LIFTING THE CURSE: DISTRIBUTION AND POWER IN PETRO-STATES

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

By

Ryan Kennedy, M.A.

*****

The Ohio State University 2008

Dissertation Committee: Approved by Professor Marcus Kurtz, Adviser

Professor Timothy Frye ______Professor Janet Box-Steffensmeier Adviser Political Science Graduate Program Professor Richard Gunther

Copyright by

Ryan Kennedy

1998

ABSTRACT

While the empirical correlation between fuel exports and authoritarianism has become conventional wisdom, influencing academics and policy-makers, the answer for why fuel exporters tend to be more authoritarian has remained elusive. This study proposes a model based on the increased importance of government decision-making in determining the distribution of economic goods in societies that are dependent on fuel exports. These increases in government distribution, as a result of fuel development revenues primarily accruing to the state, take on three forms: direct payments and subsidization, government ownership of industry, and a more arbitrary enforcement of property rights. Countries are affected differently, dependent upon both their level of dependence and the size of revenues garnered from fuel exports. In countries with high per-capita fuel export income, fuel revenues are stabilizing for authoritarian regimes, as they provide the resources for maintaining support. In countries which are heavily dependent on fuel exports in the economy, but with a relatively low per-capita income from those exports, the increased importance of distribution in the economy results in greater instability in democratic regimes, as it increases the temptation for both government and opposition parties to lock in their share of relatively scarce resources through exclusionary politics.

ii Utilizing a dataset covering 166 countries from 1965 to 2001, I demonstrate that

there is a general correlation between fuel export dependence and the importance of

government distribution. I also find that income from fuel exports is generally

stabilizing, especially for authoritarian regimes, while fuel dependence is destabilizing,

especially for democracies. In addition, the same patterns of accumulation that fuel

government distribution and resource competition, also promote incentives towards

under-provision of public goods. This study demonstrates that fuel exporting states tend

to have worse socio-economic performance than would be expected from their level of

income. Finally, this study uses an in-depth case comparison of politics in Kazakhstan,

Kyrgyzstan and Moldova to demonstrate the causal mechanisms, and to test the dynamics of the model. Based on these findings, policy recommendations are made for methods of distribution that produce incentives more consistent with democratic governance.

iii

Dedicated to my wife, Allison

iv ACKNOWLEDGMENTS

I am greatly indebted to a number of people for their help on this project. First, I

wish to thank my dissertation committee. I am grateful my advisor, Marcus Kurtz, who

has consistently pushed me on the development of my theory, has provided many insights that have been key to the project’s development, and played a critical role developing a cohesive thesis. Thanks are also in order to Timothy Frye, who first suggested that I explore the issue of oil exports and democracy, has always pushed me to develop a cohesive theory, and guided me through the frustrations associated with organizing my fieldwork. Janet Box-Steffensmeier saved me from a number of conundrums with the empirics of this project. Her direction and recommendations have not only helped this project, but also furthered my understanding of methodology immensely. Finally,

Richard Gunther was largely responsible for placing this study into the larger context of the democratization literature and suggested a number of improvements that have made the chapters much clearer.

My fieldwork in Kazakhstan and Moldova would not have been possible without several grants. For Kazakhstan these included the AGGRS grant from Ohio State’s graduate school, the Office of International Affair’s dissertation travel grant, and Ohio

State’s distinguished university fellowship. For Moldova, I am extremely grateful to the

Fulbright foundation and their associated staff.

v A number of people helped me out greatly with my fieldwork. Ustina Marcus,

Donnacha O’Beachain, and a number of other members of the faculty and staff at the

Kazakhstan Institute for Management, Economics and Strategic Research (KIMEP) helped me out immensely with finding resources and lining up interviews in Kazakhstan.

Dinissa Duvanova helped with housing and background so I could hit the ground running. I also need to thank Robert Fahs for letting me have his couch when I unexpectedly needed a place to stay.

In Moldova, Irina Nicorich, Vlad Cojuhari, Ludmila Bilevschi, Irina Colin,

Ruxanda Negru, Chris Grant, Malise Tepera, and Patricia Fogerty deserve special recognition for all their help. Also, I would like to thank the staff of the Moldova

Resource Center for Human Rights (CReDO) and the Moldova State Institute for

International Relations (IRIM). Our second family in Moldova, the Cantirs, provided some key help, in terms of contacts, resources and sanity.

I am also very grateful to a number of people who have read previous drafts of this, and related projects, for their input. Among others, these included Christina Xydias,

Leanne Powner, Ben Ansell, Michael Cohen, and Johannes Urpelainen.

More generally, my parents deserve recognition for all their support and patience through my ten years of higher education and trips abroad. They may not have always understood why I was doing what I was doing, but they were always supportive while I did it.

vi Finally, I must thank my wife, who put up with my fieldwork in Kazakhstan while we planned our wedding, and re-located with me to Moldova, Ohio and Texas. My sanity would not have survived this process without her.

vii VITA

May 1, 1979...... Born – Kansas City, Missouri, USA

2001...... B.A. political science and Russian, Truman State University

2004...... M.A. political science, The Ohio Sate University

2001-2005...... Graduate Teaching and Research Associate, The Ohio State University

2005...... Special guest researcher Kazakhstan Institute for Management, Economics and Strategic Planning

2006-2007...... Fulbright research fellow, Moldova

2007...... SBS summer survey fellow

2007-2008...... Lecturer, The Ohio State University

PUBLICATIONS

Research Publications

1. Kennedy, R. (2007). Fragments of economic accountability and trade policy. Foreign Policy Analysis, 3(2), 145-169.

2. Kennedy, R. (2006). A colorless election: The 2005 presidential elections in Kazakhstan and what it means for the future of the opposition. Problems in Post- Communism, 53, 46-58.

3. Ishiyama, J.T., & Kennedy, R. (2001). Mixed electoral systems, super presidentialism, candidate recruitment and political party development in Russia, Ukraine, Armenia and Kyrgyzstan. Europe-Asia Studies, 53(8), 1177-1191.

viii 4. Kennedy, R., & Ishiyama, J.T. (2001). Writing rights: Factors influencing the strength of rights clauses in post-Communist countries. The Journal of Political Science, 29, 1-29.

Analysis Publications

1. Kennedy, R. (2008, 2 April). Moldova: Chisinau-Gagauzia tensions remain strained. Oxford Analytica.

2. Kennedy, R. (2008, 15 January). Moldova: Old challenges represent themselves in 2008. Oxford Analytica.

3. Kennedy, R. (2007, 24 December). Moldova: Christmas row a sign of deeper tensions. Oxford Analytica.

4. Kennedy, R. (2007, 21 November). Moldova: HIV/AIDS is at point of becoming a pandemic. Oxford Analytica.

5. Kennedy, R. (2007, 8 August). Moldova: Contrary forces favour policy short- termism. Oxford Analytica.

6. Kennedy, R. (2007, 18 July). Moldova: Press freedoms remain problematic. Oxford Analytica.

7. Kennedy, R. (2007, 28 June). Moldova: Local elections show national trends. RFE/RL Belarus, Ukraine, and Moldova Report, 9(15).

8. Kennedy, R. (2007, 25 June). Moldova: Communists lose ground in local elections. Oxford Analytica.

9. Kennedy, R. (2007, 14 May). Moldova: Voronin pins hopes on encouraging investment. Oxford Analytica.

10. Kennedy, R. (2007, 3 May). Talk of Transdniestr agreement sparks speculation. RFE/RL Belarus, Ukraine and Moldova Report, 9(11). (also produced for online and radio presentation)

11. Kennedy, R. (2007, 27 April). Moldova/Russia: Reported Transnistria deal is unlikely. Oxford Analytica.

12. Kennedy, R. (2007, 5 April). Counting losses as Russian wine ban lingers. RFE/RL Belarus, Ukraine and Moldova Report, 9(10).

13. Kennedy, R. (2007, 30 March). Moldova/Romania: Visa row may result in looser ties. Oxford Analytica.

ix 14. Kennedy, R. (2007, 22 March). What's behind harsh criticism of Romania. RFE/RL Belarus, Ukraine and Moldova Report, 9(1).

15. Kennedy, R. (2007, 5 March). Moldova: Government action hinders media freedom. Oxford Analytica.

16. Kennedy, R. (2007, 22 February). Moldova's broadcast privatization -- reform or censorship? RFE/RL Belarus, Ukraine, and Moldova Report, 9(4). Re-published in The Moldova Foundation's Weekly News Bulletin, (2007, 27 February), 3(31). Re-published again in In Focus magazine. (also produced for online and radio presentation)

17. Kennedy, R. (2007, 9 February). An EU invasion waiting to happen. RFE/RL Belarus, Ukraine, and Moldova Report, 9(4). (also produced for online and radio presentation)

18. Kennedy, R. (2007, 11 January). Tricky year begins for new EU neighbour. Oxford Analytica.

19. Kennedy, R. (2007, 5 January). Unprecedented opportunities, challenges posed by $1.2 billion aid package. RFE/RL Belarus, Ukraine, and Moldova Report, 9(1).

20. Kennedy, R. (2006, 21 December). Tirsapol line hardens ahead of negotiations. Oxford Analytica.

21. Kennedy, R. (2006, 29 November). Kazakhstan learns to love Borat. Central Asia-Caucasus Analyst.

22. Marcus, U, Ibadildin, N., & Kennedy, R. (2005, 30 November). Vested interests to determine Kazakhstan's presidential race. Central Asia-Caucasus Analyst.

x TABLE OF CONTENTS

Page Abstract...... ii Dedication...... iv Acknowledgements...... v Vita...... viii

List of Tables ...... xiv List of Figures...... xvi

Chapters:

1. Introduction...... 1

1.1 Stability/instability and political regime outcomes...... 5 1.2 Instability as a result of economic crises...... 7 1.3 The difference between being dependent and rich...... 10 1.4 An overview of the theory...... 11 1.5 Socioeconomic development ...... 21 1.6 Outline for the study...... 23

2 Does “The Curse” Exist? ...... 27

2.1 Relationship between fuel and democracy, pre-1970 ...... 29 2.2 Oil shocks and the rise of the curse...... 32 2.2.1 Stability in the Middle East...... 35 2.2.2 Military regimes in Libya and Indonesia ...... 37 2.2.3 Instability in Africa...... 38 2.2.4 Cycles of Venezuela...... 39 2.2.5 Bringing “the curse” into a cross-national perspective ...... 41 2.3 What changed?...... 41 2.4 The empirical relationship between fuel dependence and democracy...... 46 2.5 Moving forward...... 57

3. Dependence, Income, and Regime Stability ...... 59

3.1 Limited explanations...... 61 3.1.1 Authoritarian stability...... 61 3.1.2 Democratic instability...... 63 xi 3.1.3 Ambiguous empirical results...... 66 3.1.4 Other explanations...... 72 3.2 Modeling the curse...... 77 3.2.1 Why are fuel exports different?...... 77 3.2.2 The importance of government distribution...... 84 3.2.3 Government distribution and democracy ...... 88 3.3 Innovations and implications ...... 98

4. Long Live the King; Or Down with Democracy?...... 101

4.1 Linking fuel exports and government distribution...... 102 4.1.1 Government consumption...... 109 4.1.2 Government ownership...... 113 4.1.3 Protection of property rights...... 115 4.1.4 Overall implications ...... 118 4.2 Authoritarian stability vs. democratic instability...... 118 4.2.1 Cross-country testing...... 119 4.2.2 Variables for testing stability ...... 120 4.2.3 Statistical methodology ...... 123 4.2.4 The effect of fuel exports on regime stability ...... 125 4.3 Effect of fuel exports on civil war...... 135 4.4 Evaluating support for the distributional model...... 141

5. Fuel Revenues and Development...... 143

5.1 Natural resource curse or disappointment?...... 147 5.2 Explaining the neighborhood effect...... 161 5.2.1 Education...... 173 5.2.2 Inequality and deprivation...... 178 5.3 Conclusions...... 187

6. Drilling for Authoritarian Resources, The Case of Kazakhstan ...... 191

6.1 Political contestation in Kazakhstan ...... 199 6.1.1 Limited pluralism, 1991-1995...... 202 6.1.2 Consolidation of power, 1995-2008...... 207 6.2 The political economy of Kazakhstan’s authoritarianism...... 219 6.2.1 Government spending...... 220 6.2.2 Government patronage and investment...... 228 6.2.3 Discretion over enforcement of property rights ...... 234 6.2.4 Results of government domination of domestic industry...... 248 6.3 Political developments in Moldova and Kyrgyzstan ...... 253 xii 6.3.1 Kyrgyzstan – liberalization, authoritarianism, and revolution ...... 257 6.3.2 Moldova – continuing competition ...... 263 6.4 Conclusions...... 269

7. Conclusions...... 272

7.1 Discussion of findings...... 272 7.2 Can we lift the curse?...... 278 7.3 Directions for future research...... 284 7.4 Concluding remarks ...... 286

Appendices:

Appendix A: List of Cases in Dataset...... 288 Appendix B: A Note on Statistical Methods in TSCS Data ...... 291

B.1 Common approaches to TSCS data...... 292 B.2 Problems with fixed effects (within) estimation for oil data...... 294 B.3 Testing the consistency of between and within estimates...... 297

Appendix C: Formalizing Government/Opposition Relations in Distribution Oriented Societies...... 302

C.1 The setup ...... 303 C.2 Democratic or authoritarian challenge ...... 306 C.3 The choice to challenge or not challenge ...... 307 C.4 Preventing a challenge...... 308 C.5 Equilibrium...... 309

Bibliography ...... 311

xiii LIST OF TABLES

Table Page

1.1 List of fuel dependent countries...... 3

1.2 A heuristic layout of the predictions of the theory...... 19

2.1 Effect of fuel exports on democracy in all states, 1965-2001...... 52

2.2 Effect of fuel exports on democracy in developing states, 1965-2001...... 53

3.1 Political outcomes in fuel dependent states, 1965-2001 ...... 67

3.2 Payoffs from the distribution game...... 91

4.1 Relationship between fuel export dependence and government consumption in developing countries ...... 111

4.2 Relationship between fuel export dependence and government ownership in developing countries ...... 115

4.3 Relationship between fuel export dependence and the protection of property right in developing countries...... 117

4.4 Effect of fuel exports on the stability of all regime types in developing countries ...... 127

4.5 Effect of fuel exports on the stability of authoritarian regimes in developing countries...... 129

4.6 Effect of fuel exports on the stability of democratic regimes in developing countries...... 131

4.7 Effect of fuel exports on civil conflicts in developing states, 1965-2001 ...... 140

5.1 Effect of fuel exports on democracy compared to regional trends in full sample, 1965-2001 ...... 151

xiv 5.2 Effect of fuel exports on democracy compared to regional trends in developing countries, 1965-2001...... 152

5.3 Economic vs. social indicators in fuel export dependent countries (and their respective cohorts), 1965-2001 ...... 157

5.4 Effect of fuel exports on education in developing countries, 1965-2001 ...... 177

5.5 Effect of fuel exports on inequality and poverty in developing states, 1965-2001 ...... 186

6.1 OSCE assessment of media coverage in the 2005 presidential election...... 210

6.2 Result of polls, 2005 presidential election ...... 215

6.3 Election results by district, 2005 presidential election...... 218

B.1 Hausman test of the fixed (within) vs. random effects model ...... 298

B.2 Hausman test of the fixed (within) vs. re-specified random effect model ...... 299

B.3 Hausman test of the fixed (between) vs. re-specified random effects model ...... 300

xv LIST OF FIGURES

Figure Page

1.1 Overview of the theory...... 16

2.1 Predicted probability of democracy based on fuel dependence for all countries, 1965-2001 ...... 54

2.2 Predicted probability of democracy based on fuel dependence in developing countries, 1965-2001...... 55

2.3 Predicted probability of democracy based on fuel export income for all countries, 1965-2001 ...... 56

2.4 Predicted probability of democracy based on fuel export income for all countries, 1965-2001 ...... 57

3.1 The variation in political regimes among fuel dependent states, 1965-2001...... 70

3.2 Location of major oil fields and ethnolinguistic groups in Nigeria ...... 79

3.3 Location of major oil fields and ethnolinguistic groups in Iraq...... 81

3.4 Crude oil prices, 1965-2001...... 84

3.5 One round of distribution game ...... 92

3.6 Effect of the percentage of income that is determined by government decision-making on the preference for democratic outcomes ...... 93

3.7 Effect of the percentage of income that is determined by government decision-making on the preference of challenging for a democratic regime or for an opposition-led authoritarian regime ...... 95

4.1 Parts of theory tested in Chapter 4 ...... 102

4.2 Effect of fuel export income on the expected hazard of regime failure in authoritarian regimes and all regimes in developing countries ...... 133

xvi 4.3 Effect of fuel export dependence on the hazard of regime failure in authoritarian, democratic, and all regime types in developing states ...... 134

5.1 Approximate oil revenue in Saudi Arabia, 1965-2001 ...... 163

5.2 Growth in GDP versus growth in debt in Nigeria, 1971-2002 ...... 169

6.1 Map of Kazakhstan and Kyrgyzstan ...... 193

6.2 Map of Moldova...... 196

6.3 Per-capita GDP in Kazakhstan, Kyrgyzstan, and Moldova, 1991-2006...... 198

6.4 Can the static model produce dynamic predictions?...... 201

6.5 Government consumption per-capita, 1991-2006...... 222

6.6 Map of the major oil and gas fields in 1995 and the primary concentrations of Russian populations in 1993 ...... 224

6.7 Importance of state investment ...... 230

6.8 Standard deviation of responses to enforcement and interpretation of regulations...... 237

6.9 Level of business influence over important economic decisions...... 239

6.10 Track of daily closing stock prices for PetroKazakhstan...... 245

6.11 Value of fuel exports per-capita, 1991-2006...... 251

6.12 Foreign direct investment (net inflow) per-capita, 1991-2006 ...... 252

6.13 Freedom House political rights scores, 1991-2006...... 254

6.14 Freedom House civil liberties scores, 1991-2006...... 254

6.15 Aid inflows per-capita, 1991-2006 ...... 258

B.1 Traditional example of OLS bias in TSCS data...... 293

B.2 Bias in fixed effects (within) models when variables move slowly...... 295 xvii CHAPTER 1

INTRODUCTION

Why are countries that export oil less likely to be democratic than their non-oil exporting counterparts? Is this because oil provides the resources for the maintenance of authoritarian regimes, or because competition over state resources makes democracies less stable? What are the implications for social wellbeing and socio-economic development in oil exporting states?

These questions form the foundation for this study, but they would have seemed strange to someone who was looking at these issues just thirty years ago. At the beginning of the oil boom in the 1970s, development economists pointed to the massive capital influx into oil exporting states, possibly the largest peacetime redistribution of capital (Amuzegar, 2001), as a historic opportunity. Developing oil exporting states, which had historically lacked the capital for investment and industrialization, were now flooded with funds. In Venezuela, Carlos Andrés Pérez proposed his plan for “La Gran

Venezuela,” meant to make the country modern and industrial in only a short period of time (see Karl, 1997, p. 123-126). In Nigeria, the government invested in everything from education and transportation to a major steel factory and automobile assembly plant in an effort to spur sustainable development (Pinto, 1982; Kretzmann and Nooruddin,

2006). Vast sums of steady income from oil exports were expected to allow Middle

Eastern economies to grow without inflation, while the incapacity of Western states to

1 maintain spending in the face of increased energy costs would devastate industrialized

economies. This newfound wealth, it was thought, would increase national living

standards, resulting in “greater individual freedoms, enhanced national political stability,

more social cohesion, improved civil society and, hopefully, both participatory

democracy and political maturity” (Amuzegar, 2001, p. x).1

Events since these writings have led most scholars to call the optimism of the

1970s an anachronism. Today the opinion of Juan Pablo Pérez Alfonso, who called oil

the “Devil's excrement,” or Sheik Ahmed, a Saudi Arabian oil minister who stated, “All

in all, I wish we had discovered water” (Ross, 1999, p. 297), are much more prominent.

Economists and political scientists have linked oil and natural gas exports to a number of

adverse outcomes, including slower economic growth (Sachs and Warner, 1995), less

democratic governance (Ross, 2001), and an increased probability of civil war (Ross,

2006).

The negative consequences of oil exports are quickly becoming conventional

wisdom, affecting policy discussions about the probability of democracy in Iraq

(Rotunda, 2004; Springborg, 2007; Birdsall and Subramanian, 2004), our policy towards

newly independent countries in Central Asia and the Caucasus (Cohen, 2006; US

Interests in Central Asian Republics, 1998; Ross, 2002), and IMF and World Bank loans

for oil development projects (Kretzmann and Noorudin, 2006). All of these attributes,

taken together, have received a popular moniker -- the “natural resource curse.” Despite

all of this, the mechanisms linking oil exports to these adverse political and economic outcomes are still a matter of debate.

1 Amuzegar (2001) is actually referencing his own work in 1975 where he reaches these conclusions (Amuzegar, 1975). 2 This study explores the link between oil exports and non-democratic political

outcomes. It argues that no single political outcome is consistent among oil exporting

states. For example, Saudi Arabia has remained stable under a traditional authoritarian regime, while Nigeria has alternated between democratic and authoritarian governance several times and experienced a severe civil conflict. Further, these variations are not a reflection of natural outliers to an underlying pattern (Smith, 2004). The real question for scholars is how the same attribute, oil exports, results in radically different political outcomes while still producing the overall pattern of non-democracy associated with the

“natural resource curse” literature.

Angola Iran Nigeria United Arab Emirates (UAE) Saudi Arabia Venezuela Azerbaijan Turkmenistan Oman Syria Kuwait Russia Trinidad Kazakhstan Gabon Indonesia Congo-Brazzaville Norway Yemen Ecuador Algeria Malaysia

Table 1.1: List of fuel export dependent countries. Fuel dependent defined as having an average of more than 5% of GDP from fuel exports during the period 1965-2001.

3 I propose that much of the variation in political outcomes can be found by looking at the difference between being dependent on fuel exports and being rich from fuel exports. While these may sound similar at first, and indeed the overlap has led to much conceptual confusion, there is a great deal of difference, especially in terms of political outcomes. One of the few areas of consensus in the oil development literature, is that, in developing states, the importance of the government in economic distribution increases as the importance of oil exports in the economy increases (Jensen and Wantchekon, 2004).

In those states which are dependent on fuel exports, but do not receive a large amount of per-capita income from those exports, the increased importance of controlling the government is not necessarily matched by increases in government capacity with which to maintain stability. This is destabilizing, especially for democracies where important groups have an incentive to lock in their share of relatively scarce resources through exclusive politics. In those states which are both dependent and rich, the importance of the government in economic distribution is matched by increased government resources to maintain stability, making these states more stable. In developing states, this effect is especially pronounced for authoritarian governments, who have little incentive to democratize, as doing so would introduce uncertainty in control of government resources, and also have the resources necessary to avoid instability.

All of this will be developed in greater detail throughout this study, but this chapter will give an overview of the arguments made in the remaining chapters.

4 1.1 Stability/instability and political regime outcomes

If indeed there is a negative relationship between oil exporting and democratic

governance, there are two ways that this relationship can come about. Fuel exports can

provide the resources that allow authoritarian governments to survive for longer, or the

competition over these resources may make democracies less likely to survive.

Some readers will recognize this as the inverse of the endogenous/exogenous

democratization literature started by Przeworski and his co-authors (Przeworski and

Limongi, 1997; Przeworski et. al., 2000). They find a difference between how variables, per-capita GDP in particular, affect the probability of democracy being established

(endogenous democratization) and how they affect the stability of democracies once they are established (exogenous democratization). While there is still quite a bit of controversy about Przeworski et. al.'s findings (see e.g. Boix and Stokes, 2003; Epstein et. al., 2006), their conceptual distinction has become fundamental to the discussion of democratization. Interestingly, most of the previous literature on the natural resource curse bifurcates neatly between those who argue that non-democratic governance comes about endogenously in oil exporting states by making democracies more fragile, and those who argue that non-democratic governance is determined exogenously and its stability is enhanced by fuel exporting.

Several causal mechanisms are put forth by those who argue that fuel exports make authoritarian regimes more stable. Oil exports are expected to provide the resources to insulate authoritarian governments from societal demands (Ross, 2001;

Mahdavy, 1970), to build institutions for stability (Smith, 2004; Smith, 2007), or to build up a stronger repressive apparatus (Scokpol, 1982; Clark, 1997). Still others, especially in the public policy literature, suggest that Western states are less likely to pressure 5 authoritarian states in oil exporting countries to change their politics (Bowman, 2005).

More recently, some scholars have argued that the resources from oil exports are

stabilizing to political regimes more generally, not just authoritarian states. Dunning

(2007) argues that oil exports provide the resources for democracies to distribute

resources to the masses without taxing elites, and, since elites pose the largest danger to

democratic stability (Acemoglu and Robinson, 2005), this makes democracies more stable. Similarly, Morrison (2007) argues that non-tax revenues, including those from oil exports, are stabilizing for both democracies and non-democracies, by allowing social spending in authoritarian states and lower taxation in democratic states. All of these arguments support the contention of Smith (2004) that there is a general pattern of stability among oil exporters.

Another area of the literature, however, argues exactly the opposite -- that oil exporters are more likely to experience social and political instability. Karl (1997) argues that fuel exporting states tend to spend for stability instead of building institutions, and, thus, when oil prices drop the state will not have the ability to maintain stability. Lam and Wantchekon (2002) suggest that democracies are destabilized by fuel exports because the current government has the ability to establish an incumbent advantage through spending fuel revenues, resulting in a one-party dictatorship. Finally, an entire

literature has arisen, arguing that there is a link between fuel exports and civil conflict

(Bannon and Collier, 2003; Ross, 2006).

Who is right? Do fuel exports provide the resources for stability of political

regimes, or do they promote competition that is destabilizing for political regimes? The

problem is that all of these theories accurately describe some subset of fuel exporting

countries during a particular period of time. While some scholars have recently 6 recognized that non-democratic outcomes in fuel exporting states come from both

democratic instability and authoritarian stability (Jensen and Wantchekon, 2004), there

has been little explication on what sets states in these various courses. In other words, the

real question is not whether fuel exports promote stability or instability, but whether we

can find some characteristic of fuel exports that explains the variety of political outcomes

in these states. This study sets about answering this question.

1.2 Instability as a result of economic crises

Throughout this study, I will be looking at political regime instability through a

political economy frame. As such, I primarily focus on how economic resources and

development affect political stability. This does not mean that other things, such as

ideology, history or even individual efforts, are not important, but that this study will be eschewing these items for intellectual clarity, treating them primarily as items to be controlled for in the models.

This being said, there is strong evidence that financial crises contribute to political regime changes. Bueno de Mesquita et. al. (2003, p. 26) state succinctly, "The survival

of leaders and of the institutions or regimes they lead is threatened when they are no longer able to provide sufficient resources to sustain political support." The danger to the

survival of a regime can come from a number of areas. Financial crises may undermine

the legitimacy of the political regime, especially if the regime justifies its political

structures on the basis of supporting economic performance (Rustow, 1970; Epstein,

1984; Linz and Stepan, 1996, p. 80). Crises may also fragment the coalition of interests

that provide support for the current political order, allowing the influence of alternative coalitions and groups (Markoff and Baretta, 1990). Conversely, it is often during times 7 of crises that competing elites have an incentive to come together in order to protect their interests (Bueno de Mesquita et. al., 2003).

Not all economic crises result in political crises, and political regimes attempt to develop other mechanisms for maintaining their resources for sustained political support.

In democracies, the rules of political competition allow competing groups to change the government without changing the political regime, in turn, allowing the institutions to maintain stability even as leaders face instability. Where conflicts over distribution are deeply rooted or where resources are scarce, however, this mechanism may have greater difficulty providing stability for the democratic regime (Powell, 1982; Powell, 1996;

Lipset, 1994).

Non-democratic states will often attempt to establish legitimacy on one of several, non mutually-exclusive, alternative grounds, including: super-natural/religious authority

(Weber, 1968; Bendix, 1978), exceptional qualities of the leader or hereditary leadership

(Weber, 1968; Benix, 1978),2 revolutionary or plebiscitory authority (Bienen and van de

Walle, 1991; Bratton and van de Walle, 1997, p. 78), or even the level of popular support based on limited competition (Bratton and van de Walle, 1997, p. 81). At different points in history, these appeals have shown varying amounts of success in establishing a pool of reserve support from which non-democratic political regimes can draw during financial crises. Since about the end of World War II, many of the traditional authoritarian appeals to legitimacy have become ineffectual. As Bendix (1978, p. 4) notes, “In our time, not only democracies but military regimes, dictatorships, and even constitutional monarchies

2 Similarly, Aristotle’s justification in The Politics for some states to have monarchies or kings is, “When therefore it comes about that there is either a whole family or even some one individual that differs from all other citizens in virtue so greatly that his virtue exceeds that of all others, then it is just for this family to be the royal family or this individual king, and sovereign over all matters” (3.11.12). Exceptional qualities of leaders often overlap with some super-natural/religious source of authority. 8 are legitimized by claims of popular mandate. Indeed, other ways of justifying authority

have become inconceivable.”

Beyond developing non-economic bases for legitimate rule, political regimes also

have different levels of resources available to maintain power during financial crisis. All states have punishment mechanisms that can be utilized for maintaining social stability.

The most obvious of these are the police and military. Some have pointed to the weakness of the government’s repressive apparatus as a necessary condition for social revolution (Skocpol, 1979). Others point to the ability of the government to exclude those who defect from the ruling coalition from receiving private economic goods (Bueno de Mesquita et. al., 2003). Greater capacity for punishment of those who challenge the current political regimes helps political systems remain in power, despite financial crises.

Finally, temporary growth crises are less crucial for those countries with access to a reserve of capital. These reserves provide insulation against short-term crises. Some states may have access to enough revenue that they run a capital-surplus during times of economic growth, and have access to enough capital to provide resources to key support groups during economic downturns (Karl, 1997). Other states retain access to lending, either through the sale of government bonds or through international lending institutions, which provide them with the resources necessary to offset economic maladies (Easterly,

2001; Bueno de Mesquita and Root, 2002). In either of these cases, the state maintains the resources to keep the political support of loyal cadres who demand “material reward and social honor” (Weber, 2002, p. 16). Thus, McGlinchey (2003, p. 25) notes the importance of “distinguishing between an economic crisis in society and an economic crisis of the state” (emphasis in original).

9

Oil exports provide resources that can impact at least two of these mechanisms.

First, where the government is able to control the distribution of oil revenues, the capacity for selective exclusion increases. Second, where oil revenues are high per- capita, states should be capable of running a capital surplus and maintaining the support of loyal cadres, even during times of lower oil prices (Karl, 1997). The amount of resources garnered by states from their oil exports, however, varies widely and is not always proportionate to the country's dependence on these exports.

1.3 The difference between being dependent and rich

Countries that are dependent on oil exports are not always rich from those exports.

Nor are countries that receive a large amount of income from fuel exports necessarily dependent on those resources. From 1965 to 2001, Angola and Nigeria were dependent on fuel exports for an average of 39 and 40 percent of their GDP respectively, and they received $106 and $200 per-capita from those exports. Compare this to Saudi Arabia, which was dependent on fuel exports for about the same amount of its GDP (36 percent), but received approximately $3,300 per-capita from those exports.3 Conversely, Norway

received an average of $2,366 per-capita from its fuel exports during this time, but this

made up a little less than eight percent of its GDP.4 Comparing the level of dependence

on fuel exports with the level of income from those exports during this period produces

only a moderate relationship.5

3 Calculated from World Bank (2003) data, see Table 3.1. 4All dollar amounts in constant 2000 US dollars. 5The Pearson's correlation between fuel exports as a percent of merchandise exports against per-capita income from fuel exports for the sample of 166 developing countries from 1965 to 2001 is a modest .557. 10 Yet, the literature on the “natural resource curse” rarely makes this distinction.

Some scholars have utilized variables that measure dependence on fuel exports, usually

indexed against GDP, but refer to those variables as export income (e.g. Ross, 2001;

Smith, 2004). Others have been more explicit in looking at dependence on fuel exports,

excluding countries with high income from fuel exports from their analysis (e.g. Karl,

1997). Still others utilized measures of per-capita income from fuel exports, arguing that

dependence is theoretically uninteresting. The most audacious statement of this came from Ross (2006, p. 276) who argues that “There is no good rationale for measuring natural resource production (or exports) as a fraction of GDP.”

This study argues that dependence on fuel exports and income from fuel exports are not the same, and that both are theoretically interesting. Indeed, they predict very different political outcomes among fuel exporting states.

1.4 An Overview of the theory

The theory for this study builds on the link between fuel exports and government resources. In the developing world, most of the funds from oil production accrue to the government. There are a number of reasons for this which will be developed in greater detail later. For now, it suffices to note that developing countries usually do not have the capacity for fuel exploitation by private domestic entrepreneurs, and, even if they did, the

resulting businesses would so distort the economy that these ventures would most likely become untenable. Because of this, most developing countries exploit their resources

through revenue or production sharing agreements with international oil companies, or

they set up a state-owned oil company. In either case, the results, as Ross (2 May 2006)

said in a speech in Azerbaijan, is that oil makes governments, but not necessarily the 11 country, rich.

Figure 1.1 shows the logical flow of incentives from this foundation. If the

country is also dependent on fuel exports for a large portion of its economy, the importance of government decision-making in the distribution of economic goods is

increased. This is manifested in three characteristics of fuel export dependent states:

increased government spending, increased importance of government investment and

business, and an increased discretion over the enforcement of property rights. Of these,

increased government spending has received the most attention from scholars, in part,

because it is the most apparent. Pictures of skyscrapers and man-made islands in the

Middle East readily come to mind. Some less dramatic associations between fuel exports

and government spending on infrastructure, social welfare programs, and outright

patronage and corruption have also been stressed by scholars (e.g. Lam and Wantchekon,

2002; Morrison, 2007; Global Witness, 2004).

The importance of government investment and business has received much less

attention, in large part because it is more difficult to quantify. Yet, case studies have long

noted that direct and indirect involvement of the government in business is a prominent

feature of fuel exporting states (see e.g. Bremmer, 2004; Salameh and Steir, 1980, p. 10-

11; Osaghae, 1998; Olcott, 2002). As the main area for the accumulation of capital, the

government also becomes the most important source of business investment. Even where

there is not explicit state ownership of oil production, the state is usually involved in

many of the large-scale businesses and industries outside of the oil sector. Indirectly, government officials and/or close relations of government officials are often involved in

the ownership and management of the country's most important industries. Some of this

stems from a desire by foreign investors to safeguard their investments by involving 12 those with close government relations in their business ventures. This relationship also stems from the access to oil revenues and investment from the government enabled by close government connections. In some cases, such as Saudi Arabia or Kazakhstan, family members of the governing family hold key positions in the major domestic companies. In other cases, such as Nigeria, those with close government connections occupy points along chains of patronage through local government positions or managing parastatal enterprises. However it comes about, the role of state decision-makers in the distribution of economic goods is enhanced.

Finally, oil resources give the state increased discretion over the enforcement of property rights. Charles Lindblom (1982) famously called free markets a prison for governments, because there was only so much that a government could do in pursuing certain courses of action without hurting economic growth. In developing countries, the enforcement of property rights has been an especially important aspect of this (North,

1981; Bauer, 1984; Bauer, 2000). Without a reliable system of property rights, businesses have little incentive to invest, since they cannot be certain of future profits.

The consensus among development economists is that an unstable business environment, characterized by non-enforced or arbitrary property rights, is harmful to investment and growth in developing states. Oil changes some of these calculations in investment, as it is a very high rent resource, meaning that once the resource is developed, the sale price is far above the cost of extraction (Dunning, 2007). Additionally, since the oil boom of the

1970s, the large multi-national oil companies have faced increased competition for oil resources from wildcat oil drilling and, more recently, national oil companies in developing countries like India and China. All of this results in companies being willing to invest large amounts of money, even when there is some uncertainty about future 13 property rights. Exxon Mobil returned to Venezuela after the oil nationalizations of

1976. The Chinese National Petroleum Company (CNPC) was willing to spend $4.2 billion to purchase PetroKazakhstan, even though the previous owners were essentially squeezed out of the Kazakhstani market by government pressure (see Chapter 7). In

2006, Royal Dutch Shell was forced to sell a majority stake in Sakhalin 2, the world’s largest combined oil and natural gas development, to Russia’s energy monopoly,

Gazprom (Kramer, 2006). Yet, foreign direct investment in Russia nearly doubled, year on year, in 2007 to almost $50 billion (RIA Novosti, 2007). New oil exploration in Sudan and Chad should also serve as evidence that international oil companies are willing to invest even in areas where stability of property rights, or stability more generally, is uncertain (Frynas and Paulo, 2007). From a technical perspective, this is not a causal relationship – there is nothing inherent in oil development that would lead a country to more weak enforcement of property rights. To the extent that oil reduces the cost of, and the government can increase its influence or revenue through such practices, however, they become more likely.

That the growth in capital associated with fuel exports accrues primarily to the government and increases the importance of government decision-making, suggests that economic growth associated with oil exploitation is qualitatively different from other types of economic growth. Perhaps the most well-known theory in comparative politics is “modernization theory,” which contends that higher levels of economic wealth, through a number of different social mechanisms, make a country more likely to be democratic (Lipset, 1959; Diamond, 1992). The evidence of a link between economic development and democracy has become so voluminous that many scholars today take it as one of the few accepted facts in political science (Reuschmeyer et. al., 1992). 14 However, one of the primary, but usually implicit, assumptions in many variants of modernization theory is that the resources from economic growth accrue to and result in the development of various social groups, whether that be the middle class (Lipset, 1960), the working class (Reuschmeyer et. al., 1992), a town-dwelling bourgeoisie (Moore,

1966), civil society (Diamond, 1992), or to an overarching “masses” (thereby decreasing inequality, Boix and Stokes, 2003).

This breaks down, however, when the state is the primary recipient and distributor of increasing capital. If both capital and labor are dependent on the state, it is unlikely that their economic interests will bring them to become forces for democratization

(Bellin, 2000). As the statist literature has pointed out, states are not value-neutral arbiters of social conflict, as they have their own incentives, often consisting of the maintenance of the current rules of the game (Evans et. al., 1985; Knight, 1992). As such, economic growth associated with oil exports is qualitatively different from economic growth in economic sectors where ownership and employment are more diffusely distributed, as in agriculture and industry.

The increased importance of the government in allocating economic goods has a predictable impact on government and social preferences for different types of political institutions. For those who are the beneficiaries of government controlled economic resources, uncertainty in who will control the distribution of those resources is less tolerable. This is especially problematic for democratic forms of government, which inherently involve some level of uncertainty over the outcomes of elections (Przeworski et. al., 2000). Thus, we would expect democratic outcomes to be less likely where the government plays a key discretionary role in the distribution of economic goods.

While the decreased tolerance for democratic uncertainty goes a long way towards 15 explaining why non-democratic outcomes are more likely in fuel export dependent states, it does not explain the very different political outcomes amongst fuel exporting states.

Here another aspect of the statism literature becomes important. Based on the desire for governments to maintain stability, their ability or inability to do so hinges on the resources available to the state (Skocpol, 1979; Bueno de Mesquita et. al., 2002). Where the state does not have the resources necessary to maintain stability through the means discussed above, or when the benefits for staying in power are not high enough, the current group in power has an incentive to either steal from the state and abandon the political scene or to offer some level of institutional reform, such as limited democratization, to sate the opposition.

16 Fuel Export Income

Accrues to

Government

When dependent Political instability (esp. in Low democracies) Increases role of government resource decisions in individual income. income

Importance in individual income Political stability (esp. in authoritarian states)

High Decreases preference for resource democratic uncertainty income

Figure 1.1: Overview of the theory.

In developing countries, this suggests a gap between those who are dependent and

rich from their fuel exports and those who are just dependent. Karl (1997, p. 17-18)

notes this difference between what she labels “capital-deficient” and “capital-surplus” countries. The former have larger populations and more diverse economies, and, because of this, tend to not only absorb all their oil revenues, but continue to be net importers of capital. Capital-surplus countries, on the other hand, cannot absorb all their revenues, and continue to run balance-of-payment surpluses, even during bust periods. As Karl

(1997, p.18) put it, “[A]lthough all oil-exporting developing countries are highly dependent on petroleum, this dependence is felt more acutely in capital-deficient countries because their opportunities are so clearly bounded.” This distinction is critical to understanding patterns of development among fuel exporters, but it has been largely

17 ignored by scholars. None of the large-N studies cited in this study mention Karl’s scope

conditions, and even Karl (1997) dedicates less than two pages to its discussion.6

Based on the level of dependence and income, the predictions from the theory laid out above are summarized in Table 1.2. Those countries receiving a large amount of income per-capita from their fuel exports and who are highly dependent on those exports are expected to be stable and authoritarian. In these countries, capital garnered from fuel exports should be high enough for the state to offset short-term liquidity and financial crises. At the same time, dependence on fuel exports, and the associated high importance of government discretion over the distribution of economic goods, decreases the preference for democratic uncertainty. Indeed, these states are often net capital exporters during times of high fuel prices, providing a pool of resources from which they can draw during price downturns. As noted in Table 1.2, this description correctly classifies many of the states associated with the rentier-states literature, including Saudi Arabia, U.A.E., and Kuwait, who have all remained stable under traditional authoritarian regimes during the period studied here.

Countries that are highly dependent on fuel exports, but do not receive a large amount of income per-capita from those exports, are what I label “vulnerable.” Such states will not always be unstable. Indeed, during times of increasing oil prices, these states may be quite stable (Karl, 1997). They are not likely, however, to have the resources available for survival during times of economic crisis, nor are they likely to be capable of distributing resources widely enough to offset other social divisions. Thus, it is in these states that the pattern of cyclical political instability and social conflict,

6 Only Ross’ (1999) review article of Karl’s makes mention of her scope conditions, but, even here, the primary focus is on her decision to only look at countries that make up a large proportion of global exports. 18 including civil war, would be most likely. In Table 1.2, Nigeria, Congo-Brazzaville,

Algeria and Angola, all of which have experienced several changes in political regime- type and substantial civil conflict based on their resource endowments (see Ross, 2003, p.21) are correctly classified. The instability predicted for countries in this category should be especially severe for democratic forms of government. As the role of the government in determining the distribution of economic goods is increased, but the resources for distribution are relatively small, there is an incentive for all groups to lock in their share of relatively scarce resources through exclusive governance.

Low Income (< $1,000) High Income (> $1,000)

Moderate Dependence Limited Resource Impact: Stable Democracy: (< 10% of GDP) Indonesia Norway Ecuador Malaysia

High Dependence Vulnerable Regimes: Stable Authoritarian: (> 10% of GDP) Nigeria Saudi Arabia Azerbaijan Oman Congo Brazzaville Kuwait Yemen Trinidad Algeria Gabon Iran UAE Venezuela Turkmenistan Syria Russia Kazakhstan

Table 1.2: A heuristic layout of the predictions of the theory.

19 For countries only moderately dependent on fuel exports, it is likely that revenues play an important role in politics, but it is not likely that they are the determining factor in determining the political regime. Norway is the only country to fall into the category of moderate dependence and high income from fuel exports. If the criterion for moderate dependence is relaxed, Canada and Great Britain would also fall into this category of countries which receive substantial income from fuel exports, but are not very dependent on those resources during the period of 1965 to 2001. Politics in these countries have certainly been influenced by oil developments. In Canada, furious debates are taking place about the proper distribution of royalties from Albertan oil sands (see e.g. CBC

News, 2007). Similarly, in Great Britain, the discovery of oil has been used by nationalists in Scotland to justify independence (Lee, 1976; Farbey et. al., 1980). For all of this, however, these countries have relatively diversified economies, negating the monopoly of government decision-making over individual income. Fragmented economic interests give justification to democratic uncertainty.

Those countries with relatively low income from fuel exports, and only moderate dependence are expected to behave in a manner similar to other developing states. The reasoning behind this is similar to the situation in oil exporting industrialized states.

While fuel exports may provide the resources for some programs, they are not likely to be enough to weather economic crises, nor is the economy so dependent on those resources that competing interests are not present. The countries which fall into this category are

Indonesia, Ecuador and Malaysia. If the threshold of moderate dependence is relaxed,

Mexico also falls into this category. Of these, Indonesia and Mexico have received the most attention from scholars of oil resources. Some scholars attribute Indonesia’s oil resources with the investment capabilities of the Suharto regime from 1967 to 1998, and 20 suggest these resources provided stability at some key moments (Robinson, 1988; Liddle,

1991). Yet, Indonesia’s change in government in 1998, brought about by a liquidity crisis, was not prevented by surplus oil revenues, nor was the outcome of the transition

consistent with the aversion to democratic uncertainty associated with highly fuel export

dependent states. Mexico has a similar story. In Mexico, a large portion of government

revenues has come from the state-owned company PEMEX, which now stands as one of

the most indebted and ill-managed companies in the world (Marcus, forthcoming).

During the 1970s, revenues from fuel exports played a key role in increasing social

spending, arguably stabilizing Mexico’s non-democratic government (Morrison, 2007).

However, as oil decreased in importance, so did the government’s role in the economy.

This is important for Mexico’s democratic development, as state dependence is a key in

explaining both capital and labor’s support for democracy (see Bellin, 2000). All of this

suggests that when a country is or becomes only moderately dependent on oil there is

enough diversity in political interests that oil is not the determinant of regime outcomes.

It is important to note that this does not mean that a decrease in dependence makes

countries more likely to democratize, only that other variables will be more important in

determining regime outcomes.

While these classifications are made primarily for heuristic purposes, the

conclusions from the theory should now be clear. Dependence on fuel exports makes

democratic institutions less likely, since the uncertainty associated with democratic

competition is less tolerable where government decision-making plays a large role in

determining individual incomes. When dependence is not accompanied with high export

income, the result is more vulnerable political institutions, especially in democracies.

Group competition for relatively scarce resources is likely to drive politics in these 21 countries. When dependence is accompanied by high export income, the result should be more political stability, especially for authoritarian regimes. In these countries, the increased importance of controlling the government is offset by the government’s ability to make widespread distributions and its access to capital to offset short-term financial crises.

1.5 Socioeconomic development

If the criticism of modernization theory from the oil resource literature is that the focus on social groups ignores the resources given to the government from development, and especially from the exploitation of oil reserves, than the criticism of the natural resource curse literature is that it does not explicitly explore the implications of the theory for socio-economic development and popular expectations. Indeed, there are important reasons for thinking that oil-based economic development will not cause a proportionate increase in socio-economic wellbeing. Thus, oil development may not be as large an exception to traditional democratization theory as it is usually portrayed.

The normal logic of economic growth in social development is that growth is like a rising tide, it lifts all boats. As national production increases, so should the resources available for health care, education, and poverty reduction. In addition, growing production should generally be reflected in growing access to employment, reducing poverty and wage disparity. There is nothing inherent in modernization theory, however, to suggest that these factors will always correspond. To the extent that economic growth does not result in proportionate increases in socio-economic development and rising social expectations, some modernization authors would contend that political change is also unlikely (Deutsch, 1991; Lerner, 1958; Diamond, 1992). Better health, higher 22 education, and lower poverty are all among the areas of broader socio-economic

development that change popular expectations of social involvement in government, and supply both a tolerance and a push for democratization. There are several reasons to

think that the link between economic growth and socio-economic development may be

weaker in oil rich states.

As has been shown in numerous studies, the extent to which growth and socio-

economic development are linked depends largely on political institutions and decision-

making. Education, health care, and poverty reduction usually take on the form of public

goods. These services are not easily divisible between individuals, governments are not

normally able to exclude people from these programs, and individuals can benefit even if

they have done nothing in exchange for those benefits. In comparison, private goods are

those that accrue to the individual, and which individuals can be excluded from if they do not provide some type of payment. As noted above, governments worried about stability will generally prefer to distribute revenues as private goods, since they can maintain the punishment mechanism of withdrawing the private good should an individual decide to defect from the ruling coalition. That many governments decide to spend, often a great deal, on public goods is a reflection of the difficulty in allocating private goods, either because it is perceived as corruption or because there are simply too many people who have demands on the government’s resources (Barr and Serra, 2006; Bueno de Mesquita et. al., 2003; Lake and Baum, 2001).

In oil exporting states, the capacity of the government to distribute private goods is enhanced, since the government is the primary recipient of oil revenues and since the demands placed on government resources from the oil industry are likely to be relatively light. 23 In addition, the government faces a challenge in dealing with severe oil price fluctuations. Some areas of socio-economic development lag behind investment in public goods. In education, for example, teachers must be trained, students must go through the system, etc. Therefore, during times of exploding revenues from rising fuel prices, these indices will likely lag well behind indicators of per-capita GDP. This combines with the problems posed when fuel prices decrease. Spending on private goods and patrimony is difficult to undo. Cuts in these areas are focused, and those hurt are usually quick to organize. This means that spending on socio-economic development is likely to be cut or shifted to private parties when difficult decisions must be made. The punchline is that oil price fluctuations are likely to undermine long-term investments in human capital. The result of this is that we would expect for growth achieved through oil development not to translate into the same scale of social gains in public concerns as we would expect in non-fuel-exporting states.

This undermines the exceptionalism of fuel exporting states, as modernization theory and other ideas of economic development usually assume that economic growth results in other social conditions which make democracy more likely (Diamond, 1992).

The contention also emphasizes the difficulties of development when economic power becomes concentrated into government hands, where the maintenance of power may change incentives for investment.

1.6 Outline for the study

The rest of this study will be laid out as follows. Chapter 2 will delve further into the foundations of the “natural resource curse.” Oil and other natural resources have not always been considered a hazard for economic and democratic development. Scholars 24 have reached varying conclusions on the effects of oil resources depending on the time

and place being analyzed. This chapter looks at the evolution of the literature on natural

resource politics over time. It reconciles the literature on early industrialization, where

natural resources are considered beneficial to political development, with the current

literature on the “natural resource curse.” This is done by highlighting the changes in the

international oil market that have negated the price advantages enjoyed by early

industrializers. It also covers the empirical outlines of the negative relationship between

fuel exports and democracy, utilizing a dataset covering 166 countries from 1965-2001.

This analysis demonstrates the negative impact of both fuel export income and

dependence on the level and probability of democratic governance.

Chapter 3 demonstrates how the development of generalizable theories to explain

the negative relationship between fuel exports and democracy closely mimics the

regional work in Chapter 2. As a result, previous studies have focused on fuel exports as

either a source of funds for political stability or a source of competition resulting in

political instability. Contrary to what previous theories would lead us to expect, I show

that there are a wide variety of regime outcomes in fuel exporting states. This motivates

a formal model based on the increased importance of government discretion in economic

distribution. This model predicts that regime outcomes are dependent on three factors:

(1) the level of dependence on fuel exports; (2) the amount of revenue garnered from

those exports; and (3) the demands placed on the government.

The formal model from Chapter 3 has several empirical implications that are

tested in Chapter 4. First, the effect of fuel exports on the importance of government distribution is tested. Here it is found that fuel exports are tied with increased government discretion in the distribution of economic resources. This confirms the basic 25 assumption of the formal model. Second, the formal model leads to the prediction that,

all else being equal, higher fuel export income should increase the lifespan of political

regimes, while dependence on fuel resources when incomes are not as high will have the

opposite effect. Cox proportional hazard models find that this is indeed the case. In

addition, they support the contention that fuel export income is especially stabilizing for

authoritarian regimes, while fuel export dependence is especially destabilizing for

democratic regimes. The chapter ends by looking at the relationship between fuel exports

and civil war. Here again, it is found that fuel export income is stabilizing, in that it

decreases the probability of civil war, while fuel export dependence is destabilizing.

Chapter 5 expounds on the relationship between fuel exports and socio-economic

development. Because fuel export revenues increase government discretion in economic

distribution, allowing the government to distribute private goods in exchange for political

loyalty, and the turbulence in oil prices underlines long-term social investments it is

expected that they will not be as good at providing public goods. Thus, even though fuel

exporters have higher per-capita GDPs than their developing colleagues, they under-

perform in a number of other, more proximate, measures of socio-economic development, such as the effective provision of health care and education or the alleviation of poverty. This calls into question whether social development in fuel exporting states is really as much of an exception to classical theories of democratization as is usually claimed.

While all of this demonstrates the theoretical consistency and general empirical support, it does not directly confirm whether the mechanisms identified above are really those that are most important in explaining political outcomes. Nor does it demonstrate that fuel exporting countries are a substantial exception to classic understandings of 26 democratization. Chapter 6 looks in-depth at the case of Kazakhstan. Kazakhstan

provides a rare glimpse into the way increased fuel exports shape politics. During this

period of rising prices, the government of Kazakhstan has been able to establish an

intricate system of government influence in the economy. The methods of this influence

range from the heavy handed (e.g. confiscating and selling the property of individuals who contribute to opposition parties) to the more indirect (building a new capital in the

North which provides construction jobs in one of the most unstable areas of Kazakhstan).

The end result is a system in which one person, Nursultan Nazerbayev, holds very broad powers with little dissention. Without the growing income from fuel exports, Kazakhstan would look much more like neighboring Kyrgyzstan, which has seen several major alternations in its government. The experience of Kazakhstan is also compared with that of Moldova, where the state’s lack of sources of capital has resulted in a much more competitive (if not entirely democratic) political system.

Finally, Chapter 7 concludes the study with a summary of the findings and a series of policy recommendations. Since the increased discretion of the government in economic distribution in fuel dependent states is responsible for both the stability of authoritarian regimes and the instability of democratic regimes, several policy recommendations are made. First, to the extent possible, transparency of revenue flows and regulation needs to be promoted in both the fuel and non-fuel sectors. Some efforts have already been made in this regard, but they need to be enlarged to include more than oil exports. Second, in states that are fuel dependent, but not fuel rich, stability can be improved through mechanisms that limit the losses of those outside of power. These mechanisms can include elite pacts or shared sovereignty (Krasner, 2005). Both of these, however, will be difficult to implement, as competing sides have incentive to defect, and 27 are most likely to occur after a serious shock, or where there is strong public pressure.

28 CHAPTER 2

DOES “THE CURSE” EXIST?

In the last decade, there has been an explosion of literature in the social sciences dealing with the various aspects of the “natural resource curse” -- the tendency for states dependent on fuel resources (also known as “point-source resources”) to have slower rates of growth (Sachs and Warner, 1995; Collier and Gunning, 1999), experience greater instability (Collier and Hoeffler, 1998; Bannon and Collier, 2003), and be less likely to be democratic than their non-fuel-exporting counterparts (Ross, 2001). The counter- intuitive conclusion of this work has been that states endowed with large natural resource wealth may be worse off than their neighbors who are resource poor. Within political science, the political aspect of the natural resource curse -- the tendency for countries that are dependent on fuel exports and/or garner a high level of per-capita wealth from these exports to be less democratic than we would otherwise expect, given their level of economic wealth -- has become conventional wisdom. Many authors who explore global patterns of democratization have either utilized dummy variables for oil exporting countries (e.g. Barro, 1999) or left the oil exporting states out of their samples altogether

(Przeworski et. al., 2000; Boix and Stokes, 2003).

The results of this conventional wisdom have been manifold. In Chad, the World

Bank has tried setting up special programs to make sure that the wealth stemming from oil development is used for poverty reduction and not to strengthen the authoritarian

27 regime (Tsalik, 2002; Zssis, 2006; Krasner, 2005). In Washington, US policy-makers have started pressuring oil companies to establish programs encouraging democratic governance (Ottaway, 2001). The Soros Institute has spent millions of dollars on programs to promote revenue transparency and good stewardship of oil revenues in

Central Asia.7 There have even been some who have raised concern about the prospects for democratic development in Iraq, based on its dependence on oil exports (e.g. Human

Rights Watch, 2003).

Such discussions are not totally new. Adam Smith, in The Wealth of Nations, stated that “[mining projects] are the projects therefore, to which of all others a prudent law-giver, who desires to increase the capitol of his nation, would least choose to give extraordinary encouragement” (1937[1776], p. 562). While Smith and others may have been dubious about the utility of natural resources in promoting economic development, it would be a mistake to assume that scholars have always viewed natural resources in the same manner. Indeed, the literature on fuel exporting and extraction can be divided into two distinct time periods, pre- and post-1970, with scholars drawing very different conclusions in these two periods. Some scholars still contend that the natural resource curse does not really exist, that many states have developed economically and politically with high levels of fuel resources, or that the natural resource curse phenomenon is simply an extension of dependency theory.

The goal of this chapter is to introduce the reader to the development of the literature on natural resources and governance, to demonstrate the existence of the political aspect of the natural resource curse, and to illustrate the phenomenon that will be the subject of this study. It will start with an overview of the literature on the natural

7For information on these projects, see www.revenuewatch.org. Currently operations are underway in Kazakhstan, Iraq and Azerbaijan. 28 resource curse. Here, it is demonstrated that hypotheses developed about the natural resource curse are largely influenced by the time and place that the researcher has studied. Given the differences in the literature over space and time, it is difficult to draw out a conclusive causal pathway for this relationship. Second, a short section will be devoted to explaining why the literature diverges so much between those who study oil developments in the 19th century and those who study development in the 20th and 21st centuries. It is suggested that there has been a fundamental shift in the global market for fuel exports, both in who is exporting and in their access to outside markets. These developments are especially problematic for those who use the lessons of development from before the 1970s oil boom, or even from the industrial developments of the 19th century, to argue about modern resource development. Third, this chapter replicates the findings of the political aspect of the natural resource curse. This is done to show the robustness of these findings, to clarify some of the issues surrounding this correlation, and to introduce the data and methods used throughout this dissertation. Finally, some of the main findings and patterns of the resource curse will be discussed before moving on to explanations of the curse in the next chapter.

2.1 Relationship between fuel and democracy, pre-1970

The presence of natural resources has not always been considered a curse. In fact, the presence of natural resources is considered a key element in economic and political development of two of the earliest democracies in modern politics: Britain and the United

States.

29 In Britain, the industrial revolution, and the social and political changes that

accompanied it, was driven by the exploitation of domestic coal reserves. In general, the

industrial revolution had two phases. The first involved the concentration of agriculture,

and the migration to cities, accompanying the enclosures movement. The second

involved moving away from an economy based on organic energy, wood and animal

power, to a mineral-based economy. It is in this latter stage that Britain's coal reserves

became an advantage. By way of comparison, the Dutch Republic urbanized earlier and

had better routes of transportation, but its economy, based on rapidly depleted peat, did not undergo the second phase of industrialization (Wrigley, 1988). The rise of industry, trade, and a bourgeoisie class, in turn, contributed to Britain's eventual democratization

(Moore, 1971).

While the United States had elections long before it became a major manufacturing power, the current economic status of the US was built on a base of intensive domestic natural resource exploitation. During the period between 1890 and

1910 the United States began exploiting its natural resources at an unprecedented rate, which, in turn, led to an explosion in US manufacturing (Wright, 1990, p. 464-468;

David and Wright, 1997; Irwin, 2003). The abundance of resources in the United States, including coal, iron and oil, provided American industry with cheap inputs, increasing the competitiveness of manufacturing in ways equivalent to decades of technological and productivity advancement (Irwin, 2003).

A number of other democratic industrialized states were also endowed with large mineral and oil reserves. It is telling that Canada and Australia are usually criticized for

not taking greater advantage of their natural resources earlier in their history (Wright and

Czelusta, 2003). Natural resource wealth during the time period from the industrial 30 revolution until World War II were considered so important to a countries' economic wellbeing that the battle for them resulted in numerous wars and drove the process of colonization.8

Even today, the top countries in terms of proved oil reserves are a motley group.

As of 2004, the United States had 29.4 thousand million barrels of proved reserves of oil, almost twice as much as Mexico's 16.0 thousand million barrels or Qatar's 15.2 thousand million barrels. Even China, who is now driving up world demand for oil to feed its growing economy, has approximately 17.1 thousand million barrels in proved reserves

(BP, 2005, p. 4). The main difference between these countries is in terms of consumption. The US consumes about 21 million barrels of oil per day, compared to

Mexico's 1.9 million and Qatar's 77,000. China has seen a boom in its use, which, in

2004 reached 6.6 million barrels a day, 15 percent more than the year prior (BP, 2005, p.9).

Even some of the major exporters of oil are not considered in the curse literature, primarily because fuel exports do not make up a large proportion of their GDP or manufacturing exports. For example, Canada exports about 2 million barrels of oil per day to the United States, making it the largest single contributor to the US oil supply (BP,

2005, p. 17). From these exports, Canada, in 2001, earned approximately $1,200 in per- capita income. However, this only made up 5% of Canada's GDP. By contrast, Algeria made approximately $704 per-capita from its fuel exports in 2001, but these made up

44% of the country's GDP (World Bank, 2003).9

8 The incidents of expansion for resources are numerous. They include early colonization for spices and trade, to the Spanish conquests in the Americas for gold (see e.g. Karl, 1997). Japan’s territorial expansion and involvement in World War II is also often attributed to the desire for the government to acquire the natural resources for expanding its military-industrial base (see e.g. Yasuba, 1996). 9These numbers are derived from the World Development Indicators. Income from fuel exports is 31 Post World War II development scholars primarily drew upon the experiences of industrialized countries, and assumed that natural resources would help developing countries to build industry, increase the wellbeing of their citizens, and eventually lead to democratization. These hopes were frustrated during the 1970s, leading to two literatures, one in the general comparative politics literature and one in the area studies literature, which are the direct predecessors to studies of the natural resource curse.

2.2 Oil shocks and the rise of the curse

The late 1960s and early 1970s were characterized by an increasing disillusionment with modernization theory. Developing countries failed to become more industrialized and often backslid from democratization. The dependency literature that evolved from this frustration led directly to the natural resource curse literature. This body of work suggests that the timing of industrialization has a significant impact on the disparity between developed and developing countries. Early industrializers are part of the “core,” a group of industrialized and predominately democratic countries, who export manufactured goods and import primary goods. Countries trying to industrialize make up the “periphery.” These countries export primary goods and import manufactured goods.

The patterns of trade between the core and periphery have the effect of keeping the developing world in a permanent state of underdevelopment. As manufactured goods cost more than their inputs, primary commodity exporters have difficulty gaining capital for industrial investment. In addition, any industry in the developing world wanting to enter the world market must compete against established manufacturing firms have

calculated by taking fuel exports as a percent of merchandise exports and multiplying it by the total value of merchandise exports (in constant 1995 USD). To derive per-capita revenue, the resulting measure of fuel income is then divided by population. To calculate fuel exports as a percentage of GDP, fuel income is divided by the total GDP in that year (in constant 1995 USD). 32 experience, technology and capitol advantages (e.g. Wallerstein, 1974).

The difficulty in industrializing also has implications for democratization.

Dependency short-circuits the process of industrialization and the development of an

entrepreneurial class that led to democracy in developed states (Frank, 1967; Chalmers,

1969). The difficulties of industrialization may also require an authoritarian decision-

making process to push through difficult stabilization measures (O'Donnell, 1973).

These works highlighted the difficulties that developing states experienced in

growing through the importation of industrial technology. Put simply, developing states

did not have the capital resources with which to purchase advanced development

technologies. Since its inception, global events have caused dependency theory to

undergo many criticisms and revisions. Cardoso (1972) argued that developing states

remain disadvantaged by the global economic system and dependent on developed states,

but that increased investment in industry has led to some level of development.

Similarly, Evans (1979) highlights a complex relationship between the state, local elites

and investors, who may all share a common interest in the development of local industry.

Nevertheless, the ideas of dependency theory played an important role in the later development of the natural resource literature. At the same time as scholars of developing states were bemoaning the capital constraints on developing countries, there was one group of primary commodity exporters that were not short on capital. The oil price shocks of 1973 and 1979 brought a surge of capital into the oil exporting countries of the Organization of Petroleum Exporting Countries (OPEC): Algeria, Indonesia, Iran

Iraq, Kuwait, Libya, Nigeria, Qatar, Saudi Arabia, United Arab Emirates (UAE), and

Venezuela. In fact, the capital surge into OPEC countries was the largest non-war

redistribution of wealth in history (Amuzegar, 2001; Tsalik, 2002). This led to the 33 expectation that these oil exporters might have finally broken the cycle of dependency.

Isbister (1993, p. 168) describes the optimism surrounding these price increases:

[I]n the immediate wake of the OPEC action in the mid-1970s, there was jubilation. A group of Third World countries had finally succeeded in turning the tables on the rich countries. Rather than see their own countries sucked dry by low-priced exports that went to fuel western industrialization, the OPEC countries were now bleeding the developing countries, showing that their former masters were vulnerable, that their prosperity depended on the resources and the cooperation of the oil exporters.

Even those developing countries hurt in the short-term by the price increases were optimistic that the oil crisis and the increased wealth of OPEC countries would help break the cycle of dependence and raise support for a New International Economic Order

(NIEO).

The increased geo-political importance of OPEC countries was accompanied by an increased interest in their politics. However, contrary to what modernization scholars

would have predicted, and development scholars would have hoped, increased wealth

was not accompanied by democratization in most of these countries. Nor did the rapid

increase in capital resources result in the overcoming of the late industrialization problem, as dependency theorists would have predicted. Indeed, many of the OPEC countries became more authoritarian and/or more unstable after the price spikes and many saw a net decline in living standards. The unexpected outcomes among these countries led to several different foci within the literature on oil states.

2.2.1 Stability in the Middle East

Within the Middle East, the primary attribute of regimes was stability. Saudi

Arabia, Kuwait, Qatar and UAE all maintained traditional monarchies led by the same families as before the oil booms. Crystal (1989, p. 427) sums up the situation very well 34 when she states:

Oil revenues preserved continuity at the apex of the political system by forcing the breakdown of the old ruling coalition and catalyzing the formation of a new pattern of political control. Oil in fact induced important changes in the ruling coalition even as it preserved continuity at the very top of the political system.

Rather than undermining the legitimacy of traditional ruling groups, oil allowed these groups to form new coalitions, and, in some cases, actually decrease calls for political participation. This contradicted the predictions of “modernization theory,” which suggested that economic development should be accompanied by increasing calls for political involvement (Lipset, 1959; Lipset, 1960; Lerner, 1958; Huntington, 1968).

Iraq emerged from a long period of instability in the 1960s, into a stable pattern of one-party rule which later evolved into a personalistic dictatorship. In 1958, the Ba'ath party regained control over the country's politics, with former general Ahmad Hasan Al-

Bakr as the president and chairman of the Revolutionary Command Council (RCC). In

July 1979, Al-Bakr handed over his position to his chosen successor, Saddam Hussein.

Hussein then went about setting up a highly personalistic dictatorship which lasted until the United States invasion in 2003.

Iran was the exception to this pattern of stability. Indeed, it was Iran's instability that resulted in the oil shock of 1979. In 1953, Shah Mohammad Reza Pahlavi was restored as monarch of Iran. At first he ruled jointly with parliament, but he eventually managed to erode parliament's authority until he exercised effective control. Along with his increased power, the Shah also introduced measures for rapid modernization. These programs resulted in the displacement of a number of traditionally powerful groups, and, combined with a faltering economy, led to violent protests which forced the Shah to flee the country on January 16, 1979. He was supplanted by an alliance, led by Ayatollah

35 Ruhollah Khomeini, who set up a theocratic state, with an elected legislature regulated by the leader and a body of senior clerics called the Guardian Council.

Thus, the focus among area scholars in the Middle East was on explaining the

relative continuity of governance in rapidly changing societies. This is the source of the

literature usually cited as the beginning of writing on the detrimental relationship

between oil and democracy -- the rentier state. Mahdavy (1970) defines a rentier state as

one that receives substantial rents from foreign individuals, concerns or governments.

This idea was drawn from the literature on foreign aid and covers more than just oil

revenues. Beblawi (1987, p. 51) made the definition a little more specific -- a rentier

state is where rents are paid by foreign actors, where they accrue directly to the states and

where "only a few are engaged in the generation of this rent (wealth), the majority being

only involved in the distribution or utilization of it."

According to the rentier hypothesis, oil revenues give the state flexibility and

insulation from the pressures of society, separating the state from society in the same way that the oil industry has few links to the rest of the economy (Mahdavy, 1970). This political insulation comes about for several reasons. Money coming to the government

from foreign actors allows the government to raise revenue without taxation (Ross,

2001). Increased revenues allow the government to buy off potential competitors

(Huntington, 1991). Rentierism also allows the government to decrease the influence of

groups that could potentially counteract the power of the ruling elite (Crystal, 1989;

Crystal, 1990).

The greatest threat to stability in these oil exporting states came when, like Iran,

they attempted to radically modernize the state rather than buy off traditional groups.

Another threat to stability might arise if the government allows their bureaucratic system 36 to atrophy to the point that it cannot react when oil revenues decline (Chaudhry, 1997).

With these two exceptions, the expectation is that authoritarian governments in fuel dependent states should be more stable.

2.2.2 Military regimes in Libya and Indonesia

The literature on the military regimes in Libya and Indonesia is similar to the rentier state literature. Both of these countries experienced long periods of rule by military officers. In Indonesia, Major General Haji Mohammad Suharto remained in power until the 1998 Asian financial collapse forced competitive elections. In Libya,

Colonel Moammar Al Qadhafi took power in 1969 and is still ruling as of this writing.

For Suharto, oil revenues provided the means for remaining in power while implementing a state-centered modernization plan. In the mid-1950s, Indonesia had few hopeful prospects for economic development (see e.g. Fryer, 1954; Hawkins, 1955).

Since that time, President Suharto was able to implement far-reaching industrialization reforms, resulting in steady economic growth. The policy decisions of Suharto have been credited, by several scholars, to the autonomous ability of the government to make major economic decisions (Robinson, 1988; Liddle, 1991). In this case, the political insulation provided by oil revenues allowed for strategic state investment, liberalization, and modernization. However, this political insulation was not strong enough to withstand the backlash against the government resulting from the 1998 currency crisis.

In Qadhafi's case, the oil revenues provided stability as he pursued a socialist agenda. Both politically and economically, Libya, when it is treated at all, is usually

lumped with the Middle Eastern states reviewed above (Collier and Gunning, 1999).

Since 1969, Qadhafi has formally pursued a revolutionary agenda, but his rhetoric has not 37 been matched by profound administrative change. In this case, while not a traditional

monarchy, Libya's oil wealth helped the Qadhafi government maintain stability, despite

corrupt government officials, exorbitant sums of money being wasted by incompetent

bureaucrats, and a resource base inadequate to absorb large-scale capital investment

(Allan, 1981).

2.2.3 Instability in Africa

In Africa, the focus has been much more on how oil has fed instability and

corruption in the region. Nigeria entered this period in a state of civil war, followed by a

succession of military regimes. Competitive elections were held in 1979. In 1983,

however, the civilian government was overthrown and followed by another succession of

military rulers. On June 12, 1993, competitive elections were again held, only for the

President to be toppled by his defense minister in November of the same year.

Democracy was re-established in Nigeria in 1999, with the election of a former military head of state -- Olusegun Obasanjo. Nigeria has become a poster child for the natural resource curse, not only because of its governmental and ethnic instability, but also because of its lack of per-capita growth since the development of its oil reserves.

Algeria has also been relatively unstable, ruled by a series of formal dictators and military personnel since gaining independence from France in 1962. The 1980s saw increased tension between the ruling National Liberation Front (FLN), which had dominated politics since independence, and the Islamic Salvation Front (FIS). This tension resulted in the banning of the FIS from politics in March 1992 and an outright civil war that characterized much of the 1990s. Competitive presidential elections were re-established in 1995, but the FIS was not allowed to participate. Henry (2004) 38 characterizes Algeria as a “bunker state” in that it rules in large part through providing

security against the insecurity it creates.

While Nigeria and Algeria have garnered most of the attention because of their membership in OPEC, Angola, Sudan, Morocco and the Republic of Congo all underwent civil wars that were, in some way, related to their oil reserves (Ross, 2003, p.

18).

The literature on African oil politics has picked up on this pattern, focusing on

oil's role in promoting and sustaining conflict (Osaghae, 1995; Henry, 2004), increasing

corruption (Ross, 2003; Global Witness, 2002), and hindering economic growth

(Khennas, 1993; Khan, 1994).

2.2.4 Cycles of Venezuela

Finally, developments in Venezuela since 1970 have sparked a literature all their

own. Since the overthrow of General Marcos Perez Jimenez in 1958, Venezuela has had

an unbroken tradition of civilian democratic rule. Until the 1998 elections, politics was

dominated by the Social Democratic Accion Democratica (AD) and the Christian

Democratic Comite de Organizacion Politica Electoral Independiente (COPEI). These

two groups functioned under a system of “pacted democracy” whereby they worked

together to form a moderate political system that would keep the Communists and the military out of politics. During this time, political scientists praised Venezuela for avoiding the pitfalls of military dictatorships in Latin America during the 1960s and 70s and talked about “Venezuelan exceptionalism” (see Hellinger, 2000). The maintenance of this system was aided by oil revenues which allowed both parties access to patronage

(Karl, 1986; Levine, 1973). 39 This began to change during the 1980s and 1990s. Years of demagoguery on the

part of political leaders and the decrease in oil revenues resulted in increased frustration

with democracy. With the oil spikes in the 1970s, President Carlos Andres Perez

attempted a radical project of modernization, which led to an economic crisis when oil

prices collapsed in the 1980s. This resulted in urban riots in 1989 and two coup attempts

in 1992. Scholarly discussion turned to the fragility of democracy in Venezuela. Articles

and books appeared with titles like “Re-arranging the Deck Chairs on the Titanic,”

Democracy for the Privileged, or The Paradox of Plenty (Romero, 1997; Hillman, 1994;

Karl, 1997).

In the December 1998 elections, Hugo Chavez, a former lieutenant-colonel who

led an unsuccessful coup attempt in 1992, was elected on a populist platform which

called for a “Fifth Republic,” involving a new constitution and a new relationship between social classes. Since then, Chavez’s consolidation of power and his intolerance

for opposition have led many to label his regime “semi-authoritarian” (see e.g. Ottaway,

2003). Reflecting this, the Freedom House classification of governance in Venezuela has changed from “free” in 1998 to “partially free” in 2006 (Freedom House, 2006).

From the perspective of writings on the oil/democracy relationship, Venezuela is now considered a primary example of the stress that boom/bust cycles of the oil economy

place on a country's stability (Karl, 1997). In this line of thought, fuel export

dependence results in hollow democratic institutions with little inherent support or legitimacy, vulnerable to populism, and fragile when faced with a downturn in oil resources (Karl, 1986, Romero, 1997).

40 2.2.5 Bringing “the curse” into a cross-national perspective

Even before the evidence from these area studies became a full-fledged cross-

national research agenda, scholars began controlling for the effect of oil in their

regression models (e.g. Barro, 1999; Przeworski et. al., 2000). The turning point came

with Ross' (2001) article, in which he tests hypotheses about fuel and mineral dependence

on a sample of 113 states from 1971-1997. Since that time, a number of publications,

working papers, and dissertations have attempted to develop an understanding of the

oil/democracy relationship which generalizes beyond one particular region. As will be

shown later, these attempts at generalization have largely been unsuccessful because they

rely on hypotheses drawn from particular parts of the regional literature, rather than on

hypotheses drawn from the attributes of fuel exports themselves.

2.3 What changed?

Some argue that the change in historical perspective on oil and democracy is

evidence that the current literature on the natural resource curse is an aberration, a

historical hiccup that will evaporate with time (e.g. Wright and Czelusta, 2003). These

authors, primarily economic historians, argue successful resource-driven industrialization

in the US, Canada and other states prove that fuel exploitation can be utilized as a

knowledge-based and growth-driving resource.

Is it possible that the entire literature on the natural resource curse is simply a

reflection of bad policy or of the unique situation brought on by the 1970 oil shocks? I

contend that this is not the case -- there has been a substantial shift in both the global oil

market and in the countries involved in the oil game. These changes make it unlikely that

we will see a return to the developmental character of natural resources in the near future. 41 First, the oil market has become much more globalized since the time of US and

British industrialization. Oil and gas prices are a good example of this. Gas prices in

Kazakhstan, a major oil exporter, are only a few cents cheaper than in the United States.10

This is not because Kazakhstan lacks the oil resources to drive the price down, but

because, like any oil-producing state, the government faces a dilemma between keeping

the price of oil products cheaper at home or bringing more money into the country by

selling it at a higher price abroad.

Oil and natural gas do not flow seamlessly across the world to the places willing

to pay the highest price, but vast improvements in the transportation of fuels and the

invention of trading tools called swaps -- where a relatively isolated country like

Kazakhstan sends oil to a more accessible oil exporting country, like Iran, for an

equivalent amount of oil to be sold abroad -- have made it much more tempting and economically viable for oil exporters to sell their product abroad than to use it at home.

This is something that is lost on many international policy-makers in debates

about the invasion of Iraq or opening up wildlife refuges to drilling. Today, unless a

project opens up a large enough reserve to increase global supply significantly, it is

unlikely to significantly affect the price of oil and gas in a particular country without the

addition of some type of subsidization. So, for example, if the United States were to

open the Anwar National Wildlife Reserve to drilling, and the resulting oil caused the

price that the US paid for crude to drop, Canada, Saudi Arabia, and other suppliers would

have an incentive to sell their products to another country that is willing to pay more for

their oil. Canada, for example, is currently considering opening up a pipeline to the

Pacific in order to gain access to the growing Chinese oil market (see e.g. Oxford

10 When I arrived in Almaty, Kazakhstan, in September 2005, I was surprised to find that gas prices were just a few cents cheaper than in my hometown of Kansas City, MO. 42 Analytica, 2005). There are many other arguments made by those who support the opening of new fields, but oil price is rarely a strong justification.

Since transportation costs are no longer as important of a factor in raw material prices, the dramatic advantages the US and Britain utilized for industrialization may not be available to the developing world. Take, for example, Wright and Czelusta's (2003, p.

13) description of how mineral discoveries changed the California market, “Spurred not just by jobs in oil but also by the dramatic fall in the cost of energy, California's share of national income doubled; contrary to Dutch disease models, the size of the state's manufacturing sector quadrupled.” Or look at how Irwin (2003, p. 368-369) describes the effect of the Messabi iron ore range on manufacturing in steel. Here he contends that the cheap access to raw materials was equivalent to about a decade's worth of technology and productivity improvement in terms of the competitiveness of US steel. In his words,

“[T]he non-tradability of iron ore resources helped to ‘crowd in’ manufactured exports, thereby avoiding some of the adverse consequences of resource abundance in the US case” (p. 364).

Absent government subsidization or legal requirements for companies to provide oil to a particular sector, it is difficult to see a country reaping the same direct benefits from oil and mineral discoveries today that were possible then. However, subsidization of natural resources, or placing requirements on companies for oil resources to be used in domestic manufacture, carries risks for producing countries as well. First, they forgo potential revenues from higher prices abroad. In some cases, selling the oil abroad and then giving the money to the industry in the form of a direct subsidy would be a cheaper alternative. In addition, it negates the incentive for the industry to become efficient in its fuel use, often delaying the adoption of more efficient technologies (Ascher, 1999). 43 When subsidies are more general, such as subsidizing the consumer price of

gasoline, it can create a political problem as well. Iran is an example of this. While gas prices have topped $3 a gallon in the US, and are even higher in Europe, the price of a gallon of gas in Tehran is about $0.40 a gallon. The result is that Iranians utilize private vehicles much more than comparable countries, they do not use fuel efficient cars, and they practice such wasteful habits as filling up tanks until gasoline spills out. The problem of waste has become so bad that the Iranian Oil Ministry has responded with billboards saying, “Oil resources are exhaustible. Let's consume correctly with a view to the future” and calling overuse of gas “A waste of national wealth” (Vick, 2005, p. A12).

Even the domestic popularity of Iran's nuclear program can be better understood in light of the extra income Iran could earn by selling its oil abroad, rather than using it domestically. However, attempts to curb gasoline use by raising the price are extremely unpopular, as demonstrated by the protests that followed the decision to decrease subsidies in 2007. Also, the Iranian government takes great pains to make sure that prices stay low during local and national election cycles, which, in Iran, come very frequently.

Aside from the globalization of the oil market, there has been a shift in the attributes of the countries involved in the oil trade. In much the same way as new technologies have made it possible to drill for oil in areas that were once impossible, the size and diversification of the largest oil companies make it possible for them to do business in places with undeveloped infrastructure, authoritarian governance, and outright instability.

In 2005, Exxon/Mobil had a profit of $36.13 billion and revenue of $471 billion

(Romero and Holusha, 2006). With these resources, they are able to take risks in 44 countries where the political and economic situation is less hospitable. Higher oil prices

also mean that investments in risky environments pay higher dividends, and projects that

were once considered unprofitable because of the risks involved are now profitable

(Isbister, 1993). In addition, dwindling reserves in the developed world increase incentives to move exploration into the developing world.

Chad and Sudan are excellent examples of both these trends in oil development.

Neither of these countries have the domestic resources for the effective exploration and

exportation of their oil reserves. Both countries have authoritarian governments, active

civil unrest, a non-existent industrial sector, and high illiteracy rates. In other words,

neither country is in a position to exploit its resources for domestic industrialization and,

through this, the development of a domestic labor and capital class. Chad and Sudan in

2006 are not the US in 1890 or even Britain in 1760. Coronil (1997, p.7) sums it up very

well: “The Dutch disease...only occasionally afflicts the resilient, diversified economies

of first-world nations, but constitutes an epidemic in the mono-culture economies of the

third world. As a colonial plague that malformed the third world into narrowly

specialized primary products exporters, the Dutch disease should be renamed the third

world or neo-colonial disease.”

It should be clear that the changes in our understanding of the oil/democracy

relationship pre- and post-1970 are not simply due to historical anomaly or poor policy

choice. The global oil market, the companies that operate in this market and the

countries that now make up most of it are all different from before the oil shocks. This is

not to say that 1970 is a hard cutoff date for these developments, as these trends have

been taking place since at least the 1930s and decolonization. They have simply become

more apparent since the 1970s oil crises. Most importantly, a number of changes in oil 45 markets militate against dismissing the negative oil/democracy relationship based on the

example of early industrializers.

2.4 The empirical relationship between fuel dependence and democracy

This section introduces the empirical regularity making up the topic of this work.

While discussion of the natural resource curse is not new, large-N analysis has only come about relatively recently, with Ross’ (2001) World Politics article as the touchpoint. The dataset used for most of this analysis will be introduced in this section, and the robustness of the negative relationship between fuel export dependence and democracy will be demonstrated.

Throughout most of this study, a dataset covering 168 countries from 1965-2001 will be utilized. This dataset combines variables from many different data sources, including Polity IV, Freedom House, Alvarez et. al. (ACLP), World Bank, Correlates of

War (CoW), etc. Because not every country was independent in 1965, some will enter the dataset at later dates. This is noted in the list of cases in Appendix A. Also, not all the data utilized in this work will cover all the countries and/or all the years in this dataset. This is regrettable, but unavoidable. The dataset will continue to expand and improve as more data becomes available. All the data is available for replication and further research use in the research section of the author’s homepage.11

Because this project utilizes data across both time and space, usually referred to as

panel, pooled, or time-series cross-sectional (TSCS) data, some special statistical tools

were used. A more detailed account of these methods, as well as a discussion of why

they were chosen, can be found by the interested reader in Appendix B. For those readers

11 http://polisci.osu.edu/grads/kennedy. 46 who are not statistically inclined, the next five paragraphs will be of little interest, and

can be skipped without losing the substantive story.

Scholars widely accept that TSCS data poses a number of problems, not all of

which have completely satisfactory solutions. As Stimson (1985, p. 916) states, “As

students of research design, we must appreciate pooled designs. But as statisticians we

are less enthusiastic. For pooled analyses, insofar as they are known at all, are known for special statistical problems." For most of this work, models are estimated using random

effects estimation, which models heteroscedasticity and autocorrelation problems as

random variables. This method was chosen because of its increased efficiency over fixed

effects models, and for its allowance of variables which remain constant within countries

over time (e.g. location and average levels), slowly move over time (e.g. institutions and

some important socio-economic indicators), and those in which the within and between

effects of the variable differ in predictable manners (labeled here as “complex movers”).

Several of the important variables in this work clearly fall into the categories in which

fixed effects estimation has difficulties. For example, the fuel dependence variable

described below has a ratio of between country to within country variation of 4.14, well

above the level of 2.8 at which the fixed effects estimator begins having severe problems

(Plümper and Troeger, 2007, p.136). Similarly, the behavior of the fuel export income

variable suggests that predictable indirect revenue mechanisms are mediating its impact.

As Plümper and Troeger (2007, p.131) demonstrate, the random effects estimator

is likely to be much more accurate than fixed effects or Hausman-Taylor at estimating the

effect of stable and slow-moving variables. This is because the inefficiency involved in

giving up information on differences between countries makes it impossible to accurately

test their effects. As Beck (2001, p. 285) correctly points out, “Although we can estimate 47 [a model] with slowly changing independent variables, the fixed effect will soak up most of the explanatory power of these slowly changing variables. Thus, if a

variable…changes over time, but slowly, the fixed effects will make it hard for such variables to appear either substantively or statistically significant.”

For some researchers, the use of stable or slow-moving variables is a reason not to

use TSCS data in the first place. Wilson and Butler (2007, p. 108) essentially accuse

researchers of artificially buttressing the statistical significance of their variables by

increasing the number of data points when the variables themselves do not change.12

Moving away from a TSCS framework simply because of the use of stable and slow-

moving variables does not make sense in this case for two reasons. First, while the

independent variables might be stable, the dependent variable is not. Thus, taking any

particular cross-section will involve losing information. Second, not all of the

independent variables are non-dynamic in their effects. Rejecting the TSCS framework

also means losing some of the flexibility to specify dynamic versus average and fixed

effects.

Because random effects has been shown to bias coefficients in faster moving

variables, tests were run to make sure that this decision did not result in bias in the

coefficients and a sampling of these tests are reported in Appendix B.

In spite of the random effects correction, there may still be problems in

calculating standard errors, since countries are not randomly selected (Beck and Katz,

1996; Beck 2001). To correct for this, in most of the models, panel-corrections were

made to the standard errors. The exception to this is where the dependent variable is

12 It should be noted that most of these results were also tested in a cross-sectional frame, using country averages, and this produced similarly significant results (for a similar application, see Sachs and Warner, 1995). 48 bivariate, and, in this case, there was not an available program for the panel correction. It

is unlikely that this made a significant difference in the hypothesis tests. It should also be

noted that some cues were taken from Beck's (2001) suggestion that country variables should be explicitly modeled in comparative politics. Where a location or country variable is especially important, it is included in the reported results.

One last note needs to be made about some of the main variables of interest in this study. In certain cases, variables that do change over time will be reduced to their mean value over the period being studied. This is done when it is found that the variable has

the attributes of what I call complex-movement, or when the variable has uncertain, but

readily apparent, time lags. Complex-movement occurs when a variable is indexed

against another correlated value. For example, to measure fuel export dependence, fuel

exports are usually indexed against either GDP or merchandise exports. There are two

reasons why this value may change: the value of fuel exports may increase, or the level of

GDP or merchandise exports may decrease. Because of the multiplication factor of

capital in the economy and because of the linkages between the fuel sector and other sectors, it is likely that a large number of increases in fuel dependency will be due to

economic crises as opposed to new finds or higher prices. As such, leaving the variable

in its untransformed state will lead to a mistaken conclusion since the effects of variable

movement within a country will produce opposite patterns from comparisons between countries, resulting in higher standard errors and a failure to reject the null hypothesis.

From a substantive angle, the relationship between dependency and crisis will likely produce very misleading results, since authoritarian regimes are particularly vulnerable to economic crises (Geddes, 1999). For fuel income, the major problem is that countries have varying capacities to smooth revenue from export income over time. This means 49 that the effect of an increase or decrease in income within a state is state is unlikely to have a proportionately sized impact on the intermediary variables involved in the theory developed in this study. Again, the likely result is an artificial increase in the uncertainty of estimates.13 In Appendix B, there is a short section demonstrating this in the case of fuel dependency and fuel export income. This pattern is also apparent when the variables are decomposed into their between and within effects (see e.g. STATA, 2005, p. 306-

308).14

Table 2.1 and Table 2.2 present the primary results on which this study builds.

Table 2.1 examines the full sample of countries, and the effect of several variables on the three most common measures of democracy. Polity measures the democratic and authoritarian aspects of institutions on a scale of -10 to 10, with 10 indicating fully democratic institutions and -10 indicating fully authoritarian institutions (Marshall and

Jaggers, 2002).15 The Freedom House measures of political rights (FHPR) looks at the

access, competitiveness and fairness of elections using surveys of country experts. It

reports scores from 1 to 6, with 1 indicating the most democratic (Freedom House, 2005).

For the purposes of this study, the scale is reversed so a higher score indicates a higher level of freedom. The final measure is taken from Alvarez et. al. (1999), and is a dummy variable indicating whether there is institutionalized turnover through elections in the country. Their data has been expanded, using the original coding rules, to cover through

13 This can be overcome to some extent by lagging the independent variable. However, there is not any a- priori theory for how long lags should be for various countries, and even lagging just five years (as Ross, 2001 does) means giving up almost all information on a number of cases. 14 The reader should note that the main results are robust to a number of different specifications, including AR(1) corrected regression with panel-corrected standard errors (Beck and Katz, 1995), GEE population averaged models (Zorn, 2001), and vector decomposition (Plümper and Troeger, 2007). These were not reported, either because of difficulties with other variables, large numbers of lost cases, or an inability to handle some of the more sparse data utilized in later chapters. 15 The Polity II was used instead of the Polity IV rankings because the gaps in the newer ratings system, to indicate unstable political time. 50 2001.16

The independent variables in Table 2.1 include the primary variables of interest,

as well as some important control variables. The log of per-capita GDP is used to control for level of economic development (World Bank, 2003).17 Later in this work, the set of

causal mechanisms surrounding the link between economic development and political

democratization will be discussed in greater detail, but, for now, it will suffice to note

that the empirical relationship between higher levels of per-capita wealth and democracy

is one of the strongest in the social sciences (Lipset, 1959; Diamond, 1994). Some of the

more explicit measures of socio-economic development, such as education or

urbanization, are not utilized because they would either dramatically cut the number of

cases or are misleading in oil states. Urbanization, for example, tends to be very high in

Middle Eastern oil states, not because of population transfers, but because of very low

population density outside of cities. A dummy variable is placed in the model indicating

whether the country-year is post-1991. This aims to capture the effect the fall of the

Soviet Union had on democratization. Not only did the countries of Eastern Europe and

the former Soviet union undergo a large shift towards democracy with the collapse of the

Communist bloc, but countries outside of the region, which had previously relied on

Soviet aid, also underwent a period of democratization either to draw replacement funds from the West or because of the discrediting of the communist ideology (Fukuyama,

1992).

The two variables of most interest are the average proportion of merchandise export made up of fuel exports and the average yearly per-capita revenue garnered from

16For more detail on the coding process see Przeworski et. al. (2000). 17The log of wealth is chosen primarily because it is the normal variable used in previous literature (see Ross, 2001). Alternative functional specifications have little or no impact on the results. 51 those fuel exports (World Bank, 2003).18 Results from these models support the general

conclusion that fuel exports are detrimental to the possibility of a country becoming

democratic. Table 2.1 looks at a full sample of countries, while Table 2.2 looks at a

restricted sample of non-industrialized countries, demonstrating that this is not due

simply to a comparison of industrial and non-industrial states.

Model 1 Model 2 Model 3 Polity FHPR ACLP

Fuel Dependence -.075 (.001) -.019 (.001) -.020 (.003)

Fuel Income -.001 (.020) -.0003 (.002) -.003 (.000)

Ln(Per-capita GDP) 2.384 (.003) 1.503 (.000) 4.177 (.000)

Post Cold War 3.792 (.000) .550 (.000) 2.958 (.000)

R2 Within .190 .059 --- R2 Between .448 .649 --- R2 Overall .350 .490 ---

N 4153 3433 4074 Groups 138 137 136 a – values are the weighted average of the within and between coefficients with p-values (1-tailed) in parentheses.

Table 2.1: Effect of fuel exports on democracy in all countries, 1965-2001.

18The measure of fuel export dependence -- fuel export revenue as percent of merchandise exports – was chosen for several reasons. First, it is generally consistent with other measures of dependency. For example, an index of fuel exports against GDP is 90 percent correlated with the variable indexing fuel exports against total merchandise exports. Second, there is more data available. The lack of GDP data for several countries causes the exclusion of large fuel exporters from analysis. For both these reasons, indexing against merchandise exports was preferred. 52 Model 4 Model 5 Model 6 Polity FHPR ACLP

Fuel Dependence -.052 (.009) -.015 (.005) -.013 (.053)

Fuel Income -.001 (.010) -.0003 (.005) -.003 (.001)

Ln(Per-capita GDP) .517 (.312) 1.015 (.001) 2.809 (.000)

Post Cold War 4.649 (.000) .669 (.000) 3.009 (.000)

R2 Within .241 .069 --- R2 Between .212 .417 --- R2 Overall .192 .240 ---

N 3354 2797 3275 Groups 115 114 113 a – values are the weighted average of the within and between coefficients with p-values (1-tailed) in parentheses.

Table 2.2: Effect of fuel exports on democracy in developing countries, 1965-2001.

The results presented in these two tables generally support the natural resource curse hypothesis. Average fuel export dependence is significantly and negatively related to political freedom, the presence of democratic institutions, and the probability of meaningful political competition. The coefficients in models 1 and 4 suggest countries with a thirteen percent higher average dependence on fuel exports will be almost a full point lower in their Polity score than their less fuel dependent counterparts. In models 2 and 5, the coefficients suggest that countries with about 50 percent higher dependence will also rate one point lower in terms of the Freedom House political rights index. The

ACLP democracy score, which only looks at whether or not the country has competitive politics, also indicates that countries with higher average fuel dependence are less likely

53 to be democratic.19 Figures 2.1 and 2.2 show this visually, by plotting the probability of a

country being democratic according to ACLP’s criteria, against the average dependence

on fuel exports. In both comparisons of all available countries and those of just

developing countries, the charts show that at low levels of dependency, the probability of

democracy is relatively evenly spread. As we move up the scale of fuel dependence,

states in the higher probability areas of democracy become increasingly rare. All of these results are within accepted bounds of statistical significance (p < .054), suggesting that it is highly unlikely that these relationships are due to chance.

1 .8 .6 .4 Pr(Democracy) .2 0

0 20 40 60 80 100 Fuel Export Dependence

Figure 2.1: Predicted probability of democracy (ACLP) based on fuel dependence for all countries, 1965-2001.

19A word on endogenaity should be placed here. It has been suggested to the author that the level of dependence may be effected by regime type, namely authoritarian regimes may promote dependence on oil. While not mentioned here, I did run tests to check this hypothesis, and there was a statistically significant correlation. However, this correlation was not substantively large, and the R squared in most models indicated that democracy explained less than 1% of the variance in dependency. Further work is being undertaken in relationship to the project to explore the political aspects of dependency, but the results to date have not undermined the correlation outlined above. 54 1 .8 .6 .4 Pr(Democracy) .2 0

0 20 40 60 80 100 Fuel Export Dependence

Figure 2.2: Predicted probability of democracy (ACLP) based on fuel dependence in developing countries, 1965-2001.

Per-capita income garnered from fuel exports also has a statistically significant and negative relationship with the level of democracy. Again, all the models are well within acceptable ranges of statistical significance (p < .021), suggesting that the results are not due to random chance. In models 1 and 3, a $1,000 increase in per-capita income is correlated with a one point decrease in the country’s Polity score. In models 2 and 4, approximately $3,000 in per-capita income from fuel exports is correlated with a 1 point decrease in a country’s Freedom House rating. Figures 2.3 and 2.4 show the effect of fuel export income on the probability of competitive politics in both the full sample and the sub-sample of developing states. As with fuel dependence, there is a clear pattern of lower probability of democracy at higher levels of fuel export income. The figures also preview some of the patterns that will be explored more in the next two chapters. The astute reader will notice that the probability of democracy remains about 20 to 50 percent 55 in even the most fuel dependent countries, whereas the very high fuel export income

countries show much less variation. Fuel exports, it will be shown in the coming

chapters, have substantially different impacts on regime outcomes and stability,

depending on whether the country is both dependent and rich from its fuel exports or if it

is just dependent on those exports.

1 .8 .6 .4 Pr(Democracy) .2 0

0 2000 4000 6000 Fuel Export Income

Figure 2.3: Predicted probability of democracy (ACLP) based on fuel export income for all countries, 1965-2001.

56 1 .8 .6 .4 Pr(Democracy) .2 0

0 2000 4000 6000 Fuel Export Income

Figure 2.4: Predicted probability of democracy (ACLP) based on fuel export income for developing countries, 1965-2001.

In general, these tests lend further support to the natural resource curse thesis.

Controlling for the level of economic development in a country, fuel exporting states are generally less democratic than their counterparts. This pattern is not something that is artificially created by comparing developing and industrialized countries. Rather, the

relationship remains constant, even when industrialized countries are excluded from the

analysis.

2.5 Moving forward

This chapter addressed several important issues providing the foundation for

analyses later in the dissertation. First, I reviewed the history of the literature on the

natural resource curse and fuel development, and found that the effects scholars predict

from fuel exports depend, to a large extent, on the time and space of their analysis. Those 57 who study fuel development prior to the 1970 oil crisis, tend to view fuel resources as

useful for a country’s ability to develop economically and politically. Those scholars

who look at the period after the oil crisis come to the opposite conclusion, arguing that

fuel exports, and the dependence on these exports, is harmful to economic development

and the prospects of democratization. In this chapter I argue that this is largely due to

changes in the market for fuel, which is now more globalized and diversified than prior to

the oil crisis, changing both the countries involved in the oil trade and their ability to use

those resources for industrialization.

In addition, this chapter reviewed how the natural resource curse is viewed in the

various country-studies and regional literatures. I argued that different regions

experienced the natural resource curse in different manners, and, as such, led scholars to

different ideas about what caused democratic underperformance in those countries. The

next chapter will show that these differences carry over into the large-N theoretical

literature.

Finally, I assessed the correlation between fuel export dependence and revenues

using a dataset, spanning 168 countries from 1965-2001. Using different tools, this test

confirmed the negative relationship between fuel exports on democracy, both in a global

sample and a sample of developing states.

How this pattern comes about, however, remains unclear. The area-studies

literature on the characteristics of fuel exporting countries shows a diverse number of

pathways, all of which appear to contribute to the general pattern of democratic under- performance. The next chapter examines these patterns more explicitly, and proposes a theory of politics and distribution in fuel exporting countries that aims to account for the diversity of these pathways, and how they constitute an overall pattern of non-democracy. 58 CHAPTER 3

DEPENDENCE, INCOME, AND REGIME STABILITY

The last chapter demonstrated that the negative correlation between fuel exports and a lack of liberalism was not illusory. Although the implications of the natural resource curse thesis have already begun to be translated into scholarly conventional wisdom and public policy, we are still a long way from understanding the causes of this phenomenon, especially in a cross-national framework. Much of the work done in this area generalizes from a particular set of cases (e.g. Loung and Weinthal, 2006; Ross,

2001; Karl, 1997). The countries affected by the natural resource curse, however, exhibit a wide variety of development patterns. Thus, the current explanations are unlikely to provide a satisfactory basis for scholarly analysis or policy-making.

In order to set this empirical phenomenon on firmer theoretical ground, this chapter provides a model using the government’s distribution imperative associated with fuel exports to explain the general lack of democracy in these states. This analysis demonstrates that fuel dependent states experience a wide variety of outcomes, that an approach based on government distribution fits the empirical patterns observed across fuel exporting countries better, and that this approach also provides a clear identification of the underlying dynamics of fuel exporting politics.

59 The beginning of this chapter will cover existing explanations for the empirical

patterns across countries and regions. After establishing some of the gaps in previous

work, the theory-building section of this chapter begins. The discussion will focus on how the attributes of fuel exports increase the importance of government distribution in

the state’s economy. From these general attributes of fuel exports, I introduce a formal

model to show how the increased emphasis on government distribution resulting from

fuel export dependence explains the statistical pattern associated with the natural resource

curse. The final section returns to the literature-to-date on the natural resource curse.

Here, I argue that the model focusing on the importance of government distribution has several important advantages over previous hypotheses, and I suggest some straightforward and novel expectations that will be tested in the next three chapters.

Several main contentions stem from this analysis. (1) As the income of potential opposition groups becomes more tied to government distribution, they are less likely to challenge authoritarian governance. (2) As incomes become more tied to government distribution, actors are less likely to prefer a democratic outcome, which would make control over government resources uncertain. (3) States with a high ratio of fuel export revenue to population, sometimes referred to as “capital surplus” fuel exporters, remain stable, as they can avoid conflict through distribution. (4) Fuel dependent states with a lower ratio of revenue to population, sometimes referred to as “capital deficient” fuel exporters, are characterized by unstable government because of conflicts over distribution of the resource.

60 3.1 Limited explanations

Scholars have posited several hypotheses to explain the pattern in the last chapter.

All of these explanations are insightful, but they fall short of providing general theory

that explains why fuel dependent states, with quite different individual characteristics,

historical legacies and political outcomes, still form an overall class of democratic under-

performers.

3.1.1 Authoritarian stability

The most cited explanations for the lack of democracy in fuel dependent states

draw from the literature on the Middle East. Not surprisingly, this literature focuses on

explaining authoritarian stability. Ross (2001) was the first to analyze whether

theoretical concepts developed in the Middle Eastern literature extend to a large-N

framework. First, he finds that fuel exporters’ ability to raise revenue without taxation

reduces calls for political reform. Second, fuel exporters have greater resources for

patronage to reduce pressure for democracy. Third, the government’s oil wealth, either purposefully or accidentally, reduces the formation of civil society groups. Together,

these hypotheses are labeled the “rentier effect” of oil revenues. He also finds some

support for the contention that fuel wealth does not result in socio-economic development

that is proportionate to the increases in per capita GDP, although he does not give an

explanation for why this is the case. In all of these explanations, the implication is that

fuel dependent states are more likely to be authoritarian, because increased government

revenues obtained independently of social collection, make authoritarian governments

more stable.

61

Most scholars, even the ones discussed below, will grant that when raw material values are high, authoritarian governments can prevent changes in the regime through

"buying off" the opposition. Many, however, do not believe that it will last when the government runs short on revenue because of low raw material prices. Benjamin Smith

(2004) argues that this is not the case. He contends that four cases, the “big four,” grab the attention of researchers -- Iran, Nigeria, Algeria and Venezuela.20 Instability in these four countries has led researchers to predict government crises as a result of falling fuel prices. However, he argues cases like Suharto in Indonesia and Saddam Hussein in Iraq, who both lasted more than thirty years, are more common (p. 242). A duration analysis of fuel dependent regimes in the developing world supports his argument. These results, contrary to what would be expected, hold up regardless of whether oil prices are high or low. He theorizes that oil dependent regimes effectively use their revenues to align social interests and create institutions to support the regime.

Another hypothesis for authoritarian stability revolves around increased spending on repression. Some scholars note a positive correlation between oil dependence and spending on military and police as a proportion of their income (Skocpol, 1982; Clark,

1997). Cross-national support for this hypothesis is generally weak (Ross, 2001; Smith,

2004). Separating spending on domestic repression from spending on national defense is also very difficult. For example, Saudi Arabia, Iran, Iraq, and Kuwait would be expected to spend more than the average amount on defense, since all of them have been threatened with foreign invasion in the past twenty years. Algeria, Nigeria, Indonesia and

Angola all have civil unrest on a scale that would require active military intervention.

20 Ross (1999) makes the case that Venezuela was actually quite stable as compared to its neighbors and actually doubts whether most oil states are all that more unstable than their neighboring countries. 62 Increases in military and police spending are more likely an effect of domestic and

international instability, rather than a cause of domestic tranquility.

Another factor receiving much attention, primarily in the media, is the international support given to oil-rich regimes. Especially in the case of Iran, observers credit US support for both the rise and fall of the Shah (e.g. Cottam, 1979). Others contend that Saudi Arabia, Kuwait and other oil rich regimes in the Middle East do not receive sufficient international pressure regarding their human rights practices, compared to that of countries without oil (Ross, 2006b). A number of oil rich countries, however, have been subject to significant international sanctions, including Iran, Iraq, Syria, Sudan,

Chad and Libya; although this has primarily been because of international aggression rather than domestic rights abuses.21

The “rentier effect,” slow socio-economic development, institution building,

repression, and the lack of international pressure are all plausible explanations for the

statistical pattern of the natural resource curse, but all of them rely on explaining the

pattern through authoritarian stability.

3.1.2 Democratic instability

A number of other scholars suggest that the main reason for the pattern of non-

democracy is the instability of governments, especially democratic ones, in fuel

dependent states. Karl’s (1997) work on Venezuela, Iran, Nigeria, Algeria, Indonesia and

Norway is the best-known example of this literature. She contends that when a

government receives a large windfall from spikes in oil prices, it will have a difficult time

21 Developed states will sometimes attempt to support/install cooperative regimes in oil exporting states. This has certainly been the case in the US support of Kuwait and Saudi Arabia in the first Gulf War. The same argument could also be made about the Shah in Iran. These actions, however, have a mixed record of success and seem to be limited to a few historic examples. 63 resisting social pressures to increase spending:

As a result, when faced with competing pressures, state officials become habituated to relying on the progressive substitution of public spending for statecraft, thereby further weakening state capacity (1997, p. 16).

Governments will remain relatively stable as long as the price of fuel continues to rise.

When the price falls, however, the fragile framework on which the government in a developing state rests, will collapse.

Chaudhry (1997) makes a similar argument about the importance of establishing government capacity to avoid instability during a drop in prices. She compares Saudi

Arabia, which depended on oil revenues, and Yemen, which depended on labor transfers, and finds that Saudi Arabia had much weaker institutional capacity to respond to crises.

The primary reason for institutional weakness, she contends, is the lack of an effective tax collection bureaucracy. Because the government of Saudi Arabia did not have incentive to tax the population, a number of other functions necessary for state control, including tracking the population and gathering demographic information, fell by the

wayside. Luong and Weinthal (2006) also cite the lack of a professional taxation

apparatus as a key aspect of institutional weakness in some oil dependent states.

Other scholars posit a pattern of democratic instability that works in reverse of

Karl’s boom-bust cycles. Lam and Wantchekon (2002) suggest that oil dependent states

in Africa experience political “Dutch disease,” whereby politicians elected to government

with high fuel revenues are able to use resources to consolidate a one-party state. They

argue that office-seeking leaders are able to use patrimonialism and corruption to lock in

an electoral advantage. Thus, as fuel export revenues rise, the danger of a democracy

being replaced by a one party state increases.

64

Still others point to an even larger pattern of political instability in fuel dependent states, especially in terms of social unrest and civil war. Ross (2003, p. 21) notes that five of the twenty most oil dependent states he examines have experienced some type of civil war since 1990 related to natural resource reserves. In a similar vein, Collier and

Hoeffler (1998; also Bannon and Collier, 2003) observe a strong correlation between dependence on natural resources and the probability of civil war. Bannon and Collier

(2003, p. 5) also note that oil production increases the risk that a civil war, when it occurs, involves secession. They posit several reasons for this. First, separatist groups must bankroll themselves, and they use extortion and the exploitation of trade in primary commodities to do it. High revenue exports are easier to use for extortion, trade, taxation and/or participation in the trade itself. Second, regions have a greater incentive to assert their independence where the distribution of resources aligns with other cleavages of conflict, such as ethnicity. Third, poor governance and corruption exacerbate civil strife, leading to perceptions of a national resource being stolen by corrupt elites.

While these authors have not drawn a direct connection, a relationship between fuel dependence and social unrest may explain the lack of democracy in these states. As

Linz and Stepan (1978, p. 18) note, “Civil war almost always results in non-democratic governance.” Wantchekon (1999) comes closest to making this link explicit, by contending that unequal distribution of oil resources results in resentment undermining stability of democratic rule. If this is the case, then fuel export dependence should be correlated with greater incidence of civil war, and this, in turn, should explain some of the democratic underperformance in these states.

65

Weak governance, decaying bureaucracies, political Dutch disease, and civil conflict all provide convincing explanations for greater democratic fragility in fuel dependent states. This predicted outcome stands in contrast, however, to the pattern of authoritarian stability predicted in the previous section.

3.1.3 Ambiguous empirical results

What is most interesting about the contradictory conclusions reached by these studies is that they all make a persuasive case, both in a regional setting and using large-

N statistical research. Why is this? An answer begins to emerge in Table 3.1, which shows some characteristics of the most fuel export dependent countries, those for whom fuel exports average more than 5 percent of total GDP from 1965-2001.22 The second column lists the percentage of GDP made up from fuel exports and the third column gives an approximation of the income from those exports per capita. The fourth column gives a listing of regime types, classified according to their Polity score as authoritarian, soft-authoritarian, semi-democratic, and democratic.23 The final column gives a listing of civil conflicts, classified as minor, intermediate or full civil war (Gleditsch et. al.,

2002).24

22 Some countries are left off of the list because their fuel income is derived primarily through refining and shipping, as determined by comparing the World Bank (2003) statistics against the BP (2005) oil production data. These are Bahrain, Lithuania, Bhutan, Singapore and Belarus. Leaving these countries out of the statistical analysis does not significantly change the results. Libya, Iraq and Qatar are not included in the table because of lack of GDP data. They are, however, included in the statistical models which index against merchandise exports. 23 Regimes were classified as follows: authoritarian is their Polity score was less than or equal to -5, soft- authoritarian if their score was greater than -5 and less than or equal to 0, semi-democratic if their score was between 0 and 6, and democratic if their score was a 6 or above. 24 Smith (2004) also uses these classifications. 66 Average Average income per dependence capita from Country on fuel fuel Polity outcomes Civil conflicts Angola 40.98% $200.61 1975-1990 authoritarian 1990-1994 war 1991 transition 1995 intermediate 1992 interregnum 1996-1997 minor 1993-1996 transition 1998 war 1997-2003 soft-authoritarian Nigeria 39.65% $106.61 1960-1965 democratic 1966 minor 1966-1977 authoritarian 1967-1970 war 1978 transition 1979-1983 democratic 1984-1997 authoritarian 1998 transition 1999-2003 semi-democratic Saudi Arabia 34.85% $3,065.24 1926-2003 authoritarian 1979 minor

Azerbaijan 34.68% $147.25 1991 soft-authoritarian 1993 minor 1992 semi-democratic 1994 war 1993-1994 soft-authoritarian 1995- minor 1995-2003 authoritarian

Oman 31.51% $1,561.06 1800-2003 authoritarian

Kuwait 29.97% $4,709.60 1963-1989 authoritarian 1990 foreign takeover 1991-2003 authoritarian Trinidad 29.31% $1,212.25 1962-2003 democratic 1990 minor

Gabon 28.46% $1,388.57 1960-1989 authoritarian 1990 transition 1991-2003 soft-authoritarian Congo 27.01% $241.18 1960-1962 semi-democratic 1993-1994 minor Brazzaville 1963-1990 authoritarian 1991 transition 1992-1996 semi-democratic 1997-2003 authoritarian

Continued

Table 3.1: Political outcomes in fuel dependent countries, 1965-2001.

67 Table 3.1 continued

Yemen 26.53% $73.94 1990-1992 transition 1994 war 1993-2003 soft-authoritarian Algeria 21.50% $346.08 1962-1988 authoritarian 1991-1992 – minor; 1989-1991 soft-authoritarian 1993-2001 – war 1992-1994 authoritarian 1995-2003 soft-authoritarian Iran 18.62% $230.39 1955-1978 authoritarian 1966-1968 1979-1981 transition intermediate 1982-1996 authoritarian 1979-1982 war 1997-2003 semi-democratic 1983-1988 intermediate 1990-1993 intermediate 1996-1997 intermediate 1999-2001 intermediate UAE 17.73% $5,895.76 1971-2003 authoritarian

Venezuela 16.67% $619.74 1958-2003 democratic 1992 minor

Turkmenistan 15.81% $190.95 1991-2003 authoritarian

Syria 13.72% $97.99 1963-2003 authoritarian 1966 minor 1979-1981 minor 1982 war

Russia 12.08% $283.30 1992 democratic 1993-1994 minor 1993-1999 semi-democratic 1995-1996 war 2000-2003 democratic 1999-2001 war Kazakhstan 11.81% $159.55 1991-2001 soft-authoritarian 2002-2003 authoritarian

Continued

68 Table 3.1 continued

Indonesia 8.23% $51.30 1959-1998 authoritarian 1965 minor 1999-2003 democratic 1967-1969 minor 1975-1978 war 1979-1989 intermediate 1990-1991 minor 1992 intermediate 1997-1998 intermediate 1999-2001 minor Norway 7.59% $2366.34 1945-2003 democratic

Ecuador 6.94% $102.70 1961-1967 soft-authoritarian 1968-1969 democratic 1970-1971 soft-authoritarian 1972-1978 authoritarian 1979-2003 democratic Malaysia 5.54% $171.37 1957-1968 democratic 1965-1966 minor 1969-2003 semi-democratic 1974-1975 minor 1981 minor

A wide variety of political outcomes are represented here, from very stable

authoritarian states and democratic states, to those which alternate regularly between

regime classifications and experience several bouts of civil unrest. Of the 22 countries

listed in this table, approximately ten have relatively stable polities, experiencing one or

fewer transitions. The other half experience two or more – as many as seven in the case

of Nigeria – significant changes in political institutions. The presence of civil unrest has

a similar pattern, with fourteen experiencing some type of civil unrest and eight experiencing no measured civil disturbance. No clear pattern emerges about the relative stability of these states. Indeed, the sample is split almost exactly in half between countries with stable political institutions and those experiencing civil unrest. This is problematic, then, for theories predicting one causal pathway to democratic under- performance in fuel dependent states. 69 Figure 3.1 places this into a visual context, plotting the Polity IV scores of the

above countries over time. The results is an absolute mess, but an informative one.

Again, no clear pattern emerges. A number of states, like Norway and Trinidad, remain

consistently at or near the highest levels of democracy. Others, like Saudi Arabia and

Kuwait, are at or near the bottom for the entire time. Still others move around the chart in an irregular pattern. Attempting to find a generalizable pattern from this chart is an exercise in futility. How then do we end up with this stability/instability bifurcation in the natural resource literature?

10

8 Angola Nigeria Saudi Arabia 6 Azerbaijan Oman Kuwait 4 Trinidad Gabon 2 Congo Brazzaville Yemen Algeria 0 Iran

7 UAE 73 83 85 95 9 9 9 9

Polity II Score 1965 1967 1969 1971 19 1975 1977 1979 1981 1 1 1987 1989 1991 1993 1 19 1999 2001 Venezuela -2 Turkmenistan Syria Russia -4 Kazakhstan Indonesia -6 Norway Ecuador Malaysia -8

-10 Year

Figure 3.1: The variation in political regimes among fuel exporting countries, 1965-2001.

70 Placing this back into the context of explanations for the natural resource curse, whether a researcher observes an underlying pattern of stability or instability as the explanation for non-democratic outcomes depends heavily on the sample. Karl’s (1997) research, for example, is explicitly limited to six states – Venezuela, Iran, Nigeria,

Algeria, Indonesia, and Norway -- which are major world exporters, have a certain level of dependence, and do not have a general capital surplus. Smith’s (2004) statistical analysis, discussed in the next chapter, is heavily influenced by the stability of a few very wealthy oil states.

That these problems have not been explicitly addressed is problematic for a number of reasons. First, most readers do not appreciate these limitations. For example, although Karl’s (1997) research is explicitly limited to a few cases, many now talk about the “paradox of plenty” without reference to the qualifying conditions. Second, limited samples prevent researchers from looking for underlying patterns that might explain both outcomes. Some underlying attributes of fuel exports that might explain the patterns presented in Table 3.1, but studies drawing from a single region fail to identify these relationships. Finally, scholarship has become stuck in a futile debate about underlying patterns of stability and instability, which, depending on the sample and methods, will produce inconsistent results. In reality, this debate appears largely illusory and diverts attention away from exploring the baseline characteristics of fuel resources that might influence governance.

71 3.1.4 Other explanations

The inconclusive results from the regional literature on the Middle East and

Africa has led some scholars to bring forth alternative hypotheses for the underlying patterns of the natural resource curse.

One such area of research focuses on the ownership of the natural resource. Ross

(1999, p. 319) suggests that state ownership of the resource might explain the pattern of success and failure in economic development among states, as parastatals tend to soften budget constraints (see also Quinn, 2006).25 Luong and Weinthal (2006) take this further, arguing that state ownership, at least in the post-Communist context, might provide an explanation for different institutional outcomes. They argue that state-owned resources reduce the need for taxation, resulting in two outcomes: (1) less social pressure associated with taxation decisions; and (2) weakened bureaucracy, as the state is not concerned with revenue extraction (see also Chaudhry, 1989). The first outcome explains the stability of some authoritarian regimes, since they face less social pressure for reform. The second outcome explains the instability of governments, since they have inadequate state capacity to respond to crises.

Generalizing this hypothesis produces a number of problems. First, the lines between state and private ownership are not entirely clear in the oil sector. Even in states that give concessions, which are increasingly rare, the government often retains legal ownership of the underground reserves. This, along with the numerous legal regulations accompanying a company’s investment into oil reserves, means that the state maintains an asymmetric relationship with the oil development company.

25 A “parastatal” is an agency or company that is owned, either in whole or in part, by the government. 72 Second, states often renege on agreements with private firms. Russia, which is

Luong and Weinthal’s primary example of domestically privatized oil development, is

rife with examples. In 2004, the state owned oil company, Rosneft, acquired the largest

private oil producer, Yukos, after questionable tax evasion and fraud charges were

brought against the company by the government.26 This move was seen by many

analysts as a response to the political activism of Mikhail Khodorkovsky, the former

owner of Yukos, against the President Vladimir Putin’s government (RFE/RL, 2006).

Currently, Khodorkovsky is in a Siberian labor camp, and his company is bankrupt.27

This is hardly a portrait of a private industry that has strong bargaining position vis-à-vis the state. Even the larger international oil firms have difficulty asserting independence. As Ottaway (2001) notes, NGOs and governments have pinned considerable hope on the ability of international oil companies to promote greater levels of democracy and transparency in oil exporting states. Yet, the power of these companies in their host states is often tightly constrained. In Russia, several major oil companies have had large investment projects re-negotiated unilaterally by the government after making significant initial investments. In the Sakhalin II oil field, Shell and its two

Japanese partners, Mitsui and Mitsubishi, were bullied into selling a majority stake in the project to Gazprom, the state-owned gas company. This prompted the European Bank for Reconstruction and Development to cancel its investment in the project (Pfeifer,

2006). British Petroleum (BP) also became a victim of nationalization in the fuel

26 Controversy still surrounds the tax evasion charges. While the accounting company PricewaterhouseCoopers has stood by its audit reports since the 2004 tax evasion charges were leveled against Yukos, they recently recanted, saying that they were provided with false information by the managers. Khodorkovsky’s lawyers blame this on pressure from the Russian government (RFE/RL, 2007b). 27 In addition to the charges of tax evasion, Khodorkovsky and his former business associate, Platon Lebedev, have now been charged with stealing $32 billion in oil (RFE/RL, 2007a). 73 industry. In June 2007, the Russian government forced the British-Russian joint venture,

TNK-BP, to cede control of the Kotytka gas field to Gazprom. The official reason for this move was that TNK-BP was purportedly not producing enough gas to comply with its production agreements with the Russian government. The Russian authorities did not mention that the company repeatedly stated that it could not meet its production agreements without a pipeline to China, and state-owned Gazprom had repeatedly blocked this project (Kupchinsky, 2007).

Oil companies make an extremely large fixed investment, and once this investment occurs, the power in negotiations shifts from the company, who provides needed investment and expertise, to the host country, which has the power to nationalize

(Vernon, 1971). The only real negotiating power that private companies have against the government derives from the fear that future investment may be less forthcoming and their ability to produce at a higher level than the state owned industries. In the latter case, many governments do not believe that international companies are really more efficient, fear that the international company will produce in a manner that is adverse to the state’s interests, or simply care less about overall production than about control over resources.

In the former case, threats of future withholding of investment are not likely to be credible, both because of the increased competition for oil reserves coming from India and China and because, even where the states confiscates or nationalizes assets, oil development projects can still be incredibly profitable (see e.g. The Economist, 2007, p.

66-67).

Oil companies also lack incentives to promote democratization. Domestic private firms have often gained their footing in the market through less-than-democratic transactions. Russia’s privatization of its fuel resources was an opaque affair, with 74 former and current government apparatchiks receiving favorable treatment (Shleifer and

Treisman, 2000). These groups often trade the continuation of their property contracts for at least tacit support of the current government. International fuel development companies also face weak incentives to promote democratization. As long as their property rights are secure, they gain little from government reform. Transparency measures are also difficult to implement, since they put the participating companies at a strategic disadvantage compared to those who are willing to forego transparency to get a contract (Ross, 2003).

Another method for explaining non-democracy in fuel exporting states is to extrapolate from more general theories about democratization. Boix’s (2003) explanation of authoritarianism, for example, also has implications for fuel exporting states. Boix contends that where domestic income inequality is high and capital assets are fixed, elites are less likely to accept democratization, because it likely involves the taxation of their profits and redistribution to less-well-off individuals. In fuel exporting states the main export, oil, is a fixed asset and inequality between those who directly benefit from the oil industry and the many who are not employed in this sector is likely to be high. This causes the elite, who benefit from oil development, to protect the authoritarian regime, since democracy would result in the redistribution of their resources.

While this explanation is intellectually compelling, especially since it brings the dynamics of fuel exporting states into line with the general democratization literature, several problems exist. First, discussions of taxation and re-distribution are not easily applied to fuel dependent states. Indeed, in many of the fuel dependent states, revenue accrues directly to the government rather than being acquired through taxation. Not only that, but this lack of taxation appears to have the effect of quelling popular unrest, not 75 causing it as Boix would predict (e.g. Ross, 2001). Boix’s explanation provides a reason

why elites would continue to support an authoritarian solution, but he does not explain

why those who are not directly involved in oil development would support the status quo.

Yet in at least half of the cases, many of which happen to be the richest fuel exporting

states, the primary interest for scholars has been the relative absence of pressures for democracy (e.g. Smith, 2004).

Boix’s explanation also does not lend itself easily to the exploration of causal pathways for political outcomes in fuel dependent states. The theory does not provide guidance for predicting the variant outcomes among fuel dependent states displayed in

Table 3.1. Stability versus instability as paths to authoritarian outcomes are obfuscated in

favor of general prediction of regime-type outcomes. This problem stems at least

partially from the theory’s decision to focus on the distribution of social, as compared to

governmental, resources.

The explanations advanced so far highlight important aspects of politics in fuel

exporting states, but fall short of explaining the general pattern presented in section 2.4.

Indeed, the reliance on generalization from a few cases or regional patterns has resulted

in a body of theory that masks the underlying variety in political outcomes across fuel

exporting states. To move away from this approach, the next section focuses on the

major attributes of fuel exports as compared to other products. The explanation for non-

democracy advanced below is complementary to many of the theories discussed in this

section, but it differs in terms of its potential application to the wide variety of fuel

exporting states. This explanation also helps identify important underlying patterns

which can be useful for both policy-makers and case-study researchers in interpreting the 76 dynamics of governance in fuel exporting states.

3.2 Modeling the curse

This section turns to establishing an explanation for the general pattern of non-

democratic governance in fuel dependent states by drawing a consistent theory from the

attributes of fuel exports, rather than generalizing from a particular case or group of

cases. This method produces more accurate predictions in the large-N framework than

have generalizations from the case literature, as demonstrated in the next two chapters.

I focus on government distribution and draw several conclusions. (1) As incomes

become more tied to government distribution, challenges to authoritarian governance are less likely. (2) As incomes become more tied to government distribution, opposition groups are less likely to prefer democratic outcomes. (3) States with a high ratio of fuel export revenue to population remain stable, since they can avoid conflict through distribution. And (4) fuel dependent states with a lower ratio of revenue to population are characterized by unstable government because of conflicts over control over distribution of the resource.

3.2.1 Why are fuel exports different?

How do fuel exports differ from the export of other products? The answer to this

question serves as the foundation for an explanation of the natural resource curse.

Low Employment Gains. Unlike agriculture, industry, and even some types of

mining, exploration and exploitation of oil does not result in large levels of employment,

and much of the employment that does take place occurs during the initial phases of

construction and development. For example, in the small country of Trinidad and 77 Tobago, in 1973, fuel exports made up about 83 percent of total merchandise exports and

about 20 percent of GDP, but employed only 4 percent of the total workforce (Black et.

al., 1976). Similarly, in Azerbaijan hydrocarbon production accounted for 30 percent of

GDP in 2000, but only 1 percent of total employment (Bagirov et. al., 2002, p. 89). Oil development companies also tend to bring in trained workers from outside of the host country, especially in developing states which do not already have a trained workforce.

The importation of trained labor further limits the employment gains from oil development. Kazakhstan is so concerned about this tendency that it has resorted to adopting a number of laws to bolster the hiring of Kazakhstani workers, including barring companies from bringing in non-Kazakhstani workers if a citizen of Kazakhstan could do the same job.

Geographic Concentration. Oil and natural gas are usually referred to as “point source” resources, because they are concentrated in specific areas of a country. This means that whatever employment and revenue gains accrue to a country will tend to be geographically concentrated unless they are re-distributed through other non-market means. The map in Figure 3.2 illustrates this in the case of Nigeria.28 The major tribal

regions are delineated by different colors on the map, while the place-markers indicate

the approximate location of major oil wells. The oil resources are predominately located

and tightly concentrated in the south of the country, along the Niger River Delta. This

has been a continuous source of problems for Nigeria, since the north has traditionally

been the more populous and politically dominant area.

28 Figure created using Google maps. 78

a -- adapted from Marin (1999) and Central Intelligence Agency (1979).

Figure 3.2: Location of major oil fields and ethnolinguistic groups in Nigeria.

79 A similar distribution problem is occurring in Iraq, where oil reserves are

predominately located in the primarily Kurdish north and the primarily Shi’a south of the

country. This has proven a major point of contention for the transition, as groups argue

over how revenues should be distributed (e.g. Weinstein, 2004; Djerejian et. al., 2002).

Figure 3.3 shows this pattern visually, marking the major oil fields in Iraq (see Horn,

2003), along with approximate dividing lines for the major population groups. As with

Nigeria, the debate over how much revenue should remain in the areas of production and how much should be distributed across regions and groups is a contentious issue, and a primary reason for the country’s continued instability. As of this writing, the government of Iraq has yet to pass a law on the administration of oil revenues.

80 a -- adapted from Horn (2003) and Central Intelligence Agency (1992).

Figure 3.3: Location of major oil fields and ethnolinguistic groups in Iraq.

81 Government Distribution. Ross (2006b) put it most bluntly -- "New oil revenues will make the government much richer, but it may not make the people richer." Oil revenues are not automatically distributed to a large portion of the population. If left to its own natural trajectory, oil revenue will accumulate in geographically concentrated outposts of oil development. For a number of political and social reasons, governments try to make the distribution of oil windfalls more widespread by signing contracts with the oil development company requiring that the government receive some of the windfall.

Several types of contracts exist. Concessions are the simplest types of contracts, and typified early oil development. Under these contracts, the government sells the rights to the land to a developer for a lump sum of money. Revenue-sharing is much more common today. Under these contracts, a pre-determined portion of the company's revenue goes to the government. Production-sharing agreements are another method for sharing revenue between the government and the oil development company. Here the government "owns" a portion of the oil mined by the company and is entitled to the revenue from that production. Both revenue-sharing and production-sharing agreements allow the government to have a much larger say in the operation of the oil company (e.g. by mandating a level of production), and provide a longer-term stream of revenue for the government. State-ownership is the most aggressive state strategy. Here a state-owned company develops the oil resources, sometimes with the help of outside contractors.

A large amount of variation exists within these types of contracts in terms of the government’s role, the level of regulation placed on the oil development company, and the revenues that accumulate to each signatory. Countries use a variety of contracts for oil field development, and sometimes a single development project can involve a mixture of contract types. For example, a country may sign a production-sharing contract in 82 extraction from a field, but will maintain state-ownership over the transportation of the extracted oil. This complicates categorizing the oil development strategies of countries into neat categories based on contract types.

Luong and Weinthal (2006) are correct to note that not all oil resources are state- owned, but the complexity of oil agreements makes drawing general conclusions about a state’s development policies difficult. The increased importance of government distribution, however, is an important and consistent aspect of fuel export development.

Where a country is dependent on oil resources, government distribution will play a much larger role in individual economic life because oil development does not result in commensurate gains in employment and geographic development.

Large Price Shifts - The Boom/Bust Cycle. An important attribute of oil export dependence is the susceptibility to price fluctuation. From January 1980 until January

2006, the real price of a barrel of oil, based on constant 2005 dollars, has had a standard deviation of $18.73, with prices ranging from $86.99 in January 1981 to $11.16 in

December 1998 (EIA, 2006).29 These fluctuations cause enormous swings in export revenues for oil exporting countries. For example, in 1980 Nigeria produced approximately 2.1 million barrels of oil per day, which, at the average yearly price of

$82.15 for that year, would have been worth $173 million a day in 2004 dollars. In 1998, the average price was $15.20 and Nigeria produced about 2.2 million barrels a day, but its production value dropped to $33 million a day despite the increased production. This amounts to a total drop in daily production value of approximately $140 million a day.30

Figure 3.4 illustrates these changes during the period from 1965-2004.

29For time-series dating back to 1861, see BP (2005) 30Calculated using data from BP (2005) 83 90.00

80.00

70.00

60.00

50.00

40.00

30.00 Price per barrel in constant 2004 USD 2004 constant in barrel per Price

20.00

10.00

0.00

6 2 6 68 70 7 74 76 78 80 82 84 86 9 9 Year 19 19 19 19 19 19 19 1 19 19 1 1988 1990 1992 1994 1996 1998 2000 Year

Figure 3.4: Spot crude oil prices, 1965-2001 (BP, 2005).

The low employment gains, geographic concentration, importance of government distribution, and price swings associated with oil development have shaped the explanations and causal pathways suggested by previous scholars, and will be a starting point for the model developed below.

3.2.2 The importance of government distribution

Of the factors established above, one requires particular attention: the importance of government distribution, defined here as economic resources over which the government has discretion in their allocation. Previously scholars have captured this in a tangential manner, by talking about public welfare programs, state ownership, or buying 84 off the opposition. The tendency for government distribution to increase in importance

for fuel dependent countries, however, can operate independent of assumptions about the methods of production or a government’s maneuvering for power. Whether from altruistic motives, the desire to spread the wealth or decrease inequality between regions, political motives, or rewarding political supporters using fuel revenues, the results across regions of the world seem to be the same. Where countries are dependent on fuel reserves the proportion of the total economy that is distributed based on the discretion of government leaders increases.

This increased dependence on distribution stems from several causes:

The need for re-distributing oil revenues. As noted above, fuel resources tend to result in low employment gains and those gains tend to be geographically concentrated.

Either for political or ideological reasons, a government may decide that revenues need to be distributed more equitably. This re-distribution can take many forms. It could come in the form of increased social services, as has been the case in Saudi Arabia and other

Middle Eastern oil states. It can also come indirectly through large state investment, such as in Venezuela and Nigeria.

Blaming leaders for these decisions is difficult, since inequality can become a divisive force. Indeed, re-distribution policies for oil revenues are not limited to developing countries. Norway distributes oil revenues from its state-owned oil company through a fund that that pays for the country's social security program. Similarly, Alaska distributes gains from its oil reserves through cash grants given directly to residents from its oil fund (see Tsalik, 2002). The key difference in these areas is that re-distribution is not directly tied to the government in power (through informal agreements in the former case and constitutional amendment in the latter). These policies are executed by 85 relatively autonomous bureaucracies under legal obligation in law-abiding states, and so

the resources are difficult for politicians to re-direct for other policy objectives.

Using oil funds for political support. A less altruistic reason for government distribution of funds from fuel exports is to gain or reward political support. In Nigeria, for example, the transfer of revenues from the oil-rich Niger Delta has reached such a high level that a large part of the Delta is still without basic utilities (Petroleum

Economist, 2006) because political leaders are rewarding their own constituencies using

oil funds. This, in turn, is a large part of the reason why the cultural affiliations of

presidential candidates and military leaders is so important in Nigerian politics – citizens

assume oil funds will be distributed to that leader’s ethnic, regional, religious or class group (Olowu, 1991). Kazakhstan has reversed this strategy, using its oil funds to build a new capital, and thus attracting large new investment, into its originally less-supportive northern region.31

“Sowing oil funds.” In developing countries, fuel exports are much more than a

promising sector of the economy. Especially in capital-deficient countries, fuel exports are viewed as a solution to the country's economic problems and a chance to use the capital to promote a modern economy. The examples are numerous. In Karl's (1997) description of Venezuela, President Carlos Andrew Perez propagated a vision of La Gran

Venezuela, where oil revenue pushed Venezuela into modernization over the course of years, rather than decades. In Nigeria, the government invested in auto manufacturing, a major steel complex, construction, transportation and primary education, going heavily into debt to finance these projects (Kretzmann and Nooruddin, 2006, p. 25). Ross (1999, p. 309) calls these “wealth-induced stupors,” where governments become so excited over

31 Chapter 6 examines Kazakhstan in greater detail. 86 the possibilities offered by resources, and are buffeted by so many demands and opinions

of how to use them, that they outspend what is sustainable. Not every country will fall

into this category. This kind of massive investment, and even going into debt against

future revenues, is much more likely in fuel dependent countries with higher populations

and lower reserves (Karl, 1997, p. 18-19).

These large-scale investment projects pose a problem for political dissent. Where

the government has the primary responsibility for investment, it also has some control

over those investments and the rules that regulate businesses. In the case of well

developed and independent bureaucracies with strong civil oversight, as in Norway, this

does not pose too large of a problem, as the use of state investments for personal or

political gain will encounter strong resistance from the professional bureaucracy,

opposition parties, and the popular press. Transparency, oversight and competition

mitigate the corrupting tendencies of state-led investment. In a state without these

institutions, the situation is very different. Large state investment projects are susceptible

to the tendencies of prebendalism, where investment, jobs, and regulation are structured

for the benefit of public officials, their constituents and their identity groups (Joseph,

1987; Joseph, 1996). To a certain extent, Luong and Weinthal (2006) are correct –

ownership does matter. However, ownership of the primary resource matters less than

how the government’s pre-eminent position in investment translates into control over the general economy.

87 3.2.3 Government distribution and democracy

A simple game theoretic model is enlightening. Below I will lay out, in intuitive

terms, the strategies and interests causing government distribution to undermine

democracy. A formal mathematical treatment is in Appendix C.

The game is played between a government (G) and a latent opposition in society

(O).32 The potential opposition receives a certain portion of their income from money

distributed to them by the government to prevent unrest, and a certain portion from

production independent of government control. In addition, the opposition has some

level of affinity for the government, modeled here as the amount of money they would

hypothetically be willing to surrender to obtain their ideal policy point. Government-

controlled distribution refers to any disbursement, employment or means of production

that the government has the ability to bestow or withhold through direct action. This

includes, for example: welfare programs; employment in state-owned enterprises; buy-

offs; public contracts; indirect ownership (through government connections); and also the

presence of regulations and/or enforcement mechanisms allowing the government to

expropriate otherwise legally-held private property.

The game is played over an infinite time period, with decisions made in the

current period setting up a discounted income stream for the future. The game starts with

the government receiving an amount of revenue from the economy, from which it assigns

an amount of distribution to the opposition. Government leaders keep revenues that are

not distributed. In addition to the revenue they are able to keep, the government receives

a certain amount of inherent benefit from being in power.

32 Collective action problems in forming the opposition are left undefined for clarity. The opposition can either be thought of as an organized opposition, which the government is interested in buying-off, or simply as a sum of social groups that has potential for challenging the government. 88 After the distribution occurs, the opposition must decide if they want to challenge

(C) or not challenge (NC) the government. This opposition will have a certain

probability, p, of winning, drawn from nature (N). This probability can be based on the

size of the opposition, the presence of outside support, the strength of the government’s

repressive apparatus, challenges to organization, or the success/failure of previous

challenges (see e.g. Boix, 2003).33

If the opposition chooses not to challenge, the status quo will continue

indefinitely. This includes the distribution from the government and the money received

from other production, deflated by their affinity for the current government. If the opposition chooses to challenge, then it receives nothing in the current period and then receives a payoff based on whether that challenge is successful. The loss of payoff in the current period is a simple method for associating some cost to the act of challenging. If the challenge is successful (win), the opposition receives a payoff depending on the type of regime that follows. If the challenge is unsuccessful (lose), the status quo resumes, but the opposition is excluded from future government distribution.

If the challenge is successful, the opposition decides whether to prefer an elected/democratic regime (E), or a non-elected/non-democratic regime dominated by the opposition (NE). If they choose the former, then elections occur, with a certain probability of the opposition (O) winning and a certain probability of the government (G) winning.34 When the opposition wins, they receive a distribution from the opposition-led government in addition to money received from non-government production. As with the

33 For simplicity, the probability of winning is exogenous to government and opposition action, since the primary interest here is on the effects of distribution. 34 The same actors compete in the democratic and non-democratic context in this model. Other actors could be substituted in with similar results, since it is the probabilistic access to distribution that is the important characteristic of democracy. However, this would involve inserting two or more actors into the model and unnecessarily complicate the notation. 89 previous government, this is deflated by their affinity for the policy outputs.35 When the government wins, the opposition receives only the money from non-government production, deflated by their affinity for the government. For simplicity, I assume that the opposition has a general idea of the performance of the economy under a democratic regime. This means that they have already taken into account the policies under probabilistic alternation in power. They therefore look at the general expected amount of money to be received from the economy under a democratic regime as opposed to any particular electoral outcome.

When the opposition chooses an opposition-led non-democratic regime (A), the result is a continual stream of distribution from the opposition, plus the amount of money from the independent economy under this new non-democratic regime, minus the affinity for the opposition’s policy outputs. Table 3.2 shows each potential outcome and the corresponding payoffs, and Figure 3.5 shows the sequence of the game.

35 Even where the opposition controls the government apparatus, it is unlikely that their policy ideal point will be achieved because of institutions, foreign pressures, or historic circumstance. 90 challenge, challenge, challenge, win, no win, government not lose elections democratic steps down challenge (DG, C, (DG, C, win, (DG, C, win, (/) (DG, NC) lose) NE) E) Opposition (O) Income Current and Zero in time Zero in time Zero in time from future t, plus future t, plus future t, plus economy in stream of stream of stream of stream of time t, plus income income income income future from from from under stream of economy economy, distributions democratic income and and G’s minus from economy, distribution distribution, affinity for opposition plus under minus G. and income probabilistic replacement affinity for from the distribution government, G. economy, when minus minus opposition affinity for affinity for wins that the elections, government. opposition. minus affinity for winner of elections in each period. Government Revenues Inherent Zero in time Zero in time Zero in time (G) from time t value of t, plus t, plus zero t, plus holding inherent in the probabilistic office, plus value of future. chance of stream of holding winning revenue left office and office and after stream of keeping distributions revenue left excess have been after revenues made. distributions when G for wins indefinite elections. future.

Table 3.2: Payoffs from the distribution game.

91 Lose Lose (1-p) (1-q)

N N C E 0≤DG≤R Win Win G O (p) O (q)

NC NE

Figure 3.5: One round of distribution game.

A solution for this game emerges through backwards induction, starting with when the opposition will prefer a democratic regime and when they will prefer an opposition-led non-democratic regime. If the opposition believes that the total economy will be more productive under democratic governance, the choice between democratic and non-democratic governance hinges on the probability that the opposition will win in elections and on how important government discretion is for income.36 Figure 3.6 shows how the probability of winning in elections (q) that is necessary for the opposition to choose democracy is related to the percentage of income that is reliant on government

36 If the opposition does not believe that the economy will perform better under democracy, or believes that it will perform better under a non-democratic government, the solution to this game becomes trivial – the opposition will always choose non-democratic outcomes, unless some new variable is added to represent civil liberties or some other value associated with democracy. For these purposes, such an addition is an unnecessary complication. The dynamics of the game will remain the same with the inclusion of additional democratic or non-democratic incentives. 92 discretion.37 The blue area in the figure is the percentage probability of the opposition

winning elections for which they choose democracy. The figure clearly shows that as the

percentage of the opposition’s income that is reliant on government transfer distribution increases, the area in which they choose democracy decreases.

1

0.9

0.8

0.7

0.6

0.5 q

0.4

0.3

0.2

0.1

0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 % Distribution

a – area in blue shows how certain a group must be of winning elections (q) for them to prefer competitive politics (democracy).

Figure 3.6: Effect of the percentage of income that is determined by government decision-making on the preference for democratic outcomes.

37The figure is calculated with the following values: the total amount available to the individual is $1,000 in authoritarian outcomes and $1,200 in democratic outcomes; diO starts at .0 and moves by .1 until it reaches 1; fiO moves inversely and is multiplied by the authoritarian total; fiD moves inversely to diO and is multiplied by the democratic total; and the affinity is set at 100 for the government and 50 for the opposition. 93 The reason for this relationship is quite intuitive. The more that income is reliant on who is in power, the less uncertainty a group is willing to allow in that outcome, even if democracy results in better overall economic performance. This makes the general argument that uncertainty over control of the government is less preferable for groups where governments are responsible for a large proportion of their income.

Backing up a node in the game, we can now calculate the conditions under which the opposition chooses to challenge the government. Economic performance in this node is roughly equivalent under government and opposition non-democratic regimes, but expected re-distribution is higher under opposition rule and, again, expected economic performance is highest under a democracy.38 The key variables of interest are the probability of the challenge being successful (p) and, as above, the percentage of total income that comes from government distribution. Figure 3.7 shows the probability of success that is necessary to challenge for a democratic regime and the probability necessary to challenge for an opposition-led non-democratic regime.39 The area above these two curves is the probability in which the opposition will choose to challenge.

Where the importance of distribution is the lowest, the opposition does not need much probability of success to challenge for a democratic regime. Little is lost from the absence of distribution if the challenge is unsuccessful, making a challenge for the better economic performance posited in democracy worthwhile. The necessary probability of success for a democratic challenge increases rapidly to the point where, when re-

38 This assumption is to provide a clear illustration for the reader, even though these settings will not be realistic for all cases. The actual results are robust for any number of assumptions about economic performance, but the results will often be trivial. In this figure, the opposition's distribution is set .05 higher than the government’s within the range 0 to 1. 39Additional settings for this simulation included setting the individual's discount rate and probability of winning elections. The discount rate was set relatively high, at 0.97, while the probability of winning elections was set at 0.5. 94 distribution makes up approximately 20 percent of total income, the opposition becomes more likely to challenge for the non-democratic outcome. In both cases, the opposition becomes less likely to challenge the current government as the importance of distribution rises. Again, the intuitive interpretation of this phenomenon is that, as the importance of distribution increases, the amount of future income being risked is also increasing, making a challenge more costly.

1

0.9

0.8

0.7

0.6

0.5 p Democratic challenge 0.4 Authoritarian challenge 0.3

0.2

0.1

0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 % Distribution a – the area above the lines indicates the certainty which an opposition would need of winning a challenge (p) for them to challenge the government for each type of regime outcome.

Figure 3.7: Effect of the percentage of income that is determined by government decision-making on the preference for challenging for a democratic regime (blue) or for an opposition-led authoritarian regime (red).

95 If the Figures 3.6 and 3.7 were taken in isolation, the theory would strongly resemble those of the rentier state hypothesis – oil revenues result in government distribution that makes governments more stable. Governments, however, also have a decision to make about whether to make such a distribution or to take all of the current revenue and give up power. Alternatively, although it is not explicit in this game, governments may decide to give in to other opposition demands (such as limited democracy) if it cannot make the distribution required for maintaining office. This decision is heavily dependent on the size of revenues available to the government.

In this game, the government’s decision on how to distribute its revenue will be predicated on how much distribution is necessary to remain in office, the amount of revenues available, and the value of remaining in office. Distributions cannot exceed the government’s total revenue.40 Here the difference between capital-rich and capital-poor countries becomes apparent. In those countries with small populations and large reserves, making the necessary distribution to remain in power is relatively easy because of the loose budget constraint against the relatively lax social demands. For countries with larger populations and smaller resource endowments, this becomes more difficult, and raiding the government coffers (or giving other concessions) when revenues are not enough to cover the necessary distribution becomes more likely.

The government must also take into account the future value of remaining in office. If the value of remaining in office, which includes the amount of non-distributed revenue and the inherent value associated with the office, is less than the value of the revenue to be confiscated by raiding the government coffers and stepping down, then the

40 In reality, governments may have some flexibility in their revenue constraints. They can, for example, borrow against future revenues. This ability to borrow, however, is not unlimited, as was amply demonstrated by Nigeria and other heavy borrowing fuel exporters. Borrowing can therefore enlarge the budget constraint, but it does not eliminate it. 96 government will choose to leave office.

While not modeled directly here, those governments for whom revenues are too

shallow to maintain authoritarian governance, or for whom the value of maintaining this

governance is too low, have the option of either leaving politics (and perhaps going into

exile) or democratizing the political system. This suggests that capital-poor countries are

more likely to undergo democratic transition. As noted above, however, preferences for

democratic outcomes are likely to be very weak so long as dependence on government distribution remains high. Thus, these democracies are likely to be relatively unstable, since each group has an incentive to lock in their share of relatively scarce resources

through exclusive governance.

The reader should draw several main points from this analysis:

• The size of the resource compared to the size of the population for distribution is

critical in explaining overall political outcomes. Where the size of revenue which

accrues to the government from fuel resources is very large compared to the

population, the government is unlikely to change greatly. When the budget

constraints are tighter, the expected pattern is one of instability, and sporadic

attempts at democratization.

• The importance of government discretion in the distribution of income is a

primary influence on the preference for democracy or non-democracy, and for the

propensity to challenge the current authoritarian government. As government

distribution becomes more important, a potential opposition has more to lose from

challenging, and is thus less likely to challenge. It also has more at risk in the

outcome of democratic elections, meaning that it will have less tolerance for

uncertainty in democratic outcomes. 97 • While not fully developed above, several other factors will be important for

stability. For example, the probability of a successful challenge can be influenced

by any number of other variables not explicitly included in the model, such as

group cohesion, access to outside support, and the strength of the government’s

repressive apparatus. In addition, in the simulations for the Figures above the

difference between an opposition-dominated non-democratic government and the

current non-democratic system was relatively minor. In the case of secession or

where strong social cleavages exist between the government and opposition, this

might not be the case. For example, politicians in secessionist provinces may

promise the total economy per capita available for distribution will be much

higher if it were to secede (Ross, 2003). This would also suggest that secessionist

movements would be most likely in the geographic areas where the resource is

concentrated.

• Finally, the probability of winning in democratic election is also important, and

suggests that small minorities represented by the opposition are unlikely to prefer

democratic outcomes. In capital-poor countries, where the government does not

have the resources to structure distribution in a way to forestall the challenge, this

poses a danger for separatist movements and may even lead to civil war.

3.3 Innovations and implications

The most obvious implication from this model is how the size of the endowment matters in the underlying causal pathway that describes the natural resource curse. This is not a completely novel insight. Karl (1997, p.17-19) suggests that there is an underlying difference between what she labels “capital-surplus” and “capital-deficient” 98 countries. Capital surplus countries are those which have relatively small populations

compared to the size of their oil endowments; they “could not possibly absorb all their

revenues and thus ran balance-of-payments surpluses until 1983, when oil prices fell

sharply” (p. 18). Capital deficient fuel exporters, on the other hand, have larger

populations compared to their resource endowment, more diversified economies and a

larger skilled workforce. As such “[t]hey appear to be able to absorb all the oil revenues

from their booms and in fact have generally been net importers of capital, except during the brief period from 1974 to 1976” (p. 18). These latter countries are where shocks are most likely to be felt, as they are much more heavily constrained by the resources available for distribution and the social demands placed on the government.

The major difference here is that, while Karl (1997) excludes capital-surplus

countries from her analysis and only dedicates a few paragraphs to the differences

between these states, the above model suggests that some of the most important dynamics

of the natural resource curse exist only in the difference between capital-surplus and

capital-deficient fuel exporting states. The model also identifies microfoundations for the

argument. Perhaps most importantly, as the next chapter will show, looking at the

systematic differences between these two types of fuel exporters shows that the debate

about the underlying dynamic of authoritarianism in fuel dependent states is largely

illusory.

The underlying causal mechanism in the model above is also important. While I

use the label “government distribution” for control of income by the government, the

reality is that this does not have to be limited to a specific type of government program.

Indeed, any type of policy making individual income more directly linked to decisions of

a particular governing power, and removable if another group controls the government or 99 if the individual withdraws support for the current government, should have the same

effect. What this means is that the focus of the literature above on specific types of

policies such as corruption, welfare spending, military spending, state-ownership,

manipulation of regulation, etc., only grasps part of the underlying cause for the natural

resource curse. The actual policy implemented matters less than the implications of that

policy for political control over income. This also helps to explain the relative success

stories among fuel exporting states. Norway, for example, uses its fuel exports to support

its pension program, but the actual distribution from this program has been supported by

a broad coalition and has not traditionally been reliant on the particular administration in

power, nor is it contingent on the individual supporting that administration.

Finally, the context under which these resources are developed is also important.

Other than their higher than expected per capita incomes, many of these countries look

very much like their less-developed counterparts. Where underlying social tensions

already exist, the geographic concentration of resources will tend to exacerbate them. In

addition, the socio-economic changes that are expected to accompany rising incomes are largely absent from fuel exporting states. Chapter 5 demonstrates that, other than their inflated per capita incomes, fuel exporting countries look and behave very much like developing countries. In particular, the ability for governments to distribute revenues from fuel exports as private (and therefore discretionary) goods makes the development of socio-economic characteristics that might offset social tensions less likely.

100 CHAPTER 4

LONG LIVE THE KING; OR DOWN WITH DEMOCRACY?

The last chapter set forth several empirical implications that could be drawn from the formal theory on government distribution in fuel exporting states. This chapter begins testing these hypotheses in large-N comparisons. First, I will establish the link

between fuel dependence and the importance of government distribution. This includes

not just distribution in a normal sense of direct monetary disbursements, but also

government methods of control over economic distribution through ownership and

regulation.

Second, the chapter will test underlying causal pathways, establishing whether

fuel export dependence and fuel export revenue make authoritarian regimes more stable,

democratic regimes less stable, or both. From the theory developed in the last chapter,

fuel export dependence is expected to be destabilizing for all political regimes, but

especially for democracies, while fuel export income is expected to be stabilizing,

especially for authoritarian regimes. Figure 4.1 highlights the parts of the theory to be

tested in this chapter. While the link between fuel exports and non-democratic

governance was established in Chapter 2, this chapter explores two other links – between

fuel export dependence and the importance of government distribution, and between the

size of fuel exports and political regime stability.

101 Fuel Export Income

Accrues to

Government

When dependent Low Political instability (esp. in resource democracies) Increases role of government income decisions in individual income.

Importance in individual income Political stability (esp. in authoritarian states)

High Decreases preference for resource democratic uncertainty income

Figure 4.1: Parts of theory tested in Chapter 4.

4.1 Linking fuel exports and government distribution

As the last chapter noted, there are a number of reasons to suspect a relationship between fuel exports and increased importance of government distribution in the economy. The last chapter argued that even benign and well-managed governments will have difficulties avoiding this relationship, given the geographic concentration and low employment gains of fuel development. In most regimes, where the motives are not so altruistic, fuel export income provides the means by which to increase government discretion over individual income. Yet, the link between fuel export dependence and the importance of government distribution has only been explored within the relatively narrow framework of social spending (Ross, 2001) or government ownership of oil production companies (Luong and Weinthal, 2006; Quinn, 2006). This section takes a

102 much broader view of government distribution, including the government’s involvement

in economic production and the arbitrary nature of regulations and enforcement of

property rights.

The broader understanding of government distribution is important for

understanding the political-economic connection in fuel exporting states. Focusing solely on social spending only captures one of many mechanisms through which the importance of government decision-making affects individual income, thus missing important aspects of how fuel exports enable greater government discretion. In Kazakhstan, for example, increased fuel exports have allowed the government to increase the pay of public workers and the pensions for retirees, but this is hardly the most important mechanism by which

President Nazarbayev has established his influence on the economy (see e.g. Kennedy,

2006; McGlichney, 2003). Financial resources from fuel exports, along with the relative importance of those resources in the overall economy, have allowed Nazarbayev to finance the control over media, oil, industrial, financial, and even charity organizations by his daughters, sons-in-law, wife, and close colleagues (see Chapter 6).

The greater involvement of government in the economic process has several implications for political opposition. First, so long as people receive an acceptable level of resources from the government-controlled system, they are unlikely to risk their income by supporting the opposition. This risk-averseness to governmental change, when government discretion over distribution is important, was demonstrated in the formal section of the last chapter. Similarly, so long as people are satisfied with the current state of economic distribution, and/or are satisfied with where their personal wellbeing (or the wellbeing of the country) are heading, uncertainty over the behavior of the opposition in distribution makes groups less likely to support government change. 103 Finally, businesses and local politicians under the direct control of the government, or controlled by close associates of the people in power, are less likely to support the opposition or to form their own opposition groups.

The focus on social spending also misses how access to fuel exports changes the incentives for establishing predictable regulation and property rights protections.

Lindblom (1987) argued that market economics function as a prison for leaders. While it allows for greater economic productivity, it also constrains policy-making because decisions which are adverse to business interests result in automatic punishment in the form of weaker economic growth. In fuel exporting states, the incentives are somewhat different. Heavy competition over fuel resources allows the government to arbitrarily enforce regulations against both domestic and foreign firms without as much concern for the effect an unstable business environment has on investment.

This effect has been demonstrated many times during the time period studied in this work. In Venezuela, for example, President Hugo Chavez has reversed many of the favorable oil contracts that his predecessor, Rafael Caldera, established during the low oil prices of the 1990s. Examples of nationalization or increased government taxation are everywhere. Among the most recent nationalized oil development projects are the

Orinoco basin (Elliot, 2007) and Jusepin oil field (International Herald Tribune, 2007).

In just one month, April 2006, the government took control of 32, mostly small, foreign- managed oil fields (Forero, 2006). Even when outright nationalization does not occur, companies face uncertainty in the levels of taxation they will face year-on-year

(Mortished, 2006). Despite this, oil companies are unlikely to stop investing in

Venezuelan fields. As a Deutsch Bank report stated, “[F]ew are likely to turn their backs on reserves of the order seen in Venezuela” (quoted in Forero, 2006, pp. C1). Similarly, 104 while increased government involvement in Kazakhstan and Russia’s oil sectors have

made investors more cautious, levels of foreign direct investment remain substantially

higher than in most of their neighbors, and have increased in Russia’s case.

Companies have a number of reasons for continuing their investment. As noted in

Chapter 2, rising oil prices have increased company profits substantially, also increasing

the amount of risk they are able to tolerate. At the same time as the resources for

investment have increased, the common perception is that the pool of oil resources is

declining (e.g. Deffeyes, 2001; Williams, 2003; Duncan, 2003). Thus, being left out of a

large new oil find is more risky than the potential losses due to a nationalization or

increased taxation. Nationalization is also not a total loss for oil companies. Companies

usually receive some type of compensation if a field is nationalized. In the case of

Venezuelan nationalizations, for example, companies have received large compensation

for loses, either in cash or with vouchers for future investment (e.g. International Herald

Tribune, 2007). Finally, if a company refuses to deal with governments who do not provide a stable business environment, there are many others who are willing to take their place. Indian and Chinese companies, in particular, appear willing to do business in a number of areas without predictable regulation or stable political environments.

An example of many of these trends, which will be developed in greater detail in

Chapter 6, is that of PetroKazakhstan. Despite numerous government lawsuits, fines, and even arrest warrants, the company made a large profit from its work in Kazakhstan. The stock price for the company went from only a few dollars a share in 1999 to $55 a share when it was sold to the Chinese National Petroleum Company (CNPC) in 2005 after a bidding war with India’s ONGC (PetroKazakhstan, 2005). The success of the company is reflected in the attitudes of the employees I spoke with shortly after the takeover. Of 105 the five interviews that I conducted with former PetroKazakhstan executives, one stayed

on with CNPC, one wanted to get involved with the next investment project of

PetroKazakhstan’s former owner, one wanted to get involved with another project just

like PetroKazakhstan (with the logic that he had gotten in too late to make the large

money on the venture), one was ready to retire or take a break from work, and one did not

know what he would do at the time I interviewed him.41

All of these mechanisms increase the importance of government discretion in

individual income, and, because fuel exports are a high rent industry in which revenues

accumulate primarily to the state, we would expect those countries that are more dependent on fuel exports to also be the countries where we see government distribution

playing a larger role in the economy.

The corresponding increase in the role of the government in economic distribution

violates one of the key implicit assumptions in traditional democratization theory. Much

of the democratization literature explicitly or implicitly assumes that economic interests

act with some independence from the government. While not usually explicit, this

assumption is in line with Marxist analysis, which assumes that political outcomes are the

result of underlying class interests. This is made explicit in Moore’s (1966) classic

analysis of social development. He argues that the development of democracy is the

result of class interactions. In particular, a large and independent business class is key for

democratization. Thus, the more complete version of Moore’s (1966, p. 418) famous

quote reads, “Without going into further evidence…we may simply register strong

agreement with the Marxist thesis that a vigorous and independent class of town dwellers

has been an indispensable element in the growth of parliamentary democracy. No

41 Names withheld at request of interviewees. 106 bourgeoisie, no democracy” (emphasis added).

Lipset (1959; 1960) and some other modernization scholars make similar

assumptions in positing a positive relationship between economic development and

democracy. They argue that economic development results in the emergence of an independent middle class, as well as other independent groups that are capable of making greater demands on the state. Two groups, in particular, receive attention. The “middle class” is identified as a moderating force for politics (Lipset, 1960, p. 51), and independent social groups (intermediary organizations) press for greater representation

(p. 52). For Lipset (1960) these demands eventually result in democratization and the stability of democracy. Similarly, Huntington (1968) argues that economic growth

results in the development of interests who press their claims against the current

government, although he believes that this process is likely to be destabilizing without

concomitant strengthening of institutions. Reuschmeyer et. al. (1992) make a similar

argument explaining the link between economic development and democracy, but focus

less on an independent bourgeoisie and more on the importance of the working class.

The latest incarnation of this literature deals primarily with inequality as an

explanation for the dynamics of political change. Boix (2003; see also Boix and Stokes,

2003) argues that increased equality and capital mobility make democracy a more likely outcome, as they decrease the redistributive effect of democratic governance. Acemoglu and Robinson (2003) make a somewhat similar argument about inequality. They, however, separate out the effects of inequality on the probability of democratic establishment and its effect on the durability of democratic governance. They argue that increased inequality increases the chance of authoritarian regimes becoming democratic because discontent among the masses will cause elites to democratize in order to avoid 107 mass unrest. Tax rates, however, are also likely to be higher in democratic states,

resulting in a danger to democratic stability, due to elite preferences to avoid large

redistribution. This means that democracies are more likely to remain stable where

inequality is relatively low.

Dunning (2007) and Morrison (2007) have extended on the work of Acemoglu

and Robinson to suggest that fuel export revenues are stabilizing for governments. Fuel

export revenues provide for lower income taxes in democracies and greater social welfare

programs in non democracies, offsetting the destabilizing effects of elites and masses respectively in both types of regimes.

One very important question, however, remains unanswered -- what happens when government distribution plays a key role in determining both the elite and non-elite income? In other words, what happens when government distribution is tied to status, as opposed to being additional to it? In states which are highly dependent on fuel export income in the economy, reduced taxation is offset by an elite that is already dependent on fuel windfalls for their position. The traditional redistributive framework of analysis breaks down where the primary role of government is not only to adjust for unequal initial distribution of goods by the economy, but also has an important role in the initial distribution. Thus, those states without resources from fuel exports for both the initial distribution and the additional distribution may be made more vulnerable by their dependence on fuel exports.

The major contention is that theories of democratization focusing on social distribution break down in oil dependent states because the state can use these resources to buttress its own stability, and the state’s distribution is likely to be central to determining social distribution. This, however, does not cover the broader socio- 108 economic or attitudinal dimensions of modernization theory, which will be discussed in greater detail in Chapter 5 (see ex. Deutsch, 1961; Deutsch, 1991; Diamond, 1992).

To capture the different aspects of increased importance of government discretion in the economy, several different measures are utilized. Because of the variety of policies involved in the importance of government distribution, any statistical tests are going to be indicative rather than conclusive. As Moore (1966, p. 520) warns, statistical measures may be a poor indication of the economic relations among actors. These measures are also likely to be very noisy; capturing aspects of the economy other than what is intended. To overcome this, several different measures are used. In all of the models discussed below, however, the relationship between the increased importance of fuel exports in the economy and the increased importance of government discretion supports the foundation of the theory, and gives some indication that the relationship between fuel dependence and the importance of government discretion in economic allocation is generalizable.

4.1.1 Government consumption

The most obvious method for measuring the importance of government distribution in the economy is to look at government consumption as a percentage of total production (GDP). These direct transfers of funds have been the primary focus of previous literature (see e.g. Ross, 2001; Jensen and Wantchekon, 2004). Table 4.1 regresses, using random effects panel estimation, fuel export dependence and fuel export income against several measures of government consumption, controlling for per-capita

GDP. Model 1 looks at the World Bank (2003) measures of general government final consumption expenditure as a percentage of GDP as reported in the World Development 109 Indicators (WDI).

Model 2 uses the measures of government consumption from the Economic

Freedom in the World index (EFW) (Gwartney et. al., 2005). This measure has a similar interpretation to the WDI indicator, but utilizes different data sources. Data is collected utilizing a worldwide network of scholars and organizations, and is indexed on a scale from zero to ten, with ten indicating “most free” – in this case less government consumption. One of the most valuable aspects of the EFW index is that, for many countries, it can be calculated back to 1970. It does have some pitfalls, however, in that the data is reported in five year intervals until 2000 (necessitating some interpolation) and data is missing for several important fuel exporters. This makes the index valuable for its temporal coverage, but less valuable in terms of spatial coverage.

Model 3 looks at the level of price control exercised by the government, as measured by the Economic Freedom Indicators collected yearly by the Heritage

Foundation (HF) (Miles et. al., 2006). This dataset has the opposite limitations of the

EFW data. It covers a much larger sample of countries, but its time series starts in 1995, allowing only six years for analysis. This data is indexed in a similar manner to the

EFW, with two differences. First, the cut-points are based on researcher decisions about the data, instead of on relative ranking. Second, the scale ranges from 1 to 5, with 1 indicating the greatest level of “economic freedom” (i.e. the least government involvement). Price controls are utilized as a proxy for government welfare, since they often represent efforts by the government to offset pressures on the price of basic items and prevent associated social protest.

In addition to the measure of fuel dependence introduced in Chapter 2, all of the models for government control of economic distribution presented here include controls 110 for the level of economic development of the country, to take into account the total size

of the economy, and the growth in GDP of the country year-on-year, to account for

spending changes based on increases in state resources. All of the models are also

limited to the sample of developing states, so as to ensure that the dependence results are not simply an artifact of comparisons between developed and developing states.

Model 1: Model 2: Model 3: Government Government Price Consumption Consumption Controls (WB) (EFW) (HF)

Fuel Dependence .042 (.040) -.026 (.003) .007 (.006) Ln(Per-capita GDP) .461 (.367) -1.121 (.017) -.607 (.000)

Growth in GDP -.109 (.000) .012 (.020) .0003 (.459)

R2 Within .027 .009 .001 R2 Between .083 .375 .232 R2 Overall .050 .289 .194

N 3202 2466 702 Groups 114 89 110

a – values are the weighted average of the within and between coefficients with p-values (1-tailed), calculated from robust clustered standard errors, in parentheses.

Table 4.1: Relationship between fuel export dependence and government consumption in developing countries.

111 Table 4.1 provides some support for the argument that fuel dependence increases

the amount of social spending by the government. In all three measures, fuel dependent

countries have higher government consumption and social welfare spending than would

otherwise be expected. Model 1 shows that, compared with countries of similar sized

economies, more countries in which fuel exports make up a higher percentage of

merchandise exports have higher levels of government consumption, although the

coefficient is relatively small (p = .040). The substantive effect is about a one percent

increase in government consumption in the economy for every 25 percent increase in fuel

export dependence. Model 2 has similar results. Dependence on fuel exports increases

the level of government consumption, causing a decrease in the country’s EFW score (p

= .003). Here, a 40 percent increase in average dependence on fuel exports results in a

one point decrease in the country’s ten point economic freedom score on this dimension.

Finally, Model 3 shows a significant relationship between fuel export dependence and government price controls. The results are well within common levels of statistical significance (p = .006), suggesting that the relationship is not due to chance, and that price controls are higher in more fuel dependent states.

These results demonstrate that there is some relationship between fuel dependence and the percentage the economy made up of government spending. The relationship is weaker, however, than previous scholarship would seem to imply. Part of this is likely due to non-reporting by a number of oil rich states, especially in the EFW measures. This also suggests, however, that the increase in government distribution is not limited to social spending and subsidization.

112

4.1.2 Government ownership

Government ownership is a second pathway through which government

distribution becomes more important in individual income. In developing states, where

independent capital is scarce, government revenues are one of the few sources for large-

scale investment. Karl (1997) argues that government investment is especially likely to

be important for capital-scarce fuel exporters, as a method for promoting sustainable

growth.

The models presented in Table 4.2 tend to support the relationship between

dependence on fuel exports and the importance of government ownership in the economy. All of these models show that countries with a high reliance on fuel exports

(40-50%) have scores indicating 5-10 percent more state ownership in industry and

banking. Model 4 looks at the importance of government enterprises as a share of the economy, as measured by the EFW dataset. The results show a strong correlation between the level of dependence on fuel exports and the importance of government enterprises (p = .000). Substantively, the results suggest that a country dependant on fuel exports for 40 percent of its merchandise exports will score around 1 point lower than a country that does not depend on fuel exports, on the ten point economic freedoms scale for this measure.

Model 5 looks at government intervention in the economy, as measured by the HF dataset. Miles et. al. (2006, p. 7) also note the relationship that oil and gas exports have on the government’s involvement in the economy, although they do not give an explanation for this link. Government intervention is operationalized much more broadly in the HF dataset. It includes “the fiscal burden of the government, which encompasses 113 income tax rates, corporate tax rates, and government expenditures as a percent of

output” (Miles et. al., 2006, p. 56). In this model, fuel dependence has a statistically

significant relationship with government involvement (p = .000). A state with

approximately 40 percent of its merchandise exports from fuel will score about a half

point higher on this scale (suggesting more government intervention), than will a state without fuel exports. Interestingly, if the measure did not include tax rates, the

relationship would likely be stronger, since the effects of fuel dependence may be

confounded by the generally lower, or nonexistant income tax rates in countries relying

on fuel exports (see e.g. Ross, 2001; Chaudhry, 1989).

Models 6 and 7 look at the level of state bank ownership. The goal here again is

to see if fuel dependence or income increase government control over capital distribution.

These models have similar results to those for government enterprise and influence. In

Model 6, which uses the EFW score for bank ownership, the effect of dependence on fuel

exports does not reach commonly accepted levels of statistical certainty ( p = .136), but

the relationship is in the expected direction – fuel export dependence is correlated with

greater state control of banks. Substantively, states which rely on fuel for about 35

percent of their merchandise exports will score about half a point, or about five percent,

lower on the EFW scale of bank ownership (more state ownership of banks). Model 7,

which uses the HF scores, and has more comprehensive coverage of major fuel exporters,

has similar results. These results are much more certain (p = .000), and suggest that a

state which is reliant on fuel exports for about 50 percent of its merchandise exports will

see a half point, or about ten percent, increase (more state ownership in banking) over a

non-fuel exporting state.

114 Model 4: Model 5: Model 6: Model 7: Government Government Bank Bank Enterprise Intervention Ownership Ownership (EFW) (HF) (EFW) (HF)

Fuel Dependence -.024 (.000) .013 (.000) -.014 (.136) .009 (.000)

Ln(Per-capita GDP) 2.560 (.000) -.473 (.001) 1.766 (.004) -.893 (.000)

Growth in GDP .011 (.056) .0009 (.452) .003 (.302) .002 (.287)

R2 Within .047 .023 .008 .000 R2 Between .209 .197 .335 .401 R2 Overall .209 .131 .311 .330

N 2459 702 2034 700 Groups 88 110 87 110

a – values are the weighted average of the within and between coefficients with p-values (1-tailed), calculated from robust clustered standard errors, in parentheses.

Table 4.2: Relationship between fuel export dependence and government ownership in developing countries.

As with the results in the first section, the models presented here provide some

support for the contention that fuel dependence increases the reliance on government distribution in the economy. In this case, the distribution is primarily due to the

government’s direct role in economic production, as opposed to its social spending

policies.

4.1.3 Protection of property rights

The government’s ability to interfere in distribution by the economy can be

promoted by more than just outright government ownership and spending. Weak

property rights promote loyalty by tying more income to government decision-making.

115 There is nothing in oil exports that would explicitly cause weaker protection of property rights; indeed the state may do an excellent job of protecting the property of some influential individuals. The resources accrued to the government through fuel export revenues, however, provide the flexibility for weaker protection and greater interference in the economy. For example, since tax revenues are not necessary for state function, rules can be utilized in a manner that promotes short-term power maximization over long-term income streams. Fuel revenues often also enrich those within the close (often familial) relations with the leadership, increasing the role of close associates of the government in the economy. The models in Table 4.3 look at the relationship between fuel export dependence and government regulations. In both models, fuel dependence is correlated with less stable property rights.

Models 8 and 9 explore the relationship with regards to property rights. In Model

8, which looks at the EFW measure of the stability of property rights, a country which depends on fuel exports for approximately 40 percent of its merchandise exports will score about a half point lower on its property rights index (p = .000). Similarly, in Model

9, a country with a higher dependence on oil demonstrates weaker protection of property rights, although this relationship does not reach commonly accepted levels of statistical certainty (p = .080).

116 Model 8: Model 9: Property Rights Property Rights (EFW) (HF)

Fuel Dependence -.013 (.000) .004 (.080)

Ln(Per-capita GDP) 1.815 (.000) -1.143 (.000)

Growth in GDP .010 (.020) .001 (.436)

R2 Within .062 .000 R2 Between .430 .500 R2 Overall .272 .423

N 1950 702 Groups 89 110

a – values are the weighted average of the within and between coefficients with p-values (1-tailed), calculated from robust clustered standard errors, in parentheses.

Table 4.3: Relationship between fuel export dependence and the protection of property rights in developing countries.

These relationships must be taken with some caution. As will be developed more fully in Chapter 6, it is the arbitrariness, not the average, level of property rights protection that is important. Governments may do a great job of protecting the property rights of those who support the government and have close connections to the leadership.

The real question is how well they protect the property rights of those in opposition and

new entrants into the market. Nevertheless, the above relationships do suggest some

level of generalizability to the theoretical linkages underpinning the theory developed in

Chapter 3.

117 4.1.4 Overall implications

The results of the previous section are more than the sum of their parts. Taken by themselves, the results in the previous three sections may be underwhelming for the reader. Government involvement in fuel dependent countries is only about five to ten percent more than their non-fuel dependent counterparts according to most of these models. When the results are taken as a whole, however, they reveal a pattern of interference in the economy, corresponding with the importance of fuel exports, which increases the general importance of government action in economic distribution.

Previously the increase in government involvement in the economy was operationalized

along a few dimensions. This analysis suggests that there are multiple dimensions of government involvement which are affected by fuel export dependence. In addition, for the numerous reasons mentioned above, it is likely that these patterns are underestimated because of the coverage and operationalization of some measures. Thus, the results seem to be in line with the major assumption of the model from Chapter 3, and helps explain why fuel exporters tend to generally be less democratic.

4.2 Authoritarian stability vs. democratic instability

As discussed in Chapter 3, two underlying causal pathways can be posited for the statistical relationship characterizing the natural resource curse. First, it could be that authoritarian states that are more dependent on fuel exports are more stable. Second, it could be that democratic regimes established in fuel dependent states are more fragile.

This distinction is important. Several scholars have pointed out the importance of distinguishing between democratic establishment and democratic stability (Bermeo,

1992, p. 297; Przeworski and Limongi, 1997). It only makes sense that the same 118 distinction is important for authoritarian regimes. The distinction set forth theoretically also has implications for what is discussed above. If the foundation of the theory, that the government’s role in economic distribution increases with the importance of fuel exports in the economy, then the size of those resources should have an impact on whether or not the political regime remains stable.

As was demonstrated in the last chapter, previous explanations of the natural resource curse have generally focused on one or the other causal pathway, greatly influencing their theories. Also in the last chapter, however, a theory focusing on the general increase in the importance of government distribution was posited. According to this theory, the debate between those who say that fuel exports make regimes more stable and those who say that they make regimes less stable is illusory. Instead it is the size of the resource in proportion to the demands of the population that will cause countries to either be stabilized or destabilized by the increased importance of government distribution accompanying large fuel export revenues. Put simply, fuel export income is generally expected to be stabilizing, especially for authoritarian regimes, and fuel export dependence is expected to be destabilizing, especially for democratic regimes.

4.2.1 Cross-country testing

Smith (2004) is the touchpoint for this debate. His study was the first to explicitly test the relationship between fuel dependence and stability in a large-N context. The results provided support for the contention that fuel exporting states have more stable polities than their non-fuel-exporting counterparts. He argues that the general pattern for fuel exporters is one of stability, with a few high profile exceptions being overly influential in the literature. 119

Karl (1997, p. 17-18), who is one of Smith's main foils, contends that a difference must be made between “capital-deficient” and “capital-surplus” countries. The former have larger populations and more diverse economies, and, because of this, tend to not only absorb all their oil revenues, but to also be net importers of capital. Capital-surplus countries, on the other hand, cannot absorb all their revenues, and continue to run balance-of-payment surpluses, even during bust periods. If, as Bueno de Mesquita et.al.

(2002) contend, surplus resources are key to maintaining power under crises, capital- surplus countries would be expected to be much more stable than their capital-deficient counterparts. As Karl (1997, p.18) put it, “[A]lthough all oil-exporting developing countries are highly dependent on petroleum, this dependence is felt more acutely in capital-deficient countries because their opportunities are so clearly bounded.”

This distinction between capital-surplus and capital-deficient fuel exporters should give some indication of which countries are destabilized by fuel exports and which find stability. The distinction is also consistent with the theory developed in

Chapter 3. However, it has yet to be tested in a large-N, cross-national framework. This will be undertaken in the next section.

4.2.2 Variables for testing stability

The dependent variables are drawn from three datasets that are already familiar to the reader, but the measures utilized will be different. The key dependent variable for these models is not the regime type (democratic vs. non-democratic), but the temporal persistence of the regime. Variables for regime stability are drawn from the Polity IV

(Marshall and Jaggers, 2002) and the ACLP dataset (Alvarez et. al., 1996; Alvarez et. al., 120 1999; Przeworski et. al., 2000). Polity looks at institutional variables to assign

democracy and autocracy scores ranging from 10 to -10. Here a three point or greater change in the Polity score is considered a substantial institutional change, and is coded as a regime change, as is a transition period defined by a lack of stable political institutions

(see Marshall and Jaggers, 2002, p. 16). Based on this, Polity provides the variable

Durable, which gives the number of years since the last regime change, and Transflag, which denotes, with a dummy variable, the first year of a new regime. For the purposes of this study, Transflag is set at t-1 to denote the end year of a regime. Year zero in this data is set at 1800, time of the last transition before 1965 or the earliest year that the country enters the dataset, depending on the country’s historical context.

The ACLP dataset, on the other hand, denotes regimes as a dummy variable of either being authoritarian or democratic. This coding is based on whether there is competition and the potential of losing in elections. The time to changes from democracy to authoritarianism, or vice-versa, is measured in Ager. The base year is generally the same as in the Polity dataset, with some differences due to the definition of regime transition and idiosyncratic differences in start date. Changes in the regimes are noted by

Ttr, and, again, this was set to t-1 to denote year of change, as opposed to year of establishment.

Both of these variables are used because each captures a slightly different aspect of regime stability. The Polity measure looks at institutional stability, not just transitions between democracy and non-democracy. It should give the reader an idea of general stability of institutions within a country, not just of the persistence, or lack thereof, of competitive elections. Its weakness, however, is that it does not draw a bright line between democracy and non-democracy, making it difficult to develop conclusions as to 121 the overall effect of the independent variables on competition, and makes the drawing of thresholds in some of the models below difficult.

The ACLP measure, on the other hand, looks specifically at democracy- authoritarian transitions. Therefore, it provides a more direct measure of the stability of authoritarian or democratic forms of rule in these countries, and clear transition points for some of the duration models where needed. On the other hand, by only focusing on democracy and authoritarianism, it often misses significant regime changes that are authoritarian to authoritarian, such as the 1979 revolution in Iran (Przeworski and

Limingi, 1997, p. 160, fn. 12).

Several control variables are utilized, depending on the model being analyzed.

Growth is the standard measure of percent growth in GDP year on year (World Bank,

2003). That countries are less stable during periods of stunted or declining growth is one of only two stylized laws in comparative politics (e.g. Geddes, 1999). The other is that countries with higher levels of per-capita GDP are more likely to be democratic (e.g.

Reuschmeyer et. al., 1992). Since this study will be working in a duration framework, and per-capita GDP is not expected to have a linear effect on stability across democracies and non-democracies, the log of this variable is only included in the models of democratic stability and civil war (see Przeworski and Limongi, 1997; Boix and Stokes,

2003).

Another variable only used in a few models is a dummy variable indicating whether the country-year is after the end of the Cold War (PostCW). The end of the Cold

War in 1991 had repercussions for the stability of authoritarian regimes, not just in the former Soviet Union, but also for all of the outside regimes dependant upon Soviet funding or on the Communist model for legitimacy (see Huntington, 1991; Fukuyama, 122 1989). Thus, this dummy variable was included in models of general regime stability and

authoritarian stability.

Percentage of the population that is urban (Urbpop) is often seen as a

destabilizing factor for regimes. Urban populations tend to demand greater resources and

have an easier time overcoming the collective action costs of mobilizing. Finally, ethno-

linguistic factionalization (ELF), measured as the probability that two randomly chosen

individuals in a country would not speak the same language (Roeder, 2001), is used to

control for ethnic tensions that might endanger regime stability (see Horowitz, 1985;

Roeder, 1999).42

4.2.3 Statistical methodology

This study uses a Cox proportional hazard duration setup for evaluating how fuel

export dependence affects stability. This setup uses time as the dependent variable and reports hazard ratios for each independent variable (see Box-Steffensmeier and Jones,

1997; 2004). Scores above 1 indicate that failure is more likely to occur, while scores below 1 indicate that failure is less likely. Substantively, the hazard ratios can be interpreted as the percentage increase or decrease in likelihood of failure against normal likelihood (1).

There are several reasons for preferring this to a panel data setup. First, by using time as the dependent variable, this model avoids duration dependence problems.

Huntington (1968) suggests that this is important because older regimes are more institutionalized. On the contrary, Kennedy (2007) offers some evidence that the opposite is the case in non-democracies, where leadership turnovers and age in the post-

42 As with any study that utilizes TSCS data there will be some cases that are not included because of lack of data. The number of cases is noted in each model. 123 World War II era generally result in legitimacy problems.43 Under either scenerio, the

duration framework allows this to be taken into account and modeled. Second, it ameliorates some of the issues of back-censoring. While there will be some censoring of

regimes that collapsed prior to 1965, using time as a variable solves the issue for regimes

that were established prior to the start of the dataset. The result is that while the dataset

itself covers 113 developing countries from 1965 to 2001, the actual time at risk is much

higher.44 Finally, the Cox framework is preferable to other duration models in this case, because the underlying hazard ratio is not clearly defined in the literature. Therefore,

using a parametric model, which imposes a structure on the data, is uncalled for and

might produce misleading results (see Box-Steffensmeier and Jones, 1997; Jones and

Branton, unpublished). For all of these reasons, the Cox model has been the workhorse for most research into regime and leadership duration.

Cox models have one other important aspect for analysis. The test assumes that the effect of variables are proportional with time – that the effect of a level of a variable at time t is the same as at time t+1. This is not always the case, especially in social science data. There is no a-priori reason to believe that the effect of fuel dependence at the start of a political regime’s duration will be the same as later in its duration. To test

this, all of the models were run with the standard variables and then the Schoenfeld

residuals of those models were evaluated using Harrel’s Rho. Those which demonstrated

significant disproportionality (p < .05) were interacted with the natural log of time and

43 Indeed, stratification based on democracy/authoritarian thresholds, so as to take into account differing hazard rates does have the effect of strengthening the results in Table 2. This is left out of this presentation, however, because these differences are explicitly tested by using separate models for democratic and authoritarian regimes. 44 Smith (2004) limits his analysis just developing countries. I used both setups and found minimal difference between models, other than a slight increase in standard errors in models which only looked at developing countries. 124 the model was run again (Box-Steffensmeier and Zorn, 2001; Box-Steffensmeier, Reitter,

& Zorn, 2003).

All of the models reported here also utilize robust clustered standard errors to account for heterogeneity in the standard errors based on unspecified country characteristics.

4.2.4 The effect of fuel exports on regime stability

Table 4.4 gives the results of the duration analysis of all regime types. The first two models show inconsistent support for Smith’s (2004) contention that fuel exporting states are more stable. The results of Model 10 give some very weak support to Smith's

(2004) contention that fuel exporters are more stable. While neither fuel dependence nor its interaction with time are statistically significant (p = .404 and p = .448), the hazard ratio indicates that transitions, as measured by the ACLP dataset, are somewhat less likely as fuel exports increase as a percent of total exports for the first five years and are increasingly destabilizing thereafter. Fuel dependence is similarly not significant in

Model 11 (p = .460) and the hazard ratio suggests that political regimes are .04 percent more likely to collapse for every one percent increase in dependence. While not reported here, models which do not take into account disproportionality yield much stronger support Smith's (2004) contention, emphasizing the importance of testing for disproportionality in event history models.

When the control for per-capita fuel export income is included the results change dramatically. In all of these models, fuel export income is stabilizing and fuel export dependence is destabilizing. Fuel export dependence and its interaction with ln(t) are within acceptable bounds of certainty in Model 12 (p = .000 and p = .001) and fuel 125 dependence comes close in Model 13 (p = .141), but both models show that fuel dependence increases the probability of regime transition under both methods of

measurement. A one percent rise in fuel dependence initially makes a political regime 10

percent more likely to collapse, according to the ACLP definition, and this decreases

slowly over time. Within the Polity definition of regime change, a ten percent increase in

dependence results in a 4 percent increase in the chance of a transition.

Per-capita fuel export income is statistically significant in both models. The hazard ratios suggest that a one dollar increase in per-capita income from fuel exports

initially decreases the likelihood of transition in the ACLP framework by 3.5 percent (p =

.007), and this effect dissipates slowly over time (p = .006). Similarly, a ten dollar

increase in per-capita income from fuel exports results in a .5 percent decrease in the probability of a substantial institutional change (p = .015).45

45 These results are not simply due to overly high collinearity. The mean VIF in these models is 2.00, with the log of per-capita GDP and percentage of population in urban centers being the two largest offenders. The respective VIF scores for fuel dependence and fuel income are 1.36 and 1.41. These are well below even the more conservative levels for collinearity concerns. 126 Model 10: Model 11: Model 12: Model 13: ACLP Polity ACLP Polity

Fuel Dependence .995 (.404) 1.0004 (.460) 1.100 (.000) 1.004 (.141)

Fuel Dependence * 1.001 (.448) --- .978 (.001) --- ln(t)

Fuel Income ------.965 (.007) .9995 (.015)

Fuel Income * ln(t) ------1.009 (.006) ---

Ln(Per-capita 25.366 (.000) 122.968 (.000) 119.522 (.000) 167.100 (.000) GDP)

Ln(Per-capita .281 (.000) .066 (.000) .237 (.000) .064 (.000) GDP) * Ln(t)

Growth in GDP .972 (.005) .965 (.000) .964 (.003) .966 (.005)

Urb Pop 1.002 (.409) .915 (.000) .993 (.252) .914 (.000)

Urb Pop * ln(t) --- 1.047 (.000) --- 1.046 (.000)

ELF 1.001 (.420) 1.001 (.382) 1.002 (.419) 1.002 (.326)

Post CW .850 (.299) .665 (.183) .932 (.410) .592 (.127)

Post CW * ln(t) --- 1.236 (.116) --- 1.282 (.083)

N 3182 2933 3177 2928 Failures 87 170 87 170 Groups 110 113 109 112 Log-likelihood -501.353 -908.530 -482.483 -905.668 Wald-Test 111.98 (.000) 288.43 (.000) 109.96 (.000) 276.96 (.000) a – values are hazard ratios with p-values, calculated from robust clustered standard errors, in parentheses. All tests are 1-tailed, except those for the Wald-Test.

Table 4.4: Effect of fuel exports on the stability of all regime types in developing countries.

127 These results are generally repeated when looking separately at the survival of

authoritarian and democratic regimes. Examining only those regimes coded as

authoritarian in the ACLP dataset or with scores less than or equal to 7 in the Polity

dataset, the effects of fuel export dependence are still heavily influenced by per-capita

fuel income.46 In Table 4.4, the first two models appear to support Smith’s (2004)

contention that fuel export dependence is stabilizing for authoritarian regimes in

developing states. In Models 16 and 17, where fuel income is included, the results for

fuel dependence reverse. In Model 16, a one percent increase in fuel dependence is

associated with a 2.5 percent increase in authoritarian instability (p = .072), while, in

Model 17, it is associated with a .4 percent increase in the instability of authoritarian

institutions (p = .196). Instead, it is per-capita fuel export income that has a consistently

positive effect on the survival of authoritarian regimes, with a dollar increase making

them 1.3 percent less likely to transition to competitive politics according to the ACLP data (p = .009) and .1 percent less likely to experience an institutional transition in the

Polity data (p = .000).

46The 7 threshold in the Polity scores has been the standard threshold in a number of studies (see e.g. Kadera et. al., 2003). 128 Model 14: Model 15: Model 16: Model 17: ACLP Polity ACLP Polity

Fuel Dependence .928 (.023) .995 (.185) 1.025 (.072) 1.004 (.196)

Fuel Dependence 1.015 (.030) ------* ln(t)

Fuel Income ------.987 (.009) .999 (.000)

Ln(Per-capita 513.066 (.000) 518.401 (.000) 3206.42 (.000) 1553.600 (.000) GDP)

Ln(Per-capita .184 (.000) .056 (.000) .159 (.000) .047 (.000) GDP) * ln(t)

Growth in GDP .977 (.031) .967 (.027) .972 (.026) .967 (.010)

Urb Pop .993 (.284) .890 (.000) .987 (.140) .884 (.000)

Urb Pop * ln(t) --- 1.053 (.000) --- 1.054 (.000)

ELF .999 (.448) 1.009 (.216) 1.003 (.344) 1.015 (.132)

ELF * ln(t) --- .998 (.350) --- .997 (.268)

Post CW 3.261 (.000) 1.837 (.092) 3.645 (.001) 1.440 (.209)

Post CW * ln(t) --- .939 (.360) --- 1.021 (.452)

N 2170 2187 2165 2182 Failures 57 151 57 151 Groups 92 99 91 98 Log-likelihood -287.941 -749.018 -270.928 -738.672 Wald-Test 58.64 (.000) 302.68 (.000) 77.84 (.000) 241.00 (.000) a – values are hazard ratios with p-values, calculated from robust clustered standard errors, in parentheses. All tests are 1-tailed, except those for the Wald-Test.

Table 4.5: Effect of fuel exports on the stability of authoritarian regimes in developing countries.

129 The flip-side of these results, and the test for the second causal pathway of the

natural resource curse, is presented in Table 4.6. Only regimes coded as democracies by

the ACLP dataset are included in models 18 and 20, and only those with Polity scores

greater than 7 are included in models 19 and 21. In the first models, there is some

suggestion in the coefficients that fuel dependency makes democracies more fragile, but these results do not reach commonly accepted levels of statistical significance (p > .200).

When fuel income is added in models 20 and 21, the relationship between fuel dependence and democratic fragility becomes much stronger. The hazard ratios suggest that a one percent increase in dependence on fuel exports initially makes a country 2.9 percent less likely to experience a transition to authoritarianism in the ACLP dataset (p =

.124), but this quickly changes over time. After about three and a third years, dependence becomes destabilizing, and this effect increases with time (p = .074). A one percent increase in dependence is also associated with a 3.2 percent increase in the likelihood of experiencing substantial institutional alteration in the Polity data (p = .030).

Fuel income, on the other hand, seems to have a generally stabilizing effect on democracies as well as authoritarian regimes, but the certainty of these results is not up to normally acceptable levels. The hazard ratios in Model 20 suggest that fuel income is initially destabilizing (p = .458) but quickly become stabilizing with time (p = .293). In

Model 21, the hazard ratio for fuel income suggests that each dollar makes a democracy

.3 percent less likely to experience a substantial institutional change (p = .140).

130 Model 18: Model 19: Model 20: Model 21: ACLP Polity ACLP Polity

Fuel Dependence 1.002 (.386) .986 (.235) .971 (.124) 1.032 (.030)

Fuel Dependence 1.007 (.263) 1.024 (.074) --- * ln(t)

Fuel Income ------1.001 (.458) .997 (.140)

Fuel Income * ------.998 (.293) --- ln(t)

Ln(Per-capita .075 (.002) .035 (.008) .080 (.004) .039 (.007) GDP)

Growth in GDP .934 (.016) .935 (.017) .934 (.013) .923 (.008)

Urb Pop 1.016 (.214) 1.015 (.337) 1.015 (.244) 1.003 (.470)

ELF 1.082 (.002) 1.052 (.083) 1.089 (.003) .963 (.017)

ELF * ln(t) .955 (.000) .961 (.001) .951 (.000) ---

Post CW .132 (.000) .446 (.175) .123 (.000) .911 (.450)

N 1012 746 1012 746 Failures 30 19 30 19 Groups 67 58 67 58 Log-likelihood -132.068 -83.638 -130.308 -98.194 Wald-Test 105.25 (.000) 74.30 (.000) 115.59 (.000) 45.15 (.000) a -- values are hazard ratios with p-values, calculated from robust clustered standard errors, in parentheses. All tests are 1-tailed except those for the Wald-Test.

Table 4.6: Effect of fuel exports on the stability of democratic regimes in developing countries.

131 Comparing the results of these models suggests that fuel export income is more stabilizing for authoritarian regimes. Under the Polity definition of regime stability, fuel income is twice as stabilizing for authoritarian regimes than for all regime-types. A $10 increase in fuel export income makes all regimes about .5 percent more stable, while that same increase makes authoritarian regimes 1 percent more stable. The effect of fuel export income on the stability of authoritarianism or all regime types, under the ACLP definition, is compared in Figure 4.2. The results are a little more difficult to interpret, both because of the smaller number of transitions in the ACLP dataset and because of the disproportionality of the main independent variables over time, but they generally support the same conclusion. While the stabilizing effect of fuel income on all regime types, in red, is larger at first, it quickly dissipates, and approximately eleven years, fuel income is more stabilizing for authoritarian regimes, in blue.

132 0 -.01 -.02 -.03 PercentageChange in the Hazard -.04 0 10 20 30 40 50 time

fuel_income_authoritarian fuel_income_overall

Figure 4.2: Effect of fuel export income on the expected hazard of regime failure in authoritarian regimes (blue) and all regimes (red) in developing countries.

Fuel dependence, on the other hand, appears to be more destabilizing for

democracies. The results for the Polity measure of regime stability are unambiguous.

The effect on democracies was the only one to reach commonly accepted levels of statistical significance and the hazard ratio is eight times larger. Figure 4.3 gives the comparisons of fuel dependence’s effect on democracies, non-democracies and all regimes under the ACLP definitions. While, initially, fuel dependence is more destabilizing for authoritarian regimes and for all regime-types, the destabilizing effect on democracies eventually becomes greater. After about ten years the destabilizing effect on democracies is greater than that on authoritarian regimes, and after about sixteen years it

133 surpasses the effect for all regimes.

.1 .05 0 PercentageChange inthe Hazard -.05 0 10 20 30 40 50 time

fuel_dependence_authoritarian fuel_dependence_democratic fuel_dependence_overall

Figure 4.3: Effect of fuel export dependence on the hazard of regime failure in authoritarian (blue), democratic (red), and all regime types (green) in developing countries.

These results paint a relatively clear picture of what is happening in the data.

Contrary to much of the current debate over the effects of fuel exports, these results suggest that fuel exports can be either a stabilizing or de-stabilizing force, depending on the magnitude of income derived from them and the level of dependence. High levels of per-capita income derived from fuel exports are found to be generally stabilizing, but more so for authoritarian regimes. Dependence on fuel exports, on the other hand, is found to be de-stabilizing, especially in the democratic context. 134

4.3 Effect of fuel exports on civil war

An interesting corollary to the theory is the effect of fuel export dependence and income civil war. That an analogous process would seem to be occurring between political regime stability and civil war seems intuitive, and this may contribute to the quickly expanding literature on natural resources and domestic conflict.

In natural resource studies, an entire literature has emerged about the impact of natural resources, and oil in particular, on civil conflict. A number of authors have argued that there is a strong correlation between the presence of substantial oil reserves, on the one hand, and the initiation, sustenance and severity of civil conflict, on the other

(Collier and Hoeffler, 1998; Bannon and Collier, 2003; Ross, 2003; Ross, 2004; Ross,

2006). These authors cite several implications of oil reserves for civil conflict. (1) Oil reserves make the value of controlling the state greater, resulting in stronger, and often violent, conflict for control of administration over those resources. (2) Oil reserves, or the right to develop such reserves, can be looted to give funding to rebel groups, sustaining and increasing the severity of conflicts (e.g. Angola, see le Billon, 2001). (3)

Oil reserves encourage succession movements in the areas where the development takes place, as politicians utilize promises of wealth to mobilize independence fighters.

Other scholars, however, have criticized this literature, arguing that the theoretical foundations of the theory are weak, that measures of fuel export dependence have problems with endogeneity (i.e. that dependence is the result of economic collapse associated with civil conflict), and that the statistical evidence is ambiguous at best (e.g.

Smith, 2004; Fearon, 2005). Interestingly, in addressing the second of these criticisms,

Ross (2006) makes the further argument that there is no theoretically interesting reason to 135 look at dependence on fuel exports in studying civil conflict.47

The arguments developed above, however, would suggest a very different conclusion. Dependence on point-source primary commodity exports, like oil, is likely to be the most important factor in influencing civil conflict, at least from a motivation and instantiation standpoint, as opposed to per-capita levels of fuel export income. There are several reasons for this. When looking at the reasons for civil conflict, scholars must not only look at the value of holding administrative control over a region, but also the

relative value of holding such control. Indeed, the psychology literature on the

development of identity from realistic resource conflicts heavily emphasizes the role of

scarcity in engendering this competition (Monroe et. al., 2000). This would suggest that

competition and separatist conflict is most likely where groups are competing for a share

of a relatively scarce resource, i.e. where fuel export resources are very important to

income but are not widespread.

To test this, I measure civil conflict in a similar way to that used in the study by

Smith (2004), which finds no relationship between fuel exports and civil conflict. It

relies on the ordinal rating of civil conflict used in Table 3.1. This defines three levels of

civil conflict, based on the number of casualties: minor (25 annual battle deaths and

fewer than 1,000 during the course of the conflict), intermediate (25 annual battle deaths,

a total of at least 1,000 deaths over the course of the conflict, and fewer than 1,000 battle

deaths in any one year) or full civil war (at least 1,000 annual battle deaths) (Gleditsch et.

al., 2002). One important change is made for this study. In addition to measuring the

variable ordinally in Model 22, this study also explores several alternative

47 It should be noted that some, but not all, of the concerns about endogeneity between fuel export dependence and civil war are attenuated by the use of average levels in this study, as opposed to yearly totals. Because they do not reflect short-term fluctuations, the within-country effects caused by the economic impacts of civil war are less likely to drive estimation results (see Appendix II). 136 conceptionalizations of the variable. Model 23 takes all three types of conflicts and

codes them as a dummy variable. Model 24 only codes "intermediate" and "war" events,

and Model 25 only looks at "war" events. The reason for this is simple. One of the

primary differences between Smith's (2004) findings and those of previous scholars

(Coller and Hoeffler, 1998) is their use of different datasets. The CoW dataset, used by

previous scholars, only codes events that are equivalent to "war" in this data (Sarkees,

2000).48 Thus, it is important to take into account how different operational definitions

might influence the results. For example, if fuel resources not only affect the probability

of civil war, but also the length and intensity (see ex. Ross, 2003), then the relationship

between fuel exports and all types of civil conflicts may be weak, at the same time as the

relationship between fuel exports and major civil conflicts is strong.49

As with regime stability, several control variables are included. The log of

economic development is included, as it is commonly thought that greater economic resources attenuate social conflicts. Similarly, economic growth is likely to attenuate

social conflicts, by increasing the opportunity costs to civil conflict. Perhaps the most

common explanation for civil conflict is ethnic factionalization, since these ethnic

divisions provide a base for mobilization of discontent and zero-sum political

competition (e.g. Horowitz, 1985; Rabushka & Shepsle, 1972). Finally, the post-1991

variable, utilized in several of the above models of regime stability, is also included here.

The end of the Cold War is noted by some scholars for marking the beginning of a period

48 The coding is a little more complicated in this regard. While the original definition set the threshold on civil war at 1,000 annual deaths, this criteria has been relaxed recently. The most recent version of CoW does not specifically articulate this threshold, but it still causes some complication in coding. A detailed discussion of this can be found in Sambanis (2004). 49 This argument could actually be taken further. While Smith (2004) argues that it is advantageous to take into account both the instance and severity of civil war in the same model, one could argue that these are two different issues that should be explored separately (see e.g. Ross, 2006). 137 of increased ethnic warfare and civil war in some countries, as the traditional centers of

power collapsed and were re-constituted (Horowitz, 1985; Huntington, 1996; Roeder,

1999).

For estimating the probability of civil war in any particular country-year, a panel

data setup was utilized, both due to the lack of time-based data and because of the lack of theoretical backing for an underlying duration dependency on the presence of civil conflict. Random effects panel data estimation is utilized for the same reasons laid out in

Chapter 2, with panel logit estimation utilized when the dependent variable is bivariate

(see STATA, 2005).

In exploring the effects of average fuel dependence and fuel income on the

probability of civil conflict, several patterns emerge. First, ordinal measures, as

expected, seem to be misleading, suggesting that incidence and severity should be explored separately. Second, consistent with the empirical and theoretical information presented so far, fuel export income makes civil conflict less likely, while fuel export dependence makes civil war more likely. Third, both fuel export income and dependence have their strongest impact in predicting moderate and war events in the data.

In Model 22, the effects of fuel dependence and fuel income on the ordinal measure of civil conflict are not statistically significant (p = .201 and p = .139). They do, however, indicate a similar directional relationship to that in the previous section. Fuel dependency seems to marginally increase the probability and severity of civil conflict, while fuel income has the reverse effect. The low R squared also suggests that fuel exports are a poor predictor within the context of the ordinal measures.

These patterns become stronger in Model 23, which ignores the intensity of conflict to focus solely on incidence of civil conflict. In this model, a one percent 138 increase in fuel dependence increases the probability of civil conflict by approximately

1.1 percent (p = .073). Conversely, a ten dollar increase in per-capita fuel income is also

associated with a two percent decrease in the probability of a civil conflict (p = .004).

Model 24, which looks only at “moderate” and “war” events, supports the contention that thresholds in measuring civil conflicts matter. Fuel dependence increases the likelihood that a country will experience a mid-range or severe civil conflict in a particular year by about 3.9 percent in this model (p = .009). Per-capita fuel income from exports, on the other hand, decreases this probability by about 11 percent for every additional ten dollars of income (p = .006).

Model 25, takes the threshold argument one step further, as it only looks at those coded as "war" incidents, approximately equal to what is coded in the CoW dataset.

Under this operationalization, the relationship between fuel dependence and civil war remains relatively stable. A one percent increase in the dependence on fuel exports

increases the chance of a major civil war by about 3 percent (p = .017), while a ten dollar

increase in per-capita fuel income is associated with a 6 percent drop in the probability of

a conflict (p = .049).

139 Model 22: Model 23: Model 24: Model 25: Civil War Civil War (all) Civil War Civil War (war) (ordinal) (moderate/war)

Fuel .002 .011 .038 .030 Dependence (.201) (.073) (.009) (.017)

Fuel Income -.00006 -.002 -.011 -.006 (.139) (.004) (.006) (.049)

Ln(Per-capita -.144 -.365 .426 -.178 GDP) (.223) (.118) (.076) (.296)

Growth in GDP -.011 -.054 -.067 -.074 (.000) (.000) (.000) (.000)

ELF .003 .014 .014 .014 (.135) (.004) (.004) (.007)

Post CW .122 .648 .652 .202 (.055) (.000) (.000) (.123)

R2 Within .025 ------R2 Between .034 ------R2 Overall .023 ------

N 3339 3339 3339 3339 Groups 114 114 114 114 Log-likelihood --- -988.196 -797.295 -529.347 Wald-Test --- 73.86 (.000) 66.20 (.000) 46.43 (.000) a -- values are the weighted average of the within and between coefficients with p-values, calculated from robust clustered standard errors, in parentheses. All test are 1-tailed except those for the Wald-Test.

Table 4.7: Effect of fuel exports on civil conflict in developing countries, 1965-2001.

140 This analysis lends even greater support to the contention that there is a difference between fuel exporting countries that are capital-surplus and those that are capital- deficient. Fuel export dependence increases the chance of civil conflict, while fuel export income decreases this probability. This effect is largest when looking at the incidence of moderate to severe conflicts. If civil instability contributes to authoritarian governance, then this is part of the explanation for why fuel dependent regimes are less likely to be democratic.

4.4 Evaluating support for the distribution model

This chapter has attempted, to the extent that comparable data has allowed, exploring whether the distributive model developed in Chapter 3 is successful explaining the behavior of fuel exporting countries and the variation among fuel exporting countries.

Overall, the results are quite encouraging. The basis for the theory, that dependence on fuel exports increases the relative importance of government discretion over the distribution of economic goods seems to have been supported. The government’s role in the economy, through social spending, government ownership and selective enforcement of property rights, is demonstrably higher in fuel export dependent states.

Similarly, the main behavioral expectations of the theory – that fuel exporting regimes will either bifurcate into more or less stable camps depending on their dependence on, and size of fuel export revenues – is also supported. Fuel export income is stabilizing to political regimes, especially for authoritarian regimes, while fuel export dependence is destabilizing, especially for democratic regimes. Not only does this impact the stability of political regimes, it also impacts the probability of social conflict.

Fuel dependence breeds resource competition and civil conflict, while high fuel export 141 income provides the resources necessary to attenuate such conflict.

Returning to the overall implications for the natural resource curse literature, these empirical results provide a basis for understanding how the general pattern of democratic underperformance arises out of a wide variety of particular case outcomes. For fuel dependent states, both rich and not-so-rich, the dependence of the country on fuel export revenues increases the importance of government decision-making, resulting in lower tolerance for democratic uncertainty. However, those countries that receive a large amount of income from their exports are likely to maintain relatively stable authoritarian governance, while those that do not receive as much income are likely to oscillate between unstable authoritarian and democratic states. Taken together, these two patterns produce the underlying governance relations that drive so much scholarly research and policy-maker interest.

142 CHAPTER 5

FUEL REVENUES AND DEVELOPMENT

Upon arriving in Almaty, Kazakhstan, one of the first things that a visitor notices is the cars. The streets are crowded with new Mercedes Benz, Cadillac, and Hummer vehicles and Russian-made Ladas are quickly becoming an endangered species (Marcus et. al., 2006). Streets which were made to handle three lanes of traffic now have four, as the infrastructure strains to cope with the doubling of traffic on the streets of Almaty since 1991.50 When combined with the new office buildings springing up all around the city, and the Western-style hotels being built for the large oil conferences and foreign businesspeople who regularly come to the city, it is tempting to conclude that the influx in oil revenues and foreign direct investment (FDI) has led to a boom in socio-economic development in Kazakhstan. Such an impression is further reinforced if one visits the capitol city of Astana, where most of the buildings are of very recent construction, and where jobs and available apartments seem to exceed the population.

If one embarks on a forty minute run outside of Almaty, however, a different picture emerges. Not far from the Al Farrabi University campus, along the dirt paths that lead into the mountains are houses connected by dirt roads, in desperate need of repair.

Further up the mountains, many teachers, police officers and manual laborers from the

50Actually, nobody knows exactly how much traffic in Almaty has grown since independence in 1991. In an interview with Robert Robertson, dean of the Bang College of Business, I was told that the number of vehicles in the city has grown by an estimated 129 percent. However, there are no official numbers to confirm this estimate. In addition, the number may be even greater because of unregistered cars that come in from outside the city. 143 city live in homes without running water or electricity. And, in a country that is as large

as the entirety of Western Europe, some remote villages are still inaccessible during the

winter months. The Council on Foreign Relations, a US-based think tank, noticed a

similar pattern, pointing out that while national poverty rates had decreased from about

30 percent to 20 percent from 2001 to 2003, and the rate fell to nearly zero in Astana and

Almaty, outside of these two major cities 30 percent of people still subsist below the

poverty line of $2 a day. It has also been noted by some agencies that the gap between

rich and poor is increasing (IRIN, 2006).

In some ways, this situation is emblematic of larger patterns in the natural resource curse literature. One of the primary drivers of the interest in the politics of fuel

exporters is their role as an exception to “modernization theory.” This theory, usually

characterized simply as the positive correlation between economic wealth and

democracy, has been one of the most durable in the social sciences (Lipset, 1959;

Diamond, 1992). In fact, Geddes (1999; see also Reuschmeyer et. al., 1992)

characterizes it as one of only two commonly accepted facts in the democratization

literature.51 Oil rich countries, however, seem to defy this theory. Crystal (1989, p. 428)

sums up the role modernization theory plays in the interest of scholars: “In the Gulf,

where the key variable [of modernization theory], capital, was available in abundance, modernization theories should have the greatest predictive capacity. Yet here they failed

most clearly.”

A closer look at the modernization literature refutes Crystal's characterization of

capital as the dominant factor in the theory. While per-capita GDP has been the most

prominent variable used to take into account the level of a country's socio-economic

51The other is that downturns in growth cause instability. 144 modernization, the main authors in the modernization literature describe it as an indirect

indicator of general social phenomena (Lipset, 1960; Diamond, 1992; Deutsch, 1991;

Lerner, 1958; Rueschemeyer et. al., 1992). These scholars suggest that it is increased

education, decreased inequality, reduction in class conflict, and increased individual

wellbeing that are the true indicators of a country's socio-economic modernization. It is

the possibility of increased wealth not contributing to these intermediate variables that has led some scholars to separate the concept of wealth, as per-capita GDP, from development, as improvement in concrete social conditions (Halliday, 1979). In this regard, the natural resource curse literature has largely been guilty of testing the effect of oil against an extremely reduced form of modernization that is economically reductionist and can be summarized simply by per-capita GDP and/or the country’s pool of capital. If modernization theory has been guilty of not taking into account how economic growth

also increases government capacity and, potentially, groups in society that are tied to

continuance of authoritarian politics, then the natural resource literature is equally guilty

of not exploring the social implications of oil-driven growth.

This naturally raises the question -- how much of an exception are oil dependent

states? Are they significant aberrations from established theory, as most of the natural

resource curse literature suggests, or are these cases just developing countries with higher

than average per-capita GDPs? Chapters 2 and 3 argued that fuel export income gives

the government increased discretion over the distribution of economic resources,

allowing them to distribute resources to key groups in society. This chapter argues that

the increased capacity for the distribution of private goods, along with some inherent

characteristics of oil-driven growth, results in lesser provision of public goods, such as

health care and education. Thus, those states with high fuel incomes significantly under- 145 perform in the provision of public services compared to what is expected from their level of economic wealth.

In oil exporting states, a number of factors work against the effective provision of public goods. First, the control by the government of economic distribution, outlined in previous chapters, encourages the provision of private goods, as patronage and corruption, over the provision of public goods such as education and poverty reduction.

Politicians, whose primary motivation is keeping power, will prefer the distribution of private goods, which can be awarded and withdrawn based on political support. Second, the tendency to view oil revenues as windfalls, rather than public money, causes them to be utilized for a number of wasteful projects. Instead of being invested, they are often distributed in the form of subsidies to politically important groups. Further, the investment that does take place often goes towards overly ambitious pet projects of political leaders, favored for their drama over their practical impact. Third, during times of decreasing prices, governments face hard choices between continuing development projects or continuing politically important patronage and subsidization projects.

Politically motivated leaders will often favor the latter over the former, magnifying the effect of oil revenue reliance on socio-economic development. Finally, even when investment does take place, there is a natural lag between this investment and the effect on social indicators. Thus, with uncertainty in revenues and a propensity towards politically motivated spending, oil exporting states would be expected to lag behind other states during this time period, compared with their exploding levels of per-capita GDP.

The empirical evidence presented in this chapter demonstrates that there is a general pattern of under-performance in the improvement of social indicators, compared with levels of per-capita GDP. In particular, it demonstrates that education, poverty 146 reduction, and improvements in social wellbeing lag behind expectations in fuel

exporting states. Even more interestingly, these statistics may mask as much as they

reveal. Several major fuel exporters do not report the demographic information on

poverty, raising questions of why this selection takes place. The estimates are also weakened by the recent rise of several former Soviet states, whose high scores on many

of these social indices are a product of institutional legacies, as major oil exporters.

Similarly, the effect of fuel dependence is likely underestimated, since the instability of

these states results in poor performance both in economic growth and social

development. Finally, the status of the large groups of migrant workers in these states,

especially in the Gulf states, is unclear.

This finding calls into question the usual framing of the natural resource literature.

While natural resources expand access to outside capital, this does not develop the

intermediate factors associated with democratization. The results also raise the spector of

a mutually reinforcing relationship between the political patterns developed in earlier

chapters and a lack of socio-economic development. This chapter will begin, however,

by motivating the modernization question. Are fuel exporting states really an exception

to classic democratization theory, or does their inflated per-capita wealth mask

underlying dynamics that are familiar in developing states?

5.1 Natural resource curse or disappointment?

Whether we make it explicit or not, the study of comparative politics is, in large

part, the study of counterfactuals. When we compare countries to determine what factors caused one outcome instead of another, we are implicitly saying that if the key variables were present/absent in another country, it would experience the same outcome. The same 147 could be said of the conclusions reached in any branch of experimental science. Where

political science differs is that we cannot directly test the counter-factual. If we could act

as gods and manipulate human societies, then maybe we could test our theories in the same way as a chemist. But even if it were possible to do so, it would be ethically dubious. So, we are left with the sub-optimal method of using heterogeneous cases with statistical or case-selection controls.

My reason for raising this issue is that the literature on the natural resource curse relies on a very counter-intuitive counter-factual -- that countries would be more likely to be democratic if they did not have oil wealth. Indeed, when all the components of the natural resource curse are taken together, they suggest that countries may be better off if they did not have any oil reserves. This counter-factual statement has evoked a strong response from some scholars. Hellinger (2000, p. 118) states that "[t]o argue that

Venezuela and other oil-exporting nations would have been better off without finding oil, as Karl does, is like arguing that a poor man would be more likely to become rich without winning the lottery than by winning it." Indeed, the larger question, which has gone unspoken by researchers in the curse literature is -- would the countries studied as part of the natural resource curse be more likely to be democratic if they had never discovered oil? Would Kuwait be more likely to be democratic if it were still dependent on pearl diving? What would Saudi Arabia's government look like if American companies had not been successful in drilling for oil in the 1930s?

Finding an answer to this question is not straightforward. A start can be made, however, by changing the question slightly. Instead of asking whether oil dependent states are less democratic than would otherwise be expected given their level of per- capita GDP, we could ask whether oil dependent states are less likely to be democratic 148 than the other countries in their geographic area.

While usually assumed to be a proxy for other variables, countries in geographic

proximity tend to share similar economic characteristics, cultures, religions, and regime

types (Inglehart and Welzel, 2005). The literature on this phenomenon is relatively

undeveloped, but there are some hypotheses for why this is the case. Bendix (1978)

suggests that reference societies are an important part of a country's becoming democratic

or authoritarian. Countries usually compare themselves to their neighbors, both because

they have more interactions and because these countries are their main competitors and/or

most important allies. Similarly, Kennedy (2008) argues that individual self- categorization, based in part on mental geographic associations, influences an individual’s political expectations and behavior. Kopstein and Reilly (2000) point to the importance of geographic proximity in explaining democratization in post-Communist countries, arguing that the level of “stocks and flows” -- economic and political interaction between countries -- causes countries in geographical proximity to adopt similar government institutions. In economics, there is a large literature on the gravity theory of trade and development, whereby transportation costs, similarity in natural resource base, and similarity in comparative advantage result in countries sharing trade partners and economic characteristics within geographical regions (see e.g. Bergstrand,

1985; Feenstra et. al., 2001; Webb, 1961).

In Tables 5.1 and table 5.2, the variables used earlier to operationalize fuel exports are coupled with a new group of control variables. The post-Cold War variable remains, since it is a temporal, rather than spatial control, but region variables are used instead of per-capita GDP. The results can be interpreted as the effect of fuel export dependence and fuel export income, given the region in which the country is located. In other words, 149 it will tell us whether fuel export dependence results in a country being more likely to be authoritarian than its neighbors who are not dependent. When this is done, the effect of fuel export dependence and, to a lesser extent, fuel export income remains intact given this alternative specification, but the effects are substantially weaker when comparing geographical cohorts than when controlling levels of per-capita GDP.

Comparing Tables 5.1 and 5.2 to Tables 2.1 and 2.2 from Chapter 2, one observes that the coefficients for the fuel dependence variable have dropped substantially. In

Model 1, which uses the Polity IV measure of democracy in the global sample, the coefficient for fuel export dependence drops from -.075 to -.034. In substantive terms, this means that the effect of fuel dependency drops by nearly 55 percent when the countries are compared to neighboring countries, rather than countries of similar per- capita GDP. The differences are slightly less pronounced in the models excluding industrialized countries, which would be expected since it limits the income comparison.

However, the coefficients are still substantially lower. Again, looking at the Polity IV model, the fuel dependence variable loses over 31 percent of its value when utilizing regional, instead of income, controls. Similar results can be observed in all the models.

150 Model 1: Model 2: Model 3: Polity IV FHPR ACLP

Fuel Dependence -.034 (.047) -.011 (.022) -.010 (.074)

Fuel Income per- -.0004 (.016) .000 (.237) -.0004 (.189) Capita

Post Cold War 4.022 (.000) .598 (.000) 3.240 (.000)

Latin America and -6.668 (.000) -2.017 (.000) -4.319 (.000) Caribbean

Middle East -12.973 (.000) -3.787 (.000) -11.455 (.000)

Eastern Europe -8.257 (.000) -2.466 (.000) -5.813 (.000)

Africa -12.667 (.000) -4.151 (.000) -8.585 (.000)

Southern Asia -12.057 (.000) -3.745 (.000) -7.192 (.000)

Eastern Asia -9.049 (.000) -3.078 (.000) -7.367 (.000)

OSCE ------

R2 Within .211 .067 --- R2 Between .613 .641 --- R2 Overall .505 .504 ---

N 4758 3845 4662 Groups 146 146 145

a – values are the weighted average of the between and within effects, with p-values in parentheses (1-tailed). All p-values are calculated using robust clustered standard errors.

Table 5.1: Effect of fuel exports on democracy compared to regional trends in full sample, 1965-2001.

151 Model 4: Model 5: Model 6: Polity IV FHPR ACLP

Fuel Dependence -.036 (.041) -.011 (.020) -.007 (.189)

Fuel Income per- -.0004 (.007) .000 (.306) -.0009 (.170) Capita

Post Cold War 4.603 (.000) .654 (.000) 3.224 (.000)

Latin America and 1.986 (.114) .982 (.019) 2.480 (.000) Caribbean

Middle East -4.149 (.024) -.756 (.073) -4.819 (.000)

Eastern Europe .124 (.471) .510 (.158) 1.049 (.089)

Africa -4.032 (.002) -1.153 (.002) -1.764 (.001)

Southern Asia -3.538 (.065) -.755 (.110) .328 (.316)

Eastern Asia ------

R2 Within .248 .070 --- R2 Between .395 .382 --- R2 Overall .311 .244 ---

N 3943 3207 3850 Groups 123 123 122

a – values are the weighted average of the between and within effects, with p-values in parentheses (1-tailed). All p-values are calculated using robust clustered standard errors.

Table 5.2: Effect of fuel exports on democracy compared to regional trends in developing countries, 1965-2001.

152 The effect of per-capita fuel income almost completely dissolves. In Models 1

and 4, which utilize the Polity IV measure of democracy, fuel export income remains

within reasonable bounds of certainty (p = .016 and p = .007 respectively), but it loses

much of its substantive effect. In both models, the coefficient drops by about 60 percent.

The models which utilize the Freedom House and ACLP measures perform even worse.

In Models 2 and 5, which utilize Freedom House, fuel export income not only falls outside reasonable bounds of certainty (p = .237 and p = .306), but it also reverses

direction. Models 3 and 6, which utilize the ACLP measure, have the expected results,

but the coefficients lose almost 87 percent of their value and the results are no longer

statistically significant (p = .189 and p = .170). This is not completely surprising, given

the heavy concentration of proved reserves in the Middle East – with 742.7 thousand

million barrels of proved reserves, compared with Eurasia’s 144.4 (BP, 2007). Yet, all of

these results suggest that the fuel export income, by itself, does not necessarily drive

countries outside of their geographic cohort.

These results are supported by qualitative comparisons of geographical regions.

Ross (1999) notes that it is striking how little fuel dependent cases differ from other

comparable cases in their region. The case of Venezuela is instructive. Chapter 2

mentioned that Venezuela is often taken as an example of a democracy that has been

destabilized by its fuel dependence. While Venezuela certainly experienced some

growing pains related to oil revenues, it was still one of the most politically stable Latin

American countries between 1974 and 1992. For example, no coups during this

timeframe were successful, although there were some significant attempts.

153 Even the current president, Hugo Chavez, pegged by some as “semi-authoritarian”

for his bending of the rule of law and personalistic leadership style (ex. Ottaway, 2003), has won regular elections that have been deemed reflective of popular support by international observers. And, more importantly, he has lost a referendum to strengthen his authority (Romero, 2007). At least on a formal level, Venezuela appears to be a relatively stable electoral democracy compared to its neighbors.

Dunning (2007) has picked up on this pattern, arguing that fuel export income can have a stabilizing effect on democratic regimes, since it alleviates the pressure of re-

distribution on the elite. In the case of Venezuela, he explicitly argues that fuel export

revenues were utilized to offset high levels of political inequality. This quieted some

calls for re-distribution from the lower classes, and eased the taxation burden on elites, making democracy more stable than other Latin American states.

Venezuela’s neighbors include Guyana, whose election in 1980 was deemed

“fraudulent in every possible respect,” Brazil, which did not hold open presidential elections until January 1985, and Colombia, which has been embroiled in a bitter civil war involving the leftist Revolutionary Armed Forces of Columbia (FARC) and several illegal counter-insurgency groups since 1966.

The average government stability rating from the World Bank's governance indicators, between 1996 and 2000, puts Venezuela comfortably in the same category as its neighbors (Kauffman et. al., 2003). Venezuela has an average score of -.41, compared to -1.27 for Colombia, -.21 for Guyana and -.06 for Brazil. All of them are below the mean for this measure and all are less than a standard deviation away from Venezuela's stability score. Further, during the entire period covered by the World Bank, only one country is ever more than a standard deviation away from Venezuela -- that is Colombia 154 who is more than a standard deviation lower in stability in 2000.

In terms of per-capita GDP, however, Venezuela acquires a slightly different

cohort. Its 1992 per-capita GDP of $5,367 is well above Guyana's $694 or Columbia's

$1,931. Brazil, with its per-capita income of $3,017 comes closer, but is still slightly below the World Bank cutoff for classification of "upper middle income" (above $3,528).

The only Latin American country with a comparable per-capita GDP in this time is

Uruguay, with $5,448. In terms of per-capita GDP, the most comparable cases are the new EU members from Central Europe -- Czech Republic, Hungary and Slovenia -- and a few smaller island states.

Table 5.3 gives some exploratory evidence that this pattern may be a more general characteristic of fuel exporting states. The table lists the 25 most fuel dependent countries, on average, from 1965 to 2001. The next three columns list their average per- capita GDP, secondary school enrollment, and infant mortality during this time.

Underneath, the three closest non-fuel-exporting countries, in terms of rank, are listed.

The results show a surprising level of disparity between the cohorts for some of the most prominent fuel exporters in terms of per-capita GDP, and their cohorts in terms of enrollments and infant mortality. Saudi Arabia, for example, has an average per-capita

GDP that puts it roughly with Cyprus, Portugal and South Korea, but its social indicators place it much more in line with developing countries like Ecuador, Turkey and Brazil.

Similarly, Kuwait’s per-capita GDP is in roughly the same class as Australia, Canada and the United Kingdom, while its social indicators are more in line with Panama, Namibia and Myanmar. Finally, the United Arab Emirates’ per-capita GDP is roughly in line with those of France, Belgium and Finland, while its social indicators are much closer to those in the Philippines, Mongolia and Bulgaria. These observations coincide with the 155 observations of the United Nations Development Programme (2002, p. 26) that “The

Arab region might be said to be richer than it is developed with respect to basic human- development indicators.” The pattern is not completely clear. The former Soviet states, in particular, seem to have much better social indicators than other oil exporters – most likely due to institutional legacies. However, a general pattern does seem to emerge.

156 Average Average infant secondary mortality, 1965- Average GDP school 2001 (infant per-capita, 1965- enrollment, deaths per Country 2001 1965-2001 1,000 births) Libya NA 82.2% 56.71 (N. Korea, (Tadjikistan, (China, Mexico, Somalia, Bulgaria, Ecuador) Afhganistan) Romania)

Saudi Arabia $ 7739.29 49.08% 66.06 (Cyprus, (Ecuador, (S. Africa, Portugal, S. Brazil, Turkey) Brazil, Korea) Dominican Republic)

Qatar NA 79.29% 21.75 (N. Korea, (Bulgaria, (Macedonia, S. Somalia, Romania, Korea, Afhganistan) Croatia) Hungary)

Iran $ 1712.70 62.09% 81.01 (Jordan, Tunisia, (Peru, Panama, (Kenya, Thailand) Namibia) Guatemala, Morocco)

Algeria $ 1547.45 53.73% 81.79 (El Salvador, (Jordan, (Kenya, Bulgaria, Columbia, Guatemala, Romania) China) Morocco)

Nigeria $ 258.94 24.69% 115.78 (Rwanda, Niger, (Cambodia, (Lesotho, Tanzania) Bangladesh, Tanzania, Guatemala) Rwanda)

Continued

Table 5.3: Economic vs. social indicators in fuel export dependent countries (and their respective cohorts), 1965-2001.

157 Table 5.3 continued

Venezuela $ 3723.65 39.23% 30.77 (Brazil, Estonia, (India, (Mauritius, Fiji, Chile) Thailand, Romania) Zimbabwe)

Azerbaijan $ 401.99 82.99% 79.72 (Laos, (Cyprus, Israel, (Papua New Kyrgyzstan, Tadjikistan) Guinea, Kenya, Pakistan) Eritrea)

Angola $ 634.07 13.73% 163.67 (Zimbabwe, (Mauritania, (Malawi, Indonesia, Malawi, Guinea, Nicaragua) Afghanistan) Liberia)

Bahrain $ 11128.46 92.21% 21.86 (Israel, Spain (US, Ukraine, (Macedonia, S. Greece) Slovenia) Korea, Hungary)

Iraq NA 42.72% 78.60 (N. Korea, (Costa Rica, (Nicaragua, Somalia, Dominican Eritrea, Papua Afhganistan) Republic, New Guinea) Nicaragua)

Oman $ 4695.30 48.20% 57.23 (Czech Republic, (Botswana, (China, Mexico, Uruguay, Croatia) Brazil, Turkey) Ecuador)

Trinidad $ 4045.24 72.56% 29.86 (S. Africa, (Guyana, Cuba, (Jamaica, Hungary, Argentina) Mauritius, Fiji) Slovakia)

Continued

158 Table 5.3 continued

Turkmenistan $ 1458.14 100.00% 73.00 (Bulgaria, (Denmark, (El Salvador, Romania, Australia, Honduras, Guatemala) Norway) Peru)

Kuwait $ 16168.51 62.23% 24.20 (Australia, (Peru, Panama, (Poland, Canada, UK) Namibia) Georgia, Malaysia)

Yemen $ 281.30 37.42% 88.8 (Madagascar, (Venezuela, (Mongolia, Bangladesh, Zimbabwe, Ghana, Turkey) Sierra Leone) Paraguay)

Gabon $ 4648.99 57.00% 80.59 (Czech Republic, (Albania, (Papua New Uruguay, Croatia) Maruitius, Guinea, Kenya, Malaysia) Guatemala)

Congo Brazzaville $ 802.23 52.78% 90.74 (Equatorial (Algeria, China, (Mongolia, Guinea, Moldova, Tunisia) Ghana, Turkey) Guyana)

UAE $ 21945.15 68.89% 20.42 (France, Belgium, (Philippines, (Cuba, Finland) Mongolia, Bulgaria, Egypt) Macedonia)

Continued

159 Table 5.3 continued

Syria $ 679.47 45.68% 55.90 (Guinea, (Turkey, (China, Mexico, Honduras, Swaziland, Ecuador) Zimbabwe) Bolivia)

Indonesia $ 671.85 42.27% 72.98 (Honduras, (Dominican (El Salvador, Zimbabwe, Republic, Honduras, Nicaragua) Nicaragua, Peru) India)

Russia $ 2424.72 85.94% 17.64 (Turkey, (Latvia, Poland, (Italy, Greece, Botswana, Italy) Belarus) Macedonia)

Ecuador $ 1398.58 51.45% 58.01 (Romania, (China, Tunisia, (China, Mexico, Guatemala, Botswana) Philippines) Dominican Republic)

Kazakhstan $ 1443.98 89.40% 64.92 (Bulgaria, (Czech (Tunisia, S. Romania, Republic, Africa, Brazil) Guatemala) Belarus, Slovakia)

Norway $ 25597.45 100.00% 8.63 (Germany, (Denmark, (Sweden, Japan, Denmark, Australia, Netherlands) Sweden) Spain)

160 One potential explanation for why fuel dependent states are more exceptional

when compared to countries with similar per-capita GDP, than when compared to regional groupings, is that increases in per-capita wealth have not been translated into corresponding increases in the intermediate variables outlined by modernization scholars.

These intermediate variables include increased education, decreased inequality, and

improvements in individual wellbeing. In turn, it is these indicators of socio-economic

development that are through to be more proximate to the social transformations and

changing political expectations associated with modernization theory.

5.2 Explaining the neighborhood effect

Why would this be the case? The model in Chapter 3, along with the empirical

results presented in Chapter 4, suggest a reason. As the importance of fuel exports

increase in the economy, so does government discretion over economic distribution. This

allows the government to distribute resources as private goods to gain and hold political

support. The empirical result of this is that, even though spending is likely to increase

with fuel income, much of this spending will be allocated towards private goods, rather

than the effective provision of public goods.

This increased capacity for the allocation of private goods combines with another

major attribute of fuel export revenues, their volatility, to make the effectiveness, and

reliability, of public goods investment more problematic for fuel exporting states. In this

section, I will contend that a large part of the explanation has to do with the combination

of increased government discretion over allocation, combined with the tendency for fuel

export revenues to rise and fall quickly.

Fuel prices over the past forty years have fluctuated wildly (for reference, see 161 Figure 3.4 in Chapter 3). Because fuel revenues tend to flow in very quickly during

periods of high prices and/or increased exploration, even a country intending to invest all of the funds into socio-economic development will find itself trying to make these investments in the absence of an institutional and regulatory framework capable of efficiently handling projects and funds of this size. If revenues then fall, either because of a drop in price or the exhaustion of resources, the government finds itself with ongoing investments and institutions that are set up with expectations of high funding and may not

be easily adjusted to periods of lower revenue.

Figure 5.1 gives an example of this in the case of Saudi Arabia. Between 1965

and 1974, the amount of money from oil exports flowing into Saudi Arabia increased by

approximately $125 billion.52 This was an average increase of about $13 billion a year.

It should be noted that this does not count any of the other money that would flow into

the country as a result of the spike.

52The numbers used here and in the chart are both approximations using data from British Petroleum (BP, 2003). They are calculated by taking total production (in thousands of barrels per day) minus consumption (in thousands of barrels per day), multiplying by the average crude price for that year, and multiplying again by the number of days in the year. 162 350000000

300000000

250000000

200000000

150000000

100000000

50000000 Approximate fuel revenues (in thousands of constant 2004 USD) 2004 constant of thousands (in revenues fuel Approximate

0

6 8 ar 68 70 72 74 80 82 84 86 88 90 92 94 96 98 96 9 9 9 9 976 97 9 9 9 9 9 9 9 9 9 9 Ye 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2000 Year

Figure 5.1: Approximate oil export revenue in Saudi Arabia, 1965-2001.

This rapid influx of funds in a developing country meets with several circumstances problematic for its administration. First, developing countries do not normally have a strong regulatory and legal framework capable of overseeing the administration of projects and to bring legal action against those who abrogate their duties related to those projects. Second, many of these countries are governed by systems that do not include strong checks and balances, and/or lack a serious opposition. This means there are few groups with the incentive, or resources, to expose waste. Third, civil society and traditions of public participation also tend to be weak, making it difficult to count on public oversight. Finally, in extreme cases, there is just too much money coming in to be spent or regulated effectively. Infrastructure investment can only go so 163 far before it runs into diminishing returns. Some infrastructure projects are contingent on other projects. For example, a hospital does not make much sense if there is not adequate transportation for patients. This has not stopped some states from trying. The expensive and modern King Khalid airport in Riyadh, Saudi Arabia, for example, has had one of its terminals mothballed since 1983. These problems are not special to fuel exporting states, they are common to all developing countries. But, not all developing countries go through the revenue swings experienced by fuel exporters.

A side note should be made at this point. There are a number of scholars who use

the experiences of established democracies and developing countries as comparisons to

the experiences of developing states with fuel revenues. The above analysis, if correct,

suggests that there is some danger in this strategy. Whereas developing countries are

unlikely to have strong traditions of oversight and opposition, developed democracies

have both, and these have a large impact on the transparency of government allocation.

In developed states, there is also likely to be a well-developed commercial class outside

of the extractive industry sector, forming a powerful lobbying group against policies of

fuel exploitation that might result in an over-emphasis on oil reserves. In other words, in

comparing Nigeria to Norway, one might miss the most important aspects of the natural

resource curse.

The results of the lack of oversight and capacity are many, but the most important

ones for our purposes have to do with the allocation of private goods. It should be taken for granted by now, that political actors have more diverse goals than simply adopting policies that satisfy a greatest good for the greatest number criterion. In order to continue in power, and continue implementing their social program, they must gain the political support of key actors. This means supporting programs that will gain support of 164 important groups. An example of such a program is the subsidization of oil for industry

and agricultural use. Subsidization of gasoline in Iran, discussed in Chapter 2, is just one

among many. In Kazakhstan, oil companies are required to give a certain amount of their

production to the state at a reduced price for use by farmers during the harvest.53 While

this may be a sub-optimal use of resources, it is a relatively costless way for the

government to win the support of agricultural workers. The need for political support can

also lead to oil revenues being used as patronage to buy support of other powerful political actors. These may include special access to government projects and/or help in commercial ventures, both of which are prevalent in fuel exporting countries.

In addition, there is a certain amount of uncertainty in any political decision.

Political actors generally have “pet projects” for their administration. These are large

projects, designed to make a large an immediate impact on their surroundings, and

establish the leader's legacy for a long time to come. These projects often involve

expense far exceeding their potential benefits, and are difficult to justify in normal

budgets. An explosion of revenue makes such projects possible with a minimum of

protest or oversight. As Ascher (1999) points out, there is still a tendency among both

political leaders and the public to view oil revenues as windfalls rather than public

money. Thus, leaders are given the opening to promote their pet projects, at a potential

cost to more modest, but effective, public investment. To put it another way, natural

resource funds are easy for political actors to exploit in a distributive, rather than re-

53One of the problems with this program is that not all of the oil intended for this purpose makes it to the farmers. Some of it is siphoned off and sold at market prices, making money for the refineries owned by individuals close to the government. PetroKazakhstan tested this by placing a dye in the oil that it contributed for the program. Later, they found the same dye in the oil products of several commercial stations being sold at market prices. As an example of how popular these programs are, however, when PetroKazakhstan refused to contribute the next year, it resulted in a rather large public outcry (interviews with PetroKazakhstan executives, 2005; PetroKazakhstan, 2005). 165 distributive fashion. Since the main costs involved in spending natural resource money in

unproductive manners are in future opportunity costs, rather than in current material cost,

most people view the policies as investment rather than re-distribution. As Lowi (1964)

points out, in the public policy sphere, distributive actions tend to win political support

because they are viewed as having only winners and no losers. Thus, a political actor may heavily mismanage a natural resource and still see his/her popularity increase.

In states where windfalls are large, corruption is much easier to hide, and even to justify. Without pre-established domestic companies to bid for projects, and with more revenue flowing in than can be effectively spent on development, the use of public funds for private gain can be viewed as a way of promoting domestic production, or as a harmless method of rewarding public servants. The Saudi Arabian ambassador to the

United States, Bandar Bin Sultan (quoted in Frontline, 2001) was surprisingly frank on this point:

In the last 30 years, we have implemented a development program that was approximately ... close to $400 billion worth... Now, look at the whole country, where it was, where it is now. And I am confident after you look at it, you could not have done all of that for less than, let's say, $350 billion. If you tell me that building this whole country, and spending $350 billion out of $400 billion, that we misused or got corrupted with $50 billion, I'll tell you, 'Yes.' But I'll take that any time. There are so many countries in the Third World that have oil that are still 30 years behind... What I'm trying to tell you is, so what? We did not invent corruption, nor did those dissidents, who are so genius, discover it. This happened since Adam and Eve. ... I mean, this is human nature. But we are not as bad as you think. ...

The "so what" attitude is much more likely when states have high revenues from oil. In cases such as Saudi Arabia, Kuwait and UAE, where, especially in boom times, the oil

revenues are more than enough to promote large social projects and fuel corruption, questions of enforcement will not loom as large. And while these social programs should result in greater socio-economic wellbeing, corruption will drag down these indicators 166 compared to other states which reached the same level of per-capita income without the exploding effect of oil revenues.

For states with lower fuel export revenues, the situation is a little more complex, but often leads to the same result. During the price spikes of the 1970s, states in this

category experienced what Ross (1999) calls “oil mania” -- a tendency to assume that: (1)

revenue increases would continue or at least remain at a higher level than previously; (2)

that the rise in revenues presented a historic opportunity to leap forward in economic

development; and (3) that rapid public investment could be absorbed into the social

system while minimizing social displacement. These assumptions led some oil states,

such as Venezuela, to borrow heavily against future oil revenues (Karl, 1997). This

essentially had the same impact as the spikes in high revenue states, with a rapid influx of

revenues coming into a system that could not readily absorb them, and the results for

socio-economic indicators were disastrous.

The impact of boom periods on debt levels in many of these countries is telling.

While those counties with extremely high revenues became the source of petro-dollars

which funded the third world debt crisis, those which were more marginal producers

became recipients. Kretzmann and Nooruddin (2006) find that increases in oil

production accompany massive increases in debt. Their study suggested that doubling a

country's output of oil increases the size of its external debt, as a share of GDP, by 43.2

percent. A large part of this is the counter-intuitive tendency of some producers to

borrow massively against future production. Since the production of oil, during periods

of high prices, increases the credit rating of these countries disproportionately to their

record on repayment, they are able to borrow much larger amounts than would have been

allowed under normal circumstances, and at lower levels of interest. In the 1970s, when 167 it appeared that oil was the way towards generating the capital necessary for

development, some of these countries borrowed massively against future oil production.

Thus, the ironic situation whereby Nigeria, Indonesia, Mexico, Venezuela, Ecuador and

Congo-Brazzaville, although oil exporters, exited the oil boom period with some of the largest debt loads in the world.

The case of Nigeria is emblematic of this trend. As Africa's largest oil producer,

it stood in what most analysts at the time would have considered an enviable position vis-

a-vis other developing countries during the 1970s oil shocks. Rising oil prices, and the

resulting increases of revenue, were expected to provide the increase in capital the

country would need to improve its infrastructure and industry. Accompanying the new

revenues was an increased pressure, from all parts of society, to spend those revenues.

The government started investments in many sectors, not just in human and physical

infrastructure. Among other projects funded by oil revenues were a major steel complex

and an automobile assembly plant. Where oil revenues fell short of funding development programs, Nigeria found many willing lenders, who figured that future oil revenues would facilitate Nigeria's ability to repay its debt. As Figure 5.2 illustrates, Nigeria's debt, as a percentage of its GDP dropped slightly in 1976, only to re-emerge and increase from 1977 onwards. This suggests that, even during the second surge in oil prices, the

Nigerian government was increasing its borrowing against future oil revenues. In other words, even those developing countries without the extreme per-capita revenues of some

Middle Eastern countries, had more than enough access to revenues to exhibit the problems outlined above.

168 160.00

140.00

120.00

100.00

80.00

60.00 Percent

40.00

20.00

0.00

r 4 5 2 a 70 71 78 8 8 86 9 93 94 00 9 9 9 Ye 19 19 1972 1973 1974 1975 1976 1977 19 1979 1980 1981 1982 1983 19 19 1 1987 1988 1989 1990 1991 19 1 1 1995 1996 1997 1998 1999 20 -20.00

-40.00 Year

Growth in GDP Growth in external debt Ratio of external debt to GDP

Figure 5.2: Growth in GDP (red) versus growth in debt (yellow) and the ratio of debt to GDP (blue) in Nigeria, 1970-2002.

Interestingly, we see many of the same attitudes emerging again in the current climate of high oil prices. Perhaps oil exporters are correct to think that this current rise of prices, driven by demand from the booming Chinese economy and instability in the

Middle East, will result in a long-term increase in the average oil price. However, history warns that such expectations are, at best, a gamble with the future of the country.

Leaders should be careful about making assumptions about the future based on current commodity trends.54

54While the point is debatable, more than one CEO of a major oil company has predicted that oil prices will fall. Jeroen van der Veer, the CEO of Shell, suggests that crude oil stocks in factories around the world are normal or above average, suggesting that the current price is driven more by geopolitical tensions and nontraditional money moving into the oil market than by a shortage or oil. John Browne, of BP, 169 Aside from the problems discussed above, there are certain areas of socio-

economic development that are just difficult to increase at the same rate as per-capita

GDP. Improving performance in education, for example, requires trained teachers along with students that have passed through the new system, wrestling with traditional elites over content, and a market demand for educational skills driving institutions towards attainment of reputation. Improvement in all these areas can take generations, and some of them are predicated on issues outside of the schools themselves, such as the power of traditional elites and market demand for skills.

Similarly, healthcare and welfare requires a bureaucracy capable of administering

the new system, physical infrastructure, and a trained professional class. The latter, in turn, relies on having an educational system with the ability to produce skilled workers.

To a certain extent, the pattern observed above could simply be due to oil revenues

increasing per-capita GDP in a manner not predicated on the development of

infrastructure, as is usually emphasized in other development plans. In this case, we would expect socio-economic indicators to lag behind to a certain extent. In other words,

in fuel exporting countries, infrastructure does not bring capital, rather capital drives

investment in infrastructure.

Periods of dropping oil prices are where policy failures become the most evident,

and where socio-economic development is often sacrificed to the needs of political

expediency. Boom periods allow governments to fund large development projects while

not worrying too much about regulating the corruption often accompanying them, and

while providing plenty of excess money for patronage. Falling oil prices force the

government into tough decisions which often involve tradeoffs between what is

goes even further, suggesting that the price will be $40 or maybe even $25-30 in the long-run (see Newsweek, 2006) 170 politically expedient and what is good for socio-economic development.

Corruption and patronage are difficult to undo. Governments usually award the

highest levels of private payments to those agents in the best position to pose a threat to

the regime's continuance. And, as Machiavelli (2003, p.72) suggested in The Prince,

“[A]bove all, he [the Prince] must abstain from taking the property of others, for men sooner forget the death of their father than the loss of their patrimony.” Thus, it is more difficult for governments to end corruption and patronage during downward cycles, while it is easy to fall into these patterns during upward cycles.

For those countries with abundant revenues, the problems are not as acute.

Certain programs may need scaling back, but the programs themselves are not in danger of discontinuance. The tendency is to continue programs in a manner posing the least amount of political risk, rather than continuing forward with more effective, but more politically and monetarily costly, programs.

More apparent and tragic examples can be found in states which do not have the same size of per-capita revenues. Continuing with the example of Nigeria, the fall in oil prices during the early 1980s spawned a rush of borrowing on the part of the Nigerian government. From 1980 until today, the external debt of Nigeria has rocketed from 13.1 percent of GDP to 163 percent, more than a twelve-fold increase. Similar trends have been noted in other oil dependent states, including Indonesia, Mexico, Congo-

Brazzaville, Venezuela and Ecuador. In Mexico, the dependence of the government on the state-owned oil company for revenue has resulted in PEMEX now being the most indebted company in the world (Markus, forthcoming). Ecuador's debt, which stood at a mere $328 million in 1972, stood at $13 billion by the end of 1998, or about two-thirds of national output (Kretzmann and Nooruddin, 2006). In Congo-Brazzaville, reckless 171 government spending, domestic conflict and shenanigans by their foreign oil partner,

ELF-Congo, resulted in the government mortgaging its projected revenues decades into the future (Global Witness, 2004). Currently, external debt in Congo-Brazzaville stands

at more than 193 percent of GDP.

Despite the seemingly contradictory nature of the statement, a strong case can be

made that some states without a very high revenue-to-population ratio in their oil development have ended up worse off than if they had not developed their oil reserves.

Nigeria's per-capita GDP has only increased from $244.67 in 1965 to $257.48 in 2001, with an increase of only .05 percent for the 36 year span, or a minuscule average per- capita growth rate of .001 percent per year. Combining this lack of growth with the manifold increase in debt load, the accusation that lending policies for natural resource exploitation by the World Bank have more to do with the energy needs of the North than the development needs of the South gains more credibility (Kretzmann and Nooruddin,

2006).

In summary, the explosive nature of oil revenues, and the propensity for their use in funding private goods, explains why we would expect fuel exporting states to have socio-economic indicators lagging behind those of other states with similar levels of per-

capita GDP. During periods of quick price increases, those countries with high revenue/population ratios find themselves with more money than can effectively be spent on development projects. This money becomes an easy area for pursuit of pet projects,

patronage and corruption. In states with lower revenue/population ratios, there is a

tendency towards “oil mania,” whereby loans are taken out against future revenue in

order to secure the funds for rapid industrialization. In both cases, the government makes

a number of politically important commitments with spending, not all of which have to 172 do with socio-economic improvement, which are difficult to break.

During a period of declining oil prices, the governments of these states face some difficult decisions. For high revenue to population states, there may not be a need to take out debt to continue social programs, but there is a political expediency demanding that programs involving patronage and corruption continue, while development programs are watered-down as a concession to traditional elites in exchange for their political loyalty.

For countries with lower revenue to population, continuing development programs usually means bringing on a large amount of debt. When these increasing debts were not offset by an eventual increase in oil prices or the increased revenue from development projects, countries like Nigeria, Ecuador, Mexico and Congo-Brazzaville found themselves in alternating periods of fiscal restraint to pay debt and political backlash against belt-tightening policies.

The next sections turn to testing this theory by looking at two areas of socio- economic development, education and individual well-being. The results provide some support for the contention that fuel export revenues do not have an impact on these measures of socio-economic development, proportionate to their per-capita GDP. This, in turn, suggests that fuel exporting countries are less of an exception to classic democratization theory, at least in terms of socio-economic conditions, than is usually argued.

5.2.1 Education

Of all the mechanisms posited to be conductive to pro-democracy attitudes, perhaps none has received as much attention in the literature as education. As Lipset

(1959, p.79) wrote: 173 Data gathered by public opinion research agencies which have questioned people in different countries with regard to their belief in various democratic norms of tolerance for opposition, to their attitudes toward ethnic or racial minorities, and with regard to their belief in multi-party as against one-party systems have found that the most important single factor in determining those giving democratic responses from others has been education. The higher one’s education, the more likely one is to believe in democratic values and support democratic practices. (emphasis in original)

There are five major hypotheses for why this is the case. First, inherent in education are the complex cognitive skills that foster an understanding and commitment to democratic values (Almond and Verba, 1963). Second, political socialization into the dominant value system of society through schooling results in more democratic values among the educated in democracies (Dahl, 1961). Third, since educated citizens are more exposed to prevailing norms conveyed through the mass media and other areas of socialization, they become more committed to democratic values (Zaller, 1992).55 Fourth, the educated are more supportive of meritocratic values, and, insofar as they associate democracy as compatible with meritocracy, they will support those democratic values (Collins, 1979).

Finally, more educated citizens are supportive of democracy because they view it as a political system in which they can promote and defend their group interests via class based organizations (Bell, 1973).

Under any of these circumstances, one would expect the level of education in a country to increase as wealth increases, since both individuals and the state can more easily afford higher levels of education, and this increase in education, in turn, makes it more likely for a country to be democratic.56

55 On the other hand, exposure to propaganda and mass media in authoritarian regimes increases support for that regime, up to the point where an individual’s education and knowledge base make them impervious to the propaganda (Geddes and Zaller, 1989). 56In another paper, I argue that the most likely reason for how education impacts this process is two-fold. First, education makes it less costly for elites to adopt democracy when a transition takes place. This occurs because it takes away the stigma associated with mass participation and tends to produce a more 174 There are reasons for doubting whether the level of education provided in oil

dependent states has risen in a consummate fashion with increased wealth. While

spending on education has increased substantially in some oil dependent states,

Waterbury (2003, p. 62; see also UNDP, 2003) points out, “There is good evidence that

the impressive increases in enrollments have come at the price of declining standards.”

Others have noted that education performance in oil dependent states tends to be lower

than would be expected, and that this tendency goes a long way towards explaining

slower growth rates in these countries (Gylfason, 2000; Gylfason and Zoega, 2001).

The theory presented in this chapter provides another reason to suspect that fuel

exporters will perform worse in producing positive educational outcomes. Since

governments in fuel exporting states have greater discretion over the allocation of

economic resources, the distribution of private goods is expected to be higher. This is

expected to trade off with investments in public education, resulting in lower than expected levels of investment and performance. This combines with the volatility of fuel

export revenues to undermine long-term investments in education, again leading us to

expect fuel exporters to under-perform.

Table 5.4 gives the results of several measures of education regressed against

measures of per-capita GDP, fuel dependence, fuel export revenue, and economic growth

for developing states. The models show a striking correlation between fuel export

income and lower-than-expected performance in education. The first indicator of

education is spending on public education as a percentage of GDP. The results presented

moderate polity. Second, it makes democracies more stable once they have been established for many of the same reasons. In addition, a better informed polity is more likely to be concerned with protecting their role in the political system. However, I contend that the logic behind arguing that increases in education will lead to a government transition is relatively weak. Indeed, increased wealth tends to make all types of regimes more stable when authoritarian-to-authoritarian regime changes are taken into account (see Kennedy, 2004). 175 in Model 7 indicate that oil dependent states generally spend a similar percentage of their

GDP on education as other countries of similar wealth. Indeed, the results are not

statistically significant and the coefficient indicates that the relationship is slightly

positive for fuel dependence. Per-capita revenue from fuel exports, however, is

marginally significant (p = .096), and the coefficient indicates that a country with about

$3,300 in fuel export income will spend one percent less of its GDP than a country of

similar per-capita GDP without fuel export income.

Secondary school enrollment demonstrates a similar pattern. The correlation between average oil dependency and secondary school enrollment is not statistically significant (p = .114), and is slightly positive. Fuel revenue per-capita is statistically

significant (p = .013) and has a negative direction -- indicating that a $1,000 increase in a country’s per-capita income from fuel exports results in a 7 percent drop in secondary school enrollment from what would otherwise be expected, given the country’s per-capita

GDP. The same is true for tertiary (college) enrollment. While the effect of fuel export dependence does not fall within the bounds of acceptable certainty (p = .427), we can be very certain that fuel export income has a negative effect on the level of tertiary enrollment (p = .000). The results suggest that a country with $1,000 in per-capita fuel export income will also have about 5 percent lower levels of tertiary enrollment.

The effect of natural resource dependence on illiteracy, measured in Model 10,

has somewhat different results. Neither fuel export dependence nor fuel export income reach acceptable levels of statistical certainty (p = .348 and p = .219 respectively). The relationship for both variables, however, is positive. This suggests that both fuel dependent and fuel rich countries have higher than expected levels of illiteracy.

176 Model 7: Model 8: Model 9: Model 10: Spending on Secondary Tertiary Illiteracy Public School School Education Enrollment Enrollment

Fuel dependence .005 (.263) .111 (.114) .006 (.427) .031 (.348)

Fuel income -.0003 (.096) -.007 (.013) -.005 (.000) .002 (.219)

Ln(Per-capita GDP) .579 (.092) 41.034 (.000) 20.400 (.000) -14.970 (.031)

Growth in GDP -.046 (.001) -.220 (.001) .027 (.165) .119 (.014)

R2 Within .012 .149 .150 .029 R2 Between .043 .516 .516 .534 R2 Overall .024 .479 .479 .486

N 1604 1345 1345 2818 Groups 113 113 113 103

a – values are the weighted average of the between and within effects, with p-values in parentheses (1-tailed). All p-values are calculated using robust clustered standard errors.

Table 5.4: Effect of fuel exports on education in developing countries, 1965-2001.

Taken together, these results generally support the argument made in the last section. Countries receiving a large income from fuel exports under-perform compared to other developing states in almost every area of education during this time period. This includes both investment in education and the outcomes from that education. A deficit in education is particularly troubling when discussing oil development. Many of the most prominent and wealthiest democracies have, at some point in their history, been substantial natural resource exporters -- including the United States, Great Britain,

Canada and Australia.57 The key to their continued growth and diversification, according

57In Australia's case, Maddison (1991, p. 45, fn. 1) notes, “In defining productivity leadership, I have ignored the special case of Australia, whose impressive achievements before the First World War were due largely to natural resource advantages rather than to technical achievements and the stock of man- 177 to some scholars, was their continued investment in exploration, technical progress and

knowledge defying the supposedly “fixed” nature of the resource (David and Wright,

1997, p. 205-212; Wright, 2001; Wright and Czelusta, 2003). In this light, the lack of success in promoting attainment of higher education in high fuel income states since

1965 is especially troubling.

5.2.2 Inequality and deprivation

The story is similar when it comes to inequality and individual wellbeing. Lipset

(1959; 1960, p. 48-50) first suggested that a reduction in inequality played a key role in explaining the positive correlation between economic wealth and democracy. He posited that where there is less disparity in income levels, the calls for redistribution are not going to be as radical. Thus the country's elites will not view democracy as a threat to their wellbeing. It also tends to correspond with the development of a strong middle class that can attenuate the more radical demands of both the lower- and upper-classes.

More recently, several scholars have developed formal models for this relationship. Boix (2003) and Boix and Stokes (2003) have developed a formal model along the lines of Lipset's initial contention, and provide empirical support for the relationship between a decrease in inequality and an increased probability of democracy.

In contrast, Acemoglu and Robinson (2001; 2005) suggest that inequality may encourage democratization, since it increases the probability of the poorer portions of society rebelling, and increases the incentive for the richer classes to prevent rebellion by democratizing, but it will be harmful to the stability of democracy. Whether the probability of democratization is increased or decreased by inequality may be an ongoing

made capital.” 178 debate, however, it seems clear from previous studies that democracies with lower levels

of inequality are more stable (e.g. Reenock et. al., 2007).

Individual wellbeing is usually discussed within the context of inequality,

however, it is worth mentioning the separate hypotheses regarding this variable. One of the great paradoxes for political scientists is the tendency for some of the least-well-off in society to also be among the most conservative. In the late 19th century, Russian revolutionaries were surprised when their attempts to recruit peasants into a revolutionary movement against the Tsar were not only unsuccessful, but some of them were turned over to the authorities by the peasants they were trying to help. Scott (1976; 1985) attributes this conservatism to what he calls the “subsistence ethic.” For individuals living on the edge of subsistence, change, even if it holds the promise of greater average returns, is often rejected for fear of instability which would drop them below an acceptable level of wellbeing. Thus, poverty would, paradoxically, decreases the likelihood of government change. On a related level, poverty and subsistence also have an effect on the ability for individuals to become involved in government. As the psychologist Abraham Maslow (1954) has suggested, there are levels of needs for individuals. Only when certain basic needs are satisfied can individuals worry about issues of self-realization, freedom, and governance. Where basic items like health care

are underprovided, and poverty is high, we are unlikely to observe individuals with a participatory outlook on governance.

The general link given between per-capita GDP and inequality is two-fold: (1) gains naturally spread through society via market distribution, and (2) greater income

makes redistribution less painful for the upper-classes. Yet, the case-study literature

suggests that the increases in wealth associated with oil based development do not 179 translate into decreased inequality. In fact, some case studies suggest that inequality may

increase with oil wealth. For example, Venezuela in 1973 experienced a large increase in growth, but this did not translate into better living standards for the poorest in society.

Karl (1997, p. 113) gives a summary of the situation during Venezuela's first oil boom:

Equity statistics were dismal. A quarter of the population was unemployed or underemployed, while almost half lacked sewage systems and running water. The average income of the poorest 20 percent was lower than that in Colombia, Argentina, Mexico or Brazil. In comparison with fifty-five other middle-income countries ranked by the World Bank, although Venezuela ranked fifth in per capita GDP, it was seventeenth in life expectancy, twenty-second in infant mortality, and fortieth in levels of caloric intake.

A number of other studies ranging from the Caribbean to the Middle East have noted

similar trends (see e.g. Auty and Gelb, 1986; Shultz, 2005, p. 39; Halliday, 1979; Ross,

2001; Davis and Tilton, 2002; Gylfason and Zoega, 2002).58

There are a few reasons why this could be happening. First, oil is a “point-

source” resources. These are resources that are extracted from a narrow geographic or

economic base (Isham et. al., 2003). While the development of these resources may

result in dramatic gains in a country's total wealth and government revenues, it does not

result in large gains in employment. Many of the workers, especially skilled workers,

come from abroad. Therefore, those gains in domestic employment that do take place,

tend to be geographically concentrated around oil field locations. This tendency may be even more pronounced with the development of off-shore oil reserves, as sites for

receiving and storing oil production can be even further concentrated. In most cases, oil development is an enclave activity, importing needed supplies with little value added domestically, since most processing takes place abroad. Thus, rapid increases in wealth

58While the claim that natural resource wealth does not necessarily help the poor is relatively uncontroversial, the claim that oil and mineral resource wealth results in greater inequality is still debated. For a refutation of Ross (2001) and others, see Davis and Tilton (2002). 180 in certain parts of a country may not be widely distributed.

Another challenge for equality posed by natural resources is the disparity in pay that develops between those working in the oil sector and those in other sectors. For example, in Trinidad and Tobago, in 1973, an elite oil worker (about 4 percent of the population) earned about $5,000 a year. By comparison, the mean pay for all other workers was $750 a year and agricultural workers only made $325 (Black et. al, 1976).

This disparity between the income of those in the oil sector and those in all other sectors, accentuated by unemployment reaching 14 percent, led Auty and Gelb (1986, p. 1163) to characterize the economy as “dualistic.”

All these tendencies are exacerbated by the problems associated with “Dutch

Disease.” This phenomena derives its name from the effects on the Dutch economy during the 1970s after the discovery of natural gas in the Groningen fields.59 Both the rising wages offered by oil and mineral producing companies to attract labor, and the appreciation of the country's currency due to dramatically increased inflows of foreign currency -- in the form of foreign direct investment and oil revenues -- hurt other sectors of the economy, such as industry and agriculture. Some scholars argue that this effect is almost unavoidable and tends to have long-lasting effects on the makeup of the economy

(see Karl, 1997; Neary and van Wijnbergen, 1986). This can become even worse if the price of the natural resource drops, since the country's traditional exports may be damaged beyond resuscitation. Other scholars contend that the effects of Dutch Disease tend to be relatively short-lived, if they manifest themselves at all, and can be avoided with prudent policy choices by the government (Gelb, 1988; Benjamin et. al., 1989;

Fardmanesh, 1991; Ross, 1999). In either case, this remains yet another explanation why

59The name “Dutch Disease” is somewhat unfortunate, since the Netherlands actually did a relatively good job in dealing with the economic adjustments involved in the expansion of natural gas exports. 181 oil dependent states may not experience corresponding decreases in inequality.

Finally, all of this is exacerbated by large groups of, often uncounted, migrant

workers who come from neighboring states to work in the oil industry or in manual labor

jobs. These numbers are not trivial. The United Nations Development Programme’s

(2002, p.36) Arab Human Development Report, for example, reports that the migrant

populations in six Gulf states expanded from 1.1 million in 1970 to 5.2 million in 1990,

now making up 2/3 of the population in these states. At the highest end of this, non-

nationals make up approximately 49 percent of the total population in Kuwait and well

over 80 percent of the population in the UAE (CIA, 2007). Some of these are expert

workers employed by oil companies, but a larger portion are from poor states and are

employed in low paying jobs and menial jobs which are “allocated” to foreigners (see

Jureidini, 2003). These workers are not generally included in the statistics analyzed

below, suggesting that the results may substantially underestimate the levels of poverty and inequality in the region. The largest group of workers come to the Middle East as

temporary foreign contract employees, who are often not protected by UN or ILO

conventions, and much less by local labor laws. Racism and xenophobia towards these

workers, especially those from Asia, remains widespread (Jureidini, 2003). Although not

analyzed here, a large group of non-citizen workers may also help to explain the relative

lack of social pressure on governments in many Gulf states.

Table 5.5 looks at the relationship between economic wealth, oil dependency and

inequality. While the models are plagued by missing data for key oil states, a pattern

emerges across the measures. Higher fuel income correlates with greater inequality and

poverty than would be expected given national wealth. Model 11 regresses the GINI

coefficient of income concentration against the independent variables used above. The 182 results are relatively inconclusive. The results for fuel dependence do not reach standard

levels of statistical significance (p = .438), nor do the results for fuel export income (p =

.216). Regressing the variables against the percent of the population living in poverty

yields similar results. Neither of the fuel export variables reach acceptable levels of

statistical significance (p = .410 and p = .201). Interestingly, however, in both models

per-capita fuel income does have the expected direction, suggesting that countries with

higher levels of fuel income do have higher levels of income concentration and poverty.

There are two explanations for these results. The first is that there is no pattern

among fuel exporters as compared to other groups of countries. In this explanation, fuel

exporters are simply too heterogeneous to draw conclusions. The other explanation is

that these two particular measures of inequality have substantial limitations.

Problems in drawing conclusions from inequality data are well known. Despite the efforts of Deininger and Squire (1996), reliable income distribution (GINI) data is simply not available for a large number of cases. In Model 11, only 297 complete cases are available for analysis, with only 81 countries having any country-year information available. Similarly, in Model 12, only 78 countries have poverty information available, with 115 resulting country-year cases for analysis. Even worse is that this data is not missing at random. Data for countries in the Middle East and North Africa are almost non-existent! Several previous authors have warned about the difficulty of drawing conclusions from currently available data on income inequality (Przeworski et. al., 2000;

Barro, 1997). Others have used different methods of imputing the missing data (Boix,

2003; Feng and Zak, 1999), but there are serious problems with these approaches. Not only do they not address the problem of systematically missing data for specific countries, but the processes of interpolation and extrapolation that they utilize is suspect 183 (King et. al, 2001). Even the most technically sophisticated methods for imputing this

missing data, such as multiple imputation, may have difficulty, as covariates for

predicting the level of inequality have not been well established.

As such, many scholars have turned to alternative measures of inequality (see e.g.

Reenock et. al., 2007; Jenkins and Scanlan, 2001). These scholars have based their

analyses on the provision of certain basic measures of needs satisfaction. These include

average life expectancy, infant mortality, and per-capita caloric intake. The logic behind

these measures is simple. Those who are relatively wealthy are likely to satisfy their basic needs no matter what the conditions of others in the society. As such, the level of healthcare they receive and the number of calories they eat is relatively invariant. By

contrast, those in poverty are much more susceptible to swings in both the health care

system and food supply. Thus, the relative levels of these variables, given the country’s

per-capita GDP, should provide an accurate measure of inequality and deprivation.

These variables have the additional advantage of being available for a wider sample of

countries and longer time periods. In Models 13, 14, and 15, the number of cases for

analysis increases by ten times, and includes thirty-three more countries.

The measures of life expectancy and infant mortality are drawn from World Bank

data (World Bank, 2003). Life expectancy is defined as the average life expectancy at

birth in years. Infant mortality is defined as the number of infant deaths per 1,000 births.

The measure of caloric intake comes from the UN Food and Agriculture Organization

(FAOSTAT, 2006). It is measured as the average daily per-capita caloric intake for the

country year.

Both of these proxy measures show a clear under-performance among fuel

exporting countries. In terms of life expectancy, Model 13, per-capita fuel income has a 184 strong (p = .003) and negative effect. The coefficient indicates that for every $1,000 increase in fuel revenue per-capita, life expectancy drops by two years. Infant mortality shows some of the same relationship (p = .029). Here, a $1,000 increase in fuel revenue results in an increase of 7 deaths per 1,000. Finally, Model 15, which looks at per-capita caloric intake, suggests that a $1,000 increase in fuel income per-capita is associated with a 191 calorie drop in the average per-capita food intake. This is likely to hit poorer individuals especially hard, and is an indication of greater inequality (Bowman, 2002).

Perhaps even more telling than the results in these models is the persistent non-reporting of statistics by some major fuel exporters. Although these variables provide much more leverage, countries like Saudi Arabia, UAE and Qatar have yet to report a single year of average caloric intake. This non-reporting by relatively wealthy states is very surprising, and potentially indicates that these models underestimate the effect of fuel exports on inequality and wellbeing. In the conclusion chapter, we will return to this problem of non-reporting of demographic statistics by fuel exporting states. In order for some level of government accountability to take place, a good point at which to begin might be to encourage the accurate measurement and reporting of demographic characteristics.

185 Model 11: Model 12: Model 13: Model 14: Model 15: Inequality Poverty Life Infant Caloric (GINI) Expectancy Mortality Intake Per- Capita

Fuel dependence -.005 -.021 .002 .078 1.651 (.438) (.410) (.471) (.238) (.125)

Fuel income .004 .007 -.002 .007 -.191 (.216) (.201) (.003) (.029) (.000)

Ln(per-capita GDP) .359 -26.826 16.162 -70.073 676.388 (.438) (.000) (.000) (.000) (.000)

Growth in GDP .038 .114 -.049 .330 -1.947 (.263) (.357) (.031) (.008) (.026)

R2 Within .000 .017 .217 .174 .079 R2 Between .026 .427 .631 .618 .552 R2 Overall .023 .397 .584 .555 .521

N 297 115 1599 1676 1226 Groups 81 78 114 114 106 a – values are the weighted average of the between and within effects, with p-values, calculated from robust clustered standard errors, in parentheses (1-tailed).

Table 5.5: Effect of fuel exports on inequality and poverty in developing states, 1965- 2001.

186 The results presented in this section are somewhat mixed. While the traditional

measures of inequality and poverty did not reveal a strong pattern with fuel exports, it is

argued that this is largely due to missing data, especially among large fuel exporters. To

correct for this, I utilized three alternative measures for inequality: average life expectancy, infant mortality, and average per-capita caloric intake. All three of these demonstrated a reliable and negative relationship with fuel export income, despite the persistent non-reporting of several major fuel exporting states. The results support the argument made in section 5.2, that fuel export income does not result in proportionate decreases in inequality and poverty. Also supported is the contention that decreases in

areas of healthcare and individual wellbeing lag behind gains in per-capita GDP among

fuel export rich states.

5.3 Conclusions

Returning to the theme introduced at the beginning of this chapter, what would fuel exporting states look like if they did not have the money from these reserves? The above analysis suggests several interesting conclusions. First, if we compare fuel export rich and fuel export dependent states to their regional counterparts, as opposed to their

economic counterparts, they are still less democratic than would be expected, but less so

than in the earlier analysis. What this suggests is that classic theories of democratization

may not be completely off in producing reasonable hypotheses for understanding the

natural resource curse. Rather, scholars should focus less on capital, of which some fuel

exporting states have in excess of what they could ever use, and look more at socio- economic development.

187 When the focus is changed in this manner, it is notable that countries with higher

levels of fuel export revenue tend to under-perform in education, inequality, and

individual wellbeing compared to countries with similar levels of per-capita GDP. Taken

together, these indicators paint a picture of states in which expectations produced by per-

capita GDP measurements paint a misleading picture of social conditions. The

qualitative evidence suggests further aspects to this tendency which might not be caught by the statistical measurements used here to generalize.

Due to the wide fluctuations in fuel prices, countries dependent on fuel revenues or receive a large amount of per-capita revenue from fuel exports, experience wide revenue swings. In periods of rapidly expanding fuel prices, the countries experience greater pressure from revenues than the political system can effectively absorb.

Increased expectations of the population, and the openings for political elites cannot

easily be constrained. The increase in revenue, combined with the government’s

discretion over its distribution, finds its own path of dispersion based, at least in part, on

political expediency. Some of this finds its way into the pockets of corrupt officials, who find it easy to enrich themselves in a system of large revenues and little effective regulation. Some of the funds find their way into patronage projects, supporting the

leadership's aspirations for a stable power base, but do not necessarily increase social

wellbeing. Some of the funds find their way into pet programs of leaders, which are

often of greater symbolic value than of practical import. And even where there is some level of good governance, the explosion of revenue is unlikely to result in immediate social gains consummate with the inflows of capital, as they quickly run into diminishing returns, with developments in education, health care, and culture taking generations to radically change. 188 During times of decreased fuel revenues, states tend to diverge between those

with very high revenue to population and those with lower revenue to population. Those

with high revenue to population will still have enough capital inflows to continue most of

their development projects, but some concessions must be made to compensate for

lessened capital, making them less likely to take on costly fights with traditional elites in

order to pursue development goals. At the same time, these states are least likely to make

cuts in the patronage and corruption networks developed during the boom times, because

these allies are even more crucial when the funds are not available to forge new networks.

For states which are not floating on a sea of oil, the situation become even more

difficult. Coming out of the 1970s, many states assumed that the price shocks indicated a

longer turning point in oil prices, and that the average oil price would rebound over time.

This assumption, along with the assumption that intense investment in development would reap returns to offset present debt, resulted in high borrowing rates against future oil revenues. This, combined with all the tendencies outlined above, result in high debt and an economic and political climate that is extremely unstable; none of which produces fertile ground for improvement in social conditions. Ironically, the reason that the models above did not note a disproportionate impact on education, inequality and individual wellbeing is likely due to the inability of fuel dependent states to produce economic growth (Sachs and Warner, 1997), making them more similar in terms of per- capita GDP to other developing states.

Placing all of this back into the context of the theory developed in Chapters 3 and

4, the results suggest that the same government discretion that is harmful to the probability of democratization is also harmful to the prospects of socio-economic modernization. Government discretion over spending, especially in non-democratic 189 states, encourages spending on goods that reward political elites for their support, and

which can be withdrawn from groups that do not support the government (Bueno de

Mesquita et. al., 2003). All of this suggests that part of the reason why fuel exporting

states are not as exceptional, when compared with their regional counterparts, is that the

same government discretion that increases the importance of controlling government

patronage, also results in an under-development of programs benefiting society as a whole. In turn, socio-economic underdevelopment likely makes access to government patronage still more important. This cycle of government discretion, group competition, and socio-economic underdevelopment helps explain why fuel exporting states are not

complete exceptions to the ideas espoused by modernization theory.

190 CHAPTER 6

DRILLING FOR AUTHORITARIAN RESOURCES, THE CASE OF KAZAKHSTAN

Previous chapters in this dissertation have made and supported several contentions. We have seen a wide variety of political regime outcomes among fuel export dependent states, although they are generally less democratic than would be expected. Chapter 3 demonstrated that an increased importance of government discretion over the distribution of economic goods can logically account for both the variety and the higher probability of authoritarian governance. Chapter 4 demonstrated that, as predicted in the model, the difference between being just dependent on fuel exports versus being both dependent and rich from those exports explains much of the variance in political regime outcomes. And the last chapter explained why fuel exporting countries under- provide public goods, compared to what would normally be expected given their level of economic affluence.

These chapters have utilized formal models to demonstrate logical consistency of the theory, and large-N statistical analysis to show how these models fit with global patterns, at least since 1960. Yet, there are some things that are difficult to demonstrate using these methods. Does government discretion over economic goods, enabled by revenues from oil exports, really drive regime variation? What are the dynamics of the relationship between fuel exports and political regimes? How do the mechanisms of this theory apply to particular cases, or are they just general categories for macro-analysis? 191 All of these questions are difficult to answer utilizing the methods from previous

chapters, as within-country variance on the main independent variables is usually low, and the statistics available for analysis only communicate so much information about each case.

I turn to a case study methodology for this last chapter to answer some of these questions. The main focus of this chapter will be on analyzing the effect of oil development on governance in Kazakhstan. Since its independence in 1991, Kazakhstan has become an important area for oil development. Its proved reserves in 2006 of 39,800

million barrels of oil, approximately 3.3 percent of the world total, place it ninth in the

world, and fourth outside of the Middle East (BP, 2007, p. 6). As an indication of how

much potential this represents, the actual production of oil in Kazakhstan in 2006,

approximately 1.4 million barrels per day, placed it 19th in the world, suggesting large

untapped potential (BP, 2007, p. 8). Similarly, Kazakhstan has the 11th largest proved

natural gas reserves (BP, 2007, p. 22). Some analysts believe these known reserves may

only be the proverbial tip of the iceberg, since much of Kazakhstan's potential oil wealth

is only beginning to be explored. One retired Chevron executive, speaking of the Tengiz

oil field, called it “a geologists dream” (quoted in Hersh, 2001). In large part because of

these reserves, and the foreign direct investment that has followed from them,

Kazakhstan’s economy has also been one of the fastest to recover in the post-Communist

area.

192

a – map acquired from the University of Texas Perry-Castaneda Library Map Collection.

Figure 6.1: Map of Kazakhstan and Kyrgyzstan.

Despite its vast resources and increased economic investment, however,

Kazakhstan has abandoned early experiments with pluralism, and is largely dominated

today by President Nursultan Nazarbayev, who “stands alone at center stage” (Kimmage,

2005). Indeed, the increasing dominance of Nazarbayev over Kazakhstan’s politics can

be directly attributed to increased economic resources available to the government.

Kazakhstan is a useful case for analysis because widespread oil development started relatively recently. Although Soviet engineers knew about oil in Kazakhstan in the 1970s

(the Tengiz field was discovered in 1979), both technical problems in extraction and strategic decisions to focus on development of Siberian reserves (Thomas, 1985) resulted in very limited development and investment until the Tengizchevroil (TCO) deal in 1993,

193 and the start of exploration in the Caspian in 1992.60 The recent nature of this oil development, combined with political changes with the collapse of the Communist system, provide a stark backdrop for examining the role of oil in the political economy of

the country.

Throughout this chapter, Kazakhstan will be held in comparison to two other

former Soviet states, Moldova and Kyrgyzstan. Both of these states experienced very

rough transitions. Kyrgyzstan started its transition in a situation of ethnic unrest between

the Uzbeks and Kyrgyz. The Osh riots, which some estimate claimed nearly 300 lives,

were emblematic of these problems (Burke, 1991). At the same time, it has experienced

serious economic upheaval, with foreign aid making up 50 to 115 percent of gross capital

formation from 1994 to 2005 (World Bank, 2007). Similarly, with a patchwork of ethnic

groups, a short but bloody civil war, high levels of poverty, and an unreformed

Communist-led government in power, Moldova would seem an unlikely place for

political contestation. While neither of these countries had much in the way of natural

resources, both have generally remained more open and democratic than Kazakhstan, and

have experienced far more political contestation.

In Kyrgyzstan, early political reforms resulted in one of the most open political

environments ever seen in Central Asia. This was followed by a period of political

chilling as President Askar Akayev began to consolidate power in his hands. Akayev,

however, was later overthrown in the March 2005 “Tulip Revolution,” a popular uprising

started by local opposition groups in protest to the results of the parliamentary elections

(see e.g. Eurasia Insight, 2005). While some today characterize Akayev’s overthrow as

60 The current development group for the Tengiz oil field includes Chevron (50%), ExxonMobil (25%), KazMunaiGaz (20%), and Russia LukArco (5%). The Kashagan field in the Caspian was discovered in 2000. The current development team consists of Eni (the operating company, 16.81%), Shell (16.81%), Total (16.81%), KazMunaiGaz (16.81%), ConocoPhillips (9.26%), and Impex (8.33%). 194 more of a coup than a revolution, it remains emblematic of a much more competitive, although not democratic, governance.61

Similarly, in Moldova, political contestation was chaotic until 2001, when the

Party of Communists of the Republic of Moldova (PCRM) was given a majority in

Parliament (see March, 2006). Yet, despite efforts by the PCRM to consolidate power through the judiciary, media, and bureaucracy, they were forced into a coalition government and their power has continued to erode. Some authors have taken to calling

Moldova an example of failed or weak-state authoritarianism, where leaders, despite their desire to limit competition, do not have the resources for doing so (Way, 2002; Levitsky and Way, 2002).

61Visitors to Eurasianet's website, for example, gave the resulting government an F in four out of seven areas of seven categories, including a C in improved civil liberties and an F in constitutional reform (Eurasianet.org, http://www.eurasianet.org/kyrgyzstan/reportcard/index.shtml). 195 a -- map acquired from University of Texas Perry-Castaneda Library Map Collection.

Figure 6.2: Map of Moldova.

196 This chapter examines why Kazakhstan has become steadily more authoritarian, while politics in Moldova and Kyrgyzstan has remained competitive, if not fully democratic, since independence. It also asks why the authoritarian leadership in

Kazakhstan has been able to strengthen its position vis-à-vis opposition groups since independence. While some of this can be attributed to legacies of independence, such as the relatively unity of the Kazakhstani elite, much of the variation must be attributed to the resources that the Kazakhstani government has received from oil and gas investments.

These have funded public works projects and large-scale government investment, allowed for more arbitrary enforcement of property rights, and given Nazarbayev’s close associates dominance in the economic sphere. As such, despite having higher levels of economic development since independence, Kazakhstan has seen steady erosion in the groups capable of challenging government authority. As Olcott (2005 , p.30) summarizes it, “In many ways, Kazakhstan is the most puzzling of the Central Asian republics, because economic growth seemed to provide little incentive for political reform; in fact, the opposite was occurring.” Comparing Figure 6.3 gives visual representation to this puzzle. While Kazakhstan has substantially outperformed Kyrgyzstan and Moldova economically, it has also been the least competitive politically, and recent economic growth has coincided with lessened levels of political competition.

197 2000 1500 1000 Per-capita GDP Per-capita 500 0

1991 1993 1995 1997 1999 2001 2003 2005 Year

Kazakhstan Kyrgyzstan Moldova a – data from the World Development Indicators and presented in constant 2007 US dollars (World Bank, 2007).

Figure 6.3: Per-capita GDP in Kazakhstan, Kyrgyzstan and Moldova, 1991-2006.

I will make the argument in several parts that oil has allowed both economic growth and authoritarianism to take root in Kazakhstan. First, I will provide a short history of Kazakhstan, showing how Kazakhstani politics has become significantly less competitive since independence. Second, I will look at the role played by oil revenues towards consolidating the authoritarian regime in Kazakhstan, and how increasing revenues have allowed President Nazarbayev to extend his dominance over Kazakhstani politics. In particular, I will be looking to see if the mechanisms posited in earlier chapters are playing an important role in Kazakhstani politics. Finally, I will compare

Kazakhstan’s political and economic development with that in Moldova and Kyrgyzstan, 198 and show how greater fragmentation among business elites has resulted in greater

political competition in the latter countries, despite what many social scientists would view as more hostile social conditions.

The information presented here is based on fourteen months of fieldwork in

Kazakhstan and Moldova, as well as secondary research on these cases and on

Kyrgyzstan.62 During my fieldwork, I interviewed business elites, oil company

executives, opposition leaders, academics, lawyers and even a former president. While it

is tempting to look at this chapter, since it comes at the end of the dissertation, as a post-

hoc illustration of theory, the reader should be aware that very little of the theory

presented earlier was developed before my fieldwork. My experiences in these countries

were critical in shaping my hypotheses and models. Also, while the information gathered

from interviews was critical in shaping the research, I will be utilizing publicly

documented evidence wherever possible, both because many of my interviewees,

especially in Kazakhstan, wished to remain anonymous, and so that others may confirm

my findings.

6.1. Political contestation in Kazakhstan

Kazakhstan’s independence from the Soviet Union came as something of a

surprise, even to the leaders of Kazakhstan. President Nursultan Nazarbayev had been a

strong supporter of a new Treaty of Union until events in the latter half of 1991 forced his

hand. Kazakhstan did not declare its independence until December 16, 1991; the last

Soviet state to do so.

62 IRB numbers 2005B0199 in Moldova and 2005B0209 in Kazakhstan. 199 Unlike many of the former Communist states, Kazakhstan’s leadership remained relatively intact during the transition. Nazarbayev became the Kazakh Communist Party

first secretary in 1990, and was later confirmed by Parliament in the title of President. In

the elections of December 1991, Nazarbayev become the first post-Soviet president of

Kazakhstan with about 98 percent of the vote. The Parliament, elected in 1990, was

largely dominated by people loyal to Nazarbayev and/or the Communist Party. This,

however, does not mean that politics in Kazakhstan has always been non-competitive.

While Nazarbayev has remained stable at the top of the leadership, the level of competition he faces has changed substantially from 1991 to 2007.

Generally speaking, Kazakhstan’s post-independence politics can be divided into two time periods. The first, lasting from 1991 until the dissolution of the Majilis

(Parliament) in March 1995, was a period of relative pluralism for Kazakhstan. During

this time, Nazarbayev faced tough investigation from the media, political protests from

various ethnic and political groups, a rebellious Majilis, and multiple defections of major

political leaders to the opposition. The second period, which continues through today, is characterized by an increasing centralization of power in the hands of the Presidency, not only in terms of formal institutional powers, but also in terms of informal economic influence. It is in this latter period, also characterized by quickly rising fuel revenues, that checks on presidential authority quickly dissolved, aided by what can best be

characterized as a “family-run state” (Olcott, 2002: 2). This pattern suggests a dynamic

component to the theory developed thus far. Figure 6.4 takes the heuristic table presented

at the introduction. A natural question to ask of the static theory presented this far is whether states can move between these categories (as the arrows indicate). The next two sections draw on the case of Kazakhstan to argue that increased oil revenues can indeed 200 increase the stability of authoritarian governance, suggesting that dynamic implications

can be drawn from the model developed earlier.

Low Income (< $1,000) High Income (> $1,000)

Limited Stable Moderate Dependence Resource ? Democracy (< 10% of GDP) Impact

? ? ?

High Dependence Vulnerable Stable (> 10% of GDP) Regimes ? Authoritarian

Figure 6.4: Can the static model produce dynamic predictions?

This section looks at Kazakhstan’s political development in greater detail, arguing that political competition has largely been stymied in Kazakhstan. The next section will then demonstrate the link between rising fuel revenues and increased authoritarianism in the case of Kazakhstan.

201 6.1.1 Limited pluralism, 1991-1995

The early period of Kazakhstani politics could perhaps best be characterized as a

time of bounded political competition.63 The post-independence constitution, not adopted

until January 28, 1993, was largely a continuation of Soviet-era institutions. Most of the

power was concentrated in the hands of the president. The constitution gave him power

to issue decrees, resolutions and instructions that were “compulsory on the whole

territory of the Republic” (Article 79), to “cancel or suspend wholly or partly the

realization of acts of the Cabinet of Ministers, ministries, state committees and

departments” (Article 78, Section 3), and to reverse actions taken by local government

officials (Article 93). The power of the president was further consolidated by the control

given to him over local government. Kazakhstan was declared a unitary state, and local

governments had a limited range of activities. Within these local governments, elected

councils, called maslikhat, were subordinated to regional governors, akims, who were, in

turn, appointed by President Nazarbayev (Article 92).64

On the surface, the powers given to the president by the constitution, along with a

generally loyal political elite from the Communist past would have seemed like a blank check for Nazarbayev’s agenda, but the 1993 announcement of Kazakhstan’s privatization program quickly turned the Parliament against itself, splitting between those who supported and those who opposed privatization. As it had in Russia, Parliament and the President found themselves at a head, and, in December 1993, the Parliament voluntarily resolved itself. Given events in Russia during October of that year, where

63 This chapter will largely adopt the anthropologists’ convention of distinguishing between “Kazakh” which references the ethnic group and “Kazakhstani” which references citizens of the country of Kazakhstan. 64 All references to the 1993 constitution are referenced to the official Russian and English translations of the document by the Republic of Kazakhstan (Blaustein and Flanz, 1994). 202 Yeltsin used tanks to disperse a disobedient legislature and then dismantled the legislators’ privileges, their decision to voluntarily step down and maintain some of their privileges is understandable (see e.g. Rubin, 1993; Hughes, 1993).65

The Majilis elections in 1994 were a tightly managed affair, with severe

restrictions placed on political competition. As a result, nearly two-thirds of the new

parliament had no previous political experience, while 90 percent of the new legislators

were officials in state or partially privatized organizations (Olcott, 2002, p. 102). Again,

this should have resulted in a relatively cooperative parliament, but it was not to be the

case.

Public opinion generally opposed privatization (Olcott, 2002), providing a focal

point for opposition in the legislature. In May 1994, the left-center opposition group

Respublika (Republic) joined with the Congress Party to pass a vote of no-confidence

against Prime Minister Tereshchenko. The vote passed 111 to 28, but Nazarbayev

refused to remove Tereshchenko until privatization was completed. A growing non-

constructive opposition formed around Respublika, re-named Otan-Otechestvo

(Fatherland in Kazakh and Russian), who called for both Tereshchenko and Nazarbayev’s

resignations. In October 1994, two of Tereshchenko’s ministers, Mars Urkumbayev

(minister of economics) and Vladimir Shumov (minister or interior), were charged with bribe-taking. After a month-long scandal, both Tereshchenko and his government

resigned, reportedly because Nazarbayev was frustrated by the slow pace of reforms

(Gall, 1994; Olcott, 2002).

65 The Kazakhstan Supreme Soviet was not the only group to step down voluntarily after the events of October. Russia’s regional legislatures also decided to voluntarily step down to avoid further confrontation. 203 The removal of the government due to pressure from the Majilis, was just the

highest profile example of a rare period of reasonably open political competition in

Kazakhstan. While it never reached the level of openness in neighboring Kyrgyzstan,

media in Kazakhstan in 1994 was relatively “free and vigorous” (Olcott, 2002, p. 104).

Major independent newspapers, including Karavan and Panorama, combined with several independent radio stations, and the independent television station, KTK, to give the parliamentary opposition news coverage. Even official media, while strongly supporting government positions, gave coverage to the opposition.

Nazarbayev’s difficulties in establishing political dominance in Kazakhstan are reflected in his difficulties, during this time, in developing a coherent political party to

support his power in government. While both the 1993 and 1995 constitutions of

Kazakhstan forbid the president from being a member of a political party (Article 77 in

1993 constitution; Article 43 section 2 in the 1995 constitution), Nazarbayev threw his

support behind several parties in an attempt to create a party with enough power to

supplant the Communist Party. In September 1991, Nazarbayev attempted to form a

socialist party as a base of support. This quickly fell apart and Nazarbayev began

supporting the already well-known People’s Congress of Kazakhstan. Again, the party

was not a reliable and successful base for Nazarbayev’s political ambitions, and it too

was abandoned. The leader of the People’s Congress, Olzhas Suleimenov, rumored to

have further political ambitions, was made ambassador to Italy in 1995 (Olcott, 2002). In

October 1992, Nazarbayev’s third attempt at forming a support party resulted in the

Union of People’s Unity for Kazakhstan (SNEK), later re-named the People’s Unity

204 Party (PNEK).66 Although Nazarbayev was not a member himself, he stated openly that

“[e]very honest citizen of the Republic of Kazakhstan must become a SNEK member”

(quoted in Petrov and Gafarly, 2001, p.39). SNEK experienced some initial success, receiving 40 percent of the vote in the 1994 Majilis elections, but it was unable to provide the base of support Nazarbayev desired. The party was the largest in the 1994 Majilis, but was unable to get its party leaders into the parliamentary speaker post, and several of the leaders of the party, Serik Abarakhmanov and Kuanish Sultanov, were rumored to have independent political ambitions.

With the Majilis in revolt, Nazarbayev’s base of power in the legislature withering, and presidential elections upcoming in late 1995 or early 1996, Nazarbayev dismissed the Majilis in March 1995 (The Economist, 1995). The legal grounding for this decision came from a surprising ruling by the constitutional court; ruling in favor of a journalist who claimed that the electoral districts for the 1994 elections had been disproportionately drawn and the votes improperly counted. Nazarbayev seized on the decision to claim that the Majilis had been illegally constituted and should therefore be dismissed.67

Nazarbayev created a People’s Assembly to work in place of the Majilis until new elections could be held. This body adopted a resolution delaying presidential elections until December 2000, which was later adopted in a nationwide referendum in April 1995 with 95.8 percent of the vote (Associated Press, 1995). A second referendum, this one in

66 SNEK is very close to the Russian word for snow, sneg. One of the jokes during the period when SNEK was falling apart was that “SNEK is melting.” 67 Some believe that the decision was actually part of a strategy to accomplish the dismissal of parliament. Indeed, the plaintiff in the case, Tatyana Kvyatkovskaya, went on to head the list of the presidentially- backed Otan party and to serve in the reconstituted Majilis (Olcott, 2002, p. 110). It should also be noted that the 1993 constitution, under Article 131, paragraph 3, allows the president or Chairman of the Supreme Soviet to raise an objection to a Constitutional Court decision, which then requires 2/3 of the Constitutional Court to override. That no such objection was raised, suggests that Nazarbayev did not overly object to the decision. 205 August 1995, adopted a new constitution, further strengthening Nazarbayev’s power and making the legislature much more of a consultative body (Reuters, 1995). Included in the

new constitution was the development of an upper-house, the Senate, structured in such a

manner that an override of presidential veto was virtually impossible.

Despite these new checks on parliamentary action and a well-stage-managed 1995 election to the Majilis and Senate, some points of contention remained. Nazarbayev’s pension reform program was initially rejected by parliament, almost resulting in a government crisis, since a second rejection would result in either the dissolution of parliament or the government. After some last-minute maneuvering, the bill was adopted the next day (Olcott, 2002). This would be one of the last acts of rebellion on the part of parliament, and it is with the 1995 parliament that one begins to see the turn towards a much less competitive political system in Kazakhstan.

Although Kazakhstan never reached commonly accepted definitions of free and fair competition, this early period in Kazakhstan’s independence is marked by a surprising amount of pluralism, both in parliament and in the media. Had presidential elections been held in 1995, Nazarbayev would probably still have won, but would have faced a difficult contest, even with the usual electoral manipulation. The next period of

Kazakhstan’s history is characterized by a consolidation of Nazarbayev’s position, and a general erosion of the opposition.

206 6.1.2 Consolidation of power, 1995-2008

After the 1995 dismissal of the Majilis, Nazarbayev began utilizing the government’s resources and patronage to weaken opposition. One of the first areas targeted was the media. Utilizing privatization as a justification, the government required that all television and radio companies bid for their licenses in order to use broadcast frequencies. The starting bids for these licenses were set at $150,000 for TV and $50,000 for radio. These higher limits caused the closure of 31 stations unable to raise the required amount (Olcott, 2002, p. 105). One of the president’s daughters, Darigha

Nazarbayeva, won the license for the national independent television broadcast frequency in this bidding. Similarly, in the print media, Karavan was purchased in a very generous buy out and transferred in stages to the Nazarbayev media empire. Panorama also came under increasing pressure, as authorities pressured its publishers not to publish the newspaper. In addition, pressure, arrests, and confiscation became a regular part of opposition print media. Mukhtar Abylazov, formerly a major player in transport, communications and consumer industry, was arrested and imprisoned on charges of corruption. His ownership of Channel 31 was then sold to Bulat Utemuratov, the current

Secretary of Kazakhstan's Security Council, at a discount rate. As discussed further below, the ability for the government to charge such high rates, and, more importantly, the ability of Nazarbayev’s close associates to pay these fees, stems from the importance of such close government connections for economic wealth and the government’s ability to arbitrarily enforce property rights.

The result of this media clampdown can clearly be seen in the last presidential election, held in December 2005. During the election, most independent monitoring agencies noted a large discrepancy in the coverage of the President and the opposition. 207 The best situation was in the print media, where three pro-opposition newspapers are

allowed to operate. However, even in this area of the media, there are numerous

problems. Two months before the election, for example, the printing press which printed

three opposition newspapers, Vremya Printing, decided to stop accepting these

publication contracts. Bakhytzhan Mukashev, editor-in-chief of Epocha newspaper, compared the decision of the printing press to drop paying clients to “bees versus honey or customs officers against bribery” (Yuritsyn, 2005). After a short hunger strike, the contract was picked up by another printing press with strong ties to the government. This allowed the government to receive advanced notice of possibly sensitive articles, and to confiscate them. This occurred three times -- twice during the election and once directly afterwards. The official reason for the confiscations was violation of Article 100 of the

Code on Administrative Violations, protecting the honor and dignity of the President.

One of these newspapers, The Zuma Times, was banned from publications at the end of

December for violations of this code.

Average citizens, however, get their news from television , which is dominated by close associates of the President. While having to leave her post as chairperson of the state news agency “Khabar” after being elected to the Majilis, Darigha Nazarbayeva still held a great deal of influence in the programming of both Khabar, which was run by her close associates, and KTK. Of all the television stations, several members of the opposition identified Channel 31 as the most accommodating to the opposition.68 Even this channel, however, was owned by Bulat Utemuratov, the current Secretary of

Kazakhstan's Security Council, who purchased it after Abylazov was jailed.

68 This assessment was repeated by two campaign workers in For a Just Kazakhstan, although both still felt that the station was biased in favor of the government. 208 The results of monitoring by the OSCE are telling. Table 6.1 summarizes the results of the OSCE's media monitoring. It shows that 49-77% of media stories were dedicated to the president on the major television channels. It also suggests a qualitative difference between the portrayal of Nazarbayev, which ranges from “neutral” to

“overwhelmingly positive,” and the portrayal of the main opposition candidate,

Zharmakhan Tuyakbai, which ranges from “neutral or positive” to “negative” (OSCE,

2005a). These results are bolstered by similar results from other organizations, including the CST monitoring mission.69 The only report that contradicted this pattern came from

the Kazakhstan Ministry of Culture, Information and Sport, which suggested that, among state media, only 26% of the articles on candidates were dedicated to the President, while Tuyakbai recieved 25% (Embassy of the Republic of Kazakhstan, 2005). This figure is highly suspect, however, as the CST and OSCE found very different results during the same period of time, it only covered state media, and it is unclear how coverage of the President acting in an official capacity was treated.70 As the OSCE

described Khabar's coverage, “[N]ews coverage of several events which were presented

as regular activities of the President was overly extensive, and the events resembled

campaign events” (OSCE, 2005b).

69 All CST reports on media coverage are available, in Russian, on Jennifer Wilson’s Kazakhstan elections website: http://groups.yahoo.com/group/ Kaz-Elections/files/. 70 According to his campaign, none of President Nazarbayev’s speeches or actions during the campaign were actually “campaign activities,” but were part of his duties in his capacity as president. The only items which were considered “campaign activities” were those conducted by his supporters. It is likely that the Ministry of Culture, Information and Sport was using this standard. 209 Coverage of Portrayal of Portrayal of Nazarbayev Nazarbayev Tuyakbai

Khabar 49% (but coverage of “neutral or positive” “negative, and…often presidential activities distorted.” “overly extensive” and “resembled campaign events.”)

TV Kazakhstan-1 59% "neutral or positive" "neutral or positive"

KTK 77% “overwhelmingly “mostly negative” positive” Channel 31 74% “neutral or positive” “slightly negative”

Table 6.1: OSCE assessment of media coverage in the 2005 presidential election.

It should also be noted that several major newspapers, television and radio

stations refused to run opposition campaign ads that were considered to violate rules

against insulting the honor and dignity of the president. Khabar, Eurasia ORT, and KTK

also refused to run some opposition ads because of a newly adopted policy against

running pre-election ads shorter than thirty seconds. The opposition ads in question were

around twenty two seconds in length. This is the only time I am aware of, where a

television channel has refused to accept payment for running an ad because it did not take

up enough of their airtime.

Beyond simple counts of the amount of time given to candidates, there were some

qualitative differences that are important to note. Satire and negative documentaries were

heavily focused on the opposition. For example, a puppet satire on KTK, modeled after

the British show “Spitting Image,” portrayed opposition candidates as fascists or

streetwalkers, and consistently portrayed Tuyakbai as a woman. Unlike its British

counterpart, there were no characters from the current administration (Marcus, 2005).

210 News stories also demonstrated subtle biases. Stories on the opposition were

usually limited to saying who conducted the meeting, where it was, and the general topics

discussed (the environment, the economy, etc.). There were rarely direct quotations from

the candidates or specifics about opposition proposals. At the same time, coverage of the

President and his colleagues was much more extensive and included direct quotations

from the people involved. There were also a number of highly negative news stories on

the opposition's campaign conduct, including numerous stories on the fears of regional

ministers that the opposition was going to cause post-election instability, and a news

story which claimed that For a Just Kazakhstan (FJK) had paid campaign workers in the

Atyrau district with forged currency.71 News stories often tracked with the President's campaign platform. For example, the portrayal of the situation in Kyrgyzstan was

heavily negative. Several Kyrgyz students, who were attending KIMEP at the time I was

there, said they were offended by the portrayal of their country in Kazakhstan's media.

Quotations by Kyrgyz citizens and lawmakers that were disparaging of the “Tulip

Revolution” often made it on the air without an opposing viewpoint. At the same time,

coverage of events in Kazakhstan was overwhelmingly positive.

At the same time that the media became more accommodating, Nazarbayev was

finally able to establish the hold he desired over Kazakhstani politics, starting with his

election win in 1999. This election victory was not without manipulation. Legislation

introduced on October 7, 1998, extended the president’s term from five to seven years

and brought forward the date for elections. The new election date was set for January 19,

1999, far too soon for opposition parties, who anticipated elections in December 2000, to

gather the support to mount an effective campaign. The one person who was well-placed

71 News broadcast on CaspioNet, November 9, 2005. 211 to challenge Nazarbayev, former Prime Minister Akazhan Kazhegeldin, was prevented

from doing so, because his holding of an unsanctioned election rally resulted in a

criminal record. Another major contender, the former mayor of Almaty, Zamanbek

Nurkadilov, was bought off with an appointment to head the Almaty oblast (district).

Nazarbayev won, against Communist Party leader Serikbolsyn Abdildin and two other relative unknowns, with 81 percent of the vote. The elections were heavily criticized by

the OSCE for “the refusal of the registration of two candidates because of minor

administrative convictions; a media environment inadequate for free electoral process;

and allegations of intimidation of voters in order to secure support for the incumbent president” (OSCE, 1999, p. 3). Indeed, the elections were considered so flawed that the

Office for Democratic Institutions and Human Rights sent out a memo on the third of

December, 1998 in which they proposed that the election be postponed.

The elections to the Majilis began on October 10, 1999. Previous to the elections,

Nazarbayev abandoned the PNEK to head a new party coalition, Otan (fatherland), with

former Prime Minister Tereshchenko. The move paid off, with the pro-Nazarbayev Otan and Civil Party gaining 57 percent of the vote and 80 percent of the seats. The OSCE

(2000, p.2) noted that these elections were undermined by:

(1) illegal interference by executive authorities; (2) unfair campaign practices by parties closely associated with existing power structures; (3) threats of bureaucratic, administrative, and judicial measures jeopardizing media operations; (4) bias by lower level election commissions for candidates and parties favored by regional and local officials; and (5) intimidation and obstruction of the electoral campaign of opposition parties and candidates. In particular, widespread violations during the vote count and tabulation of results...were a serious setback.

The observers also noted that the judiciary was unable to address these violations of law.

The results were to set the stage for further consolidation in later elections.

212 In the 2004 Majilis elections, the concept of controlled opposition entered

Kazakhstani politics. Nazarbayev's Otan Party won 60.61 percent of the list vote, with the opposition Ak Zhol Party receiving the second highest percent, 12.04 percent (OSCE,

2004). The surprisingly strong performance of an opposition party in the list votes, however, only resulted in one seat in the Majilis, which the leader of the Ak Zhol Party,

Alikhan Baimenov, refused to take for two years. Part of this was because Otan dominated the single mandate elections, winning 35 of the 67 single mandate seats. In addition, two putatively “opposition” parties gained seats in the Majilis. The Asar Party, headed by the President's daughter, Darigha Nazarbayeva, won four seats. The Agrarian-

Industrial Union of Workers (AIST) bloc, which included the Agrarian and Civil Parties, won 11 seats. These two parties were opposition in name only, and primarily represented attempts by Darigha and other business elites to ensure their influence on Nazarbayev though the legislature. In addition, 18 self-nominated candidates received mandates, but these candidates were already associated with Otan (OSCE, 2004, p. 24). The consolidation of the Civil Party and Agrarian Party into Otan on December 22, 2006, followed by the Darigha's Asar Party in early 2007, resulted in an expanded pro- presidential party, Nur-Otan (The Fatherland's Ray of Light). Nazarbayev had finally attained the strong legislative power he had fought for since independence. The last

Majilis elections, held in 2007, further confirmed this hold on power, with pro- presidential parties winning all of the seats. Nur-Otan won 88.41 percent of the vote.

The second closest party, the Social-Democrats, won only 4.54 percent of the vote, not enough to pass the electoral threshold (Central Electoral Commission, 2007). Combined with the single-mandate districts, Nur-Otan won all of the 98 mandates in the Majilis.

213 Similarly, the latest presidential elections, held in December 2005, demonstrated

the weakness of opposition groups. Although Nazarbayev’s 91 percent of the vote was

probably inflated, the underlying results suggest a real weakness in the opposition. The

margin of victory was much higher than Nazarbayev's previous vote share of 79 percent in the 1999 presidential elections, but the fairness of these elections was judged more

favorably (OSCE, 2005b).

Exit polls completed shortly after the election, demonstrate that the president’s

vote total was probably exaggerated, but they also suggest that he had relatively

widespread support over the opposition. As shown in Table 6.2, the official results do

not coincide with any of the exit polls conducted in this election. All of the exit polls

place the President’s share of the vote between 83 and 87 percent.

214 Nazarbayev Tuyakbai Baimenov Abylkasimov Eleusizov

Official 91.15% 6.61% 1.61% 0.34% 0.28% Results

Association 86.90% 8.41% 2.51% 1.40% 0.78% of Politologists and Sociologists (AS&P)

Center for 87.43% 8.54% 2.04% 1.10% 0.89% Political Technology

Eurasia 84.55% 9.58% 3.46% 1.43% 0.98% Rating

Gallup & 83.2% 9.9% n/a n/a n/a International Republican Institute (IRI) a – polling results drawn from Central Electoral Commission, 2006; Reuters, 2005; Megapolis, 2005.

Table 6.2: Result of polls, 2005 presidential election .

215 This discrepancy could partially be explained by methodological problems in the

exit polling, such as under-sampling rural areas. While it is impossible to completely rule these out without access to detailed methodological accounts, these polling agencies have proved sufficiently reliable in the past and it is unlikely that all four would be so far off.

Other methodological problems, such as individuals refusing to participate in the exit polls or lying about who they voted for, would tend to increase, not decrease, the vote share for Nazarbayev.

Even if the exit polls are more accurate than official vote totals, they still put

Nazarbayev's vote total above 80 percent. The higher than expected victory margin in these exit polls is not simply due to methodological problems with the exit polling, as the opposition contends. Rather, it looks like a large portion of the opposition’s supporters did not show up to vote on election day.

Table 6.3 shows the proportion of the vote garnered by each candidate, the percent of invalid ballots and the voter turnout, broken down by district. The indented numbers indicate where the proportion in that category is higher than the average for all districts. It is striking that in seven out of the eight districts where there was higher than

average voter turnout, Nazarbayev also received a higher than average vote share. In four

out of six districts where Tuyakbai receives higher than average vote share, the turnout is

lower than the mean.

If this were because due to stuffing and counting irregularities, it would tend to

depress the percentages for other opposition candidates, and the proportion of invalid

ballots as well. This is not the case. In three of the districts where there was higher than

average turnout and presidential vote share, the proportion of invalid ballots was also

higher than average, and in five of these seven districts at least one opposition candidate 216 also received higher than average vote share.

Absent polling data, it is impossible to say for sure what turnout among opposition supporters looked like, but these district-level results, when placed in the context of the pre-election and exit polling, seem to indicate that many opposition supporters stayed home. That voting turnout was lowest in Almaty city, which is commonly believed to be the strongest area of support for the opposition, is also suggestive of the opposition’s lack of mobilization. For a Just Kazakhstan contends that turnout in Almaty was even lower, around 30-32 percent.72

So, generally speaking, there were a number of irregularities in the election, but these irregularities are not enough to explain the scale of the opposition's defeat. Even more troubling for the opposition, it appears that they have a problem, not only in terms of the size of their support base, but also in their ability to mobilize their supporters into action. Given that the opposition seems to have trouble getting their supporters to the polls, the post-election statement that the opposition could “mobilize ten thousand people into the streets if it chose” seems rather hollow, and explains why, instead of street protests, the opposition chose to place 1,204 lawsuits in the Kazakh courts.73 As Ian

Bremmer noted, “[T]here's a lot of distance between revolution and the kind of stability that Nursultan Nazarbayev enjoys in Kazakhstan” (Bremmer, 2005).

72 This according to a press release from the press service of the central headquarters of For a Just Kazakhstan, entitled “Brief Overview of Major Systematic Violations on the Voting Day,” released on 7 December 2005. 73 This according to a press release from the press service of the central headquarters of For a Just Kazakhstan, entitled “Opposition Appeals Against Illegal Actions of Election Commission,” released on 14 December 2005. 217 218 A paradox exists in summarizing Kazakhstan's political development since

independence. At the same time as the OSCE has continually reported improvements in

the legal regulations governing elections, and the Kazakhstani government has given

repeated emphasis to the process of democratization in the country, political competition

has become weaker. While legal conditions may have improved, economic conditions

substantially recovered, and international monitoring persisted, Kazakhstan's politics has

come to be even more dominated by President Nazarbayev. How has this happened?

The next section argues that government resources derived from massive

international investments in the oil resources sector provided Nazarbayev and his

supporters with the resources for enhancing authoritarian control.

6.2. The Political Economy of Kazakhstani Authoritarianism

As with any case study, there are multiple reasons for the outcomes observed in

Kazakhstan. Among them, the persistence of clan networks (Schatz, 2004), the relative

unity of the Communist government after independence (McGlichney, 2003), lack of a

supportive political culture (Kennedy, 2006), and the maintenance of Soviet patronage

and judiciaries (McGlichney, 2003; Kennedy, 2006) have received the most attention

from political scientists. In past work, I have suggested that most of these explanations

have some validity (Kennedy, 2006). Of these explanations, however, the political

economy explanation would seem to be the most critical and does the best job explaining the dynamics of Kazakhstan’s politics since independence. Other countries had similar backgrounds in terms of clan networks, recent political history, and political culture, but experienced very different outcomes (e.g. Kyrgyzstan and Turkmenistan). In addition, only the political economy explanation provides a solid explanation for both why the 219 Communist elite remained relatively unified during the transition and for the dynamics of

competition after the transition.

I am not the first to suggest that access to fuel export revenues played a critical

role in the maintenance of authoritarian government in Kazakhstan, as compared to the

relative competitiveness of politics in neighboring Kyrgyzstan (see e.g. McGlichney,

2003), but the mechanisms which I posit are somewhat different than those previously

espoused, and seem to have more generality to other petro-states. Below, I will lay out

the case that oil revenues have allowed the government in Kazakhstan to increase direct government subsidies to buy political support, to be more arbitrary in enforcing the property rights of foreign and domestic investors, and to establish a business and political

interest environment dominated by Nazarbayev, his family, and his closest associates.

These characteristics, in turn, deprive the opposition of the resources and people it needs

to be competitive.

6.2.1 Government spending

The moves of Nazarbayev, for example, in dismissing the 1994 Majilis and ruling

by decree could not have been done without his relatively widespread popularity. Indeed,

Nazarbayev has remained the most popular political figure in Kazakhstan. In late 1995,

one poll showed 72.3 percent of Kazakhs and 55.1 percent of Russians as naming the

president as the one institution doing the most to solve the country’s problems (Olcott,

2002, p. 108). His popularity since that time has only increased. According to data

collected before the 2005 presidential elections, approval of Nazarbayev’s performance

as president ranges between 70 and 83 percent. A poll done by KazRating, the week

before the election, placed Nazarbayev’s approval rating at 76% (Lyubinaite, 2005). In 220 addition, a poll by the Russian VTsIOM agency completed in November found 82%

support for Nazarbayev (Risa and Weir, 2005). The U.S.-based survey group,

Intermedia, found 70% support for the President in its pre-election poll (Kimmage,

2005). Only one poll, published in the opposition newspaper Free Word, had the election

closer. It put the likely vote at 46% for Nazarbayev and 44% for the main opposition candidate, Tuyakbai.74 I, however, have yet to find an independent analyst who believes

the results of this last poll. In addition, the sheer magnitude of difference between its

results and those of every other major poll conducted in the region puts its results under

suspicion.75

Much of this popularity is attributable to increased government payments for

various projects, funded, in large part, by oil revenues. A look at Figure 6.5 tells part of the story. While government revenues fell across the former Soviet Union with the loss of patronage from Moscow, they fell much less severely and recovered faster in

Kazakhstan than in either Kyrgyzstan or Moldova. Kazakhstan’s general government consumption expenditure (in 2007 US dollars) per-capita was about 2.53 times as high as in Kyrgyzstan and 2.88 times as high as in Moldova.76 This spending has remained

higher during the entire period since independence, and the gap has increased exponentially since 2000. Partly as a result of this, those who are in occupations receiving direct or indirect government payment -- teachers, agricultural workers, and pensioners -- have seen their wages rise.77 Government grants have also spiked. Just

74 Poll obtained from press office of For a Just Kazakhstan, November 2005. 75 It is a common mistake among regional analysts to blindly believe what is printed in the opposition media. However, as one campaign worker for For a Just Kazakhstan warned me, these sources of news have their bias and are sometimes just as distorted as the pro-government media. 76 Calculated from World Bank (2007) data. 77 Interview with Mereurt Makhmutova, Director of Public Policy Research Center in Almaty, Kazakhstan, 17 October 2005. This point was also illustrated in a widespread campaign poster from the Nazarbayev campaign in the 2005 elections. The poster showed bar charts of pay for public workers from 1991 and 221 from 2003 to 2004, the number of grants to families with children increased by 44.4 percent, from 165,246 to 238,636 (Shokamanov, 2005). This data may underestimate the

difference in spending between Kazakhstan and other post-Communist governments,

since much of Kazakhstan’s government expenditure, especially in oil revenues, is “off

budget.”

600 400 200 Government Consumption Per-capita 0

1991 1993 1995 1997 1999 2001 2003 2005 Year

Kazakhstan Kyrgyzstan Moldova

A – data from the World Development Indicators and presented in constant 2007 US dollars (World Bank, 2007).

Figure 6.5: Government consumption per-capita, 1991-2006.

demonstrated a marked increase, especially since 2000. 222 Averting ethnic tension with oil spending, the city of Astana

One of the largest examples of off budget expenditure and of utilizing government

resources to avert popular dissent is the new capital city of Astana. At the time of

independence, Kazakhstan faced a difficult confluence of ethnicity, geography and

resources. The Russian population, which made up 40 percent of Kazakhstan’s total

population before the breakup of the Soviet Union (Petrov and Gafarly, 2001), was

concentrated primarily in the northern portion of the country. This area is also the

primary area of the old Soviet industry, which collapsed along with the fall of the Soviet

Union (Olcott, 1995, p.273). Because of this, Russians were disproportionately affected by the economic downturn of the early 1990s (Olcott, 1995, p.283). Additionally, as the

reader can see in Figure 6.6, the oil reserves that were to eventually bring in the money

for Kazakhstan’s recovery were primarily located in the East, near the Caspian Sea, and

in the South. The primary area for capital development was in the then capital city of

Almaty (formerly Alma-Ata). For a country as large as Kazakhstan, with land mass

approximately equivalent to all of Western Europe, geography worked against simple

solutions to the disparity of capital flows.

223 a – map created using Google maps with overlays from the University of Texas Perry Casteneda Map Collection. b – major oil and gas fields (blue) are labeled as of 1996. c – ethnolinguistic concentrations (green) are labeled as of 1993.

Figure 6.6; Map of major oil and gas fields and the primary concentrations of Russian populations.

224 The identity problems were further agitated by the not-so-subtle Kazakhification

of public life. In 1989, legislation was passed making Kazakh the official state language.

The legislation allowed some exceptions, giving the most heavily Russified areas fifteen years to comply. This deadline has repeatedly been pushed back, as the difficulties in enforcing the language change have become apparent. The 1993 constitution defined

Kazakh as the state language, with Russian given the ambiguous status as “language of international discourse” (Preamble). This is preserved in the 1995 constitution, with the additional guarantee that all state services will be made available in Russian (Article 7,

Section 1). Kazakh, however, has increasingly come to dominate education. As of 2000,

Kazakh was the language of instruction in 3,357 schools. This compares to 2,402 schools

that use Russian, and 2,153 that use both Russian and Kazakh. In colleges and universities, the use of Kazakh for instruction has also increased substantially (Olcott,

2000). Similarly, government positions, which were previously dominated by the

Russian elite, have now become reserved for those of Kazakh nationality. For example the position of president remains off limits to those who cannot demonstrate “perfect command of the state language [Kazakh]” (Article 41, Section 2, 1995 Constitution).

The position of chairperson in both the Senate and Majilis are similarly restricted to those with perfect command of Kazakh (Article 58, Section 1).

Between the economic situation and what some interpreted as hostility to their ethnic group, many Russians either migrated or advocated separation from Kazakhstan.

Migration of Russians was large, especially in the early years of independence. The

National Committee for Migration and Demography estimated that, from 1993 to 1998,

1.2 million people left Kazakhstan for Russia, including 854,000 Russians (cited in

Petrov and Gafarly, 2001). In Russia, Viktor Kozlov, a scholar at the Institute of 225 Ethnography of the USSR’s Academy of Sciences in Moscow, argued that Russia had a

just claim to territory in eastern Kazakhstan. Similarly, Aleksandr Solzhenitsyn

advocated Russian claims to northern areas of Kazakhstan (see Olcott, 1995, p. 261).

Within Kazakhstan, the Uralsk city Soviet seated a bloc that favored annexation by

Russia in the December 1989 elections. So salient were these concerns about ethnic

conflict in Kazakhstan that Nazarbayev said in 1991, “God grant that no one should stir

up Kazakhstan on ethnic grounds. It would be far worse than Yugoslavia” (quoted in

Olcott, 1995, p. 298).

In 1994 Nazarbayev proposed, and the parliament approved, the movement of the

capital from Almaty to the city of Akmola, in the northern part of the country. The name

of the city was changed from Akmola, sometimes translated as “White Tombstone” or

“Holy Place,” to Astana, which means “capital” in Kazakh.78 The new capital’s

construction was almost wholly funded through “off budget” money diverted from oil

contracts. The establishment of Kazakhoil to manage state assets in oil enterprises played

a key role in this. Since the company was officially private, it had to pay taxes on revenues, but the revenues it received from oil sales were not officially government revenues. This arrangement allowed the government to pay for projects, including

Astana, without officially registering them as uses of government budget revenues (Peck,

2004, p. 150).79

The result was a surge of economic activity surrounding the construction and

population of the new capital city. As of 2002, the construction in Astana accounted for

three percent of the country’s GDP and 20,000 jobs were created in the construction

78 The name of the Astana has changed its name four times in the past five decades, and there is some current speculation that it might be named after Nazarbayev in the near future. 79 Kazakhoil was later merged with the national oil and gas pipeline company to form Kazmunaigaz. 226 industry alone. From 1997 to 2002, some 600 new shops and 200 new cafes were opened

(Embassy of the Republic of Kazakhstan, 2002).

Major construction projects continue in the capital city. In 2003, the government

announced the opening of “Duman,” a 110,000 square meter amusement park featuring an aquarium, a 3-D movie theater, two restaurants, a mall, a theme park, a nightclub, a hotel, and what was described as “a local Las Vegas” (Embassy of the Republic of

Kazakhstan, 2003). In 2006, Nazarbayev unveiled plans for the Khan Shatyry (Khan’s

Tent), which can best be described as a “giant transparent tent” built over a part of the city. The tent, which will cover an area the size of ten football stadiums, is made to create a summer environment year-round in the often harsh Central Asian steppe.

According to government estimates, Astana has cost over $15 billion to construct, with even more to come (see Antelava, 2006). In the next five years, planned construction projects are expected to double the size of the city (Lillis, 2007).

The attempt to create jobs in the northern part of the country appears to have been successful. As Margulan Rakhimbekov, a company manager who moved to Astana to further his career, remarks, “There are lots of towns around Astana and they are developing in parallel” (quoted in Lillis, 2007). Indeed, while I was in Kazakhstan, there were several lengthy news reports leading up to the election talking about how the areas of northern Kazakhstan, once decimated by the fall of the Soviet Union, were now full of

jobs and opportunity. This has translated into greater support for Nazarbayev. A pre- election poll conducted by the Russian newswire Infotag reported that support for

Nazarbayev was strongest in the northern regions of the country, where 84 percent of the

population reported that they planned to vote for him (Infotag, 2005).

227 All of this spending would not have been possible without Kazakhstan’s massive natural resource endowments. According to official statistics, fuel exports make up about

72 percent of total exports and about 50 percent of government revenues. The ability of the government to spend for support has contributed to its much greater level of stability and the higher levels of public support that have been maintained by Nazarbayev.

6.2.2 Government patronage and investment

Another important aspect of Kazakhstan’s economic development contributing to the continued authoritarian governance is the patrimonial system of economic distribution

that Nazarbayev has established. The vast majority of political and economic power

flows directly from Nazarbayev, and a small circle that surrounds him. Judges, local

Akims (governors), and even university regents are all appointed and dismissed by the

President.80

On the economic front, while ownership remains opaque in many Kazakh

industries, it is clear that a large portion of economic activity is controlled directly or

indirectly by individuals close to the President. In 1997, the Institute of Research on

Contemporary Political Issues, a Moscow think tank, identified seven major business groupings in Kazakhstan (quoted in Dave, 2005). Of these groupings, two of them were

under the charge of Timur Kulibayev and Rakhat Aliev, sons-in-law of Nazarbayev, and

two more were held by government officials close to the President.81 There are serious

80 For example, judges at all levels serve at the pleasure of the president. The one exception is Constitutional Court appointees, who must be approved by the Senate. However, this check is relatively weak given the dominance of parties supporting the president and the ability of the president to appoints part of the senate. 81 Rakhat Aliev has recently fallen out of favor with the Kazakhstani government. He was recently accused of kidnapping and beating two bank officials, and an international warrant has been issued for his arrest. He is currently living in Austria (where he was serving as ambassador to Austria and the OSCE). His wife, Darigha Nazarbayeva, has also separated from him, cutting his major family connection. 228 doubts as to how independent some of the other groupings identified in this study were.

For example, several scholars have suggested that Timur Kulibayev owns around 36% of

KazKommertzBank, listed as an independent group in this study, through Central Asian

Industrial Holdings (CAIH) -- a holding company registered in Netherlands Antilles. He is also thought to have close connections to the Eurasia Group in a similar manner (Peck,

1995, p. 6-7). Bryan D. Stirewalt, a former adviser to the Kazakhstan Central Bank, sums up the situation very concisely: “Getting into any business, you had to have [the

Nazarbayev family's] permission” (quoted in Kramer, 2005). Similarly, “[a]necdotal evidence suggests that those with medium-sized and even small enterprises who lack

official connections are at a disadvantage from the start because of the difficulties of

getting capital and the need to purchase protection” (Olcott, 2002, p. 142).

Moving beyond anecdotes, survey evidence suggests that the role of state investment in business is higher than in either Kyrgyzstan or Moldova. The World

Bank’s BEEPS (2000) survey, conducted in 1999 through 2000, surveyed businesses and

asked what proportion of the firm’s fixed investment came from the government. The

surveyed businesses in Kazakhstan reported, as shown in Figure 6.7, on average that

10.89 percent of investment came from the state (with standard deviation of 29.61). This

is substantially higher than in Moldova, where businesses reported an average of 5.8

percent (standard deviation of 21.21), and Kyrgyzstan, where businesses reported an

average of 7.54 percent (standard deviation of 23.61). What makes these differences

even more interesting is that, as noted earlier, the size of the economy and the level of

foreign direct investment are substantially higher in Kazakhstan, which would normally

Interestingly, his fall from grace was followed shortly thereafter by the government reclaiming his partial ownership of Channel 31 (Institute for War and Peace Reporting, 2007) and MMG (Chazan, 2007), supporting the suspicions of many that he had a large ownership share in oil and the media. 229 lead one to expect government investment to be proportionately lower on average for

businesses.

Q38bSta

Blue = Kazakhstan, Red = Moldova, Green = Kyrgyzstan 0 2 4 6 8 10 Mean Percentage of Fixed Income from the State

a -- Q38bSta – What proportion of your firm’s fixed investment have been financed from each of the following sources over the past year? The State. (BEEPS, 2000).

Figure 6.7: Importance of state investment.

The ability for the government, and for Nazarbayev’s family to be involved in the

business sector to such a heavy degree is hinged on outside revenue flows. Government

investment, foreign direct investment and money from very profitable oil sales have

bolstered the government’s ability to maintain state/family influence over major

corporations, and, as oil prices have risen, to re-assert state influence in major economic sectors. Kazakhstan’s third stage of privatization, begun in 1996, was to be the stage in 230 which large-scale industrial, mineral, and petroleum assets were put up for privatization.

Originally, those assets which were deemed as of national significance were not to be sold, but shortage of government revenue resulted in some of these being offered for sale as well (Olcott, 2002: 139). While financial resources provided a significant impetus for privatization, the impetus was not so strong that contracts were offered on a competitive bidding basis. Rather, most contracts were decided on a case-by-case basis in deals with individual investors. This allowed the process to proceed in a very corrupt manner, and allowed interests close to Nazarbayev to gain access to major economic assets.

With the turnaround in Kazakhstan’s fiscal situation, driven in large part by rising oil prices, Kazakhstan has also changed its attitude towards privatization. In the oil sector in particular, Kazakhstan has begun demanding a larger stake for state ownership in development projects. This has included a law passed requiring that KazMunaiGaz, the state-owned oil company, hold a 50 percent-plus-one share equity stake in all new offshore oil and gas projects, and a right of first refusal for the sale of any oil assets by a private company. This was followed by attacks on the Eni-led operations in the

Kashagan field in the Caspian Sea, which began when Prime Minister Karim Masimov accused unnamed foreign companies of not living up to their contracts, and, in February

2007, with greater restrictions on foreign investment and requirements for training of

Kazakh workers (PRS Group, 2007). In addition, Chevron now faces $609 million in fines, and AES, a power generation company, is faced with $200 million in fines (Central

Asia Review, 2007). These steps are not entirely unforeseen, as the State Program for the Development of the Kazakhstan Sector of the Caspian Sea, published in 2003, calls for “import substitution” policies favoring Kazakh producers (p. 23), “eventual substitution” of foreign workers with “qualified local specialists,” and “national 231 company’s stake in the project (not less than 50%)” (p. 12) (Ministry of Energy and

Mineral Resources of the Republic of Kazakhstan, 2003). That these programmatic goals are being applied to already existing contracts is indicative of Kazakhstan’s increasing confidence in its access to revenues, independent of previous foreign direct investment.

On the individual level, the development of oil revenues allowed Nazarbayev and his family to accrue a huge personal fortune which has fostered their role in private industry, the media, and finance. The most famous example of this is the so-called

“Kazakhgate” scandal, in which Mobil (now ExxonMobil) and an independent merchant banker, James Giffen, collaborated in a system of kickbacks and offshore money laundering that benefited Nazarbayev and his family. The huge amounts of money that changed hands resulted in indictments against Giffen in the largest foreign corruption case in US history, and a guilty plea by J. Bryan Williams III, a former Mobil executive, on charges of evading taxes on $7 million in unreported income (Global Witness, 2004).

At least part of the money, about $1.1 billion, was discovered in a secret fund, held by

President Nazarbayev, in a Swiss bank account. The account was discovered after

Nazarbayev asked Belgium officials to investigate any illegal holdings of Kashegeldin, his then political rival. The Belgium officials did not find such funds for Kashegeldin, but did find accounts of other Kazakhstani officials and referred the case to the Swiss

(Global Witness, 2004). Eventually, the Kazakhstani government admitted to the fund’s existence in April 2002. Then Prime Minister Imangali Tasmagambetov admitted that there was a fund containing “hundreds of millions of dollars,” but that the fund was set up by the President as an emergency reserve to preserve Kazakhstan’s independence. He went so far as to argue that the fund played a critical role in preventing the government from going bankrupt in 1997 and again in 1998. He continued by arguing that any 232 foreign bank accounts in Nazarbayev’s name were set up by the opposition to besmirch

his good name (Pannier, 2002). Even today, discussion of “Kazakhgate” is extremely sensitive to the administration. Several newspapers have been dismantled or had prints

confiscated due to running articles, even reprints of foreign press reports, dealing with

Kazakhgate. Such reports are considered to violate provisions protecting the honor and

dignity of the president. While Kazakhgate was discovered, it is highly unlikely in the murky world of Kazakhstani business negotiation that this was the only deal from which the Nazarbayev family personally benefitted. Such deals likely played a large role in

Nazarbayev’s family and his close acquaintances accumulating vast personal fortunes that were used during the privatization process to position themselves in key positions at the head of Kazakhstan’s economy.

All of this suggests that the role of the state in business is high in Kazakhstan, and is enabled by funds directly tied to oil development. Formally, this takes place by allowing the flexibility in revenues, allowing the government to directly invest in businesses and maintain state control over some major industries. Indirectly, this has taken place through corrupt practices that have allowed Nazarbayev’s family and close associates to accumulate vast fortunes that, in turn, have allowed them to position themselves in key positions in the economy.

6.2.3 Discretion over enforcement of property rights.

Just as important as the concentration of economic power in the hands of the close associates of Nazarbayev and his family, is the fungibility of economic power based on one's relationship to the administration. Although the right to private property is guaranteed in the 1995 constitution (Article 6), the enforcement of those property rights, 233 especially for domestic entrepreneurs has been uneven. The protection of property often hinges on connections to the government. For example, Zhamanbek Nurkadilov, once a close associate of the President and a major player in the construction and agriculture sector, fell out with the President and joined the opposition. Since that time, he has lost most of his previous political and economic power. During the 2005 presidential elections, he was found dead in his residence with two shots to the chest and one in the head, in what the Interior Ministry ruled a suicide. Mukhtar Abylazov, formerly a major player in transport, communications and consumer industry, was arrested and imprisoned on charges of corruption. His assets were sold at cut-rate prices to close associates of the

President. Similarly, Nazarbayev’s opposition in the in the 1999 presidential election,

Serikbolsyn Abdildin, and former Prime Minister Akezhan Kazhegeldin have both found their finances under official scrutiny (Olcott, 2002: 161). These confiscations and investigations are clearly tied to the fallout of these figures with Nazarbayev. More generally, as one high-profile consultant told me, “If your [joint-venture] partner supports the opposition, that's not a good sign. If he is well-connected in the government, that's very useful.”82

This instability in property rights is reflected in the BEEPS (2000) survey. One of the problems with normal business surveys on property rights and political risk is that they tend to be biased by recent economic performance, the so-called “halo effect” (see

Kurtz and Schrank, 2007a).83 Even more problematic for this study is that the government may do a relatively good job of protecting the property rights of those companies who support government policy and/or have connections with the government

82 Personal interview in Kazakhstan. Name withheld at request of informant. 83 There is some debate on this issue. See the debate in The Journal of Politics (Kaufmann et. al., 2007; Kurtz and Schrank, 2007b) . 234 and/or are large foreign investors, but arbitrarily violate the property rights of those who

do not support government policy. To counter this perception problem in using the mean

reported level of property rights protection, Figure 6.8 reports the standard deviation of a

battery of questions dealing with enforcement and judicial rulings in protecting private

property. The intuition being that if the government is being arbitrary in its enforcement

of property rights, there should be a greater diffusion of answers (Flores, 2008).

The results of this comparison generally support the conclusion that the

enforcement of property rights is more arbitrary in Kazakhstan. Question 16a in the

BEEPS survey asks respondents whether the interpretations of rules affecting their

business are predictable. As the bar chart shows, there is substantially more variance in the responses in Kazakhstan, compared with either Kyrgyzstan or Moldova. This

disparity between the quality of laws and enforcement of laws is important. As a tax

lawyer who worked for a company representing many large firms in Kazakhstan noted,

while Kazakhstan has some of the best tax laws in the former Soviet Union, the

authorities are given wide latitude over the enforcement of those rules. This included, for example, reports of auditors who came into their investigations with a pre-determined

amount of taxes owed and conducted their investigations to meet that goal, as opposed to letting the collected data determine taxes owed.

Question 22 asked respondents how often they would associate the country’s legal system with being “fair and impartial.” Again, Kazakhstan has the largest dispersion of answers, although the differences are not as great between it and Moldova in this particular instance. Unfortunately, the question does not further inquire why the respondent does or does not think the judicial system is fair and impartial. Literature on the court system in Moldova suggests a number of problems with impartiality, especially 235 when it comes to issues of libel laws in the media (see Oxford Analytica, 2007b), but

these do not seem to be as systematic as the judicial bias in Kazakhstan, and deal with general problems of corruption in Moldova.

Finally, question 23a asks businesses how confident they are that their property rights will be upheld. In this case, the dispersion of responses is much higher in

Kazakhstan than in Moldova, but closer to that in Kyrgyzstan. Again, this suggests that

protection of private property is problematic in Kazakhstan, and that there is a difference

in how property rights are enforced.

236 Q16a

Q22

Q23a

Blue = Kazakhstan, Red = Moldova, Green = Kyrgyzstan 0 .5 1 1.5 Standard Deviation of Responses a -- Q16a – To what degree do you agree with the following statement? “Interpretation of regulations affecting my firm are consistent and predictable.” 1 = fully agree, 2 = agree in most cases, 3 = tend to agree, 4 = tend to disagree, 5 = disagree in most cases, 6 = strongly disagree. b -- Q22 – Now, thinking about your country’s legal system, how often do you associate the following description with the court system in resolving business disputes? – Fair and impartial. 1 = always, 2 = usually, 3 = frequently, 4 = sometimes, 5 = seldom, 6 = never. c -- Q23a – To what degree do you agree with the following statement? “I am confident that the legal system will uphold my contract and property rights in business disputes.” 1 = fully agree, 2 = agree in most cases, 3 = tend to agree, 4 = tend to disagree, 5 = disagree in most cases, 6 = strongly disagree. (BEEPS, 2000).

Figure 6.8: Standard deviation of responses to enforcement and interpretation of regulations.

237 Taken separately, none of these questions provide definitive distinctions in the

enforcement of private property, but taken together they suggest that Kazakhstan’s

system of legal enforcement is much less predictable than either Kyrgyzstan or Moldova.

Ultimately, they suggest that the government of Kazakhstan is arbitrary in the

enforcement of property rights, in that some businesses, and particularly those with

strong government connections, receive better protection than others.

Combined with the anecdotal evidence on the effect of property confiscations, it

seems reasonable to argue that this has a chilling effect on opposition activities by

businesses. Uncertain property rights decrease the influence of business on government

policy. Indeed, the reasonable expectation is for government actions to have a large

effect on business behavior, but for businesses to have a relatively small influence on

important economic decisions. This is again borne out in the BEEPS (2000) data. Figure

6.9 shows the mean responses to questions asking businesses how much influence they

have over different levels of government on decisions affecting their businesses. In all of the branches of government – executive, legislative, government ministries, and regulatory agencies – businesses in Kazakhstan report having less influence. The stars indicate the level of statistical significance, and, as you can see, most of these differences are also highly significant. In particular, Moldova scores substantially higher than

Kazakhstan in all these areas of business influence. Kyrgyzstan shows greater levels of influence in all these areas, but the differences do not appear to be significant in terms of influencing the executive. These results are consistent with a business environment where government officials, rather than private enterprise, hold the upper-hand in political influence. Cutler (2004) makes the case that, “The greatest fundamental restraint on democratization in Kazakhstan is continuing restriction on the growth of 238 socio-economic strata interested in and capable of supporting real alternative parties.”

a -- * p<.05, ** p<.01, ***p<.001 b -- Q34a -- When a new law, rule, regulation, or decree is being discussed that could have a substantial impact on your business, how much influence does your firm typically have at the national level of government to try to influence the content of that law, rule, regulation or decree? 1 = not applicable, 2 = never influential, 3 = seldom influential, 4 = influential, 5 = frequently influential, 6 = very influential. (BEEPS, 2000).

Figure 6.9: Level of business influence over important economic decisions

Even foreign oil companies, which, because of their access to outside support and expertise, would be less prone to adverse government actions against their interests, have faced difficulties in doing business in Kazakhstan when their interests are not in line with the government’s. Most recently, these conflicts have spread to the development of

Kazakhstan’s Kashagan reserves in the Caspian Sea, the largest oilfield development

239 since the 1960s. Here, the Eni-led development group was threatened with the loss of its

development contract by the government for failing to start production on the

government’s assigned timetable and running over cost (see e.g. Crooks, 2007).

Profitability in unstable property, the case of PetroKazakhstan

The situation of PetroKazakhstan (formerly Hurricane Hydrocarbons) is emblematic of both the problems companies face with regards to their property, and of how oil resources have enabled this behavior. This section gives an overview of the

PetroKazakhstan situation to illustrate many of the problems associated with property rights in oil producing states. Not only does the case of PetroKazakhstan demonstrate the arbitrary enforcement of property rights, but it also illustrates why oil producing countries like Kazakhstan do not need to worry as much about gaining a reputation for an unstable legal environment. Despite being targeted by the government and being forced out of the Kazakhstani market, PetroKazakhstan was immensely profitable and other companies competed for the ability to buy its resources. Put simply, the government was not punished for, and, in fact, substantially benefitted from, violating property agreements.

PetroKazakhstan had an inauspicious start. Olcott (2002, p. 156) cited the company as an example of how “[s]maller companies can be eaten alive by the conditions

of doing business in Kazakhstan.” Indeed, less than three years after its purchase of

Yuzhneftegaz’s development rights to Kumkol Munai in 1996, the company found itself

heavily in debt and on the verge of collapse. Lower oil prices and the natural monopoly

of the Shymkent refinery in the area, combined to place the company in default on its

loans. In an agreement to keep the company alive, they agreed to buy the Shymkent 240 refinery for a third of its shares and $54 million in cash. When the CEO of the company

stepped down in 1999, the company was a mess. They company’s stock was only worth

a few dollars, they were millions in debt, and millions more dollars were missing.84

The projections for PetroKazakhstan changed quickly with the entrance of its new

President and CEO, Bernard Isautier, and the rising oil prices of the early 2000s. He

brought in many new employees, including a new Executive Vice President, Marlo

Thomas, who, at over 300 pounds, was described to me as “a big bear of a man from

Canada” with a personality to match.85 The employees I interviewed attributed most of

PetroKazakhstan’s comeback to Isautier and Thomas’ aggressive business style and their

ability to get others to work as hard as they did. These same qualities, however, also led

to conflicts with the government of Kazakhstan. For example, Kazakhstan required all

oil companies in the country to turn over part of their production for sale at a greatly reduced cost to farmers for the harvest. PetroKazakhstan suspected that much of this production was being stolen by government officials and re-sold at market prices. To test this, they put dye into the products given to the government for the harvest, and later checked to see if they were being sold on the market. The findings supported their suspicions, and the next year PetroKazakhstan refused to provide oil for the government program, a decision cited by both the government and representatives in the Majilis, to suggest that PetroKazakhstan was not fulfilling its social responsibilities.86 As one

PetroKazakhstan executive explained it, “He [Isautier] was convinced that the company

could be run like a Western oil company. He would not pay bribes, and would only do

what was required by law. He was seen by the government [of Kazakhstan] as very

84 The executive who described this situation suggested that the previous leadership had essentially been “raping” the company resources. Name withheld at request of informant. 85 Interview with PetroKazakhstan executive. Name withheld at request of informant. 86 Interview with PetroKazakhstan executive. Name withheld at informant’s request. 241 combative and inflexible.”87

PetroKazakhstan’s actions, and its attitude towards government regulation,

eventually resulted in it being targeted by actors within the government. Legislation

passed in December 2004 called for the elimination of gas flaring in the production of oil

except under limited circumstances. PetroKazakhstan was singled out for violations of

this statute, although the development of PetroKazakhstan’s fields started well before the

law’s passage. Additionally, PetroKazakhstan was not close to being the largest in terms

of gas flaring. A company memo circulated by Isautier cited statistics putting

PetroKazakhstan in the bottom third of companies operating in Kazakhstan for gas flared

by production.88 In the second quarter of 2005, PetroKazakhstan reported a $1.30 per

barrel increase in administrative costs over the same period in 2004. Of that, $0.47 per

barrel represented new legal costs and $0.28 per barrel was due to decreased production

in response to the new flaring regulations (PetroKazakhstan, 2005, p. 8). In addition to

flaring, PetroKazakhstan faced a dizzying array of other lawsuits and regulatory actions,

laid out in the company’s second quarter 2005 report (PetroKazakhstan, 2005).89 In

2004, the company’s 50/50 joint-venture with Lukoil, Turgai, was subject to a tax audit for 2002-2003 and received a tax assessment of $150 million (p. 40). In April 2005, two of the company’s subsidiaries, PKKR and PKOP, were fined $73.6 million for transfer pricing assessments in 2002 and 2003. In late 2003, PetroKazakhstan’s joint venture

Kazgermunai received a tax assessment for $9.6 million and a transfer pricing assessment of $6.1 million (p. 41). The Kazakhstani government’s Agency for Regulation of Natural

Monopolies and Protection of Competition claimed two separate violations of

87 Interview with PetroKazakhstan executive. Name withheld at informant’s request. 88 Information given to me by a PetroKazakhstan executive. Name withheld at informant’s request. 89 I would like to thank a lawyer for PetroKazakhstan for pointing out the location of this list. 242 Kazakhstan’s competition law by group distribution companies, costing $32.7 and $97.3

million respectively (p. 41). After numerous court cases, the Finance Police, in April

2005, placed a $98.9 million lein on the property of PetroKazakhstan’s subsidiary, PKOP

(p. 42). In February 2005, the Kyzylorda Regional Oblast filed suit against

PetroKazakhstan’s joint venture, Kazgermunai, for failure to live up to infrastructure

agreements. The suit claimed $102 million related to infrastructure charges and interest

(only $31.2 million was actually related to the projects themselves, p. 42).

The Kazakhstani government was not the only one involved in forcing

PetroKazakhstan out of the market. PetroKazakhstan’s 50/50 joint venture with Lukoil,

Turgai, became a serious drain on the company’s resources. In May 2005, Turgai filed

suit for $188.2 million, claiming that PetroKazakhstan had produced crude originating

from Turgai’s license area. Aside from the strangeness of a company being sued by

another company of which it owns 50 percent, the lawsuit was also strange because all

subsoil property rights are guaranteed to the government by the 1995 constitution.90 That same month, Turgai also sued for 3.1 million barrels of crude oil, which it claimed should not have been shipped to the Shymkent refinery (p. 42). Another claim, in February

2005, asked for $18.3 million in damages related to curtailment of production in Turgai in 2004 (p. 43). Two more claims, in April 2005, were filed by Turgai with respect to the price paid by a PetroKazakhstan subsidiary, PKOP, for purchase of crude oil. Finally, in

June 2005, Turgai sued for $3.5 million for actions by PetroKazakhstan to invalidate certain export contracts.

PetroKazakhstan retaliated, taking Lukoil into international arbitration three times in Paris, Amserdam, and Stockholm (p. 43-44). The feud with Lukoil was damaging to

90 A lawyer for the company characterized the arguments made by Lukoil on behalf of Turgai “absurd.” 243 the company, causing corporate costs to rise by $2.2 million in the second quarter of

2005, compared with the same period in 2004 (PetroKazakhstan, 2005, p. 21).

Given the sheer immensity of the claims against it, and the, not unwarranted,

perception that the company was being forced from the market, Isautier and the company

decided to sell their shares on the Canadian Stock Exchange. Despite its legal problems,

the company’s stock price only experienced a relatively minor dip before the decision to sell was announced, as seen in Figure 6.10. After a bidding war over the company involving China’s CNPC and India’s ONGC, the company’s stock bounced back. On

October 18, 2001, the company called a meeting in which it advised its stockholders to accept an arrangement with 818 Acquisition Inc., an indirect wholly-owned subsidiary of

CNPC, under which they were to receive $55 a share.91 The total price of the sale for

CNPC was $4.2 million, the largest overseas acquisition by a Chinese company. But

even the sale was not without controversy, as PetroKazakhstan first had to defend its right to sell its share of Turgai without Lukoil receiving pre-emption rights.92 CNPC also

had to make some concessions to the Kazakhstani government for gaining a foothold in

the Kazakhstan market. This included the sale of 33 percent of the acquisition at a much

lower price to KazMunaiGaz, the state-owned oil company, including the right to joint

management of the Shymkent refinery and its products (Xinhua News Agency, 17 July

2006).

91 “Management Information Circular of PetroKazakhstan Inc.,” 16 September 2005. 92 Court of Queen’s Bench of Alberta, “PetroKazakhstan Inc v. Lukoil Overseas Kumkol B.V., 2005 ABQB 789.” Date: 20051025. Docket: 0501 13439. Registry: Calgary. 244 70

60

50

40

30 Closing price (US dollars) (US price Closing

20

10

0 8/13/1996 4/13/1997 8/13/1997 4/13/1998 8/13/1998 4/13/1999 8/13/1999 4/13/2000 8/13/2000 4/13/2001 8/13/2001 4/13/2002 8/13/2002 4/13/2003 8/13/2003 4/13/2004 8/13/2004 4/13/2005 8/13/2005 12/13/1996 12/13/1997 12/13/1998 12/13/1999 12/13/2000 12/13/2001 12/13/2002 12/13/2003 12/13/2004 Date a – data compiled from PetroKazakhstan’s investor information, made available on their website as of December 21, 2005.

Figure 6.10: Track of daily closing stock prices for PetroKazakhstan (in US dollars).

The motives attributed to the government for pushing PetroKazakhstan out of the

market are many. Some individuals I interviewed pointed to the attitude of

PetroKazakhstan CEO, Isautier, as one of the primary reasons. His decision to publicly

challenge government rules about oil distribution may have raised the ire of government

regulators. Unlike other CEOs, who preferred to strike a bargain with regulators, Isautier

regularly went to court and international mediation. PetroKazakhstan brought a number

of the government’s cases to higher court in very public battles.93 Others suggested that

93 One interviewee suggested that it was partially the propensity of the company to take cases to court that 245 PetroKazakhstan was targeted for a takeover from either its partner in the Turgai development, Lukoil, who would play by the rules set by the government, or by the

Kazakhstan government itself. These informants suggested that the purpose of the lawsuits was to drive the price of PetroKazakhstan down and either seize the property or buy it at a bargain price. From the perspective of these informants, Isautier’s decision to sell on the Canadian Stock Exchange was a brilliant move to avoid such an outcome.

Still others cited the government’s desire to control the country’s refineries as a major reason for the increased regulation. Indeed, after the CNPC takeover, the government of

Kazakhstan did receive a large portion of PetroKazakhstan’s resources, including the

Shymkent refinery, and Lukoil did receive greater control over the Turgai project. The net effect, however, was that the government of Kazakhstan gained much more pliant partners in these oil developments, control over the largest oil refinery, and sent a warning about the stability of oil property rights in the country.

The case of PetroKazakhstan illustrates the methods utilized by the government, both regulatory and judicial, for expropriating property, but perhaps more importantly, it illustrates why Kazakhstan, and other oil rich states, do not pay a penalty for having an unstable business environment. Despite all the instances of government expropriation, doing business in Kazakhstan, and especially in the oil industry, is extremely profitable right now. In fact, of the five interviews that I conducted with former PetroKazakhstan executives, one stayed on with CNPC, one wanted to get involved with Isautier’s next investment project, one wanted to get involved with another project just like

PetroKazakhstan (with the logic that he had gotten in too late to make the large money on the venture), one was ready to retire or take a break from work, and one did not know

hurt its relationship with the government. 246 what he would do at the time I interviewed him. That half of them wanted to stay in that

market or enter it in another manner, after all they had been through in the previous year, illustrates the argument made here. Because oil is such a high-rent commodity, even at much lower prices than is currently the global average, investment is unlikely to dissipate significantly despite the unstable business environment.

This also points out the problem with expecting foreign companies to exert a democratizing influence in oil rich states like Kazakhstan. Some analysts have suggested that foreign investors could utilize their influence in lieu of a strong independent domestic sector to encourage liberalization. The problem is that these companies have their own interest in not taking such actions. As Ottaway (2001, p. 53) argues, “Oil companies may be ‘organs of society,’ but they are highly specialized ones, and their strengths lie not in devotion to democracy and human rights but in finding, extracting and distributing oil.” Similarly, Global Witness (2004) has expressed its frustrations with encouraging foreign investors to divulge how much they pay for oil shares. This is because there is a competitive advantage to gaining a reputation for not sharing this information, especially in non-democratic polities. Perhaps the strongest statement on the subject came from Nurbulat Masanov, the president of the Kazakhstan Political

Science Association, during a personal interview. He argued that Nazarbayev was purposefully selling assets to foreign companies as a way to avoid creating strong domestic pressure groups. In his words, “Nazarbayev knows that when people become rich they want a say in their government.”94

94 Interview with Nurbulat Masanov, 4 November 2005. 247 This section argued that oil exports have allowed the government to be more arbitrary in its enforcement of property rights. Businesses, who have reason to worry about whether they will be able to maintain their property if they criticize the government or fund the opposition, are unlikely to substantially influence government decision- making or to become active in liberalizing the political atmosphere.

6.2.4 Results of government domination of domestic industry

The results of both direct government involvement in business and the arbitrary enforcement of property rights can be illustrated with reference to the 2005 presidential campaign. During the election, there were a number of instances where local Akims and the regents of universities, dependent on Nazarbayev’s patronage for their position and funding, posted campaign advertisements for the president “on their own initiative.” In

Almaty alone, there were large pictures of Nazarbayev and/or campaign banners place outside Almaty State University, the Kazakhstan State Technical University, the

Kazakhstan University of International Business, the Kazakhstan State Conservatory, and the Kazakhstan Pedagogical Institute. There was also a large discrepancy in the approval of billboard space for the opposition as compared to the government. At the same time as

100 billboards and lightboxes were approved for the President's campaign, local Akims gave permission for only 24 billboards and 9 banners for Tuyakbai.95 The overall

dominance of the President's billboards was noted by several international organizations,

including the CIS-EMO (Shidenove, 2005).

In addition, there were several instances of rectors using their influence to

encourage students to vote for the President. At a police training college in Almaty,

95 This according to a press release from the press service of the central headquarters of For a Just Kazakhstan, entitled “Statement of December 2, 2005,” released on 2 December 2005. 248 around 5,000 students were told that they could be expelled if they did not vote for the

President. At the Almaty School of International Business, it was reported that students

received a talk from the Dean, who advised them to vote for Nazarbayev (Buckley and

Ostrovsky, 2005). One colleague even reported being advised how important it was that

he and other parents vote for the President during a parent-teacher conference at his

daughter's school. Local officials campaigned tirelessly for Nazarbayev, utilizing both

their political influence and their own patronage networks, largely funded through

transfers from the central government. Advisor to the president, Yermukhamet

Yertysbayev, pointed out the importance of this support, especially in northern

Kazakhstan in a telephone interview with Kazakhstan’s Ekspress-K:

Why did nearly 82 percent of the voters who took part in the polls in North Kazakhstan Region vote for Otan…? Because the regional akimat administration) and all the region’s activists were engaged in organizational work night and day! The region’s Akim, Tair Mansurov, personally conducted some 200 meetings with voters during the [2004] parliamentary election campaign… Any speech by the president is published in brochure form in the region in massive print runs and is widely distributed among the public. It is not just distributed, but publicized and explained, and propagandists and canvassers travel to the remotest rural areas (Ekspress-K, 2005).

This amount of support among regional and local leaders has been a key part of

Nazarbayev’s increasing hold over Kazakhstani politics.

Government resources from investment, particularly in the oil sector, have given

Nazarbayev a degree of flexibility on social spending vastly superior to that of

Kyrgyzstan or Moldova. Seven months before the 2005 elections, the Majilis passed an additional $891 million in extra spending, specifically targeted at raising pensions, social spending, and state employee wages (Turkish Daily News, 2005). Indeed, as Respublika

(2005), pointed out, “however much the opposition promises voters to raise the standard of living, it loses out, a-priori, to the current president. He is, after all, able to increase 249 state budget spending by just over 100bn tenge [about 770m dollars], so as to boost

pensions, benefit payments and wages, replace education credits with grants, and so on.”

The importance of government connections in business and politics also deprives

the opposition of a number of resources, ranging from campaign funds to simple window

space. It was striking to note how many businesses had at least one poster for the

President in their window, while only a few had Tuyakbai posters, and some of these

were obviously placed over Nazarbayev posters by opposition supporters. Campaign

funding for Nazarbayev’s political activities is anything but transparent. One of his

advisors, Aleksandr Pavlov, argued in an interview that all activities were “financed only

from legal sources” including membership fees in Otan, donations, and revenues from

commercial activities in the party’s charter (Vremya, 2005). Although transparent

numbers are not available, the sheer scale of pro-government campaigns makes it clear

that it possesses a large advantage in resources. Indeed, one of the largest complaints of

the opposition is that businesses are afraid to offer support for the organization, and those

that do are often subject to official harassment.96

It is difficult to imagine how Nazarbayev would have been capable of achieving the consolidation of power that he has without the resources he has garnered from the

investment and sale of fuel resources. Even more importantly, explanations that do not

make reference to the increased price of and profits from oil, cannot explain how

Nazarbayev was able to increase his hold over Kazakhstan’s governing institutions over

time. Figures 6.11 and 6.12 illustrate the explosion of revenues from both fuel exports

and foreign direct investment. These patterns bear an uncanny resemblance in their

timing to the pattern of Nazarbayev’s political consolidation.

96 Interview with Adil Nurmakov, 11 November 2005. 250 80000 60000 40000 20000 Value of Fuel Exports Per-capita 0

1991 1993 1995 1997 1999 2001 2003 2005 Year

Kazakhstan Kyrgyzstan Moldova a – values calculated from the World Development Indicators and presented in constant 2007 US dollars (World Bank, 2007).

6.11: Value of fuel exports per-capita, 1991-2006

251 300 200 100 Value of FDI Per-capita FDI of Value 0

1991 1993 1995 1997 1999 2001 2003 2005 Year

Kazakhstan Kyrgyzstan Moldova a – values calculated from the World Development Indicators and presented in constant 2007 US dollars (World Bank, 2007).

6.12: Foreign direct investment (net inflow) per-capita, 1991-2006.

252 Billions in lightly regulated revenues have fueled the corruption, cronyism,

spending, and economic discretion that have played a key role in raising Nazarbayev’s

popularity, and hindering the operation of the opposition. If revenues had remained

lower, the decreased revenue, along with the high cost of oil development in Kazakhstan,

would have most likely resulted in only limited gains in revenue and economic growth, resulting in much less stable governance.97 In the next section, Nazarbayev’s politics of

plenty is compared against two countries without the benefit of these resources,

Kyrgyzstan and Moldova.

6.3 Political developments in Moldova and Kyrgyzstan

Neither of the comparison states used in this analysis have had an unambiguously democratic record since independence. Figures 6.13 and 6.14 show the trends in

Moldova and Kyrgyzstan according to Freedom House (2007) from 1991 to 2006, and compares these trends with those in Kazakhstan. Figure 6.13 looks as the level of political rights (open elections, political competition, etc.), with scores closer to 1

indicating greater political freedom, Figure 6.14 does the same with civil liberties

(freedom of speech, organization, press, etc.). In both of these it is clear that the level of

political and civil freedoms since independence have generally been higher in both

Kyrgyzstan and Moldova, although there are some substantial changes that are worth

noting.

97 In the early part of its oil development Auty (1997) predicted that the cost of oil development would limit its utility for Kazakhstan, unless prices increased substantially. As can be seen by the government spending chart above, increased prices since 2000 have resulted in an upsurge of government consumption. The flexibility that this allows in crisis situations was recently demonstrated when the US housing crisis caused a liquidity crisis in Kazakhstan’s banking system. The ability for the Kazakhstani government to respond effectively is largely attributable to its revenue reserves from oil development. 253 6 5 4 3 Freedom House Political Rights Score Rights Political House Freedom 2

1991 1993 1995 1997 1999 2001 2003 2005 Year

Kazakhstan (PR) Kyrgyzstan (PR) Moldova (PR)

6.13: Freedom House political rights score, 1991-2006 (Freedom House, 2007).

5 4 3 Freedom House Civil Liberties Score Liberties Civil House Freedom 2

1991 1993 1995 1997 1999 2001 2003 2005 Year

Kazakhstan (CL) Kyrgyzstan (CL) Moldova (CL)

6.14: Freedom House civil liberties scores, 1991-2006 (Freedom House, 2007).

254 Kyrgyzstan has undergone several shifts since independence. Figure 6.13

suggests that political liberalization occurred initially, although at a gradual pace.

Political rights peaked in 1994 and began declining in 1998. By 2000, Kyrgyzstan was receiving the same political rights score as Kazakhstan. In terms of civil liberties, Figure

6.14 shows that Kyrgyzstan liberalized very quickly, but also fell back in the mid-1990s.

Kyrgyzstan's civil rights score peaked at 2, almost fully free, the year after independence

and slowly became less free. By 1995, Kyrgyzstan was overtaken by Moldova, and, by

1998, it was again at a similar rating to Kazakhstan. Both of these scores improved

marginally in 2005, with the “Tulip Revolution” (which will be discussed more below).

These trends are in line with the qualitative literature, which suggests that Kyrgyzstan

went through an initial period of political opening, followed by a period of greater

authoritarianism, and then another period of minor liberalization (McGlichney, 2003;

Olcott, 2005).

Moldova has undergone a gradual liberalization on both the political and civil

front. Protection of civil liberties have not changed much in Moldova since

independence. This is consistent with the qualitative literature which suggests that

Moldova continues to struggle in areas such as freedom of press, as illustrated by its

dismal record in front of the European Commission on Human Rights, ruling against

Moldova over 70 times and resulting in more than 800,000 euros in fines (Oxford

Analytica, 2007a; Oxford Analytica, 2007b). Nevertheless, political rights in Moldova

consistently improved from 1991 to 1995, peaking at a level of 2, indicating near full

freedom in this area. In addition, despite warnings that the victory of the Party of

Communists of the Republic of Moldova (PCRM) in 2001 would result in the

consolidation of a non-democratic regime (see Mazo, 2004), the political rights score has 255 only dropped slightly.

Beyond the broad comparisons in these Freedom House numbers, it is clear from both the literature on these states, as well as my own work in Moldova, that their politics is much more competitive than in Kazakhstan. In both of these states, business interests have effectively formed opposition parties and unseated executives. In Kyrgyzstan, despite the increasingly authoritarian politics of Askar Akayev, and the growing personal fortune of both him and his wife, opposition groups were able to strip him of his power following a flawed parliamentary election. It would be difficult to locate similarly placed groups in Kazakhstan. In Moldova, no government was able to stay for more than one term until the PCRM successfully took power in 2001. Yet, despite having superior resources, party discipline, and a loyal base of support, the PCRM had to form a coalition government with their arch-nemesis, the Christian Democrats, after the 2005 elections, and have most recently seen their numbers slip in the 2007 local elections and public opinion polls (Oxford Analytica, 2007b).

This section of Chapter 6 will take a somewhat closer, although still relatively superficial look at these cases, with a focus on two questions. Why has politics in both

Moldova and Kyrgyzstan remained relatively competitive, while it has not in

Kazakhstan? Can the changes in politics in these two countries be explained by the state resources framework utilized in the rest of this study? The contention will be that a large influx in aid resources in Kyrgyzstan helped Akayev to consolidate power, but both the restrictions on aid usage and its relative limitation, prevented him from consolidating his power in society in the same manner as Nazarbayev. Similarly, the lack of resources in

Moldova has consistently undermined efforts to consolidate power. This contention has serious implications for recent aid programs to Moldova by the EU. 256

6.3.1 Kyrgyzstan – liberalization, authoritarianism, revolution

Kyrgyzstan’s exit from communism was not nearly as orderly as Kazakhstan’s.

Unlike in Kazakhstan, where the elite remained relatively intact, Kyrgyzstan’s independence followed shortly after the Osh riots, and deep divisions had already formed among the elite. Reflecting this, the first president of the country, Askar Akayev, was not the first secretary of Kyrgyzstan’s communist party, but rather the president of the

Kyrgyz Academy of Sciences. Only when neither the President of the Council of

Ministers of the Kyrgyz SSR, Apas Jumagulov, nor the First Secretary of the Communist

Party of the Kyrgyz SSR, Absamat Masaliyev, were able to gain a majority from the

Supreme Soviet for election to the newly created post of president, was Akayev put forward as a compromise candidate.

Partially as a reflection of his uncertain status in office, Kyrgyzstan underwent a dramatic period of liberalization. The media in Kyrgyzstan during this time has been described as “virtually untrammeled and decidedly raucous” (Olcott, 2002, p. 104).

Western leaders loudly praised both Akayev and the regime that was created in

Kyrgyzstan. US Deputy Secretary of State Strobe Talbott called Kyrgyzstan a

“remarkable and very promising and admirable” country, and characterized Akayev as “a true democrat” (quoted in Cranston and Green, 1994). Similarly, the 1992 State

Department survey of world human rights conditions ranked Kyrgyzstan as the best of the former Soviet republics (Cranston and Green, 1994). The characterization of Akayev as the “prodemocratic physicist” was widespread, and not totally unwarranted

(Mandelbaum, 1994, p. 9). As a result of this, Kyrgyzstan was targeted by Western nations for massive international aid. Figure 6.15 shows this rapid expansion of foreign 257 aid, which, at certain points, made up more than 100 percent of gross capital formation in the country (World Bank, 2007).

60 40 20 Value of Foreign Aid Per-capita 0

1991 1993 1995 1997 1999 2001 2003 2005 Year

Kazakhstan Kyrgyzstan Moldova a – values from the World Development Indicators and presented in constant 2007 US dollars (World Bank, 2007).

Figure 6.15: Aid inflows per-capita, 1991-2006

By the mid-1990s, however, the “island of democracy” in Central Asia, as

Akayev (quoted in Smith, 2005) would describe Kyrgyzstan, had begun to be pulled in the same superpresidential direction as its neighbors. In the parliamentary elections held in 1995, Akayev himself admitted to irregularities, telling Interfax that the elections

“were free and open but not entirely fair” (Deutsch Presse-Agentur, 1995). Western observers noted that elections were run by local Akims, appointed by Akayev, and that 258 these leaders were manipulating the results to favor supporters of the president (Levine,

1995). Much like other Central Asian leaders during this period, Akayev attempted to avoid his own electoral challenge by passing a referendum extending his term until 2001.

Despite collecting 1.2 million signatures in favor of the referendum, the upper house of parliament voted that the action would be unconstitutional, and presidential elections were scheduled for December 1995. The elections were set for December 24, but competition against Akayev was minimal. Indeed, as of November 28, Akayev was the only registered candidate (Interfax, 1995). Not surprisingly, Akayev won the election easily with 73.9 percent of the vote.

Consolidation of Akayev’s power continued, with the president increasingly using referendums to subvert the power of the parliament. In the 2000 parliamentary elections, the OSCE noted “a series of negative trends,” including “a high degree of interference in the process by state officials, a lack of independence of the courts resulting in a selective use of legal sanctions against the candidates, and a bias in the state media” (OSCE, 2000, p. 1). The once heavily critical and independent press was also affected. The OSCE noted that the state media “showed an overt bias in favour of pro-government parties,” while the private media were “vulnerable to pressure from the authorities, creating a climate of self-censorship” (p. 2). Yet, despite all of these negative trends in Kyrgyz politics, human rights watchdogs argued that “although Akayev has failed to live up to the hopes engendered by his early political career, he has pursued a far softer authoritarianism than leaders in several neighboring states” (Freedom House, 2005).

The continued effectiveness of opposition to Akayev was dramatically illustrated in the 2005 “Tulip Revolution.” Leading up to the 2005 parliamentary elections, rumors abounded about Akayev’s plan for succession when his term expired in that year. From a 259 constitutional standpoint, he was not allowed to run for another term, but opposition

leaders feared that either he would abrogate such provisions (as Nazarbayev has done) or would hand power to either his son, Aidar Akayev, or his daughter, Bermet Akayeva, both of whom won office in the 2005 parliamentary elections. Shortly after the elections, which the OSCE evaluated as “more competitive than previous elections” but undermined by “widespread vote buying, de-registration of candidates, interference with independent media, and a low level of confidence in electoral and judicial institutions”

(OSCE, 2005, p.1), protests broke out in Osh and Jalal-Abad.98 The protests gained

momentum, and the government offered to start negotiations. Before negotiations could

take place, however, on March 24 protesters at a large opposition rally in the capital city

of Bishkek stormed the presidential compound, only to find that Akayev and his family

had already fled to Russia.

The implications of the “Tulip Revolution” are still in doubt. Instability and

protest after Akayev fled resulted in the death of three deputies, and the legislature voted to allow members to carry firearms for protection. By year’s end, Freedom House (2006) characterized the situation, saying “there was much talk about reforms and accountable government but little opportunity and very little progress towards fulfilling these pledges.” Akayev went even further, calling the actions by opposition leaders “a coup” that resulted in the downfall of “the island of democracy” in Central Asia (quoted in

Solovyov, 2006; Smith, 2005).

This very short overview of Kyrgyz politics is offered as a contrast to the politics of Kazakhstan. Kyrgyzstan’s initial liberalization and later retreat into soft-authoritarian practices stands in contrast to Kazakhstan’s path of authoritarian consolidation. Even

98 A colleague from Osh told me with pride during the leadup to the Kazakhstani presidential elections, “All revolutions in Kyrgyzstan start in Osh.” 260 more importantly, even during the period of Akayev’s increasing dominance over the

Kyrgyz political system and media, he was never able to establish the kind of widespread

support or political dominance enjoyed by Nazarbayev in Kazakhstan. What explains

these divergent trends?

The most obvious factor in all of this is the relative autonomy of Kyrgyz

economic interests, compared to those in Kazakhstan, and the more restrictive

dependence on foreign aid, rather than foreign investment. In terms of economic

interests, Kyrgyzstan’s economy during the period from 1991 to 2006 could best be

described as stagnant. Yet, despite the lack of growth, Kyrgyzstan’s economic

environment allowed business interests to develop without the necessity of supporting the

government administration. The events of the Tulip Revolution were largely the result of

protests by influential elites in impoverished regions. Adakhan Madamarov and

Kurmanbek Bakiev, who lost seats that they were expected to win and accused Akayev of

fraud, were particularly important in this regard. Olcott (2005, p. 5-6) testified to these new areas of independent influence in her testimony before the Commission of Security and Cooperation in Europe:

In his final few years in power former president Askar Akayev sought to restrict the power of key patronage groups, and as a result he drove more of the country's leading political figures into opposition... Economic reforms had led to a small group of independent businessmen, who while willing to pay some "tribute" also wanted market conditions to regulate economic opportunities.

Even after the events of 2005, influential business leaders continue to pursue their interests primarily through sponsoring opposition parties. The current president of

Kyrgyzstan, Kurmanbek Bakiyev, characterized the situation aptly, “If one has a business and can afford financing, one forms a party” (Interfax, 2006). One would be hard pressed to find similarly positioned business interests in Kazakhstan, where “[p]olitical patronage 261 is so fully controlled by the office of the president that those who seek to create an

independent political power base have little to offer to attract support” (Olcott, 2002, p.

95).

Another important factor in these diverging patterns of liberalization is the role of foreign aid versus foreign investment. McGlichney (2003) makes the persuasive case that Akayev was capable of establishing greater presidential powers largely due to the boom of Western aid to the country. This aid, in turn, permitted many of the same practices characterizing Kazakhstan’s oil economy. For example, weak control and monitoring allowed the government to divert much of the aid towards patrimonial programs that undermined the independence of the judiciary and local authorities. There are two large limitations, however, to utilizing international aid. First, the amounts are

usually far less than those generated by oil development. In Kazakhstan, foreign direct

investment alone reached over $250 per-capita, while, even at its peak, aid to Kyrgyzstan

only reached about $60 per-capita (World Bank, 2007). The current government accuses

the Akayev family of acquiring about $2 billion in corrupt capital during his time in

office (Interfax, 2005b). While this is certainly a nontrivial amount of money, especially in a country as poor as Kyrgyzstan, it is nowhere near the estimates for Nazarbayev’s family, some of which place him as the eighth richest man in the world (e.g. Evans,

1999).

Second, even when there is weak oversight, some regulation of aid distribution still occurs, limiting government action. The amount Akayev could redirect into patronage programs was limited by concerns that actions severely undercutting his government’s image would jeopardize future aid. Indeed, most of this aid was a reward for the economic and political reforms of the early 1990s, and was tied to continued 262 progress in these areas. Because of the concern about future aid, Akayev’s ability to restrict the opposition and use aid for patronage purposes was limited. Importantly, these distinctions, along with some selection effects since the collapse of the Soviet Union, suggest a reason for why aid and oil, although they both increase government resources, have such a divergent reputation in the political science literature. This is illustrated by two recent articles, one entitled “Does Oil Hinder Democracy,” and the other entitled

“Does Foreign Aid Promote Democracy” (Ross 2001; Knack, 2004 respectively).

6.3.2 Moldova – continuing competition

A first glance at Moldova would make it seem an unlikely place for democratic governance. Moldova's economy has yet to recover from the post-independence collapse.

Today, Moldova has the dubious distinction of being the poorest country in Europe. This is compounded by deep social divisions, which, although they cannot be reasonably classified as "ethnic" in nature, have still developed into competition among several national identities (Russian, Romanian, Gagauz, Transnistrian, etc.) (King, 2000;

Furman, 2007). The most obvious implication of these divisions found expression in the short, but bloody, civil conflict in Transnistria (which aligned itself heavily with the

Soviet Union and then Russia), and the autonomy agreement with Gagauzia (a Turkish- speaking region in the south of the country). A long literature in comparative politics has pointed to the dangers of economic collapse (e.g. Remmer, 1991) and plural politics (e.g.

Rabushka and Shepsle, 2007) for the survival of competitive governance. Resource scarcity, especially when combined with underlying national differences, usually results in a zero-sum politics uninviting for insitutionalized political competition.

Despite this, Moldova's politics has remained competitive, if not fully democratic, 263 since independence. In October 1990, Mircea Snegur of the Moldovan Popular Front was elected president, and under his leadership, Moldova declared independence in

August 1991. Divisions, however, quickly appeared among the membership of the

Popular Front. Some of the Popular Front's support wanted immediate unification with

Romania. Although actively seeking Western support for Moldova, Snegur did not support immediate unification. Despite the split in his support, he was still able to win in the December 1991 presidential election, after running unopposed.

The February 1994 parliamentary elections were judged free and fair by observers and resulted in the return of most of the first post-independence leadership, with nationalist forces suffering serious losses (Angeli, 1994). In contrast, the presidential elections produced a surprise. Petru Lucinschi, the former First Secretary of the

Moldovan SSR, defeated Snegur in the second round of the elections. Again, international observers felt "confident that the citizens of Moldova have been able to express their will in the elections and that the results reflect the opinion of the voters on election day" (OSCE, 1996, p. 3).99

It is during the presidency of Lucinschi that many scholars believe Moldova faced

its most distinct threat of a stronger presidency establishing a less competitive political

system (Way, 2002). With the 1998 parliamentary election, again judged "as a whole

satisfactory" by observers (OSCE, 1998, p. 1), Lucinschi faced mounting opposition to

privatization and economic liberalization. Lucinschi argued that Moldova, like Russia,

Ukraine and the other former Soviet republics, needed a stronger presidential system to

push through needed reforms. He gave several justifications for these actions, arguing

99 In all of these descriptions, I am purposefully ignoring the results in Transnistria. The central government of Moldova, in Chisinau, has no de facto control over this area of the country. Rather, it has, since 1992, been controlled by a separate government in Tiraspol. 264 that the original constitution set forth for Moldova was meant to reflect conditions in

European countries, that it was "more European than Moldovan" and that Western

investors felt that the changes were necessary to carry out the needed economic reforms

and to guarantee investments (Botsan, 2001). In my interview with Lucinschi, he argued

that the reform was necessary because “The president is the only one elected from the

entire country, and people look to the president to solve the problems of the people…

But according to the constitution, the president does not have any mechanism to influence

legislative actions.” He also argued that parliamentary systems were only a characteristic of “highly developed democracies.”100

The strengthening of presidencies in post-communist countries was not unusual

during this time, but did not occur in Moldova. Indeed, Moldova was one of only two

post-Communist countries, along with Croatia, to weaken the formal powers of the

presidency. Rather than gaining political power, the parliament amended the constitution

such that the president would be elected from the parliament, rather than by popular vote.

Lucinschi's defeat on the issue was due to several factors. First, many viewed the effort

as hypocritical, since, as speaker of parliament, Lucinschi had argued for greater parliamentary authority vis-a-vis the president. Second, the constitutional structure of

Moldova allowed parliament to amend the constitution, giving them the power to weaken

the president’s constitutional authority (Nezavisimaya Gazeta, 2000). A largely ignored

item in studies of intensified presidential power, the ability for a recalcitrant parliament to

block the president from utilizing popular referendum was crucial in the outcome.

Finally, and perhaps most importantly, parliamentarians feared the president and the

100 Interview with Lucinschi, 22 January 2007. Lucinschi also downplayed the significance of the reforms, arguing that there was only one reform – the ability for the president to dismiss the legislature or government if they are deadlocked – and that this was just to allow the president to act as a broker between competing interests. He drew a comparison to the office of president in Portugal. 265 interests who supported him gaining too much power. As Igor Botsan, a long-time

Moldovan political observer, argues, parliamentarians did not want “all power in the hands of the president’s [economic] clan” (Botsan, 2001, p. 11). Lucinschi describes the situation differently, but the conclusion is similar. In our interview he talked about how parliamentarians were serving their “personal interests” of “maintaining their jobs and influence.” The lesson he drew from this is, “If you want to understand the reason for an action, look for personal interest.”101 The sentiment in both these cases was that the diversity of economic interests represented in the parliament at the time prevented the expansion of presidential power.

Members of international democracy organizations and human rights groups who

I talked with in Moldova, viewed February 21, 2001 as a turning point in Moldova's politics. Like all previous elections in Moldova since 1991, international observers ruled that these elections “met international standards for democratic elections, consolidating a trend that should continue” (OSCE, 2001, p. 1). This election was different, however, in that a well-disciplined and cohesive political party, the Party of Communists of the

Republic of Moldova (PCRM), was elected to a majority, winning 71 of the 101 available seats. This gave the PCRM enough power to elect its leader, Vladimir Voronin, to the presidency. This was the first, and only, time that a country in East Europe democratically elected an unreformed communist political party.

For some observers, the election of the PCRM spelled the end of pluralism in

Moldova. Mazo (2004, p. 3) wrote an article discussing how “a parliament dominated by one party has been able to exclude political opponents in Moldova in a way that was

101 Interview with Petru Lucinschi, 22 January 2007. President Lucinschi also argued that the current compromise between the PCRM and the CDPP was illustrative of how legislators were serving their business interests over their ideology or the interests of the state. 266 impossible under presidentialism.” Former President Lucinschi argued that, contrary to the working of the constitution, the PCRM victory, and the election of their leader as president, had resulted in “rule by one man.”102 The PCRM was accused of undermining the independence of the judiciary, freedom of the press, and the legitimacy of local and electoral administration.103 The PCRM even reinstated celebrations of the October communist revolution. Despite what was widely seen as an attempt to consolidate more authoritarian rule, pluralism in Moldova continued. In 2002, the Christian Democratic

People's Party (CDPP), the successor to the Popular Front, held massive demonstrations against PCRM efforts to introduce official use of the Russian language in education and public life. As many as 30,000 people took to the street in Chisinau, in protests that lasted from January 9 to April 29. Such demonstrations continued to play an important role in Moldova's politics, also playing a role in pushing Voronin away from the acceptance of the Kozak Memorandum, an agreement that would have brought

Transnistria back under control of the Moldova central government, but only under conditions that would have kept Russian soldiers stationed on Moldova's territory and would have delayed any ambitions for European integration indefinitely.

Despite efforts by the PCRM to consolidate its control over the government, its support steadily eroded from 2001. The 2005 parliamentary elections were criticized on several grounds: “[T]hey fell short of meeting some [commitments and other international standards] that are central to a genuinely competitive election process. In particular, campaign conditions and media access were not satisfactorily equitable, and in this regard, the negative trends noted already in the 2003 local elections were confirmed”

(OSCE, 2005, p. 1). Yet, despite complaints of media bias, intimidation, and bias in the

102 Interview with Petru Luchinschi, 22 January 2007. 103Interview with Lilia Carasciuc, 2 June 2006. 267 Central Electoral Commission, the PCRM failed to win a large enough majority need to

unilaterally re-elect Voronin president. In a surprising agreement, the PCRM and CDPP,

previously mortal enemies, agreed to the election of Voronin as president in exchange for

a commitment to European integration and repaired relations with Romania. The Red-

Orange coalition has been a surprising and unique feature of Moldova's politics.

The continuing erosion of the PCRM's support was observed in the latest

nationwide local elections in 2008. Despite spending 83 percent more during the

campaign than their closest rival, the Party of Social Democracy, the PCRM lost about 22

percent of its local council seats and 25 percent of its municipal seats. In addition, in the

capital city of Chisnau, the heavily supported PCRM candidate was defeated by 28-year-

old human rights activist, Dorin Chirtoaca. Although opposition parties remained heavily

fragmented, ballots in the Gagauz region ran to over four feet in length, it seems clear

that the PCRM has not been able to consolidate its advantage in either the regions or the capital city.

Freedom House (2006) provides an excellent statement of Moldova's political history since independence: “Since declaring independence in 1991, Moldova has been one of the most pluralistic post-Soviet states, even if at times it has oscillated between nonconsolidated democracy and nonconsolidated authoritarianism.”

Much like in Kyrgyzstan, the consistently cash-starved Moldovan government never had access to the resources necessary to consolidate authoritarian governance.

Corruption in Moldova has always been a problem, but it remains primarily petty and disorganized, caused by the lack of public pay and oversight.104 Rumors about the

personal dealings of Voronin's family are almost laughable compared with those of

104 Interview with Lilia Carasciuc, 2 June 2006. 268 Nazarbayev's family holdings, including, for example, that Voronin's son took part

ownership in a local pizza chain, Andy’s Pizza, and had a plot to corner public

transportation by phasing out unofficial shuttles (Cioranica, 2007). Business interests in

Moldova remained diffuse, especially in large cities, while regional governors and local

administrators remained strongly independent.105

A recent explosion of aid to Moldova from the European Union, making Moldova

the second largest recipient of EU aid behind the Palestinian territories, should encourage caution from this comparison. Some observers have noted a disconcerting centralization of finances in the hands of the central government. This has combined with Voronin's threat to not work with local governments run by the opposition, to reinforce the point that Western donor states must make sure that aid does not go to centralize power.

Nevertheless, Moldova's broad support for closer relations with the EU, as well as the growing disenchantment with Voronin and the PCRM, make it unlikely that political competition could be effectively limited.

6.4 Conclusions

This chapter has given an overview of politics in three former Soviet republics:

Kazakhstan, Kyrgyzstan, and Moldova. Despite similar institutional histories, these countries have experienced very different political outcomes. In Kazakhstan, wealth from oil development and investment has consistently strengthened the hand of President

Nazarbayev. The massive inflow of revenues and resources allowed Nazarbayev's government to assert itself in economic distribution in three ways. Directly, this allowed the government to increase subsidies to public officials and maintain patronage networks

105 For example, in Chisinau, the capital and largest city, despite heavy spending by the PCRM, a young mayor from the Liberal Party, which is strongly opposed to the PCRM, won election. 269 among local officials. More indirectly, the inflow of resources has allowed greater direct

government investment in the economy, both through increased importance of

government capital in private investment and direct government ownership. Finally, oil

resources have allowed Nazarbayev, his family, and close associates to gain a

commanding position in most major areas of Kazakhstan's economic production and

banking. This has left few independent areas of wealth from which the opposition can

draw for support. As oil prices have increased, Nazarbayev's initially vulnerable position

after independence has turned into an increasing dominance over Kazakhstan's politics.

Many would suggest a similar process has taken place in Russia, as Putin has turned oil

and gas revenues into a much more solid hold on political leadership than that enjoyed by

Yeltsin in the 1990s (e.g. Hanson, 2007).

Kyrgyzstan and Moldova, by contrast, have had much more competitive political

systems since independence. In Kyrgyzstan, international aid allowed Akayev to slowly

constrain an initially raucous media and civil society. However, the scope and

restrictions on this aid limited Akayev's ability to consolidate his position, and relatively

powerful business interests remained in opposition to the government. Thus, even though

Kyrgyz politics has not returned to the level of openness that it experienced in the early

1990s, it remains much more competitive than in neighboring Kazakhstan.

Moldova's politics has remained very competitive since independence. Freedom

of press, independence of judiciary, and building of an independent civil society have always posed problems. Yet, every election from 1991 to 2001 was judged to have lived up to OSCE standards of open competition. Even after an increase of concern, following

the election of the PCRM in 2001, the opposition remained successful in winning office

and influencing government action through protest. Indeed, the PCRMs position in 270 Moldovan politics has consistently eroded since 2001. Unlike in Kazakhstan, or even in

Kyrgyzstan, economic resources in Moldova remained diffuse. Business interests remain

relatively independent of government influence, and frequently support opposition

groups. Local leaders also remain relatively independent, often standing in virulent

opposition to the PCRM.

Taken together, these comparisons support the contention made in the previous

chapters of this study. Revenues from oil and gas developments result in a centralization

of resources, and these resources, when they are large enough, allow incumbents to

solidify their control over the political system. When these centralized resources are

relatively small, however, they form a basis for political competition. The results of this

chapter also suggest that this approach can be applied to other areas of outside funding, although these resources, such as international aid, tend to be smaller and carry restrictions limiting their impact.

271 CHAPTER 7

CONCLUSIONS

At the beginning of this study, I posed three inter-related questions for the study of politics in oil exporting countries: (1) Why are countries that export oil less likely to be democratic than their non-oil exporting counterparts? (2) Is this because oil provides the resources for the maintenance of authoritarian regimes, or because competition over state

resources makes democracies less stable? (3) What are the implications for social

wellbeing and socio-economic development in oil exporting states?

This final chapter returns to these issues and attempts to show what these findings

mean, both for policy-makers and scholars interested in oil politics. The first section will

review the findings and how they build on our previous understanding of politics in oil

exporting states, the second section will look at the policy implications of these findings,

and the third section will lay out an agenda for further research. A wide range of topics

have been covered in this study, but there is still much more that can be done to further our understanding of politics in oil exporting states.

7.1 Discussion of findings

This study has attempted to address a relatively broad topic, involving politics in

numerous countries. Some simplification is required in any such undertaking.

Nevertheless, this study has presented a number of interrelated arguments and, as such,

272 taking stock of the main arguments and their supporting evidence is worthwhile.

Writing on the effects of oil exports has a long and distinguished heritage, dating

back as far as Adam Smith’s Wealth of Nations, but, as shown in Chapter 2, the conclusions reached by these studies have varied dramatically. More troubling is that

studies of oil’s effects vary in a systematic manner, depending on the time period in

which they were written and the selection of cases. This study posited a few underlying

changes in oil markets that have driven these shifts in the literature. The first was the

decline in transportation costs during the last two centuries. When transportation costs

are relatively high, areas located near natural deposits of fuel resources can substantially

improve their productivity through markedly cheaper inputs without recourse to

technological innovation (Wright, 1990, p. 464-468; David and Wright, 1997; Irwin,

2003). When transportation costs are relatively low, however, profit margins from

exporting the raw material increase and will usually exceed those of selling to inefficient

local industries (Ascher, 1999). In non-industrialized fuel exporting states, this has

forced government subsidization to replace physical distance as the primary means by

which fuel costs are kept lower than in states without large fuel resources. Most of these

programs have proven highly inefficient, and have just as often been sources of cronyism

and patronage for the government.

The effects of the incentive to export fuel resources, rather than to industrialize

with them, were magnified by other historic events. Decolonization unlinked relatively

wealthy, industrialized states from areas that served primarily to produce raw materials.

The inability of many of these states to effectively utilize domestic raw material reserves

for industrialization led to the dependency literature of the 1970s. At the same time as

the dependency literature, which focused on the lack of capital in developing states, 273 evolved, the oil booms of 1973 and 1979 brought much greater attention to the group of

oil exporting states. When democratic reform and economic prosperity did not materialize as expected, the “natural resource curse” was born. As Chapter 2 demonstrates, countries with higher average dependence and countries with higher average income from fuel exports from 1965 to 2001 are substantially less democratic than we would otherwise expect from their other macro-characteristics (see also Ross,

2001). In other words, modernization theories, focusing on social development as a result of socio-economic change, and dependency theory, focusing primarily on capital and domestic ownership, both failed to take into account the importance of government discretion over economic distribution. This last attribute, associated with the non- diversified economies of developing states, is the primary characteristic of oil exports

that drives the findings of this study.

Despite the widespread acceptance of the authoritarian influence of fuel exports, it

is only recently that scholars have attempted to form a generalizable theory about the

effects of fuel exports. Unfortunately, as was demonstrated in Chapter 3, most of these

theories draw directly from the area studies research and bifurcate along the same lines

(e.g. Smith, 2004). Scholars studying the effect of fuel exports have tended to focus on either the stabilizing or destabilizing effects of these resources, and empirical support has been found for the results of both sides, depending on their frame of analysis. This is problematic because fuel exporting states experience a wide variety of regime outcomes.

The real question, which to this point has been largely ignored, is what characteristics of fuel exports result in such dramatic differences among these states.

274 Chapter 3 developed a theory based on the commonly accepted tendency for fuel

exports to accumulate resources primarily to the central governments of developing states

(e.g. Ross, 2006b; Jensen and Wantchekon, 2004). In these states, the economy becomes structured around government distribution, and this distribution takes on a formal structure in government grants, subsidies, and investment. It also has an informal structure through the influence of government leaders on business and the increased ability for the government to arbitrarily enforce property rights. Potential opposition groups become less likely to challenge the government because independent sources of wealth, outside of government control, are limited.

The capabilities derived from oil revenues, however, vary dramatically among countries (Karl, 1997, p. 18-19). From this perspective, those countries who are both dependent on and rich from fuel exports show greater stability than normal, because dependence breeds risk aversion, while the size of revenues should allow the government to make the necessary disbursements for stability. Stability is an especially strong characteristic of authoritarian regimes in fuel export rich countries, since these regimes

have the most discretion over allocation of private goods. In contrast, those dependent on

fuel exports, but not rich from those exports, are expected to show greater levels of

instability. While control over government resources becomes more valuable in these

states, central authorities do not necessarily possess the resources for maintaining

stability. This produces contradictory incentives for actors to lock-in their share of a

relatively scarce resource through the exercise of exclusive governance, and for

governments to cede some level of control when they do not have the resources to

maintain exclusive governance. It is in these states that we find the largest propensity for

alternation between democracy and authoritarianism, and the greatest danger of civil 275 conflict.

Utilizing a variety of methods and measures, Chapter 4 verified the tendency for

increased government control of economic distribution in fuel exporting states. It also

demonstrated that fuel export dependence is generally destabilizing, especially for

democracies, while fuel export income is generally stabilizing, especially for authoritarian states.

Thus, while fuel exports in developing states generally correlate with lower levels of democracy, the way that correlation comes about differs dramatically between countries made rich by their fuel exports and those who are only dependent on their fuel exports.

This theory also has implications for socio-economic development in fuel exporting countries. If, as the theory suggests, fuel income provides the government with the capabilities for increased distribution of private goods, it would also suggest that governments in fuel exporting states will tend to under-invest in public goods (education, health care, etc.). This combines with other factors, such as the volatility of oil revenue and natural lags in development, to result in a pattern of socio-economic under- performance. If modernization is about social transformation, it is likely that these social changes also lag behind growth in oil states. This is something that previous literature on the natural resource curse has not adequately addressed. Indeed, Chapter 5 demonstrates that the level of democracy in fuel exporting states is not nearly as different from their regional neighbors as it is from countries in similar income brackets. While these states have substantially higher incomes than their neighbors, their socio-economic development lags behind their per-capita income. In terms of education, poverty reduction and health care, countries with higher incomes from fuel exports underperform 276 compared with non-fuel exporting countries with similar per-capita wealth. Even more

interestingly, these results may under-estimate the real impact, since a number of fuel

rich states do not provide statistics related to many of these measures of social wellbeing.

For example, Saudi Arabia, Kuwait and Libya have never reported poverty headcount

data to the World Bank. Literacy rates have also been reported sporadically, at best.

Kuwait and Saudi Arabia have published literacy statistics three times since 1965, and

Libya has published them twice. Even GDP was not reported for Libya from 1988 to

1998 (World Bank, 2007).

Finally, the last chapter looked at a comparison of three post-Soviet countries:

Kazakhstan, Kyrgyzstan and Moldova. Despite many similarities in previous institutions,

ethnic heterogeneity and cultural background, these states have developed in very

different directions since their independence in 1991. Kazakhstan experienced a

relatively constant strengthening of semi-authoritarian politics, and, in particular, the

power of President Nazarbayev over the political system has increased substantially. The

strengthening of the authoritarian system under Nazarbayev is largely attributable to the

lack of independent power centers in the economy, enabled by investment in fuel reserves

following independence and rising fuel prices since 2000. After an initial period of liberalization, Krygyzstan saw a decline in civil and political liberties, largely due to the

establishment of patronage systems through the use of international aid. International

aid, however, offered neither the scale nor the flexibility for Kyrgyzstan’s leaders to

solidify their position like in Kazakhstan. As a result, Kyrgyzstan’s political landscape is

still dotted with a number of competing groups, although civil liberties and stability

remain elusive. Moldova is somewhat unusual, in that it has many of the attributes

usually associated with non-democratic governance, but elections have remained 277 relatively free and contested, though not totally fair, since independence. In part, the lack

of resources for the central government has resulted in multiple competing economic

interests. These competing interests are reflected in the diverse group of political parties

that are still able to effectively contest Moldova’s elections and the inability for any one

party, even the relatively unified Communists, to establish political dominance, despite

weakened civil liberties.

These cases seem to confirm two critical areas of the theory which could not be

addressed in the large-N framework: (1) they provide a more direct link illustration of the causal mechanism – government distribution does indeed seem to be one of the main

enabling forces for authoritarian dominance of the political landscape; and (2) they

suggest a dynamic relationship between fuel development and stability – increased

resources make authoritarian regimes more stable. The dynamics of this relationship are

difficult to verify in the large-N setting, given the number of mechanisms states have at

their disposal for income smoothing. Nevertheless, the dynamics developed in these

cases would seem to suggest that weak state authoritarianism, built on high oil prices, in

places like Russia, Kazakhstan, Turkmenistan and Venezuela will find their position

quickly eroded if oil prices fall by any substantial margin (Hanson, 2007).

Overall, the theory and empirical evidence developed in this study suggests the

natural resource curse is not illusory, but nor is it automatic. Indeed, the mechanisms

producing the natural resource curse findings seem to be regular and predictable. This

naturally leads to the question of whether the “curse” can be lifted.

278 7.2 Can we lift the curse?

No universal and easily implementable fix for the natural resource curse is likely.

Those familiar with the policy literature on this front will immediately recognize the

paradox arising from this study. Most recent research has centered around the establishment on natural resource funds (NRFs) for stabilizing government revenues and/or saving revenues from depleting oil supplies for development, investment or for future use. The philosophy behind NRFs is that they force leaders to treat oil revenues as public money, and bind them to long-term, responsible management of those revenues.

These funds have demonstrated some success in smoothing spending, although the results vary substantially from case to case (e.g. Davis et. al., 2001). Some of these funds, such as the Alaska Permanent Fund and Norway’s State Petroleum Fund have been highly successful in managing oil resources, and have developed substantial reserves. Many funds, however, have not been as successful. The Alberta Heritage Savings Fund,

modeled on the Alaska program, saw its savings deplete rather than expand because the

government spent 92 percent of oil revenues (Scoffield, 2006). Venezuela’s Stabilization

Investment Fund has failed to either protect the raiding of oil funds for temporary political purposes or to substantially increase the standard of living for most Venezuelans

(Karl, 1997).

Tsalik (2002, p. 19) suggests three factors for the success of NRFs: checks and

balances, transparency, and public involvement. All of these are in short supply in most

of the countries in which this study focuses. Still more troubling, the results of this study suggest that oil wealth may hinder the very processes that Tsalik argues are necessary for successful management of oil revenues. NRFs are one of the few mechanisms developed to keep governments from having discretion over the spending of oil revenues. Leaders, 279 seeking power as their primary motive, have an incentive to undermine best-practices

with regards to NRFs, and use discretionary power over the allocation of oil revenues to

expand their hold on power.

Any solution in developing countries will require bringing the incentives of government leaders to remain in office in line with responsible management of oil

revenues. Such a process is not uncomplicated. Kazakhstan, Turkmenistan, Venezuela

and Nigeria, not to mention most of the countries in the Middle East and North Africa,

have very strong executives with discretion over government distribution and revenues.

These states also tend to have influential groups benefiting from preferential government

treatment, and populations that do not have a strong sense of public ownership of natural

resources. In many of these states, constituencies for non-discretionary methods of

distribution will have to be developed, not found.

Standards for successful development of NRFs in developed states still apply:

independent monitoring of the funds, transparency of fund management, and binding

restrictions on fund use. In developing states, however, several additional

recommendations would seem necessary.

First, restrictions must be even more binding than is usually the case in developed

states. Alaska’s model of utilizing constitutional amendment to formalize the rules may

seem strict to many scholars of the developing world, since it would not allow

governments to utilize oil funds to make up for shortfalls in government spending. Yet,

the history of oil dependent states suggests that the opposite danger is much greater –

states become dependent on oil revenues to cover spending and resort to borrowing when

oil revenues run low. Even those with NRFs designed to smooth consumption over time,

such as Venezuela, consistently fall victim to the discretion, or lack thereof, of various 280 presidents who used the funds to purchase popular support and to a public demanding the

spending of oil revenues on current needs. Similarly, maintaining integration of resource

funds within the budget is a laudable goal (e.g. Davis et. al., 2001), but in states without

publicly responsible legislatures, too much discretion over fuel revenues, rather than not

enough control, is the larger danger. As such, independent entities for resource

management and limits on allocation of funds through the regular budget may be

preferable to maintaining revenue integration.

No binding restriction, even constitutional amendment, will work if it does not

have a broad constituency. Thus, it is necessary to develop broad support for the NRF’s

policies and independence. Norway is often held as the model for managing natural

resources. Indeed, Kazakhstan’s NRF is based on the Norwegian model. Yet most of the

success of the Norwegian fund has less to do with its structure and goals, as it is with the

restraint of the country’s legislators (Tsalik, 2002, p. 37). The process of allocating the

share of oil revenues to be deposited in the fund through an annual legislative process

would not work in most developing states, especially where the legislature is dominated by the executive. In some ways, Norway stands as a warning – even in a firmly established, wealthy, and diversified country whose NRF has been cited as a model for

resource management by experts from Alberta to Kazakhstan, the temptation to raid the

resource funds is strong, with political parties increasingly arguing for using oil funds for

current spending or tax cuts (e.g. Underhill, 2006).

Broad public attention and support is unlikely where demobilized groups see only

potential future gains (as in Norway’s savings for future pension outlays) or where

revenues are allocated for stabilization of the government during a future rainy day (such

as Venezuela’s stabilization fund). In this regard, Alaska’s model for direct distribution 281 would seem to be the best model for promoting public interest and accountability, by

combining current benefits with public investment in the future of the fund. Indeed, in a

1999 vote to gauge support for using the fund for offsetting Alaska’s budget deficit,

rather than for dividends, 83 percent of voters cast a “no” vote (Tsalik, 2002, p. 25).106

Some have even suggested a similar system for Iraq as a way to obviate some of the

factional tensions surrounding management of oil reserves, and to engender a spirit of

public ownership (The Economist, 2003). In the particular case of Iraq, bureaucratic and

stability shortfalls are so severe that such a solution may not be practical. But, in many

states such a model seems more politically feasible, even if not economically ideal.

There are restrictions on this model, however. It works best where the population is relatively low compared to oil revenues, allowing for the quick development of relatively large dividends.

The size of resources raises a third point. Results of this study suggest that countries receiving a relatively large per-capita amount of revenue from oil exports will have the least incentive to change oil management. Even in developed states, successful

NRFs are usually the result of crisis. In the 1970s, when debate over Alaska’s NRF began, about 40 percent of the workforce received some form of public support and the state faced impending financial crisis. Similarly, Norway faced acute problems with inflation, wages, and external debt prior to the establishment of its NRF (Tsalik, 2002).

In countries where oil revenues are high enough to maintain government stability, only increased social demands, outside pressure, economic crisis due to poor management, or some combination of these factors would seem likely to open the possibility of more

106 Similarly, the policies suggested by Ak Zhol in Kazakhstan’s 2004 parliamentary election were very similar to the Alaska model. The part won 12.04% of the party-list vote, although, due to Kazakhstan’s electoral rules, it only received one seat in parliament. 282 transparent management.

Reform seems much more likely in the more unstable dependent-but-not-rich oil

exporting countries. Even here, the window for reform may be slight. Democratic regimes, for whom dependence is especially destabilizing, face the largest incentives

toward establishment of a strong, broadly supported, NRF. Democratic spells of politics

also provide the best conditions for debate over and the adoption of binding NRFs. The

IMF, World Bank and others should take advantage of these windows of opportunity for

developing wide ranging debate and participation in the distribution of oil revenues.

Unfortunately, normal legislative mechanisms are likely inadequate for this process

compared with broader popular referendums. The results of this study should provide

added impetus for dependent-but-not-rich oil exporting democracies to develop popular

mechanisms for oil revenue management, since it demonstrates that cronyism becomes

destabilizing and dangerous for governments in these states, after only a short period of

time.

Finally, transparency is essential, even in countries with high oil revenues. One of

the most striking features of the data dealt with in this study is the lack of reporting of even basic measures of social wellbeing by a number of oil rich states. Even the reports of oil revenues are estimates, made with great trepidation. This lack of information is unacceptable and inhibits popular governance. NGOs have focused on a number of methods for overcoming this lack of transparency. The “Publish What You Pay” movement, encompassing over 300 NGOs worldwide, seeks to mandate that oil and mining companies publish the money that they give to governments for natural resource extraction (see www.publishwhatyoupay.org). However, oil companies have incentives to not report these earnings, especially when in competition for oil contract with 283 companies like CNPC who do not report such revenues. Governments face similar disincentives, both out of concern for domestic companies and because of increased competition for natural resources. Nevertheless, pressure on oil companies from NGOs resulted in the acceptance of substantial regulation in oil developments in Chad and could be successful elsewhere. Other efforts to educate reporters in oil exporting countries on how oil contracts work and what they should ask their officials may also contribute to a more informed public (e.g. Tsalik and Schiffrin, 2005; Shultz, 2005).

This study, however, also argues that international organizations collecting global demographic data (such as the World Bank, IMF and UN) and their member states should pressure oil exporting states to report their demographic statistics in an accurate manner.

As the United Nations Development Programme (2002, p. 25) has noted, “The Arab region suffers from a severe shortage of detailed data and information necessary to undertake a comprehensive examination of human development.” While these do not provide direct information about oil revenues and expenditure, they would provide an indicator of whether increased oil prices and revenues are resulting in concrete socio- economic gains for the population. Ideally, a process of independent monitoring would provide social groups with all the information they needed to assess the use of oil revenues at every stage of their development, but broad-based social indicators of wellbeing would at least provide a start at such assessment.

Public ownership and responsibility for oil resources would seem to be the only antidote to government dominance of economic distribution, but the opportunities for developing such oversight and responsible management are relatively rare. Policy- makers must remain diligent, consistently pressing for the information that can be used for public evaluation, and prepared to assist in broad-ranging debate and decision on oil 284 management when the possibility arises.

7.3 Directions for future research

As with any research of this kind, there are as many routes for new research

suggested by the results as there are answers to pressing questions in the literature. One

of the primary questions raised deals with the application of this distinction between

being dependent on and being rich from oil exports.

Much has been made of the link between oil exports and poor governance. Most of this literature has focused on how oil resources cause poor governance because of the deterioration of taxation bureaucracies in oil rich states (Chaudhry, 1997). The lack of stability in oil dependent-but-not-rich, states suggest another potential mechanism for this finding. As the formal model in Chapter 3 suggests, this group of states are the most likely to be destabilized by their reliance on oil resources. This yields an incentive to take current resources, rather than to invest them. Similar to the difference between roving and stationary bandits (Olson, 1993; McGuire and Olson, 1996), this would suggest that governance may be driven by a stability mechanism, rather than a revenue mechanism. Certainly, the influence of these mechanisms needs to be explored, especially with the controversy over how critical taxation is for government capacity

(Henry, 2004).

Similarly, in the area of corruption, the incentives for corruption, at least as a proportion of oil revenues, would seem to be larger in states where oil revenues do not

provide enough resources for stability. States that become wealthy from their oil exports

would seem to have greater incentive for investment that might yield higher returns over

the medium- to long-term. This suggests that the corrupting effect of oil revenues may 285 not be solely due to the size and non-transparency of such revenues, but also due to the

competition that they engender in unstable states.

A third area of interest deals with the perception of public ownership of oil revenues. To what extent are these attitudes present in oil exporting states? Are there programs encouraging this type of attitude? Unfortunately, the intersection of oil politics and political behavior is substantially underdeveloped. Yet, the political economy

literature on oil development consistently suggests that popular attitudes are important to

the success of oil management strategies. Exploration of these attitudes in the survey

literature may yield surprising insight into the possibility of successful oil management in

the developing world.

Finally, research into the effect of oil on popular demands needs to take place.

Chapter 5 suggests that the socio-economic conditions modernization scholars associate

with greater demands for popular participation may not develop in a proportionate

manner with oil-based growth. If these effects are also observable on the level of

individual demands and expectations, this would be a significant contribution to the

natural resource literature, and would bring this literature into closer alignment with

broader democratization theory.

7.4 Concluding remarks

Thornton Wilder’s analysis of wealth in The Matchmaker may hold some wisdom

for developing states with oil resources, “Money is like manure; it’s not worth a thing

unless its spread around encouraging young things to grow.” Similarly, the evidence

presented herein argues that the concentration of oil export wealth in the hands of central

authorities in developing states has had a deleterious effect on the representativeness and 286 stability of government, not to mention socio-economic development. Karl (2005, p. 21)

sums up the gloomy outlook for oil exporting states: “[C]ountries that depend on oil for

their livelihood are among the most economically troubled, the most authoritarian, and

the most conflict-ridden in the world.” These problems of distribution remain primarily a political problem, and are difficult to solve solely with appeals for responsible resource

management.

The most obvious implication of these findings is that the economics of oil

revenue needs to be changed in developing states. Oil revenue should be distributed in

such a manner that it does not result in dominance over economic distribution in society

by the current administration. Such mechanisms, however, are difficult to design without

a change in the political incentives of people with the power to make the rules. These

incentives are difficult to modify without some change in the external circumstance, such

as economic crisis, political crisis, and/or a change in public demands.

Transparency of oil revenues and a popular attitude of ownership with regard to

natural resources would seem to be the best antidote to the incentives of politician’s to

use the resources for cronyism and personal enrichment. Even the most thought-out

institutional mechanisms for oil revenues are unlikely to be satisfactory barriers to

political and personal ambitions, without popular oversight. NGOs, states, and

international organizations can all help to engender these attitudes through sponsoring

popular education and involvement programs dealing with oil development.

The fact that oil reserves have not always been considered a curse, and that they

provide resources in otherwise cash-starved societies, should give everyone hope that the

current period of the natural resource curse will one day be seen as a similar historical

exception. Some of these trends can already be seen as oil exporting states adjust to “the 287 strains imposed on institutional structures and mechanisms by the combination of the expectations of a better-informed and more knowledgeable public at home together with external forces such as globalization and increasingly rapid technological progress”

(UNDP, 2002, p. 5). The best hope for such development it seems will have to come, like the oil itself, from below.

288

APPENDIX A

LIST OF CASES IN DATASET

289 Country, Years

United States, 1965-2001 Russia, 1992-2001 Morocco, 1965-2001 Canada, 1965-2001 Estonia, 1991-2001 Algeria, 1965-2001 Cuba, 1965-2001 Latvia, 1991-2001 Tunisia, 1965-2001 Haiti, 1965-2001 Lithuania, 1991-2001 Libya, 1965-2001 Dominical Republic, 1965-2001 Ukraine, 1991-2001 Sudan, 1965-2001 Jamaica, 1965-2001 Belarus, 1991-2001 Iran, 1965-2001 Trinidad, 1965-2001 Armenia, 1991-2001 Turkey, 1965-2001 Mexico, 1965-2001 Georgia, 1991-2001 Iraq, 1965-2001 Guatemala 1965-2001 Azerbaijan, 1991-2001 Egypt, 1965-2001 Honduras, 1965-2001 Finland, 1965-2001 Syria, 1965-2001 El Salvador, 1965-2001 Sweden, 1965-2001 Lebanon, 1965-2001 Nicaragua, 1965-2001 Norway, 1965-2001 Jordan, 1965-2001 Costa Rica, 1965-2001 Denmark, 1965-2001 Israel, 1965-2001 Panama, 1965-2001 Guinea-Bissau, 1974-2001 Saudi Arabia, 1965-2001 Colombia, 1965-2001 Equatorial Guinea, 1968-2001 North Yemen, 1965-1990 Venezuela, 1965-2001 Gambia, 1965-2001 Yemen, 1990-2001 Guyana, 1966-2001 Mali, 1965-2001 South Yemen, 1967-1990 Ecuador, 1965-2001 Senegal, 1965-2001 Kuwait, 1965-2001 Peru, 1965-2001 Benin, 1965-2001 Bahrian, 1971-2001 Brazil, 1965-2001 Mauritania, 1965-2001 Qatar, 1971-2001 Bolivia, 1965-2001 Niger, 1965-2001 UAE, 1971-2001 Paraguay, 1965-2001 Ivory Coast, 1965-2001 Oman, 1965-2001 Chile, 1965-2001 Guinea, 1965-2001 Afghanistan, 1965-2001 Argentina, 1965-2001 Burkina Faso, 1965-2001 Turkmenistan, 1991-2001 Uruguay, 1965-2001 Liberia, 1965-2001 Tajikistan, 1991-2001 United Kingdom, 1965-2001 Sierra Leone, 1965-2001 Kyrgyzstan, 1991-2001 Ireland, 1965-2001 Ghana, 1965-2001 Uzbekistan, 1991-2001 Netherlands, 1965-2001 Togo, 1965-2001 Kazakhstan, 1991-2001 Belgium, 1965-2001 Cameroon, 1965-2001 China, 1965-2001 France, 1965-2001 Nigeria, 1965-2001 Mongolia, 1965-2001 Switzerland, 1965-2001 Gabon, 1965-2001 Taiwan, 1965-2001 Spain, 1965-2001 Cent. African Republic, 1965-2001 North Korea, 1965-2001 Portugal, 1965-2001 Chad, 1965-2001 South Korea, 1965-2001 Germany, 1990-2001 Congo (Brazzaville), 1965-2001 Japan, 1965-2001 West Germany, 1965-1989 Congo (Kinshasa), 1965-2001 India, 1965-2001 East Germany, 1965-1989 Uganda, 1965-2001 Bhutan, 1965-2001 Poland, 1965-2001 Kenya, 1965-2001 Pakistan, 1965-2001 Austria, 1965-2001 Tanzania, 1965-2001 Bangladesh, 1971-2001 Hungary, 1965-2001 Burundi, 1965-2001 Myanmar, 1965-2001 Czechoslovakia, 1965-1992 Rwanda, 1965-2001 Sri Lanka, 1965-2001 Czech Republic, 1993-2001 Somalia, 1965-2001 Nepal, 1965-2001 Slovakia, 1993-2001 Djibouti, 1977-2001 Thailand, 1965-2001 Italy, 1965-2001 Ethiopia, 1965-2001 Cambodia, 1965-2001 Albania, 1965-2001 Eritrea, 1993-2001 Laos, 1965-2001 Macedonia, 1991-2001 Angola, 1975-2001 Vietnam, 1965-2001 Croatia, 1991-2001 Mozambique, 1975-2001 South Vietnam, 1965-1975 Yugoslavia, 1965-1991 Zambia, 1965-2001 Malaysia, 1965-2001 Bosnia, 1992-2001 Zimbabwe, 1965-2001 Singapore, 1965-2001 Yugoslavia (Serbia and Montenegro), Malawi, 1965-2001 Philippines, 1965-2001 1991-2001 South Africa, 1965-2001 Indonesia, 1965-2001 Slovenia, 1991-2001 Namibia, 1990-2001 East Timor, 1999-2001 Greece, 1965-2001 Lesotho, 1966-2001 Australia, 1965-2001 Cyprus, 1965-2001 Botswana, 1966-2001 Papua New Guinea, 1975-2001 Bulgaria, 1965-2001 Swaziland, 1968-2001 New Zealand, 1965-2001 290 Moldova, 1991-2001 Madagascar, 1965-2001 Fiji, 1970-2001 Romania, 1965-2001 Comoros, 1975-2001 USSR, 1965-1991 Mauritius, 1968-2001

291

APPENDIX B

A NOTE ON STATISTICAL METHODS IN TSCS DATA

292 Time-series cross-sectional (TSCS) data -- also sometimes referred to as pooled

data, panel data or longitudinal data -- is some of the most promising and problematic

data used in the social sciences. On the one hand, it is the most promising way to both

increase the number of cases and to draw inferences that are generalizable across both

time and space. However, the handling of TSCS data raises a number of issues that vary widely across datasets. The data analyzed here presents several unique problems. This

appendix will discuss some of the controversies related to the use of TSCS data. It will

then discuss some of the unique challenges posed by fuel export data and explain why

this study chooses to focus on the between-country effects of fuel exports. Finally, it will

present some statistical tests to support the underlying logic of these decisions.

B.1 Common approaches to TSCS data

The most common approach for using CSTS data by statisticians and

methodologists is a fixed effects (within) model (Wilson and Butler, 2007; Acemoglu et.

al., 2007). The intuitive reasoning behind this model can be illustrated with a simple

example. In Figure B.1, there are two cases, measured over ten time points. The broken lines show the relationship between x and y within each country, while the solid line

shows the relationship predicted by ordinary least squares (OLS) regression. As can be

seen, the OLS model produces biased results, suggesting a negative relationship between

the dependent and independent variables, while both cases suggest a positive relationship.

The equation for fixed effects is:

293 ______( y − y ) = (x − x ) β + (ε − ε ) it i it i it i (1)

The fixed effects model solves the bias problem in OLS, due to country heterogeneity, by

estimating the coefficient as the weighted average of the coefficients within each case.

This is why it is sometimes referred to simply as the “within” estimator. The effect is

similar effect to adding a dummy variable for each case in the sample (least-squares with

dummy variables (LSDV), see Stimson, 1985).

Y

X

Figure B.1: Traditional example of OLS bias in TSCS data

The other part of any TSCS dataset will be the variation between countries, giving

rise to the “between” estimator. As the reader can probably deduce, the between

estimator is simply the inverse of the within estimator, eliminating the variance within cases and estimating differences between country means: 294 _ _ _

y i = α + x β + vi + ε i (2)

Generally, the between estimator has been relegated to a minor role in TSCS estimation unless substantive reasons guide us to estimate it.

The random effects technique reports the matrix weighted average of the within and between estimates, and estimating the error components as a random variables.

Mathematically this looks like the following (STATA, 2005, p.288):

_ _ _

(yit −θ y i ) = (1−θ )α + (xit −θ x i )β + [(1−θ )vi + (ε it −θ ε i )] (3)

2 2 where θ is a function of σ v and σ ε . The advantage to the random effects estimator is that it makes more efficient use of data, since it includes information about both within and between country variance. Its main drawback, however, is that it might not eliminate bias due to systematic unit heterogeneity. An assumption of the random effects estimator, and of OLS based techniques (see Beck and Katz, 1996), is that a unit increase in the independent variable(s) will have approximately the same effect whether it occurs over time or between units. As such, the standard method of testing for bias in the random effects estimator is to test its estimates against those of the fixed effects (within) estimator. While this is appropriate under a number of circumstance, the data analyzed here falls squarely within a subset of data characteristics that are problematic for fixed effects (within) estimation.

B.2 Problems with fixed effects (within) estimation for oil data

The exceptions to fixed effects have not been thoroughly explored in the political science methodology literature, since more often the focus has been on eliminating unit heterogeneity, but there are a number of potential pitfalls. First, if the independent 295 variable does not move, or moves slowly, within units, the fixed effects estimator may produce biased results. Figure B.2 gives a visual example of this. The dotted line shows the within effects estimator, which gives opposite predictions of the clear effect between units. Even if we assume that a one unit increase has the same effect within and between countries, the estimates can be incorrect where there is not enough within variance.

Plümper and Troeger (2007) demonstrate more generally that these problems can be substantial for TSCS estimation. This is clearly the case in the fuel dependence variable which has a ratio of between country to within country variation of 4.14, well above the level of 2.8 at which the fixed effects estimator begins having severe problems (Plümper and Troeger, 2007, p.136).

Y

X

Figure B.2: Bias in fixed effects (within) models when variables move slowly

296 Second, a variable may exhibit characteristics of what I label, for lack of a better term, “complex movement.” This is where a variable, usually an indexed variable

(against income, population, etc.), has two underlying reasons for temporal movement that might have dramatically different effects. In this dataset, fuel export dependence, the percent of merchandise exports consisting of fuel exports, exhibits signs of complex movement. A country can show greater dependence in one of two cases: (1) an increase in fuel exports, while merchandise exports remain relatively constant; or (2) a decrease in merchandise exports, while fuel exports remain relatively constant. These two reasons for change have very different implications for analysis. In the first case, the temporal movement is consistent with discussion of the effect of fuel exports. The second case, however, is more likely the result of other economic trends, such as economic collapse, which have very different implications for the likelihood of democracy. As Geddes

(1999) points out, democratization is more likely during times of economic trouble.

Thus, if whatever temporal movement that does occur over time in the fuel dependence variable is due to movement in merchandise exports because of economic problems, it would not capture the underlying of interest, and would produce very different hypotheses.

Finally, there may be substantive issues underlying differences in the within and between estimates. In a classic economics example, spending is considered a function of income. It seems logical, however, to suggest that the effect of a transitory change in income may not be the same as a change in average income, with the latter being much larger (e.g. STATA, 2005, p. 288). Lumping these two together would underestimate the effect of average income, and, depending on how the dataset is weighted, may produce misleading hypothesis tests. Similarly, the effect fuel export income, especially in 297 countries that receive a very high amount of income per-capita, will not likely be manifested primarily within short temporal proximity. As Karl (1997) points out, fuel export rich countries are likely to run capital surpluses, even during times of decreased prices. This difference between average changes and transitory changes in fuel export income is magnified by the inclusion of the fuel dependence variable, which, by capturing the fuel-dependent-but-not-rich group of countries, makes this variable even more focused on countries that are likely to run a capital surplus based on their average fuel export income, as opposed to being effected by short-term price shifts.

Some of this can be overcome by using long lags for the independent variables

(e.g. Ross, 2001 uses 5 year lags). However, these lags simply denatures the problems outlined above, and do not respond in a theoretically rigorous way to the proponents of the fixed-effects (within) model. The next section demonstrates the need to examine the differences between the between and within estimates and make the argument for using the between estimator within the context of random effects estimation for the two primary independent variables in this study.

B.3 Testing the consistency of between and within estimates

The normal method for testing for bias in TSCS models is to run a Hausman

(1978) test for consistency between the estimates of fixed effects and those of random effects. This section utilizes the Hausman test in the context of the models presented in

Table 2.1. Other models produce similar results, but, for space reasons, this is used as a typical example for illustration. Table B.1 compares the estimates produced by the fixed effect (within) estimator and the random effects estimator with the yearly variables in

Model 1 of Table 2.1. The Husman tests null hypothesis that the difference in the 298 coefficients is not systematic. As can be seen in Table B.1, this is not the case for the normal country-year models. The chi2 is 104.50 (p < .000) giving strong evidence that the null hypothesis can be rejected, and, by implication, that the difference in the coefficients is systematic. Even without the statistical confirmation, it is easy to see that there are substantial differences between the estimates of the two models – both fuel dependence and the log of per-capita GDP switching direction.

Within Random effects effects Standard coefficient coefficient Difference error Fuel dependence .015 -.001 .017 .002 Fuel income -.0001 -.0002 .00008 .000008 Ln(per-capita GDP) -.884 3.178 -4.062 .500 Post-Cold War 3.744 3.377 .367 .043 a -- H0: Difference in coefficients is not systematic. b -- chi2 = 104.50 c -- p = .000

Table B.1: Hausman test of the fixed (within) vs. random effects model

Those who militantly believe that only fixed effects (within) models produce unbiased results might take the results as evidence of bias in the model, but the analysis provided in the last section should encourage further investigation. To test whether these results are due to slow movement, complex movement, or some other type of systematic, and theoretically interesting, difference between average and temporal effects, I take a further test of the sources of the Hausman test’s discontent (see STATA, 2005, p. 307-

309). Breaking the variance of fuel dependence, fuel income, and GDP into their component parts, the average over the time period and the variance of each yearly value from the mean, the models were re-run and tested against both the within effects 299 estimator and the between effects estimator. Tables B.2 and B.3 demonstrate that the results in the last table were due to systematic differences between the between and within effects of these main variables. Table B.2 compares the deviation of the fuel dependence, fuel income, and GDP variables to the within estimator. These produce very similar coefficient estimates. In this case, the chi2 value is only 3.07 (p = .381) and the null hypothesis cannot be rejected under normally acceptable limits. Similarly, Table B.3 the comparison with the between estimator cannot reject the hypothesis that the differences in coefficients are not systematic. The chi2 value is only 2.94 (p = .817), which is well above a standard rejection level

Within Random effects effects Standard coefficient coefficient Difference error Dev. fuel dependence .015 .016 -.0002 .0002 Dev. fuel income -.0001 -.0001 -.000001 .000002 Dev. ln(per-capita GDP) -.883 -.977 .093 .065 Post-Cold War 3.744 3.756 -.012 .018 a -- H0: Difference in coefficients is not systematic. b -- chi2 = 3.07 c -- p = .381

Table B.2: Hausman test of the fixed (within) vs. re-specified random effect model

300 Between Random effects effects Standard coefficient coefficient Difference error Avg. fuel dependence -.072 -.061 -.011 .008 Dev. fuel dependence .395 .015 .380 .350 Avg. fuel income -.002 -.002 .0004 .0005 Dev. fuel income -.0007 -.0001 -.0006 .009 Avg. ln(per-capita GDP) 6.798 7.028 -.231 .155 Dev. ln(per-capita GDP) -14.726 -.977 -13.749 8.382 Post-Cold War 3.854 3.756 .099 1.327 a -- H0: Difference in coefficients is not systematic. b -- chi2 = 2.94 c -- p = .817

Table B.3 Hausman test of the fixed (between) vs. re-specified random effects model

All of this demonstrates that the differences between the within estimates and the random effects estimates is a substantive, not a statistical, artifact. As such, it seems more likely that the differences are due to the substantive reasons outlined in the last section than because of some type of unobserved country-level heterogeneity that only happens to manifest in fuel exporting states.

Two objections could be raised to the use of only the average of the fuel variables in the models for this study. First, that it does not give the full information available from the models, because the deviations do communicate some (even if somewhat trivial) information. This criticism has some grounding. The decision to concentrate only on the averages of the fuel variables in the study was made for stylistic reasons, namely to make the discussion clearer for the reader, and not to add confusing discussion about the technicalities of data distribution that would not have contributed much to the reader’s understanding of the problem. I am not the first to make such a decision, and, indeed,

Sachs and Warner (1995) use average fuel dependence in their models as well for similar

301 reasons. A second criticism is that, if the between-country effects predominate and are more accurate for estimation, why utilize a panel data setup (Wilson and Butler, 2007)?

It is true that much of this could be estimated using global averages for the variables.

However, I hesitate to give up the power of panel data, not to mention the ability to include some variables, like the post-Cold War variable, that are not constant over time.

Indeed, panel data estimation does not give any more of an advantage in hypothesis testing than common methods of weighting based on the number of observations that make up a time averaged variable, and it provides greater flexibility for the estimation of other variables.

Speaking more generally, these results should give warning to those who assume that either fixed effects (within) estimation will automatically yield unbiased results. The results demonstrate that substantive variation is something to be explored and explained, not disregarded.

302

APPENDIX C

FORMALIZING GOVERNMENT/OPPOSITION RELATIONS IN DISTRIBUTION-ORIENTED SOCIETIES

303 C.1 The setup

This game is played between a government decision-maker, G, and a latent opposition, O, in society. The government has a certain amount of revenue R, which it can distribute to the opposition 0≤DG ≤ R. Whatever the government does not spend on distribution, it can keep for itself. The government also gains a certain inherent value of

Φ for being in power and being able to implement its policies.

After the distribution has been made, the opposition decides whether they are going to challenge the government, C, or not challenge the government, NC. If they decide to challenge, their probability of winning is based on a draw from nature, N, whereby the challenge has a probability of winning of p and a probability of losing of 1- p.

If the opposition chooses not to challenge, they assumes that they will continue receiving the utility from the current government for the foreseeable future. This utility includes government distribution, DG, plus income garnered from economic activities independent of the government, FG. In addition, the opposition has some level of ideological affinity towards the current government, aG. This last variable can be viewed simply as the amount of current income the opposition is willing to forgo in order to bring the government into line with their ideal point ideologically. This is loosely consistent with Kuran's (1991) contention that psychological costs shape tolerance for a government. It also explains why some groups are willing to forgo all current utility to protest against a government, while others only act against a government when they are

304 sure that their actions will succeed.

While the government and opposition are making decisions at the current time, it is assumed that they expect the payoffs from that decision to continue indefinitely. In some ways this is similar to Boix's (2003) assumption that individuals choose regimes once during their lifetime. Unlike Boix, however, individuals in this model expect repeated distributions from the economy. This allows the individual to take actions that may result in greater average utility over time, rather than focusing on current utility.

Since people's generally value current consumption more than future consumption, an individual's utility at each time point is multiplied by a deflation factor of σє[0,1). Thus, the expected utility stream for i for the status quo (SQ) is EU(O:NC) = [u(SQ), σu(SQ), ...

, σt(SQ), ... ].

This simplifies to:

DG + FG − aG EU(O : NC) = [1] 1−σ

If the opposition challenges, and is not successful, L, they receive no income for the current time, and only those payments that are independent of the government, fiG, and their ideological affinity for the government, aiG, for the foreseeable future. If the opposition challenges, and is successful, W, they receive no utility for the present time, but have the ability to re-establish the government, either as a democracy, where the opposition competes with the government for political power, or as an authoritarian government with the current opposition as the new government. In both cases, the opposition receives zero utility for the current period, followed by one of three outcomes.

305 If the opposition challenges and loses, their utility is:

σ (FG − aG ) EU(O : C,L) = 0 + [2] 1− σ

If they challenge, win, and a democracy is established, the future government is decided by elections (E), which the opposition wins with probability q and loses with probability 1-q. If the opposition wins, they receive a government distribution, Do, in addition to their distribution that is independent of the government. If the current government wins elections, then they do not receive a private distribution from the government, but continues to receive the independent distribution from the performance of the economy. For simplicity, the opposition expects an overall performance for the economy under elections, resulting in a consistent payoff, FE. This not only simplifies the equations, but is consistent for the comparisons made in choosing a regime-type as opposed to a policy formulation. When the choice is one of regime, the view of performance will usually be of that regime, not just of performance under a particular party under that regime, since the inherent risk of losing office is already taken into account (Przeworski, 1991). This is also consistent with expectations of aid from outside sources and greater accountability that are expected of democracy.

Thus, the total expected payoff from a transition to a democratic regime is:

σ[q(DO − aO ) − (1− q)(aG ) + FE ] EU(O : C,W ,E) = 0 + [3] 1− σ

Finally, if the opposition challenges, wins, and establishes an unelected regime

(NE), they receives zero utility for the present time, followed by a continuous stream of private distributions, Do, a distribution from the economy under the opposition's dictatorship, FNE, and their affinity for policy outcomes. This case also works well for

306 the situation where the opposition challenges in support of a separatist region in the country, as the current government will effectively be excluded from competing in the newly independent area. The total expected payoff for this situation is:

σ (DO + FO − aO ) EU(O : C,W , NE) = 0 + [4] 1−σ

The overall game was laid out above in figure 3.4. This intuitive setup can be solved by simple backwards induction, and the results will be analyzed in the following sections.

C.2 Democratic or authoritarian challenge

The analysis of this game begins at the end, asking when the opposition, if their challenge is successful, will choose a democratic or authoritarian government. The opposition will choose to support democratization when the value of equation [3] is greater than the value of equation [4].107 Solving for q, the probability of the opposition winning elections, yields the following solution:

DO + (FO − FE ) + (aG − aO ) q > [5] DO + (aG − aO )

Note that as the importance of government distribution relative to distribution from the economy increases, the acceptability of democracy decreases. As the relative size of distribution, DO, increases relative to economic distribution, FNE and FE, a change in economic importance decreases in importance. Put another way, the more important government distribution is compared to economic distribution for the opposition, the less likely they are to accept uncertainty in that distribution, even if it results in greater overall

107It is assumed that, in the case of a tie, the opposition prefers less uncertainty, and therefore the authoritarian solution. 307 economic performance.

C.3 The choice to challenge or not challenge

Now the choice of the opposition of whether to challenge, C, or not challenge,

NC, can be explored. Here, opposition must also calculate in their probability of winning

(p) or losing (1-p) if they challenge.

When this probability of winning or losing is included in an opposition's calculation, the total expected utility for challenging (C) in favor of a democratic regime is the following combination of equations [2] and [3]:

 σ [q(DO − aO ) − (1− q)(aG ) + FE ]  σ (FG − aG ) EU(O : C,E) = p0 +  + (1− p)0 +  [6]  1−σ   1− σ 

The total expected utility to challenge in favor of a non-democratic regime is a combination of equations [2] and [4]:

 σ (DO + FO − aO )  σ (FG − aG ) EU(O : C, NE) = p0 +  + (1− p)0 +  [7]  1− σ   1−σ 

These two outcomes must now be weighed against the expected utility of maintaining the status quo, which is given in equation [1]. Solving for values of p yields the following inequalities. When choosing between the status quo and challenging for a democratic regime, the opposition will not challenge if:

(DG + FG − aG ) −σ (FG − aG ) p < [8] σ[q(DO − aO ) − (1− q)(aG ) + FE − (FG − aG )]

Put into words, this equation suggests that the status quo will be preferred if the probability of winning is less than the value of the status quo minus the results of losing, divided by the expected value of successful democratization minus the utility under the

308 government if a challenge is lost. The intuitive interpretation of this is that the opposition will not challenge if the probability is less than the utility risked by challenging, divided by the expected gain of successfully challenging.

Similarly, comparing the expected utility from challenging for a non-democratic regime versus staying with the status quo results in the following:

(DG + FG − aG ) −σ (FG − aG ) p < [9] σ[(DO + FO − aO ) − (FG − aG )]

The intuitive interpretation of equation [9] is the same as in equation [8], with the opposition weighing the utility risked by a challenge against the gains from an authoritarian regime.

All other things being equal, increases in distributions controlled by the government, DG, have a greater stabilizing effect than economic growth not controlled by the government. This is because the government is able to take away those gains if the opposition challenges and loses. Knowing this, the opposition is risking more when the government has a high level of distribution to withhold. If both DG and FG come out of the same pot (ex. if we are discussing proportion of ownership in industry), than the higher the proportion of income that comes from distribution, the riskier a challenge becomes.

C.4 Preventing a challenge

At this point, the government must make some decision about distribution to the individuals in society within its budget constraint. To prevent the opposition from challenging for an authoritarian regime, the government must distribute to O:

309 * p[σ (DO + FO − aO )] + (1− p)[σ (FG − aG )] DG ≥ [10] FG − aG

And to prevent a challenge for a democratic regime, the government distribution must be:

* p[σ (q(DO − aO ) − (1− q)(aG ) + FE )] + (1 − p)[σ (FG − aG )] DG ≥ [11] FG − aG

C.5 Equilibrium

Given the above equations, the government must make a decision over how much to spend on distribution, within its revenue constraint, R. The government will choose to

* make the necessary distribution to the opposition DG if two conditions are met:

* (1) DG ≤ R and

Φ + R − D* (2) G ≥ R 1−σ

The first inequality simply states that the government cannot overspend its budget constraint. The second inequality suggests that the government faces a choice in whether or not to make the necessary distribution to the opposition. If the government makes the distribution, it receives the value of being in leadership and the remainder of revenues after the distribution has been made, and this is repeated into the indefinite future. If the government does not form a coalition, however, it is able to keep all the revenues at time t and leave. Thus, the value gained from forming a winning coalition must be higher than the value of raiding the state coffers.

310 The equilibrium in this case is:

1. Where both of the government's above conditions are met, the government forms a winning coalition and the status quo remains in place.

2. Where condition (1) does not hold, the government chooses not to form a winning coalition, raids the state coffers, and the opposition either forms a democratic or opposition-dominated regime based on equation [5].

3. Where condition (2) does not hold, the government chooses not to form a winning coalition, raids the state coffers, and the opposition forms either a democratic or opposition-dominated regime based on equation [5].

311 BIBLIOGRAPHY

Acemoglu, D., & Robinson, J.A. (2005). Economic origins of dictatorship and democracy. Cambridge, UK: Cambridge University Press.

Acemoglu, D., Johnson, S., Robinson, J.A., & Yared, P. (2007). Reevaluating the modernization hypothesis. NBER Working Paper No. 13334.

Allan, J.A. (1981). Libya: The experience of oil. Boulder, CO: Westview Press.

Almond, G.A., & Verba, S. (1963). The civic culture. Boston: Little Brown.

Amuzegar, J. (2001). Managing oil wealth: OPEC’s windfalls and pitfalls. London: I.B. Taurus.

Angeli, F. (1994, March 4). Results of elections in Moldova. ITAR-TASS. Retrieved from Lexis-Nexis database.

Antelava, N. (2006, December 9). Giant tent to be built in Astana. BBC News. Last retrieved from BBC news archives January 16, 2008, from http://news.bbc.co.uk/2/hi/asia-pacific/6165267.stm

Aristotle. (1944). The Politics. In H. Rackman (Trans.), Aristotle in 23 Volumes, Vol. 21. Cambridge, MA: Harvard University Press.

Ascher, W. (1999). Why governments waste natural resources. Baltimore, MD: Johns Hopkins University Press.

Associated Press. (1995, May 1). Kazakh President hails vote extending his run. The New York Times, p. A5

Auty, R., & Gelb, A. (1986). Oil windfalls in a small parliamentary democracy: Their impact on Trinidad and Tobago. World Development, 14,1161-1175.

Bagirov, S., Akhmedov, I., & Tsalik, S. (2002). State oil fund of the Azerbaijan Republic. In S.Tsalik, Caspian oil windfalls: Who will benefit? (pp. 89-125). New York: Open Societies Institute.

Bannon, I., & Collier, P. (Eds.). (2003). Natural resources and violent conflicts: actions and options. Washington, DC: World Bank.

312 Barr, A., & Serra, D. (2006). Culture and corruption. Global Poverty Research Group Working Paper, GPRC-WPS-040.

Barro, R.J. (1999). Determinants of democracy. The Journal of Political Economy, 107,158-183.

Bauer, P.T. (1984). Reality and rhetoric: Studies in the economics of development. London: Weidenfeld and Nicolson.

Bauer, P.T. (2000). From subsistence to exchange and other essays. Princeton: Princeton University Press.

Beblawi, H. (1987). The rentier state in the Arab world. In H. Beblawi & G. Luciana (Eds.). The rentier state (pp. 49-62). New York: Croom Helm.

Beck, N. (2001). Time-series-cross-section data: What have we learned in the past few years? Annual Review of Political Science, 4, 271-293.

Beck, N., & Katz, J.N. (1995). What to do (and not to do) with time-series-cross- sectional data in comparative politics. American Political Science Review, 89, 634-647.

Beck, N. & Katz, J. (1996). Nuisance vs. substance: Specifying and estimating time series cross section models. Political Analysis, 6, 1-34.

Bell, D. (1973). The coming of post-industrial society: A venture in social forecasting. NY: Basic Books.

Bellin, E.R. (2000). Contingent democrats: Industrialists, labor, and democratization in late-developing countries. World Politics, 52, 175-205.

Bendix, R. (1978). Kings or people: Power and the mandate to rule. Berkeley, CA: University of California Press.

Benjamin, N.C., Devarajan, S., & Weiner, R.J. (1989). 'The Dutch Disease' in a developing country: Oil reserves in Cameroon. Journal of Development Economics, 30, 71-92

Bennett, D., & Sharpe, K. (1980). The state as banker and entrepreneur: The last-resort character of the Mexican state's economic intervention, 1917-76. Comparative Politics, 12, 165-189.

Bergstrand, J.H. (1985). The gravity equation in international trade: Some microeconomic foundations and empirical evidence. The Review of Economics and Statistics, 67, 474-481.

Bermeo, N. (1992). Democracy and the lessons of dictatorship. Comparative Politics, 24, 273-291. 313

Bienen, H., & van de Walle, N. (1991). Of time and power. Stanford, CA: Stanford University Press. le Billon, P. (2001). Angola’s political economy of war: The role of oil and diamonds, 1975-2000. African Affairs, 100, 55-80.

Birdsall, N., & Subramanian, A. (2004). Saving Iraq from its oil. Foreign Affairs, 83, 77-89.

Black, J.K., Blutstein, H.I., Johnston, K.T., & McMorris, D.S. (1976). Area handbook for Trinidad and Tobago, 1st ed. Washington, DC: US Government Printing Office.

Boix, Carles. (2003). Democracy and Redistribution. Cambridge, UK: Cambridge University Press.

Boix, C., & Stokes, S. (2003). Endogenous democratization. World Politics, 53, 517- 549.

Botsan, I. (2001). Republic of Moldova between transformation and status quo: Analysis of ten years of independence. In proceedings of conference ‘Republic of Moldova in the European Context.’

Bowman, B.L. (2005). Realism and idealism: US policy toward Saudi Arabia from the Cold War to today. Parameters, 35, 91-105.

Box-Steffensmeier, J.M., & Jones, B.S. (1997). Time is of the essence: Event history models in political science. American Journal of Political Science, 41, 1414-1461.

Box Steffensmeier, J.M., & Jones, B.S. (2004). Timing and political change: Event history modeling in political science. Ann Arbor, MI: University of Michigan Press.

Box-Steffensmeier, J.M., & Zorn, C. (2001). Duration models and proportional hazards in political science. American Journal of Political Science, 45, 972-988.

Box-Steffensmeier, J.M., Reiter, D., & Zorn, C. (2003). Nonproportional hazards and event history analysis in international relations. Journal of Conflict Resolution, 47, 33- 53.

Bratton, M., & van de Walle, N. (1997). Democratic experiments in Africa. Cambridge, UK: Cambridge University Press.

Bremmer, I. (2004). The Saudi paradox. World Policy Journal, 21, 23-30.

Bremmer, I. (2005, December 27). Not all nonrevolutions are created equal: Why unremarkable elections still matter. Slate. Retrieved December 27, 2005 from http://www.washingtonspeakers.com/prod_images/pdfs/BremmerIan.NotAllNonrevolutio 314 nsAreCreatedEqual.12.27.05.pdf

Buckley, N., & Ostrovsky, A. (2005, December 6). Former Soviet Republics Use Dirty Tricks to Sway Polls. Financial Times. Retrieved from Financial Times online archive on December 7, 2005, from http://news.ft.com/cms/s/b6d64434-65fe-11da-8f40- 0000779e2340.html.

Bueno de Mesquita, B., Morrow, J.D., Siverson, R.M., & Smith, A. (2002). Political institutions, policy choice and the survival of leaders. British Journal of Political Science 32, 559-590.

Bueno de Mesquita, B., Morrow, J.D., Siverson, R.M., & Smith, A. (2003). The logic of political survival. Cambridge, MA: MIT Press.

Bueno de Mesquita, B. & Root, H.L.. (2002). The political roots of poverty. National Interest, 68, 27-38.

Burke, J. (1991, December 5). Once a Soviet flashpoint, fragile ethnic peace comes to Kirghizia. Christian Science Monitor, 4.

Cananche, D. (2002). From bullets to ballots: The emergence of popular support for Hugo Chavez. Latin American Politics and Society, 44, 69-90.

Cardoso, F. (1972). Dependency and development in Latin America. New Left Review, 74, 83-95.

Saskatchewan coming into its own as oil power: U.S. expert. (2007, 13 December). CBC News. Retrieved from CBC News online archive on December 17, 2007, from http://www.cbc.ca/canada/saskatchewan/story/2007/12/13/oil-saskatchewan.html

What’s going on in Kazakhstan? (2007, October 7). Central Asia Review. Retrieved February 25, 2008, from http://centralasiareview-blog.blogspot.com/2007/10/eni-led- kashagan-project-under-threat.html.

Central Intelligence Agency. (2007). World Factbook. Washington, DC: CIA.

Chalmers, D.A. (1969) Development on the periphery. In J.N. Rosenau (Ed.). Linkage politics: Essays on the convergence of national and international systems (pp. 67-93). New York: Free Press.

Chaudhry, K.A. (1997). The price of wealth: Economics and institutions in the Middle East. Ithaca, NY: Cornell University Press.

Chaudhry, K.A. (1989). The price of oil wealth – Business and state in labor remittance and oil economies. International Organization, 43, 101-145.

Chazan, G. (2007, September 26). Kazakh oil gem is ripe for takeover. Wall Street 315 Journal. Retrieved from Wall Street Journal online archive on February 25, 2008, from http://online.wsj.com/article/SB119074819583138965.html?mod=googlenews_wsj.

CIA. (1979). Nigeria: Ethnolinguistic groups from map no. 504014 1979. Retrieved from the Perry-Casteñada Map Collection, University of Texas Online Library on March 13, 2007, from http://www.lib.utexas.edu/maps/nigeria.html.

CIA. (1992). Iraq: Ethnoreligious groups from a map folio. Retrieved from the Perry- Casteñada Map Collection, University of Texas Online Library on March 13, 2007, from http://www.lib.utexas.edu/maps/iraq.html.

Cioranica, R. (2007, March). “Interview with Oleg Voronin.” VIP Magazine. Retrieved from VIP Magazine online archive on March 3, 2008, from http://www.vipmagazin.md/index.php?mod=article&id=82

Clark, J. (1997). Petro-politics in Congo. Journal of Democracy, 8, 62-76.

Collier, P., & Hoeffler, A. (1998). On the economic causes of civil war. Oxford Economic Articles, 50, 563-573.

Collier, P., & Gunning, J.W. (1999). Why has Africa grown slowly? The Journal of Economic Perspectives, 13, 3-22.

Collins, R. (1979). The credential society: A historical sociology of education and stratification. NY: Academic Press.

Coronil, F. (1997). The magical state: Nature, money, and modernity in Venezuela. Chicago: University of Chicago Press.

Cottam, R. (1979). Goodbye to America's Shah. Foreign Policy, 34, 3-14.

Cranston, A., & Green, M.D. (1994, 4 May). “Kyrgyzstan Takes Quiet Path to Democracy.” Christian Science Monitor, pp. 14.

Crooks, E. (2007, December 4). Kashagan becomes a thorn in Eni's side. Financial Times, pp. 20.

Crystal, J. (1989). Coalitions in oil monarchies: Kuwait and Qatar. Comparative Politics, 21(4), 427-443.

Crystal, J. (1990). Oil and politics in the Gulf: Rulers and merchants in Kuwait and Qatar. Cambridge, UK: Cambridge University Press.

Cutler, R.M. (2004, October 6). Kazakhstan holds elections for new parliament. Central Asia-Caucasus Analyst. Retrieved from Central Asia-Caucasus Analyst online archive on February 18, 2007, from http://www.cacianalyst.org/view_article.php?articleid=2744

316 Dahl, R.A. (1961). Who governs? Democracy and power in the American city. New Haven, CT: Yale University Press.

Dave, B. (2005). Kazakhstan's 2004 parliamentary elections: Managing loyalty and support for the regime. Problems in Post-Communism, 52, 6.

David, P.A., & Wright, G. (1997). The genesis of American resource abundance. Industrial and Corporate Change, 6, 203-245.

Davis, G.A., & Tilton, J.E. (2002). Should developing countries renounce mining? A perspective on the debate. International Council on Mining and Minerals Working Paper.

Davis, J.M., Ossowski, R., Daniel, J., & Barnett, S. (2001). Stabilization and savings funds for nonrenewable resources. IMF Occasional Working Paper, no. 205.

Deffeyes, K.S. (2001). Hubbert’s peak: The impending world oil shortage. Princeton, NJ: Princeton University Press.

Kyrgyz second round elections held, President cites irregularities. (1995, February 19). Deutsch Presse-Agentur. Retrieved from Lexis-Nexis database.

Deutsch, K.W. (1991). Social mobilization and political development. The American Political Science Review, 55, 493-514.

Diamond, L. (1992). Economic development and democracy reconsidered. In G. Marks & L. Diamond (Eds.), Reexamining democracy: Essays in honor of Seymour Martin Lipset (pp. 93-139). Newbury Park, CA: Sage Publications.

Djerejian, E.P., Wisner, F.G., Bronson, R., & Weiss, A.S. (2003). Guiding principles for U.S. post-conflict Iraq. Retrieved from Iraq Foundation online archive on December 15, 2007, from, http://www.iraqfoundation.org/studies/2003/afeb/iraq.pdf

Duncan, R.C. (2003). Three world oil forecasts predict peak oil. Oil & Gas Journal, 101, 18-21.

Dunning, T. (2007). Crude democracy: Natural resource wealth and political regimes. Manuscript: Yale University.

Easterly, W.R. (2001). The elusive quest for growth. Cambridge, MA: MIT Press.

Central Asia: The shadow of dictatorship. (1995, March 18). The Economist, pp. 40.

The people’s oil. (2003, April 12). The Economist. Retrieved from Lexis-Nexis database.

BP: Some you win, some you lose. (2007, June 2). The Economist, pp. 66-67.

317 Ekspress-K. (2005, March 6). Nazarbayev ready for presidential elections – advisor. BBC Monitoring Central Asia Unit. Retrieved from Lexis-Nexis database.

Elliot, L. (2007, May 2). Venezuela seizes foreign oil firms. The Guardian (London), pp. 25.

Embassy of the Republic of Kazakhstan. (2002, December 17). Astana celebrates 5th anniversary as capital. Kazakhstan News Bulletin, 4(12). Retrieved from the Embassy of the Republic of Kazakhstan’s website on January 16, 2008, from http://www.kazakhembus.com/121702.html.

Embassy of the Republic of Kazakhstan. (2003, January 8). Disneyland and Las Vegas come together in Astana. Kazakhstan News Bulletin, 1(1). Retrieved from the Embassy of the Republic of Kazakhstan’s website on January 16, 2008, from http://www.kazakhembus.com/010803.html.

Embassy of the Republic of Kazakhstan. (2005, November 9). Major news media cover Nazarbayev, Tuyakbai, Baimenov. Kazakhstan News Bulletin, 5(50). Retrieved from the Embassy of the Republic of Kazakhstan’s website on January 20, 2006, from http://www.kazakhembus.com/112905.html.

Epstein, E.C. (1984). Legitimacy, institutionalization and opposition in exclusionary bureaucratic-authoritarian regimes: The situation in the 1980s. Comparative Politics, 17, 37-54.

Epstein, D., Bates, R., Goldstone, J., Kristensen, I., & O’Halloran, S. (2006). Democratic transitions. American Journal of Political Science, 50, 551-569.

Akayev administration collapses in Kyrgyzstan, sending tremors across Central Asia. (2005, March 24). Eurasia Insight. Retrieved from Eurasianet’s online archive on February 10, 2008, from http://www.eurasianet.org/departments/civilsociety/articles/eav032405.shtml

Kyrgyzstan: Revolution revisited, report card. (2008). Eurasianet. Retrieved January 16, 2008, from http://www.eurasianet.org/kyrgyzstan/reportcard/index.shtml.

Evans, P.B. (1979). Dependent development: The alliance of multinational, state, and local capital in Brazil. Princeton, NJ: Princeton University Press.

Evans, P.B., Rueschemeyer, D., & Skocpol, T. (Eds.). (1985). Bringing the state back in. Cambridge, UK: Cambridge University Press.

Evans, T.B. (1999, December 20). Dining with dictators: White House fetes Kazakh President. The Washington Times. Retrieved from Eurasianet’s press archive on February 2, 2007, from http://eurasia.org.ru/archive//1999/press_en/12_20_12_20_1999.htm

318 Farbey, B.A., Mitchell, C.R., & Webb, K.. (1980). Change and stability in the ideology of Scottish nationalism. International Political Science Review, 1, 405-424.

Fardmanesh, M. (1991). Dutch Disease economics and the oil syndrome: An empirical study. World Development, 19, 711-717

Fryer, D.W. (1954). Indonesia's economic prospects. Far Eastern Survey, 23, 177-182.

Fearon, J.D. (2005). Primary commodity exports and civil war. Journal of Conflict Resoution, 49, 483-507.

Feenstra, R.C., Markusen, J.R., & Rose, A.K. (2001). Using the gravity theory to differentiate among alternative theories of trade. The Canadian Journal of Economics, 34, 430-447.

Forero, J. (2006, June 1). As prices soar, the Orinoco belt becomes and economic battleground. The New York Times, pp. C1.

Frank, A.G. (1967). Capitalism and underdevelopment in Latin America: Historical studies of Chile and Brazil. New York: Monthly Review Press.

Freedom House. (2005). Country report, Kyrgyzstan. Retrieved from the Freedom House organization webpage on February 29, 2008, from http://www.freedomhouse.org/template.cfm?page=47&nit=370&year=2005

Freedom House. (2006). Country report, Kyrgyzstan. Retrieved from the Freedom House organization webpage on February 29, 2008, from http://www.freedomhouse.org/template.cfm?page=47&nit=385&year=2006.

Freedom House. (2006). Country report, Moldova. Retrieved from the Freedom House organization webpage on March 2, 2008, from http://www.freedomhouse.org/template.cfm?page=47&nit=396&year=2006

Frontline. (2001). Saudi Time Bomb? Interview: Bandar Bin Sultan. Retrieved from Frontline’s online archive on March 23, 2005, from http://www.pbs.org/wgbh/pages/frontline/shows/terrorism/interviews/bandar.html.

Fryer, D.W. (1954). Indonesia's economic prospects. Far Eastern Survey, 23, 177-182.

Frynas, J.G., & Paulo, M. (2007). A new scramble for African oil? Historical, political and business perspectives. African Affairs, 106, 229-251.

Fukuyama, F. (1989). The end of history? The National Interest. Retrieved online on March 8, 2007, from www.unc.edu/~rlstev/Text/Fukuyama%20End%20of%20History.pdf

Furman, D.E. (2007). Amazing Moldova. Russia in Global Politics, 5. Retrieved from 319 Russia in Global Politics’ online edition on March 3, 2008, from http://www.globalaffairs.ru/numbers/28/8551.html.

Gall, C. (1994, October 12). Kazakh regime resigns. The Moscow Times. Retrieved from Lexis-Nexis database.

Geddes, B., & Zaller, J. (1989). Sources of popular support for authoritarian regimes. American Political Science Review, 33, 319-347.

Gelb, A. (1988). Oil Windfalls: Blessing or Curse? New York: Oxford University Press.

Gerschenkron, A. (1966). Economic backwardness in historical perspective. Cambridge, MA: Harvard University Press.

Global Witness. (2002). All The President's Men. London: Global Witness. Retrieved August 13, 2005, from http://www.globalwitness.org/media_library_detail.php/85/en/all_the_presidents_men

Global Witness. (2004). Time for transparency: Coming clean on oil and gas revenues. London: Global Witness. Retrieved August 13, 2005, from http://www.globalwitness.org/media_library_detail.php/115/en/time_for_transparency.

Gylfason, T. (2000). Natural resources, education and economic development. Presented at the Annual Congress of the European Economic Association, Bolzano, Italy.

Gylfason, T., & Zoega, G. (2001). Natural resources and economic growth: The role of investment. CEPR Discussion Paper no. 2743.

Gylfason, T., & Zoega, G. (2002). Inequality and economic growth: Do natural resources matter? CESifo Working Paper no. 712.

Hanson, S.E. (2007). The uncertain future of Russia’s weak state authoritarianism. East European Politics and Societies, 21, 67-81.

Hawkins, E.D. (1955). Prospects for economic development in Indonesia. World Politics, 8, 91-111.

Hellinger, D. (2000). Understanding Venezuela's crisis: Dutch Disease, money doctors and magicians. Latin American Perspectives, 27, 105-119.

Henry, C.M. (2004). Algeria's agonies: Oil rent effects in a bunker state. Journal of North African Studies, 9, 68-81.

Hersh, S. (2001, July 9). The price of oil. The New Yorker, pp. 48-50.

Hillman, R.S. (1994). Democracy for the privileged: Crisis and transition in Venezuela. 320 Boulder: Lynne Rienner.

Horn, M.K. (2003). Giant oil fields, 1868-2003 (CD-ROM). In M.K. Halbouty (Ed.). Giant Oil and Gas Fields of the Decade, 1990-1999. AAPG Memoir 78.

Horowitz, D.L. (1985). Ethnic groups in conflict. Berkeley, CA: University of California Press.

Hughes, C. (1993, October 17). Yeltsin dismantles parliament’s empire. The Associated Press. Retrieved from Lexis-Nexis database.

Human Rights Watch. (2003). Considerations for the management of oil in Iraq: A Human Rights Watch background briefing. Retrieved from Human Rights Watch online February 5, 2007, from http://www.cfr.org/publication/10532/chads_oil_troubles.html

Huntington, S.P. (1991). The Third Wave: Democratization in the Late Twentieth Century. Norman, OK: University of Oklahoma Press.

Huntington, S.P. (1996). The clash of civilizations and the remaking of world order. New York: Simon and Schuster.

Kazakhs see no alternative to Nazarbayev. (2005, November 9). Infotag. Retrieved from Lexis-Nexis database.

Bakiyev unconcerned by growing number of parties. (2006, August 31). Interfax. Retrieved from Lexis-Nexis database.

Interfax. (1995, November 28). President Akayev remains only registered candidate. BBC Monitoring Central Asia Unit. Retrieved from Lexis-Nexis database.

Restrictions on companies linked to Akayev to be lifted. (2005, June 27). Interfax. Retrieved from Lexis-Nexis database.

Venezuela reaches deals on oil settlements. (2007, May 6). International Herald Tribune, pp. 12.

Institute for War and Peace Reporting. (2007, September 13). Kazakhstan/Russia: Channel 31 takeover to ‘obstruct’ press freedom. BBC Monitoring World Media. Retrieved from Lexis-Nexis database.

Integrated Regional Informational Network (IRIN). (2006). Kazakhstan: Year in Review 2005 -- Energy Brings Growth and Stability. Retrieved from IRIN online archives August 14, 2006, from http://www.irinnews.org/report.aspx?reportid=33638

Irwin, D.A. (2003). Explaining America's surge in manufactured exports, 1880-1913. The Review of Economics and Statistics, 85, 364-376.

321 Isbister, J. (1993). Promises not kept: The betrayal of social change in the Third World, 2nd ed. West Hartford, CT: Kumarian Press.

Isham, J., Woolcock, M., Pritchett, L.,& Busby, G. (2003). The varieties of resource experience: How natural resource export structures affect the political economy of economic growth. Middlebury College Discussion Paper No. 2003-08.

Jensen, N., & Wantchekon, L.. (2004). Resource wealth and political regimes in Africa. Comparative Political Studies, 37, 816-841.

Jones, B.S., & Branton, R.P. (2003). Beyond logit and probit: Some alternative duration modeling strategies for state policy adoption. Working Paper: University of Arizona.

Joseph, R.A. (1987). Democracy and prebendal politics in Nigeria: The rise and fall of the Second Republic. Cambridge, UK: Cambridge University Press.

Joseph, R.A. (1996). Nigeria: Inside the dismal tunnel. Current History, 96, 193-200.

Jureidini, R. (2003). Migrant workers and xenophobia in the Middle East. United Nations Research Institute for Social Development Working Paper PP-ICC-2.

Karl, T.L. (1986). Petroleum and political pacts: The transition to democracy in Venezuela. In G. O'Donnell & P. Schmitter (Eds.). Transitions from authoritarian rule: Latin America (pp. 196-236). Baltimore: Johns Hopkins University Press.

Karl, T.L. (1997). The paradox of plenty: Oil booms and petro-states. Berkeley: University of California Press.

Karl, T.L. (2005). Understanding the resource curse. In Svetlana Tsalik & Anya Schiffrin (Eds.), Covering oil: A reporter’s guide to energy and development (pp. 21- 30). New York: Open Society Institute.

Kaufmann, D., Kraay, A., & Mastruzzi, M. (2007). Growth and governance: A reply. The Journal of Politics, 69, 555-562.

Kennedy, R. (2004). Re-conceptualizing the social requisites of democracy: A conditional probability interpretation of modernization. Presented at the Annual Meeting of the Midwest Political Science Association, Chicago, IL.

Kennedy, R. (2008). Supranational identity, political attitudes, and political participation: The case of Moldova. Presented at the Annual Meeting of the Midwest Political Science Association, Chicago, IL.

Khan, S.A. (1994). Nigeria: The political economy of oil. New York: Oxford University Press.

Khennas, S. (Ed.). (1993). Industrialization, mineral resources and energy in Africa. 322 Oxford: Africa Books Collective.

Kimmage, D. (2005, December 6). Kazakhstan: Nazarbayev landslide buries future problems. Radio Free Europe/Radio Liberty. Retrieved December 7, 2005, from http://www.rferl.org/featuresarticle/2005/12/180c3c44-f5ed-4b16-aca4- 0296e27b0f58.html

King, C. (2000). The Moldovans: Romania, Russia and the Politics of Culture. Stanford, CA: Hoover Institute.

Knack, S. (2004). Does foreign aid promote democracy? International Studies Quarterly, 48, 251-266.

Knight, J. (1992). Institutions and social conflict. Cambridge, UK: Cambridge University Press.

Kopstein, J., & Reilly, D.A. (2000). Geographic diffusion and the transformation of the postcommunist world. World Politics, 53, 1-37.

Kramer, A.E. (2005, December 23). Kazakh leader's family widens clout. International Herold Tribune. Retrieved from International Herald Tribune online archive January 5, 2005, from http://www.iht.com/articles/2005/12/23/business/kazakh.php

Kramer, A.E. (2006, December 22). Russians buy control of oil field. New York Times. Retrieved from New York Times online archive December 14, 2007, from http://www.nytimes.com/2006/12/22/business/worldbusiness/22shell.html?_r=1&oref=sl ogin

Krasner, S.D. (2005). The case for shared sovereignty. Journal of Democracy, 16, 69- 83.

Kretzmann, S., & Nooruddin, I. (2006). Drilling into debt: An investigation into the relationship between debt and oil. Washington, DC: Oil Change.

Kupchinsky, R. (2007, June 13). Russia: BP-TNK, Gazprom, and the Kovytka Gas Field. Radio Free Europe/Radio Liberty. Retrieved February 8, 2008, from http://www.rferl.org/featuresarticle/2007/06/0397b691-67a9-4c20-9833- 9b4b26396cca.html.

Kuran, T. (1991). Now out of never: The element of surprise in the East European revolution of 1989. World Politics, 44, 7-48.

Kurtz, M.J., & Schrank, A. (2007). Growth and governance: Models, measures, and mechanisms. The Journal of Politics, 69, 538-554.

Kurtz, M.J., & Schrank, A. (2007b). Growth and governance: A defense. The Journal of Politics, 69, 563-569. 323

Lake, D., & Baum, M.A. (2001). The invisible hand of democracy: Political control and the provision of public services.” Comparative Political Studies, 34, 587-621.

Lam, R., & Wantchekon, L. (2002). Political Dutch Disease. Working Paper: Yale University.

Lee, G.W. (1976). North sea oil and Scottish nationalism. The Political Quarterly, 47, 307-317.

Lerner, D. (1958). The passing of traditional society. NY: Free Press.

Levi, M. (1989). Of rule and revenue. Berkeley: University of California Press.

Levine, D. (1973). Conflict and political change in Venezuela. Princeton, NJ: Princeton University Press.

Levine, S. (1995, February 21). Kyrgyzstan poll set to boost reform: If he wins Akayev will be pressed for more change.” Financial Times, pp. 4.

Levitsky, S., & Way, L. (2002). The rise of competitive authoritarianism. Journal of Democracy, 13, 51-65.

Liddle, R.W. (1991). The relative autonomy of the third world politician. International Studies Quarterly, 35, 403-427.

Lillis, J. (2007, July 9). Kazakhstan’s capital holds lavish celebration. Eurasia Insight. Retrieved from Eurasianet’s online archive February 6, 2008, from http://www.eurasianet.org/departments/insight/articles/eav070907.shtml.

Lindblom, C.E. (1982). The market as prison. Journal of Politics, 44, 324-336.

Linz, J.J., & Stepan, A. (1978). The breakdown of democratic regimes. Baltimore, MD: Johns Hopkins University Press.

Linz, J., & Stepan, A. (1996). Problems of democratic transition and consolidation: Southern Europe, South America and post-Communist Europe. Baltimore: Johns Hopkins University Press.

Lipset, S.M. (1959). Some social requisites of democracy: Economic development and political legitimacy. American Political Science Review, 53, 69-105.

Lipset, S.M. (1960). Political man: The social bases of politics. Garden City, NY: Doubleday.

Lipset, S.M. (1994). The Social Requisites Revisited. American Sociological Review, 59, 1-22. 324

Lowi, T.J. (1964). American business, public policy, case-studies, and political theory. World Politics, 16, 677-715.

Luong, P.J., & Weinthal, E. (2006). Rethinking the resource curse: Ownership structure, institutional capacity, and domestic constraints. Annual Review of Political Science, 9, 241-263.

Lyubinaite, M. (2005, November 25). Sluggish pre-election race. Kazakhstan Monitor, pp. 3.

Mahdavy, H. (1970). The patterns and problems of economic development in rentier states: The case of Iran. In M.A. Cook (Ed.), Studies in Economic History of the Middle East (pp. 428-467). London, UK: Oxford University Press.

March, J.G., & Olsen, J.P. (1985). The new institutionalism: Organizational factors in political life. The American Political Science Review, 78, 734-749.

March, L. (2006). Power and opposition in the former Soviet Union: The Communist Parties of Moldova and Russia. Party Politics, 12, 341-365.

Marin, C. (1999, February). Battle for oil in Nigeria. Le Monde Diplomatique. Retrieved June 12, 2007, from http://mondediplo.com/maps/africanigeriamdv51.

Markoff, J., & Baretta, S.R.D. (1990). Economic crisis and regime change in Brazil: The 1960s and 1980s. Comparative Politics 22(4), 421-444.

Markus, U. (2005, November 22). Kazakhstan's uneventful election campaign. ISN Security Watch. Retrieved January 20, 2006, from http://www.isn.ethz.ch/news/sw/details.cfm?ID=13590

Maslow, A. (1954). Motivation and personality. New York: Harper and Row.

Mazo, E. (2004). Post-Communist paradox: How the rise of parliamentarism coincided with the demise of pluralism in Moldova. CDDRL Working Papers: Stanford, CA.

McClure, C.E. (1991). Income tax reform in Colombia and Venezuela: A comparative history. Hoover Institution, Stanford, California.

McGlinchey, E.M. (2003). Paying for patronage: Regime change in post-Soviet Central Asia. Dissertation: Princeton University.

McGuire, M., & Olson, M. (1996). The economics of autocracy and majority rule. Journal of Economic Literature, 34, 72-96.

Mandelbaum, M. (1994). Central Asia and the World. New York: Council on Foreign Relations Press. 325

Ministry of Energy and Mineral Resources of the Republic of Kazakhstan. (2003). State Program for the Development of the Kazakhstan Sector of the Caspian Sea. Astana, Kazakhstan: Ministry of Energy and Mineral Resources of the Republic of Kazakhstan.

Monroe, K.R., Hankin, J., & Van Vechten, R.B. (2000). The psychological foundations of identity politics. Annual Review of Political Science, 3, 419-447.

Moore, B. (1966). Social origins of dictatorship and democracy: Lord and peasant in the making of the modern world. Boston, MA: Beacon Press.

Morrison, K. (2007). Non-tax revenue and the redistributional foundations of regime stability. Working Paper: Duke University.

Mortished, C. (2006, May 9). Venezuela doubles tax on oil majors. The Times (London), pp. 51.

Neary, P., & van Wijnbergen, S. (Eds.) (1986). Natural resources and the macroeconomy. Cambridge: MIT Press.

Why oil will get cheaper. (2006, July 17). Newsweek, pp. 62.

Nezavisimaya Gazeta. (2000, October 21). Moldovan President accuses MPs of upsetting balance of forces. BBC Monitoring Former Soviet Union. Retrieved from Lexis-Nexis database.

North, D.C. (1981). Structure and change in economic history. New York: W.W. Norton & Company.

O'Donnell, G. (1973). Modernization and bureaucratic-authoritarianism: Studies in Latin American politics. Berkeley, CA: Institute of International Studies.

Olcott, M.B. (2000). Regional study of human development and human rights. Human Development Report 2000 Backgrounder. Retrieved February 6, 2008, from http://hdr.undp.org/en/reports/global/hdr2000/papers/martha%20olcott.pdf

Olcott, M.B. (2002). Kazakhstan: Unfulfilled promise. New York: Carnegie Endowment.

Olcott, M.B. (2005). Central Asia’s second chance. Washington, DC: Carnegie Endowment for International Peace.

Olcott, M.B. (2005, April 7). Lessons of the ‘Tulip Revolution.’ Testimony prepared for the Commission of Security and Cooperation in Europe Hearing on Kyrgyzstan Revolution: Causes and Consequences.

Olowu, D. (1991). Centralization, self-governance and development in Nigeria. In J.S. 326 Wunsch, & D. Olowu (Eds.). The failure of the centralized state: Institutions and self- governance in Africa (pp. 193-227). Boulder, CO: Westview Press.

Olson, M. (1993). Dictatorship, democracy, and development. American Political Science Review, 87, 567-576.

Osaghae, E.E. (1998). The crippled giant: Nigeria since independence. Bloomington, IN: Indiana University Press.

Osaghae, E.E. (1995). Structural adjustment and ethnicity in Nigeria. Uppsala Research Report, no. 98.

Ottaway, M. (2001). Reluctant missionaries. Foreign Policy, 125, 44-54.

Ottaway, M. (2003). Democracy challenged: The rise of semi-authoritarianism. Washington, DC: Carnegie Endowment for International Peace.

Oxford Analytica. (2005). Canada to Compete in Oil Market. Forbes Online. Retrieved from Forbes Online archive January 20, 2007, from http://www.forbes.com/energy/2005/02/17/cz_0217oxan_canadaoil.html.

Moldova: Government Action Hinders Media Freedom. (2007, March 5). Oxford Analytica.

Moldova: Communists Face Setbacks in Local Elections. (2007, June 25). Oxford Analytica.

Moldova: Press Freedoms Remain Problematic. (2007, July 18). Oxford Analytica.

Pannier, B. (2002, 5 April). Kazakhstan: Prime Minister offers explanation of President’s alleged ill-gotten gains. Radio Free Europe/Radio Liberty. Retrieved February 19, 2008, from http://www.rferl.org/features/2002/04/05042002091023.asp

Nigeria: From Bad to Worse. (2006, November 3). Petroleum Economist, pp. 4.

Petrov, N.I., & Gafarly, M.S. (2001). The course towards political stability. In A. Vassiliev (Ed.)., Central Asia: Political and economic challenges in the post-Soviet era. London, UK: Saqi Books.

Pfeifer, S. (2006, December 30). EBRD set to deny loan to Sakhalin. The Telegraph. Retrieved from The Telegraph online archive April 16, 2007, from http://www.telegraph.co.uk/money/main.jhtml;jsessionid=IJTXN023IAKI1QFIQMFSFG GAVCBQ0IV0?xml=/money/2006/12/31/cnebrd31.xml.

Pinto, B. (1987). Nigeria during and after the oil boom: A policy comparison with Indonesia. World Bank Economic Review, 1, 419-445.

327 Plümper, T., & Troeger, V.E. (2007). Efficient estimation of time-invariant and rarely changing variables in finite sample panel analyses with unit fixed effects. Political Analysis, 15, 124-139.

Kazakhstan (South and Central Asia) comments and analysis. (2007, March 1). Political Risk Services (PRS) Group. Retrieved from Lexis-Nexis database.

Powell, G.B. (1982). Contemporary democracies: Participation, stability, and violence. Cambridge, MA: Harvard University Press.

Powell, G.B. (1996). Political cleavage structure, cross-pressure processes, and partisanship: An empirical test of the theory. American Journal of Political Science, 20, 1-23.

Press Service of Central Headquarters. (2005, December 2). Statement of December 2, 2005. For a Just Kazakhstan.

Press Service of Central Headquarters. (2005, December 7) Brief overview on major systematic violations on the voting day. For a Just Kazakhstan.

Press Service of Central Headquarters. (2005, December 15). Opposition appeals against illegal actions of the election commissions. For a Just Kazakhstan.

Przeworski, A. (1991). Democracy and the market. Cambridge, UK: Cambridge University Press.

Przeworski, A., & Limongi, F. (1997). Modernization: Theories and facts. World Politics, 49, 155-183.

Przeworski, A., Alvarez, M.E., Cheibub, J.A., & Limongi, F. (2000). Democracy and development: Political institutions and well-being in the world, 1950-1990. Cambridge, UK: Cambridge University Press.

Quinn, J.J. (1999). The managerial bourgeoisie: Capital accumulation, development and democracy. In D.G. Becker & R. Sklar (Eds.), Postimperialism and world politics (pp. 219-252), Westport, CT: Praeger Publishing.

Quinn, J.J. (2002). The road oft traveled: Development policies and majority state ownership of industry in Africa. Westport, CT: Praeger.

Quinn, J.J., & Conway, R.T. (2006). The resource curse revisited: What role ownership? Annual Meeting of the African Studies Association, San Francisco.

Rabushka, A., & Shepsle, K.A. (2007). Politics in plural societies: A theory of democratic instability. New York: Longman.

Remmer, K.L. (1991). “The Political Impact of Economic Crisis in Latin America in the 328 1980s.” American Political Science Review, 85, 777-800.

Respublika. (2005, April 22). Kazakh President’s election campaign starts. BBC Monitoring Central Asia Unit. Retrieved from Lexis-Nexis database.

Reuschmeyer, D., Stephens, E.H., & Stephens, J.D. (1992). Capitalist Development and Democracy. Chicago, IL: University of Chicago Press.

Reuters. (1995, August 31). New Kazakh constitution. The New York Times, p. A3.

Yukos declared bankrupt. (2006, August 1). Radio Free Europe/Radio Liberty. Retrieved February 17, 2007, from http://www.rferl.org/featuresarticle/2006/08/453ab5fd-285f-454a-8418- bf529e3dc7c2.html.

Prosecutors say Khodorkovsky, partner stole $32 billion in oil. (2007, February 9). Radio Free Europe/Radio Liberty. Retrieved February 18, 2007, from http://www.rferl.org/featuresarticle/2007/02/4b4c2f52-9c4f-4dfd-a07a- b695a4d3fcc7.html.

Accounting firm withdraws Yukos audit reports. (2007, June 25). Radio Free Europe/Radio Liberty. Retrieved February 18, 2007, from http://www.rferl.org/featuresarticle/2007/6/750DBCA9-3C2B-4FEE-85B9- F6210BB72744.html.

Foreign direct investment could reach $50 bln. in 2007. (2007, November 13). RIA Novosti. Retrieved December 14, 2007, from http://en.rian.ru/russia/20071113/87872457.html

Robinson, R. (1988). Authoritarian state, capital-owning classes, and the politics of newly industrializing countries: The case of Indonesia. World Politics, 41, 52-74.

Roeder, P.G. (1999). Peoples and states after 1989: The political costs of incomplete national revolutions. Slavic Review, 58, 854-882.

Romero, A. (1997). Re-arranging the deck chairs on the Titanic: The agony of democracy in Venezuela. Latin American Research Review, 32, 7-36.

Romero, S., & Holusha, J. (2006, January 20). Exxon Mobil post largest annual profits for US company. New York Times. Retrieved October 23, 2006, from http://www.nytimes.com/2006/01/30/business/30cnd- exxon.html?ex=1296277200&en=8ec83a7f4025b22b&ei=5088&partner=rssnyt&emc=rs s

Romero, S. (2007, December 3). Venezuela hands narrow defeat to Chavez plan. The New York Times, pp. A1.

329 Ross, M.L. (1999). The political economy of the resource curse. World Politics, 51, 297-322.

Ross, M.L. (2001). Does oil hinder democracy. World Politics, 53, 325-361.

Ross, M.L. (2002). Oil, democracy, and the resource curse. ASNews, 7, 1.

Ross, M.L. (2003). The natural resource curse: How wealth can make you poor. In I. Bannon and P. Collier (Eds.), Natural Resources and Violent Conflict: Options and Actions (pp. 17-42). Washington, DC: World Bank.

Ross, M.L. (2004). What do we know about natural resources and civil war? Journal of Peace Research, 41, 337-356.

Ross, M.L. (2006). A closer look at oil diamonds and civil war. Annual Review of Political Science, 9, 265-300.

Ross, M.L. (2006, May 2). Can Azerbaijan avoid the resource curse? Presentation at the Azerbaijani Alumni Association, Baku, Azerbaijan.

Rotunda, R.D. (2004, April 23). Iraq, oil and democracy. Cato Institute. Retrieved December 13, 2007, from http://www.cato.org/pub_display.php?pub_id=2623.

Rubin, J. (1993, October 8). Report says Russian hard-line leaders to face treason charges. The Associated Press. Retrieved from Lexis-Nexis database.

Rueschemeyer, D., Stephens, E.H., & Stephens, J.D. (1992). Capitalist development and democracy. Chicago, IL: University of Chicago Press.

Rustow, D.A. (1970). Transitions to democracy: Toward a dynamic model. Comparative Politics 2, 337-363.

Sachs, J.D., &Warner, A.M. (1995). Natural resource abundance and economic growth. NBER Working Paper, no.5398.

Salameh, G., & Steir, V. (1980). Political power and the Saudi state. MERIP Reports, 91, 5-22.

Sambanis, N. (2004). What is civil war?: Conceptual and empirical complexities of an operational definition. Journal of Conflict Resolution, 48, 814-858.

Schatz, E. (2004). Modern clan politics: The power of ‘blood’ in Kazakhstan and beyond. Seattle: University of Washington Press.

Scoffield, H. (2006, October 21). Too much, too fast: Can Alberta learn from Norway? The Globe and Mail, pp. B7. Retrieved from Lexis-Nexis database.

330 Schultz, J. (2005). Follow the money: A guide to monitoring budgets and oil and gas revenues. New York: Open Society Institute.

Scott, J.C. (1976). The moral economy of the peasant. New Haven: Yale University Press.

Scott, J.C. (1985). Weapons of the weak: Everyday forms of peasant resistance. New Haven: Yale University Press.

Shidenova, A. (2005, November 11). CEC Needs Aspirin. Kazakhstan Monitor, pp. 1.

Shleifer, A., & Treisman, D. (2000). Without a map: Political tactics and economic reform in Russia. Cambridge, MA: MIT Press.

Skocpol, T. (1979). States and social revolutions: A comparative analysis of France, Russia and China. Cambridge, UK: Cambridge University Press.

Skocpol, T. (1982). Rentier state and Shi’a Islam in the Iranian revolution. Theory and Society, 11, 265-282.

Smith, A. (1937[1776]). An inquiry into the nature and causes of the wealth of nations. New York: The Modern Library.

Smith, B. (2004). Oil wealth and regime survival in the developing world. American Journal of Political Science, 48, 232-246.

Smith, B. (2007). Hard times in the land of plenty: Oil politics in Iran and Indonesia. Ithaca, NY: Cornell University Press.

Smith, C.S. (2005, March 30). Ousted Kyrgyz leader hints he might officially resign. The New York Times, pp. A6.

Solovyov, V. (2006, November 8). Akayev: 'I don't understand their stubbornness.' Kommersant: pp. 1. Retrieved from Lexis-Nexis database.

Springborg, R. (Ed.). (2007). Oil and democracy in Iraq. London: Saqis.

STATA. (2005). Longitudinal/Panel Data, Release 9. College Station, TX: STATACorp LP.

STATA. (2005). Survival Analysis and Epidemiological Tables, Release 9. College Station, TX: STATACorp LP.

Stimson, J.A. (1985). Regression in space and time: A statistical essay. American Journal of Political Science, 29, 914-947.

Thomas, J.R. (1985). Natural resources in Soviet foreign policy. NY: National Strategy 331 Information Center.

Tsalik, S. (2002). Caspian oil windfalls: Who will benefit? NY: Open Society Institute.

Tsalik, S., & Schiffrin, A. (Eds.). (2005). Covering oil: A reporter’s guide to energy and development. New York: Open Society Institute.

Kazakh Parliament okays $891 million extra spending. (2005, 7 May). Turkish Daily News. Retrieved from Lexis-Nexis database.

Underhill, W. (2006, December 4). Trouble in paradise: Too much thrift is a bad thing, more Norwegians say. Newsweek, pp. 10.

US Interests in Central Asian Republics: Hearing before the Subcommittee on Asia and the Pacific of the Committee on International Relations House of Representatives, 105th Congress, 2nd Session. (1998).

Venn, F. (2002). The oil crisis. London: Longman.

Vernon, R. (1971). Sovereignty at bay. NY: Basic Books.

Vick, K. (2005, July 4). Iran guzzles gas at its own cost. Washington Post, pp. A12.

Wallerstein, I. (1974). The modern world system I: Capitalist agriculture and the origins of the European world-economy in the sixteenth century. New York: Academic Press.

Vremya. (2005, March 8). No illegal funding behind ruling party’s election campaign – Kazakh politician. BBC Monitoring Central Asia Unit. Retrieved from Lexis-Nexis database.

Wantchekon, L. (1999). Why do resource dependent countries have authoritarian governments? Leitner Working Article 1999-11.

Waterbury, J. (2003). Hate your policies, love your institutions. Foreign Affairs, 82, 53- 65.

Way, L. (2002). Pluralism by default in Moldova. Journal of Democracy, 13, 127-141.

Webb, M.J. (1961). Economic geography: A framework for a disciplinary definition. Economic Geography, 37, 254-257.

Weber, M. (2002). Politics as vocation. In C. Basteman (Ed.), Violence: A reader (pp. 1-12). New York: New York University Press.

A Modest Proposal for Bribe Takers. (2005, July 1). The Week, pp. 18.

332 Weinstein, M.A. (2004). The uncertainty of Iraq’s transition. Power and Interest News Report. Retrieved December 1, 2006, from http://www.pinr.com/report.php?ac=view_printable&report_id=188&language_id=1

Williams, B. (2003). Debate over peak-oil issue boiling over, with major implications for industry, society. Oil & Gas Journal, 101, pp. 18-29.

Wilson, S.E., & Butler, D.M. (2007). A lot more to do: The sensitivity of time-series cross-section analyses to simple alternative specifications. Political Analysis, 15, 101- 123.

Wright, G. (1990). The origins of American industrial success, 1879-1914. American Economic Review, 80, 651-658.

Wright, G. (2001). Resource-based growth then and now. Prepared for the World Bank Project ‘Patterns of Integration in the Global Economy.’

Wright, G. (2003). Mineral resources and economic development. Prepared for the Conference on Sector Reform in Latin America, Stanford Center for International Development.

Wright, G., & Czelusta, J. (2003). The myth of the resource curse. Working Paper: Stanford University.

Wrigley, E.A. (1988). Continuity, chance and change: The character of the industrial revolution in England. Cambridge, UK: Cambridge University Press.

CNPC Transfers 1/3 Stake in PK to KazMunaiGas. (17 July 2006). Xinhua News Agency. Retrieved from the China Institute, University of Alberta, February 26, 2008, from http://www.uofaweb.ualberta.ca/chinainstitute/nav03.cfm?nav03=48109&nav02=43876 &nav01=43092.

Yasuba, Y. (2006). Did Japan Ever Suffer from a Shortage of Natural Resources Before World War II? The Journal of Economic History, 56, 543-560.

Yuritsyn, V. (2005, September 30). Bees vs. Honey. Kazakhstan Monitor, pp. 2.

Zaller, J. (1992). The nature and origins of mass opinion. Cambridge, UK: Cambridge University Press.

Zorn, C.J.W. (2001). Generalized estimating equations models for correlated data: A review with applications. American Journal of Political Science, 45, 470-490.

Zssis, C. (2006). Chad's oil troubles. Council of Foreign Relations Backgrounder. Retrieved February 16, 2006, from http://www.cfr.org/publication/10532/chads_oil_troubles.html 333

Statistical Resources and Documentation

Alvarez, M., Cheibub, J.A., Limongi, F., & Przeworski, A.. (1996). Classifying political regimes. Studies in Comparative International Development, 31, 3-36.

Alvarez, M., Cheibub, J.A., Limongi, F., & Przeworski, A.. (1999). ACLP political and economic database codebook. Retrieved March 8, 2007, from http://www.ssc.upenn.edu/~cheibub/data/Default.htm

British Petroleum. (2005). Putting energy in the spotlight: BP statistical review of energy, June 2005. Retrieved January 7, 2006, from http://www.bp.com/statisticalreview

British Petroleum. (2007). BP statistical review of world energy, June 2007. Retrieved March 21, 2008, from http://www.bp.com/statisticalreview

Central Electoral Commission. (2006). Official results of 2005 presidential election. Retrieved January 15, 2006, from http://election.kz/portal/page?_pageid=73,88928&_dad=portal&_schema=PORTAL

Central Electoral Commission of the Republic Kazakhstan. (2007). Election of deputies to the Majilis of the Republic of Kazakhstan. Retrieved February 10, 2008, from http://election.kz/portal/page?_pageid=73,473388&_dad=portal&_schema=PORTAL

Energy Information Agency (EIA). (2006). Short term energy outlook: Real petroleum prices. Retrieved February 20, 2007, from http://www.eia.doe.gov/emeu/steo/pub/fsheets/PetroleumPrices_files/v3_document.htm.

FAOSTAT. (2006). FAOSTAT integrated core database, food consumption series. Retrieved June 16, 2007, from http://faostat.fao.org/

Freedom House. (2007). Freedom in the world: historical ratings. Retrieved March 10, 2008, from http://www.freedomhouse.org/template.cfm?page=5.

Gleditsch, N.P., Wallersetein, P., Eriksson, M., Sollenberg, M., & Strand, H. (2002). Armed conflict 1946-2001: A new dataset. Journal of Peace Research, 39, 615-637.

Gwartney, J., & Lawson, R. (2005). Economic freedom in the world: 2005 annual report. Vancouver, BC: The Frasier Institute. Retrieved March 21, 2008 from www.freetheworld.com

Hellman, J., Jones, G., & Kaufmann, D. (2000). Seize the state, seize the day: State capture, corruption, and influence in transition economies. World Bank Policy Research Working Paper No. 2444. Retrieved March 21, 2008, from http://info.worldbank.org/governance/beeps/

334 Kaufmann, D., Kraay, A., & Mastruzzi, M. (2003). Governance matters III: Governance indicators for 1996-2002. World Bank Working Papers and Articles. Retrieved April 1, 2008, from http://www.worldbank.org/wbi/governance/pubs/govmatters3.html.

Marshall, Monty G and Keith Jaggers. (2002). Polity IV Project: Political Regime Characteristics and Transitions, 1800-2002. Retrieved February 12, 2006, from http://www.cidcm.umd.edu/inscr/polity

The Decision is Made. (2005, December 5) Megapolis, pp. 1.

Miles, M.A., Holmes, K.R., O’Grady, M.A. (2006). 2006 index of economic freedom. Washington, DC: The Heritage Foundation. Retrieved April 21, 2008, from http://www.myheritage.org/Features/HMN/2006/winter/IndexFreedom.asp.

OSCE. (1996). Preliminary Statement Presidential Elections of the Republic of Moldova -- First Round 17 November 1996. Warsaw, Poland: OSCE.

OSCE. (1998). Preliminary Statement issued on the 23 March 1998 by the Election Observation Mission. Warsaw, Poland: OSCE.

OSCE. (1999). The Republic of Kazakhstan Presidential Election, 10 January 1999, Assessment Mission. Warsaw, Poland: OSCE.

OSCE. (2000). Republic of Kazakhstan Parliamentary Elections, 10 and 24 October 1999, Final Report. Warsaw, Poland: OSCE.

OSCE. (2000). Kyrgyz Republic Parliamentary Elections – 20 February & 12 March 2000. Warsaw, Poland: OSCE.

OSCE. (2001). Statement of Preliminary Findings and Conclusions: Republic of Moldova elections to the parliament, 25 February 2001. Warsaw, Poland: OSCE.

OSCE. (2004). Republic of Kazakhstan Parliamentary Elections, 19 September and 3 October 2004. Warsaw, Poland: OSCE.

OSCE. (2005). Presidential Election, Republic of Kazakhstan -- 4 December 2005: Statement of Preliminary Findings and Conclusions. Warsaw, Poland: OSCE.

OSCE. (2005). Statement of Preliminary Findings and Conclusions: Parliamentary elections, Republic of Moldova – 6 March 2005. Warsaw, Poland: OSCE.

OSCE. (2005). Statement of preliminary findings and conclusions by the OSCE/ODIHR election observation mission on the parliamentary elections held in Kyrgyzstan on 27 February 2005. Warsaw, Poland: OSCE.

OSCE (2005, December 5). Kazakh Elections Flawed Despite Some Administrative 335 Improvements. Retrieved from OSCE homepage January 20, 2006, from http://www.osce.org/item/17236.html.

Nazarbayev Wins Kazakh Vote with 83 pct. (2005, December 4) Reuters. Retrieved December 6, 2005, from http://today.reuters.com/news/newsArticle.aspx?type=worldNews&storyID=2005-12- 05T033301Z_01_KWA512690_RTRUKOC_0_US-KAZAKHSTAN- EXITPOLL.xml&archived=False.

Roeder, P.G. (2001). Ethnolinguistic Fractionlization (ELF) Indices, 1961 and 1985. Retrieved March 23, 2006, from http://weber.ucsd.edu/~proeder/elf.htm

Sarkees, M.R. (2000). The correlates of war data on war: An update to 1997. Conflict Management and Peace Science, 18, 123-144.

Shokamanov, Y.K., Amirkhanova, M.M., Kalmanova, N.K., & Massabek, E.N. (Eds.). (2005). Women and Men of Kazakhstan: Gender Statistics. Almaty, Kazakhstan: Agency of the Republic of Kazakhstan on Statistics.

United Nations Development Programme. (2002) Arab Development Report. New York: United Nations.

World Bank. (2003). Governance Indicators for 1996-2002. Washington, DC: World Bank. Retrieved March 21, 2005, from http://www.worldbank.org/wbi/governance/pubs/govmatters3.html

World Bank. (2003). World Development Indicators. Washington, DC: World Bank, CD-ROM.

World Bank (2007). World Development Indicators. Washington, DC: World Bank. Retrieved March 17, 2008, from http://ddp-ext.worldbank.org/ext/DDPQQ/ member.do?method=getMembers&userid=1&queryId=6

UNDP. (2003). The Arab Human Development Report 2003: Building a Knowledge Society. New York: United Nations Development Programme.

Selected Interviews and Conferences

Kazakhstan

Anton Artemyev, program director of Kazakhstan Revenue Watch, September 27, 2005.

Alikhan Baimenov, candidate for President of the Republic of Kazakhstan, leader of Ak Zhol Party, “Meeting with Presidential Candidates of the Republic of Kazakhstan,” Meeting with Klub Polyton discussion group, November 24, 2005.

336 Kanat Barantayev, economist and co-founder of Public Policy Research Center in Almaty, Kazakhstan, November 14, 2005.

Mels Eleusizov, candidate for President of the Republic of Kazakhstan, leader of environmental movement of Kazakhstan, “Meeting with Presidential Candidates of the Republic of Kazakhstan,” Meeting with Klub Polyton discussion group, November 4, 2005.

Ertysbaev Ermuhamet, political advisor to the President of the Republic of Kazakhstan, “N.A. Nazarbayev’s Pre-election Campaign and the Course of the Election,” Meeting with Klub Polyton discussion group, November 25, 2005.

Kazakhstan Oil and Gas Exhibition (KIOGE), October 4-7, 2005.

Nurbulat Masanov, President of the Kazakhstan Political Science Association, November 4, 2005.

Meruert Makhmutova, director of the Public Policy Research Center in Almaty, Kazakhstan, October 17, 2005.

Adil Nurmakov, head of external relations department of opposition group For a Just Kazakhstan, November 10, 2005, December 12, 2005.

Zharmakhan Tuyakbai, candidate for President of the Republic of Kazakhstan, leader of For a Just Kazakhstan political movement, “Meeting with Presidential Candidates of the Republic of Kazakhstan,” Meeting with Klub Polyton discussion group, November 21, 2005.

Robert Robertson, dean of the Bang College of Business, October 28, 2005.

4 partners in law firms representing businesses in Kazakhstan

2 partners with major tax auditing and representation firms

3 current and former oil executives

5 executives and lawyers for PetroKazakhstan

Moldova

Arcadie Barbarosie, director of the Moldova Institute of Public Policy, November 26, 2006.

Lilia Carasciuc, director of Transparency International, Moldova, June 2, 2006.

337 Valentin Crilov, executive secretary of the Socialist Party of Moldova, ‘Patria-Rodina’, May 30, 2006, June 5, 2006.

Victor Draguton, member of the Central Committee of the Party of Communists of the Republic of Moldova, December 4, 2006.

International Scientific Conference ‘Process of European Integration: Opportunities, Costs, Risks, Advantages.’ Moldova State Institute of International Relations, May 5, 2006.

Petru Lucinschi, former president of the Republic of Moldova, January 22, 2007.

Serghei Ostaf, head of Moldova Resource Center for Human Rights (CReDO), June 26, 2006.

338