well as the opportunities inherent in democratic government. The remainder of this chapter examines what the model suggests about some the common practices and concerns in this field. Assessing the Practice of Democracy Assistance

An approach to democracy assistance that concentrates on shaping the strategic incentives confronting political organizations in the recipient country would run contrary to key assumptions underlying much of the work done today by Western governments and other international actors in this area. A relatively new invention in the field of international development, democracy assistance refers to aid targeted directly at fostering the establishment, survival, and deepening of democracy.2 The idea of foreign development assistance aimed specifically at effecting

Implications for Democracy Promotion

127

political rather than economic change really gained currency in the 1980s, when international financial institutions became increasingly aware of the effects of political institutions on economic growth. The practice was boosted further by the collapse of the Soviet Union and the flurry of new attempts at democracy that ensued in countries eager to integrate further with the West. By the 1990s, many policymakers, scholars, and development practitioners were drawing stronger connections between democracy and other desired outcomes, including economic growth, but many had also grown increasingly confident in casting political development as a good in its own right. In a widely cited 2004 essay, Thomas Carothers, a longtime and generally sympathetic observer of the practice of democracy promotion, incisively critiqued this field for its teleological bent.3 As Carothers describes, countries moving away from specific authoritarian regimes are assumed to be transitioning toward democracy, and that transformational process is expected to unfold as a sequence of distinct stages, each of which involves certain “challenges” which external interventions can be designed to help overcome. Countries may stall or stray in their journey—words like “feckless” and “backsliding” are often used to describe cases which spend more than the desired amount of time constructing or consolidating democracy—but once that journey has started, liberal democracy is presumed to be its only possible end. While I concur with Carothers’ critique of this teleological mindset, I am equally concerned about the field’s functionalist foundations. Closely related to the transitions-as-sequences concept that emerged from the “transitology” school is the assumption that democratic government consists of a set of modular but interconnected systems, each of which serves a specific function required to sustain the health of the system as a whole. Thus, for example, the website for the Democracy and Governance (D&G) office in the U.S. Agency for International Development (USAID) describes how its work is organized around four technical areas it considers to be the “building blocks” of democracy: Elections and Political Processes, Rule of Law, Civil Society, and Governance. These technical areas and their components are often described much in the way that an auto mechanic might speak of a car’s drive train, cooling system, and exhaust system. Each element of the system is assumed to have universal features and functions, and the performance of those elements is assumed to be a technical matter amenable to technical solutions. Taken together, these assumptions comprise a mindset that bears a closer resemblance to civil engineering than it does to the practice of politics. This engineering mindset, in turn, suggests that democracy can

128

Dilemmas of Democratic Consolidation

be promoted most effectively by identifying which of the systems have flaws or deficits and then applying technical knowledge and material resources to repair those particular flaws. Some practitioners may argue that the sequencing of these improvements matters, but the basic logic remains the same: as the deficits in specific functional areas decline, democracy becomes more consolidated, and the odds that democracy will survive improve.4 The trouble with this mindset is that it is not sufficiently sensitive to the capacity for strategic behavior by individuals and organizations in the targeted societies and elsewhere whose interests are affected by these well-intentioned efforts. As a result, those efforts often fail to give adequate heed to the possibility of unintended consequences. Unlike the parts of car engines, human individuals and organizations do not respond mechanically to stimuli; instead, they think and plan and learn. In a word, they have agency. This human capacity for agency means that attempts to influence the course of human social development cannot be reduced to a series of isolated diagnoses and treatments. These “building blocks” are abstractions of our own creation, and the facts of human agency and intelligence mean that the development of democracy is not as mechanical and modular as the conventional view assumes. How did we get here? In a review of U.S. democracy-promotion efforts in the 1980s and 1990s, Carothers (2004: 5) suggests that the current approach to democracy assistance resulted from a confluence of trends in scholarly work on democratization and real-world regime changes. His description of the learning dynamic in this field is worth quoting at length: Every few years some new idea is embraced as the key to unlocking the democracy puzzle. Since the late 1980s, one enthusiasm after another has enjoyed a brief, intense run—including elections, civil society, rule of law, decentralization, and anticorruption. As a fad takes hold, aid groups rush to create programs in that area, often shifting resources from other work and investing in the new activities with great expectations. In parallel fashion, enterprising people in recipient countries demonstrate a newfound interest in the topic and quickly start up work on it, or strategically re-label what they are already doing to fit the new fashion. A boom period follows, but then within a few years the hoped-for dramatic results do not appear and cracks in the edifice start to show. Restless aid providers move on in search of a new romance. Each fad rests on some degree of insight…. But the urge to embrace fads reflects an unhelpful attachment to the idea that democratization is amenable to magic bullets and the lack of a well-grounded and well-accepted base of knowledge about the process.

Implications for Democracy Promotion

129

For Carothers, then, the field is plagued by a quixotic search for the “keys” to democratic development—a search that he believes is driven by the an absence of reliable technical knowledge coupled with an optimistic view of the possibilities for social engineering. I agree with Carothers that the paucity of relevant research and penchant for optimism have had important effects on the practice of democracy promotion. At the same time I believe the current can-do, technical approach also stems from issues of organizational design. In his landmark study of public administration, Wilson (1989) shows how the behavior and effectiveness of government agencies depends substantially on the way they are organized. Wilson is particularly concerned with how agencies define their critical task and then coordinate and execute the activities they undertake to fulfill that task. In analyzing the effectiveness of U.S. agencies, he underscores the importance of constraints placed on those agencies by Congress and the president. In the American system of government, agencies are not free to define their missions and design their tasks as they see fit; instead, they are usually subject to extensive efforts by policymakers and interest groups to control their actions in order to satisfy an array of sometimesconflicting interests. Viewed through the lens of Wilson’s theory, USAID faces a steep challenge in its efforts to design programs that actually help vulnerable democratic regimes survive. The trouble starts with the fact that the agency’s mission in this area—“promoting and consolidating democracy worldwide,” according to the web site of USAID’s Democracy and Governance (D&G) office—is inherently vague. According to Wilson (1989: 34), when goals are vague, what operators actually end up doing to accomplish those goals will be shaped “by the circumstances they encounter at the job, the beliefs and experiences they bring to the job, or the external pressures on the job.” In the realm of circumstances encountered, Wilson emphasizes the importance of technology, by which he means not computers and such but available tactics or tools of all kinds. Put simply, operators within agencies will tend to use the tools most readily available to them, even when those tools may not be the most effective means to accomplish the original objective. Carothers’ argument about the influence of the “transitology” perspective on strategies for democracy promotion speaks nicely to Wilson’s point about the effects of prior beliefs and experiences on organizational behavior. Here, though, I am more concerned with Wilson’s observation about the influence of technology. For a variety of bureaucratic and legal reasons, the core tactic available to USAID’s D&G office is programmatic assistance: allocating money to fund self-

130

Dilemmas of Democratic Consolidation

contained, short-term programs with specific technical objectives. These programs are not carried out directly by USAID but are instead designed and implemented by contractors—what USAID’s D&G office refers to as “implementing partners”—who bid for slices of the funds in USAID’s D&G budget in response to requests for proposals from that office.5 As independent organizations, those contractors bring their own interests to the table, and those interests are not always neatly aligned with USAID’s fuzzy goals of expanding or extending democracy. This contractual relationship leaves USAID with what economists call a principal-agent problem, which refers to the difficulties involved in “motivating one person or organization to act on behalf of another” (Milgrom and Roberts 1992: 214). Because USAID does not directly execute the programs it funds, it has to concern itself primarily with selecting contractors (its “agents”) and then monitoring their performance and compliance. One of the most common strategies for monitoring performance in a situation like this, and one on which USAID relies heavily, concentrates on making a programs’ objectives as concrete and measurable as possible: hold X seminars, train Y judges, buy Z ballot boxes. This approach can help overcome on part of the principal-agent problem by making it easier for the agency to monitor its contractors’ compliance. In so doing, however, it can also distract the parties to that exchange from the larger question of how those actions might or might not be serving the larger goal of developing democracy. In short, the basic challenge for democracy assistance undertaken by USAID—and, presumably, by similar agencies in other countries and related non-governmental organizations—is that money cannot buy democracy per se. What money can buy is people’s time and materials in support of actions that are presumed to lead to democracy or to help sustain it once established. Just how democracy works and how these arrangements arise, however, is still poorly understood. This scarcity of systematic knowledge means that operators in this field end up devoting their time to supporting narrowly defined projects with practical outputs that can be linked only loosely to the original task. The National Research Council’s USAID-funded Committee on Evaluation of USAID Democracy Assistance Programs (2008: 26) summarized the situation as follows: USAID and other democracy assistance agencies therefore face a difficult task. Practitioners’ reflections present informed viewpoints and policy research often presents thoughtful and systematic analysis, but their judgments about program success or failure are not rigorously tested according to academic standards. Yet since academic debates

Implications for Democracy Promotion

131

regarding democratization remain largely unresolved, they offer little practical guidance on what to do in a given country to build or sustain democracy. Policy professionals working in democracy assistance have therefore formed their own ‘practical wisdom,’ based on elements drawn from their readings of the academic and think tank literature, their own experiences, and what they glean from other practitioners. Policy professionals thus often describe democracy promotion as ‘more of an art than a science,’ where policy choices must depend on intuition and personal judgment as much or more than on any scientific guidelines.

Here again, a thoughtful and sympathetic observer has described the practice of democracy promotion in a way that is not especially flattering. Given the inherent complexity of democratic development and the fundamental challenges of public administration, however, it is hard to imagine how government agencies could approach democracy promotion any other way. In other words, the modular and technical character of democracy assistance is largely built into the system. Unfortunately, this approach does not seem to be as effective as we would hope, at least on the scale on which it has been attempted so far. A USAID-funded but independently conducted study on the impact of U.S. assistance in this area found a statistically significant effect from some types of D&G assistance on recipient countries’ subsequent levels of democracy. As that study’s authors (Finkel et al. 2008: 2) describe, The positive impact is such that $10 million of USAID DG funding would produce an increase of more than one-quarter of a point…on the 13-point Freedom House democracy index in a given year—or about a five-fold increase in the amount of democratic change that the average country would be expected to achieve, ceteris paribus, in any given year.

In short, this study found that U.S. DG assistance does have a statistically significant effect on the development of democracy in recipient countries. What the preceding excerpt does not make clear, however, is that the effect they describe is substantively small.6 A quarter-point increase on a 13-point scale is probably not the kind of change that would catch most observers’ eyes, and it certainly is not indicative of the kind of transformational outcomes that democracy promoters aim to achieve.7 The modesty of that effect is more even apparent when we recognize that $10 million is not a trivial share of USAID’s DG budget. According to the data analyzed by Finkel et al., total USAID spending on DG aid (in constant 1995 dollars) was just $106 million in 1990, the first year in their period of observation.

132

Dilemmas of Democratic Consolidation

USAID DG spending had risen to $830 million by 2003, but by that time the office was spreading its assistance across scores of countries each year, and a smaller number of high-priority countries were receiving a disproportionate share of those resources. Meanwhile, perhaps the only large-scale research project to date that has systematically investigated the impact of democracy assistance in post-conflict countries—an subject of particular concern in the 1990s and 2000s—reached a more pessimistic conclusion about the effects of democracy assistance in these cases. According to de Zeeuw (2005: 481): Post-conflict democracy programmes consist mainly of technical, material, and financial assistance as well as short-term project aid. This aid may have spurred the growth of many training activities and NGOs that excel in organizing workshops and seminars, but proves unsustainable and largely insignificant in the wider process of democratization.

As the author somewhat wryly notes (p. 499), “The creation of recipient organizations is often mistaken for democratic institution building.” This statement echoes Carothers and Ottaway’s observation (2000), based on a comparison of roughly a dozen cases from all parts of the world, that Western efforts to foster the growth of non-governmental organizations and other elements of “civil society” have not had the democratizing impact the heady rhetoric on this topic sometimes seems to promise. Donor governments are increasingly sensitive to these problems, and some have tried to overcome it in recent years by demanding that their bureaucratic agents do a better job assessing the effectiveness of their work. According to Green and Kohl (2007: 2), “Legislatures are increasingly requiring that government aid agencies be able to measure the results of their programs, thereby demonstrating a ‘return on investment’ that would guide future assistance.” For its part, USAID’s D&G office has undertaken a long-term and costly initiative, called the Strategic and Operational Research Agenda (SORA), to start doing just that: measuring its impact to date and developing systematic knowledge about the effectiveness of specific forms of assistance. As of mid-2008, this initiative had produced two major and impressive pieces of work: a meticulous but still very general statistical analysis of the effects of D&G assistance on the level of democracy in recipient countries (the aforementioned Finkel et al. 2007); and a report prepared by a panel of experts convened by the National Research Council of the National Academy of Sciences to help USAID think about how to assess its own

Implications for Democracy Promotion

133

performance in this area (Committee on Evaluation of USAID Democracy Assistance Programs 2008). This call for accountability, however, cannot resolve the difficulties inherent in designing practical tasks to achieve fuzzy goals on behalf of principals with diverse and sometimes even conflicting interests. After all, how exactly can we measure “return on investment” in an area such as democratization when the nature of democracy itself remains subject to intense debate, even among its advocates? The push for accountability from legislators inevitably confronts many of the same problems USAID faces in dealing with its “implementing partners” in this area, and a devotion to metrics as the solution to these problems produces the same distorted incentives. If the practitioners of democracy assistance are held strictly accountable for trends in quantitative measures of democratic development, they will focus their work on moving those measures, whether or not that movement actually produces the broader outcomes their principals originally set out to achieve. Given the complexity of the problem and the lack of reliable knowledge about how external interventions affect political change, it is difficult to see how a rush to metrics-based accountability would improve the efficiency of spending in these areas or lead to better outcomes. Finally, I think it is important to acknowledge that the search for a “well-accepted base of knowledge about the process” of democratization may be quixotic as well. The technical challenges involved are certainly real, and efforts to carefully analyze the impact of past projects—a task in which USAID began to invest more heavily in the mid-2000s— probably can improve the effectiveness of work in this area. That said, who does the assisting, who gets assisted, and how that assistance is implemented all matter, too. The process is not modular, and the bigpicture effects of a particular kind of intervention are usually contingent on a host of forces to which project-specific assessments and “best practices” thinking generally cannot be sensitive. Assessing Specific Types of Intervention

If the current approach suffers some important flaws, what should policymakers and practitioners seeking to promote the survival of democratic regimes do? I will not pretend that my theoretical model offers a clear and simple answer to that question.8 Unfortunately, these challenges inhere in a process in which governments must inevitably use narrow technical tools to try to bring about grand political ends. Instead, I can use that model as a lens through which to re-examine the promise and pitfalls of some of the most commonly used or intensely debated

134

Dilemmas of Democratic Consolidation

approaches in the practitioner’s democracy-promotion “toolkit.” The remainder of this chapter takes up that task. Strengthening the Opposition

Since the early 1990s, a growing share of international democracy assistance has been devoted to fostering the development of so-called civil society organizations (CSOs) and, in some cases, to strengthening specific political parties. For diplomatic reasons, US government agencies generally restrict their activities to the former, but a variety of international NGOs and partisan foundations engage in the latter.9 The distinction drawn in the policy community between these two forms of assistance is based, in part, on the notion that assistance aimed at civil society is inherently less political and therefore more defensible in a world in which direct engagement in the domestic politics of other countries is considered undiplomatic. Support for CSOs is said to increase popular participation and strengthen accountability by supporting the growth of groups doing uncontroversially “good” things.10 According to the web site of USAID’s D&G office, Strengthening civil society is increasingly seen as a way to counterbalance the exercise of excessive authority by governments and economic and political elites, and as a way to encourage more open dialogue about public policy matters too often decided behind closed doors. A vibrant civil society can even provide recourse to justice through the work of human rights groups, especially in post-conflict situations.

Support for partisan activity is less common because it is hard to distinguish from efforts to manufacture specific political outcomes. Even with aid to individual political parties, however, donors typically justify the decision to provide assistance on technical grounds. The logic is as follows: political parties play an essential role in representative government as vehicles for the organization and expression of popular interests; ipso facto, countries with stronger parties have stronger democracies (Kumar 2004: 2). The idea that party strength is inherently good for the development of democracy rests on the premise that the problem in fragile democracies is to constrain the government and increase popular participation. While normatively persuasive, that argument is strategically flawed. In effect, it privileges the opposition’s view. If we widen our lens to consider the concerns of the incumbent party and the

Implications for Democracy Promotion

135

military, we can see how increased party strength might appear threatening, not reassuring, and that threat might motivate a coup that otherwise might not have happened. As deterrence theory suggests (Schelling 1980), when the goal is to prevent conflict and preserve cooperation, relative capabilities are sometimes more important than absolute ones. Even the supposedly apolitical assistance aimed at building civil society can have political repercussions that affect these strategic concerns. As a political currency, organization is fungible, and the term “empowerment” means just that. Even groups with nonpartisan charters can directly or indirectly facilitate collective action and advocate for certain policy outcomes, and their missions will often align them in some predictable way with one or another of the major political parties. As a result, even support to supposedly nonpartisan organizations inevitably influences the dynamics of partisan rivalries and thus the balance of political power in democratic regimes.11 These effects can help to produce cooperative outcomes, but they can also exacerbate conflict and heighten other actors’ fears of exploitation. What’s more, these programs do not need to be large or effective to have counterproductive consequences. Apart from the actual effects such programs might have, incumbent parties and the military may perceive the act of promoting opposition parties and NGOs as a signal that foreign governments endorse their rivals and would support those rivals in any future confrontation. In the language of this book’s gametheoretic model, these programs can strengthen rivals’ beliefs about not only the capacity of the opposition to rebel, but also the likelihood that such a rebellion would prevail if attempted, due partly to foreign support for those challengers. Under some circumstances, this change in beliefs could reduce the temptation to attempt an executive coup or military coup as intended. Under other circumstances, however, those changes in beliefs could strengthen incentives to usurp power. In sum, my model suggests that the “more is better” logic often used to justify this kind of assistance does not always hold. Under certain circumstances, actions that increase the capacity of opposition parties and nongovernmental organizations could increase the risk of executive coups and military coups by heightening incumbents’ and officers’ fears of exploitation. Contrary to the conventional wisdom, this result implies that the strategic goal of supporting democracy may not always be served by the tactic of strengthening parties and civil society organizations. As such, international actors engaged in democracy promotion should carefully consider domestic actors’ perceptions of

136

Dilemmas of Democratic Consolidation

how that capacity might be used before pursuing these kinds of programs. Events in Russia after the Cold War illustrate the problem. Although concerns about foreign support for domestic political opposition and non-governmental organizations were probably neither necessary nor sufficient causes of the erosion of democracy that occurred in Russia in the 2000s, those concerns do seem to have played a role. When Vladimir Putin assumed the presidency in 1998, Russia was hardly a model democracy, but it arguably met the minimal qualifications summarized in this book’s Introduction. After legislative elections in 2003 and a presidential election in 2004, however, it had clearly slipped below that threshold as a result of actions taken by the Putin government against regional authorities, independent journalists, and nongovernmental organizations—actions that led some Russia watchers to describe that country’s regime as “competitive authoritarianism” (McFaul and Goldgeier, 2005). After the disintegration of the USSR in the 1991, Western governments sought very publicly to promote democracy in Russia, and programs aimed at strengthening political parties and civil society organizations were an important part of their strategy.12 President Putin evidently cast a jaundiced eye of these activities from the start, but the success of the so-called Color Revolutions in Georgia, Ukraine, and Kyrgyzstan appears to have helped solidify his distrust of them. Georgia’s so-called Rose Revolution, in autumn 2003, was led by a movement, Kmara (Enough), that benefited significantly from Western financial and logistical support, and that point was not lost on the leaders of governments in neighboring countries (Beissinger 2007). The Rose Revolution and Orange and Tulip Revolutions that followed in Ukraine and Kyrgyzstan, respectively, appear to have solidified a view among incumbent governments in the region that democracy-promotion activities, even ones aimed at ostensibly neutral civil-society organizations, were Trojan horses for Western governments bent on expanding their influence in Eurasia by installing more sympathetic governments. The thinking of those incumbents may also have been influenced by the public pronouncements of US President George W. Bush who, for example, in a December 2003 speech at the National Endowment for Democracy, described the US invasion of Iraq as the beginning of a “global democratic revolution.” 13 As McFaul and Goldgeier (2005) describe it, the diminution of democracy in Russia began before the Color Revolutions, so those events and the democracy-promotion programs that facilitated them cannot be blamed for Russia’s backsliding. Still, the slide into

Implications for Democracy Promotion

137

authoritarianism does appear to have accelerated in the wake of those revolutions (Hassner 2008: 10), sometimes in ways that seem specifically to be reactions to the potential threats posed by Western democracy and governance assistance. Most significant, the Russian government in January 2006 passed a law requiring charities and nongovernmental organizations to register and inform it in advance of any projects they plan to conduct. In defending the legislation, President Putin said, “The continuing financing of the political activity from abroad should be, I think, in the state’s field of vision, especially if this financing is carried out through the state channels of other countries” and if the organizations “are, in fact, used as a tool of the foreign policy of other states.”14 What’s more, once the requirement had been adopted, the Russian government appeared to go out of its way to make the registration process as cumbersome as possible.15 Again, there is no question that the impulses driving the diminution of democracy in Russia under President Putin predated the Color Revolutions and extended beyond concerns about the threat of defeat, by means of election or rebellion, at the hands of what was then a relatively weak opposition. The push against foreign NGOs in the mid-2000s came at a time when the Putin government was also narrowing press freedom and taking dubious legal actions against oil tycoon Mikhail Khodorkovsky and other prominent Russians who dared to support political activities outside the government’s purview. Nevertheless, the Russian government’s concerns about foreign democracy-promotion programs illustrate the possibility that these activities can exacerbate the uncertainty and fear that are identified by my theory as an important incentive for some organizations to try to rig the system in their own favor or otherwise usurp power. Building Institutions

Another major channel for Western democracy assistance involves efforts to accelerate what some scholars call a process of “political institutionalization.” Diamond (1999: 75) describes this process as “strengthening the formal representative and governmental structures of democracy so that they become more coherent, complex, autonomous, and adaptable and thus more capable, effective, valued, and binding.” This process is said to have numerous virtues. According to Diamond, political institutionalization facilitates the development of trust among political actors and helps to deepen democracy by proscribing the power of the state and the military and strengthening representative organs such as the legislature.

138

Dilemmas of Democratic Consolidation

In practice, the judicial and legislative branches are major targets of assistance aimed at accelerating the “institutionalization” process. Support for these organs of government often falls under different headings—in USAID’s D&G office, assistance to the judiciary is categorized as “Rule of Law” activity while assistance to the legislature is identified as part of its “Governance” work—but the basic logic is the same. Assistance to the judiciary often involves guidance on code reform as well as training for judges, lawyers, and police with the goal of increasing their effectiveness and freedom from political influence. Meanwhile, support for legislatures is typically intended to enhance that body’s technical capacity by such means as improvements in infrastructure, changes in internal organization, standardization of procedures, and increased staff sizes. This book’s game-theoretic model supports the general idea that external support aimed at improving the functioning of some institutions can improve the prospects for democracy’s survival. What seems especially likely to reduce the temptation to coups and rebellion are programs that improve the technical capacity and political independence of bodies charged with overseeing and enforcing the processes and defending the enabling conditions on which democratic accountability depends. The administrative capacity of a country’s judiciary is fundamental to the credibility of the threat that individuals who attempt to usurp power will suffer significant punishment if they fail. When that threat is credible, it deters coups and rebellions by reducing the expected gains from those actions. This deterrent effect is enhanced when the credibility of the threat does not depend on the outcome of future elections—in other words, when the police and judiciary, as agencies, are politically independent. In this sense, the model implies that a capable and autonomous judiciary is a crucial bulwark against the subversion of democracy. Encouragingly and in contrast to many of the structural factors affecting the relevant actors’ expected payoffs, the capability and autonomy of the judiciary are qualities that external actors may be able to help enhance in a relatively short time, because they depend so directly on money. Such measures are unlikely to tip the balance in favor of democracy’s survival in cases where other factors are heavily stacked against it, but they could play an important role in more ambiguous cases. USAID’s empirical assessment of the impact of its democracy and governance (D&G) assistance in the post–Cold War period provides some corroborating evidence for this assertion (Finkel, Perez-Liñan, and Seligson 2006: 67–79). Looking at the effects of D&G aid by sector on various measures of democratization, the study found that U.S. spending

Implications for Democracy Promotion

139

on Rule of Law programs—the title USAID gives to its efforts to promote judicial effectiveness and independence—had a statistically significant and positive impact on a scalar measure of free and competitive elections, along with aid focused on Elections and Political Processes. By contrast, spending on Civil Society and Governance programs did not have a statistically significant effect on this outcome.16 One entity that has traditionally received much less attention from international donors than legislatures and judiciaries but may be equally if not more important to the survival of democracy is the election management body, or EMB. Surprisingly little empirical research has been devoted to assessing the effects of the design and functioning of these bodies on the quality of electoral administration. One reason may be that EMBs are a relatively recent innovation, or at least new to catch on. According to data reported by López-Pintor in 2000, elections in roughly one-fifth of all democracies worldwide are still run directly by a branch of the government, the practice that generally prevailed before World War II (Birch 2007); in one-quarter of democracies, the government runs elections with supervision from an (allegedly) independent body; and in about one-half of all democracies, an (allegedly) independent commission actually runs the elections. The lack of attention paid in the academic literature to the institutional aspects of election administration is disappointing because management practices in this area so heavily influence actors’ perceptions of the fairness of elections, and claims that elections were unfair are toxic to democratic consolidation.17 As Pastor (1999) observes, technical irregularities are almost inevitable in modern elections—especially but not exclusively in poorer countries—and the perception that such irregularities resulted from political malfeasance can be the catalyst to a regime breakdown. As a result, Pastor (p. 5) argues that “the character, competence, and composition of Electoral Management Bodies (EMBs) can determine whether an election is a source of peaceful change or a cause for serious instability.” Consistent with Pastor’s argument, my game-theoretic model suggests that effective and independent EMBs can improve the prospects for democracy’s survival in a couple of ways. First and most obvious, effective EMBs should reduce the risk of an executive coup by making it harder for the incumbent party to manipulate electoral outcomes in its favor. Election manipulation is only one of a few modes by which incumbent governments can dismantle democracy, so sound and evenhanded election administration does not cure a regime of the risk of this form of failure. That said, election manipulation is often the subtlest path to an executive coup, so the costs associated with carrying

140

Dilemmas of Democratic Consolidation

out this kind of usurpation will generally be higher, and thus the incentive to do so will be lower, when that path is blocked. Second, by making it harder to cheat, effective EMBs can also reduce the opposition and the military’s uncertainty about the prospect of an executive coup. This reduces those actors’ fear of exploitation, a fear that can propel them to attempt a military coup or rebellion. In cases where independent EMBs are organized only after democracy’s onset, the very act of establishing that body may have a powerful effect on actor’s beliefs as well. In game-theoretic terms, that establishment or enhancement of an EMB represents a costly signal from the incumbent to the opposition and the military that the incumbent party is not planning to attempt an executive coup. The costliness of this signal should enhance the credibility of the incumbent’s commitment to democracy and thus facilitate cooperation among the contending parties. The hypothesis that independent and effective EMBs improve the odds that democracy will survive is at least tentatively supported by the small body of empirical work conducted on this topic to date. In one relevant study, Pastor (1999) examined fifty first, or “transitional,” elections in developing countries during the period 1985–1994 and found a strong association between the presence of an independent electoral commission (EC) and the quality of those elections. Of twentythree elections that “failed,” only three were overseen by independent commissions. Of the twenty-three that “succeeded,” all were overseen by electoral commissions, twelve of which were deemed (by Pastor) to be fully independent and eleven of which were deemed partially independent. Based of this evidence, Pastor (1999: 18) concluded that, “The absence of independent ECs is likely to lead to ‘flawed’ elections, but the establishment of independent ECs is not sufficient to assure successful elections.” In a second relevant study, Birch (2007) analyzed the impact of EMBs in twenty-four fledgling democracies in Eastern Europe and the former Soviet Union during the period 1995–2005. The comparability of these cases is enhanced by the fact that thet came into being around the same time and, because of their common Communist legacy, confronted similar challenges in the development of their practices of election administration. Using a statistical model that also controlled for the type of electoral system and the general level of administrative corruption in the country, Birch found that the design of the EMB affects the level of electoral corruption much more than its nominal independence does. Specifically, her model shows a statistically significant, strong, and negative effect from the number of parties that participate in appointing the members of an election commission on observed levels of electoral

Implications for Democracy Promotion

141

corruption. She also found a significant but less powerful effect from the number of branches of government involved in those appointments. The evidence on which these two studies are based is far too limited to draw firm conclusions about the impact of EMBs on the survival of democracy. That said, the model developed in this book is consistent with a growing body of theoretical work on election administration in suggesting that the presence of an electoral management body which is administratively effective and politically independent (or at least multiparty) should improve the prospects for democracy’s survival under most conditions, and the empirical analysis reported so far supports that claim. Bearing this in mind, practitioners of democracy promotion may wish to consider devoting more of their resources to this element of the democratic process than they currently do. Observing Elections

For some of the same reasons that capable and independent electoral management bodies should improve the odds of democracy’s survival, the presence of international elections observation missions can also help reduce the risk of reversion to authoritarian rule. As with EMBs, however, the impact of international election observation missions on players’ incentives to usurp power depends on those missions’ ability to observing manipulation if it happens and on the credibility of their commitment to reporting it when they do. Once again, there is surprisingly little rigorous empirical analysis on this question, but the best analysis so far supports the hypothesis that international election observation missions can constrain electoral fraud. In a district-level analysis of one recent Armenian election that provided a rare version of a natural experiment, Hyde (2007) found that that the presence of international observers did lead to a significant, albeit marginal, reduction in the amount of electoral fraud that occurred. As noted in Chapter 3, this effect probably helps explain the improved survival rates of democracies in the post–Cold War period, a time when international observation of elections was transformed from an occasional practice into a global norm. Meanwhile, Pastor (1999: 18) suggests an indirect and conditional effect in which the presence of election monitors enhances the efficacy of electoral management bodies. Specifically, he cites two comparative studies showing that “international monitors and mediators have played pivotal roles supporting, advising, and enhancing the credibility of ECs.”18 Of course, international observer missions are political as well as technical endeavors, and their political nature often shapes their

142

Dilemmas of Democratic Consolidation

reporting in ways that can undercut their effectiveness as checks on electoral malfeasance.19 The summary judgments issued by these missions sometimes whitewash or downplay the problems they see, and as a result the missions can appear to endorse elections in cases where significant fraud or abuse plainly occurred. Those endorsements sometimes allow “cheaters” to avoid costly consequences for their actions.20 Anticipating this soft response, would-be cheaters are less likely to be deterred from courses of action that could derail democracy. And each time this happens, the currency of international election observation is somewhat devalued. Bearing this in mind, policymakers interested in putting international elections observation firmly in the service of democratic development might consider establishing and funding an electionsauditing body that is both technically effective and insulated from the political and diplomatic concerns that weaken the mandate of many of the existing ad hoc missions. This process only deters cheating when political parties genuinely fear that they will be caught and called to task, and that fear will be stronger when international observers are not obliged to weigh their sponsoring organizations’ diplomatic concerns when designing their missions and issuing their reports. If international election observation were performed more rigorously by an independent entity, actors bent on cheating would presumably opt out of the system more often than they do now. Skeptics of this proposal might view that as a bad thing, but I would argue that the resulting outcomes in many cases would be no worse than they are under the current arrangement, and in some cases the outcome would probably be better. Politicians determined to cheat in front of observers in the current system would probably continue to cheat in the new one, only without the observers there. Meanwhile, politicians who were on the fence about cheating would have an added incentive to run a clean election, or at least to cheat in subtler and therefore probably less effective ways. This incentive would arise because the decision to opt out of a more robust system would send a clearer signal to interested parties about the incumbent’s intentions—a signal that might therefor prove to be costly in and of itself (Hyde and Marinov 2008). Conditioning Foreign Aid

A growing number of scholars and advocates is urging Western governments to strictly condition their foreign aid and development assistance on the establishment or maintenance of “good governance” (Diamond 2008), or even more specifically on the protection of civil

Implications for Democracy Promotion

143

liberties (Bueno de Mesquita and Downs 2005). The basic rationale for these proposals is that conditional aid will give rulers a stronger incentive to establish “good governance” practices—such as rule of law, control of corruption, and the promotion of social welfare—that are thought to help democratic government emerge and survive. In fact, the U.S. Government already conditions its foreign assistance on democratic governance in a variety of ways. Launched in 2002 by the Bush Administration, the Millennium Challenge Account (MCA) is one of the newest and probably the most visible policy manifestation of this idea, but it is not the only one.21 Under Section 508 of its annual appropriations act, the U.S. Congress also mandates the suspension of assistance to countries whose elected head of government has been deposed by a coup d'etat or decree. The law makes exceptions for certain democracy-promotion and humanitarian activities and allows for exemptions in the national interest, but the implementation of these sanctions is more or less automatic.22 Recent coups that triggered Section 508 sanctions include ones in Pakistan in 1999, Cote d’Ivoire in 2000, and Fiji and Thailand in 2006. As far as I can determine, no other government automatically imposes sanctions in response to a usurpation of power, but coups can trigger responses from some intergovernmental organizations. The Organization of American States (OAS) probably goes furthest in this regard with Resolution 1080, which was adopted in 1980 and requires the organization’s Permanent Council to meet within ten days of a coup in a member state and develop a response “consistent with” the OAS’s stated commitment to democracy. The Commonwealth of Nations, composed mostly of former British colonies, also usually responds to coups with suspensions of membership, a penalty it imposed most recently on Pakistan in 1999 and Fiji in 2006. The game-theoretic model developed in this book supports the premise that outside actors can improve the odds that democratic regimes will survive by threatening to impose sanctions in response to coups and offering concrete rewards for sustaining democracy. If properly designed, these kinds of commitments can affect actors’ beliefs about the payoffs associated with different outcomes and the costs of trying and failing. At the same time, the model reminds us that these threats and promises must be both credible and costly in order to be effective. Credibility means the actors whose behavior is targeted must believe it is in the interest of the originator of those commitments to carry them out, either because doing so will leave that originator better off or because reneging on the commitment will impose its own direct costs on him or her. Reputation matters in this regard, and this is a currency that gets diluted when threats and promises are made but not

144

Dilemmas of Democratic Consolidation

executed or are frequently overruled in pursuit of other national interests. Finding an effective level and mix of sanctions and rewards is another significant challenge. The fact that usurpations of power have continued to occur since the addition of Section 508 to the foreignappropriations act and the establishment of the Millennium Challenge Account shows that the current combination of carrots and sticks is not costly enough to deter all coups. Actors’ expectations about their relative welfare under democracy and dictatorship are shaped powerfully by the nature of the domestic economy, so opportunities for rewards or sanctions will often be constrained by structural factors. In addition to being credible, promises of financial rewards or sanctions must be carefully targeted to have the desired effect. The national welfare is not the same thing as the welfare of the relevant players, so outside actors considering this strategy must think carefully about how the impact of their efforts will be distributed. Economic sanctions or suspensions of foreign aid are blunt tools that may not alter the payoffs for the specific actors enough to make a meaningful difference. Preventing Polarization

This book’s model can also be used to help anticipate and think about how to prevent the political polarization that typically precedes violent conflict and authoritarian coups in new democracies. To identify strategies for ameliorating this problem, I turn to existing work on preventing interstate conflicts and settling civil wars in which a security dilemma appears to be operating. In order to extend those models to cases of democratization, I assume that rival parties in new democracies are like rival states or warring parties in intrastate conflicts, and that democratic regimes are functionally similar to cooperative regimes in the international system or negotiated settlements in civil wars. Just as international regimes and settlements in civil wars are susceptible to reneging, new democracies fall prey to coups or rebellion when parties choose to defect from their deal. If polarization arises from commitment problems stemming from mutual insecurity, then Fearon’s widely cited 1998 analysis of structural conditions influencing the risk of ethnic conflict should apply to prospects for polarization among political parties more generally. According to Fearon, the most powerful means for overcoming this mutual insecurity are credible guarantees by third parties for the protection of vulnerable groups from predation by potential rivals. In line with Fearon, Walter (1999: 129–130) argues that finding solutions

Implications for Democracy Promotion

145

to the substantive disputes motivating conflict is not the hard part; instead: The greatest challenge is to design a treaty that convinces the combatants to shed their partisan armies and surrender conquered territory even though such steps will increase their vulnerability and limit their ability to enforce the treaty’s terms. Groups that obtain third-party security guarantees for the treacherous demobilization period following the signing of an agreement, and internal political, military, or territorial guarantees will implement their settlements. If an outside state or international organization is not willing or able to provide such guarantees, the warring factions will reject a negotiated settlement and continue their war.

She adds (p. 135) that vital concerns for third parties providing security guarantees include reducing the risk of a surprise attack after demobilization and reducing the prospect of permanent exclusion from power. This line of thinking about how to avoid or end civil wars suggests that outside actors can improve the chances that democracy will survive by providing credible guarantees to protect the personal security and material interests of electoral losers. In Jervis’ (1978) terms, the goal of these guarantees is to lower the expected costs of being exploited. “States that can afford to be cheated in a bargain or that cannot be destroyed by a surprise attack can more easily trust others and need not act at the first, and ambiguous, sign of menace,” he writes (p. 172). When those costs are sufficiently low, parties can spend less time worrying about their potential victimization and concentrate instead on pursuing gains through cooperation. This logic may help to explain how democracy survived through a second election in Sierra Leone, where outside powers, led by the United Kingdom, intervened on a large scale after a decade of civil war marked by almost unimaginable brutality. Sierra Leone’s initial attempt at democracy came at independence from Great Britain in 1961, but an intensifying rivalry between parties with distinct ethnic and regional bases helped lead to a military coup after elections in 1967. The elected government was restored in 1968, but another coup attempt in 1971 prodded the government of Siaka Stevens, with support from neighboring Guinea’s army, to claim dictatorial powers. An aborted transition in 1991 sparked a civil war that destroyed the next attempt in 1996, when mutinous soldiers allied with rebel Revolutionary United Front (RUF) forces to overthrow elected president Ahmed Tejan Kabbah.

146

Dilemmas of Democratic Consolidation

In the late 1990s, international military intervention turned the tide against the RUF, which then negotiated a ceasefire with government forces that led to a disarmament process starting in 2001. In elections held on 10 May 2002 under the watch of 17,000 U.N. peacekeepers, eleven political parties, including the Revolutionary United Front (RUF), competed for the presidency and seats in the National Assembly. President Ahmed Tejan Kabbah won the post of chief executive with 70 percent of the vote, while his party, the Sierra Leone Peoples’ Party (SLPP), won a majority of seats in the legislature. Foday Sankoh, leader of the RUF, had been declared ineligible and did not contest the presidency. Democracy in Sierra Leone then survived the next several years, and second elections were successfully held in 2007. That balloting produced a turnabout in the electoral fortunes of the country’s leading political parties as Ernest Bai Koroma of the All Peoples’ Congress (APC) defeated SLPP candidate Solomon Berewa for the presidency in a September runoff. What is remarkable about Sierra Leone’s experience since 2002 is not only that democracy has stuck and power has peacefully passed from one party to another, but also that political polarization seems to have eased in the process. The continuing presence of a large and effective foreign security force acts as a powerful disincentive to renewed fighting, and the United Nations’ prominent role in the administration of Sierra Leone’s 2002 and 2007 elections goes a long way toward ensuring the fairness of those contests. Nevertheless, in spite of a variety of structural conditions generally considered conducive to polarization and democratic breakdown—endemic poverty, “lootable” natural resources, long-running regional animosities rooted in perceptions of injustice and inequity, and a large pool of young men skilled in violence—political tensions had cooled enough by late 2007 that Polity, a leading source of data on political regimes worldwide, no longer considered politics in Sierra Leone to be acutely polarized.23 Sierra Leone offers a good-news/bad-news story for democracy promoters. On the upside, the case of Sierra Leone shows how credible third-party guarantees for the security of political rivals can short-circuit the downward spiral toward factionalism and thereby help to prevent democracy’s demise, even in what conventional theories of democratization would regard as infertile conditions. This country’s recent experience suggests that the kinds of third-party guarantees described by Fearon and Walter can play a role not only in forestalling armed conflict but also in defusing other threats to democracy historically associated with factionalism, such as executive coups and popular uprisings. On the down side, Sierra Leone also suggests that

Implications for Democracy Promotion

147

such guarantees are expensive to provide. The United Nations and the United Kingdom have invested many tens of millions of dollars and committed precious forces to this small country for more than six years now, and most accounts seem to suggest that ending that support abruptly might scuttle the whole endeavor. Taken to its logical end, the idea that third-party guarantees can play an important role in supporting democracy in developing countries suggests that democracy could be advanced on a global scale through the establishment of an international security force with the capacity and neutrality to protect political losers and enforce constitutional bargains in democracies. Effective enforcement requires reliable observation of violations, so the credibility of this security force’s guarantees might depend, in turn, on the establishment of a more effective international regime for observing democratic elections. These ideas are politically unrealistic, of course. Nevertheless, showing how they emerge from the strategic logic that drives many failures of democracy can help clarify how difficult it is to promote democracy on a global scale. Summary

Above all, this chapter has aimed to remind individuals and organizations engaged in democracy promotion to assess carefully the ways their actions can affect the strategic concerns of key political actors in the countries they aim to help. In this chapter, I have argued that democracy assistance as currently practiced puts too much emphasis on technical issues and often fails to consider possible unintended consequences of well-intentioned programs. I have also posited that this technical approach to democracy promotion is, at least in part, a consequence of the bureaucratic organization of aid-giving by foreign governments. At the same time, this chapter has also suggested that practitioners in this field might achieve better results by refocusing their energies from the “softer” side of democracy promotion—things such as promotion of civil society and civic education—to areas where they can most directly affect the probability that usurpations of power will succeed and the expected payoffs if they do. Drawing on my game theoretic model, I have identified support for judicial reform, the creation of effective and impartial electoral management bodies, the development of a more rigorous and independent international electionobservation capability, and credible threats of costlier sanctions as four promising approaches along these lines. These strategies probably would not prove effective under all conditions, nor are they the only ways to accomplish the goal of democracy promotion. Nevertheless, the theory

148

Dilemmas of Democratic Consolidation

and evidence discussed here give policymakers and advocates good reason to consider focusing additional resources in these directions.

Notes 1 I am using the term “belief” in the game-theoretic sense, referring to an actor’s expectations about the intentions and capabilities of the other actors with which it is interacting strategically. These beliefs are generally represented in game theoretic models as probabilities that other actors will choose a particular course of action or will succeed in that course of action should they try it. 2 As used here, this term is not meant to encompass other ways in which foreign governments try to effect democratization in “target” countries, including high-level diplomacy, trade sanctions, and military intervention. 3 For a cogent overview of the organization of democracy-promotion work in the United States, see Melia (2006). 4 See http://www.usaid.gov/our_work/democracy_and_governance/ technical_areas/. To be fair to the D&G office, its core statement on developing assistance strategies for specific countries, “Conducting a DG Assessment: A Framework for Strategy Development” (November 2000), explicitly discusses the need to consider how interests and institutions combine to create “incentives for democratic or anti-democratic behavior.” My criticism is that this element is not emphasized enough in practice. 5 As of mid-2008, a list of the agency’s “implementing partners” could be found at www.usaid.gov/our_work/democracy_and_governance/ technical_areas/dg_office/partners.pdf. 6 I do not mean to suggest that the study’s authors obscure this point. Finkel et al. (2008: 6) use the term “modest” to describe the substantive impact, and I think that is a fair characterization. 7 The study also considers the cumulative impact of DG aid and concludes that on average a sustained investment of $10 million per annum for five years is associated with an increase of about one-half a point on the 13-point Freedom House scale compared with the expected growth absent that investment. In substantive terms, this effect is comparable to the aforementioned annual one. 8 For an intriguing general proposal, see Melia (2006). Among other things, he recommends that the U.S. Government get out of the business of direct programmatic assistance and instead provide longer-term block grants to a dozen or so of the nongovernmental organizations that have proven to be most effective in this area. 9 Carothers and Ottaway (2000) thoughtfully describe and analyze the rise of civil-society assistance, while Kumar (2004) offers a careful review of assistance to political parties. 10 See Grugel (2002) and Chapter 6 in Diamond (1999) for strong claims about the importance of civil society to the functioning of democracy. 11 That donor agencies are sometimes selective about the kinds of CSOs they will support suggests they are sensitive to this point, their rhetoric notwithstanding. In the Middle East, for example, virtually none of the West’s assistance to CSOs goes to what are arguably that region’s most powerful and popular social organizations, namely, Islamist movements (Al-Sayyid 2000).

Implications for Democracy Promotion

12

149

On this topic, see Henderson (2003). As cited in Beissinger (2007: 261). 14 “Putin Defends Bill Cracking Down on NGOs,” The New York Times (November 25, 2005). 15 “Foreign NGOs Complain Russia Hinders Registration,” The New York Times (August 21, 2006). 16 It should be noted, though, that this study included countries with all kinds of political regimes, so it is impossible to isolate in their results the effects of Rule of Law assistance on the persistence of democracy in countries where it was already established. 17 Work that has begun to fill this gap includes Elklit and Reynolds (2002) and a special issue of the journal International Political Science Review (Mozaffar and Schedler 2002) in addition to the research cited here. 18 Those studies are Vikram K. Chand, “Democratization from the Outside In: NGO and International Efforts to Promote Open Elections,” Third World Quarterly Vol. 18, No. 3 (1997): 543–561; and Jennifer McCoy, “Monitoring and Mediating Elections During Latin American Democratization,” in Kevin Middlebrook (ed.), Electoral Observation and Democratic Transitions in Latin America (La Jolla. CA: Center for US-Mexican Studies, 1998). 19 This politicization is openly acknowledged by some of the organizations that provide these missions. For example, the Commission of the European Communities stated in a 2000 white paper on EU electoral assistance and observation that criteria used to assess the validity of any election “should be applied with flexibility,” and that post-conflict elections and first-time elections require a “softer” approach. This paper is available online at http://europa.eu/eur-lex/en/com/cnc/2000/com2000_0191en01.pdf. 20 In reaching this judgment, I disagree with Hyde and Marinov (2007: 1), who state that “a typical election monitoring mission…is usually willing to condemn an election which is found to be fraudulent.” While I cannot provide a systematic comparison of electoral conduct and observer statements, I can say that, in researching the quality of many specific elections, I have often found the summary judgments in observers’ reports—the part that usually gets trumpeted in the international press—to be disconnected from the facts described in the full accounting. 21 The MCA offers grants to countries that meet certain criteria on sixteen performance indicators grouped under the headings “Governing Justly,” “Investing in People,” and “Promoting Economic Freedom.” Countries that score above the median on half or more of the criteria in each of the three areas are identified as “better performers” who are then eligible for selection by the MCA’s Board of Directors for funding. For more information on MCA, see www.mca.gov. 22 Here is the actual language from Section 508 of the 2005 appropriations act: “None of the funds appropriated or otherwise made available pursuant to this act shall be obligated or expended to finance directly any assistance to the government of any country whose duly elected head of government is deposed by military coup or decree: Provided, that assistance may be resumed to such Government if the President determines and certifies to the Committees on Appropriations that subsequent to the termination of assistance a democratically elected government has taken office; Provided further, that the provisions to this 13

150

Dilemmas of Democratic Consolidation

section shall not apply to assistance to promote democratic elections or public participation in democratic processes; Provided further, that funds made available pursuant to the previous provisos shall be subject to the regular notification procedures of the Committee on Appropriations.” 23 See www.systemicpeace.org/polity/polity4.htm. Polity identifies acutely polarized countries as “factional,” indicated by a score of 3 on its PARCOMP indicator, and it records factionalism in Sierra Leone as having ended in 2007.

7 Conclusion

In this book, I have developed and applied a game-theoretic model that is meant to help us understand how and why attempts at democracy so often fail. Consistent with conventional thinking on the subject, my model suggests that democracy is fragile, in part, because relevant organizations often covet state power as a means to improve or defend their personal welfare. The spoils of power often give political parties and the military strong incentives to attempt coups or mount rebellions. As Hardin (1997: 249) so cogently observes, “Democracy is essentially a member of the mutual-benefits class of theories. If political divisions cut very much deeper than the marginal issues on which we can democratically compromise, democracy may no longer seem to produce mutual benefits. It then produces major—not marginal—winners and losers. Big disagreements bring us down.” More novelly, though, my model also shows how political organizations might sometimes try to usurp power because they are unsure of their rivals’ intentions and capabilities. More specifically, the model predicts that political parties and the military will sometimes attempt coups and rebellions because they fear that their political rivals may be plotting to seize power, and they badly want to avoid that outcome. Combined with the more conventional temptations of state power, this strategic uncertainty makes democratic regimes even more fragile than most theories of consolidation recognize. So, to Przeworski’s (1991: 19) statement that “political forces comply with present defeats because they believe that the institutional framework that organizes the democratic competition will permit them to advance their interests in the future,” my model implies that we must add some provisos. For those forces to comply, they must also believe that democratic competition itself will survive or that they are not strong enough to prevent its usurpation or to steal power on their own. In contrast to Przeworski’s emphasis on future election outcomes as the key source of political uncertainty in democracies, my model suggests

151

152

Dilemmas of Democratic Consolidation

that actors fears about their rivals’ intentions will often overshadow their beliefs about their prospects in future elections in shaping their own decisions about whether to attempt a coup or rebellion. The representation of democracy encapsulated in this book’s gametheoretic model suggests another way to think about the larger process of democratic consolidation. Following Schedler (1998), we can describe a consolidated democracy as one that most observers, including the key actors themselves, expect to survive indefinitely. In the framework described here, that means that none of the key actors is expected to attempt to usurp power, no matter what the outcome of the most recent or impending elections, barring some significant exogenous shock. Most obvious, this expectation can arise when the expected payoffs from upholding democracy are substantially better than the ones following an attempt at usurpation. That outcome does not require any communication or coordination among the players, and it seems intuitively like an unrealistic prospect in many young democracies. There is another possibility, however. Because the game I describe is repeated indefinitely, and because the players can communicate with one another, they may learn from each others’ behavior and develop coordinated strategies. This kind of coordination can produce different outcomes from the one we would expect if we only assessed players’ expected payoffs on the basis of the game’s initial conditions. Through learning and communication, players can find their way to an equilibrium path that might appear to be at odds with strictly welfaremaximizing behavior. This possibility suggests that we might also think of consolidation as that process of coordination, a process that can allow democracy to persist under what might otherwise appear to be inauspicious conditions. This process is constrained by the broader conditions affecting actors’ expectations about the probabilities and payoffs associated with various courses of action, but it is not wholly determined by them. Ironically, then, a model that underscores the inherent fragility of democracy also shows how this form of government might sometimes survive even in what seems to be inhospitable soil. In that respect, the theory of democratic consolidation that arises from my game-theoretic model diverges in a subtle but important way from the traditional emphasis in democratization research on the quality of governance as the key to the survival of new democracies. Diamond offers a cogent restatement of this argument, which traces its lineage at least to Linz (1978), when he stated in a 2008 Foreign Affairs article that, “For democratic structures to endure—and to be worthy of endurance—they must listen to their citizens’ voices, engage their participation, tolerate their protests, respect their freedoms, and respond

Conclusion

153

to their needs.” Without disputing the normative thrust of Diamond’s assertion, I hope that the model developed in this book has shown how key actors in democratic systems confront a set of strategic dilemmas that are connected to questions of economic performance, social justice, and personal freedom but are not determined by them. Moreover, I would posit that these dilemmas play a more direct role in determining the survival or demise of democratic regimes than the broader but more amorphous question of how well their governments are meeting the needs of their citizens, or how those citizens feel about democracy as a form of government. Put another way, I am asserting that the strategic concerns of specific organizations, not popular frustrations, are the proximate cause of most attempts to usurp power in democracies. Just as dictatorship often persists even when it is profoundly unpopular, democracy can too, because regime change requires collective action that is usually difficult and costly. If this is true, then efforts to shore up vulnerable democracies would do well to consider ways to ameliorate those strategic tensions instead of or in addition to trying to deepen democracy in a more general fashion. Areas for Further Research

This book has attempted to develop a new theory of democratic breakdown and consolidation, and it has used analytic narratives to probe that theory’s plausibility and test its explanatory power relative to some prominent prior theories. In light of the newness of the model and the limited scope of the empirical analysis undertaken here, it is not hard to imagine a few ways in which this theory could benefit from further extensions of the theoretical model and additional empirical testing. One area of enduring interest to scholars of comparative democratization to which this book has devoted little attention is the effect of institutional design on the survival of democratic regimes. Democracies come in many forms, and those variations in form should shape the incentives to which political parties and the military respond in choosing whether or not to try to usurp power. For example, democracies vary widely in how they distribute power among branches of government, and different distributions—presidentialism, parliamentarism, and the many mixed versions thereof—could make executive coups easier or more difficult to carry out, thereby affecting the likelihood that one will be attempted. Democracies also vary widely in the rules they use to determine election winners, and these variations should have some effect on actors’ assessments of their prospects in and payoffs from future elections, and thus their temptation to usurp power.

154

Dilemmas of Democratic Consolidation

The combinations of these elements should matter as well. For example, a democracy that concentrates power in a president’s hands and elects that president under first-past-the-post rules (in which the candidate who receives the most votes wins, whether or not he or she received a majority of votes) seems more likely to lead outnumbered opposition parties to conclude that they will benefit little from the survival of democracy than a parliamentary system that uses proportional representation to select its legislators, thereby ensuring that opposition party at least a foothold in the democratic government. I chose not to dig into this topic in this book, in part, because there are so many possible variations that it is extremely difficult to make the sharp comparisons required to test the resulting claims. Nevertheless, this is a topic that other scholars have fruitfully examined, and it would be very interesting to explore its implications for the model sketched in this book.1 My game-theoretic model could also be extended through a more detailed consideration of collective action. The processes by which organizations form, mobilize resources, and act publicly in pursuit of their interests are crucial to understanding several elements of the theory, including where the political parties that comprise key players in the formal model come from; how routine interactions within those organizations and the military shape the preferences they pursue; and how these organizations can try to usurp power, and thus how costly those attempts are and how likely they are to succeed. In this initial formulation of the theory, I have represented the costs of collective action as an exogenous constraint on the players’ choice of strategies. An alternative formulation might represent that parameter as a choice each player can make regarding how much to invest in collective action aimed at usurping power. This approach would allow us to consider how players might hedge risks with something subtler than an all-or-nothing approach to defection, and it could help to explain collective behaviors that might otherwise seem irrational, such as sustained but small protest movements aimed at toppling elected governments. More careful treatment of collective action points toward another potentially rewarding area for further research: the effects of what Tsebelis (1990) calls “nested games” on the strategic choices this theory considers. As Tsebelis shows, embedded, connected, or encompassing interactions can alter players’ incentives, and sometimes in counterintuitive ways, by shaping their preferences, imposing additional constraints, or offering additional rewards. The case of Cyprus, considered in Chapter 4, offers an excellent example. The study sketched there notes that politics in Cyprus after independence were not just a two-way rivalry between the Greek Cypriot and Turkish Cypriot

Conclusion

155

communities living on the island; instead, it argues that they were heavily conditioned by interactions with three other national governments, each of which had its own interests in that country’s trajectory. More explicit treatment of these interactions through a nested-game design might help to explain why the two communities on Cyprus wound up fighting a series of predictably costly battles over the years, only to wind up stuck with a partition that could have been produced at much lower cost in the initial negotiations on the island’s status in 1959. On the empirical side, our ability to rigorously test this theory would be greatly improved by the collection of cross-national timeseries data on attempted as well as successful military coups, executive coups, and rebellions. As noted in Chapter 3, while several existing data sets capture failed as well as successful military coups and some record violent conflicts and non-violent protest events that may be indicative of the type of rebellion anticipated in my model, currently available data sets do not provide the kind of information required for dynamic, crossnational statistical analysis of that model. This is especially true of executive coups, which by their nature are probably more difficult to observe when they fail than military coups and rebellions because that failure often happens behind the scenes. Although it would require a substantial investment of time and resources to get there, creation of a global, time-series, cross-sectional data set that included unsuccessful as well as successful attempts to usurp power would allow us to draw more confident inferences about the effects of other covariates on the behaviors with which my model is concerned. It would also allow us to study the conditions under which such attempts are likely to succeed or fail and how those processes interact—research that could be especially helpful to policymakers who sometimes have to consider how to react to these events as they unfold in real time. Future Trends

The poor survival rates for democracies in the past half-century that I documented in Chapter 3 seem especially stark when we recall that many of these episodes include a decade that represents something of a “golden age” for democratization, namely, the time between the end of the Cold War and the terrorist attacks on the United States of September 11, 2001. The revolutions in Eastern Europe in 1989 and ensuing collapse of the Soviet Union spawned a period of profound optimism in the West about the spread of elected government to new corners of the world. That optimism saw its ideological reflection in the emergence of

156

Dilemmas of Democratic Consolidation

what seemed to be a global norm favoring popular rule (Franck 1992, McFaul 2004, Sen 1999). That optimism also gained concrete expression in a proliferation and strengthening of pro-democracy intergovernmental and non-governmental organizations, changes that were both catalyzed and complemented by a sharp increase in official democracy assistance from Western governments.2 In the 2000s, however, conditions for democracy appear to have become less auspicious. Military coups in Bangladesh, Fiji, and Thailand; subversions of democracy by elected leaders in places such as Kenya, Russia, Senegal, and Venezuela; the persistence of authoritarian rule in China, Cuba, Iran, Myanmar, and Zimbabwe, to name just a few cases; and the continuing failure of elected governments backed by Western armies and billions in aid to bring stability to Afghanistan and Iraq have put democracy as a concept, and the local and global activists who agitate for it, on the defensive. Several global and regional dynamics appear to be driving this trend. For one, China has increasingly parlayed its phenomenal economic growth into international aid and investment that offer authoritarian rulers in many developing countries a no-strings-attached alternative to the conditional support they have long received from Western governments and international financial institutions. Meanwhile, rising global demand has pushed prices for oil and natural gas to record levels, providing numerous authoritarian governments with windfall revenues they have used to help sustain the loyalty of security forces and key elites. Russia has used its newfound hydrocarbon wealth to reassert its leadership in the “near abroad” and, in so doing, has undercut Western democracy-promotion efforts in many of the Soviet successor states. Venezuela’s President Hugo Chávez has attempted to play a similar role in Latin America, if less effectively. Taken together, these and other developments have rendered the international environment substantially less favorable to the spread of democracy than it was in the several years immediately following the Soviet Union’s collapse.3 These trends have led some observers to speak of a “democratic recession” (Diamond 2008) and even “the end of the end of history” (Kagan 2008). While I agree that the short- and even medium-term outlook has dimmed somewhat, I remain optimistic about the long-term prospects for the global spread and deepening of democracy. First of all, the current pushback from countries such as Russia and Venezuela rests on shaky foundations. Whatever the predictions about “peak oil,” the inexorable logic of markets—what goes up must come down—means that the price of oil and natural gas will eventually fall. The timing of

Conclusion

157

that eventual decline is the subject of intense debate among experts who study the behavior of commodities markets, the geology of hydrocarbon resources, and the prospects for new technologies that would significantly and rapidly alter the demand side of this equation. Still, the basic logic seems undeniable, as evidenced by other recent boom-andbust cycles in equities markets and real estate. And when oil and gas prices fall, that decline will deliver a blow to the anti-democratic efforts of some authoritarian regimes by reducing the revenue streams on which those efforts depend. The consequences of this decline are impossible to predict with precision, but one result almost certainly will be another moment of opportunity to establish democratic governments in oil-rich autocracies and other countries that depend on their largesse. A second source of long-term optimism stems from the continuing trend toward deeper integration into the liberal global economy. Since the end of World War II, and especially in the past two decades, trade flows have swelled, international organizations have proliferated, and more countries have attempted democracy. The nexus of these trends almost certainly lies in the functional links between democracy and economic development—links that promote exactly the kind of positive feedback loops Pierson (2005) identifies as a key mechanism for pathdependent change in political institutions. When governments discover they cannot survive by force alone, they must find ways to secure the habitual, quasi-voluntary compliance of the populations they seek to rule.4 To secure that compliance, they need to promote prosperity and remove incentives to rely on force as a means to effect political change. As I read the evidence, the second half of the twentieth century demonstrated convincingly that the combination of democratic governance with market-based economies offers the most effective means to achieve those ends in a durable way (Ulfelder 2008). The combination of democracy and the market does not always produce immediate gains, but at present there appears to be no sustainable alternative, so polities that try democracy and fail almost invariably try again. As Bates (2001: 107) argues, “The creation of limited government may not be sufficient to secure high levels of investment, much less the growth of national economies. But assurances to investors surely are necessary to secure the formation of capital,” and thus to allow economic growth to occur. As technological and political developments have expanded the possibilities for global exchange, governments have increasingly reached out to one another in an effort to create new opportunities for growth and then to help their citizens realize the resulting gains. Thus, even as many countries’ domestic political regimes remain volatile, the historical trajectories point

158

Dilemmas of Democratic Consolidation

decidedly toward a world increasingly composed of states with elected governments linked by dense networks of economic exchange and political and legal entanglements. A crucial point is that this trend does not derive fundamentally from changes in norms and values. Instead, it depends most of all on the enduring and interlocking human desires for security, freedom, and well-being—the “life, liberty, and pursuit of happiness” identified by the drafters of the U.S. Declaration of Independence as the most essential of all human rights. Consequently, the apparent reemergence of an authoritarian model of development poses less of a challenge to the long-term trajectory of political change that some observers are claiming (Kagan 2008). As I have already noted, the rise of states willing to provide foreign investment and assistance without conditions certainly can affect the short-term incentives for political parties and militaries to uphold democratic procedures and the civil liberties required to energize them. My contention, however, is that those patrons’ authoritarian power rests on shaky foundations because it depends on the exploitation of resources that are exhaustible and not inherently valuable. When oil prices finally fall or workers and entrepreneurs in places such as China finally organize to demand new opportunities and to seek protection from the corruption that saps the returns on their investments, this authoritarian revival will founder, and another wave of new attempts at democracy will probably ensue.

Notes 1 On presidential vs. parliamentary arrangements, see Cheibub (2007). On electoral systems, see Norris (2004). On other forms of power-sharing, see Reynolds (2002). 2 Francis Fukuyama’s 1992 book End of History and the Last Man offered what is probably the best-known assertion that the end of the Cold War would usher in an “age of democracy,” but other prominent work in this vein includes Inglehart and Weltzel (2005), Plattner (2002), McFaul (2004), and Sen (1999). On the expansion of US Government assistance for democracy and governance, see Finkel et al. (2007). 3 For a concurring view, see Diamond (2008). 4 On this point, see Franck (1992) and Levi (1988).

Appendix Episodes of Democracy, 1955–2007

Country Albania Albania Argentina Argentina Argentina Argentina Armenia Australia Austria Azerbaijan Bangladesh Bangladesh Belgium Benin Benin Bolivia Bolivia Bosnia Botswana Brazil Brazil Bulgaria Burkina Faso Burma Burma Burundi Burundi Cambodia

Start 1991 1997 1958 1963 1973 1983 1991 1901 1946 1992 1972 1991 1944 1960 1991 1956 1982 1992 1966 1946 1985 1990 1978 1948 1960 1993 2005 1993

End 1996 2007 1962 1966 1976 2007 1996 2007 2007 1993 1974 2007 2007 1963 2007 1964 2007 1995 2007 1964 2007 2007 1980 1958 1962 1996 2007 1997

159

Dur. 6 11 5 4 4 25 6 107 62 2 3 17 64 4 17 9 26 4 42 19 23 18 3 11 3 4 3 5

Type of Failure Exec. Coup . Mil. Coup Mil. Coup Mil. Coup . Exec. Coup . . Mil. Coup Exec. Coup Mil. Coup . Mil. Coup . Mil. Coup . Other . Mil. Coup . . Mil. Coup Mil. Coup Mil. Coup Mil. Coup . Exec. Coup

Total Episodes 1 2 4 5 6 7 1 1 2 1 1 2 3 1 2 1 2 1 1 1 2 2 1 1 2 1 2 1

160

Dilemmas of Democratic Consolidation

Country Canada Cen. Afr. Rep. Chile Chile Colombia Comoros Comoros Congo-Brazzaville Congo-Brazzaville Costa Rica Croatia Croatia Cyprus Cyprus Czech Republic Czechoslovakia Denmark Dominican Rep. Dominican Rep. Dominican Rep. East Timor Ecuador Ecuador Ecuador Ecuador El Salvador Equatorial Guinea Estonia Fiji Fiji Fiji Finland France Gambia Georgia Georgia Germany

Start 1868 1993 1952 1990 1958 1996 2002 1960 1992 1949 1991 2000 1968 1983 1993 1990 1945 1963 1978 1996 2002 1948 1966 1979 2003 1982 1968 1991 1970 1992 2001 1918 1946 1965 1992 1992 1949

End 2007 2003 1973 2007 2007 1999 2007 1963 1997 2007 1995 2007 1973 2007 2007 1992 2007 1963 1990 2007 2007 1961 1970 2000 2007 2007 1969 2007 1987 2000 2006 2007 2007 1970 1992 2000 2007

Dur. 140 11 22 18 50 4 6 4 6 59 5 8 6 25 15 3 63 1 13 12 6 14 5 22 5 26 2 17 18 9 6 90 62 6 1 9 59

Type of Failure . Mil. Coup Mil. Coup . . Mil. Coup . Mil. Coup Rebellion . Exec. Coup . Exec. Coup . . Other . Mil. Coup Exec. Coup . . Mil. Coup Mil. Coup Rebellion Exec. Coup . Exec. Coup . Mil. Coup Mil. Coup Mil. Coup . . Exec. Coup Rebellion Exec. Coup .

Total Episodes 1 1 1 2 2 1 2 1 2 3 1 2 1 2 1 3 2 1 2 3 1 1 2 3 4 1 1 2 1 2 3 1 3 1 2 2 2

Appendix

Country Ghana Ghana Ghana Ghana Greece Greece Guatemala Guinea-Bissau Guinea-Bissau Guinea-Bissau Guyana Guyana Haiti Haiti Haiti Honduras Honduras Honduras Hungary India India Indonesia Indonesia Iraq Ireland Israel Italy Jamaica Japan Kenya Kenya Laos Latvia Lebanon Lebanon Lesotho Lesotho

Start 1957 1969 1979 1992 1944 1974 1986 1994 2000 2005 1966 1992 1991 1994 2006 1957 1971 1982 1990 1950 1977 1955 1999 2006 1922 1948 1948 1962 1952 1963 2002 1954 1991 1943 2005 1966 1994

End 1958 1972 1981 2007 1967 2007 2007 1999 2003 2007 1968 2007 1991 1999 2007 1963 1972 2007 2007 1975 2007 1957 2007 2007 2007 2007 2007 2007 2007 1966 2007 1960 2007 1976 2007 1970 1994

Dur. 2 4 3 16 24 34 22 6 4 3 3 16 1 6 2 7 2 26 18 26 31 3 9 2 86 60 60 46 56 4 6 7 17 34 3 5 1

Type of Failure Exec. Coup Mil. Coup Mil. Coup . Mil. Coup . . Mil. Coup Mil. Coup . Exec. Coup . Mil. Coup Exec. Coup . Mil. Coup Mil. Coup . . Exec. Coup . Exec. Coup . . . . . . . Exec. Coup Exec. Coup Mil. Coup . Rebellion . Other Other

161

Total Episodes 1 2 3 4 4 5 2 1 2 3 1 2 2 3 3 1 2 3 2 1 2 1 2 1 1 1 2 1 1 1 2 1 2 1 2 1 3

162

Dilemmas of Democratic Consolidation

Country Lesotho Liberia Liberia Lithuania Macedonia Madagascar Madagascar Malawi Malaysia Mali Mauritania Mauritius Mexico Moldova Mongolia Montenegro Mozambique Namibia Nepal Nepal Netherlands New Zealand Nicaragua Niger Niger Nigeria Nigeria Norway Pakistan Pakistan Pakistan Panama Panama Papua New Guinea Paraguay Peru Peru

Start 1994 1997 2006 1991 1991 1960 1993 1994 1957 1992 2007 1968 1997 1997 1990 2006 1994 1990 1959 1991 1945 1908 1984 1993 1999 1960 1979 1945 1948 1972 1988 1952 1989 1975 1993 1956 1963

End 2007 1998 2007 2007 2007 1965 2007 2004 1969 1997 2007 2007 2007 2007 2007 2007 2004 2007 1960 2002 2007 2007 2007 1996 2007 1966 1983 2007 1958 1977 1999 1968 2007 2002 2007 1962 1968

Dur. 14 2 2 17 17 6 15 11 13 6 1 40 11 11 18 2 11 18 2 12 63 100 24 4 9 7 5 63 11 6 12 17 19 28 15 7 6

Type of Failure . Exec. Coup . . . Exec. Coup . Exec. Coup Exec. Coup Exec. Coup . . . . . . Exec. Coup . Other Other . . . Mil. Coup . Mil. Coup Mil. Coup . Exec. Coup Exec. Coup Mil. Coup Mil. Coup . Other . Mil. Coup Mil. Coup

Total Episodes 3 1 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 2 1 2 2 1 2 3 1 2 1 2 2 3

Appendix

Country Peru Peru Philippines Philippines Poland Portugal Romania Russia Russia Senegal Senegal Serbia Serbia & Monten. Sierra Leone Sierra Leone Sierra Leone Sierra Leone Slovakia Slovenia Solomon Islands Somalia South Africa South Korea South Korea Spain Sri Lanka Sri Lanka Sudan Sudan Sudan Swaziland Sweden Switzerland Syria Syria Taiwan Tanzania

Start 1980 1995 1944 1986 1989 1976 1996 1991 1993 1960 2000 2006 2000 1961 1968 1996 1998 1993 1991 1978 1960 1994 1960 1988 1977 1948 1994 1956 1965 1986 1968 1919 1849 1954 1961 1992 1995

End 1992 2007 1972 2004 2007 2007 2007 1993 2003 1963 2007 2007 2006 1967 1971 1997 2007 2007 2007 2007 1969 2007 1961 2007 2007 1982 2007 1958 1969 1989 1973 2007 2007 1958 1962 2007 2007

Dur. 13 13 29 19 19 32 12 3 11 4 8 2 7 7 4 2 10 15 17 30 10 14 2 20 31 35 14 3 5 4 6 89 159 5 2 16 13

Type of Failure Exec. Coup . Exec. Coup Mil. Coup . . . Exec. Coup Exec. Coup Exec. Coup Exec. Coup . Other Mil. Coup Exec. Coup Rebellion . . . . Mil. Coup . Mil. Coup . . Exec. Coup . Mil. Coup Mil. Coup Mil. Coup Other . . Other Mil. Coup . .

163

Total Episodes 4 5 2 3 2 2 1 1 2 1 2 1 1 2 3 4 1 1 1 1 1 1 2 2 1 2 1 2 3 1 1 1 3 4 1 1

164

Dilemmas of Democratic Consolidation

Country Thailand Thailand Thailand Trinidad Turkey Turkey Turkey Turkey Uganda Ukraine Ukraine United Kingdom Uruguay Uruguay Venezuela Zambia Zambia Zambia Zimbabwe

Start 1975 1983 1992 1962 1946 1961 1973 1983 1962 1991 2005 1901 1942 1985 1959 1964 1991 2006 1980

End 1976 1991 2006 2007 1960 1971 1980 2007 1966 1998 2007 2007 1972 2007 2005 1968 1996 2007 1987

Dur. 2 9 15 46 15 11 8 25 5 8 3 107 31 23 47 5 6 2 8

Type of Failure Mil. Coup Mil. Coup Mil. Coup . Mil. Coup Mil. Coup Mil. Coup . Exec. Coup Exec. Coup . . Exec. Coup . Exec. Coup Exec. Coup Exec. Coup . Exec. Coup

Total Episodes 1 2 3 1 1 2 3 4 1 1 2 1 2 3 2 1 2 3 1

References

Acemoglu, Daron and James A. Robinson. 2006. The Economic Origins of Dictatorship and Democracy. New York: Cambridge University Press. Al-Sayyid, Mustafa Kamel. 2000. “A Clash of Values: U.S. Civil Society Aid and Islam in Egypt,” in Thomas Carothers and Marina Ottaway, eds. Funding Virtue: Civil Society Aid and Democracy Promotion. Washington, DC: Carnegie Endowment for International Peace. Alexander, Gerard. 2002. The Sources of Democratic Consolidation. Ithaca, New York: Cornell University Press. Allison, Paul D. 1984. Event History Analysis: Regression for Longitudinal Event Data. Newbury Park: Sage Publications. Armony, Ariel C. and Hector E. Schamis. 2005. “Babel in Democratization Studies.” Journal of Democracy Vol. 16, No. 4 (October): 113–128. Bates, Robert H. 2001. Prosperity and Violence. New York: Norton. ———. 2008. When Things Fall Apart. New York: Cambridge. Bates, Robert H., Avner Grief, Margaret Levi, Jean-Laurent Rosenthal, and Barry R. Weingast. 1998. Analytic Narratives. Princeton, New Jersey: Princeton University Press. Beissinger, Mark. 2007. “Structure and Example in Modular Political Phenomena: The Diffusion of Bulldozer/Rose/Orange/Tulip Revolutions.” Perspectives on Politics (June): 259–276. Betz, David J. 2004. Civil-Military Relations in Russia and Eastern Europe. New York: Routledge. Birch, Sarah. 2007. “Electoral Management Bodies and the Electoral Integrity: Evidence from Eastern Europe and the Former Soviet Union.” Paper presented at the Annual Meeting of the American Political Science Association, Chicago, Illinois. Bode, Carl, ed. 1947. The Portable Thoreau. New York: Penguin Books. Boix, Carles. 2003. Democracy and Redistribution. New York: Cambridge University Press. Bojcun, Marko. 1995. “The Ukrainian Parliamentary Elections in March-April 1994,” Europe-Asia Studies Vol. 47, No. 2 (March): 229–249. Borowiec, Andrew. 2000. Cyprus: A Troubled Island. Westport, CT: Praeger. Box-Steffensmeier, Janet M. and Bradford S. Jones. 2004. Event History Modeling: A Guide for Social Scientists. New York: Cambridge University Press. Bratton, Michael. 1998. “Second Elections in Africa.” Journal of Democracy Vol. 9, No. 3 (July): 51–66. Brown, Andrew. 2007. “The Thaksinization of Thailand,” Journal of Contemporary Asia Vol. 37. Bueno de Mesquita, Bruce and George W. Downs. 2005. “Development and Democracy.” Foreign Affairs (September/October).

165

166

Dilemmas of Democratic Consolidation

Bueno de Mesquita, Bruce, James D. Morrow, and Ethan R. Zorick. 1997. “Capabilities, Perception, and Escalation.” American Political Science Review Vol. 91, No. 1 (March): 15–27. Bunce, Valerie. 2000. “Comparative Democratization: Big and Bounded Generalizations.” Comparative Political Studies Vol. 33, No. 6/7 (August/September 2000): 703–734. Carey, John M. 1997. “Institutional Design and Party Systems,” in Larry Diamond, Marc F. Plattner, Yun-han Chu, and Hung-mao Tein, eds. Consolidating the Third Wave of Democracies: Themes and Perspectives. Baltimore: The Johns Hopkins University Press. Carey, John and Andrew Reynolds. 2007. “Parties and Accountable Government in New Democracies.” Party Politics Vol. 13, No. 2: 255–274. Carothers, Thomas. 2002. “The End of the Transition Paradigm.” Journal of Democracy Vol. 13, No. 1 (January): 5–21. ———. 2004. Critical Mission: Essays on Democracy Promotion. Washington, DC: Carnegie Endowment for International Peace. Carothers, Thomas and Marina Ottaway, eds. 2000. Funding Virtue: Civil Society Aid and Democracy Promotion. Washington, DC: Carnegie Endowment for International Peace. Chalk, Peter. 2008. “The Malay-Muslim Insurgency in Southern Thailand: Understanding the Conflict’s Evolving Dynamic.” RAND Counterinsurgency Study—Paper 5. Santa Monica, California: RAND Corporation. Chand, Vikram K. 1997. “Democratization from the Outside In: NGO and International Efforts to Promote Open Elections.” Third World Quarterly Vol. 18, No. 3: 543–561. Cheibub, José Antonio. 2007. Presidentialism, Parliamentarism, and Democracy. New York: Cambridge University Press. Committee on Evaluation of USAID Democracy Assistance Programs, National Research Council, National Academy of Sciences. 2008. Improving Democracy Assistance: Building Knowledge through Evaluations and Research. Washington, DC: The National Academies Press. Coppedge, Michael. 1994. “Prospects for Democratic Governability in Venezuela.” Journal of Inter-American Studies (Summer). D'Anieri, Paul, Robert Kravchuk, and Taras Kuzio. 1999. Politics and Society in Ukraine. Boulder, Colorado: Westview Press. Dahl, Robert A. 1971. Polyarchy. New Haven, CT: Yale University Press. De Zeeuw, Jeroen. 2005. “Projects Do Not Create Institutions: The Record of Democracy Assistance in Post-Conflict Societies.” Democratization Vol. 12, No. 4 (August): 481–504. De Zeeuw, Jeroen and Krishna Kumar, eds. 2006. Promoting Democracy in Post-Conflict Societies. Boulder, Colorado: Lynne Rienner. Diamond, Larry. 1999. Developing Democracy: Toward Consolidation. Baltimore: The Johns Hopkins University Press. ———. 2008. “The Democratic Rollback: The Resurgence of the Predatory State.” Foreign Affairs (March/April). Diamond, Larry, Juan J. Linz, and Seymour Martin Lipset, eds. 1989. Democracy in Developing Countries: Latin America. Boulder, Colorado: Lynne Rienner.

References

167

Durutalo, Alumita L. 2000. “Elections and the Dilemma of Indigenous Fijian Unity.” In Lal, Brij V., ed., Fiji Before the Storm: Elections and the Politics of Development. Asia Pacific Press. Ellingsen, Tanja. 2000. “Colorful Community or Ethnic Witches’ Brew? Multiethnicity and Domestic Conflict During and After the Cold War.” Journal of Conflict Resolution Vol. 44, No. 2: 228–249. Epstein, David L., Robert H. Bates, Jack Goldstone, Ida Kristensen, and Sharyn O’Halloran. 2006. “Democratic Transitions.” American Journal of Political Science Vol. 50, No. 3 (July): 551–569. Fearon, James D. 1995. “Rationalist Explanations for War.” International Organization Vol. 49 (Summer): 379–414. ———. 1998. “Commitment Problems and the Spread of Ethnic Conflict,” in David A. Lake and Donald S. Rothchild, eds. The International Spread of Ethnic Conflict. Princeton, New Jersey: Princeton University Press. Fearon, James D. and David D. Laitin. 2008. “Integrating Qualitative and Quantitative Methods.” In The Oxford Handbook of Political Methodology. Janet Box-Steffensmeier, Henry E. Brady, and David Collier, eds. (New York: Oxford University Press): 756–776. Finkel, Steven E., Aníbal Perez-Liñan, and Mitchell A. Seligson. 2007. “The Effects of U.S. Foreign Assistance on Democracy Building, 1990-2003.” World Politics Vol. 59, No. 3 (April): 404–440. Finkel, Steven E., Aníbal Perez-Liñan, Mitchell A. Seligson, and C. Neal Tate. 2008. “Deepening Our Understanding of the Effects of U.S. Foreign Assistance on Democracy Building: Final Report.” Unpublished manuscript dated January 28. Fraenkel, Jon. 2001. “The Alternative Vote System in Fiji: Electoral Engineering or Ballot-Rigging?” Commonwealth and Comparative Politics Vol. 39, No. 2 (July): 1–31. Fraenkel, Jon and Bernard Grofman. 2006. “Does the Alternative Vote Foster Moderation in Ethnically Divided Societies?” Comparative Political Studies Vol. 39, No. 5: 623–651. Franck, Thomas M. 1992. “The Emerging Right to Democratic Governance.” The American Journal of International Law Vol. 86, No. 1 (January): 46– 91. Friedman, Thomas L. 2000. The Lexus and the Olive Tree. New York: Anchor Books. Ganguly, Sumit. 2007. “India’s Unlikely Democracy: Six Decades of Independence.” Journal of Democracy Vol. 18, No. 2 (April). Geddes, Barbara. 1999. “What Do We Know about Democratization after Twenty Years?” Annual Review of Political Science Vol. 2: 115–144. Gibbons, Robert. 1992. Game Theory for Applied Economists. Princeton, New Jersey: Princeton University Press. Gilman, Nils. 2003. Mandarins of the Future: Modernization Theory in Cold War America. Baltimore: The Johns Hopkins University Press. Goldstone, Jack A. and Jay Ulfelder. 2005. “How to Construct Stable Democracies.” The Washington Quarterly Vol. 28, No. 1 (Winter 2004– 2005), pp. 9–20. Green, Andrew T. and Richard D. Kohl. 2007. “Challenges of Evaluating Democracy Assistance: Perspectives from the Donor Side.” Democratization Vol. 14, No. 1 (February): 151–165.

168

Dilemmas of Democratic Consolidation

Grugel, Jean. 2002. Democratization: A Critical Introduction. New York: Palgrave Macmillan. Haggerty, Richard A., ed. 1993. Venezuela: A Country Study. Washington, DC: Federal Research Division, Library of Congress. Hardin, Russell. 1997. “Democracy on the Margin,” Chapter 12 in Albert Breton, Gianluigi Galeotti, Pierre Salmon, and Ronald Wintrobe, eds., Understanding Democracy: Economic and Political Perspectives. New York: Cambridge University Press. Hassner, Pierre. 2008. “Russia’s Transition to Autocracy.” Journal of Democracy Vol. 19, No. 2 (April): 5–15. Henderson, Sarah. 2003. Building Democracy in Contemporary Russia. Ithaca, New York: Cornell University Press. Hyde, Susan D. 2006. “Observing Norms: Explaining the Causes and Consequences of Internationally Monitored Elections.” Doctoral Dissertation. Department of Political Science, University of California, San Diego. ———. 2007. “The Observer Effect in International Politics: Evidence from a Natural Experiment.” World Politics Vol. 60, No. 1 (October): 37–63. Hyde, Susan D. and Nikolay Marinov. 2007. “An Informational Theory of Election Observers, Allowing for Domestic and International Audiences.” Paper presented at the National Meeting of the Midwest Political Science Association, Chicago, Illinois (April). Inglehart, Ronald and Chris Weltzel. 2005. Modernization, Cultural Change and Democracy. New York: Cambridge University Press. Kagan, Robert. 2008. “The End of the End of History.” The New Republic (April 23). Karl, Terry Lynn. 1987. “Petroleum and Political Pacts: The Transition to Democracy in Venezuela.” Latin American Research Review Vol. 22, No. 1: 63–94. Kumar, Krishna. 2004. “International Political Party Assistance: An Overview and Analysis.” Working Paper 33, Netherlands Institute of International Relations ‘Clingendael’, Conflict Research Unit (October). ———. 2006. Promoting Independent Media: Strategies for Democracy Assistance. Boulder, CO: Lynne Rienner. Lal, Brij V., ed. 2000. Fiji Before the Storm: Elections and the Politics of Development. Asia Pacific Press. Levi, Margaret. 1988. Of Rule and Revenue. Berkeley: University of California Press. Linz, Juan J. 1978. The Breakdown of Democratic Regimes: Crisis, Breakdown, and Reequilibration. Baltimore, Maryland: The Johns Hopkins University Press. ———. 1990. “The Perils of Presidentialism.” Journal of Democracy Vol. 1, No. 1: 51–69. Linz, Juan J. and Alfred Stepan, eds. 1978. The Breakdown of Democratic Regimes: Latin America. Baltimore, Maryland: The Johns Hopkins University Press. Linz, Juan J. and Alfred Stepan. 1996. Problems of Democratic Transition and Consolidation: Southern Europe, South America, and Post-Communist Europe. Baltimore, Maryland: The Johns Hopkins University Press.

References

169

Linz, Juan J. and Arturo Valenzuela, eds. 1995. The Failure of Presidential Democracy. Baltimore, Maryland: The Johns Hopkins University Press. Lipset, Seymour Martin. 1981. Political Man: The Social Bases of Politics [expanded edition]. Baltimore, Maryland: The Johns Hopkins University Press. López-Pintor, Rafael. 2000. Electoral Management Bodies as Institutions of Governance. New York: UN Development Programme. Marinov, Nikolay and Hein Goemans. 2008. “What Happened to the Coup d’Etat? International Responses to the Seizure of Executive Power.” Paper presented at the National Meeting of the Midwest Political Science Association, Chicago, Illinois (April). McCoy, Jennifer. 1998. “Moderating and Mediating Elections during Latin American Democratization,” in Kevin Middlebrook, ed., Electoral Observation and Democratic Transitions in Latin America. La Jolla, California: Center for US-Mexican Studies. McFaul, Michael. 2001. Russia’s Unfinished Revolution: Political Change from Gorbachev to Putin. Ithaca, New York: Cornell University Press. ———. 2004. “Democracy Promotion as a World Value.” The Washington Quarterly Vol. 28, No. 1 (Winter). ———. 2005. “Transitions from Postcommunism.” Journal of Democracy Vol. 16, No. 3 (July): 5–19. McFaul, Michael and James Goldgeier. 2005. “What to Do about Russia.” Policy Review (October/November). McMillan, John. 2002. Reinventing the Bazaar: A Natural History of Markets. New York: W.W. Norton & Co. Mehta, Bhanu. 2007. “India’s Unlikely Democracy: The Rise of Judicial Sovereignty.” Journal of Democracy Vol. 18, No. 2 (April): 70–83. Melia, Thomas O. 2006. “The Democracy Bureaucracy.” The American Interest Vol. 1, No. 4 (Summer): 122–130. Milgrom, Paul and John Roberts. 1992. Economics, Organization, and Management. Englewood Cliffs, New Jersey: Prentice-Hall. Morrison, Andrea L. 2008. “Democratic Breakdown Reconsidered.” Paper presented at the National Meeting of the Midwest Political Science Association, Chicago, Illniois (April). Norris, Pippa. 2004. Electoral Engineering: Voting Rules and Political Behavior. New York: Cambridge University Press. North, Douglass C. 1990. Institutions, Institutional Change, and Economic Performance. New York: Cambridge University Press. North, Douglass C., John Joseph Wallis, and Barry R. Weingast. 2009. Violence and Social Orders: A Conceptual Framework for Interpreting Recorded Human History. New York: Cambridge University Press. O’Donnell, Guillermo A. 1973. Modernization and BureaucraticAuthoritarianism: Studies in South American Politics. Berkeley, CA: Institute for International Studies, University of California, Berkeley. O’Donnell, Guillermo. 1998. “Horizontal Accountability in New Democracies.” Journal of Democracy Vol. 9, No. 3 (July): 112–126. O’Donnell, Guillermo and Philippe C. Schmitter. 1986. Transitions from Authoritarian Rule: Tentative Conclusions about Uncertain Democracies. Baltimore, Maryland: The Johns Hopkins University Press.

170

Dilemmas of Democratic Consolidation

Olson, Mancur. 1993. “Dictatorship, Democracy, and Development.” American Political Science Review Vol. 87, No. 3 (September): 567-576. Pastor, Robert A. 1999. “The Role of Electoral Administration in Democratic Transitions: Implications for Policy and Research.” Democratization Vol. 6, No. 4 (Winter): 1–27. Phatharathananunth, Somchai. 2008. “The Thai Rak Thai Party and Elections in North-East Thailand.” Journal of Contemporary Asia Vol. 38. Phongpaichit, Pasuk and Chris Baker. 2008. “Thaksin’s Populism,” Journal of Contemporary Asia Vol. 38. Pierson, Paul. 2005. Politics in Time: History, Institutions, and Social Analysis. Princeton, New Jersey: Princeton University Press. Plattner, Marc. 2002. “Globalization and Self-Government.” Journal of Democracy Vol. 13, No. 3 (July): 54–67. Posen, Barry R. 1993. “The Security Dilemma and Ethnic Conflict.” In Michael E. Brown, ed. Ethnic Conflict and International Security. Princeton, New Jersey: Princeton University Press. Przeworski, Adam. 1991. Democracy and the Market. New York: Cambridge University Press. ———. 2005. “Democracy as an Equilibrium.” Public Choice Vol. 123: 253– 273. Przeworski, Adam, Michael E. Alvarez, José Antonio Cheibub, and Fernando Limongi. 2000. Democracy and Development: Political Institutions and Well-Being in the World, 1950–1990. New York: Cambridge University Press. Rapoport, David C. 1968. “The Political Dimensions of Military Usurpation.” Political Science Quarterly Vol. 83, No. 4 (December): 551–572. Reynolds, Andrew, ed. 2002. The Architecture of Democracy: Constitutional Design, Conflict Management, and Democracy. New York: Oxford University Press. Roe, Paul. 1999. “The Intrastate Security Dilemma: Ethnic Conflict as a ‘Tragedy’?” Journal of Peace Research Vol. 36, No. 2 (March): 183–202. Salih, Halil Ibrahim. 1978. Cyprus: The Impact of Diverse Nationalism on a State. Tuscaloosa, Alabama: University of Alabama Press. Schedler, Andreas. 1998. “What Is Democratic Consolidation?” Journal of Democracy Vol. 8, No. 2 (April): 91–107. Schelling, Thomas C. 1980. The Strategy of Conflict (1980 edition). Cambridge, Massachussetts: Harvard University Press. Schmitter, Philippe and Terry Lynn Karl. 1990. “What Democracy Is…and Is Not.” Journal of Democracy. Schmitter, Philippe C. and Javier Santiso. 1998. “Three Temporal Dimensions to the Consolidation of Democracy.” International Political Science Review Vol. 19, No. 1: 69–92. Sen, Amartya. 1999. “Democracy as a Universal Value.” Journal of Democracy, Vol. 10, No. 3 (July). Simon, Herbert A. 1976. Administrative Behavior: A Study of Decision-Making Processes in Administrative Organization (Third Edition). New York: The Free Press. Sontag, Deborah. 2003. “The Erdoğan Experiment.” The New York Times Magazine (May 11).

References

171

Svolik, Milan. 2008. “Authoritarian Reversals and Democratic Consolidation.” American Political Science Review Vol. 102, No. 2 (May): 153–168. Tsebelis, George. 1990. Nested Games: Rational Choice in Comparative Politics. Berkeley: University of California Press. ———. 2002. Veto Players: How Political Institutions Work. Princeton, New Jersey: Princeton University Press. Ulfelder, Jay. 2005. “Contentious Collective Action and the Breakdown of Authoritarian Regimes.” International Political Science Review Vol. 26, No. 3 (July): 311–334. ———. 2007. “Natural-Resource Wealth and the Survival of Autocracy.” Comparative Political Studies Vol. 40, No. 8 (August): 995–1018. ———. 2008. “International Integration and Democratization: An Event History Analysis.” Democratization Vol. 15, No. 2 (April). Ulfelder, Jay and Michael Lustik. 2007. “Modelling Transitions To and From Democracy.” Democratization, Vol. 14, No. 3 (June): 351–387. Valenzuela, Arturo. 2004. “Latin American Presidencies Interrupted.” Journal of Democracy Vol. 15, No. 4 (October): 5–19. Van Zon, Hans. 2000. The Political Economy of Independent Ukraine. New York: St. Martin's Press. Walter, Barbara F. and Jack Snyder, eds. 1999. Civil Wars, Insecurity, and Intervention. New York: Columbia University Press. Wantchekon, Leonard. 2004. “The Paradox of ‘Warlord’ Democracy: A Theoretical Investigation.” American Political Science Review Vol. 98, No. 1 (February): 17–33. Weingast, Barry R. 1997. “The Political Foundations of Democracy and the Rule of Law.” American Political Science Review, Vol. 91, No. 2 (June): 245–263. Welzel, Christian and Ronald Inglehart. 2008. “The Role of Ordinary People in Democratization.” Journal of Democracy Vol. 19, No. 1 (January): 126– 140. Wilson, James Q. 1989. Bureaucracy: What Government Agencies Do and Why They Do It. New York: Basic Books. Wood, Elisabeth Jean. 2003. Insurgent Collective Action and Civil War in El Salvador. New York: Cambridge University Press. ———. 2005. “Challenges to Political Democracy in El Salvador.” In Frances Hagopian and Scott P. Mainwaring, eds. The Third Wave of Democratization in Latin America: Advances and Setbacks. New York: Cambridge University Press. Zagare, Frank C. and D. Marc Kilgour. 2000. Perfect Deterrence. New York: Cambridge University Press.

Index Acemoglu, Daron, 11, 13, 16, 78, 101, 107 Alexander, Gerard, 103–104, 105, 106, 107, 110 analytic narratives, 20–21, 74–75, 111, 113, 153 Argentina, 34

democracy, consolidation of, 11, 152; definition of, 4–6; economic development and, 63–67; observation of, 56–57; prevalence of, 58– 59 Dahl, Robert A., 4–5 De Zeeuw, Jeroen, 132 Diamond, Larry, 8, 11, 18, 137, 142, 152–153, 156

Bangladesh, 33–34, 155 Bates, Robert H., 15, 20–21, 65, 66, 74, 157 Beissinger, Mark R., 136 Birch, Sarah, 86, 88–89, 139, 140 Boix, Carles, 12–13, 16, 78 Bratton, Michael, 6 Bueno de Mesquita, Bruce, 143 Bush, George W. (president), 136, 143

Ecuador, 18 El Salvador, 41–42 election management body (EMB), 44, 139–141 election observation missions, 69–70, 141–142 executive coup, frequency of, 60–61; nature of, 16–17; observation of, 57

Carothers, Thomas, 127, 128–129, 132 Chávez, Hugo (president), 48, 116– 117, 155 civil liberties, 5, 17, 43, 117, 157 civil society, 2, 68, 120, 128, 132, 134–136, 147 Cold War, 67–68, 70 collective action, 17, 27–28, 36, 52, 154 Color Revolutions, 136–137 Coppedge, Michael, 114, 116 consolidation, see democracy, consolidation of coups, see military coup or executive coup credible commitments, 13, 37–42, 52, 64, 113 Cyprus, 93–101, 154–155

Fearon, James, 74, 97, 144 Fiji, 6, 75–84, 143, 156 Finkel, Steven, 68, 131–132, 138 Franck, Thomas M., 156 Friedman, Thomas L., 66 game theory, 12–13, 20, 38–39; Prisoner’s Dilemma, 37; Stag Hunt, 32 Ganguly, Sumit, 44 Geddes, Barbara, 9, 15 Georgia, 90, 136 Greece, 46–47 Green, Andrew T., 132 Haggerty, Richard A., 114 Hardin, Russell, 66, 151 Honduras, 47–49 horizontal accountability, 42–44 Hyde, Susan, 6, 69, 141

de Figueiredo, Rui, 12, 99

173

174

Dilemmas of Democratic Consolidation

IMF (International Monetary Fund), 79, 87–88 India, 43–44 Inglehart, Ronald, 7–8, 9 instrumental rationality, 27 Kagan, Robert, 156, 158 Karl, Terry Lynn, 5, 21, 38, 114, 115 Kenya, 2, 156 Kilgour, D. Marc, 27 Kohl, Richard D., 132 Lijphart, Arend, 4 Linz, Juan, 5, 8, 10–11, 103, 107, 109, 152 Lipset, Seymour Martin, 7–8, 9, 63 Makarios III (archbishop), 94, 98, 100 Marinov, Nikolay, 142 McFaul, Michael, 89–90, 136, 156 McMillan, John, 67 military coup, frequency of, 60–61; nature of, 18–19; observation of, 57 Milgrom, Paul, 14, 130 Millennium Challenge Account, 143 modernization theory, 7–9, 123 nested games, 154–155 Niskanen, 14 North, Douglass C., 65–67 OAS (Organization of American States), 47–48, 143 O’Donnell, Guillermo, 34, 38–39, 43, 125 organizations, 13–14 OSCE (Organization for Security and Cooperation in Europe), 85, 87 Ottaway, Marina, 132 pacts, 38–42, 53, 113; Moncloa (Spain), 106; Punto Fijo (Venezuela), 114–116

Perez-Liñan, Anibal, 68, 131–132, 138 Pierson, Paul, 66, 157 Posen, Barry R., 97 presidentialism, 8, 153 principal-agent problem, 14, 130 Przeworski, Adam, 9, 11, 12–13, 14, 16, 118, 151 random narratives, 74 Rapoport, David C., 16 rebellion, frequency of, 60–61; nature of, 17–18; observation of, 58 rents, 44, 64–65, 87 Roberts, John, 14, 130 Robinson, James A., 11, 13, 16, 78, 101, 107 Russia, 69, 88, 90, 92, 136–137, 156 Rustow, Dankwart, 9–10, 101 Schedler, Andreas, 11, 152 Schelling, Thomas C., 40, 135 Schmitter, Philippe C., 21, 38, 39, 67, 125 Seligson, Mitchell, 68, 131–132, 138 Sen, Amartya, 156 Sierra Leone, 145–147 Simon, Herbert A., 14 Sontag, Deborah, 50 Soviet Union, see USSR Spain, 102–110, 114 Stepan, Alfred, 8, 38, 103, 107, 109 Svolik, Milan, 11 Thailand, 1–2, 117–123, 143, 156 Tsebelis, George, 5, 154 Turkey, 49–51, 118 Ukraine, collective action problems in, 89–92 Ulfelder, Jay, 65, 67, 157 USAID, 68, 127, 129–133, 138–139 USSR, 61, 84 Venezuela, 6, 48, 70, 113–117, 156 veto player, 5, 27

Index 175

Wallis, John Joseph, 64, 67 Walter, Barbara F., 97, 144, 146 Wantchekon, Leonard, 41 Weingast, Barry R., 12, 38, 39, 42, 64, 67, 99

Wilson, James Q., 129 Wood, Elisabeth Jean, 41 Zagare, Frank, 27

About the Book

Why have so many attempts at democracy in the past half-century failed? Confronting this much discussed question, Jay Ulfelder offers a novel explanation for the coups and rebellions that have toppled fledgling democratic regimes and that continue to threaten many new democracies today. Ulfelder draws on an original dataset of 110 democratic failures spanning 1955–2007 and also presents analytic narratives for six cases (Cyprus, Fiji, Spain, Thailand, Ukraine, and Venezuela) to illustrate why some governments survive while others collapse. Focusing on political parties and the military as key players in the "democracy game," he sheds light on the pathways by which new democracies slide all too often from founding elections to polarization and breakdown. Jay Ulfelder works for Science Applications International Corporation (SAIC) as director of research for the Political Instability Task Force.

177