Bureaucratic Discretion, Regulatory Uncertainty, and Private Investment in the Russian Federation

RISK IN THE REGIONS: BUREAUCRATIC DISCRETION, REGULATORY UNCERTAINTY, AND PRIVATE INVESTMENT IN THE RUSSIAN FEDERATION

DISSERTATION

Presented in Partial Fulﬁllment of the Requirements for the Degree Doctor of

Philosophy in the Graduate School of the Ohio State University

Quintin Hayes Beazer, B.A./M.A.

Graduate Program in Political Science

The Ohio State University

2011

Dissertation Committee:

Irfan Nooruddin, Advisor

Timothy Frye, Advisor

Marcus Kurtz

William Minozzi c Copyright by

Quintin Hayes Beazer

2011 ABSTRACT

This dissertation identifies bureaucratic discretion – agents’ leeway in making subjective determinations about when and how rules apply – as a primary source of uncertainty that deters long-term investors by undermining the predictability of firms’ regulatory environment. Using a principal-agent framework, I argue that where regulatory bureaucrats exercise greater discretion in interpreting and applying laws, economic actors experience greater uncertainty about how those policies will be put into practice. This unpredictability deters investment by making it difficult for economic actors to predict how the regulatory environment will affect their projects’ future returns and costs. Whereas existing studies tend to focus on the economic effects of particular policies or government decisions, the dissertation’s focus on the predictability of policy application calls attention to the consequences for private actors related to how government agents carry out those decisions.

In arguing that bureaucratic discretion can be a source of uncertainty for investors, my argument challenges an existing literature based upon studies within developed democracies that emphasize the economic beneﬁts of insulating state actors, such as bureaucrats, from the pressures of the political arena. Faced with this apparent conﬂict, I provide a theoretical framework for explaining why bureaucratic discretion might create more uncertainty in some locations than in others. Arguing that bureaucratic discretion should generate greater or lesser uncertainty depending upon

ii the broader institutional context in which it is granted, I call attention to one particular conditioning factor: the level of political competition. I argue that political competition makes policy application more predictable by making politicians more responsive to constituents’ concerns about bureaucratic discretion and by spreading the costs of monitoring bureaucrats across non-state actors and supporting institutions. In contrast, economic actors bear the full brunt of regulatory uncertainty in politically-uncompetitive regions — investors face more unpredictable behavior from discretionary bureaucrats and fewer formal channels for handling disputes that may arise.

I develop this argument by examining private investment across the regions of the Russian Federation. Using a survey of Russian firm managers (Frye 2006), I find that firm managers who perceive bureaucrats to have high discretion are less likely to plan fixed-capital investments for the immediate future, ceteris paribus. In addition, I rely on qualitative evidence from field interviews with Russian firm managers, business association leaders, and policy experts to provide insight into exactly how uncertainty over bureaucrats’ application of regulatory laws shapes firms’ decisions about where and how to invest. After finding that, on average, discretion corresponds with reduced incentives for firms to invest, I merge the firm-level survey data with regional data on political competition to demonstrate that the negative relationship between discretion and investment is most pronounced in regions of Russia where surrounding institutions limit political competition. Among firms that perceive bureaucrats to have high discretion, those located in politically-uncompetitive regions have a much lower probability of investing than their counterparts in regions with high political competition. Additional analyses bolster confidence that these results do arise via the theory’s proposed causal mechanisms; in the final empirical chapter, I

iii ﬁnd evidence supporting the link between political competition on the one hand and increased government responsiveness and better-behaved bureaucrats on the other.

iv For my parents, Sherman and Lorilee Beazer,

and my wife, Brooke Petersen Beazer.

v ACKNOWLEDGMENTS

This dissertation would not have been happened without the help and support of many people. While it would be impossible to name all those whose help has been invaluable along the way, I wish to express my gratitude to a number of individuals that have made a diﬀerence to me, both professionally and personally.

Among my blessings in life, I count as especially choice the opportunity that I have had to work with my committee members. As scholars and as individuals, there are few people with whom I’d rather spend my time. I took my ﬁrst graduate seminar from Marcus Kurtz, and since that time I have looked to him as a guide and authority on all things comparative. I have always beneﬁted by heeding Marcus’s trademark skepticism and by listening to his generous advice. Even before William Minozzi ever became part of my committee, he had spent hours helping me frame my dissertation ideas and recognize their contribution to broader literatures. I thank William for going above and beyond the call of duty in all respects and for opening my eyes to what it means to be a social scientist. I am especially grateful to Timothy Frye and

Irfan Nooruddin for their work in co-chairing this dissertation. They are role models for me in the truest sense, and I look forward to continuing to work closely with them for many more years to come. Tim’s political economy research in Russia and the post-communist countries stands as the gold standard for what I hope my work can one day become. Tim deserves special thanks, not only for the use of his survey data

vi in this dissertation, but for bringing me to Ohio State and providing an open door whenever I have needed guidance. Even after his departure to Columbia University, I have continued to proﬁt from his patience, enthusiasm, and unerring instincts about which ideas were dead-ends and which ideas were worth pursuing. Similarly, I am forever grateful to Irfan for taking me under his wing. Irfan has taught me about research, about the discipline, and about what it means to give freely to others. As an advisor, he has encouraged, restrained, corrected, questioned, demurred, vetoed, demanded, promoted, pushed, chastised, defended, and applauded at all the right times. Irfan’s indelible mark is evident on every page of this dissertation, just as

I hope it will be in all the academic research that I will ever do. He has been an exceptional mentor, and I consider Irfan among my closest of friends.

In addition to the members of my committee, I owe a debt of gratitude to Sarah

Brooks, Mike Neblo, Massimo Morelli, Philipp Rehm, Alex Thompson, Herb Weis- berg, Alan Wiseman and many others in the Political Science Department at Ohio

State for their superb instruction, their sage advice, and their open doors. Spe- cial thanks go to Craig Volden and Jeremy Wallace for numerous conversations at multiple stages of this project; their thoughtful comments and suggestions have left lasting impressions on the way I see both my own research and the discipline in general. One of the deﬁning moments of my graduate career was the chance to work with Janet Box-Steﬀensmeier as the Junior Fellow in the Program in Statistics and

Methodology. A consummate professional, a committed parent, an enthusiastic researcher, and a warm friend – Jan continues to be one of my heroes. Beyond Ohio

State, I also thank Noah Buckley, Scott Cooper, Michael Findley, Ethan Bueno de

Mesquita, Scott Gehlbach, Bonnie Meguid, Graeme Robertson, Konstantin Sonin,

Joshua Tucker, Katia Zhuravskaya, and three anonymous reviewers at the Journal of

Politics for comments, questions, and encouragement that have helped to move this

vii project forward. Likewise I am indebted to the Horowitz Foundation for Social Policy, the Mershon Center, the Ohio State University and the Frances Aumann family for helping to fund travel for the dissertation ﬁeld research.

In Russia, my deep thanks go to Sergei Guriev, Maria Bolotskaya, and the Cen- tre for Economic and Financial Research (CEFIR) at the New Economic School for granting me oﬃce space in Moscow and access to one of the most vibrant research environments one could imagine. I am very grateful to those experts and professionals who generously shared their time to participate in my interviews. The pain of leaving my family at home during my ﬁeld work in Russia was made tolerable by good friends and surrogate family: to Noah Buckley, Kolya Makarov, Misha Moiseev, and my Moscow branches of the Church of Jesus Christ of Latter-Day Saints, I say spasibo. Finally, I express my love to Lyuba and Lena Morozova, my adopted sister and mother in Russia, for opening their home to me during my returns to Moscow as eagerly as they did nearly a decade earlier.

Graduate study at Ohio State has been a journey ﬁlled with fabulous traveling companions from whom I continue to learn much, including Soundarya Chidambaram,

Michael Cohen, Dinissa Duvanova, Ryan Kennedy, Carolina Mercado, Jason Mor- gan, Banks Miller, Yoonah Oh, Autumn Lockwood Payton, Allyson Shortle, Anand

Sokhey, Sarah Wilson Sokhey, Dana Wittmer, Byungwon Woo, and Kadir Yildirim.

In particular, graduate school has provided me with three remarkable friends and fellow musketeers – Daniel Blake, Dino Christenson, and Scott Powell – upon whom

I rely daily for honesty, laughs, and sound advice.

There is no amount of thanks that I can give to my parents, Sherman and Lorilee

Beazer, that could ever sufficiently convey the depth of my gratitude for what they have given me. Throughout my life, my parents have been the very models of selfless sacrifice, unconditional love, and unwavering dedication. Their support, along with

viii that of my two sisters, Mary and Alynne, has been a endless source of motivation.

During my seven years of graduate school, parenthood has brought me three little assistants: Lydia, Corbin, and Graham. They have been very patient with a daddy who needed to work before he could play, and I look forward to celebrating their accomplishments with them as heartily as they have celebrated mine with me. Finally, no one has suﬀered more for this dissertation than my wife, Brooke. She is a faithful friend and unfailing teammate who knows when to listen sympathetically and when to push me. Her wisdom, grace, and humor make life wonderful. For this and a million reasons more, my deepest gratitude and my warmest aﬀection belong to Brooke.

ix VITA

1980 ...... Born in Sunset, Utah, USA

2004 ...... B.A. in English, International Studies; mi- nor in Russian, Utah State University

2006 ...... M.A. in Political Science, The Ohio State University

2009 ...... M.A. in Economics, The Ohio State Uni- versity

2004-Present ...... Graduate Teaching/Research Associate, The Ohio State University

FIELDS OF STUDY

Major Field: Political Science

Specialization: Comparative Politics

Specialization: Political Economy

x TABLE OF CONTENTS

Abstract ...... ii

Dedication ...... iv

Acknowledgments ...... vi

Vita...... x

List of Tables ...... xiv

List of Figures ...... xvi

CHAPTER PAGE

1 Introduction ...... 1

2 A Theory of Bureaucratic Discretion and Investment Behavior . . . . . 12

Introduction ...... 12 Political Explanations of Investment ...... 20 Policy Content ...... 20 Policy Stability ...... 22 Policy Application ...... 26 Bureaucratic Discretion and Investment: A Theoretical Framework . . 29 Bureaucracy and Regulation ...... 30 Principal-Agent Relationships, Discretion, & Uncertainty . . . . 34 Competing Views of Bureaucratic Discretion ...... 40 Bureaucratic Discretion and Political Context ...... 44 Hypotheses ...... 51 Conclusion ...... 53

3 Bureaucratic Discretion and Business Investment: Evidence from Rus- sian Enterprises ...... 56

Introduction ...... 56 Quantitative Analyses of Discretion & Investment ...... 57 Independent Variables ...... 59

xi Qualitative Evidence from Field Interviews ...... 70 Conclusion ...... 82

4 Does Political Competition Make Bureaucratic Discretion Better for In- vestors? An Empirical Test ...... 84

Introduction ...... 84 Political Competition & Russia’s Regions ...... 85 Testing the Conditional Theory of Bureaucratic Discretion & Investment 90 Results from Multilevel Statistical Analyses ...... 90 Robustness Checks & Additional Analyses ...... 99 Alternate Measures of the Key Independent Variables ...... 99 Alternate Measures of Investment ...... 103 Conclusion ...... 110

5 How Does Political Competition Make Bureaucratic Discretion Better for Investors? Testing the Mechanisms ...... 114

Introduction ...... 114 Political Competition & Responsive Politicians ...... 116 Political Competition & Regulatory Bureaucrats ...... 132 Corruption & Discretion ...... 134 Disputes & Discretion ...... 139 Conclusion ...... 143

6 Conclusion ...... 145

Goals of the Dissertation ...... 146 Empirical Findings of the Dissertation ...... 148 Implications of the Dissertation ...... 151

Bibliography ...... 159

CHAPTER PAGE

A Additional Material for Chapter 3 ...... 169

B Additional Material for Chapter 4 ...... 175

C Additional Material for Chapter 5 ...... 186

D Additional Tests: Missing Mechanisms, Measurement, and Validity . . 189

Introduction ...... 189 Bureaucratic Discretion & Regulatory Uncertainty ...... 189 Internal Validity & Perceptions of Discretion ...... 192 External Validity ...... 204

xii Validating Perceptions of Discretion ...... 204 Validating Intent to Invest ...... 207 Summary ...... 208

xiii LIST OF TABLES

TABLE PAGE

2.1 Private Investment Varies Across Russia ...... 14

3.1 Summary Statistics for Variables from Frye (2006) Data ...... 59

3.2 Perceived Bureaucratic Discretion Associated with Less Investment . 61

3.3 Discretion Associated with Lower Probability of Investment . . . . . 64

4.1 Regional Democracy Scores, 2000-2004 (Regional Monitoring Project) 87

4.2 Institutional Context Aﬀects Discretion’s Relationship with Invest- ment ...... 91

4.3 Political Context Conditions Investors’ Response to Discretion . . . 93

4.4 Robustness Check: Alternate Measures of Independent Variables . . 100

4.5 Robustness Check: Investment-Related Dependent Variables . . . . . 106

5.1 Assessment of Elected Oﬃcials: Diﬀerence of Means ...... 118

5.2 Lobbying Eﬀorts: Diﬀerence of Means ...... 121

5.3 Lobbying Response to Discretion Depends upon Political Environment 126

5.4 Comparing Strategies for Contesting Bad Claims: Responses to a Hy- pothetical Scenario ...... 130

5.5 Lobbying Eﬀorts: Diﬀerence of Means ...... 133

5.6 Discretion Less Associated with Corruption in High Competition Regions136

5.7 Discretion’s Association with Disputes Depends upon Context . . . 141

A.1 Descriptive Statistics of Frye (2006) Survey Sample ...... 169

A.2 Firm-Level Analyses: Robustness Check (Ordinal DV) ...... 171

xiv A.3 Firm-Level Analyses: Alternate Model Speciﬁcations ...... 172

B.1 Summary Statistics for Region-Level Variables ...... 176

B.2 Robustness Check: (Alternate Model Speciﬁcations) ...... 177

B.3 Robustness Check: Bayesian Hierarchical Analysis with Diﬀuse Priors 178

B.4 Robustness Check: Dropping Outlying Region ...... 179

B.5 Multilevel Analyses: Disaggregating Political Competition Index . . . 180

D.1 Controlling for Lack of Political Knowledge ...... 195

D.2 Controlling for Experience with Government & Job ...... 197

D.3 Controlling for Experience Outside the Region ...... 201

D.4 Controlling for Respondent Optimism ...... 203

D.5 Testing Regional Diﬀerences in Perceptions of Discretion: ANOVA . 204

xv LIST OF FIGURES

FIGURE PAGE

2.1 Private Investment Spread Unevenly Across Russia, Even Excluding Oil-Rich Regions ...... 16

2.2 Private Investment Varies Widely Across the Russian Federation . . 17

2.3 Firm Managers Spend a Signiﬁcant Amount of Time with Regulators 32

2.4 Businesses Subjected to Multiple State Inspections on a Regular Basis 33

3.1 Discretion Associated with Large Decreases in Probability of Invest- ment (Estimated Eﬀects) ...... 68

4.1 Sampled Regions & Variation in Political Competition ...... 89

4.2 Investors’ Response to Bureaucratic Discretion Shaped by Institu- tional Context (Predicted Probabilities) ...... 97

4.3 Predicted Probability Plots: Proportional Representation as Measure of Political Competition ...... 104

4.4 Predicted Probability Plot: Investing in Labor Skills (Alternate DV) 108

4.5 Predicted Probability Plot: Extending Credit to Buyers (Alternate DV)...... 109

4.6 Predicted Probability Plot: Conducting Marketing Study (Alternate DV)...... 111

5.1 Marginal Eﬀects of Bureaucratic Discretion and Political Competition on Probability of Hiring Manager with Government Ties ...... 127

5.2 Marginal Eﬀects of Bureaucratic Discretion and Political Competition on Bribes’ Severity ...... 138

5.3 Marginal Eﬀects of Bureaucratic Discretion and Political Competition on Probability of Dispute with Government ...... 142

xvi 6.1 Scatterplot of Regional Private Investment & Aggregated Measure of Bureaucratic Discretion ...... 153

6.2 Scatterplot of Regional Private Investment & Legislative Constraints on Discretion ...... 155

B.1 Predicted Probability Plots: Index Component – Openness ...... 181

B.2 Predicted Probability Plots: Index Component – Elections ...... 182

B.3 Predicted Probability Plots: Index Component – Pluralism ...... 183

B.4 Predicted Probability Plots: Dichotomous Measure of Bureaucratic Discretion ...... 184

B.5 Predicted Probability Plots: Dichotomous Measure of Political Com- petition ...... 185

C.1 Diﬀerence between High- and Low-Competition Regions: Comparing Across Subsets ...... 187

C.2 Diﬀerence between High- and Low-Competition Regions: Comparing Across Subsets ...... 188

D.1 Policy Uncertainty & Variance in Time with Oﬃcials (Russian BEEPS Data) ...... 191

D.2 Policy Uncertainty & Variance in Time with Oﬃcials (WBES Data) 193

D.3 Perceptions of Discretion versus Regulatory Executive Directives . . 206

D.4 Intentions to Invest and Regional Private Investment (2005) . . . . . 208

xvii CHAPTER 1

INTRODUCTION

Two fundamental tasks of government are to make and implement policy, and the manner in which governments carry out both tasks is crucial to the private actors that are governed by the state. This is especially true with regards to businesses and their response to states’ economic policies. As such, the question of what determines whether policy environments attract or deter investment is central to research on the interaction between states and markets. While the quality of governments’ economic policies and their eﬀects on market actors have been well-studied, this dissertation examines instead how the second half of government’s task – policy implementation – shapes the investment decisions that, over the long-run, determine nations’ economic performance.

This dissertation identiﬁes bureaucratic discretion – agents’ leeway in making subjective determinations about when and how rules apply – as a primary source of uncertainty that deters long-term investors by undermining the predictability of

ﬁrms’ regulatory environment. By highlighting bureaucrats’ role in determining how and when regulatory laws apply to businesses, the theory emphasizes a group of political actors that many traditional explanations of state-business relations overlook.

Using a principal-agent framework, I argue that where regulatory bureaucrats exercise greater discretion in interpreting and applying laws, economic actors experience

1 greater uncertainty about how those policies will be put into practice. This unpredictability deters investment by making it diﬃcult for economic actors to predict how the regulatory environment will aﬀect their projects’ future returns and costs. This project contributes to debates on the role of the state in the economy by highlighting how bureaucratic discretion shapes the dynamics of state-business interactions.

Whereas existing studies tend to focus on the economic eﬀects of particular policies or government decisions, the dissertation’s focus on the predictability of policy application calls attention to the consequences for private actors related to how government agents carry out those decisions.

The dissertation makes a second theoretical contribution. In arguing that bureaucratic discretion can be a source of uncertainty for investors, my argument challenges an existing literature that emphasizes the economic beneﬁts of insulating state actors, such as bureaucrats, from the pressures of the political arena. For some, leaders’ political incentives to reverse, manipulate, or renegotiate policy create a pressing need to put policy management in the hands of independent or insulated bureaucrats (Rogoﬀ

1985, Levy & Spiller 1994, Miller 2000). Such theories may ﬁnd support from studies within developed democracies, but often their predictions appear to be at odds with institutional and administrative realities of the developing world.

Faced with this apparent conﬂict, I provide a theoretical framework for explaining why bureaucratic discretion might create more uncertainty in some locations than in others. Arguing that bureaucratic discretion should generate greater or lesser uncertainty depending upon the broader institutional context in which it is granted, I call attention to one particular conditioning factor: the level of political competition. I argue that political competition makes policy application more predictable by making politicians more responsive to constituents’ concerns about bureaucratic discretion and by spreading the costs of monitoring bureaucrats across non-state actors and

2 supporting institutions. This diﬀuse monitoring helps to stabilize investors’ expectations about the regulatory environment. In contrast, a lack of competition in the political environment dulls leaders’ incentives to respond to regulatory problems that arise from their agents’ behavior. Furthermore, such environments restrict economic actors’ access to the policy-making process and reduce their ability to help monitor state agents. Thus, economic actors bear the full brunt of regulatory uncertainty in politically-uncompetitive regions — investors face more unpredictable behavior from discretionary bureaucrats and fewer formal channels for handling disputes that may arise.

I develop this argument by examining private investment across the regions of the Russian Federation. I pursue a multi-methods approach that employs quantitative data from multiple levels of analysis alongside qualitative evidence from field interviews. Using a survey of Russian firm managers (Frye 2006), I test the theory’s micro-level arguments about the link between bureaucratic behavior and firm investment decisions. I find that firm managers who perceive bureaucrats to have high discretion are less likely to plan fixed-capital investments for the immediate future, ceteris paribus. In addition, I rely on qualitative evidence from field interviews with

Russian ﬁrm managers, business association leaders, and policy experts to complement the statistical analyses; these interview data provide insight into exactly how uncertainty over bureaucrats’ application of regulatory laws shapes ﬁrms’ decisions about where and how to invest.

After ﬁnding that, on average, discretion corresponds with reduced incentives for

firms to invest, I merge the firm-level survey data with regional data on political competition to demonstrate that the negative relationship between discretion and investment is most pronounced in regions of Russia where surrounding institutions limit political competition. Among firms that perceive bureaucrats to have high discretion,

3 those located in politically-uncompetitive regions have a much lower probability of investing than their counterparts in regions with high political competition. More- over, I show that this result holds for ﬁrm activities that should follow a similar logic to ﬁxed capital investment, such as investing in workers’ skills through labor training programs, extending credit to buyers, and conducting extensive marketing research.

Finally, as a follow-up to these key findings, I use a series of statistical analyses on multiple measures of firms’ attitudes and reported behavior to bolster confidence that the results of the conditional model do arise via the theory’s proposed causal mechanisms. In line with the implications of the dissertation’s theoretical framework, I

ﬁnd supporting the link between political competition on the one hand and increased government responsiveness and better-behaved bureaucrats on the other.

Taken together, the empirical ﬁndings provide a nuanced understanding of how bureaucratic discretion aﬀects economic actors and their investment environment.

The findings offer strong support for the argument that bureaucratic discretion is a troubling source of unpredictable regulation that can deter investors, but emphasize at the same time that the negative effects of bureaucratic discretion are conditioned by the broader political environment and the degree to which political competition creates space for business actors to identify and address problems created by discretionary agents.

The remainder of this dissertation proceeds according to the following outline.

Chapter 2 is the theoretical chapter, establishing the causal argument that links private investment, bureaucratic discretion, and the conditioning effects of the broader political environment in which investors’ find themselves. Chapters 3 through 5 present the empirical analyses, with each chapter testing a different set of predictions generated by the theoretical framework developed in Chapter 2. Following the empirical chapters, Chapter 6 concludes.

4 My dissertation explores the eﬀects of bureaucratic discretion on investment.

Given the importance of investment for economic growth and development, identifying the political factors that inﬂuence investors is a central research agenda in comparative and international political economy. Why do some governments succeed in creating the conditions that promote long-term investment while others attract very little private investment? More generally, what determines whether policy environments attract or deter investment? Previous social science research on the political determinants of investment sorts roughly into two approaches that either focus on the speciﬁc incentives created by policies’ content (Hines 2001), or else stress the need for credible commitments to stable policies (Aizenman & Marion 1993,1999; North

& Weingast 1989). Both camps have contributed greatly to our understanding of how political factors shape investment, yet each side has overlooked a fundamental concern for ﬁrms that, ultimately, must satisfy legal requirements: how are laws and regulations applied? Overlooking policy application’s importance to investors has led scholars to focus disproportionately on politician-investor interactions and underesti- mate the way that bureaucratic agents shape the policy environment.

Chapter 2 presents a theory of investment behavior that highlights the role of bureaucrats in creating regulatory conditions that influence business actors’ willingness to invest. Profit-motivated firms need to know about the actual application of laws in order to predict how government regulations will affect their business interests. For investors contemplating a long-term project, uncertainty about policy application raises the risk that unanticipated losses will erode their profits. Using a principal-agent framework, I argue that discretion deters investment because it increases uncertainty about the way that regulatory bureaucrats interpret and apply laws, thus making it more difficult for investors to predict how the regulatory environment will affect their business plans. When bureaucrats apply statutes consistently

5 and predictably, on the other hand, business actors can incorporate regulation into their investment formula as a known parameter.

For some scholars, politicians’ control over the state apparatus is the main contributor to unpredictability in the policy environment. Politicians respond to political rather than economic incentives, and their eﬀorts to please constituent groups or further their political goals can make it hard for economic actors to predict changes in government policy that might aﬀect their business interests (Levy & Spiller 1994,

Miller 2000). According to such arguments, one solution lies making bureaucrats more independent as a means to insulate policies on the ground from the political arena (Rauch & Evans 1999, Bertelli & Whitford 2009). To address the apparent disagreement between my argument and scholarly accounts that emphasize the economic beneﬁts of cultivating an insulated bureaucratic corps, I extend the theory to explain how institutional context aﬀects the level of uncertainty that bureaucratic discretion creates for investors.

In Chapter 2, I call attention to the beneﬁcial role of political competition in mitigating the uncertainty that businesses associate with bureaucratic discretion.

Institutions that encourage political competition make policy implementation more predictable in at least two ways. First, competitive politics help to keep politicians attentive, effectively raising the principals’ costs from disputes caused by their agents’ behavior. Politicians worried about losing office have a pressing electoral incentive to monitor the bureaucracy and reduce regulatory disputes in order to avoid scandal or prevent political opposition from capitalizing on constituents’ dissatisfaction with inconsistent legal requirements or aggravating treatment by state officials. Sec- ond, by giving economic actors more opportunities to influence regulatory policy and voice concerns about about administrative practices, political competition helps to spread the costs of monitoring bureaucrats across actors and stabilize expectations

6 about how bureaucrats will use their discretion to interpret and apply regulations.

This discussion leads to the testable prediction that bureaucratic discretion should not obstruct business investment where political competition is high because investors should have clearer expectations about what regulators want and how agencies’ discretion will aﬀect their business interests. But, where the political environment restricts competition, there are fewer active limitations on how agents will use their discretion, and bureaucratic discretion remains a stumbling block for investors since they cannot rely on institutional constraints to enhance the predictability of the regulatory environment.

Focused on economic actors’ perceptions and behavior, Chapter 3 examines the theory’s individual-level logic using quantitative data from business surveys as well as qualitative evidence from field interviews. To test the prediction that firms’ perceptions of bureaucratic discretion affect their decisions about investment, I use a survey of 666 Russian firm managers conducted in 2005 by Timothy Frye (see Frye 2006). In particular, this survey inquires about how firms perceive various institutional actors, including regulatory bureaucrats. This makes the data a natural fit for testing the argument’s microeconomic predictions about individual actors’ subjective investment decisions. Using logit models of firm investment, I find strong support for my core hypothesis that bureaucratic discretion deters investment. Specifically, I demonstrate that firm managers who perceive regional bureaucrats to have high levels of decision- making discretion are less likely to plan fixed-capital investments for the immediate future.

To complement to the statistical analyses, I present qualitative evidence gathered from over forty ﬁeld interviews with ﬁrm directors, heads of business associations, policy advocates, and legal experts in Russia. These interviews help by corroborating the theory’s proposed causal mechanisms with observations from people who know

7 and operate within Russian markets. In particular, interviews reveal that: ﬁrst, investors do indeed associate discretion with uncertainty about how bureaucrats will apply regulatory laws to their business; and second, unpredictable regulatory assessments create losses that can drain ﬁrms’ resources and derail their investment plans.

Consistent with my theoretical argument, the qualitative evidence from these interviews indicates that investors in Russia regard unpredictable policy application as a primary concern and that ﬁrms consider carefully the behavior of regional regulatory bureaucrats when researching new investment locations.

While Chapter 3 establishes empirical support for the theory’s main behavioral hypothesis, Chapter 4 investigates the dissertation’s unique claim that the broader institutional context can either mitigate or magnify the economic uncertainty created by bureaucratic discretion. Combining the ﬁrm survey data with regional data on political competition from the Moscow Carnegie Center’s Regional Monitoring Project,

I use multilevel models of firm investment to test the prediction that discretion’s negative effect on investment is strongest in regions where political competition is restricted. I find the negative relationship between firm’s plans to invest and perceived bureaucratic discretion to be especially pronounced for firms in regions of Russia with limited political competition. These results suggest that greater independence for agents in such circumstances only heightens investors’ uncertainty about the policy environment. In the sample’s most restrictive political environment, a firm manager who perceives regional bureaucrats to have high discretion has a predicted probability of investment that is nearly 50 percentage points lower than a hypothetical counterpart who describes regional bureaucrats as having no discretion. Additional analyses provide similar results for firms’ other investment-related activities such as investing in worker-training programs, extending credit to their buyers, or conducting extensive marketing research.

8 These findings help resolve the apparent tension between my argument and the literature’s seemingly sanguine attitude towards bureaucratic discretion. In the developed democracies that inform a majority of studies, high-quality institutions attenuate investors’ uncertainty about independent bureaucrats’ behavior. Conversely, economics actors in the developing world often find themselves in the low-institutional- quality, high-discretion scenario where they face increased uncertainty over the application of regulatory rules yet have limited institutional channels for resolving their difficulties. The findings also contradict conventional assertions that, given unconstrained political leaders, investors should welcome greater independence for executive agents as a type of substitute for political constraints on leaders’ ex-post behavior.

Instead, the analyses suggest the exact opposite: discretion must accompany robust political institutions rather than substitute for them. Ironically, delegation to independent agents may be counterproductive in precisely those places where the literature expects it to help most.

To bolster conﬁdence in the dissertation’s theoretical framework, Chapter 5 tests observable implications of the theory’s proposed causal mechanisms. In developing an explanation for why the broader political context shapes investors’ response to bureaucratic discretion, the theory posits at least two mechanisms. High levels of political competition are expected to help reduce investors’ uncertainty about the regulatory environment both by making leaders more attentive to constituents’ concerns as well as by minimizing agents’ divergent behavior by sharing the task of monitoring across multiple actors.

Turning again to the multilevel dataset, I ﬁnd support for the observable implications of a link between political competition and more attentive political leaders; ﬁrms in politically-competitive regions display greater approval of local elected leaders and are more likely than their low-competition counterparts to respond to bureaucratic

9 discretion by seeking political influence, either through organized lobbying or informal government contacts. Firms in low-competition regions, on the other hand, are less likely to turn to political leaders if they perceive bureaucrats to have discretion; instead, they eschew formal government channels in favor of resolving disputes through alternative methods, such as paying bribes. Chapter 5 also contains a second series of tests, which provide indirect evidence supporting the claim that better monitoring in politically-competitive environments reduces firms’ uncertainty by diminishing agents’ errant behavior. The analyses demonstrate that, in high-competition regions, discretion is associated with high approval of administrators, fewer problems with bribery, and fewer disputes between firms and government agencies.

Chapter 6 concludes the dissertation with a summary of the empirical findings and ways in which this research contributes to the literature. Empirically, this dissertation identifies and demonstrates that the link between bureaucratic discretion and reduced investment depends upon the political environment. Using survey data from over Russian enterprises, I find that perceptions of bureaucratic discretion are negatively associated with firm managers’ willingness to invest; moreover, this effect is strongest in regions where the institutional environment discourages political competition. Theoretically, the dissertation calls attention to the importance of predictable policy application and emphasizes the political environment’s influence on how bureaucratic discretion affects economic actors’ expectations. In providing an explanation for why investors might associate bureaucratic discretion with greater uncertainty in some locations while not in others, this dissertation helps to explain conflicting scholarly accounts about the effects of greater bureaucracy independence.

Studies based upon the experiences of developed democracies have drawn their inferences from robust institutional environments with vibrant political competition; in

10 such environments, the surrounding context plays a critical role in reducing the uncertainty that investors associate with bureaucratic discretion. In marked contrast to developed democracies, regulatory oversight in many developing countries is typically associated with limited political competition and inaccessible political institutions.

In these locations, the political environment does little to alleviate businesses’ uncertainty about the way in which discretionary bureaucrats choose to implement policy, nor do it provide economic actors with many viable options to handle disputes should they arise.

11 CHAPTER 2

A THEORY OF BUREAUCRATIC DISCRETION AND

INVESTMENT BEHAVIOR

Introduction

Following a ﬂood of incoming petrodollars and decades of pent-up demand, Russia’s retail sector has experienced an investment boom over the last nine years. Hoping to capitalize on the growth in consumer spending, Swedish retail giant IKEA moved aggressively into Russia, opening eleven stores to the tune of $4 billion since 2000.

IKEA’s ﬁrst stores appeared in Moscow and St. Petersburg, but as consumer incomes began to rise, the company began subsequently to locate in the big cities of Russia’s regions, opening shopping centers in Kazan, Yekatirinburg, and Nizhniy Novgorod.

However, on June 23, 2009, IKEA’s rapid expansion across Russia came to an abrupt halt.

Despite seeing annual sales increases of nearly twenty percent each year for the ﬁrst decade since their arrival, IKEA headquarters announced a suspension of all further investment in the country, postponing plans for an additional thirty or so stores

(Anishyuk 2009a). In his statement about the investment freeze, Per Kaufman, then- head of IKEA operations in Russia, linked the decision to problems surrounding an

IKEA investment in Samara, a medium-sized city located in one of Russia’s southern regions along the Volga River. According to the company, investment would be

12 suspended pending the resolution of a conﬂict between IKEA and the Samara regional government.

What prompted IKEA’s move to halt investment? With the future of a mutually- beneﬁcial, multi-million-dollar project hanging in the balance, how did the Samara government provoke IKEA to reconsider its investment program? More generally, questions from this speciﬁc episode echo a core question in the political economy literature: How do some governments create the conditions that foster long-term investment?

Given its importance for an economy’s long-term performance, questions of where and why private investment accumulates address a topic at the forefront of public discussion in many developing countries. Having witnessed during the Asian ﬁnancial crisis of 1997 the speed with which mobile capital from global investors can evac- uate a country and paralyze markets, developing countries have increasingly made encouraging long-term, ﬁxed capital investments a pillar of their growth strategies

(Jensen 2006). Whether domestic or foreign in its origin, long-term investment lays the foundation for economic growth, which, in its own turn, opens new economic opportunities for people and can result in higher living standards. In this way, the factors that draw or deter investors to a particular area matter a great deal to those interested in the welfare gains associated with increased economic performance.

The Russian Federation provides substantial variation across regions in private investment. Over the last decade, private investment and economic development have spread in an uneven manner across Russia. Table 2.1 compares average annual investment per capita for the top and bottom ﬁfteen regions.

Whether normalized by population or reported in raw totals, the numbers show that private investment is nowhere near uniformly distributed across the regions, even when we discount the regions that draw heavy investment to develop their natural

13 Region Per Capita Regional Regional Share Regional Oil Investment Investment of Investment Production mil. rubles / 1000 mil. rubles % of national % of national Top 20 Kamchatka Oblast 126.72 47450 6.59% 0.00% Nenets AO 75.32 3256 0.45% 1.56% Yamalo-Nenets AO 72.99 37377 5.19% 10.60% Khanty-Mansisk AO 31.04 43895 6.09% 55.67% Sakhalin Oblast 30.76 17592 2.44% 0.79% Tyumen Oblast 27.47 89664 12.45% 66.45% Primorsky Krai 24.65 52173 7.24% 0.00% Magadan Oblast 20.11 4147 0.58% 0.00% Amur Oblast 15.39 14452 2.01% 0.00% Sakha Republic 11.47 11220 1.56% 0.09% Khabarovsk Krai 9.95 14634 2.03% 0.00% Komi Republic 7.02 7462 1.04% 2.46% Leningrad Oblast 5.53 9195 1.28% 0.00% Belgorod Oblast 5.51 8257 1.15% 0.00% Yaroslavl Oblast 5.18 7174 1.00% 0.00%

Bottom 20 Ulyanovsk Oblast 1.46 2048 0.28% 0.11% Buryat Republic 1.32 1321 0.18% 0.00% Republic of Kalmykia 1.32 397 0.06% 0.08% Pskov Oblast 1.31 1012 0.14% 0.00% Tula Oblast 1.25 2113 0.29% 0.00% Kurgan Oblast 1.21 1255 0.17% 0.00% Chita Oblast 1.20 1435 0.20% 0.00% Ivanovo Oblast 1.13 1323 0.18% 0.00% Republic of North Ossetia 1.09 749 0.10% 0.00% Bryansk Oblast 1.09 1517 0.21% 0.00% Ust-Ordyn Buryat AO 1.00 138 0.02% 0.00% Evenki AO 0.96 18 0.00% 0.01% Republic of Ingushetia 0.46 188 0.03% 0.04% Chechen Republic 0.37 382 0.05% 0.27% Tuva Republic 0.33 102 0.01% 0.00%

Averages of annual ﬁgures from 1995-2007; ﬁgures taken from various years of Rosstat publications. Oil production data for Tyumen Oblast includes production for Khanty-Mansisk (an autonomous district within Tyumen).

Table 2.1: Private Investment Varies Across Russia

14 resource wealth. Comparing the lowest of the top fifteen regions with the highest region of the bottom fifteen, the average total investment per capita in Yaroslavl is roughly 3.5 times higher than in Ulyanovsk.1 Figure 2.1 presents a graphical summary of how investment compares between the two groups. Even excluding oil-producing regions, thirty percent of Russia’s average annual private investment comes from the regions in the top half of Table 2.1, as compared to barely three percent located in the bottom group.2 Plotting the regional averages for private fixed capital investment per capita from 1995 to 2007, Figure 2.2 provides the full range of variation across the regions.

Undoubtedly, a strictly economic logic explains a part of the variation that we see. By virtue of their natural resource abundance (e.g., Tyumen Oblast in Western

Siberia, and Sakhalin in the Far East) or their colossal market size (Moscow, and to a lesser degree St. Petersburg), several regions draw investment to their jurisdictions in familiar and unsurprising ways. Other patterns, however, do not explain themselves so easily. Political economy provides a useful set of tools for thinking about the varied political and administrative factors outside economic geography that contribute to increasingly heterogeneous investment climates in Russia’s regions.

1The unequal spread of investment is likely a primary driver of the mounting economic disparities across the Federation. While Muscovites live comparably to citizens of the Czech Republic or Malta, life in the poorest regions resembles more closely Mongolia, Guatemala, or Tajikistan (UNDP 2007).

2Within the last few years, the financial crisis of 2008-2009 stirred worries that dissatisfaction in the economically-depressed regions might reach destabilizing levels. In June 2009, leaders received a small taste of how quickly dissatisfaction with economic underdevelopment can manifest itself; after unpaid workers from a small town blocked a major highway in the Leningrad Oblast, dissatisfied citizens in the Altai Krai and Kurgan followed suit shortly thereafter. With the direct intervention of high government officials and concerted state efforts, political leaders were able to mollify protesters, but worries remain that such protests could someday erupt on a larger scale (“In the Trap of Populism” 2009, “Wage Arrears Cause Protests” 2009). To avoid exactly this type of civil disruption and any chance of more violent protests, Russian federal officials instructed that regional leaders in poorer regions take extra measures to improve economic conditions (Von Twickel 2008). Yet, despite internal and external pressures to cultivate economic growth, major progress eludes many regions that lag behind in attracting investment.

15 Comparing Private Investment across Russia's Regions, Top and Bottom Quartiles

20 Percent of Total Private Investment

0 Top 20, Top 20 Excluding Oil− Bottom 20 Regions Rich Regions Regions

Averages of annual ﬁgures from 1995-2007; ﬁgures taken from various years of Rosstat publications.

Figure 2.1: Private Investment Spread Unevenly Across Russia, Even Excluding Oil- Rich Regions

16 Regional Private Investment in the Russian Federation Per Capita, Averaged from 1995−2007

Tuva Republic Chechen Republic Republic of Ingushetia Evenki AO Ust−Ordyn Buryat AO Republic of North Ossetia Bryansk Oblast Ivanovo Oblast Chita Oblast Kurgan Oblast Tula Oblast Pskov Oblast Republic of Kalmykia Buryat Republic Ulyanovsk Oblast Mari El Republic Altai Krai Agin−Buryat AO Kirov Oblast Republic of Adygea Altai Republic Vladimir Oblast Kabardino−Balkar Republic Smolensk Oblast Republic of Dagestan Karachay−Cherkess Republic Penza Oblast Kaluga Oblast Chukotka AO Republic of Mordovia Udmurt Republic Republic of Khakassia Voronezh Oblast Ryazan Oblast Orenburg Oblast Oryol Oblast Novosibirsk Oblast Tambov Oblast Kostroma Oblast Republic of Karelia Saratov Oblast Novgorod Oblast Kursk Oblast Rostov Oblast Republic of Bashkortostan Omsk Oblast Tver Oblast Chuvash Republic Stavropol Krai Irkutsk Oblast Murmansk Oblast Volgograd Oblast Chelyabinsk Oblast Krasnoyarsk Krai Sverdlovsk Oblast St. Petersburg Nizhny Novgorod Oblast Lipetsk Oblast Samara Oblast Jewish Autonomous Oblast Republic of Tatarstan Kaliningrad Oblast Moscow Taimyr AO Perm Krai Moscow Oblast Kemerovo Oblast Koryak AO Astrakhan Oblast Arkhangelsk Oblast Krasnodar Krai Vologda Oblast Tomsk Oblast Yaroslavl Oblast Belgorod Oblast Leningrad Oblast Komi Republic Khabarovsk Krai Sakha Republic Amur Oblast Magadan Oblast Primorsky Krai Tyumen Oblast Sakhalin Oblast Khanty−Mansi AO

0 5 10 15 20 25 30

Private Fixed Capital Investment (millions of rubles per 1000 people)

Averages of annual ﬁgures from 1995-2007; ﬁgures taken from various years of Rosstat publications. The top three outlying observations dropped to reduce scaling distortions: Kamchakta Oblast (126.72), Nenets AO (75.32), Yamalo-Nenets AO (72.99).

Figure 2.2: Private Investment Varies Widely Across the Russian Federation

17 With policymakers setting policy and creating legislation at various levels of government, Russia’s federal structure allows political factors to vary meaningfully across regional territories. Most observers would describe the Russian Federation as a “centralized” federal system that is biased towards primary policy-making at the federal level, especially over the last decade (Remington 2007). Indeed, the federal government does produce a sizable amount of economic legislation that applies universally across and takes precedence within the eighty-ﬁve “Subjects of the Federation,” the formal and constitutionally-proper term for the regions.3 The sovereignty and prolif- eration of federal law does not mean that regions are powerless to shape their own economic environments. In Russia, regional governments still have the ability to pursue policies that inﬂuence directly the investment decisions of economic actors.

For example, regions have jurisdiction over important areas that shape the agenda for investment promotion, including responsibilities over property taxes and levies,

ﬁnancial aid and subsidies, natural resource use, licensing, and government procure- ment practices. Thus, despite the centralized bias of Russia’s federal system, regional governments retain meaningful administrative and legislative tools to set conditions that encourage business growth. Recognizing their scope of authority, we ought to expect regional governments to play an important role in explaining why investment climates vary across regions of the same country. Given pressures to achieve continuing economic growth and stability, what explains whether government actions foster or frighten private investment? What determines whether policy environments attract or deter long-term investment?

3This dissertation refers to Russia’s sub-national constituent units by the common term “regions,” rather than their (unwieldy) constitutional title: “Subjects of the Federation.” The total number of these regions have changed over the last few years, dropping from eighty-nine throughout the 1990’s to eighty-ﬁve at the time this dissertation is being written. Several smaller regions were folded into neighboring larger regions in a Kremlin-backed consolidation process that began in 2004. While speculations about further consolidations appear from time to time, no more oﬃcial plans have been announced at the time of this writing.

18 This dissertation identifies bureaucratic discretion – agents’ leeway in making subjective determinations about when and how certain regulations apply – as a primary source of uncertainty that deters long-term investors by undermining the predictability of firms’ regulatory environment. By highlighting bureaucrats’ role in determining how and when regulatory laws apply to businesses, the theory emphasizes a group of political actors that traditional explanations of investment have overlooked. Using a principal-agent framework, I argue that where regulatory bureaucrats exercise greater discretion in interpreting and applying laws, economic actors experience greater uncertainty about how those policies will be put into practice. This unpredictability deters investment by making it difficult for economic actors to predict how the regulatory environment will affect their projects’ future returns and costs.

Applying insights from the bureaucracy literature, the theory also explains why investors should associate bureaucratic discretion with greater uncertainty in some locations than in others. Political institutions that encourage political competition encourage leaders to be more attentive to the concerns of their constituents in the business community, thus providing economic actors with more responsive political channels to use should regulatory discretion create a problem. Moreover, by increasing the access that constituents have to policymakers and the policy process, politically- competitive environments help to spread monitoring costs across non-state actors and supporting institutions, thereby helping to stabilize investors’ expectations about the regulatory environment. In regions where uncompetitive politics limit this diﬀuse monitoring and reduce leaders’ need to attend to constituents, economic actors bear the full brunt of regulatory uncertainty: they have less overall information about discretionary bureaucrats’ behavior and fewer formal channels for handling disputes that may arise. Thus, in terms of ex-ante expectations as well as ex-post responses, the political context of bureaucratic discretion matters for investors.

19 The next section examines the literature on the political determinants of investment, identifying where adherence to dominant approaches has led us to overlook features of the policy environment that investors find very salient. After review- ing the literature, I present the full theoretical argument in two parts. In the first part, I explain why bureaucratic discretion deters investment by discussing in detail the causal logic that links bureaucrats’ policy discretion and economic actors’ uncertainty about the application of regulatory rules. To complete the argument, I examine countering views about the economic effects of bureaucratic discretion and develop an explanation for how features of the political environment affect economic actors’ response to bureaucratic discretion. Following this discussion, the concluding section provides a set of empirically-verifiable propositions suggested by the theory and outlines the dissertation’s empirical strategy for testing those hypotheses.

Political Explanations of Investment

Policy Content

Political explanations for outcomes such as investment accumulation or economic growth fall into two broad categories. The ﬁrst category emphasizes the importance of adopting policies that appeal to economic actors and set the stage for economic growth. Whether targeting tax-shy ﬁxed capital from multinationals or coaxing latent entrepreneurs to venture into business within a new market economy, policymakers must choose the “right” policies to entice asset holders into some desired economic activity, according to such arguments.

Examples of policy-centric approaches pervade political economy. For instance, scholars of foreign direct investment (FDI) have emphasized the role of favorable tax incentives and regulatory regimes for countries seeking to lure international investors

(Dunning 1988, Caves 1996, Li 2006, Hays & Peinhardt 2009). These kinds of policy 20 incentives often ﬁgure prominently in debates over whether or not globalization forces countries into competitive races with each other in order to provide desirable investment conditions, despite the short-term ﬁnancial costs these policy positions might entail for state budgets or social programs (Vogel & Kagan 2004, Easson 2004, Jensen

2006). In an entirely diﬀerent way, the literature on the so-called “Asian miracle” and state-led economic growth also adopts the position that choosing the “right” policies eventually produces favorable economic outcomes. Anchored by research from Robert

Wade (1990), Alice Amsden (1989), and Chalmers Johnson (1982), this literature champions heavy government intervention as the cause for the East Asian economies’ expansion throughout the twentieth century’ later decades. Finally, academic debates over what policy strategy or which reform package produced greater benefits dominated the post-communist literature for much of its first decade. Scholars argued over the pace of reform, whether economies benefited more from gradual reform

(Murrell 1993, Roland 2002) or from “shock therapy” and other rapid approaches

(Aslund 2002, Sachs 1994). Similarly, researchers spent considerable effort linking privatization policies to outcomes from aggregate economic performance to firm-level productivity and profitability (Brown, Earle & Telegdy 2006, Shleifer 1998, Guriev

& Megginson 2007).

Arguments grounded in the importance of policy content do not always make specific policy prescriptions or even an explicit link between specific policies and economic outcomes. Many arguments in this tradition build their theories upon an implicit assumption that some particular set of macroeconomic and regulatory policies attracts economic activity, although perhaps at the expense of certain subgroups in society. Written in the 1990’s era of big reforms across the globe, influential works on the politics of economic reform studied how governments negotiated painful reforms in the midst of organized political opposition (Williamson 1994, Przeworski

21 1991, Haggard & Kaufman 1995). Although written first and foremost to call our attention to dynamics wrought by domestic politics, these arguments contain a basic premise that the reforms needed for countries looking to develop economically have the same basic components everywhere: anti-inflationary measures predicated upon government austerity, liberalization of capital, consolidation of economic property into private hands versus state ownership, and the elimination of tariffs to allow freer access for global markets.

Policy Stability

A second set of explanations for investment-related outcomes separates policy content from the question of policy stability. According to this view, irrespective of whether specific policies enter investors’ calculus in a meaningful way, the predictability of the policy regime has a direct and dramatic impact on economic actors who are seeking to invest in long-term projects. Before tying up their asset in the present for the sake of expected profits at some time in the future, economic actors make calculations of expected returns and costs, both of which are affected by the economic policies that prevail over the period of investment. When investors perceive that laws and regulations may change frequently or without warning, uncertainty about the policy environment that will dictate those future payoffs makes them hesitant to part with their resources (Aizenman & Marion 1993, Alesina et al. 1996, Alesina & Perotti 1996).

According to Timothy Frye (2002b, 2010), uncertainty about future economic policies explains patterns of economic growth across the post-communist countries in the early years of their economic transition. When both pro-reform and communist- successor parties receive electoral support in roughly equal measure, policies are expected to vacillate from one extreme to the other as political turnover moves diametrically-opposed groups in and out of oﬃce. Consequently, national economic

22 performance suﬀers as uncertain economic conditions cause businesses to shy away from long-term investment. On the other hand, when one side tends to dominate, countries experience more economic growth since investors can plan for the future without worrying about large swings in policy.

The large literature on credible commitments provides another prominent example of setting aside the question of which policies are “right” to focus on whether or not policies can be expected to stay the same. In fact, business-government interactions over policy have long served as an archetypal example of non-credible commitments.

In order to attract investors to their jurisdiction, politicians can promise legal environments that favor increased returns, such as secure property rights, desirable regulation, preferential treatment, etc. Politicians have incentives to do so because capital investment can cultivate additional tax revenue streams for politicians and beneﬁt society by creating economic growth and increased employment. Of course, political pressures can also push politicians to act in ways that cut into businesses’ proﬁts as well: crises may prompt increased government intervention in markets, voter demands for redistribution can lead politicians to raise tax rates, and rising unemployment can spark calls for labor protections at employers’ expense. The literature asserts that, recognizing that politicians’ short-term and long-term incentives often diverge, investors doubt the credibility of today’s pro-business promises (however sincerely made) since they know that politicians may have strong temptations to reverse them tomorrow (Levy & Spiller 1994). If an unconstrained government can defect easily from ex-ante agreements in order to pursue ex-post incentives, businesses will forgo costly investments since they perceive any policy commitments to be non-credible.

Starting with seminal works by Vernon (1971) and Kydland and Prescott (1977), credible commitment arguments have enjoyed considerable popularity for explaining

23 economic outcomes. North and Weingast (1989) argue that formal institutional checks on the English Crown after the Glorious Revolution were the commitment mechanism that spurred the capital formation Britain needed for the Industrial Revolution. Be- cause ﬁxed-capital investments made by multinational ﬁrms become vulnerable to expropriation once located in a host country, credible commitment explanations play a central role in the literature on the political determinants of FDI (see Jensen 2006, Li

& Resnick 2003). Using surveys from Russian firms, Frye (2004) shows that when they perceive that state commitments to uphold legal rights are non-credible, firms invest less. Within authoritarian regimes, where commitments’ credibility are un- dermined by the lack of checks on extremely powerful leaders, scholars have located partial solutions in the presence of representative legislatures (Wright 2008) or in- stitutionalized party structures (Gehlbach & Keefer 2008). Investors even benefit from unintended credibility mechanisms; as Nooruddin’s (2011) study of economic growth volatility demonstrates, the politics of compromise engendered by coalition and minority governments lend stability to the investment environment by providing

“credible constraints” on unexpected policy changes.

From this dissertation’s standpoint, the problem is that, although they both have much to offer, neither approach fits well with the Russian case. In terms of policy content and policy stability, regional policy environments appear to have little heterogeneity that could produce the patterns we observe in private investment across the Russian Federation. For example, it would be difficult to imagine that the details of economic policies have no influence on investors’ decisions within Russia, but the low regional variation in terms of legal requirements or special policy incentives for business limits this explanation’s usefulness.4 Investment experts within Russia

4In a few notable cases, regional governments have been able to use policy tools to woo very large investors. For example, the Kaluga Oblast attracted a multimillion-dollar Volkswagen assembly

24 indicate that most regions have very similar laws on their books in terms of content

(author interview, 2 Jul 2009). Additional interviews with firm directors and business associations in Russia confirm that, looking across regions, investors often see very little difference among the economic policies that set regions’ regulatory agendas.

The policy stability camp makes a convincing claim that no matter how attractive the laws may be today, it becomes diﬃcult and costly to attract investors when those policies are seen as unstable and potentially ephemeral. Given that Russia’s second decade after communism’s collapse has been remarkably stable both politically and economically, we do not observe the conditions typically thought to engender policy volatility. For most of the last decade, Vladimir Putin’s continued dominance of the political scene has translated into commanding and easily-anticipated electoral vic- tories in the regions for his supporters in the United Russia political party. Until the worldwide drop in oil prices in August 2008, consumers, investors, and oﬃcials in

Russia showed little indication of worry that the extremely favorable economic conditions would ever change. Reform advocates complain that this complacency has led to stagnation and inertia in policy-making. Instead of seizing a golden opportunity, they claim, regional and national leaders did very little but ride the wave of petrodollars that was washing over Russia (“Diversiﬁed Economy Not for Rent” 2008).

The dominant approaches’ inability to explain this variation in investment suggests that our theories have overlooked an important factor that inﬂuences investors’ behavior in developing countries such as Russia. This presents an opportunity to identify yet another dimension of the policy environment that has a signiﬁcant impact on whether or not locations attract or deter investment.

plant in 2007 by subsidizing the construction of their production site. Such examples, however, are rare and represent a very small share of overall private investment in Russia.

25 Policy Application

Explanations that privilege either policy content or stability contribute to our understanding of investment, but restricting our analytical focus to a choice of one or the other assumes improbably that all salient political uncertainties have been resolved for investors once policies are both “right” and robust to political reversals. I argue that to understand investors’ response to regulatory environments, we should consider investors’ concerns about policy application, or the deliberate actions and decisions that put authorized policies into effect. Although it receives little scholarly attention, policy application is critical to investors because, no matter how attractive or stable they may be, laws on the books do not affect firms’ bottom lines until they become applied legal practices. Thus, scholarly preoccupation with policy content and stability loses sight of the fundamental reason why entrepreneurs care about policies: at some point, those rules will be applied to their activities and influence how they conduct their business.

A brief discussion about investors and their motivations should help clarify why they would be so concerned about policy application. The term “investor” encom- passes a varied group of foreign and domestic actors with economic resources that could potentially be plowed into entrepreneurial ventures: firms, banks, venture capital funds, and private individuals.5 Whatever their organizational differences, investors are assumed to want to maximize their profits.6 Investors can pursue future

5In the Russian economy, the bulk of investors deciding how to allocate their accumulated capital are domestic firms and entrepreneurs. According to Rosstat, the government statistical agency, the number of private domestic enterprises grew from 1.43 million in 1996 to 3.8 million one decade later (169% increase); during the same period, only 55,000 new foreign- or jointly-owned firms were registered (26% increase). Moreover, domestic firms lead foreign firms in their total share of investment. While foreign investment’s share of total investment did grow impressively from 3% to 19% from 1996-2006, domestic private investment outpaced this growth, increasing from 16% to 49% of total investment.

6Although treated together here, international and domestic ﬁrms have distinct diﬀerences that

26 gains by expanding operations, opening new product lines, conducting research and development, buying new machinery, etc., or they can forgo investment and use that capital for current consumption. By their nature, these long-term investment projects require that investors pay extensive costs initially and then wait for returns to mate- rialize several time periods in the future.

When assessing a project’s potential profitability, economic actors try to account for various factors that will influence their business operations, including government policies that will affect marginal costs and gains of the proposed activity. Certainly, investors prefer policies that reduce costs or open up new revenue sources, but because investors must make costly decisions based upon forecasts about future conditions, the predictability of policy is just as important – perhaps even more so – for investors’ calculations. As long as they are predictable, undesirable policies can be handled by investment strategies that account for the associated challenges. Empirical studies showing that investment responds negatively to volatility and the threat of unpredictable changes underscore the point: whatever the policy content, investors want predictable investment environments and avoid locations characterized by high policy uncertainty.7

Acknowledging the importance of policy application provides us with a more complete picture of how policy environments shape investment by reminding us that investment depends upon more than just which regulations prevail and whether they are expected to change. Proﬁt-motivated ﬁrms must pay close attention to the actual

should be noted. As outsiders to a host country, international firms typically suffer from greater informational disadvantages in regards to non-market factors, such as regulatory regimes, than do their local, domestic counterparts. On the other hand, domestic firms do not usually have the same broad range of alternative options (such as exit) that are open to multinational corporations, leaving local firms more exposed to the effects of an desirable policy environment. In this light, theorizing more about how these differences affect firms’ responses to bureaucratic uncertainty suggests itself as a fruitful avenue for further research.

7See Aizenman and Marion (1993, 1999), Bechtel and F¨uss(2008), and Nooruddin (2011).

27 application of laws in order to predict how their business interests will be aﬀected

by government regulations. For investors deciding whether or not to commit their

assets to a long-term project, uncertainty about policy application raises the risk that

unanticipated losses will erode their proﬁts. When regulatory bureaucrats interpret

and execute policies predictably, investors can treat that anticipated behavior as a

known parameter in their cost-beneﬁt calculations. On the other hand, when reg-

ulatory guidelines are applied inconsistently, economic actors invest less in order to

protect their assets from unpredictable regulatory risks.

The literature’s lack of attention to policy application corresponds to a similar

myopia about which government actors matter for economic development. In the ex-

isting literature, tensions between politicians and investors take center stage. Politi-

cians value investment because it opens up new streams of tax revenue that they can

use to provide both public and private goods for their constituents. Moreover, invest-

ment has the added bonus of bringing jobs and economic expansion, which often help

reduce demands on political leaders.8 For politicians, the trick is ﬁnding a way to encourage business activity and attract investment without jeopardizing any of their core political interests in the process.

Despite the usefulness of studying politician-investor dynamics, the predominant focus on these two groups all too often leads us to overlook the exact government representatives with whom businesses interact most frequently: bureaucratic agents.

8In the case of the Russian regions, governors are no longer elected directly by the public since 2005, but instead nominated by the country’s president, then ratiﬁed by the regional legislature. While this makes the situation slightly more idiosyncratic, basic motivations for increasing investment remain very similar given the federal government’s intense interest in defusing economic tensions and governors’ accountability to the Kremlin for handling economic development (Borisov 2009). Even in the era of nominated governors, regional executives must still rely on state resources to provide for citizens’ demands, and the taxation of regional businesses remains high on the list of potential revenue sources for government spending.

28 Policy application is the purview of regulatory bureaucrats. Overlooking bureaucrats makes an implicit (and tenuous) assumption that no slippage exists between laws as drafted by politicians and those policies’ implementation on the ground. In contrast, a focus on policy application emphasizes that the economic impact of politicians’ decrees depends largely upon the way bureaucrats translate legislative writ into practice. Everywhere, developed and developing countries alike, authorized representatives pass judgment as to whether people’s performances meet certain standards or fall short. Executive agents such as police oﬃcers, license examiners, auditors, and inspectors all refer to guiding principles or policies, and then make a subjective decision about how the current circumstances map onto the criteria. Bureaucratic discretion – agents’ leeway in making those subjective determinations about how and when certain rules apply – lies at the heart of this dissertation’s argument for why some locations are able to foster conditions that attract business investment while others have not.

Bureaucratic Discretion and Investment: A Theoretical Frame- work

I argue that regulatory bureaucrats with more discretion create greater uncertainty over how policies – independent of content – will be applied. In turn, higher uncertainty about the application and interpretation of existing regulations deters investment. Depending upon the extent that they decentralize monitoring costs and open up avenues for dispute resolution, supporting institutions can either attenuate or exacerbate investors’ overall level of uncertainty concerning discretionary bureaucrats.

Thus, economic regulation’s political context plays an important role in shaping investors’ response to bureaucratic discretion, both in ﬁrms’ initial willingness to invest

29 as well as their additional strategies for coping with uncertain regulatory environments.

This section explicates the argument’s logic. The theoretical explanation begins by describing regulatory bureaucrats’ extensive involvement in economic affairs. Next, after a brief discussion of the principal-agent relationship between politicians and bureaucrats, I present the argument for how bureaucratic discretion shapes investment decisions by undermining the predictability of regulatory policies’ implementation, complete with an illustrative example from the Russian context. After presenting the basic argument, I consider how it departs from standard accounts of bureaucracy and economic development that draw extensively upon the experience of the developed world. In response, I refine the theory further to explain why bureaucratic discretion in some contexts creates more uncertainty for investors than in others. Lastly, to round out the theoretical treatment, I discuss the observable implications of the mechanisms behind the conditional argument and indicate how firms’ choice of strategies relates back to their political environment.

Bureaucracy and Regulation

Bureaucrats form the backbone of modern government, acting as executive agents for policymaking political leaders. Delegation to bureaucrats by lawmakers is ubiquitous because it is necessary; government leaders and politicians cannot hope to enforce and implement laws themselves without any assistance (Shipan 2004). Once policymakers have created legislation, regulatory bureaucrats and public administrators take over by interpreting, implementing, and enforcing those laws that make it onto the books.

Considering that every law passed by politicians must go through their hands before it goes into eﬀect, bureaucrats play a critical part in determining how laws govern society (Lipsky 1980).

30 As far as business activity in Russia is concerned, the scope of bureaucratic au-

thority in the regional economies covers an expansive range. Regional bureaucratic

agencies interpret and apply an array of statutes, including registration laws, zoning

ordinances, and safety regulations (Brown, Searle & Gehlbach 2009, Solomon 2008).

Given the importance of these policy areas for proﬁtability, the manner in which

regulatory bureaucrats carry out their duties matters greatly for investors in Russia.

For a sense of bureaucracy’s prevalence in business matters, a quick look at data

from cross-national enterprise surveys underscores the frequent contact between reg-

ulatory bureaucrats and ﬁrms in Russia and beyond. Figure 2.3 reﬂects the average

estimated time that ﬁrm management spends dealing with government oﬃcials over

the application and interpretation of laws.

First, the graph indicates that ﬁrm managers in Russia spend a signiﬁcant amount

of time dealing with authorities; on average, they spend ﬁve percent more time with

state oﬃcials than do their counterparts in the European Union countries of Germany,

Greece, and Portugal.9 Secondly, the graph shows that businesses in other countries also deal frequently with officials. At nearly five percent of management time devoted to dealing with officials, the averages for both the rest of the former Soviet Union republics (FSU) and East European countries are just a few percentage points behind the Russian average. Reporting the average number of inspections made by various state agencies per year, Figure 2.4 confirms that, both inside and outside of Russia,

ﬁrms undergo regular scrutiny by bureaucratic agents. Together, these ﬁgures emphasize that, whether they welcome it or not, business actors cannot help but come into contact with regulatory bureaucrats on a frequent basis.

9In substantive terms, a five percent increase works out to two additional hours per week spent of management time spent with dealing with government officials, assuming a forty-hour work week. Over the course of a year, that extra time amounts to more than two full weeks of time that Russian firms managers spend on average with officials that managers in the surveyed EU countries do not. Statistically, the difference between the two groups is significant as well (t = 13.25, p < 0.001).

31 Average Time Spent by Management Dealing with Officials about Application of Laws & Regulations 7

3 Percent of Management Time Percent 2

0 EE & Russia FSU Baltic States EU

Data from EBRD-World Bank Business Environment and Enterprise Performance Survey (BEEPS), 2005. The Baltic states are included in the Eastern European category rather than the Former Soviet Union (FSU) category; FSU also excludes Russia. The EU category uses responses from ﬁrms in Germany, Portugal, and Greece that were collected during a limited BEEPS wave conducted in 2004 for comparative purposes.

Figure 2.3: Firm Managers Spend a Signiﬁcant Amount of Time with Regulators

32 Average Number of Annual Regulatory Inspections

3.2 Russia FSU EE 2.8

2.4

2.0

1.6

Number of Inspections 1.2

0.8

0.4

0 Tax Fire & Bldg. Labor Sanitation Inspections Safety Inspections Inspections Inspections

Data from EBRD-World Bank Business Environment and Enterprise Performance Survey (BEEPS), 2005. The Baltic States are included in the Eastern European category rather than the Former Soviet Union (FSU) category; FSU also excludes Russia.

Figure 2.4: Businesses Subjected to Multiple State Inspections on a Regular Basis

33 As the ﬁgures suggest, economic activity attracts bureaucratic oversight. Regula- tory statutes exist for nearly all facets of business, stipulating how and where ﬁrms enter the market, restricting the types of goods and services available for sale, and even specifying the processes that can be used to produce and supply those products.

And for every individual statute, a regulatory agent decides whether ﬁrms are in compliance or not. In large part, the impact of laws and regulations depend upon those bureaucratic judgments, meaning the greater the latitude with which bureaucrats can interpret and implement those policies, the higher the uncertainty for investors about how their business interests will be aﬀected now and in the future.

This observation, that variance in policy application can impede investment, goes virtually unnoticed in the literature’s prevailing explanations. By overlooking bureaucrats, extant arguments either make heroic assumptions that politicians’ ﬁnalized policies specify all the details that ﬁrms care about or else that bureaucrats implement all directives from their political bosses in a straightforward, predictable manner. A large literature on the principal-agent problems between politicians and bureaucrats suggests that such assumptions pose a grave oversight.

Principal-Agent Relationships, Discretion, & Uncertainty

Developed in economics for use in organizational and contracting theories, the classic principal-agent problem highlights speciﬁc tensions that plague nearly all hierarchical relationships. In this framework, principals are those in authority positions, while agents are their subordinates. The setup begins with certain premises: principals have goals that they cannot accomplish without relying on agents to act on their behalf. Monitoring agents is always costly and imperfect, so although principals have authority over their agents, agents can take hidden actions or make use of information

34 that remains hidden from the principal (Grossman & Hart 1983, Przeworski 2003).

This gives agents an informational advantage over the principal.

From the principal’s point of view, the crux of the problem is that agents can exploit the resulting informational asymmetry and proﬁt at their superiors’ expense.

Principals have two basic methods to combat agents’ defection: the first requires ex- post control through increased monitoring; the second uses ex-ante enticements to provide incentives for agents to stay in line with principals’ desires (Holmstrom 1979). By these means, principals can sometimes find arrangements that minimize agents’ abuse, but they invariably require a trade-off that puts the principal’s first-best outcome out of reach (Miller 2005). Commonplace examples of principal-agent dynamics include employers seeking to maximize employees’ efforts (Laffont & Tirole 1988), insurers trying to minimize the risky behavior of their clients (Spence & Zeckhauser 1971), or chief executives acting on behalf of their constituencies (Downs & Rocke 1994).

Beginning with two papers by Barry Weingast (1984; also Weingast and Moran

1983), political scientists have used the principal-agent framework fruitfully to analyze interactions between bureaucratic agents and their political principals. Although politicians are elected to engage directly in policy making, underneath them work countless cadres and subordinate state oﬃcials – bureaucrats – whose job it is to execute enacted policies. Scholars argue that, given this hierarchical structure, the politician-bureaucrat relationship should suﬀer from familiar principal-agent problems of information asymmetry and imperfect control, allowing agents to manipulate the policy process (Epstein & O’Halloran 1994).

True to the classic principal-agent setup, politicians have authority over bureaucratic agents, but bureaucrats enjoy signiﬁcant informational advantages. Accord- ingly, researchers have expressed doubts that bureaucrats’ actions consistently reﬂect the preferences of political leaders (Shipan 2004). For instance, bureaucrats work in

35 specialized agencies, often for many years, giving them specialized knowledge about issue areas that the average lawmaker lacks. If their preferences diverge from principals’ interest, agents can use their hidden information to misdirect policymakers, leading them to choose policies closer to agents’ ideal points. Moreover, since bureaucrats take written policies and implement them within society, agents have tremendous scope for hidden action. The impossibility of monitoring the myriad of daily bureaucratic decisions ensures that politicians have a hard time telling whether or not bureaucrats are acting in a manner that advances principals’ goals.10 This is especially true when, along with their speciﬁc duties, principals grant to their agents large amounts of discretion.

In the regulatory context, bureaucratic discretion in interpreting and applying laws hinges upon the specificity with which legislative statutes lay out their constituent procedures and practices (Huber & Shipan 2002). Lawmakers limit agents’ discretion when directives set specific parameters on the manner in which a particular law should be enforced. On the spectrum’s other end, vaguely-worded policies and broad mandates open the door for regulatory bureaucrats to exercise significant discretion over how much, when, and whom to regulate. For their part, government leaders may concede more leeway for bureaucratic decision-making, even when they might prefer otherwise, because bureaucrats sometimes need discretion in order to execute their duties in complex contingencies (Bawn 1997, Huber & Shipan 2002). In fact, principals may see discretion as an incentive for bureaucrats to acquire specialized

10This point highlights a key distinction between my argument and established commitment explanations. In the traditional argument, commitments lack credibility when politicians cannot guarantee that policies will remain unchanged after economic actors have made investments that are costly to undo. Once regulatory agents enter the analysis as a separate group of actors, however, we can and should consider a separate possibility: rather than the threat of future political reversals, some commitments may fail to attract investment because of politicians’ immediate inability to commit on behalf of their bureaucratic agents. That agents may undermine their principals’ commitments to other actors is a novel and useful insight that remains distinct from conventional credibility explanations.

36 expertise that will ultimately work to politicians’ beneﬁt (Gailmard & Patty 2007).

In such cases, agents’ discretion has potential value for their principals.

Granting discretion can help principals to capitalize on bureaucratic expertise, but the danger exists that bureaucrats will use their discretionary privileges to cir- cumvent directives or undermine principals’ goals (Heclo 1977, Wilson 1989, Epstein

& O’Halloran 1994, Huber & Shipan 2002). Recall that the principal-agent problem hinges upon two main factors: diﬀering preferences between agent and principal, and the agent’s ability to take hidden action or exploit hidden information. Discretion does not necessarily create agents’ informational advantages, but it does allow agents greater opportunity to put those advantages to use.

Bureaucrats, including those in Russia, make regular use of any discretion granted to them. Researchers have noted that vague laws with extensive delegation to executive agencies are common in Russian legislation (Solomon 2008). Remington (2006) makes the connection between such laws and uncertain outcomes: “Many legislative acts in Russia have been so vague that the bureaucracy has been able to eviscerate, reinterpret, or ignore them freely” (p. 278). Indeed, cross-regional studies of regulatory reforms indicate that regional bureaucrats guard this autonomy; ﬁrm surveys provide evidence that regulators in many regions have resisted enforcing federal reforms that would standardize small business regulations (Yakolev & Zhuravskaya 2008).

From bureaucrats’ standpoint, discretion is desirable because it gives agents the ability to carry out their functions as they see ﬁt with few external limitations.11 By easing constraints on agents’ behavior, however, greater discretion for bureaucratic decisions may exacerbate the existing problems surrounding agents’ asymmetrical informational advantages.

11See Heclo (1977) and Carpenter (2001) for examples in the American context of agencies ﬁghting against legislative and administrative reforms that would curb their autonomy; Oleinik (2008) provides more Russian examples.

37 As a thought experiment, imagine the impracticality of completely removing bureaucrats’ decision-making discretion. For any given statute, lawmakers would need to codify detailed procedures for every contingency and specify the appropriate behavior for every possible circumstance – such a task lies well beyond lawmakers’ constraints on time, energy, and information. Just as monitoring costs prevent principals from knowing fully about agents’ actions, the prohibitive costs of entirely eliminating bureaucratic discretion mean that political principals must tolerate some discretion for their agents, despite the possibility that bureaucrats may use it to move policy outcomes away from principals’ ideal (Moe 1990, Epstein & O’Halloran 1994).

Why should principal-agent dynamics and bureaucratic discretion matter to investors? Scholars rarely consider the consequences of principal-agent dynamics for parties outside the hierarchical relationship, but leaders imperfect control over their agents actually holds important implications for entrepreneurial activity. The arguments above suggest strongly that, contrary to implicit assumptions in the investment literature, politicians do not have perfect control over what bureaucrats do. Thus, the very same information asymmetry that troubles the principal-agent relationship impedes investors’ ability to forecast regulatory conditions. Increased discretion widens the range of potential policy outcomes that bureaucrats may produce, making the policy environment less predictable. This creates situations where investors may know a particular policy’s content, but given a high degree of bureaucratic discretion, they can still have relatively little information about how bureaucrats may choose to interpret and implement that policy, either today or in the future.

An Illustrative Example from IKEA

Returning to IKEA’s struggles in Russia helps illustrate the kind of subjective and arbitrary policy application that impedes investment. In its decade in Russia, the fur- niture chain has had several high-proﬁle disputes with local and regional authorities. 38 Inspectors from the Emergency Situations Ministry closed IKEA’s Nizhniy Novgorod

outlet for the 2006 holiday season after slamming the company with an alleged 800-

plus ﬁre-safety violations. Two years previously, oﬃcials in Moscow halted an outlet’s

grand reopening just minutes before the ceremony’s start, claiming that the parking

lot’s location over a high-pressure gas pipe posed a public danger (“IKEA Halts New

Investment in Russia” 2009). The ﬁnal straw came with the company’s diﬃculty

in getting ﬁnal permission to open recent outlet in the city of Samara. Nearly two

years after its completion in 2007, the 1.4 million-square-foot IKEA mall in Samara –

including restaurants, skating rink, and nearly 200 shops – sat empty. The reason for

this costly hold-up: citing “deﬁciencies” in the IKEA construction, oﬃcials demanded

that the store be able to withstand hurricane-force winds, even though the Samara

region has no history of extreme weather (Anishyuk 2009a).

The uncertainties of dealing with bureaucratic decision-making were dramatic enough to halt IKEA’s rapid expansion in Russia for several months. As mentioned in the introduction, in June 2009, the home-furnishings company announced plans to suspend all future projects in Russia, with company representatives stating that no new investments would be considered until after the Samara store opened (Anishyuk

2009b). According to country manager, Per Kaufman, the drastic move was inspired by diﬃculties in coping with “the unpredictability of the administrative processes in some regions” (Bush 2009, italics added).

The IKEA example highlights the uncertainty created by bureaucrats exercising their discretionary authority. Undoubtedly, the inspectors’ wild-sounding claims have grounding in genuine laws; violation citations reference specific articles written in existing fire and building codes. On the other hand, few observers believe that officials were not applying those statutes to IKEA in a highly arbitrary and idiosyncratic manner. If, in fact, IKEA’s Samara outlet had undergone fitting in order to withstand

39 potential hurricanes, it might very well have been the only such structure in the whole region ever to meet that criterion. It is exactly this type of conspicuous subjectivity in applying laws that can result in economic loss and diminished investment. While the company spent the several years seeking the necessary permissions to open its Samara outlet, the project’s costs doubled to $253 million (Anishyuk 2009a). Combined with the lost revenues from an empty mall, the high opportunity costs of such massive overrun cut investors deeply.

Thanks to the scope and size of its investments, IKEA’s unhappy experiences may differ from those of an average firm in important ways, but they illustrate clearly the chilling effect of regulatory uncertainty on investment. For sure, IKEA’s plights make for eye-catching headlines, but on a daily basis, scores of similar situations for both domestic and foreign investors go unpublicized. If anything, with a plethora of attractive external investment alternatives and deep financial reserves, IKEA can likely deflect the costs of regulatory uncertainty better than small- or medium-size investors.

The bottom line is that, whether large or small, businesses respond to uncertainty about regulations’ application in a similar way. In locations where regulatory uncertainty is acute, market actors invest less because they cannot clearly determine the future costs or returns to their long-term investments.

Competing Views of Bureaucratic Discretion

The claim that bureaucratic discretion deters investment runs counter to several established arguments that see the economic gains from discretion as outweighing the potentially negative side-eﬀects. Proponents of discretion claim that it allows agencies to recruit and retain career-minded professionals, ostensibly improving the economic climate through higher-quality governance and fewer oﬃcials that use their authority irresponsibly (Rauch & Evans 1999). Arguments of this sort often cast bureaucracy

40 as the state’s “neutral competence” (Kaufman 1956), portraying bureaucratic discretion as a way to prevent elected oﬃcials from derailing eﬀective public policy in their pursuit of particularistic goals (Mashaw 1999). Thus, discretion is good because it allows bureaucrats to be technocratic experts instead of political stooges.

Other scholars have suggested that bureaucratic discretion can make the policy environment more predictable, rather than less. Anchored in the idea that autonomous bureaucracies insulate government programs from major political changes

(Carpenter 2001, Lewis 2003), one set of arguments portrays discretion as a source of continuity amidst government turnovers (Miller 2000). In this view, discretionary bureaucrats help investors by uncoupling policy application from the inherent changes associated with political competition.12 In a similar fashion, some have argued that commitment at the policy implementation stage ceases to be a problem if political leaders can “lock-in” policies by handing the management of policies to autonomous bureaucratic actors who have diﬀerent preferences (Miller 2005). Delegation-as- commitment arguments appear regularly in political economy studies of independent central bankers (Rogoﬀ 1985, Cukierman, Web & Neyapti 1992), independent judiciaries (Levy & Spiller 1994), and independent regulatory agencies (Bertelli &

Whitford 2009). Again, granting discretion to bureaucratic agencies is argued to stabilize investors’ expectations by insulating policy application from the messiness of politics.13

12This logic informs the “Bureaucratic Quality” measure of the International Country Risk Guide (ICRG), a set of cross-national risk ratings that is prominent in the economic development literature. Studies use this measure both explicitly, beginning with Knack and Keefer (1995) as well as implicitly, since the World Bank uses ICRG data to build its own governance indicators (Kaufmann, Kraay & Mastruzzi 2009). Consider the ICRG’s reasoning: “Bureaucracy is another shock absorber that tends to minimize revisions of policy when governments change...[High- quality] bureaucracy has the strength and expertise to govern without drastic changes in policy or interruptions in government services. In these low-risk countries, the bureaucracy tends to be somewhat autonomous from political pressure” (ICRG 2010).

13It is important to note that using the term “discretion” involves some slippage. In addition to

41 The main problem with explanations linking bureaucratic discretion to beneﬁcial economic outcomes is sample bias. Like the institutional design literature, such theories rely heavily on the experience of developed countries (Levitsky & Murillo 2009).14

This over-representation of developed democracies threatens to skew our understanding of how investors react to policy environments. The beneﬁts of bureaucratic discretion may have some plausibility in countries with robust political and administrative institutions, but those conditions are lacking in much of the world. Without evidence that bureaucracies in other institutional environments behave similarly, we should exercise extreme caution generalizing arguments about the economic beneﬁts of bureaucratic discretion to developing and transition economies.

Theoretically, the danger in borrowing too heavily from the experiences of developed democracies is that their well-functioning institutions may obscure relevant facets of the relationship between investment and bureaucratic discretion. For example, touting insulated bureaucracies as a solution to commitment problems glosses over the fact that delegation does not eliminate moral hazard so much as relocate it from politicians to bureaucrats. Bureaucratic control only aids investors as long as they can anticipate the outcome from bureaucrats’ involvement; otherwise, discretion merely increases uncertainty about the application and interpretation of laws at the hands of unelected oﬃcials. Similarly, reaping the beneﬁts of bureaucrats’ technical expertise presupposes technocrats of the “Weberian” sort, yet many governments

discretion, scholars across subﬁelds and disciplines have used a variety of terms — delegation, autonomy, independence, insulation — when describing a similar phenomenon. Despite the pro- liferation of terms and diﬀerent types of state agents to which these ideas have been applied from central bankers down to line-level bureaucrats there is a general logic to these arguments that rests upon bureaucratic actors independence from political interference. The core logic is this: if bureaucrats are allowed to do their job without inference, they will produce better and more stable outcomes.

14With its championing of bureaucratic autonomy as a key contributor to the economic success of developmental states in East Asia, the literature on state-led development stands out as an important exception (Johnson 1982, Evans 1995, Amsden 1989).

42 struggle to recruit and train such experts. Consequently, bureaucratic discretion backfires when agencies do not possess sufficient human or financial resources to implement policies effectively (Huber & McCarty 2004). Given their unique institutional environment, developed democracies’ successes in substituting bureaucratic control for political control are quite likely to be the exceptions rather than the rule.

In this regard, post-communist Russia is an excellent location for examining investors’ response to bureaucratic discretion. From a research design aspect, this dissertation has two useful methodological traits. First, by studying Russia, this research presents us with an opportunity to observe investors’ response to bureaucratic discretion in an institutional setting that is closer to the typical non-OECD country.15 Secondly, variation in the Russian Federation’s regional political environments provides us with sub-national variation in institutions. Russia’s federal structure also helps to mitigate some of the challenges posed by studying bureaucracy and regulation in a cross-national setting, where measurement problems and unobserved heterogeneity in countries’ legal or administrative factors are often diﬃcult to address.

In this regard, regions’ shared administrative legacies and the overarching sovereignty of Russia’s federal law help to hold constant many potentially confounding institutional and policy factors that might vary cross-nationally. To make full use of these opportunities, the following section expands upon the dissertation’s basic theory to develop an argument for why the problems associated with bureaucratic discretion should be particularly acute in certain political contexts.

15For the case that Russia’s stiﬂed political freedoms, weak institutions, social problems, and economic insecurity put it squarely in the company of other middle-income countries, see Shleifer and Treisman (2004) or Shleifer (2005).

43 Bureaucratic Discretion and Political Context

The claim that bureaucratic discretion deters investment rests upon two arguments.

First, studies of investment unambiguously point to uncertainty as a deterrent for investors. Second, the principal-agent framework implies that uncertainty about policy application should increase with the amount of discretion delegated to agents. Yet, as noted in the previous section, predictions that discretion deters investment disagree with scholarly accounts of bureaucratic discretion in the developed world. This incongruity suggests that we need further theorizing to understand why investors’ response to bureaucratic discretion might diﬀer across institutional settings.

Several plausible mechanisms could make uncertainty about bureaucratic discretion higher in developing economies. At first blush, differences in bureaucratic corruption present one possible explanation. Observers of developing countries regularly equate greater discretion with increased opportunities for bureaucrats to take or ex- tort bribes. Unfortunately, a mixed empirical record makes it difficult to draw con- clusions about whether corruption hurts or helps investment (Treisman 2007).16 In fact, several promising studies suggest that corruption’s predictability matters more to investors than its costs (Wei 1997, Campos, Lien & Pradhan 1998, Malesky &

Samphantharak 2008), implying that those who wish to use corruption to predict diﬀerences across the two groups must also supply a theoretical argument for why discretion should lead to corruption that is more predictable in developed countries than in developing countries.

Variation in state capacity represents another plausible alternative, positing that

16Theoretically, scholars have taken both sides of the debate: while some scholars see bribe payments as “unofficial” taxes that allocate resources inefficiently and crowd out beneficial economic institutions (Vishny & Shleifer 1993, Mauro 1995), others have linked corruption to additional investment through corruption’s ability to help firms bypass onerous regulatory demands or obtain other preferential treatment (Leff 1964, Stigler 1971, Slinko, Yakolev & Zhuravskaya 2005).

44 bureaucratic management produces unanticipated outcomes when agencies lack suf-

ﬁcient resources to carry out their assignments. Such an explanation highlights the tension in many locations between governments’ ambitious economic policies and the challenges of translating them into practice. In locations such as sub-Saharan Africa,

India, and the post-communist states of Eastern Europe, scholars have identiﬁed inadequate local cadres and tight resource constraints as main contributors to the inconsistent implementation of national policies (Varese 2001, Easter 2002, Kohli 2010).

In contrast to developed democracies, where a major concern is that zealous bureaucrats will use their discretion to realize extreme policy preferences (Bendor, Glazer &

Hammond 2001), the worry in many developing countries is that overwhelmed bureaucracies may use their discretion to reduce administrative burdens by adopting half- measures and cutting corners. If greater discretion widens the range of outcomes that bureaucrats may produce while attempting to manage burdensome obligations, the combination of resource-poor agencies and high discretion may present an especially unattractive option for economic actors who are seeking predictable environments.

This dissertation concentrates on a third explanation, that political institutions can facilitate additional oversight of government agents, thereby reducing investors’ uncertainty about bureaucrats’ discretionary behavior. The principal-agent literature shows that, although discretionary agents are hardly predictable automatons, politicians can still take costly actions to inﬂuence how bureaucrats interpret and apply delegated policies. In particular, politicians increase the likelihood that bureaucrats will implement policies in a predictable manner by monitoring agents closely, but the resource-intensive nature of monitoring often makes it unattractive to leaders. Con- veniently, lawmakers can keep bureaucrats in check without directly paying all the costs themselves by allowing interested constituents and other institutional actors to help in the monitoring process.

45 The literature on “ﬁre-alarm” oversight provides a speciﬁc example of how such

decentralized monitoring of bureaucrats can take place. In studies of American bu-

reaucracy, ﬁre-alarm oversight refers to a system of rules and administrative processes

that create opportunities for the public to examine administrative practices, question

agency decisions, and identify actions that violate legislative goals (McCubbins &

Schwartz 1984). Such arrangements allow politicians to reduce their own monitoring

and rely instead on self-interested constituents to report troubling bureaucratic be-

havior. In addition, by providing dissatisﬁed actors with clear, formalized channels

for seeking recourse, politicians can further shift enforcement costs oﬀ onto support-

ing institutions, such as courts (McNollgast 1987). Generalizing the logic of diﬀuse

monitoring beyond the speciﬁc example of ﬁre-alarms and the American Congress,

we see that broader political institutional arrangements wield this same mechanism

and often do so in a way that makes the regulatory environment more predictable for

investors.

In this study, I call attention to the beneﬁcial role of political competition in miti-

gating the eﬀects of uncertainty that businesses associate with bureaucratic discretion.

Institutions that encourage political competition make policy implementation more

predictable in at least two ways. First, competitive politics help to keep politicians

attentive, effectively raising the costs to principals associated with not monitoring their agents. Politicians worried about losing office have a pressing electoral incentive to avoid scandal and prevent political opposition from capitalizing on constituents’ dissatisfaction with inconsistent legal requirements or aggravating treatment by state officials. Likewise, electorally-sensitive leaders should respond more readily to constituents’ attempts to influence and clarify regulatory rules in reaction to disputes that arise from agents’ discretion. Second, by giving economic actors direct and indirect opportunities to influence regulatory policy or complain about administrative

46 practices, political competition helps to spread the costs of monitoring bureaucrats across actors in society. Similarly, politically-competitive environments often give rise to a wide array of institutional watchdogs, such as political parties and independent courts, that share in the monitoring process and can provide formal, predictable channels for settling disputes that arise. This improved monitoring reduces the chances that bureaucrats will engage in errant behavior and helps economic actors know better what to expect out of their high-discretion environment.

With their robust institutions and vibrant political competition, developed democracies are in the best position to foster an eﬀective, polity-wide realization of the bureaucracy literature’s ﬁre-alarm oversight. In developed democracies, the institutional environment actively encourages organized political opposition, free media, and politically-savvy interest groups to help monitor the behavior of state agents. More- over, the adversarial nature of politics in electorally-competitive systems also keeps politicians attentive in the event that problems do arise. It is hardly surprising, then, that studies of investment based upon developed democracies have a sanguine attitude about bureaucratic discretion. In such environments, the surrounding institutional context plays a critical role in reducing the uncertainty that investors associate with bureaucratic discretion.

In marked contrast to developed democracies, regulatory oversight in developing countries such as Russia is typically characterized by limited political competition and inaccessible political institutions. Given unstable and toothless institutions in many parts of the world (Levitsky & Murillo 2009), political and legal institutions in developing countries often cannot play the supporting role required to make bureaucratic discretion more predictable. For example, the diﬀuse monitoring of state agents cannot work well if institutions restrict citizens’ ability to participate in political life or shut organized societal groups out of the policy process. Furthermore,

47 without intense political competition, leaders face little electoral pressure to respond to constituents’ complaints. Thus, in addition to reducing the channels for the effective monitoring of discretionary bureaucrats, institutions that discourage political competition also weaken politicians’ incentives to respond to concerns when they are raised. In contrast to the dynamics within politically-competitive environments, low-competition settings do little to alleviate businesses’ uncertainty about the way in which discretionary bureaucrats choose to implement policy, nor do they provide economic actors with many viable options to handle disputes should they arise.

If correct, this argument predicts that delegation to independent agents will be counterproductive in precisely those regions of Russia where the delegation-as- commitment literature expects it to help most. Conventional commitment arguments suggest that political leaders’ moral hazard problem is high where politicians face few constraints on ex-post behavior (Miller 2000, Jensen 2006). This implies that delegation to independent agents should encourage firm investment by acting as an autonomous counterbalance to political leaders who lack organized political opposition or have little threat of losing office. In direct contrast, I argue that discretion is most problematic in regions where leaders are the least constrained by political competition because closed political processes and unaccountable government inhibit effective oversight and restrict economic actors’ access to the policy-making process.

This discussion helps us understand why bureaucratic discretion in some regulatory environments presents a particularly sharp problem for economic actors. The political institutional context of economic regulation aﬀects the level of uncertainty that bureaucratic discretion creates for investors. Businesses should more readily invest where political institutions spread monitoring costs across multiple non-government and institutional actors, and then allow those same parties to be involved in the policy process. Under such circumstances, investors have clearer expectations about what

48 regulators want and how agencies’ discretion will aﬀect their business interests. But, where political institutions restrict political participation or give leaders’ few incentives to respond to regulatory concerns, bureaucratic discretion remains a stumbling block for investors since they cannot rely on institutional constraints to enhance the predictability of the regulatory environment.

Firm Responses to Regulatory Uncertainty

To clarify, the theory does not predict that bureaucratic discretion leads to no investment. Holding other factors constant, an unpredictable regulatory environment raises the risk that investors will lose some or all of their assets, requiring higher marginal returns in order balance out these risks. As the risks associated with uncertainty rise, they crowd out investment below a particular rate of return, thus driving a wedge between investors’ optimal investment levels and the amount that they are willing to venture amidst heightened uncertainty. Instead of quitting the game completely, investors can re-optimize their strategies and adopt second-best alternatives. In many circumstances, that second-best alternative may well be finding a substitute location, such as a neighboring region or even another country with a more predictable regulatory regime. Some investment locations, however, have no substitute: whatever Moscow’s regulatory uncertainties, no other consumer market in Russia even approaches its size, and for many firms, that market’s potential rewards dwarf nearly all risks. For investors committed to a particular location, then, “second-best” means expending extra firm resources on strategies to cope with the attending uncertainty created by bureaucratic discretion.

Firms respond to uncertainty by taking active measures to protect their assets.

Protective measures include hedging strategies that sink a portion of resources into lower-risk, but less profitable, projects. Investing in more liquid assets as opposed to investing in fixed assets represents one such hedging strategy (Baldwin & Meyer 49 1979). Investors that seek greater predictability can lobby politicians to clarify laws or carve out individual protections.17 Research suggests that firms often turn to business associations and inter-firm relationships in the uncertain environment of post-communist transition (Duvanova 2007, Pyle 2009). In seeking greater certainty about how policy application will affect their particular interests, firms may engage in questionable practices such as cultivating personal relationship with political leaders or bringing influential people into the business in order to leverage their connections.18

Bribery and attempts to buy preferential treatment from oﬃcials represent outright illegal, although not uncommon, strategy for dealing with regulatory uncertainties

(Satarov et al. 2007).

In claiming that investors’ political environment affects their response to bureaucratic discretion, the previous section argues that political institutions mitigate uncertainty by increasing governments’ responsiveness to economic actors’ concerns and by allowing more actors to take part in monitoring state agents. If these mechanisms are at work, we should expect to see the differences in political environments reflected in business actors’ strategies for handling bureaucratic discretion and greater regulatory uncertainty. In other words, firms choose their response to bureaucratic discretion based upon the political institutional resources at their disposal. Because business actors in high-competition regions should have greater influence on political leaders and easier access to the policymaking process, we should expect to see firms in such locations pursue strategies such as lobbying or cultivating government connections in

17According to an analysis of preferential laws in Russia’s regions by Slinko et al. (2005), ﬁrms that are able to secure special treatment in regional laws do, in fact, perform better relative to their counterparts.

18Former government officials often find work in the Russian private sector easily because demand among firms is high for connections and expertise that will help businesses avoid legal problems. Creating an extreme conflict of interest, some entrepreneurs have reportedly gone so far as to bring local bureaucrats into their business as partners in charge of handling regulatory affairs (Mereu 2008).

50 order to make use of those opportunities. Alternatively, investors in low-competition regions have fewer eﬀective political channels available for making their regulatory environment more predictable; given their limited options, ﬁrms in such locations are more likely to eschew political channels in favor of alternate mechanisms, such as bribery.

Whether successful or not, pursuing these strategies comes at a price for investors, in both terms of both money and time. Mitigating the dangers of an unpredictable investment climate can bring beneﬁts to the ﬁrm, but it also requires diverting resources away from other desirable uses, such as production or investment. So although these activities draw upon available resources to make at least some level of investment possible, they replace economic activity that could have been possible had not those resources been spent protecting their assets and trying to make the policy environment more predictable. In essence, these expenditures are deadweight losses that represent the costs of uncertainty about regulatory application.

Hypotheses

In this section, I present the empirical predictions that follow from the theoretical framework laid out in the preceding discussion. The theory generates several empirical implications about patterns of investment in Russia’s regions that we should observe at the micro level of ﬁrm behavior. I discuss each of these predictions in turn.

Starting with an assumption that firms seek profits and the observation that predictable policy regimes help investors to evaluate the viability of potential projects, the argument links regulators’ behavior and investors’ decisions. Specifically, the theoretical framework identifies bureaucratic discretion as a primary source of uncertainty about the rules that will determine long-term investors’ future returns. Thus, the baseline hypothesis of investors’ behavior appears as follows:

51 Hypothesis 1: The more that a ﬁrm perceives that regulatory bureaucrats interpret and apply laws in a highly discretionary manner, the less that ﬁrm will be willing to invest, ceteris paribus.

The theory also presents a rationale for why the political context of economic regulation can aﬀect investors’ uncertainty about policy application at the hands of discretionary bureaucrats. In particular, surrounding institutions that engender political competition reduce investors’ uncertainty by granting more access to the policy process and helping economic actors know better what to expect out of their high-discretion environment. Political environments that insulate political leaders from competitive electoral pressures and restrict oversight by non-government actors and deny investors that additional stability. Thus, the following empirical prediction that political context conditions the relationship between bureaucratic discretion and investment:

Hypothesis 2: Holding other factors constant, the relationship between bureaucratic discretion and investment should be more negative in regions characterized by low political competition than in regions with high political competition.

In addition to claims about the direct effect of bureaucratic discretion on firm investment, the theory’s conditional logic can generate additional predictions that relate to the argument’s causal mechanisms for why political competition reduces the effects of regulatory uncertainty. The claim that political environments differ in terms of their responsiveness and openness has observable implications, both for the way

firms perceive political actors and for the strategies that they adopt to mitigate uncertainties associated with bureaucratic discretion. For example, in high-competition regions, we would expect firms to pursue mitigating strategies that engage with the policy arena and seek to influence government leaders. In a similar fashion, the theory implies that improved monitoring within politically-competitive environments guards 52 against agents acting contrary to their political principals’ expressed goals. If true, we would expect to observe stark differences in the interactions between firms and discretionary bureaucrats, depending upon the political environment. In a later chapter (Chapter 5), I derive specific predictions to test the observable implications of the theory’s causal mechanisms.

Conclusion

This dissertation makes at least two theoretical contributions to our understanding of the political determinants of investment. First, it highlights the crucial role played by bureaucratic agents in determining whether policy environments attract or deter investment. The observation that delegation alters policy outcomes is the starting point for principal-agent theories of bureaucratic politics, yet the idea is novel within the literature on economic development and investment. The argument reminds us that businesses care about the actual application of policies, not just their adoption or durability. Considering investors’ desire for predictable application lends perspective on the reigning theoretical arguments. For example, in the wrong context, an exclusive focus on the content and stability threatens to create Potemkin policies that look attractive on paper but have little actual eﬀect on the practical conditions that matter most for economic actors. Likewise, the dissertation’s argument suggests a twist to the standard logic of credibility problems: some policy commitments are non-credible less because of potential policy reversals in the future, and more because of politicians’ inability to commit on behalf of their bureaucratic agents. Thus, developing a principal-agent framework enriches our understanding of the politics of investment by inserting bureaucrats into the government-business equation alongside politicians and ﬁrms.

53 This dissertation makes a second theoretical contribution in highlighting the importance of the institutional context in which discretion is granted to regulatory bureaucrats. By diﬀusing the costs of monitoring and giving leaders an incentive to respond to triggered ﬁre alarms, institutional environments that encourage political participation and competition help to reduce investors’ uncertainty about how discretionary bureaucrats will apply policy. This logic suggests that existing scholarship’s seemingly sanguine attitude towards bureaucratic discretion comes from focusing on the best-case scenario. In the developed democracies that inform a majority of studies, high-quality institutions attenuate investors’ uncertainty about what to expect from independent bureaucrats’ behavior. Yet, economic actors in much of the world

(Russia included) find themselves in a very different situation – high uncertainty over the application of regulatory rules coupled with political institutions that limit investors’ options for resolving their difficulties.

The rest of the dissertation seeks to test the argument’s empirical predictions with a mixed-methods approach and using different types of data. Immediately following this chapter, Chapter 3 begins investigating the theory’s micro-level implications. In quantitative analyses of data from enterprise surveys of Russian firms, I test the argument’s baseline prediction and find that perceived bureaucratic discretion has a strong negative association with firms’ plans to invest. I then use qualitative data from over forty field interviews in Russia to supply the context-specific detail about relevant actors and processes that go missing from the large-N study of firms’ behavior. These interview data collect responses from firm directors, policy experts, and business associations to provide insight into how regulatory bureaucrats influence the decision-making calculus of business actors in Russia. Chapter 4 investigates the hypothesis that regional political institutions affect firms’ investment response to bureaucratic discretion. In multilevel analyses of firm-level survey, I show that the

54 negative association between firm investment and perceptions of bureaucratic discretion is increasing in restrictiveness of the regional political environment. I continue the empirical investigation in Chapter 5 by examining whether empirical evidence supports the theory’s proposed causal mechanisms. I show that, depending upon whether they operate in a high- or low-competition region, firms that perceive bureaucratic discretion have differing views of their elected leaders and display different propensities to engage with the state through lobbying and the cultivation of government connections. In low-competition regions, bureaucratic discretion is associated with resolving disputes through alternate mechanisms, such as business associations and bribery. Moreover, I find that discretion within high-competition regions is associated with better-behaved bureaucrats, resulting in fewer problems with bribe-taking and fewer reported disputes with government agencies. Finally, Chapter 6 provides a summary of the implications of my empirical findings and concludes the dissertation.

55 CHAPTER 3

BUREAUCRATIC DISCRETION AND BUSINESS

INVESTMENT: EVIDENCE FROM RUSSIAN

ENTERPRISES

Introduction

Focused on the argument’s hypotheses about the perceptions and behavior of economic actors, this chapter employs a mixed-methods approach to investigating the theory’s logic at the microeconomic level of the firm. Section I uses data from a 2005 survey of 666 Russian enterprise managers to test the prediction that firms’ perceptions of bureaucratic discretion affect their decisions about investment. Following the statistical analyses, Section II presents qualitative evidence from field interviews with business elites and policy experts in Russia regarding the motivations, perceptions, and mechanisms behind investors’ response to the unpredictable application of regulatory laws. After the empirical analyses, the chapter then concludes with a brief summary of the findings.

56 Quantitative Analyses of Discretion & Investment

The empirical analysis begins by testing Hypothesis 1. This prediction encapsulates

the theoretical argument’s central behavioral insight: arbitrary and subjective behav-

ior on the part of regulatory bureaucrats creates risks that diminish economic actors’

incentives to invest. The hypothesis is phrased as follows:

Hypothesis 1: The more that a ﬁrm perceives that regulatory bureaucrats interpret and apply laws in a highly discretionary manner, the less that ﬁrm will be willing to invest, ceteris paribus.

To test this empirical prediction, I conduct statistical analyses using data from a survey of Russian enterprise managers conducted in 2005 by Timothy Frye (see

Frye 2006). As a data source, several features of the Frye survey make it attractive for this analysis. Because the survey was commissioned speciﬁcally for studying how ﬁrms respond to political factors within the Russian Federation, the survey instruments are sensitive to the conditions and practices that matter for entrepreneurs in the

Russian context. The survey also contains items on ﬁrm investment and, importantly,

finely-grained questions about managers’ perceptions of various specific institutional actors, including regulatory bureaucrats. Thus, the individual-level data on how firms perceive their regulatory environment make a natural fit for testing the argument’s microeconomic predictions about individual actors’ subjective investment decisions.

The survey was conducted by a well-respected Moscow polling ﬁrm, the Levada

Center, polling 666 ﬁrm managers in eleven diﬀerent regions.1 Pollsters interviewed

firms’ chief executive officers, chief financial officers, or chief legal officers – the vast

1For more detail on the strategy behind the stratiﬁed sampling design, see Frye (2006). Including at least one region from each of the seven federal districts, the survey sampled ﬁrms in the following cities: Ekaterinburg, Khabarovsk, Moscow, Nizhniy Novgorod, Novgorod, Omsk, Rostov, Smolensk, Tula, Voronezh, and Ufa.

57 majority of whom were male (79 percent) and, on average, middle-aged (average respondent age was forty-seven). The mean size of firms is 727, a number that is skewed upwards since half the firms in the sample reported having fewer than 125 workers. In regards to size and sector, the sample distribution of firms approximates the national population of enterprises.2 With only twelve percent majority state-owned and five percent foreign-owned, the modal firm is a domestic, private enterprise.3

Dependent Variable: Firm Investment

The dependent variable is investment by firms. Following Frye (2006), I operationalize investment using firm managers’ response to the following item: “Do you plan to make any large investment in the next twelve months for the development of your firm

(i.e., construction, reconstruction, capital renovation of the building or surroundings, equipment updates, etc.)?” 4 I collapse answers on the question’s original four-point scale (“yes”, “likely yes,” “likely no”, and “no”) into a dichotomous variable that gives a value of 1 for the two aﬃrmative answers and assigns a 0 to the remaining, negative answers.5 Firms in the sample lean towards no investment plans, with sixty

2A caveat that applies to this survey, as well as any other enterprise survey that only includes operating firms: the sample is truncated because it contains no data for potential firms who were deterred from initial investment or those short-timers who entered the market but then closed their doors. This survivor-bias implies that respondents are firms who have higher tolerance for risk or have managed to cope with existing barriers. Any bias of this type that might exist works against my hypotheses, making make it harder to find a negative relationship between bureaucratic discretion and firm investment.

3Appendix A provides descriptive statistics on the basic characteristics of the sample.

4The survey does contain an additional item asking managers about any fixed capital investment over the last two years. However, to be consistent with the argument’s emphasis on the importance of predictability to firms’ present calculations about future returns and costs, I choose the forward- looking item over the more retrospective alternative. Using this other variable in place of the dependent variable in this section’s analyses, however, causes no meaningful difference for the empirical results.

5The results do not depend on this coding choice; all the relationships hold with the original coding.

58 percent of respondents answering negatively. Table 3.1 provides descriptive statistics for responses on this variable and all others included in this section’s analyses.

Variable N Median Mean Std. Dev. Min/Max Firm Investment 645 0 0.40 0.49 0/1 Bureaucratic Discretion 576 2 1.96 0.89 1/4 Changes to Laws 656 4 4.01 1.15 1/5 High Tax Rates 663 4 4.09 1.10 1/5 Regional Administration 601 3 3.00 0.92 1/5 Regional Courts 565 3 3.22 0.83 1/5 Regional Governor 604 3 3.12 1.02 1/5 Access to Finance 629 4 3.43 1.49 1/5 Labor Shortages 662 4 3.71 1.35 1/5 Competitive Pressures 657 3 3.38 1.35 1/5 Privatized Firms 666 1 0.59 0.49 0/1 Annual Sales 609 1 0.65 0.63 -1/1 Firm Size 666 4.84 4.99 1.53 1.39/11.16 Private ﬁrms 666 1 0.88 0.33 0/1 Bureaucratic Corruption 523 1 0.55 0.50 0/1 Past Investment 660 1 0.55 0.49 0/1 Note: Survey data from Frye (2006).

Table 3.1: Summary Statistics for Variables from Frye (2006) Data

Independent Variables

To test the prediction that ﬁrms’ investment decisions correlate negatively with their perceptions of bureaucratic discretion, I operationalize the independent variable using an item about the independence of regional bureaucrats’ decision-making: “To what degree is independent decision-making, separate from other government bodies,

59 characteristic of bureaucrats, administrators, and various inspectors in your region?”6

Respondents’ answers range on a four-point scale that evaluates regional bureaucrats’ autonomy as characteristic “to a high degree,” “essentially to a high degree,” “to a lesser degree,” or “completely uncharacteristic.”7 For ease of interpretation, I in- vert the original scale so that higher scores reﬂect high perceptions of bureaucratic discretion.8

Contrary to explanations that posit a positive relationship between bureaucratic discretion and investment, my argument predicts that this variable should correlate negatively with firms’ plans to invest, and Table 3.2 shows that it does. A comparison of the column percentages in Table 3.2 reveals a relationship between perceptions of bureaucratic autonomy and firms’ plans to invest. Among respondents that see little independence for regional regulatory bureaucrats, forty-four percent indicate that they will invest in their firms soon. In contrast, among managers that identify regulatory bureaucrats as independent in their decision-making, just thirty-two percent of firms plan to invest (the Pearson chi-squared statistic for the paired observation is

χ2 = 5.954, p = 0.015). To ensure that this result is not spurious, I add a battery of control variables that have been identiﬁed in existing research.

6Because Russia’s regions have diﬀerent titles, such as oblast, territory, republic, and district, the item’s wording changes slightly to ﬁt the appropriate regional designation, but the character trait and actors in question do not change depending on respondent location.

7Because this key question relies upon respondents’ perceptions of bureaucratic discretion, I provide evidence in Appendix D supporting this measure’s internal and external validity. A extensive set of additional analyses show that the dissertation’s ﬁndings are robust to controlling for a host of outside factors that might shape respondents’ answer to the prompt, such as respondents’ political knowledge, optimism, experience interacting with the government, familiarity with the region, or sectoral/legal characteristics. Furthermore, empirical analyses demonstrate that, in aggregate, respondents’ perceptions correlate as expected with external measures of bureaucratic discretion and private investment at the region level.

8Recoding this measure into an dichotomous indicator of low versus high bureaucratic discretion creates no substantive change for the results.

60 Does ﬁrm have Do regional bureaucrats plans to invest make decisions independent in the next of other gov’t bodies? 12 months? No Yes Total

Yes 193 39 232 (43.96%) (31.71%) (41.28%)

No 246 84 330 (56.04%) (68.29%) (58.72%)

Total 439 123 562 (100%) (100%) (100%) Note: Survey data from Frye (2006). Column percentages in parentheses.

Table 3.2: Perceived Bureaucratic Discretion Associated with Less Investment

Controlling for certain regulatory-related variables is of particular interest in order to test my argument against established explanations as well as ensure that my measure of bureaucratic discretion does not tap other policy concerns beyond investors’ worries about how regulations are interpreted and applied. Other features of the policy environment may also drive investment, such as the actual policies of regulatory regimes or the frequency with which laws change. In terms of policy content, scholars have paid special attention to the idea that high tax rates may discourage investment (Dunning 1988, Li 2006, Jensen 2006). Accordingly, I include respondents’ evaluations about the degree to which high tax rates hinder their ﬁrm’s development.

Frequent legislative changes could deter investors with uncertainty about the regulatory environment much in the same way that bureaucratic discretion does over the application of those policies. To control for policy volatility’s potential relationship

61 to both bureaucratic discretion and ﬁrm investment decisions, I include managers’ response regarding the extent to which frequent changes in legislative and statutory acts are an obstacle to their ﬁrms. The answer scale for both variables takes values between 1 and 5, with 1 = “they do not create any obstacles,” and 5 = “they create very serious obstacles.” Conventional arguments would anticipate that both these variables should associate negatively with the dependent variable.

As a check for whether my measure merely reﬂects investors’ antipathy towards corruption, I include in one model a dummy indicator for company managers’ perceptions that regional bureaucrats are corrupt. This variable assigns a value of 1 to responses that characterize bureaucrats as corrupt, 0 otherwise. As noted earlier in the theory chapter, the literature disagrees on this variable’s expected sign, but if corruption truly represents the heart of investors’ concerns about discretion, then we should expect its inclusion into the model to overshadow the bureaucratic discretion variable.

Since models rely on respondents’ subjective perceptions of bureaucratic discretion, we might worry that such perceptions may be colored heavily by general attitudes towards the region or its government in a way that also correlates with firms’ disposition towards future investment. As a precaution, I include variables that measure separately firms’ assessments of the regional political institutions to help account for regional “halo effects” that might vary with both perceptions of bureaucratic discretion and plans to invest. These controls capture respondents ratings of the regional administration, regional arbitration courts, and governor; we would expect more favorable assessments of these institutions to correlate positively with firm investment.

I also add controls for key economic factors that might similarly influence firms’ decisions to invest. I control for difficulties in getting access to finance with a measure of

ﬁrms’ problems obtaining credit. High levels of economic competition may pressure

62 ﬁrms to invest in innovative processes and products when they might not otherwise want to do so; accordingly, I control for respondents’ views about competition as a problem for their business growth. Similarly, managers may refrain from investment projects that they think will be hobbled by a poor labor pool. Thus, I include a measure for managers’ view about the shortage of skilled labor as a hindrance to business.

In addition to the perception-based measures, I control also for firm-specific characteristics that might affect both firms’ investment plans as well as their regulatory experience. Declining sales might stall investment plans, so I include a control variable for whether firm sales have increased, decreased, or stayed the same over the past three years. Small firms might have less need or ability to invest; as such, I control for

firm size with the (logged) number of employees. Finally, to control for potentially higher investment among private and start-up enterprises, I include dummy variables for private firms (as opposed to state-owned) as well as a dummy for firms that were privatized following the collapse of communism.

I estimate the model predicting intentions to invest using logistic regression. To capture heterogeneity in any unobserved region-specific factors that may influence the relationship between firms’ perceptions of bureaucratic discretion and their investment plans, I estimate random-intercept, random-coefficient models that allow coefficient estimates on the intercept and bureaucratic discretion variable to vary with a region-specific error.9 Results from these analyses appear in Table 3.3.

9Results do not change substantively if models drop the random eﬀects and use standard errors that are clustered by region instead. Similarly, results hold using a standard logit model with (or without) Huber-White “robust” standard errors to correct for potential heteroskedasticity. Results available in Appendix A.)

63 Firm Investment dummy, 1 = ﬁrm plans to invest during coming year (1) (2) (3) (4)

Bureaucratic Discretion -0.458*** -0.612*** -0.573*** -0.606*** 1 = no discretion, 4 = high discretion (0.145) (0.161) (0.174) (0.191) Frequent Changes to Laws 0.101 0.066 0.139 1 = no obstacle, 5 = very serious obstacle (0.112) (0.118) (0.125) High Tax Rates -0.392*** -0.437*** -0.404*** 1 = no obstacle, 5 = very serious obstacle (0.123) (0.131) (0.147) Regional Administration 0.509*** 0.447** 0.477** 1 = poor job, 5 = excellent job (0.195) (0.210) (0.229) Regional Courts -0.150 -0.181 -0.041 1 = poor job, 5 = excellent job (0.147) (0.152) (0.173) Regional Governor 0.037 -0.324* -0.284 -0.437** 1 = poor job, 5 = excellent job (0.106) (0.184) (0.197) (0.222) Access to Finance -0.087 -0.034 -0.026 0.035 1 = no obstacle, 5 = very serious obstacle (0.067) (0.082) (0.086) (0.096) Labor Shortages -0.028 0.001 0.066 1 = no obstacle, 5 = very serious obstacle (0.086) (0.091) (0.103) Competitive Pressures 0.107 0.127 0.050 1 = no obstacle, 5 = very serious obstacle (0.085) (0.089) (0.101) Privatized Firm 0.043 0.044 0.209 dummy, 1 = privatized, former SOE (0.273) (0.282) (0.346) Annual Sales 0.476*** 0.442** 0.330* 0.258 -1 = decreasing, 1 = increasing (0.166) (0.182) (0.190) (0.217) Firm Size 0.335*** 0.298*** 0.284*** 0.337*** number of employees (logged) (0.073) (0.084) (0.088) (0.117) Private Firm 0.721** 0.694 0.542 dummy, 1 = private ownership (0.337) (0.439) (0.443) Bureaucratic Corruption -0.015 dummy, 1 = perceived as corrupt (0.254) Past Investment 2.086*** dummy, 1 = invested in last 3 yrs. (0.294) Constant -1.940*** -0.594 -0.023 -2.523 (0.674) (1.031) (1.151) (1.712) Dummies for Sector & Legal Form No No No Yes Log-likelihood -290.198 -240.182 -220.298 -201.094 AIC 600.396 514.363 476.595 472.188 No. of Cases 470 403 365 402 Note: Survey data from Frye (2006). Coefficients represent estimates from multilevel logistic regressions with with a random coefficient for the bureaucratic discretion variable and random intercepts at the region level; standard errors in parentheses. Out of space concerns, region random effects and sector/legal-form variables not reported. * p<0.10, ** p<0.05, *** p<0.01

Table 3.3: Discretion Associated with Lower Probability of Investment

64 Results

To demonstrate that estimated relationships hold across various model speciﬁcations,

Table 3.3 reports the results from multiple models. Model 1 estimates a baseline specification that includes only the independent variable and a few controls for the investment environment and firm-specific factors. Next, Model 2 tests the results’ robustness in a full set of controls, followed by Model 3, which examines discretion’s relationship with investment once we control for firm managers’ perceptions of bureaucratic corruption. Model 4 adds a control for recent past investment as well as dummies for firms’ sector and legal form in order to control for a wider range of firm- level characteristics that might determine firms’ contact with regulatory agencies as well as shape their investment plans.

From a substantive standpoint, the statistics in the first two rows display the most important empirical result: perceived decision-making autonomy for regulatory bureaucrats correlates with a lower probability that firms plan to make any large investment over the coming year. Whether in sparse models or exposed to a large number of possible confounding variables, the coefficient estimates on the bureaucratic discretion variable in all models are negative and statistically significant from zero at conventional levels. This leads us to reject the null hypothesis that perceived bureaucratic discretion has no relationship to firms’ investment plans. Substantively, these findings show that perceptions of increased bureaucratic discretion are associated with reduced incentives for firms to invest for future returns. After discussing the models’ basic insights with regards to the other variables, I return to the question of the relative size of this estimated effect.

Among the other variables directly related to the policy environment, ﬁrm problems with high tax rates are also a strong predictor of ﬁrm investment decisions.

Supporting arguments that stress the economic incentives of speciﬁc policy content,

65 the coeﬃcients on all tax rate variables are statistically signiﬁcant and negative, sug-

gesting that high tax rates provide disincentives to managers to sink capital into new

investments.10 Together, the strong performance of both the bureaucratic discretion and tax rate variables suggests that policy environments can affect investment via multiple channels and that investors’ concerns about content and application might reinforce each other. In contrast, neither frequent changes to laws nor bureaucratic corruption (Model 3) bear statistically significant relationships to firm investment in these data, countering the expectations of conventional explanations.

The models reveal a nuanced relationship between government institutions and the investment climate. In regards to economic courts, the analyses reveal no significant association between respondents’ assessments of regional economic courts and decisions to invest. This runs counter to expectations from other research that suggests that trust in judicial institutions fosters investment by bolstering asset holders’ conﬁdence in their property rights (Levy & Spiller 1994, Frye 2004, World Bank 2005).

In contrast, the probability of ﬁrms’ intent to invest has a signiﬁcant, positive association with expressed support for the regional administration in all models in which the variable appears. Futhermore, Models 2 and 4 suggest that, ceteris paribus, sup-

portive views of the regional governor and investment plans correlate negatively. Such

ﬁndings could arise if ﬁrms’ non-investment relates to frustrations that a poor-quality

regional bureaucracy is hampering the governor’s good policies.11

The economic variables have mixed success in predicting investment. Problems with ﬁnance, competitive pressures, labor shortages and status as a privatized ﬁrm

10Within studies of post-Soviet Russia, other recent ﬁndings in favor of policy incentives include studies of FDI (Iwasaki & Suganuma 2005) and growth among regional enterprises (Berkowitz & DeJong 2003).

11Additional analyses support this interpretation. Regressing the dependent variable on a measure for the diﬀerence between two ratings shows that higher ratings for the governors relative to the regional administration correlate negatively with ﬁrms’ investment plans (p = 0.06).

66 have no statistically significant relationship with investment in any model. Private ownership and increasing sales trends, on the other hand, do have significant, positive relationships with investment in earlier models, although uncertainty about this relationship grows as more controls are added. Of the main economic variables, only larger firm size is consistently estimated to have a significant, positive association with firms’ plans to invest, reflecting perhaps smaller firms’ difficulties with funding the expansion of their business operations.

To test the sensitivity of these findings to modeling assumptions and specifications, I conduct a number of robustness checks. Investigating whether the estimated relationship between perceived discretion and firms’ intent to invest is spurious due to unobserved sector- or firm-specific factors, Model 4 controls for economic sector or

ﬁrms’ legal form. Model 4 also adds a dichotomous indicator for past investment, out of concerns that past investment gives occasion for encounters with regional bureaucracy and aﬀects investment strategies for the future. The inclusion of these controls do not change the results meaningfully. Results are also robust to alternate modeling strategies, such as basic logistic regression, as well as to the inclusion of additional substantive controls. These additional robustness checks are reported in Appendix

Figure 3.1 shows the substantive effect of the estimates from Table 3.3. In the graph, dots represent the estimated change in a hypothetical firm’s predicted probability of investing that is associated with manipulating each variable by some chosen amount while holding all other variables at their median values. The lines around the dots indicate 95% confidence intervals. For the sake of space, I focus my comments on variables that have estimated effects distinguishable from zero.

Figure 3.1 indicates the large decrease in ﬁrms’ decisions to invest that is associated with greater bureaucratic discretion. Holding other variables at their median,

67 Bureaucratic Discretion ●

High Tax Rates ●

Regional Governor ●

Regional Courts ●

Changing Laws ●

Credit Problems ●

Corruption ●

Labor Shortages ●

Privatization ●

Competitive Pressures ●

Annual Sales ●

Firm Size ●

Private Firm ●

Regional Administration ●

−30 −20 −10 0 10 20 30

Percentage Change in Probability of Firm Investment

Firm investment model using survey data from Frye (2006). Dots represent first differences in predicted probabilities from manipulating the indicated variable while all other variables held at median values. Bureaucratic discretion moves its median (“ 2 = low degree of independence”) to the maximum ( 4 =“completely independent”). The trichotomous sales variable moves from “no change” to “increasing sales,” and the logged measure of firm size increases from the median (4.84) by one standard deviation to 6.37. By necessity, dummy variables (corruption, privatization, private ownership) move from 0 to 1. All other variables are five-point ordinal measures that move from their middle category (“moderate obstacle” or “neutral assessment”) to their maximum values. Lines represent 95% confidence intervals calculated via simulation in R.

Figure 3.1: Discretion Associated with Large Decreases in Probability of Investment (Estimated Eﬀects)

68 a firm manager who perceives that regulatory agents have complete independence in decision-making has a predicted probability of investing that is 26 percentage points lower than a identical respondent who reports that regional bureaucrats have a low degree of independence. Only the negative effect associated with high tax rates has a similarly sized effect, with a 21 percentage-point difference between firm managers that identify high tax rates as a moderate problem and those that see high tax rates as a very serious obstacle. At the 95% confidence level, two other variables to have a clear non-zero estimated effect. Going from 125 to 665 full-time employees, a one- standard-deviation increase from the median, corresponds with an estimated increase of 11 percentage points in the predicted probability of firm investment. Going from a neutral assessment of the regional administration to the view that the administration is doing a very good job is associated with an increase of 20 percentage points in the predicted probability of a firm’s intent to invest.

The implications of Figure 3.1 are twofold. First, discretion’s large estimated eﬀect suggests strongly that ignoring the impact of discretionary bureaucrats’ behavior on investment decisions blinds us to meaningful dynamics that shape the investment climate. Secondly, that bureaucratic discretion stands out even when controlling for high tax rates and vice versa tells us something about investors view deﬁning traits of the policy environment in relation to each other: concerns about policy application and implementation appear to run parallel, rather than in competition, to concerns about other dimensions of policy, such as content.

Summary of Quantitative Findings

The theoretical framework in Chapter 1 argues that business investors care just as much or more about how policies get interpreted and applied by regulatory agents as they do about the details of the policy content. This section has used firm-level survey data to test the prediction that firm perceptions of regulatory agents’ discretion 69 affects firms’ plans to invest in the immediate future. Via crosstabs and maximum likelihood models that control for region- and sector-specific effects, I demonstrate that bureaucratic discretion bears a robust, negative relationship firm investment.

Firm managers who believe regulatory bureaucrats to make decisions independently of other government bodies have a lower predicted probability of investing in ﬁxed capital assets in the coming year. Building upon these analyses, the next section uses qualitative evidence to cross-validate the quantitative results. Using data from ﬁeld interviews in Russia, I examine both the mechanisms and motivations that lead to the statistical relationships that we observe in the survey data.

Qualitative Evidence from Field Interviews

To test further the relationship between bureaucratic discretion and private investment, I interviewed ﬁrm directors, heads of business associations, policy advocates, and legal experts in Russia. Conducted during the summers of 2008 and 2009, these interviews support the claim that bureaucratic discretion deters investors by creating uncertainty about the application of regulatory rules. In particular, the qualitative evidence from these interviews complements this chapter’s statistical analyses by:

first, providing insight into the relative weight that investors in Russia attach to concerns about unpredictable policy application; second, elaborating the mechanisms by which bureaucratic discretion hinders investment; and third, showing that firms’ desire to avoid uncertain regulatory environments affects their overall investment strategy. Below, I highlight details from the interviews.

Before presenting the ﬁndings, it is useful to describe the interviews themselves.

Over the period of 2008-2009, I traveled to Russia on two separate six-week research trips to conduct interviews with business elites and informed experts about private

70 investment and regional regulatory environments. Interviews typically lasted from 40-

60 minutes and followed a semi-structured format that asked a series of open-ended questions about general obstacles to business development, ﬁrms’ process for making investment decisions, experiences with various regulatory and political institutions, and regional variation in the desirability of investment climates. Although the vast majority of interviewees represented organizations with operations in multiple regions, most interviews took place in or around Moscow at the organizations’ headquarters.

Additional locations included Kazan, Nizhniy Novgorod, Vladimir, and Zelenograd.

Out of forty interviews, two interviewees preferred to speak English; I conducted the rest of the interviews in Russian without an interpreter. For this analysis, I have translated all Russian-language responses in English. With the exception of two phone interviews and one written correspondence, all interviews involved face-to-face conversations. To preserve their conﬁdentiality and anonymity, I do not report the names of my interviewees or their organizations. For citation purposes, however, I refer to my personal interviews throughout this chapter according to an interview label which indicates the type of interviewees organization, plus an assigned interview number and date of conversation. For more information on recruitment and composition of the sample, please see Appendix A.

Regulatory Obstacles to Investment

A starting point for examining the theory’s micro-level claims is to investigate what business representatives and policy experts identify as the major obstacles to investment. To assess how concerns about regulatory uncertainty compare to other explanatory factors, I asked interviewees to list in order of importance the factors that inﬂuence ﬁrms’ decision over whether or not to invest in a given region. The results were remarkably consistent across interviews. The majority of responses placed

“administrative concerns” about the regulatory environment very high on the list, 71 second only to economic concerns about market demand and labor supply in the

region. One interviewee explained investors’ approach to investing in the regions:

“[Investors] look at business ﬁrst, then taxes after. Business ﬁrst, then they look

at: ‘Is regulation a barrier?’ And if the barrier is high enough, then maybe another

region looks attractive. But these are all considerations when you are there [in the

regions]” (consulting executive, interview 32-090723). Of course, concerns about the

regulatory environment are not the sole factors driving investment, but the interview

responses revealed that they do feature prominently in investors’ decision-making.

Given conventional stereotypes about businesses’ aversion to state oversight, we

might not be surprised that economic actors express misgivings about government

regulation, but examining which particular aspects of the regulatory environment

investors choose to discuss can help to adjudicate between rival hypotheses. What

speciﬁc challenges do regulatory environments present for investors in Russia? While

most ﬁrm managers mentioned that stability and predictability are important for

business, only one made a direct complaint that regulatory legislation changed too

frequently.12 In the other cases where respondents mentioned legislative turnover, comments typically provided examples of updates or alterations to existing laws that they welcomed as improvements over the previous set of governing rules. Judging from the lack of complaints about this particular issue, policy instability does not appear to be a pressing concern for investors in Russia.

In contrast to their silence about policy stability, business leaders and policy experts displayed greater interest in discussing policy content. My discussions showed business actors to have strong opinions about existing legislation in their sector as well as clear ideas about what kinds of laws would beneﬁt their business interests.

12In fact, this particular comment was made in the context of the respondent’s discussion of Russia’s tumultuous history of reforms and seemed less connected to current conditions (entrepreneur and head of national business association, interview 10-080828).

72 Yet, for all their legislative expertise, not one interviewee offered examples of specific policies motivating investors away from one location towards another. When asked directly about policy content’s influence on businesses’ decisions over where to invest, respondents downplayed its overall effect on regional investment patterns.13

One investment consultant asserted that, although important to investors, regulatory statutes themselves are rarely the decisive factor: “I don’t know of anyone who says: ‘Well, I’ll go to Nizhniy Novgorod because I’ll get one point oﬀ my tax rate”’

(interview 32-090723). Such a lukewarm endorsement suggests that other regulatory factors may have a larger impact on investors in Russia.

The dissertation’s theory argues that, whatever importance investors attach to policy content, the details of legal requirements only matter inasmuch as they become standard operating practices: investors’ fundamental concern is policy application. In their responses to open-ended questions about the regulatory environment, business actors and policy experts did, in fact, reveal that problems with the application of regulatory policy pose a major obstacle to investors. In the words of one NGO director: “We have lots of very good laws...The whole problem is implementation. We

[the Russian business community] do not have problems with legislation – we have problems with the inconsistency and selectivity of applying legislation” (interview

24-090713). The sentiment that laws’ content takes a backseat to implementation issues found strong support among business actors as well. Consider the following statement by a foreign executive with nearly a decade of experience in the Russian economy:

13This attitude may relate back to a perceived homogeneity in the economic laws that govern the regions. Discussions with firm directors and business associations suggested that, looking across regions, investors often see little meaningful difference among the economic policies that set regions’ regulatory agendas. Similarly, policy experts were doubtful that regions could adopt platforms that were different enough to distinguish themselves from rival regions (interview 35- 090629, interview 22-090702).

73 “On the face of it, there are bits of legislation that are ok. There are even bits of it that might be progressive...The big problem lies in the implementation and enforcement...There is a legal system, but it doesn’t act exactly the way you’d expect one to work in the States or in Canada or in Western Europe. It doesn’t have that level of reliability because it’s poorly drafted, diﬃcult to implement and hard to enforce. Actually, it’s easy to get things enforced around here; it’s just that what they’re enforcing may be a little more subjective” (interview 32-090723).

By identifying implementation as a major concern for investors in Russia, the interview data support the theory’s emphasis on policy application and justify further investigating the discretionary behavior of those actors who interpret and apply laws

– regulatory bureaucrats.

Vague Laws, Bureaucratic Discretion, & Investor Uncertainty

This chapter’s statistical evidence provides robust support for the argument that bureaucratic discretion deters investment. Due to the limitations of the available survey questions, however, the quantitative analyses tell us very little about why

bureaucratic discretion troubles investors. This section uses interview evidence to

corroborate the theory’s proposed causal mechanisms with observations from those

who know and operate within Russian markets.

A common theme emerges from interviews with businesses and policy experts in

Russia that echoes the theory’s motivating logic: uncertainty about the application

and implementation of regulatory policies is a formidable barrier to investment. In

the opinion of policy consultants and NGO representatives, ambiguous and contradic-

tory statutes are a common feature of Russian regulation that lead to unpredictable

regulatory decisions (researcher at business-related NGO, interviews 31-090626; pol-

icy consultant, 18-090729). When asked about regulatory conditions that aﬀect busi-

ness development, respondents complained speciﬁcally about the discretion that these

74 vague laws grant to regulatory bureaucrats. According to a lobbyist for the food processing industry who once worked in the presidential administration, vague statutes allow regulatory agencies to append or extend laws in ways that are difficult for businesses to follow, let alone anticipate: “Before technical regulation comes into effect, obligatory requirements in this sphere pass through one department or another in the ministries...Frequently these departments each add their own requirements that effectively disagree with each other, creating a confused picture. It is a highly confused situation” (interview 8-090716). A researcher at one prominent NGO pointed out how the resulting confusion affects businesses: “[Companies] carefully study regional laws before deciding whether or not to enter a region...Companies in a majority of cases try to obey these laws, but quickly become convinced that fulfilling all the requirements is impossible” (interview 31-090626).

According to interviewees, much of businesses’ confusion and frustration with economic regulation comes from trying to anticipate how regulatory officials will apply ambiguous statutes. According to business representatives, vague regulatory laws create “gray areas” that bedevil investors by making them vulnerable to authorities’ whims (government affairs officer at international food corporation, interview 25-

090716). The director of a medium-sized meat plant explained ﬁrms’ predicament in the following words: “When a [regulatory] point can be interpreted broadly or narrowly, and that decision lies with a bureaucrat, then there can be obstacles. All he has to do is start taking a broader interpretation, and you’re done for....The regulation’s formulation allows oﬃcials to interpret it such that any claim against you has legal grounding” (interview 28-090720). Given investors’ worries over these

“gray areas” and discretionary decisions, minimizing opportunities for bureaucrats to make subjective interpretations, according to an advisor to Moscow’s mayor, should

75 be a high priority task for any government body that wishes to facilitate investment

(interview 15-090722).

Field interviews provide further qualitative support for the theory’s speciﬁc claim

that bureaucratic discretion deters investors by creating uncertainty about regulatory

expectations. During interviews, ﬁrm directors and business representatives discussed

how unpredictable policy application impairs investors’ ability to assess projects’ fea-

sibility. An executive of an international consulting company explained the challenges

of trying to plan when regulatory authorities have high amounts of discretion: “No-

body knows where a tax assessment is going to come from. They generally come in

completely from left-ﬁeld. You try and analyze all risks and try and prepare for those

that you think are probable, but in the end – the tax department is going to come

at you from wherever” (interview 32-090723). Not only does regulatory discretion

makes it harder for businesses to predict what type of performance counts as comply-

ing with the law, but interviewees pointed out that a temporal dimension complicates

the problem further – when faced with the same circumstances, will discretionary bu-

reaucrats apply regulations consistently across time periods? On this point, a senior

researcher of regional development at a prominent NGO remarked:

“It is better to take [regulatory] questions and include some regional of- ﬁcials; ask them to clarify as needed...The formal, juridical statutes are ambiguous. So if you have come to an agreement [with regulators] about a plan of action for how you will run your business, then you can operate in peace. Naturally, that is only for as long as those administrators stay in place. The second they leave, then the whole system can change. That risk is very high. In my opinion, this is one of the factors that deters investment ” (interview 31-090626).

Thus, even when discretion might possibly help investors by providing bureaucrats with the flexibility to handle complex questions or accommodate firms’ individual circumstances, respondents’ experience suggest that such benefits depend upon a fragile arrangement that is subject to unpredictable changes over time. 76 The Costs of Bureaucratic Discretion

The theoretical claim that bureaucratic discretion deters investment rests upon the

notion that unpredictability in the application of regulatory policies puts businesses at

greater risk of experiencing economic losses that can render investments unproﬁtable.

Evidence from ﬁeld interviews sheds light upon the types of losses that investors fear

may arise from regulations being applied in a discretionary manner. Conversations

with business representatives show that, in the extreme case, investors worry that

discretionary bureaucrats may interpret regulations in a way that will lead to closing

business operations: “It is always possible that oﬃcials will ﬁnd some problem and

halt your production” (executive of large seafood company, interview 39-090720).

Even if short-lived, forced closures can harm the viability of investment projects by

disrupting supply relations, creating bad press, and choking oﬀ revenue streams.

Interviews also reveal that the losses that business actors associate with bureau-

cratic discretion can be much more mundane and yet still hinder investment. If

reoccurring or drawn-out over a long period of time, unexpected regulatory diﬃcul-

ties act as a steady drain on businesses’ ﬁnite resources and erode investors’ proﬁts.

According to an international retail consultant, even “winning” against discretionary

regulators has this sapping eﬀect:

“Because of the gray areas, because of the level of interpretation, [tax authorities] have more latitude than most jurisdictions to be able to come up with assessments. And some of those assessments are not necessarily well-reasoned or technically accurate or technically supportable, but it’s an assessment anyway. So I’ve seen some ridiculous tax assessments, things that are clearly wrong...multi-million dollar assessments that come out to a couple thousand dollars of ﬁnes in the end. Now that wastes a lot of time and eﬀort” (interview 32-090723).

In fact, investors need not have actual dispute with regulatory oﬃcials in order to accrue losses from bureaucratic discretion; the threat of such problems alone is often

77 enough to force ﬁrms to take on additional costs, such as extra legal or accounting

help, in the hopes of protecting their business from unforeseen claims. The consulting

executive continues: “If your document is stamped down here instead of up here, that

can be a reason for rejecting it. So you have a tremendous cost for compliance...You’ve

got to have a lot of people managing the paperwork” (interview 32-090723). From a

business standpoint, the opportunity cost of such preventative measures is quite high

– resources spent trying to defend against potential regulatory claims are resources

that are unavailable for investment.

In addition to noting direct monetary costs, the responses of business actors and

experts highlight the large amount of time required to deal with the challenges of an

unpredictable regulatory environment. An investment expert described the temporal

costs of discretion by relating the following scenario:

“To sell alcohol, you need to get a license once every six months...You barely get your license, make a deal with suppliers, work for two or three months, and then – time to go again! And it is not a given that they will give it to you immediately. That means for a period of time it is not clear what you are selling or how you are getting along. But that does not bother anyone...they can always just say: ‘That’s it! I’m not giving you a license’...and no one will especially explain to you why. Sure, you can try to slug it out in court, and maybe eventually you’ll get to the end [of the court proceedings] – in a year, year and a half. And it you aren’t sick of it by that point, then they will get around to telling you why you do not get a license” (interview 31-090626).

These delays and drawn-out proceedings represent more to investors than mere inconvenience or distraction. For businesses looking to make a return on investment, the connection between time and money is concrete. The head of a regional food processing company identified the time demands as the most off-putting part of dealing with unpredictable regulatory requirements: “It’s all about time...if you have borrowed credit, then you are definitely making payments in the meantime because your interest is running up every month. If you are using your own money, well...money brings 78 in more money, but not if your capital is stuck in this project and you cannot even get started” (interview 28-090720). Interviews suggest that by creating uncertainty about how standards apply and to resolve problems that inevitable arise, bureaucratic discretion generates cumulative costs in both time and money that either drive away or crowd out investment.

Investors’ Response to Regulatory Uncertainty

Having established the relative importance of predictable policy application to investors and illustrated the mechanisms by which bureaucratic discretion impedes their investment activity, the interview evidence can also give us insight into how concerns about regulatory environment shape the process by which investors make their investment decisions. How do concerns about predictable policy application inﬂuence investors’ behavior? Interviews show that investors’ worries about the implementation of regulation leads them to investigate regional bureaucracies before making large investment decisions. To begin with, businesses looking at new investment locations typically gather information about the behavior of regulatory oﬃcials by speaking with other businesses already operating in the area, inquiring with local and national business associations, or hiring consultants with experience in the potential region.

Conducting due diligence in this manner allows investors to hear about consistent problems and avoid those regions where regulatory oﬃcials have acquired reputations for discretionary decision-making (interview 28-090720, interview 8-090716).

Nearly all respondents agreed that, in addition to seeking secondhand reports about regions’ regulatory environment, serious investors meet personally with regional bureaucrats to discuss their proposed projects prior to actual investment. One source explained the logic behind such efforts: “What investors like is certainty...Once you’re set up, you have a long-term investment. You need to know that there’s some kind of stability, some kind of regularity, and some kind of reliability so that you’re 79 not spending your whole time fighting a system that’s a moving target” (consulting executive, interview 32-090723). A string of positive interactions may sufficient to put investors’ fears to rest, according to one regional expert: “When they [investors] can see eye-to-eye and build working relations with regional administrations, it turns out that there are good opportunities to invest, build factories, etc.” (researcher at

NGO, interview 31-090626). Alternatively, witnessing the bureaucratic apparatus in action can also cause investors to walk away from a region completely: “If the regional oﬃcial is reasonable, then you can ﬁnd a common language. If he acts without any real constraints, then the investment environment will be third-rate” (director of nation-wide organization that represents small businesses, interview 30-080826).

Investors’ eﬀorts to collect all available information about the predictability of regulatory behavior in candidate regions indicates that investors recognize bureaucracy’s inﬂuence on the viability of their long-term projects and that they base their choice of investment location partially upon their assessment of that regulatory policies will be predictably applied for the foreseeable future.

In discussing the relationship between bureaucratic discretion and investment, respondents relayed anecdotal accounts of unusual investment patterns that they attributed to investors prioritizing predictable regulatory arrangements over other concerns. According to members of a national association of meat processors, the

Voronezh region in Russia’s South is blessed with agricultural conditions that are

“paradise” for raising livestock; despite these favorable conditions, however, the region has attracted surprisingly little of the growing investment in agriculture and food processing due to the diﬃculties of dealing with regulatory oﬃcials there. Instead of locating in Voronezh, which arguably has the best farmland in Russia, many investors

80 have chosen second-best sites in the neighboring regions of Penza and Belgorod because regional administrations there have taken considerable eﬀorts to make regulation transparent and predictable (interview 19-090725).14 One regional investment expert alluded to similar dynamics in the auto industry. According to the interviewee, predictable regulatory requirements and the ability to work amicably with regional oﬃcials help to explain why BMW and KIA produce cars in Kaliningrad, an isolated territory of the Russian Federation that is separated from the rest of the country by

Latvia, Lithuania, and Belarus. Rather than locate somewhere within Russia closer to

Moscow, the country’s largest and dominant consumer market, these firms prized the benefits of increased regulatory certainty in Kaliningrad enough that they are willing to incur the cost and inconvenience of transporting finished goods across international borders to reach their customer base (interview 31-090626). Such accounts suggest that investors’ concerns over bureaucratic discretion and unpredictable regulatory environments can play a significant role in shaping businesses’ investment plans.

In summary, the interview data presented in this section provide qualitative evidence that investors in Russia recognize bureaucratic discretion and unpredictable policy application as obstacles to investment. Across a wide spectrum of economic sectors and organization size, interviewees in the business world reported that uncertainty over the application of regulatory policy hinders business activity and shared examples of how bureaucratic discretion creates obstacles for investors. Conversely, no business representative made a case in favor of bureaucratic discretion or indicated that investors seek it out in potential investment locations. Policy experts in the legal, academic, and NGO community pointed to ambiguous legislation and contradictory

14According to my interlocutors, the governor of Belgorod, Yevgeny Savchenko, has even been known to call representatives from each of the relevant government agencies together into meetings with potential investors in order for each regulator to tell investors – in front of the governor – exactly what procedures are required and the proper way to get investment projects oﬀ the ground without delays or setbacks (interview 19-090725).

81 laws as contributors to bureaucratic discretion and regulatory uncertainty. And, while many experts highlighted perceived negative consequences of discretionary regulation, none pointed to it as a beneﬁcial tool for regions to attract investors. Reports from interview respondents even indicated that that some business-savvy regional administrations have begun using transparent regulatory processes as a selling point to investors (interview 15-090722); on the other hand, there were no observations or reports of regional governments advertising discretionary regulators as an enticement to potential investors.

Conclusion

This chapter tests empirical predictions of the dissertation’s theory at the microeconomic level. As part of my empirical strategy to test multiple predictions that follow from Chapter 1’s claims, I leverage data from surveys of Russian firm managers. Sec- tion I focuses on the main hypothesis. Using logit models of firms’ investment plans, I demonstrate that firm managers’ perceptions that regional bureaucrats have decision- making autonomy has a robust, negative relationship with firms’ plans to invest in

ﬁxed capital assets in the immediate future. Furthermore, the in-depth, qualitative investigation in Section II of investors’ response to their regulatory environment help to validate and provide insight into these robust statistical patterns. Interviews with

ﬁrm directors, heads of business associations, and policy experts conﬁrm that, not only are investors concerned about the unpredictable application of laws by regulatory bureaucrats, but this regulatory uncertainty can have a meaningful impact on businesses’ decisions over where and how to invest.

Taken together, the empirical analyses in this chapter oﬀer a wide-ranging test of the micro-level logic that lies at the heart of the dissertation’s theoretical framework.

The chapter provides evidence that bureaucrats’ discretionary behavior inﬂuences

82 firms’ decisions about whether or not to invest. The next chapter begins testing another key claim of the dissertation, namely that discretion and investment’s relationship is affected by the broader institutional environment – depending upon the institutional context in which they find themselves, investors should associate bureaucratic discretion with greater uncertainty in some locations than in others. Combining this chapter’s survey data with data from the Moscow Carnegie Regional Monitoring

Project on regional variation in political competition, Chapter 3 tests the prediction that the negative relationship between discretion and investment is strongest in regions where restrictions on political participation prevent eﬀective ﬁre-alarm oversight of discretionary bureaucrats.

83 CHAPTER 4

DOES POLITICAL COMPETITION MAKE

BUREAUCRATIC DISCRETION BETTER FOR

INVESTORS? AN EMPIRICAL TEST

Introduction

The previous chapter has demonstrated a robust association between ﬁrm managers’ perception of regulatory bureaucrats’ discretion and lower probabilities that they plan to invest any time in the coming year. The apparent contradiction between these

ﬁndings within data from the Russian Federation and existing research that draws heavily from the experience of developed democracies begs the following question: to what extent does the relationship between discretion and investment depend on the broader institutional context? In making the case that uncertainty over the application of laws deters investors who seek predictable environments, I have claimed that uncertainty surrounding bureaucratic discretion is a particularly acute problem for economic actors in environments that lack political institutions that encourage robust competition. The following hypothesis identiﬁes the empirical implications of this claim that political competition provides economic actors with a more supportive institutional environment:

Hypothesis 2: Holding other factors constant, the relationship between

84 bureaucratic discretion and investment should be more negative in regions characterized by low political competition than in regions with high political competition.

This chapter tests that prediction using region-level data on firms’ political environments in conjunction with the Russian enterprise-level survey data from the previous chapter. The first section introduces the measure of political competition and provides descriptive statistics. Following the description of the region-level data, the middle section uses a series of statistical analyses to conduct the main empirical tests of Hypothesis 2. The penultimate section explores the main findings’ robustness to alternate measures and model specifications. After the robustness checks, the chapter concludes with a brief summary and discussion of the findings’ implications.

Political Competition & Russia’s Regions

The empirical results from the previous chapter show that, on average, firms’ perceptions of higher bureaucratic discretion correlate with lower probabilities of firm investment in fixed capital assets. According to the theory, however, political competition can play an important role in helping to make policy application by discretionary bureaucrats more predictable for investors. Politically competitive environments encourage elected leaders to be attentive to constituents’ concerns, and in doing so, they create more opportunities for economic actors to have their voices heard within the policy-making process. Moreover, to the extent that political competition does bring these additional voices into the policy-making process, political competition spreads the costs of monitoring bureaucrats across multiple actors. These features of political competition give businesses predictable channels for challenging regulatory enforce- ments with which they disagree and help to reduce investors’ uncertainty about what to expect from a high-discretion regulatory environment. Thus, the theory’s conditional logic implies that the empirical relationships observed in the previous chapter 85 should vary in predictable and systematic ways with the level of political competition in respondents’ region. This is the observable implication that I test in this chapter.

To measure political competition in ﬁrms’ regions, I use component scores from a regional democracy index created by Nikolai Petrov and Alexei Titkov as part of the

Moscow Carnegie Center’s Regional Monitoring project. The original index records experts’ assessment of Russia’s regions along ten diﬀerent dimensions for four diﬀer- ent time periods between the years 1991-2006. For this analysis, I use the time period immediately prior to the survey, 2000-2004. To measure political competition for the sampled regions, I construct an additive index that captures three dimensions of the political environment: representative elections (existence of free and fair elections, few limitations on political rights), the openness of political life (the extent of transparency and public involvement in the political sphere), and pluralism (participation by stable parties or legislative factions before and after elections).1 Each of these three scores can range from 1 to 5 (ordered worst to best); the lowest additive score is a 6 (Khabarovsk Krai and the Republic of Bashkortostan) and the highest is 15

(Sverdlovsk Oblast).2 For ease of interpreting its interaction with bureaucratic discretion, the actual analyses transform this variable by subtracting the index’s mean

1The original Petrov and Titkov measure also includes component scores for economic liberalization, corruption, civil society, elite recruitment and coordination, municipal governance, media freedom, and “regional political structure.” To maintain conceptual focus on electoral accountability and political participation, I leave these components out of the analysis. Results are robust, however, to using the full set of measures. For a English-language discussion of data collection and basic descriptions, albeit regarding earlier versions of the dataset, see McMann and Petrov (2000).

2In addition to this additive speciﬁcation, I have also used principal component factor scores as a more precise estimate of latent levels of political competition. All results hold if this factor-based index is used in place of the additive scale. Diagnostic tests show that the three components scores are closely related; the average interitem covariance is 0.59, and the Cronbach’s alpha statistic for scale reliability is α = 0.93.

86 (mean = 9). Regions’ scores for the political competition index and components appear below in Table 4.1. Summary statistics for all region-level variables appear in

Appendix B.

Region Political Openness Elections Pluralism Competition component component component index score score score Sverdlovsk 14 5 4 5 Nizhniy Novgorod 13 4 5 4 Moscow 11 4 3 4 Novgorod 10 4 3 3 Voronezh 9 3 3 3 Omsk 9 3 3 3 Smolensk 9 3 3 3 Tula 9 3 3 3 Rostov 8 3 2 3 Khabarovsk 6 2 2 2 Bashkortostan 6 2 2 2 Note: Regional democracy scores from the Moscow Carnegie Center’s Regional Monitoring project, taken for the period 2000-2004. Data on GDP per capita is averaged from 2002-2004, taken from annual Rosstat publications. Regions are those sampled by Frye (2006).

Table 4.1: Regional Democracy Scores, 2000-2004 (Regional Monitoring Project)

While shared legacies and the overarching sovereignty of Russia’s federal laws allow us to control for a host of administrative and legal factors that would be diﬃcult to account for in a cross-national study, political and economic factors vary meaningfully across regional territories for the period under study.3 This diversity is reﬂected in the measures of political competition reported in Table 4.1. As a visual reference, Figure

3Very shortly after the survey was conducted in 2005, then-president Vladimir Putin pushed through reforms that replaced direct gubernatorial elections with a system wherein regional legislatures ratify candidates that have been nominated by the Russian president. Although these 87 4.1 maps out the regions and overlays the measure of political competition. At one end of the spectrum, the sample includes regions that were among the Regional Monitoring

Project’s top scorers: Sverdlovsk and Nizhniy Novgorod (ranked second and ﬁfth, respectively.)4 At the spectrum’s other end, the sample also includes Khabarovsk and Bashkortostan, locations well-known for their leaders’ tight control over regional politics through the use of strong-arm tactics and clientelistic political machines. As a direct result, the survey provides the widest possible range of variation in Russia’s regional politics for investigating how political conditions aﬀect investors’ response to bureaucratic discretion.

At the same time, variation in the competitiveness of regional politics also allows for a direct test of the delegation-as-commitment logic. Conventional credible commitment arguments suggest that political leaders’ moral hazard problem is high where politicians face few constraints on ex-post behavior (Miller 2000, Jensen 2006).

This logic implies that, by providing an autonomous counterbalance to relatively unconstrained leaders, delegation to independent agents should encourage ﬁrm investment in those places where political leaders lack organized political opposition or have little threat of losing oﬃce. Thus, a conventional commitment argument would

reforms undoubtedly hold long-term implications for the dynamics of regional political competition in Russia, the changes come after the data were collected and thus do not directly concern these analyses.

4The idiosyncrasies of Russian politics may introduce some ceiling eﬀects on the analysis – being “politically competitive” in Russia still does not necessarily put a region’s political environment on par with the advanced democracies of the OECD. Witness, for example, complaints about the heavy-handed governing style of Sverdlovsk’s popular governor, Eduard Rossel’ (Golosov 2004) or the bitter in-ﬁghting among Nizhniy Novgorod’s prominent regional elites (Sharafutdinova 2007). However, in the late 1990s and early 2000s, these regions did display traits that we associate with more highly competitive environments, such as high levels of popular support for democracy and strong local parties that competed intensely – and successfully – with national parties (Sverdlovsk) or hotly-contested gubernatorial elections that resulted in the surprise victory for an outside candidate (Nizhniy Novgorod).

88 Note: Region-level political data from the Moscow Carnegie Center’s Regional Monitoring project. Map shows regions sampled by Frye (2006). From left to right: Novgorod, Smolensk, Moscow (city), Tula, Voronezh, Rostov, Nizhniy Novgorod, Republic of Bashkortostan, Sverdlovsk, Omsk, Khabarovsk Krai.

Figure 4.1: Sampled Regions & Variation in Political Competition

predict that, in regions with closed and uncompetitive politics, perceptions of bureaucrats’ independent decision-making should correlate positively with firm managers’ investment. In contrast, I argue that closed political processes and unresponsive governments restrict economic actors’ channels for resolving their regulatory difficulties and inhibit their ability to serve as effective monitors of state agents. This generates the empirical prediction that more discretion to regulatory bureaucrats in politically- uncompetitive environments will deter investors by increasing policy uncertainty.

89 Testing the Conditional Theory of Bureaucratic Discretion &

Investment

Is there evidence that bureaucratic discretion presents a particular problem for economic actors in settings with less democratic institutions? Even before turning to the results of the controlled statistical models, a basic descriptive comparison of investment and bureaucratic discretion in high- versus low-competition regions suggest this to be the case.

Table 4.2 presents crosstabs relating firms’ plans to invest and their perceptions of bureaucratic discretion, grouped by high- versus low-levels of regional political competition. In highly competitive regions, there is no statistically significant difference between managers with differing perceptions of bureaucratic discretion (p = 0.658).

In regions with little political competition, however, only 25.8% of ﬁrm managers that perceive bureaucrats to have high discretion report to have investment plans for next year, as compared to 42.8% of managers who see bureaucrats as having low discretion

( χ2(1) = 8.281, p = 0.004). Although ﬁrms in high-competition regions seem less concerned about bureaucrats’ discretion over policy, ﬁrms’ perceptions of regional bureaucrats’ discretion in applying and interpreting laws appear to be strongly related to investment decisions in low-competition regions.

Results from Multilevel Statistical Analyses

The crosstabs and scatterplots above provide empirical evidence that the negative relationship between bureaucratic discretion and investors’ willingness to invest depends upon the broader political institutional environment. As a more rigorous test of the hypothesis, I now turn to estimating statistical models that control for potentially confounding factors and competing explanations. Speciﬁcally, I re-estimate

90 In regions with low levels of political competition.

Does ﬁrm have Do regional bureaucrats plans to invest make decisions independent in the next of other gov’t bodies? 12 months? Yes No Total

Yes 24 109 133 (25.81%) (42.75%) (38.22%)

No 69 146 215 (74.19%) (57.25%) (61.78%)

Total 93 255 348 (100%) (100%) (100%)

In regions with high levels of political competition.

Does ﬁrm have Do regional bureaucrats plans to invest make decisions independent in the next of other gov’t bodies? 12 months? Yes No Total

Yes 15 84 99 (50.00%) (45.65%) (46.26%)

No 15 100 115 (50.00%) (54.35%) (53.74%)

Total 30 184 214 (100%) (100%) (100%) Note: Survey data from Frye (2006). Column percentages in parentheses. High versus low restrictions on political competition are relative to the mean of regional score. For upper table, the Pearson chi-squared statistic for the paired observation is χ2 = 8.281, p = 0.004. For lower table: χ2 = 0.192, p = 0.658.

Table 4.2: Institutional Context Aﬀects Discretion’s Relationship with Investment

91 the main models from Chapter 2, albeit this time with an eye towards understanding how economic actors’ response to bureaucratic discretion changes across political environments.

The models in Table 4.3 extend the analysis on firms’ intentions to invest to a multilevel logistic regression model of firm investment.5 As before, the new interactive model adopts a random-intercept, random-coefficient specification to account for unobserved heterogeneity that might vary across region.6 Models also include the same control variables as in earlier analyses to account for factors that are likely to affect both firms’ attitudes towards future investment as well as their perceptions of bureaucratic discretion; among others, these controls include measures for firms’ perceptions of the policy environment, attitudes towards regional political institutions, and various firm-level characteristics that likely affect both firms’ investment strategies and their contact with the state’s regulatory apparatus.7 In addition, all models now include three region-level variables: the measure of regional political competition described above, its cross-level interaction with managers’ perceptions of bureaucratic discretion, and a (logged) variable for average regional GDP per capita (2002-2004) to help control for regions’ economic development.8

5Multilevel models provide flexibility in modeling hierarchical relationships, helping us to make inferences about variables that we observe at different levels of analysis (Gelman & Hill 2007). The nested relationship of firms within regions makes this a natural modeling choice here.

6The empirical relationships do not depend on this chosen modeling strategy. Analyses produce similar results in non-multilevel logistic regression with region covariates or, alternatively, multilevel analysis via Bayesian estimation with diﬀuse priors. Results provided in Appendix B.

7For a detailed description of the full list of control variables, please see the discussion in the previous chapter.

8Just as the results in Table 4.3 are robust to a variety of estimation strategies, they also hold for model speciﬁcations that include a richer set of region-level control variables. For example, following Brown, Searle, and Gehlbach (2009), I control for regional bureaucrats per capita to check the possibility that bureaucracy size may inﬂuence investors’ opinion of both bureaucratic discretion and regional investment climate, but neither this variable nor additional region-level

92 Firm Investment dummy, 1 = ﬁrm plans to invest during coming year (1) (2) (3) (4) Bureaucratic Discretion -0.474*** -0.645*** -0.614*** -0.672*** 1 = no discretion, 4 = high discretion (0.135) (0.157) (0.168) (0.184) Frequent Changes to Laws 0.115 0.082 0.149 1 = no obstacle, 5 = very serious obstacle (0.112) (0.118) (0.125) High Tax Rates -0.385*** -0.417*** -0.379*** 1 = no obstacle, 5 = very serious obstacle (0.122) (0.130) (0.144) Regional Administration 0.508*** 0.433** 0.450* 1 = poor job, 5 = excellent job (0.196) (0.210) (0.230) Regional Courts -0.149 -0.169 -0.022 1 = poor job, 5 = excellent job (0.148) (0.152) (0.174) Regional Governor 0.003 -0.372** -0.328* -0.441** 1 = poor job, 5 = excellent job (0.104) (0.185) (0.198) (0.222) Access to Finance -0.087 -0.027 -0.023 0.039 1 = no obstacle, 5 = very serious obstacle (0.067) (0.083) (0.086) (0.096) Labor Shortages -0.046 -0.007 0.042 1 = no obstacle, 5 = very serious obstacle (0.086) (0.091) (0.103) Competitive Pressures 0.089 0.109 0.027 1 = no obstacle, 5 = very serious obstacle (0.085) (0.089) (0.100) Privatized Firm 0.078 0.081 0.206 dummy, 1 = privatized, former SOE (0.275) (0.283) (0.346) Annual Sales 0.469*** 0.459** 0.357* 0.275 -1 = decreasing, 1 = increasing (0.165) (0.182) (0.191) (0.219) Firm Size 0.338*** 0.306*** 0.294*** 0.333*** number of employees (logged) (0.072) (0.083) (0.087) (0.115) Private Firm 0.766** 0.774* 0.620 dummy, 1 = private ownership (0.336) (0.441) (0.444) Bureaucratic Corruption -0.071 dummy, 1 = perceived as corrupt (0.252) Past Investment 2.049*** dummy, 1 = invested in last 3 yrs. (0.295) Constant -3.081*** -2.149 -1.577 -3.920* (1.155) (1.436) (1.565) (2.133)

GDP per capita 0.330 0.446* 0.421 0.445 in constant 2000 rubles per 1000 persons (logged) (0.245) (0.252) (0.268) (0.308) Regional Political Competition -0.139 -0.185 -0.148 -0.188 index, -3 = uncompetitive, 5 = highly competitive (0.111) (0.127) (0.130) (0.149) Competition × Discretion 0.124** 0.144** 0.124* 0.172** interaction (0.059) (0.066) (0.066) (0.080) Dummies for Sector & Legal Form No No No Yes Log-likelihood -286.738 -235.432 -216.731 -195.699 No. of Cases 470 403 365 402 Note: Survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center and economic data from Rosstat. Coefficients represent estimates from multilevel logistic regressions with a random coefficient for the bureaucratic discretion variable and random intercepts at the region level; standard errors in parentheses. Out of space concerns, random effects and sector/legal-form variables not reported. * p<0.10, ** p<0.05, *** p<0.01

Table 4.3: Political Context Conditions Investors’ Response to Discretion 93 Table 4.3 reports the results from multiple models in order to demonstrate that estimated relationships hold across various model specifications. As a baseline specification, Model 1 includes only the independent variables and a very limited number of controls for the investment environment and firm characteristics. Subsequent models test the results’ robustness to controlling for an extended set of controls (Model 2), perceptions of bureaucratic corruption (Model 3), and factors specific to firms’ economic sector and legal form (Model 4). The final column of Table 4.3 also examines the findings’ sensitivity to controlling for fixed capital investments in the recent past.

Before turning to the interaction between perceived discretion and regional political competition, consider briefly the findings related to the control variables. Consis- tent with earlier results and standard expectations from economic theory, problems with high tax rates display a negative and statistically significant association with

firms’ intent to invest across all models. As before, the other policy environment variables – perceptions of bureaucratic corruption and problems with frequently changing laws – do not bear any statistically significant relationship to firms’ investment plans.

The analyses’ results also indicate that respondents’ attitudes towards regional political institutions relate directly to their decisions about whether or not to invest in the future. Consistent with the scenario under which firms’ non-investment relates to potential frustrations with the poor execution of governors’ policies that respondents actually support, we see that higher support for the regional administration has a positive and statistically significant correlation with firms’ investment plans, while support for the governor displays a negative and significant coefficient. Regard- ing economic courts, however, the analyses reveal no significant association between respondents’ assessments of regional economic courts and decisions to invest. Hold- ing other factors constant, the analyses also fail to find any statistically significant

controls, such as population or regional infrastructure (measured by railway density) change the results. Tables for these analyses appear in Appendix B.

94 relationship between ﬁrms’ intent to invest and managers’ perceptions of problems associated with access to credit, shortages of skilled labor, and intense economic competition.

Finally, the models show that several variables measuring ﬁrm characteristics help to predict plans to invest in the immediate future. In particular, larger ﬁrm size and

fixed capital investment in the past three years are both associated with a higher probability of investment. Non-decreasing sales trends and private ownership also display positive relationships with firms’ intent to invest, although the statistical uncertainty around the coefficient estimates increases as more controls enter the analyses.

We now return our attention to the test of the theory’s empirical prediction of political competition’s mediating eﬀect on the negative relationship between bureaucratic discretion and economic actors’ willingness to invest. Across all models in

Table 4.3, the bureaucratic discretion variable has a negative and statistically significant coefficient estimate, indicating that when the mean-centered political competition variable takes a value of zero (i.e., “medium” levels of political competition), perceptions of higher bureaucratic discretion are associated with a decrease in the predicted probability of firm investment. Moreover, the coefficient on the interaction between firms’ perceptions of bureaucratic discretion and the competitiveness of regional politics is positive and statistically significant in all models; for rising levels of political competition, perceptions of higher discretion are expected to correlate with an increasingly less negative (perhaps eventually positive) change in the probability of investment. According to these results, we should reject the null hypothesis that bureaucratic discretion’s relationship to investment plans is not conditioned by levels of political competition within the region. Since nonlinear models increase the challenge of interpreting interacted variables’ substantive effects, Figure 4.2 uses a predicted probability graph to discuss the estimated relationship between perceived

95 regulatory discretion, institutional constraints, and ﬁrms’ intentions to invest in ﬁxed capital assets.9

Figure 4.2 highlights the change in the predicted probability of firms’ investing in the next twelve months associated with a change in perceived bureaucratic discretion at a given level of political competition. Holding all variables at their sample medians, I manipulate whether a hypothetical firm perceives that regional bureaucrats have high or low discretion and then predict the probability of firm investment for different levels of regional political competition. In the sample’s least politically- competitive environment, a hypothetical firm manager who perceives regional bureaucrats to have high discretion (the solid line) has a predicted probability of investment that is more than 50 percentage points lower than an identical counterpart who reports regional bureaucrats as having no discretion (the dotted line). As the level of regional political competition increases, however, this gap between the two hypothetical firms diminishes as the high-discretion profile’s predicted probability of investing converges steadily towards the that of the low-discretion profile. In fact, the overlapping confidence intervals suggest that on the upper ends of the scale, where regional institutions foster open and competitive politics, the difference between a high-discretion and low-discretion scenario is negligible from a statistical viewpoint.

To reiterate the point, these results support my argument’s empirical predictions.

9Recent scholarship by Ai and Norton (2003) and Berry et al. (2010) remind scholars about general principles for interpreting interaction effects within nonlinear models. Together, they show that a simple t-test on the coefficient estimate of the interaction term does not necessarily test an interaction effect’s statistical significance. This problem arises because nonlinear models make the interaction effect conditional on the covariate profile of the other independent variables (as is also the case with marginal effect on uninteracted variables in nonlinear models). Therefore, an exhaustive descriptive of the interaction effect would require plotting the estimated effect of the interaction for every possible profile within the data. For convenience’s sake, I present the interaction effect indirectly by comparing predicted probabilities for the theoretically-relevant cases of high- versus low-discretion while holding the all other variables at their median values.

96 ● ● ● ● ●●●● ●●●●● ●● ●●●● ●● ●●●● ●● ●●● ●●●● ●● ●●●●●● ●● ●●●● ●●● ●● ●●●●● ●● ●● ●●● ● ●●● ●●●●● ●● ●●● ● ●● ●●●● ● ●●●●● ● ● ● ● ●●●● 1.0 ●● ● ●●● ●● ●●● ●●● ●●●● ●● ●●●● ●●●●● ●● ● ● ●●● ● ●●● Low Discretion●●● ●●●●● ●●● ● ●●●● ●● 0.9 High Discretion

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

Predicted Probability of Firm Investment ●● ●●● ●● ● ● ● ●●●●● ● ●●●●●● ●●●● ● ●●●● ●● ●●●● ●●● ●●●●●● ●●●● ●● ● ● ●●●● ●●●●● ●●● ●●●●●● ●●● ●●●● ●●● ●● ●●● ●●●● ●●●●● ●●●● ●●●● ● ●●● 0.0 ●●●● ●●● ●●●●●● ●●● ●● ●● ● ●●●●● ●●● ●●●●●●● ● ● ● ●● ●●●● ●●●● ● ●●●●●● ●●●● ●● ●●● ●● ●●●●●● ●● ●●●● ●●● ●●●●● ●● ●●●

Low Medium High Regional Political Competition

Note: Firm-level survey data from Frye (2006), region-level political variables for the eleven sampled regions come from the Moscow Carnegie Center and economic data from Rosstat. Bold lines represent the predicted probability that a hypothetical ﬁrm would invest, given the level of regional political competition; discretion is manipulated from minimum (1) to maximum (4) while holding all variables at their median value. Small dashed lines represent 90% conﬁdence intervals obtained via simulation in R.

Figure 4.2: Investors’ Response to Bureaucratic Discretion Shaped by Institutional Context (Predicted Probabilities)

97 Institutions that engender competitive politics make government more open and responsive in a way that helps to reduce economic actors’ uncertainty about the discretionary behavior of regulatory bureaucrats; where that supportive institutional environment is lacking, bureaucratic discretion generates uncertainty over the application of laws that remains a troubling obstacle to investors. It is worth noting that these results run directly counter to conventional logic that, given unconstrained political leaders, investors should welcome greater independence for executive agents as a type of substitute for political constraints on leaders’ ex-post behavior. Quite the opposite appears true in these data – the good that substitutes for desirable political institutions is less discretion rather than more. This finding puts scholars’ seemingly sanguine attitude towards bureaucratic discretion into perspective. In the developed democracies that inform a majority of studies, high-quality institutions attenuate investors’ uncertainty about independent bureaucrats’ behavior. On the spectrum’s other end – where a sizable portion of the world lives – economic actors in the low- democracy, high-discretion scenario find themselves in the worst of all worlds: high uncertainty over the application of regulatory rules coupled with low-quality regional institutions that offer limited political channels for resolving investors’ difficulties.

The results in Table 4.3 are robust to a variety of estimation strategies and model speciﬁcations. This fact is witnessed by a number of additional analyses in Appendix

B demonstrating the ﬁndings’ resilience to additional control variables or the exclusion of outlying regions. The next section continues the robustness checks, examining the sensitivity of Table 4.3’s ﬁndings to measurement choices associated with the analysis’s dependent and key independent variables.

98 Robustness Checks & Additional Analyses

This section conducts a series of additional analyses aimed at demonstrating the resiliency of the chapter’s main empirical finding: supportive political environment attenuate investors’ negative response to high bureaucratic discretion. The section begins by investigating the results’ robustness to adopting alternate specifications and measures of the key independent variables. Following these analyses, I present a second set of empirical models in which the dependent variable, firms’ intent to invest, is replaced by other reported firm behaviors that should follow a similar logic.

After presenting the results of these robustness checks, the chapter concludes.

Alternate Measures of the Key Independent Variables

The results of the previous analyses show that, in regions characterized by restrictions on political competition, firm managers’ perceptions of high bureaucratic discretion are correlated with a decrease in the probability that a firm has plans to invest in the coming year; in high-competition regions, this negative relationship subsides dramatically. Continuing the analyses, Table 4.4 presents four separate models that probe the robustness of these findings to various changes in the specification of the key independent variables. For the sake of standardization, all models include the same set of control variables as used in Model 2 above.

Table 4.4’s ﬁrst column reports the results of re-estimating the model using a alternate coding of the measure of perceived discretion. Because ordinal variables have potentially unequal spacing between ordered categories, it may be wrong to assume that movement from one ordered category to the next has a linear and additive eﬀect (Alvarez, Bailey & Katz 2011). As one way to deal with this concern, I collapse the four-point ordinal measure into a binary indicator of perceived discretion. This newly-coded variable takes a value of 1 if respondents perceive bureaucrats to make

99 Firm Investment Binary Binary Civil Proportional dummy, 1 = ﬁrm plans to invest in coming year Discretion Political Society Representation Measure Comp. (Comp.) (Comp.) Bureaucratic Discretion -0.773*** -0.577*** -0.863*** 1 = no discretion, 4 = high discretion (0.187) (0.152) (0.238) Bureaucratic Discretion -1.006*** dummy, 1 = high discretion (0.333) Frequent Changes to Laws 0.109 0.118 0.108 0.096 1 = no obstacle, 5 = very serious obstacle (0.110) (0.113) (0.112) (0.113) High Tax Rates -0.347*** -0.394*** -0.385*** -0.394*** 1 = no obstacle, 5 = very serious obstacle (0.119) (0.122) (0.121) (0.123) Regional Administration 0.431** 0.508*** 0.522*** 0.564*** 1 = poor job, 5 = excellent job (0.194) (0.193) (0.195) (0.195) Regional Courts -0.18 -0.128 -0.152 -0.151 1 = poor job, 5 = excellent job (0.147) (0.148) (0.148) (0.149) Regional Governor -0.328** -0.387** -0.374** -0.387** 1 = poor job, 5 = excellent job (0.184) (0.182) (0.183) (0.181) Access to Finance -0.013 -0.041 -0.033 -0.031 1 = no obstacle, 5 = very serious obstacle (0.082) (0.083) (0.083) (0.083) Labor Shortages -0.046 -0.035 -0.032 -0.015 1 = no obstacle, 5 = very serious obstacle (0.085) (0.086) (0.085) (0.085) Competitive Pressures 0.088 0.094 0.085 0.089 1 = no obstacle, 5 = very serious obstacle (0.084) (0.085) (0.085) (0.085) Privatized Firm 0.051 0.078 0.086 0.094 dummy, 1 = privatized, former SOE (0.272) (0.274) (0.274) (0.273) Annual Sales 0.466* 0.471** 0.459** 0.465** -1 = decreasing, 1 = increasing (0.182) (0.183) (0.182) (0.182) Firm Size 0.312*** 0.305*** 0.301*** 0.295*** number of employees (logged) (0.083) (0.083) (0.083) (0.085) Private Firm 0.829* 0.778* 0.786* 0.719 dummy, 1 = private ownership (0.434) (0.440) (0.441) (0.439) Constant -3.409** -1.712 -2.365* -1.999 (1.363) (1.514) (1.431) (1.470)

GDP per capita 0.484** 0.365 0.472* 0.489* in rubles per 1000 persons (logged) (0.246) (0.289) (0.246) (0.251) Political Competition 0.032 -0.688 -0.402 -0.671 various measures (0.052) (0.627) (0.333) (0.607) Competition × Discretion 0.398** 0.558* 0.304* 0.451 interaction (0.155) (0.312) (0.172) (0.302) Log-likelihood -238.531 -236.822 -236.893 -237.745 AIC 517.063 513.643 513.786 515.49 No. of Cases 403 403 403 403 Note: Survey data from Frye (2006). Region-level political variables come from the Moscow Carnegie Center’s Regional Monitoring Project. Data on GDP per capita from annual Rosstat publications. Coefficients from multilevel logistic regression with random coefficient for the discretion variable and random intercepts at the region level; standard errors in parentheses. Out of space concerns, random effects not reported. * p<0.10, ** p<0.05, *** p<0.01

Table 4.4: Robustness Check: Alternate Measures of Independent Variables 100 decisions independently of other government bodies and a 0 otherwise. Following the same logic, the second column in Table 4.4 replaces the political competition index, an ordinal variable with nine categories, with a dichotomous indicator of political competition that assigns a value of 1 to regions with scores above the sample mean and 0 to all others. The results in the table’s ﬁrst two columns shows that the initial ﬁndings are robust to these changes; whether using a binary measure for high perceived discretion or for high political competition, both variables continue to be statistically distinguishable from zero and signed in the expected directions. The predicted probability plots associated with Table 4.4 appear in Appendix B.

While the political competition measure singles out elections, openness, and pluralism as the most important indicators of regional political competition, other observable features of the political environment might capture latent levels of political competition equally well.10 The strength of civil society is a particularly good example. For some scholars, the degree to which societal groups organize in order to promote common goals is one of the hallmark of well-functioning political institutions

(Putnam, Leonardi & Nanetti 1993, North 1990). Moreover, this type of organization of societal and economic interests plays an implicit role in the theory’s diffuse monitoring mechanism – strong civil society can help by diffusing the costs of monitoring state agents and by giving businesses the organizational resources that they need to reduce uncertainty surrounding policy application or fight disputed regulatory

10As a side note on the political competition index, using the three components separately instead of the entire index creates no substantive changes in the outcomes, although the estimate on the interaction term is associated with higher statistical uncertainty in models using the index’s election component only. Associated tables and ﬁgures appear in Appendix B.

101 claims.11 Accordingly, the third column of Table 4.4 replaces the political competition index with a measure of civil society that also comes from the Moscow Carnegie

Regional Monitoring project. After mean-centering the scores for interpretation’s sake, the data range from a minimum value of -1 (Smolensk, Bashkortostan, and

Khabarovsk) up to a maximum value of 2 (Sverdlovsk). Table 4.4’s column shows that expectations that civil society should function similarly to political competition are borne out by the data – the coefficient on the interaction between civil society and bureaucratic discretion is positive and statistically significant, and the bureaucratic discretion variable keeps its negative and statistically significant coefficient.12

Predicted probability graphs conﬁrm that the estimates from this model function

very similarly to those presented above; the associated ﬁgure appears in Appendix

B. Not only do these ﬁndings for the civil society variable show that the chapter’s

main results do not depend on a speciﬁc measurement choice, but they also provide

greater conﬁdence in the theory’s proposed mechanism of diﬀuse monitoring.

Finally, the last column of Table 4.4 re-estimates the model including an alternate

measure of political competition that is based upon exogenous variation in a relevant

feature of institutional design: regional legislatures’ use of proportional representation

(PR) to allocate some portion of legislative seats.13 Compared with institutions that rely exclusively upon ﬁrst-past-the-post electoral rules, proportion representation

11In fact, as the above description implies, this particular analysis might just as easily be interpreted as a test of one of the theory’s proposed causal mechanisms. For the goals of this chapter, however, I treat the measure of civil society as an alternative indicator of the latent levels of competitiveness within ﬁrms’ political environment. Chapter 5 deals exclusively with testing the proposed mechanisms by which political competition reduces the eﬀects of regulatory uncertainty.

12Besides the strength of civil society, free media might be yet another feature of the political environment that taps a dimension of latent levels of political competition. Additional analyses show that results are robust to the use of scores for regional media freedoms, as well.

13The analyses’ reliance upon the subjective scores by regional experts might raise concerns that raters’ own subjective biases may be endogenous to the quality of regional investment climates or perhaps otherwise unduly influence the empirical findings. Scholars have voiced related concerns 102 rules can be viewed as enhancing political competition by creating opportunities for a greater diversity of parties to win seats in the legislature. The presence of these additional voices in parliament is likely to enhance diffuse monitoring of bureaucratic agents and provide greater opportunities for businesses to seek redress should disputes arise.

Substantively, the estimates in the table’s final column are consistent with the conditional hypothesis; the bureaucratic discretion variable reports a negative, statistically significant coefficient and the interaction term is positive, albeit with greater levels of statistical uncertainty than tolerated by conventional significance levels

(p=0.13).Given the nonlinearity of the logit model, however, statistical signiﬁcance is not necessarily a prerequisite for variables to interact meaningfully in inﬂuencing the probability of observing the outcome of interest (Berry, DeMeritt & Esarey 2010).

For this reason, Figure 4.3 below plots the differences in the predicted probabilities of future investment associated with opposing perceptions of bureaucratic discretion for both the PR and non-PR institutional settings. As anticipated, perceptions of high discretion in the non-PR context correlate with a distinctly lower predicted probabilities of firms’ intent to invest. In regions that include a PR seating rule, the difference between high- and low-discretion profiles is statistically indistinguishable.

Alternate Measures of Investment

The previous set of analysis demonstrate the ﬁndings’ robustness to alternative measures of the key independent variables. As a ﬁnal test of the conditional hypothesis, I investigate whether the models used to test the theory’s empirical predictions about

ﬁxed capital investment are useful in predicting related ﬁrm behavior. In this set of

with regards the subjectivity of cross-national democracy scores (Bollen 1993, Munck & Verkuilen 2002, Treier & Jackman 2008) and governance indicators (Kurtz & Shrank 2007).

103 ● ● ●●●●●●● ● ●●●●● ● ●●● ●●● ●●● ● ●●● ●●●●●●●●●●● ●●●●● ● ●●●●● ●●●●● ●●● ●●●●● ●● ●● ● ●●● ● ●● ● ● ●●●● ●●●●●● ●●●● ●●●● ●●● ●●●●●● ● ● ●●●●● ●● ● ●●●●●●●●●●●●●●●● ●●● 1.0 ●●● ● ●● ● ● ●● ●●● ●●●●● ●● ●●●●●●● ● ●●●●● ●● ● ●●●●●● ● ●●● ●●● ● Low● Discretion●● ●●●●●●●●●●●● High Discretion 0.9

0.8 ● ● 0.7

0.6

0.5 ●

0.4

0.3 ● 0.2

0.1 ● ● ●● ●●●● ● ●●● ●●●●●●● ●●●● ●●●●●●●● ●●● ●●● ●●●●●●●●●●●●●●●●● ● ●● ●●●●●●●●●● ● ●●●● ●●●●●●●●●●●●●●●● ● ● ●●●●●●● ●● ●●●●●●● ●●● ●●●●●● ●●●●●●●●●● ●●●●●●●●● ●●●● ●●●●●●●●● ●●●●●●● ●●● ●● ●●● ● ●●●● ●●●●●●●●●●●●● 0.0 ●●●●●●●●●●●● ●●● ●●●●●● ●●●●●●●●●●●● Predicted Probability of Firms' Intent to Invest ●●● ●●●●● ●●● ● ●●●●● ●●●●●●●●●● ●●●●●●●●●● ● ●●●●●●● ●●●●●●● ●● ●●●●● ● No PR Rule Includes PR

Institutional Design of Regional Legislature

Note: Firm-level survey data from Frye (2006), data on regional institutional design for the eleven sampled regions coded by author. Dots represent the predicted probability that a hypothetical ﬁrm would invest, given the level of regional political competition; discretion is manipulated from minimum (1) to maximum (4) while holding all variables at their median value. Solid lines represent 95% conﬁdence intervals obtained via simulation in R.

Figure 4.3: Predicted Probability Plots: Proportional Representation as Measure of Political Competition

104 analyses, I drop the prospective measure of firm investment in fixed assets in favor of three separate dichotomous dependent variables that capture firm managers’ recent engagement in costly investment-related behavior: undertaking programs to improve workers’ skills, extending credit to buyers, and conducting marketing studies. Ad- mittedly, these variables measure a diverse set of activities, but at their heart, they share with investment the same calculus of paying costs now in order to reap returns in the future. Thus, we would expect these new models to produce statistical results that look much like the chapter’s previous findings. Specifically, the conditional theory predicts that, in low-competition regions, higher bureaucratic discretion should correlate negatively with these activities due to firms’ increased uncertainty about policy application; as regional political competition increases, however, the theory anticipates that the negative effects of bureaucratic discretion should attenuate.

The results of the additional tests appear in Table 4.5.14 The table’s first column takes as its dependent variable a binary indicator for whether or not firms have undertaken programs to improve their workers’ skills in the past two years. Because such programs incur real costs for firms, both in terms of direct monetary payments as well as in opportunity costs, managers who perceive that their expected gains are threat- ened by uncertain regulatory obstacles should be less likely to engage in this activity unless that uncertainty is mitigated by the broader institutional environment. The statistical results support this claim, reporting a negative coefficient estimate for the bureaucratic discretion variable and a positively-signed coefficient for the interaction term. Figure 4.4 provides a graphical representation of these same results. Holding all other variables at their sample means, a hypothetical firm manager in a region with very low political competition who perceives bureaucrats to have high discretion has a

14As before, all models report coefficient estimates from multilevel logistic regressions using a random-intercept, random-coefficient specification. Because the tests’ primary goal is comparing with previous analyses, I focus the discussion completely upon the main variables of interest.

105 Investment-Related Activities Worker Extending Marketing dummy, 1 = speciﬁed activity within the past 2 yrs. Training Credit Study

Bureaucratic Discretion -0.243 -0.383** -0.309 1 = no discretion, 4 = high discretion (0.203) (0.158) (0.253) Frequent Changes to Laws 0.075 -0.148 0.139 1 = no obstacle, 5 = very serious obstacle (0.158) (0.104) (0.132) High Tax Rates 0.012 -0.131 0.026 1 = no obstacle, 5 = very serious obstacle (0.165) (0.112) (0.140) Regional Administration -0.385 -0.121 -0.062 1 = poor job, 5 = excellent job (0.276) (0.182) (0.239) Regional Courts 0.208 0.033 0.443** 1 = poor job, 5 = excellent job (0.221) (0.143) (0.186) Regional Governor 0.448* -0.138 -0.011 1 = poor job, 5 = excellent job (0.268) (0.178) (0.234) Labor Shortages -0.071 0.014 -0.065 1 = no obstacle, 5 = very serious obstacle (0.137) (0.085) (0.113) Competitive Pressures 0.006 0.173** 0.070 1 = no obstacle, 5 = very serious obstacle (0.127) (0.082) (0.105) Privatized Firm 0.108 -0.562** -0.007 dummy, 1 = privatized, former SOE (0.391) (0.268) (0.350) Annual Sales 0.556** 0.040 0.238 -1 = decreasing, 1 = increasing (0.236) (0.169) (0.208) Firm Size 0.490*** 0.215*** 0.205* number of employees (logged) (0.141) (0.081) (0.105) Private Firm -0.144 1.189*** 1.102** dummy, 1 = private ownership (0.652) (0.431) (0.483) Constant -2.529 -0.956 -1.887 (2.436) (1.578) (2.066)

GDP per capita 0.428 0.244 -0.000 in constant 2000 rubles per 1000 persons (logged) (0.505) (0.319) (0.427) Regional Political Competition -0.058 -0.276** -0.376* index, -3 = uncompetitive, 5 = highly competitive (0.208) (0.128) (0.213) Competition × Discretion 0.182* 0.195*** 0.302*** interaction (0.099) (0.069) (0.117) Log-likelihood -124.661 -254.503 -174.803 No. of Cases 415 411 415 Note: Survey data from Frye (2006). Region-level political variables come from the Moscow Carnegie Center’s Regional Monitoring Project. Data on GDP per capita from annual Rosstat publications. Coefficients from multilevel logistic regression with random coefficient for the discretion variable and random intercepts at the region level; standard errors in parentheses. Out of space concerns, random effects not reported. * p<0.10, ** p<0.05, *** p<0.01

Table 4.5: Robustness Check: Investment-Related Dependent Variables

106 predicted probability of investing in workers’ skills that is 45 percentage points lower than the low-discretion counterpart (45% versus 90%). This gap quickly disappears as the level of political competition approaches the sample’s mean levels of political competition. In harmony with the predictions of the conditional theory, Figure 4.4 shows that in the absence of a supportive institutional environment, bureaucratic discretion is associated with reduced incentives for ﬁrms to invest in improving the skills of their labor force.

Following the model of investment in workers’ skills, the second column in Table

4.5 investigates the effects of bureaucratic discretion on firms’ willingness to supply their goods or services to customers on credit. Within the statistical model, the credit- lending measure takes a value of 1 if the firm manager reports to have extended credit to buyers in the last two years and a 0 if the firm has not. Because firms’ decision to extend credit entails much the same dynamics as capital investment – immediate production costs with payment and interest materializing in the future – the logic of the conditional theory expects that, in low-competition regions, bureaucratic discretion should be negatively correlated with firms’ credit-lending behavior. The findings in Table 4.5 are consistent with this prediction. Once again, we see that marginal relationship between perceptions of higher discretion and the dependent variable has two components: a large negative term associated with the bureaucratic discretion variable and a smaller positive term that represents the mediating effects of political competition. As evidenced by the predicted probability plots in Figure 4.5, the expectations of the conditional theory are borne out in full by the empirical results of this test.

Finally, we turn to the last column of Table 4.5 and its corresponding model of

ﬁrms’ decision to conduct extensive market research over the past two years. Much like the other two alternate dependent variables, market research can be considered

107 ●●●● ●● ●●●●● ●● ●●● ●●● ●●● ● ●●●● ● ●●●● ●● ●●● ●● ● ●●●● ● ● ●● ● ● ●●● ●●● ●●●● ●●●●● ● ●● ●●● ●●● ●●● ●●● ●● ●●● ●● ●● 1.0 ●● ●●● ●●●●● ●● ●● ●●● ●●●● ●●●● ●● ●●●●●● ●●● ●●● ●●● ●●● ●●●● ●● ●●●●● ●●● ●●● ●● ●● 0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2 Low Discretion High Discretion 0.1

Predicted Probability of Skills Training for Workers for Predicted Probability of Skills Training ● ●●● ● ●●● ●●●●● ●●●●●● ●● ●●● ● ● ●● ●●●●● ●●●● ●●●●●● ●●● ● ● ●●●● ●●●● ●●●●● ●● ●●●●●● ●●●● ●● ●● ●● ●●●● ●● ●●●●●● ●●●●● ● ●●●● ● 0.0 ●●●● ●●●● ●●●●● ● ●●● ●● ●●●● ●●● ●●●● ●●●●● ●●●● ●●● ●●● ●●●● ●●●● ● ●●●●●● ●●● ●● ●● ●●● ●●●●● ●●● ●●●●●● ●●●● ●●● ●●● ●●

Low Medium High

Regional Political Competition

Note: Firm-level survey data from Frye (2006), region-level political variables for the eleven sampled regions come from the Moscow Carnegie Center and economic data from Rosstat. Bold lines represent the predicted probability that a hypothetical ﬁrm would invest in skills training for its workers, given the level of regional political competition; discretion is manipulated from minimum (1) to maximum (4) while holding all variables at their median value. Small dashed lines represent 90% conﬁdence intervals obtained via simulation in R.

Figure 4.4: Predicted Probability Plot: Investing in Labor Skills (Alternate DV)

108 ● ● ● ●● ● ● ● ●● ●● ●●●●● ● ●● ●● ● ●●● ●●●● ●●●●●● ●●● ● ●● ● ●● ●● ●●●●● ●●● ● ●● ●●● ●●● ●● ●●●●● ●● ●●●● ●●● ● 1.0 ●●●●● ●●● ●●●●● ● ● ●● ●● ● ●●● ●● ●●●●● ● ●●●● ●●● ●●●● ●●● ● ●●●●● ● ●● ●●●● ●●●●● ●●●● ●● ●●●● ●●● ●●● ●● ●●● 0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2 Low Discretion High Discretion 0.1

Predicted Probability of Extending Credit to Buyers ● ● ●●● ●● ●●● ●●●●● ●●● ●●●●●● ●●●●● ● ●●● ●● ●●●● ●● ●●●●●● ●● ●●● ● ●● ●●● ●●● ●●●●●● ●●● ● ●●● ●●● ●●● ●●●● ●●●●●●● ●●●● ●●●● ● ●●● 0.0 ●●●● ● ●●●●● ●●●● ● ●● ●● ●●●●● ●● ●●●●● ●●● ●●●● ●● ●● ●●● ●●●● ●●●●●● ●●●● ● ●● ●●● ●●●● ●●● ●●●●●● ●●● ●●● ●●●● ●●●

Low Medium High

Regional Political Competition

Note: Firm-level survey data from Frye (2006), region-level political variables for the eleven sampled regions come from the Moscow Carnegie Center and economic data from Rosstat. Bold lines represent the predicted probability that a hypothetical ﬁrm would allow customers to buy on credit, given the level of regional political competition; discretion is manipulated from minimum (1) to maximum (4) while holding all variables at their median value. Small dashed lines represent 90% conﬁdence intervals obtained via simulation in R.

Figure 4.5: Predicted Probability Plot: Extending Credit to Buyers (Alternate DV)

an alternative measure to capital investment in the sense that it is often very costly to firms in the short-term but has the potential to create valuable and lucrative business opportunities for firms in the future. As such, a general expectation of the theoretical argument is that, if unmitigated by the factors associated with an open and competitive political environment, bureaucratic discretion should be a source of uncertainty that can deter economic actors from engaging in this type of behavior. 109 The statistical results in Table 4.5 and the corresponding graph in Figure 4.6 suggest this to be true. In regions with very little political competition, the estimated effects of high discretion are striking: holding all other variables at their sample medians, the predicted probability of having undertaken marketing research is almost 90% for a firm that perceives bureaucrats as having no independent decision-making ability whereas the predicted probability is only 18% for an identical firm that perceives bureaucrats as having extensive independence in decision-making. At the other end of the competitiveness spectrum, we observe an inversion of these positions – although the predicted probability of the low-discretion profile essentially remains constant, the high-discretion profile is associated with a predicted probability of marketing research that is close to 1 in the highest-competition region.

The empirical analyses presented above as well as in Appendix B show that the

findings in support of Hypothesis 2 hold up under a variety of conditions. The findings do not depend, for example, on specific codings of the bureaucratic discretion or political competition measures, nor are they dependent upon the subjective measure of political competition used in the main analyses. As an additional way to bolster confidence in both the findings’ robustness as well as the general flexibility of the theory’s logic, I have also shown that the model successfully explains other dependent variables that should be closely related to firms’ decisions over capital investments.

Taken together, the results of this chapter’s additional analyses provide strong support for the theory’s conditional argument.

Conclusion

This chapter tests the theory’s hypothesis of the conditional relationship between bureaucratic discretion and investment by combining enterprise survey data with region- level information on the competitiveness of regions’ political environment. Consistent

110 ● ● ● ● ●●●●● ●●● ● ●● ●● ●●● ● ●●●●● ● ●●●● ●●●● ● ●●● ●●●●●● ● ●●● ●● ●●● ●●● ●● ●●●● ●●● ●●● ●●● ●●●● ●●●●● ●●●● ●●●●●●● ● ●●●● ● ● 1.0 ●●● ●● ●●●●●● ●●● ●●● ●●● ●● ●● ● ●●● ● ●●●● ● ●●●● ●●● ●● ●●●● ● ● ● ●●●● ●●● 0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2 Low Discretion High Discretion 0.1 ●●● ● ●●●●●● ●● ●●●●●● ●●●● ●●● ●● ●● Predicted Probability of Conducting Marketing Study Predicted Probability of Conducting Marketing ● ●● ● ● ● ●●●● ●●● ●●●●●● ●●●● ● ● ●●●●● ●●●● ●●● ● ●●●●● ●●● ●●●● ● ●● ●●●●●● ●●● ●●●●●● ●●● ●●● ●● ●● 0.0 ●●●● ●●● ●●●●●● ●● ● ●●● ●● ●●● ●● ●●●●●●● ●●● ●● ●●● ●●● ●●●●● ●●●● ●●●●●● ●●● ●●●● ●●● ●●● ●●●●● ●●●●● ●●●●●● ●● ●●● ● ●●●

Low Medium High

Regional Political Competition

Note: Firm-level survey data from Frye (2006), region-level political variables for the eleven sampled regions come from the Moscow Carnegie Center and economic data from Rosstat. Bold lines represent the predicted probability that a hypothetical ﬁrm would invest in extensive marketing research, given the level of regional political competition; discretion is manipulated from minimum (1) to maximum (4) while holding all variables at their median value. Small dashed lines represent 90% conﬁdence intervals obtained via simulation in R.

Figure 4.6: Predicted Probability Plot: Conducting Marketing Study (Alternate DV)

with the theory’s empirical predictions, in the restricted political conditions that prevail for a majority of the respondents in the survey, ﬁrms’ perceptions of bureaucratic discretion are associated with substantial decreases in the probability that ﬁrms will choose to invest in the near future. High-competition political environments mitigate this negative relationship, supporting the theoretical claim that political competition

111 reduces regulatory uncertainty associated with bureaucrats’ discretion in applying and interpreting laws.

The results of this chapter’s analyses have important implications for political economy. For literatures where bureaucracy already plays a central role, this paper adds nuance to our understanding of how bureaucracy influences economic behavior. I offer post-Soviet studies of bureaucracy as an example. Within the literature on post-communist development, bureaucratic institutions have been characterized as “grabbing hands” that drive entrepreneurs into the unofficial economy

(Frye & Zhuravskaya 2000) and force businesses to pool resources in common defense

(Duvanova 2007, Pyle 2009). At the same time, others have defended bureaucrats as

“helping hands” that provide infrastructure and support to companies making their way in a new economic system (Brown, Searle & Gehlbach 2009). This article highlights a unifying framework for the opposing camps — bureaucrats’ ability to shape the predictability of ﬁrms’ regulatory environment — and encourages us look for variation in the broader institutional context that will help explain when bureaucratic discretion should mitigate or magnify economic uncertainty.

Moreover, in ﬁnding that the negative relationship between discretion and investment is strongest in regions where leaders are least constrained by political competition, this research produces a result that is somewhat counterintuitive from the standpoint of conventional arguments: delegation to independent agents may be counterproductive in precisely those places where the literature expects it to help most.

Rather, the ﬁndings suggest the beneﬁts of extensive delegation may only begin to outweigh the negatives where surrounding institutions can help reduce investors’ uncertainty about how regulatory bureaucrats will use their discretion.

In the next chapter, I build upon these ﬁndings by exploring additional observable implications generated by the theory. If politically-competitive environments mitigate

112 the negative effects of bureaucratic discretion, then we should be able to observe systematic differences across political environments in the manner that firms respond to bureaucratic discretion. Chapter 4 shows that this is the case: within regions with high levels of political competition, firms that perceive bureaucratic discretion seek to mitigate their losses from regulatory risks by adopting strategies that make use of the political channels provided by their surrounding environment.

113 CHAPTER 5

HOW DOES POLITICAL COMPETITION MAKE

BUREAUCRATIC DISCRETION BETTER FOR

INVESTORS? TESTING THE MECHANISMS

Introduction

In testing the predictions of the dissertation’s theoretical framework, previous chapters have established that bureaucratic discretion is negatively associated with ﬁrms’ willingness to engage in investment behavior, and that this negative relationship is strongest within political environments characterized by low levels of political competition. This chapter continues the empirical analysis by testing observable implications of the theory’s proposed causal mechanisms.

Why should political competition reduce the eﬀects of investors’ uncertainty about policy application at the hands of discretionary bureaucrats? Within Chapter 2’s theoretical discussion, I supplied two complementary mechanisms. First, competitive politics reduce the eﬀects of investors’ uncertainty about policy application by making politicians more attentive to constituents, including those in the business community. Given electorally-sensitive leaders’ increased incentives to appease their constituents, economic actors interested in changing and clarifying regulatory rules should have more opportunities to do so within politically-competitive environments.

114 By providing more open access to policymakers and the policy process, political competition should increase business actors’ incentives to use political channels to address problems arising from bureaucratic agents’ discretion. Such logic yields empirically- testable predictions about ﬁrms’ attitudes and behaviors. For example, as political competition increases, ﬁrms should be more likely to engage in lobbying-related activities, particularly as part of their strategy for making bureaucratic discretion more predictable.

Second, political competition reduces regulatory uncertainty by encouraging political principals to watch their bureaucratic agents more carefully and allowing additional parties to assist in monitoring state agents. More widespread monitoring helps economic actors know better what to expect out of their high-discretion environment and reduces the chances that bureaucrats will engage in behavior that conﬂicts with principals’ goals. By this logic, improved monitoring leads to fewer problems with discretionary bureaucrats in politically-competitive environments. If true, we should observe that discretion and disputes with regulatory oﬃcials are more likely to be correlated in low-competition regions than in high-competition regions.

This chapter’s goal is to test the observable implications that follow from the argument’s causal mechanisms. Using the 2005 Frye survey of enterprise managers in

Russia alongside data on regions’ political competition, I investigate whether firms’ attitudes and reported behaviors vary systematically across high- and low-competition regions in ways that are consistent with the theory’s implied expectations. I divide this chapter into two main sections. In the first section, I examine the proposed link between political competition and attentive political leaders. I find that in regions with high levels of political competition, firms display greater approval of certain

115 elected leaders and are more likely than their low-competition counterparts to respond to bureaucratic discretion by seeking political inﬂuence, either through organized lobbying or informal government contacts. In contrast, evidence indicates that

firms in low-competition regions are more likely to respond to regulatory disputes by offering bribes or simply acquiescing to agents’ dubious claims. The second section provides an indirect investigation of the claim that better monitoring of state agents in politically-competitive environments reduces firms’ uncertainty. Examining whether discretionary bureaucrats are “better behaved” in regions with high political competition, the analyses indicate that in high-competition regions discretion is associated with high approval of administrators, fewer problems with bribery, and fewer disputes between firms and government agencies. After presenting the results in these sections, the chapter concludes.

Political Competition & Responsive Politicians

Political competition reduces the effects of economic actors’ uncertainty about bureaucratic discretion through multiple pathways. My first goal is to shed light on whether high levels of political competition prompt governments to be more open and responsive to the concerns of business actors. Given the survey’s lack of explicit questions about politicians’ responsiveness, I examine instead responses within the survey that indicate the degree to which firms perceive their governments to be ac- cessible and open to influence. I begin with a brief examination of variation across political environments in firms’ approval of various political leaders. Next, based upon the logic that firms’ behavior should be adapted to make use of opportunities offered by responsive leaders, I investigate how firms’ choice to engage with the state via lobbying and organized interests varies across political environments. Taken together, the results of this section offer strong support for the proposed mechanism.

116 Approval of Elected Leaders

Firms’ attitudes towards their elected oﬃcials can potentially reveal evidence of a con-

nection between political competition and more responsive political leaders. If politi-

cians in high-competition regions do attend more closely to constituents, including

those within the business community, one possibility is that this higher attentive-

ness could be detected in more positive attitudes towards elected oﬃcials among ﬁrm

managers in high-competition regions. Table 5.1 uses respondents’ assessments of

leaders in various political oﬃces to compare the average approval for leaders in high-

competition regions versus those in low-competition regions.1 The ﬁrst two variables

report respondents’ assessment of regional politicians, the governor and the regional

legislature, while the second pair of variables report assessment of local elected oﬃ-

cials, the mayor and city legislature.

Do ﬁrms in high-competition regions hold more positive views about their elected

oﬃcials? The results in Table 5.1 provide qualiﬁed support for that proposition.

At the region level, a diﬀerence of means test for high- and low-competition groups

fails to reject the null hypothesis that average assessments of regional politicians

are the same across across groups. At the city level of government, however, we do

see diﬀerences that are statistically distinct: compared to their counterparts in low-

competition regions, ﬁrms in high-competition regions have more positive evaluations

of their mayors and city legislatures (p = 0.09 and p = 0.00, respectively).2 Thus, although far from oﬀering any conclusive evidence, the ﬁndings in Table 5.1 yield

1Survey respondents answered the following question: “Based upon your own personal experience or the experiences of business colleagues, what is your assessment of your [name of political oﬃce] in your region [city]?” Answers range from “very poor” to “very good” on a ﬁve-point ordinal scale.

2A Kolmogorov-Smirnov test comparing the distributions of these variables provides a similar intuition. Distributions are statistically distinct at the city-level for mayors and city legislatures (p = 0.04 and p = 0.0001, respectively), but not at the region-level for governors or region legislatures (p = 0.74 and p = 0.99, respectively). 117 Political Competition Variable Low High df

Regional Governor 3.13 3.10 491.55 1 = very poor, 5 = very good Regional Duma 2.82 2.86 427.02 1 = very poor, 5 = very good Mayor 2.93 3.07 350.52 1 = very poor, 5 = very good City Duma 2.80 3.09 368.60 1 = very poor, 5 = very good

Note: Survey data from Frye (2006). Bold ﬁgures indicate a p- value of p < 0.10 rejecting the hypothesis test that the two means are equal.

Table 5.1: Assessment of Elected Oﬃcials: Diﬀerence of Means

two observations. First, at a basic level, at least some political leaders in high- competition regions are better on average at garnering favor with economic actors.

Second, to the extent that politicians who acknowledge constituents’ concerns receive higher approval ratings, the ﬁndings in Table 5.1 suggest that locally-elected leaders are more responsive than their peers in low-competition environments.

Obviously, myriad factors can affect the public’s assessment of politicians’ performance (MacKuen, Erickson & Stimson 1992, Enns 2007). Under the very best of circumstances, respondents’ subjective ratings of elected officials are likely to be a noisy signal of whether the political environment provides businesses with more opportunities to raise concerns about the application and interpretation of laws by discretionary bureaucrats. As such, I turn next to examining firms’ reported behavior for greater insight into whether high-competition environments increase businesses’ voice with government leaders.

118 Lobbying & Government Connections

The theoretical framework implies that discretion generates less uncertainty in regions with high political competition because firms there have both more opportunities to shape the policy-making process as well as better institutional channels for mitigating problems that might arise. If that is true, then we would expect to observe that, in response to perceptions of high bureaucratic discretion, firms should adopt certain types of strategies more frequently in politically-competitive environments than in non-competitive environments. In this section, I investigate this observable implication by looking for empirical evidence that, if faced with a highly-discretionary regulatory environment, firms in high-competition regions seek to engage policy-makers via lobbying and the cultivation of government ties. The decision to examine firms’ lobbying behavior is based upon the premise that seeking to effect change on their uncertain regulatory environment via organized interest groups or government connections is less likely to be an effectual choice for firms where political leaders are unresponsive to those lobbying efforts. Thus, as political competition increases, firms should be more likely to engage in lobbying activities, particularly as part of their strategy for making bureaucratic discretion more predictable.

To test this proposed mechanism, Table 5.2 compares the average response to a set of lobbying-related questions for firms in high-competition versus low-competition regions. The responses themselves are all dichotomous indicators with a value of 1 representing the presence of a specified, lobbying related activity that will be discussed below. For each of the four chosen questions, I compare mean responses in three different samples – the full sample, a subset of only firms that perceive bureaucrats to have low discretion, and the subset of firms that perceive bureaucrats to have high discretion. Earlier in the dissertation, I argued that bureaucratic discretion creates uncertainty for business actors about how regulatory agents will interpret and apply

119 the law. This argument implies that, of all firms in the sample, firms in the high- discretion subset should have the greatest motivation to use lobbying and government connections to seek policy changes that would make regulatory rules more transparent and less open to interpretation. Given that we might expect all firms in this category to want to pursue lobbying as a strategy for making their regulatory environment more predictable, any differences we observe across groups among high-discretion responses sends strong signals about whether or not the political environment makes pursuing this strategy worthwhile. I now introduce and discuss the results for each variable in turn.

The first two items come from a section on firms’ involvement in business associations. As part of the survey, respondents who identified themselves as members of a business association were queried on the benefits of joining such an organization.3

Respondents were able to choose multiple answers from among a set of potential beneﬁts, including an option for lobbying, or as worded in the survey: “the opportunity to defend business interests in legislative and executive jurisdictions.” Of the

245 ﬁrms responding to the question, roughly half (n=129) picked this option as one of the beneﬁts of association membership. The proposed mechanism implies that

firms in high-competition regions should disproportionately regard lobbying as a useful part of associational membership because political competition in their localities encourages politicians to respond to organized constituencies. The analyses confirm the prediction: 61% of responding firms in high-competition regions see lobbying as

3The question’s exact wording is as follows: “In your opinion, what does membership in such associations provide your firm?” Respondents were allowed to choose all applicable answers from eleven different options, including the two used in the analyses above. The options were less relevant for the analysis at hand, including alternatives such as “the opportunity to find new partners in Russia,” “the opportunity to obtain information about the reliability of business partners,” and “access to services.”

120 Political Competition Variable Low High df

Beneﬁts of Associations: Lobbying dummy, 1 = agrees full sample 0.49 0.61 150.40 subset: low discretion 0.50 0.59 114.87 subset: high discretion 0.39 0.82 20.90

Beneﬁts of Associations: Help Set Rules dummy, 1 = agrees full sample 0.24 0.31 136.54 subset: low discretion 0.26 0.24 116.51 subset: high discretion 0.21 0.64 14.79

Attempts to Lobby Governor dummy, 1 = seeks to inﬂuence governor full sample 0.05 0.11 383.79 subset: low discretion 0.07 0.10 361.34 subset: high discretion 0.03 0.19 33.75

Management with State Experience dummy, 1 = former SOE manager full sample 0.61 0.65 509.97 subset: low discretion 0.68 0.61 391.06 subset: high discretion 0.52 0.81 62.88

Note: Survey data from Frye (2006). Bold ﬁgures indicate a p-value of p < 0.10 rejecting the hypothesis test that the two means are equal.

Table 5.2: Lobbying Eﬀorts: Diﬀerence of Means

121 a beneﬁt to association membership as opposed to only 49% in low-competition re-

gions.4 Among ﬁrms that perceive bureaucrats to have high discretion – the most

theoretically-relevant subset of ﬁrms – the diﬀerence between the high- and low-

competition groups is even larger (82% versus 39%, p=0.01).

The second variable in Table 5.2 reports ﬁrm managers’ responses concerning

associations’ involvement in setting industry rules and standards.5 Such opportunities range from associations’ participation in government-business roundtables and reform-oriented advisory boards to the adoption and enforcement of industry standards by associations acting parallel to or in place of the state. Dissatisfaction with the application and enforcement of regulatory rules provides a strong motive for businesses to be involved in establishing clearer guidelines, yet such opportunities are limited and depend heavily upon state actors’ willingness to invite the business community into policy discussions or legitimize their self-regulation effort.6 Thus, the proposed mechanism implies that, as a strategy for dealing with regulatory uncertainty and discretionary rules, this feature of associational membership should be relatively more attractive to firms in high-competition regions, where politicians are expected to be more responsive to businesses’ efforts to take an assertive role in shaping industry guidelines. Once again, according to the data, this is the case for firms

4The p-value for the diﬀerence in means is p>0.07.

5The survey presented this potential beneﬁt of associational membership to respondents as “the opportunity to participate in establishing new moral-ethical norms in the business sphere.”

6Indeed, during my ﬁeldwork in Russia, several business associations brought up varying degrees of self-regulation as a strategy for reducing frictions with regulatory bureaucrats. In addition to the usual obstacles to collective action, association leaders mention diﬃculties in obtaining the requisite permission from the government as one of the biggest obstacles to their grand ambitions. As a result, only a very small number of the business associations that I interviewed could boast of having successfully won wide autonomy from the government.

122 with the high-discretion subset. In regions with low political competition, the opportunity to set business standards is seen as a benefit by only 21% of firms. Among such firms in high-competition regions, however, the number is three times as high

– 64% perceive participation in rule-making a beneﬁt to membership in a business association.7

We observe similar patterns if we expand the analysis to include those firms that do not belong to business associations. Table 5.2 summarizes by group whether or not firms report to lobbying their governor as a way to influence the content of news and statutes.8 Looking at lobbying patterns across high- and low-competition groups, do we see evidence that political competition encourages businesses to approach their regional executive? For all three subsets, firms in high-competition regions claim to lobby their governor more than in low-competition regions, although the difference is not statistically significant in the low-discretion subset. Among firms that perceive regional bureaucrats to have high discretion, 19% of firms in high-competition regions reportedly lobby their governor, yet only 3% of such firms in the low-competition group pursue a similar strategy.9 Here again, the results support the proposed mechanism. Where bureaucrats’ discretion in applying and interpreting laws can potentially create obstacles for business, those firms in high-competition regions appear to regard their political leaders as a resource for mitigating their difficulties in way that

ﬁrms in low-competition regions do not.

7With a p-value of 0.02, the diﬀerence in means across these two groups is statistically signiﬁcant.

8Specifically, the survey asks: “If your firm attempts to influence the content of new laws or statutory acts which are important for your business, in what ways does it go about this?” In this analysis, I focus on the following option: “consultations with the governor (republic president, mayor of Moscow).” Forty-eight firm managers respond affirmatively to this option, or roughly 7% of the full sample.

9The p-value associated with this diﬀerence of means test is p=0.04.

123 Finally, because questions about lobbying activities may be so sensitive that some

ﬁrms might not answer truthfully, I also examine an indirect measure of government

connections – ﬁrm management’s past leadership experience in a state-owned enter-

prise.10 Existing research in Russia has shown that state-owned enterprises (SOE)

and ﬁrms with historical connections to the state (i.e. those privatized ﬁrms that were

once state-owned) receive better treatment from government oﬃcials (Gehlbach 2008).

Earlier studies have also linked such ﬁrms to successful lobbying within Russia’s re-

gions (Frye 2002a). Given historically-close relationships between government oﬃcials

and state-owned enterprises, ﬁrms grappling with an uncertain regulatory environ-

ment may see employing a former SOE manager as an eﬃcient strategy for acquiring

lobbying contacts within the regional government. The data are consistent with this

scenario. Within the high-discretion subset of ﬁrms, a full 81% of ﬁrms in politi-

cally competitive regions have among their senior management a former SOE leader,

as compared with only 52% ﬁrms in the group of regions with less political compe-

tition (p<0.01). The proposed mechanism provides a ready-made explanation: for

ﬁrms seeking to reduce the uncertainty of their regulatory environment, the potential government contacts that come with employing a former SOE manager are more beneﬁcial in regions where political competition encourages leaders to respond when these former associates reach out to them.

Table 5.3 shows how these results fare in more rigorous analyses that can control for potentially confounding factors. In order to maximize the number of observations,

I use as dependent variables only those questions that were asked of all respondents – those about lobbying governors and hiring former SOE managers. In addition to the measures of bureaucratic discretion and regional political competition, I control for

10Respondents were asked the following question: “At any time in the past, have you worked as a leader at any level of a state enterprise?” Sixty-two percent of ﬁrm managers responded positively.

124 additional attitudinal measures that might correlate with both firms’ propensity to engage in lobbying activity and their perceptions of bureaucratic discretion, including: perceptions of bureaucratic corruption, problems with frequently changing laws and high tax rates, and firms’ separate assessments of the governor, regional administration and economic courts. I also control for potentially influential firm characteristics with indicators for economic sector and status as a private firm, as well as a measure to control for firm size. As elsewhere, I estimate the multilevel logistic regressions use a random-intercept, random-coefficient specification to account for unobserved heterogeneity that might vary across region.

Looking at results in the first column, the coefficient on the interaction term between political competition and perceived discretion is signed according to predictions, but fails to obtain conventional levels of statistical significance (p = 0.19). According to this model, then, we cannot reject the null hypothesis that political competition does not condition the relationship between perceived discretion and firms’ reported lobbying of their governor.11 Turning to the indirect measures of firms’ government contacts, the second model in Table 5.3 supports the contention that firms facing high bureaucratic discretion are more likely to try to cultivate better relations with regional governments where political competition is high. The interaction between the discretion and competition variables is positive and statistically significant (p=0.02), suggesting that as political competition increases, the correlation between perceptions of bureaucratic discretion and the practice of employing former SOE leaders in firm

11These null ﬁndings depend upon the keeping the perceived discretion variable’s original four-point scale. Re-estimating the model using a dichotomous coding of perceived bureaucratic discretion yields results that support a conditional relationship. Although these conﬂicting results data do not provide us with clear answers on this front, additional analyses that drop the interactive terms do provide evidence supporting a positive relationship between political competition and lobbying.

125 Lobby Manager w/ Governor State Ties Bureaucratic Discretion -0.132 -0.015 1 = no discretion, 4 = high discretion (0.294) (0.160) Bureaucratic Corruption -0.982** -0.077 dummy, 1 = perceived as corrupt (0.461) (0.249) Frequent Changes to Laws -0.044 0.308*** 1 = no obstacle, 5 = very serious obstacle (0.203) (0.110) High Tax Rates -0.464** -0.039 1 = no obstacle, 5 = very serious obstacle (0.200) (0.116) Regional Administration -0.355 -0.269 1 = poor job, 5 = excellent job (0.354) (0.199) Regional Governor 0.457 -0.027 1 = poor job, 5 = excellent job (0.362) (0.193) Regional Courts 0.294 -0.135 1 = poor job, 5 = excellent job (0.295) (0.155) Firm Size 0.769*** 0.079 number of employees (logged) (0.193) (0.093) Private Firm -0.06 -0.343 dummy, 1 = private ownership (0.624) (0.375) Constant -8.012*** 0.915 (3.022) (1.804)

GDP per capita 0.372 -0.059 in constant 2000 rubles per 1000 persons (logged) (0.484) (0.321) Regional Political Competition -0.098 -0.274*** index, -3 = uncompetitive, 5 = highly competitive (0.218) (0.127) Competition × Discretion 0.134 0.152*** interaction (0.103) (0.066) Log-likelihood -90.395 -238.364 No. of Cases 400 402

Note: Survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center and economic data from Rosstat. Coefficients represent estimates from multilevel logistic regressions with a random coefficient for the bureaucratic discretion variable and random intercepts at the region level; standard errors in parentheses. All models also include dummy variables for respondents’ economic sector. Out of space concerns, region random effects and sector variables not reported. * p<0.10, ** p<0.05, *** p<0.01

Table 5.3: Lobbying Response to Discretion Depends upon Political Environment

126 management becomes increasingly positive. Figure 5.1 shows the estimated marginal eﬀects for each of the interaction’s components.12

1.5 1.0

1.0

0.5 0.5

0.0 0.0

−0.5 −0.5

−1.0

−1.0 −1.5 Marginal Estimate of Bureaucratic Discretion (dy/dx1) Marginal Estimate of Bureaucratic Marginal Estimate of Regional Political Competition (dy/dx2) Marginal Estimate of Regional Political Low Medium High Low Medium High

Regional Political Competition Perceived Bureaucratic Discretion

Note: Firm-level survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center and economic data from Rosstat. Plots generated using coefficients from the interactive model reported in Table 5.3 indicating whether firms employ someone with past experience as management within a state-owned enterprise. Bands represent 90% confidence intervals.

Figure 5.1: Marginal Eﬀects of Bureaucratic Discretion and Political Competition on Probability of Hiring Manager with Government Ties

The figure’s left-hand plot demonstrates the main finding, that in regions with high levels of political competition, perceptions of higher bureaucratic discretion are associated with a higher probability that firms hire managers that are likely to have government ties. In contrast, higher bureaucratic discretion corresponds with fewer incidences of this strategy in regions with very low levels of political competition. In

12In this chapter, I focus solely on the estimated marginal effects of perceived bureaucratic discretion across the levels of regional political competition observed in the sample. Of course, the estimated substantive impact associated with these marginal effects depends upon an observation’s unique covariate profile and requires computing cross derivatives or cross differences (Ai & Norton 2003).

127 total, these ﬁndings are consistent with the claim that politically-competitive environments provide businesses with mechanisms for redress that are either unavailable or less eﬀective in locations where political leaders face fewer competitive pressures.

Comparing Firms’ Strategies for Resolving Disputes

This section set out looking for empirical evidence of increased responsiveness to the business community within politically-competitive regions. Responsive leaders should raise business actors’ expected utility from lobbying policymakers to make changes to

ﬁrms’ uncertain policy environment. This implies that, as a response to bureaucratic discretion, lobbying behavior is more likely in regions with high political competition.

The analyses offer broad support for this proposed link, confirming the intuition that seeking help from political leaders makes the most sense for firms in places where their government leaders have more pressing incentives to be helpful. Within this line of inquiry, however, a lingering question remains. What about economic actors in the politically-uncompetitive environments? If they are less likely to seek support from political leaders, what strategies do they favor in dealing with the regulatory uncertainty generated by bureaucratic discretion? Table 5.4 provides some preliminary answers to these questions.

Table 5.4 compares firm managers’ responses within high- and low-competition regions regarding a hypothetical scenario about a disputed regulatory claim.13 In this scenario, the survey first asked a randomly-selected group of respondents to imagine that a hypothetical firm had been wrongly assessed fines for breaking environmental laws, and then asked them to speculate about each of four distinct strategies that

13In addition to anchoring respondents’ thoughts in a specific situation, hypothetical scenarios can often elicit a greater response rate to potentially sensitive questions. Although this particular hypothetical scenario is taken from within a set of experimental questions, I limit my analysis to firms that received a specific set of conditions that best fit with the current inquiry. See Frye (2006) for an example of inference using survey experiments from this same dataset.

128 the ﬁrm could take in response to the charges.14 The stipulation that the imaginary

firm has been fined despite not actually breaking the law matches well with economic actors’ worst-case scenario: even after ostensibly doing their best to follow regulatory requirements, firms can still run afoul of the rules’ application at the hands of discretionary bureaucrats.

The first column in Table 5.4 reports for both high- and low-competition subsets the percentage of respondents indicating that the hypothetical firm would “definitely” or “most likely” turn to a business association for assistance in resolving the dispute.

The second column offers a comparison to the percentage of firms responding affirmatively to one of the other possible strategies for dealing with the disputed claim: paying the fine, attempting to bribe officials, or going to court. Interestingly, the comparisons show that firms in both high- and low-competition subsets agree that going to court is a more likely strategy for fighting disputed charges than seeking assistance from a business association. In regards to the other strategies, however, patterns diverge. In high-competition regions, response rates are statistically indistinguishable between seeking help from business associations (42%) and the other two strategies, bribery (42%) and acquiescing to the contentious fine (36%). In comparison, in regions with low political competition, seeking help from the business association is a dominated strategy; difference of means tests show that respondents in the low-competition regions chose bribing officials (49%) and paying the disputed

ﬁne (44%) in relatively greater proportions than turning to organized interest groups

14The prompt was: “A firm’s leadership is certain that the law has not been broken. In the past, it has never broken any laws regarding environmental protection. To resolve the dispute, will the business [...]?” The following choices were presented in separate questions: “turn to a business association for assistance in resolving the dispute,” “pay the fine without any further action,” “attempt to give officials a bribe to avoid a fine,” and “file a claim in arbitration courts.” For each separate strategy, respondents could answer that the hypothetical firm: “will take this action”, “more than likely will take this action”, “more than likely will not take this action”, “will not take this action”, or “difficult to say.”

129 Seek Help Alternate Via Association Strategy df Alternate: Pay Fine Anyways

low competition 31% 44% 164.49 high competition 42% 36% 110.29

Alternate: Attempt to Bribe

low competition 31% 49% 156.06 high competition 42% 42% 108.00

Alternate: File a Claim in Court

low competition 31% 74% 164.29 high competition 42% 78% 105.89

Note: Survey data from Frye (2006). Percentages correspond with the proportion of respondents answering affirmatively to the separate choices offered after the following hypothetical situation: “A firm’s leadership is certain that the law has not been broken. In the past, it has never broken any laws regarding environmental protection. To resolve the dispute, will the business [...]?” The following choices were presented in separate questions: “turn to a business association for assistance in resolving the dispute,” “pay the fine without any further action,” “attempt to give officials a bribe to avoid a fine,” and “file a claim in arbitration courts.” Bold figures indicate a p-value of p < 0.10 rejecting the hypothesis test that the two means are equal.

Table 5.4: Comparing Strategies for Contesting Bad Claims: Responses to a Hypo- thetical Scenario

130 (31%).15 These findings suggest at least one response to the question of what businesses in politically uncompetitive regions do when faced with problems stemming from the unpredictable application of regulatory laws. In such locations, firms appear to ready to remedy the situation through bribe payments or even paying the questionable fine before they would seek help from business associations and, by extension, their resources for lobbying.

This chapter’s expressed goal is to investigate the causal mechanism by which political competition conditions investors’ reaction to bureaucratic discretion. Accord- ingly, this ﬁrst half has focused on one potential mechanism – that by encouraging responsiveness among political leaders and granting more opportunities to inﬂuence economic policy, political competition gives businesses an expanded set of political tools that they can use to settle regulatory disputes and reshape policy arrangements.

The empirical results have shown that firms in politically-competitive regions display higher approval of elected leaders and a greater willingness to engage in activities that bring them in contact with policy-makers. This appears particularly true for those firms that perceive regional bureaucrats to have high discretion. Taken together, these findings suggest strongly that politically-competitive environments do, in fact, create more attentive political leaders as well as increase businesses’ representation in the policy-making process. Having found evidence that political competition provides investors’ with political tools for dealing with difficulties that arise from bureaucratic discretion, the analyses now turn to investigating whether or not politically-competitive environments make such difficulties less likely in the first place.

15The p-values associated with these independent t-tests are p=0.02 and p=0.09, respectively.

131 Political Competition & Regulatory Bureaucrats

In addition to making claims about political competition’s effect on the responsiveness of government leaders, the theoretical framework implies that political competition makes bureaucratic discretion less of an obstacle for investor by involving more parties in monitoring state agents. The logical result of this should be that, within highly competitive regions, less costly and more vigilant monitoring reduces undesirable agent behavior. Because data limitations prevent me from observing the amount of monitoring that takes place with a given region, this section concentrates its analytic efforts instead on testing the mechanism’s predicted outcome – political competition should lead to better-behaved bureaucrats. In doing so, I confront the challenge that “undesirable” agent behavior is often context-specific and identifying it requires knowing principals’ specific goals and directives.16 As a result, I investigate three general indicators of bureaucratic behavior: respondents’ assessments of administrators, reports of severe bribe-taking among regulatory officials, and disputes with government agencies over the application and interpretation of laws. The remainder of the chapter looks at each of these variables in turn, comparing firms’ responses across political environments to test the observable implications of this causal mechanism.

Approval of Administrators & Discretion

Borrowing a play from earlier analyses in this chapter, I begin by examining ﬁrms’ attitudes towards various unelected oﬃcials. I do this to probe the plausibility that discretionary bureaucrats in politically-competitive regions are less likely to carry

16For example, within the classical principal-agent literature on American bureaucracy, several types of agent behavior could qualify as undesirable to political principals; depending upon the situation, agents can defect by using expertise to mislead principals about policy outcomes or, alternatively, not investing eﬀort into developing suﬃcient expertise. Similarly, depending upon principals’ goals, agents can frustrate principals by both shirking as well as by enforcing standards too rigorously. For a comprehensive review of the delegation literature, see Bendor et al. (2001).

132 out their duties in ways that would scare away investors. According to the implied mechanism, political competition encourages political principals to watch their bureaucratic agents more carefully and allow additional parties to assist in monitoring state agents. If that logic holds, we might expect discretionary bureaucrats to have higher approval (or less disapproval) among ﬁrms in political environments with high levels of competition. Table 5.5 below shows the results from testing this observable implication.

Assessment of Oﬃcials Political Competition 1 = very poor, 5 = very good Low High df

Regional Administrators full sample 2.97 3.06 458.90 subset: low discretion 2.93 2.98 368.58 subset: high discretion 3.11 3.48 52.79

City Administrators full sample 2.80 3.09 494.38 subset: low discretion 2.78 3.02 388.31 subset: high discretion 2.90 3.55 54.93

Tax Administrators full sample 3.26 3.29 491.55 low discretion 3.26 3.23 390.39 high discretion 3.35 3.54 47.16

Note: Survey data from Frye (2006). Bold ﬁgures indicate a p-value of p < 0.10 rejecting the hypothesis test that the two means are equal.

Table 5.5: Lobbying Eﬀorts: Diﬀerence of Means

Table 5.5 compares respondents’ assessments of three diﬀerent sets of bureaucrats:

133 administrators at the regional level, administrators at the local level, and representa-

tives of the federal tax administration. For each set of bureaucrats, the table presents

comparisons for the full sample as well as for subsets depending upon whether ﬁrms

perceive bureaucrats to have high or low discretion. For exposition’s sake, I focus

on the high discretion subset as the most theoretically meaningful. Looking at the

table, we see that among ﬁrms that perceive high discretion, those respondents in

high-competition region have a much higher evaluation of their regional and local

administrators.17 The difference is particularly large within the high discretion subset at the local level; on a five-point scale, firms in politically-competitive regions rate their administration on average almost two-thirds of a point higher. In contrast, none of the differences across groups are significant for firms’ assessments of federal tax administrators. In general, bureaucratic discretion appears to meet with higher approval among firms in politically-competitive regions than among firms in regions with low levels of political competition. This is in harmony with the claim that political competition encourages better monitoring and, consequently, better behavior on the part of discretionary bureaucrats. The next analysis delves deeper, giving us an example of errant behavior that might be constrained by improved monitoring.

Corruption & Discretion

In order to test the observable implication that political competition leads to better- behaved agents, this section ﬁrm managers’ response to inquiries about demands for bribes by state oﬃcials – a behavior that, at least ostensibly, should be frowned upon by political principals, regardless of context. In emerging economies such as Russia,

17For both sets of bureaucrats in the high discretion subset, the diﬀerence in high- versus low- competition regions is statistically signiﬁcant (p=0.07 and p<0.001, respectively), indicating that we reject the null hypothesis that the means are indistinguishable.

134 worries about bureaucratic discretion are often tied to corruption. Greater discretion, the argument goes, may give rent-seeking bureaucrats extra leverage and more opportunities to use their authority for personal proﬁt.18 It is a view that pervades in Russia: a 2005 survey conducted by the Information Science for Democracy Foun- dation (INDEM) found that ninety-one percent of Russian businesspeople surveyed believed that corruption has its origin in “the vague character of laws, which gives government an opportunity to interpret them widely” (Popov 2006).

Since empirical research provides no clear answers about whether corruption in and of itself depresses business investment (Treisman 2007), the emphasis here is on bribe- taking as an example of deviant behavior on the part of state agents that ought to be curbed by improved monitoring.19 In this light, the claim that politically-competitive environments provide better monitoring of agents leads to the prediction that in high- competition regions, discretion should be less associated with bureaucratic corruption.

To the extent that discretion and corruption do have a positive relationship, we would expect it to be most prevalent in political environments where diﬀuse monitoring is less likely to occur, i.e. regions with low levels of political competition.

I test these predictions using ﬁrms’ perceptions about the severity of bribe-taking

18In a 2007 report written by anti-corruption watchdogs and sponsored by the State Duma Anti- Corruption Committee, Russian policy experts spell out the perceived dangers of bureaucratic discretion: “There are areas in which it is essential that government employees be given as little discretion as possible...Corruption risk is greatest in situations where the law gives several possible options without clearly deﬁning the criteria for deciding one way or the other” (italics added, Talapina & Yuzhakov 2007, 35).

19For an example of ambivalence towards bribery, consider the sentiments of one Western consultant in the Russian retail market: “[Bureaucratic corruption] is actually quite manageable...It is a cost of doing business. [If] official rates are low, [then] so is the cost of doing business...In an environment like this, if you are not paying, somebody somewhere is...Investors have to come to terms with that” (author interview, 23 July 2009). In accordance with analyses in this first half of this chapter, paying bribes may possibly be one of firms’ preferred strategies for dealing with discretionary bureaucrats. The implication of the improved monitoring mechanism is that, as political competition increases, bribe-seeking should decrease among more-carefully monitored agents.

135 Are Bribes a Serious Obstacle? (Region) (City) (Inspectors) Bureaucratic Discretion 0.008 -0.086 -0.058 1 = no discretion, 4 = high discretion (0.144) (0.145) (0.128) Frequent Changes to Laws 0.110 0.104 0.049 1 = no obstacle, 5 = very serious obstacle (0.113) (0.118) (0.102) High Tax Rates 0.022 0.084 0.135 1 = no obstacle, 5 = very serious obstacle (0.120) (0.127) (0.110) Regional Administration -0.268 -0.614*** -0.372** 1 = poor job, 5 = excellent job (0.201) (0.209) (0.183) Regional Governor -0.127 0.102 0.058 1 = poor job, 5 = excellent job (0.186) (0.186) (0.167) Firm Size -0.054 -0.016 -0.035 number of employees (logged) (0.091) (0.096) (0.083) Private Firm 0.417 0.370 0.649* dummy, 1 = private ownership (0.400) (0.430) (0.371) Constant -1.013 2.771 -1.727 (1.871) (2.608) (1.466)

GDP per capita 0.320 -0.858 0.341 in constant 2000 rubles per 1000 persons (logged) (0.387) (0.624) (0.250) Regional Political Competition 0.285* 0.365*** 0.291** mean-centered index, -3 = uncompetitive, 5 = highly competitive (0.155) (0.136) (0.122) Competition × Discretion -0.162** -0.172*** -0.179*** interaction (0.067) (0.067) (0.060) Log-likelihood -241.000 -216.200 -273.200 No. of Cases 432 397 446

Note: Survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center and economic data from Rosstat. Coefficients represent estimates from multilevel logistic regressions with a random coefficient for the bureaucratic discretion variable and random intercepts at the region level; standard errors in parentheses. All models also include dummy variables for respondents’ economic sector. Out of space concerns, region-specific effects and sector variables not reported. * p<0.10, ** p<0.05, *** p<0.01

Table 5.6: Discretion Less Associated with Corruption in High Competition Regions

136 among various groups of state oﬃcials. Table 5.6 presents that results of regression analyses for three separate dependent variables related to corruption.The models in

Table 5.6’s three column analyze respondents’ perceptions of the severity of bribe- taking among specific groups: bureaucrats at the regional-level of government, city- level bureaucrats, and regulatory inspectors.20 The independent variables are firms’ perceptions of bureaucratic discretion, regional political competition, and their interaction. Once more, I control for firm characteristics, such as sector, firm size and private- versus state-ownership. I also control for firms attitudes towards changing laws and tax rates, as well as their assessment of their governor and regional administration. To estimate the models, I use multilevel logistic regression with the same random-slope, random-intercept specification used elsewhere in the analyses.

In terms of signs and statistical signiﬁcance, the models in Table 5.6 provide consistent results for the variables of interest. The interaction between discretion and political competition is negative and statistically signiﬁcant. Substantively this tells us that, as regional politics become increasingly competitive, perceptions of higher bureaucratic discretion are less positively associated with serious bribery problems.

To provide better intuition about how the relationship between ﬁrms’ perceptions of discretion and severe corruption changes across political environments, Figure 5.2 plots out the marginal eﬀects.21

Focusing on Figure 5.2’s plot of the bureaucratic discretion’s marginal eﬀect, we

20The exact wording of the question is: “For firms operating in your sphere of business, how severe are bribes required by officials from the regional level [officials from local city administration, from representatives of from regulatory authorities (inspectors)]?” Response choices ranged on a five-point scale from “this problem does not occur” to “extremely severe problem.” As with the other variable, I have recoded this variable as a binary indicator, with a value of 1 representing views that bribes are perceived to be “severe” or “extremely severe.”

21Specifically, Figure 5.2 uses the coefficient estimates from the model that focuses specifically on bribe-taking among regulatory inspectors. The results are similar across all three models.

137 1.0

1 0.5

0.0 0

−0.5

−1 −1.0

−1.5 Marginal Estimate of Bureaucratic Discretion (dy/dx1) Marginal Estimate of Bureaucratic −2 Marginal Estimate of Regional Political Competition (dy/dx2) Marginal Estimate of Regional Political Low Medium High Low Medium High

Regional Political Competition Perceived Bureaucratic Discretion

Note: Firm-level survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center and economic data from Rosstat. Plots generated using coeﬃcients from the interactive model reported in Table 5.6 indicating whether bribes demanded by regulatory inspectors pose a severe obstacle for businesses’ development. Bands represent 90% conﬁdence intervals.

Figure 5.2: Marginal Eﬀects of Bureaucratic Discretion and Political Competition on Bribes’ Severity

see some interesting results. Confirming the fears of anti-corruption advocates, at low- levels of political competition, the results show that that bureaucratic discretion and bribe-taking are indeed positively correlated. The relationship flips, however, as regional political competition increases. In regions with moderately-high and high levels of political competition, perceptions of higher bureaucratic discretion are negatively associated with bribe payments being a severe problem for businesses. These findings point fall in line with the improved monitoring mechanism: politically-competitive environments produce better-behaved state agents.

138 Disputes & Discretion

Thus far in the chapter, the empirical analyses have provided evidence supporting the

proposition that competitive politics create responsive governments. I have also pro-

vided evidence that political competition reduces the relationship between discretion

and agent behavior that is “undesirable” or inconsistent with principals’ directives.

Taken together, these results bolster the theoretical argument that the eﬀects of bu-

reaucratic discretion should depend upon the political context in which discretion is

granted.

Ideally, we would want to go one step further and test the micro-level implication

that, by the various pathways identiﬁed in this chapter’s analyses, high political com-

petition severs the link between bureaucratic discretion and investors’ uncertainty

about how the regulatory application will aﬀect their business interests. Unfortu-

nately, the survey provides no question that asks ﬁrms explicitly about policy uncer-

tainty or about the predictability with which laws are interpreted and applied. Instead

of a direct assessment of ﬁrms’ uncertainty about regulatory matters, the survey pro-

vides an alternative. This alternative measure comes from an experienced-based item

asking ﬁrms whether or not they have had a dispute with a government agency in the

last two years.22 Although not perfect, disputes with the government are a plausible measure of regulatory uncertainty in as much as they indicate that ﬁrms have been caught oﬀ-guard, disagree with, or felt themselves to be treated wrongly by regulatory agents.23 Assuming that disputes with government are indeed symptomatic of

22Speciﬁcally, the survey asks respondents the following yes/no question: “Over the last two years, have any economic disputes arisen between your ﬁrm and government agencies or local self- governing municipalities?” Roughly 26% percent of respondents replied “yes.”

23Of course, businesses who know that they have violated laws may still have an incentive to dispute claims in order to escape their actions’ legal consequences. In fact, in the experimental part of the survey, firm managers acknowledge as much – 48% of respondents affirmed that a hypothetical firm would file a claim in court contesting an environmental claim while knowing secretly that

139 unpredictable policy application, the theoretical framework implies that discretion should correspond with higher incidences of disputes in politically-uncompetitive regions. In contrast, the theoretical framework anticipates the opposite within regions with high political competition; there bureaucratic discretion should correlate with fewer disputes.

For the chapter’s ﬁnal analysis, I test these predictions using a multilevel logistic regression model of ﬁrms’ disputes with government agencies in the last two years.

First and foremost, the model contains the measures of perceived bureaucratic discretion, regional political competition, and their interaction. Keeping with other analyses in the chapter, I control for potentially confounding variables that could covary with both discretion and disputes, including problems with high tax rates or frequently changing laws, perceptions of bureaucratic corruption, and assessments of the regional administration and governor. Given that close to 60% of the reported disputes involve the application of tax charges, I include firms’ assessment of tax administration as a precaution against omitted variable bias. Similarly, to account for the possibility that strategically-important firms may be treated differently by regulatory officials, I include a dummy indicator for whether a firm has received help from the regional government in the last two years. I also control for firm size, non-state ownership, and economic sector – all factors that might covary with firms’ contact with bureaucrats as well as their propensity to enter a dispute with government agencies. As a robustness check for the model of general disputes, I also model disputes that are specific to the application of tax legislation. Table 5.7 reports the results from the analysis.

the charges were legitimate. However, in the case that the hypothetical ﬁrm believes itself to be innocent, 75% of respondents perceive that disputing charges is likely. These proportions are virtually identical across theoretically-relevant subsets of the data. Assuming a representative sample, a rough extrapolation of these data would suggest that a vast majority of disputes – roughly two-thirds – are pursued sincerely rather than disingenuously.

140 Dispute with Government Agency General Dispute Tax Dispute dummy, 1 = dispute in last 2 yrs. Bureaucratic Discretion -0.450*** -0.449** 1 = no discretion, 4 = high discretion (0.174) (0.197) Bureaucratic Corruption -0.262 -0.518* dummy, 1 = perceived as corrupt (0.267) (0.302) Frequent Changes to Laws 0.149 0.053 1 = no obstacle, 5 = very serious obstacle (0.119) (0.135) High Tax Rates 0.077 -0.026 1 = no obstacle, 5 = very serious obstacle (0.126) (0.141) Regional Administration -0.046 -0.002 1 = poor job, 5 = excellent job (0.221) (0.254) Regional Governor -0.100 -0.047 1 = poor job, 5 = excellent job (0.209) (0.237) Tax Administration -0.301** -0.351** 1 = poor job, 5 = excellent job (0.147) (0.170) Government Help 0.543 0.709 dummy, 1 = recent government help (0.343) (0.364) Firm Size 0.489*** 0.518*** number of employees (logged) (0.105) (0.119) Private Firm 0.617 0.927* dummy, 1 = private ownership (0.420) (0.528) Constant -1.975 -3.902 (2.342) (2.634)

GDP per capita 0.294* 0.229 in constant 2000 rubles per 1000 persons (logged) (0.151) (0.180) Regional Political Competition -0.064 -0.309 index, -3 = uncompetitive, 5 = highly competitive (0.496) (0.508) Competition × Discretion -0.185** -0.146* interaction (0.082) (0.089) Log-likelihood -224.238 -182.227 No. of Cases 435 435

Note: Survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center and economic data from Rosstat. Coefficients represent estimates from multilevel logistic regressions with a random coefficient for the bureaucratic discretion variable and random intercepts at the region level; standard errors in parentheses. All models also include dummy variables for respondents’ economic sector. Out of space concerns, region-specific effects and sector variables not reported. * p<0.10, ** p<0.05, *** p<0.01

Table 5.7: Discretion’s Association with Disputes Depends upon Context

141 1.0 1.0

0.5 0.5 0.0

0.0 −0.5

−1.0 −0.5

−1.5 −1.0 −2.0

Marginal Estimate of Bureaucratic Discretion (dy/dx1) Marginal Estimate of Bureaucratic −2.5 −1.5 Marginal Estimate of Regional Political Competition (dy/dx2) Marginal Estimate of Regional Political Low Medium High Low Medium High

Regional Political Competition Perceived Bureaucratic Discretion

Note: Firm-level survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center and economic data from Rosstat. Plots generated using coeﬃcients from the interactive model reported in Table 5.7 indicating whether respondents have had a economic dispute of any kind with a government agency in the last two years. Bands represent 90% conﬁdence intervals.

Figure 5.3: Marginal Eﬀects of Bureaucratic Discretion and Political Competition on Probability of Dispute with Government

As predicted by the logic of the theoretical framework, bureaucratic discretion appears to be less associated with unpredictable regulatory enforcement in high- competition regions that in low competition regions. The negative, statistically significant coefficient estimate on the interaction between political competition and perceived bureaucratic discretion indicate that as political competition increases, the correlation between perceived discretion and disputes with the government becomes increasingly more negative. Figure 5.3 provides a visual representation of this effect.

When political competition is high, perceptions of higher discretion correspond with a reduction in the probability of a reported dispute between a given ﬁrm and government agencies. On the low-competition end, the 90% conﬁdence bands overlap zero, preventing a clear statistical interpretation of the direction of the correlation

142 between discretion and disputes in politically-uncompetitive regions. The estimates on the low-competition end still do provide us with useful information. Comparing the value on the lower conﬁdence interval for the minimum level of political competition (-0.41) with the value for the upper conﬁdence interval in the maximum level of political competition (-0.66), we see that the two do not overlap. Statistically speaking, increases in perceived discretion within high-competition regions correspond with lower probabilities of disputes than identical increases in perceived discretion within low-competition regions.

These ﬁndings provide us with useful insights into the theory’s causal mechanisms.

To the extent that disputes with the government reflect firms’ inability to predict the application of regulatory rules ex-ante, then these results substantiate the claim that political competition reduces the uncertainty that businesses associate with bureaucratic discretion. Even disregarding a direct link between disputes and regulatory uncertainty per se, the findings provide a powerful follow-up to the previous analyses. These are the exact results that we would expect to see if, in equilibrium, discretionary agents in politically-competitive locations are better-monitored and, consequently, less likely to undertake actions that would give grounds for dispute.

Conclusion

I have argued that high levels of political competition help to reduce investors’ uncertainty about the regulatory environment by making leaders more attentive to constituents’ concerns and by minimizing agents’ divergent behavior by sharing the task of monitoring across multiple actors. Accordingly, the analyses in this chapter have focused on testing the observable implications of these two mechanisms.

So, are governments more responsive to business in political competitive regions?

Does political competition lead to “better-behaved” bureaucrats? Patterns in the

143 data provide broad support for both claims. Regarding government responsiveness, I have found that ﬁrms in politically-competitive regions evaluate their locally-elected leaders more positively, and they are more likely than counterparts in low-competition regions to respond to bureaucratic discretion by seeking political inﬂuence, both through organized lobbying as well as via informal government contacts. In additional analyses, within uncompetitive regions where leaders have little incentive to respond to economic actors’ concerns, I have found organized interest groups to be

firms’ last choice for handling regulatory problems, coming in behind other options such as bribing officials or even paying unjust fines.

Corroborating the observable implications of the theory’s improved monitoring mechanism, the data suggest that bureaucrats in politically-competitive environments are less likely to use their discretion in ways that create uncertainty in the regulatory environment. In these regards, my analyses have shown that discretion in high-competition regions is associated with more positive evaluations of administrators, fewer problems with bribery, and fewer disputes between ﬁrms and government agencies.

In the next chapter, I summarize the dissertation’s theoretical argument and review the ﬁndings of the empirical analyses. I then conclude the dissertation by discussing its implications for the discipline and identify avenues for future research.

144 CHAPTER 6

CONCLUSION

Given its importance for an economy’s long-term performance, questions of where and why private investment accumulates address a topic that is at the forefront of public discussion in many developing countries. Long-term investment lays the foundation for economic growth, which, in its own turn, opens new economic opportunities for people and can result in higher living standards. In this way, the factors that draw or deter investors to a particular area matter a great deal to those people interested in the welfare gains associated with increased economic performance. Fueled by these same concerns and motivation, I have directed my dissertation research to studying how key political and institutional factors shape the predictability of investors’ regulatory environment.

For investors who seek predictable environments, the state is a double-edged sword. On one hand, states’ involvement in economic aﬀairs is commonly justiﬁed as a way to increase the predictability of markets and market activity. Businesses and consumers look to the state to help correct market failures, supply necessary infrastructure, and reduce the transaction costs of measurement and enforcement.

On the other hand, the actors who control and represent the state respond to their own set of political and organizational incentives. This can often make it harder for economic actors to predict changes to their policy environment that might aﬀect their business interests. This dissertation contributes by identifying the conditions under

145 which granting bureaucratic agents discretion might be a boon to investors and when it is likely to only generate more unpredictability.

In this concluding chapter of the dissertation, I revisit the study’s goals and review its empirical ﬁndings. I then end the chapter by discussing both the prospects for extending the theory to other levels of analysis as well as the implications of the results for existing debates in political science.

Goals of the Dissertation

The dissertation sets out to explain the eﬀects of bureaucratic discretion on private investment across Russia’s regions. By their nature, long-term investment projects require that investors pay extensive costs initially and then wait for returns to mate- rialize several time periods in the future. Because their investment decisions depend upon their ability to forecast future costs and expected returns, investors seek predictable investment environments and avoid locations characterized by high policy uncertainty (Aizenman & Marion 1993, Aizenman & Marion 1999, Nooruddin 2011).

This dissertation identiﬁes bureaucratic discretion — agents’ leeway to make decisions and act independently of political bodies — as a potential source of uncertainty that can deter investors.

Within the principal-agent framework, granting discretion increases the inherent risk that bureaucratic agents may deviate from their principals’ expressed goals, yet politicians may extend discretion because it allows bureaucrats to execute their duties in complex cases (Bawn 1997) and encourages specialization (Gailmard &

Patty 2007). Leaders’ decision to reduce control over their bureaucratic agents has important implications for entrepreneurial activity. Increased discretion widens the range of potential outcomes that bureaucrats may produce, making the policy environment less predictable. In this dissertation, I argue that, if unchecked by other

146 factors in the political environment, discretion deters long-term investors by making it harder for them to predict how bureaucratic agents will interpret and apply the laws that will govern their business interests. A primary goal of the dissertation was to test this prediction.

In arguing that bureaucratic discretion can be a source of uncertainty for investors, my argument contradicts an existing literature that emphasizes the economic benefits of insulating state actors, such as bureaucrats, from the pressures of the political arena. For some scholars, putting policy management in the hands of independent or insulated bureaucrats is an attractive strategy for reducing unexpected changes that can arise due to political leaders’ potential incentives to reverse, manipulate, or renegotiate policy (Rogoff 1985, Levy & Spiller 1994, Miller 2000). Implicitly, such arguments rely upon bureaucratic agents who perform their delegated duties in a manner that investors can anticipate. Predictable bureaucracy, however, is scarce in many countries outside the developed world. The improbability that the predictability assumption holds within the Russian Federation or, for that matter a wide variety of developing and transition economies, has prompted the dissertation to pursue a second question: how does the broader institutional context shape bureaucratic discretion’s effect on investment?

In answering this question, I have called attention to political competition as a crucial feature of investors’ political environment that serves to condition their response to bureaucratic discretion. I have argued that political competition makes policy application more predictable by making politicians more responsive to constituents’ concerns about bureaucratic discretion and by spreading the costs of monitoring bureaucrats across non-state actors and supporting institutions. This diﬀuse monitoring

147 helps to stabilize investors’ expectations about the regulatory environment. In contrast, a lack of competition in the political environment dulls leaders’ incentives to respond to regulatory problems that arise from their agents’ behavior.

The theoretical framework that was developed in Chapter 2 brings together existing explanations of investment behavior and insights from the principal-agent literature to improve our understanding of why bureaucratic discretion stands as a potential threat to investment in some locations (such as those that prevail across many of Russia’s regions), yet may be beneﬁcial in others. Furthermore, in developing the logic associated with these claims, the theoretical argument generates a number of empirical predictions and observable implications, which I test to bolster conﬁdence in the argument’s validity. I turn now to a discussion of those tests.

Empirical Findings of the Dissertation

At the theoretical level, my argument predicted that bureaucratic discretion should correspond with lower levels of private investment, and that this negative relationship should be especially strong in regions of Russia characterized by low levels of political competition. Chapter 3 tested the first half of this claim using both quantitative data from business surveys and qualitative evidence from field interviews to show how firms’ perceptions of bureaucratic discretion affect their decisions about investment. As part of my empirical strategy to test the dissertation’s theoretical claims, I leveraged data from a survey of over 660 Russian firm managers. Using logit models of firms’ intent to invest, I demonstrated that bureaucratic discretion bears a robust, negative relationship to firm investment. Firm managers in the survey who believe regulatory bureaucrats to make decisions independently of other government bodies have a lower predicted probability of investing in fixed capital assets in the immediate future. Whereas later analyses explore how regions’ political environment conditions

148 these results, this ﬁrst set of analyses is useful in establishing a empirical baseline for how economic actors in Russia to respond to bureaucratic discretion.

As a complement to the statistical analyses, I also presented qualitative evidence gathered from ﬁeld interviews with ﬁrm directors, heads of business associations, policy advocates, and legal experts. Conducted during the summers of 2008 and

2009, the interview data were analyzed as a way to corroborate the theory’s proposed causal mechanisms with insights from those who know and operate within Russian markets. Revealing additional information about the patterns and practices of Rus- sian business actors, the qualitative evidence complemented the statistical analyses by: ﬁrst, providing insight into the relatively high weight that investors in Russia attach to concerns about unpredictable policy application; second, elaborating the link between bureaucratic discretion and investors’ uncertainty about how rules will govern their business interests, both now and in the future; and ﬁnally, showing that

ﬁrms’ desire to avoid uncertain regulatory environments aﬀects their overall investment strategy of where to locate and what activities to pursue. All in all, these interviews strengthen the analyses by support the claim that bureaucratic discretion deters investors by creating uncertainty about the application of regulatory rules.

As a way to establish a baseline response to bureaucratic discretion among economic actors in Russia, Chapter 3 pooled together respondents, irrespective of political environment. However, one of the main theoretical contributions of the dissertation is the argument that the relationship between discretion and investment is likely to depend on the broader institutional context. Within this dissertation, I have claimed that uncertainty surrounding bureaucratic discretion is a particularly acute problem for economic actors in environments that lack supportive political institutions. This is, perhaps, the key empirical prediction of the dissertation.

149 Accordingly, Chapter 4 tested the theory’s hypothesis of the conditional relationship between bureaucratic discretion and investment by combining enterprise survey data with region-level information on the competitiveness of regions’ political environment. Consistent with the theory’s empirical predictions, the statistical analyses demonstrated that, in regions with restricted political competition, firms’ perceptions of bureaucratic discretion are associated with substantial decreases in the probability that firms will choose to invest in the near future. The results also showed that high- competition political environments mitigate this negative relationship – as regional political increases, perceived bureaucratic discretion has less negative relationship with firms’ intent to invest within the coming twelve months.

To bolster confidence in both the findings’ robustness as well as the general flexibility of the theory’s logic, I also showed that the model successfully predicts other dependent variables. Politically-competitive environments mitigate the negative relationship between firms’ perceptions of high bureaucratic discretion and other activities that should be closely related to firms’ decision to invest in fixed capital assets. These activities include undertaking worker training programs, extending credit to buyers, and conducting marketing research. Taken together, Chapter 4’s analyses offer strong support for the theory’s argument that political competition affects economic actors’ response to unpredictable elements in their regulatory environment.

Chapter 5 utilizes the multilevel data to investigate two causal pathways by which the theory purports political competition to be reducing the regulatory uncertainty that investors associate with bureaucratic discretion. To whit, the theoretical framework asserts that political competition makes policy application more predictable by decentralizing the costs of monitoring state agents and providing investors with more responsive leaders and more opportunities to approach the state regarding regulatory disputes. So, are governments more responsive to business in political competitive

150 regions? Does political competition lead to “better-behaved” bureaucrats? The statistical analyses in Chapter 5 support both claims.

Regarding government responsiveness, I have found that firms in politically competitive regions evaluate their locally-elected leaders more positively. Perhaps more importantly, the analyses demonstrated that firms are more likely in politically- competitive regions to respond to bureaucratic discretion by seeking political influence, either through organized lobbying or employing managers with informal government contacts. Furthermore, within politically-uncompetitive regions where leaders should be less attentive to economic actors’ concerns, I found evidence indicating that, as a means for handling regulatory problems, turning to organized interest groups comes in far behind other options, such as bribing officials or even paying unjust fines. The data also substantiated the observable implications of the theory’s claim that politically-competitive environment provide improved monitoring. Multi- ple analyses demonstrated that bureaucrats in politically-competitive environments are less likely to use their discretion in ways that create uncertainty in the regulatory environment. Along these lines, the concluding analyses of Chapter 5 show that discretion in high-competition regions is associated with more positive evaluations of administrators, fewer problems with bribery, and fewer disputes between firms and government agencies.

Implications of the Dissertation

In this ﬁnal section, I provide a preliminary evidence of that the dissertation’s micro- level logic can help explain investment patterns at the macro-level in Russia and beyond. I then conclude by considering four implications of my research for political science and call attention to open avenues for future inquiries based upon my research.

The data that I have used in this dissertation test the argument at the micro level

151 using enterprise surveys and interviews with business actors and policy advocates.

A natural and desirable extension of my efforts would be to examine whether these microeconomic dynamics help to explain macro-level patterns of private investment across Russia’s regions. While an extensive treatment is outside the scope of this dissertation, I provide a glimpse here of what such an inquiry is likely to find. To probe the plausibility that the micro-level argument developed in the dissertation explains macro-level phenomena, I investigate the degree to which the interaction between bureaucratic discretion and political competition predicts observed investment patterns across regions. I do this using two different measures of bureaucratic discretion, one that comes from aggregating survey respondents’ perceptions of discretion by region and another that originates outside the survey completely.

Figure 6.1 provides an encouraging answer to the question of whether aggregate data would support the implications of the dissertation’s micro-level analyses. For the regions sampled by the survey, aggregate perceptions of bureaucratic discretion do map onto observed growth in private investment exactly as my argument’s conditional framework would predict. In keeping with the spirit of the dependent variable from the survey regarding plans to invest over the next twelve months, the y-axis reports growth in regions’ private ﬁxed capital investment for the year directly following the survey (i.e., percent change from 2005 to 2006). The results are right in line with the theory’s empirical predictions. In regions with low political competition

(denoted by the “L” markers), discretion is associated with lower investment growth.

This contrasts starkly with the high-competition regions (the “H” markers), where discretion appears to have the positive relationship with investment that is anticipated by the extant literature based upon developed democracies.

Figure 6.2 tells a similar story, only this time using an external measure of bureaucratic discretion that is completely outside the survey. To capture the statutory

152 Keeping Outliers Dropping Outliers H H

40 40

H H 30 H 30 H H H L L 20 20 L L L L (2005−2006) L (2005−2006) L L L 10 10

L L

Percent Growth in Private Investment in Private Growth Percent 0 Investment in Private Growth Percent 0 10 20 30 40 50 60 10 20 30 40 50 60 Percent of Firms Perceiving Percent of Firms Perceiving High Bureaucratic Discretion High Bureaucratic Discretion

Note: Survey data from Frye (2006); percent growth in private investment taken from Rosstat. Red lines and the letter “H” refer to high-competition regions; blue lines and “L” refer to low-competition regions.

Figure 6.1: Scatterplot of Regional Private Investment & Aggregated Measure of Bureaucratic Discretion

constraints placed upon bureaucrats’ freedom to interpret and apply laws in a discretionary manner, I code the number of legal documents issued by the governor and regional administration over the three years prior to the survey, 2002-2004.1 The logic behind this measure is that bureaucrats’ discretion in regulating economic activity should relate inversely to the extensiveness with which the body of legal code speciﬁes procedural behavior and decision-making rules.2 As the number of directives from

1These data were collected from Konsultant-Plus, a Moscow-based legal services company that archives Russian legislation. The counted documents are legally-binding documents entitled postanovleniya; while I have translated this as ‘directive,’ alternate translations might also include ‘resolution’ or ‘statement.’ They cover a wide range of issues, including important economic policies such as taxes, environmental and safety standards.

2While prominent research has used a similar argument for focusing on the speciﬁcity of individual

153 regional executives grows, the increasing body of detailed guidelines should place greater constraints regulatory bureaucrats’ ability to make decisions in interpreting and applying laws that are independent of other government bodies. As the theory’s conditional framework would predict, the scatterplot shows that increasing legislative constraints on bureaucratic discretion in the politically-uncompetitive regions is associated with higher growth in investment.3 Once again, in regions where political competition is high, the data tell a story that is consistent with the conventional wisdom within the developed democratic setting; in such regions, constraining bureaucratic discretion is associated with lower rates of growth in private investment.

Although preliminary, both ﬁgures hint at the explanatory power of the dissertation’s theoretical framework in predicting actual investment patterns outside the survey.

Admittedly, these results in Figures 6.1 and 6.2 are based upon a very small number of observations, but they are highly suggestive that we could learn more by testing the empirical implications of my argument in diﬀerent data and at diﬀerent level of analysis.

The results of this chapter’s analyses have important implications for political economy. First, to the general political economy literature, my dissertation reintro- duces a group of oft-overlooked government actors – bureaucrats – and highlights potential beneﬁts to considering more frequently their role in shaping economic outcomes. Scholars of investment have not fully appreciated the implications of political principals’ imperfect control over their bureaucratic agents, despite the fact

pieces of legislation (Huber & Shipan 2002), a similar logic holds for the completeness of the body of law within a particular policy environment. Given extensive and oft-updated guidelines, the parameters for individual interpretation or arbitrary application should be narrower; in contrast, policy application is likely to remain much more open to interpretation in regions characterized by little law-making activity.

3Appendix D shows the relationship between regional executive directives and aggregated perceptions of discretion.

154 H

H 30 H

L 20

(2005−2006) L L L L

10 Percent Growth in Private Investment in Private Growth Percent

L 0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 Number of Executive Directives, logged (2002−2004)

Note: Survey data from Frye (2006); data on volume of legal documents by regional executive collected by author from Konsultant-Plus. The letter “H” refer to high-competition regions; the “L” refer to low-competition regions. The low-competition regression line drops Omsk Oblast, the outlier at the bottom left of the graph; including Omsk only improves the ﬁt.

Figure 6.2: Scatterplot of Regional Private Investment & Legislative Constraints on Discretion

that principal-agent theory appears regularly in political economy arguments. For instance, in a conventional treatment of non-credible commitments, the theoretical emphasis lies on politicians’ time-inconsistent preferences and their inability to commit credibly to not reverse policy after investors have sunk their quasi-irreversible investments. Rather than focus on politicians’ inter-temporal dilemma, my argument makes a separate point rooted in the division between the negotiating parties and those charged with implementing the deal: even without the time-inconsistency problem, political promises may not yield predictable results if principals have weak control over the agents charged with carrying out those promises. To the extent such commitments are also “non-credible,” it suggests that there are gains to be had from

155 conceptualizing a broader family of credibility problems that would include a role for agency issues.

Second, political economy research could gain much from making bureaucracy less of a black box. This dissertation has made inroads by identifying how bureaucratic discretion can deter investors (by creating uncertainty about the application and interpretation of laws) as well as how political competition can attenuate investors’ misgivings about bureaucrats (through improved monitoring and actors’ enhanced access to the policy process). Even then, the dissertation’s theoretical framework treats bureaucrats less like self-directed actors and more like policy-implementation machines that function with varying degrees of predictability. Whether through care- ful survey work or extensive qualitative treatment, a better understanding of bureaucrats’ incentives and constraints within a given system can only improve our theories ability to explain and predict the interactions between state and market actors.

On this same note, in setting up part of my theoretical argument, I have used existing studies as touchstones to highlight an idea that pervades certain literatures

– that policy outcomes will be qualitatively better if bureaucratic experts responsible for overseeing and implementing those policies are protected from the interference of meddling politicians. Yet clearly, there are reasons to believe that, on some level, giving more independence to bureaucrats in charge of applying and enforcing regulatory rules has different implications for an economy than does granting greater flexibility to ministerial technocrats who make strategic decisions about macroeconomic development (Evans 1995, Kohli 2010) or central bankers who must try to control inflation

(Rogoﬀ 1985, Maxﬁeld 1998).

By delving deeper into the varying levels and functions of bureaucratic person- nel, future research has the opportunity to use create more ﬁnely-grained theories

156 about the economic eﬀects of granting bureaucrats greater leeway. For instance, future research could explore how policy domain or bureaucrats’ responsibilities within the policy process conditions business actors’ response to bureaucratic discretion.

Alternatively, further studies could beneﬁt by distinguishing between agents’: independence in applying and interpreting rules (what I have called discretion), ability or authority to draft and establish the rules themselves (delegation or autonomy), and invulnerability to political pressure (insulation or political independence). Given the general nature of the argument, the theoretical framework is already amenable to a cross-national analysis outside the Russian Federation, but a cross-national setting would be ideal for analyzing how democratic political institutions and competitive politics aﬀect investors’ response to various independent state actors, such as regulatory agencies, judiciaries, or central banks.

Third, for literatures where bureaucracy already plays a central role, this dissertation offers a way to reconcile conflicting views over the economic influence of bureaucratic institutions. On one side, there are concerns that bureaucratic involvement can obstruct private economic activity by raising the costs of doing business through red tape and corruption (World Bank 2005, Kaufmann, Kraay & Mastruzzi 2009). Such sentiments are echoed within the literature on post-communist development, where studies have portrayed bureaucrats as “grabbing hands” that drive entrepreneurs into the unofficial economy (Frye & Zhuravskaya 2000) and force businesses to pool resources in common defense (Duvanova 2007, Pyle 2009). At the same time, others point to bureaucracy’s efficiency and expertise as the bedrock upon which successful economic policy is built (Johnson 1982, Rauch 1995, Rauch & Evans 1999). For instance, a recent cross-regional study within Russia takes this view, arguing that larger bureaucracies extend a “helping hand” of critical infrastructure and support that helps privatizing companies making their way in a new economic system (Brown,

157 Searle & Gehlbach 2009). This dissertation contributes by shifting the debate away from vague questions of whether bureaucracy hurts or harms economic development and focusing instead upon one specific dimension of bureaucrats’ involvement: under what conditions does bureaucratic involvement make the regulatory environment more predictable? This line of thinking encourages us to look for sources of variation in the broader institutional context that will help explain when bureaucrats’ role in the policy process should mitigate or magnify economic uncertainty. While this research has focused on the mitigating effects of political competition, future research can build upon these findings by elaborating additional mechanisms by which supporting political institutions can affect the predictability of policy implementation.

Finally, my work implies that delegation to independent agents may be counterproductive in precisely those places where the literature expects it to help most.

The analyses in Chapter 4 find that the negative relationship between discretion and investment is strongest in regions where leaders are least constrained by political competition. These results run counter to the standard credibility story wherein, given unconstrained political leaders, investors should welcome greater independence for bureaucratic agents as a type of substitute for political constraints on leaders’ ex-post behavior. Rather, my findings suggest the opposite – the benefits of extensive delegation may only begin to outweigh the negatives where surrounding political institutions can help reduce investors’ uncertainty about how regulatory bureaucrats will use their discretion.

158 BIBLIOGRAPHY

Ai, Chunrong & Edward C. Norton. 2003. “Interaction Terms in Logit and Probit Models.” Economics Letters 80:123–129.

Aizenman, Joshua & Nancy Marion. 1993. “Macroeconomic Uncertainty and Private Investment.” Economic Letters pp. 207–210.

Aizenman, Joshua & Nancy Marion. 1999. “Volatility and the Invertment Response.” Economica 66:157–179.

Alesina, Alberto & Roberto Perotti. 1996. “Income distribution, political instability, and investment.” European Economic Review 40(6):1203–1228.

Alesina, Alberto, Sule Ozler,¨ Nouriel Roubini & Phillip Swagel. 1996. “Political instability and economic growth.” Journal of Economic Growth 1(2):189–211.

Alvarez, R. Michael, Delia Bailey & Jonathan N. Katz. 2011. “An Empirical Bayes Approach to Estimating Ordinal Treatment Eﬀects.” Political Analysis 19(1):20– 31.

Amsden, Alice. 1989. Asia’s Next Giant: South Korea and Late Industrialization. Oxford University Press.

Anishyuk, Alex. 2009a. “IKEA Reaches Deal on Samara Outlet.” The Moscow Times.

Anishyuk, Alex. 2009b. “IKEA Upbeat after 10 Years of Problems.” The Moscow Times.

Aslund, Anders. 2002. Building Capitalism. Cambridge University Press.

Baldwin, Carliss & Richard Meyer. 1979. “Liquidity Preference Under Uncertainty: A Model of Dynamic Investment in Illiquid Opportunities.” Journal of Financial Economics 7(4):347–374.

Bawn, Kathleen. 1997. “Choosing Strategies to Control the Bureaucracy: Statutory Constraints, Oversight, and the Committee System.” Journal of Law, Eco- nomics, and Organizations 13(1):101–126.

159 Bechtel, Michael & Roland F¨uss. 2008. “When Investors Enjoy Less Policy Risk: Divided Government, Economic Policy Change, and Stock Market Volatility in Germany, 1970-2005.” Swiss Political Science Review 14(2):278–314.

Bendor, Jonathan, Amihai Glazer & Thomas Hammond. 2001. “Theories of Delega- tion.” Annual Review of Political Science 4:235–69.

Berkowitz, Daniel & David DeJong. 2003. “Policy Reform and Growth in Post-Soviet Russia.” European Economic Review 47(2):141–156.

Berry, William D., Jacqueline H.R. DeMeritt & Justin Esarey. 2010. “Testing for Interaction in Binary Logit and Probit Models: Is a Product Term Essential.” American Journal of Political Science 54(1):248–266.

Bertelli, Anthony & Andrew Whitford. 2009. “Perceiving Credible Commitments: How Independent Regulators Shape Elite Perceptions of Regulatory Quality.” British Journal of Political Science 39:517–537.

Bollen, Kenneth A. 1993. “Subjective Measures of Liberal Democracy.” American Journal of Political Science 33(1):58–86.

Borisov, Sergei. 2009. “Younger Governors for Russian Regions.” Russia Today. Archived at: http://www.russiatoday.com/Politics/2009-07-21/ROAR Younger governors for Russian regions.html/.

Brown, David, John S. Earle & Almos Telegdy. 2006. “The Productivity Eﬀects of Privatization: Longitudinal Estimates from Hungary, Romania, Russia, and Ukraine.” Journal of Political Economy 114(1):61–99.

Brown, David, John Searle & Scott Gehlbach. 2009. “Helping Hand or Grabbing Hand? State Bureaucracy and Privatization Eﬀectiveness.” American Political Science Review 103(2):264–83.

Bush, Jason. 2009. “Why IKEA is Fed up with Russia.” Business Week.

Campos, J. Edgardo, Donald Lien & Sanjay Pradhan. 1998. “The Impact of Corrup- tion on Investment: Predictability Matters.” World Development 27(6):1059– 1067.

Carpenter, Daniel. 2001. The Forging of Bureaucratic Autonomy: Reputations, Networks, and Policy Innovation in Executive Agencies, 1862-1928. Princeton University Press.

Caves, Richard E. 1996. Multinational Enterprise and Economic Analysis. Cam- bridge University Press.

160 Cukierman, Alex, Steven Web & Bilin Neyapti. 1992. “Measuring the independence of central banks and its effect on policy outcomes.” The World Bank Economic Review 6(3):353–398. “Diversified Economy Not for Rent”. 2008. The Moscow Times. Downs, George & David Rocke. 1994. “Conflict, Agency, and Gambling for Res- urrection: The Principal-Agent Problem Goes to War.” American Journal of Political Science 38(2):362–380. Dunning, John. 1988. Explaining International Production. Unwin Hyman. Duvanova, Dinissa. 2007. “Bureaucratic Corruption and Collective Action: Busi- ness Associations in the Post-communist Transition.” Journal of Comparative Politics 39:441–61. Easson, Alex. 2004. Tax Incentives for Foreign Direct Investment. Kluwer Law International Press. Easter, Gerald. 2002. “Politics of Revenue Extraction in Post-Communist States: Poland and Russia Compared.” Politics and Society 30(1):599–627. Enns, Patrick. 2007. “The Micro Foundations of Presidential Approval.” Unpub- lished manuscript. http://falcon.arts.cornell.edu/pe52/Enns PresApp.pdf. Epstein, David & Sharyn O’Halloran. 1994. “Administrative Procedures, Informa- tion, and Agency Discretion.” American Journal of Political Science 38(3):697– 722. Evans, Peter. 1995. Embedded Autonomy: States and Industrial Transformation. Princeton University Press. Frye, Timothy. 2002a. “Capture or Exchange? Business Lobbying in Russia.” Europe-Asia Studies 54(7):1017–1036. Frye, Timothy. 2002b. “Perils of Polarization: Economic Performance in the Post- communist World.” World Politics 54(3):308–377. Frye, Timothy. 2004. “Credible Commitment and Property Rights: Evidence From Russia.” American Political Science Review 98(3):453–466. Frye, Timothy. 2006. “Original Sin, Good Works, and Property Rights in Russia.” World Politics 58(8):479–504. Frye, Timothy. 2010. Building States and Markets after Communism: The Perils of Polarized Democracy. Cambridge University Press. Frye, Timothy & Ekaterina Zhuravskaya. 2000. “Rackets, Regulation, and the Rule of Law.” Journal of Law, Economics, and Organization 16(2):478–502. 161 Gailmard, Sean & John W. Patty. 2007. “Slackers and Zealots: Civil Service, Pol- icy Discretion and Bureaucratic Expertise.” American Political Science Review 51(4):873–889.

Gehlbach, Scott. 2008. Representation through Taxation: Revenue, Politics, and Development in Postcommunist States. Cambridge University Press.

Gehlbach, Scott & Phillip Keefer. 2008. “Investment without Democracy: Ruling- Party Institutionalization and Credible Commitment in Autocracies.” Manuscript available at: http://ssrn.com/abstract=1080169.

Gelman, Andrew & Jennifer Hill. 2007. Data Analysis Using Regression and Multi- level/Hierarchical Models. Cambridge University Press.

Golosov, Grigorii V. 2004. Political Parties in the Regions of Russia: Democracy Unclaimed. Lynne Rienner Press.

Grossman, Sanford J. & Oliver D. Hart. 1983. “An Analysis of the Principal-Agent Problem.” Econometrica 51:7–45.

Guriev, Sergei & William Megginson. 2007. Privatization: What Have We Learned? In Annual World Bank Conference on Development Economics Regional: Be- yond Transition, ed. Francois Bourguignon & Boris Pleskovic. World Bank.

Haggard, Stephan & Robert Kaufman. 1995. The Political Economy of Democratic Transitions. Princeton University Press.

Hays, Jude & Clint Peinhardt. 2009. “Geographical Limitations on the Utility of Investment Liberalization.” Paper presented at annual meeting of the Inter- national Studies Association. http://www.allacademic.com/meta/p313711_ index.html.

Heclo, Hugh. 1977. A Government of Strangers. Brookings Institute.

Holmstrom, Bengt. 1979. “Moral Hazard and Observability.” Bell Journal of Eco- nomics 10(1):74–91.

Huber, John D. & Charles R. Shipan. 2002. Deliberate Discretion? The Institutional Foundations of Bureaucratic Autonomy. Cambridge University Press.

Huber, John D. & Nolan McCarty. 2004. “Bureaucratic Capacity, Delegation, and Political Reform.” American Political Science Review 98(3):481–94.

Iarossi, Giuseppe. 2006. The Power of Survey Design: A User’s Guide for Managing Surveys, Interpreting Results, and Inﬂuencing Respondents. The World Bank.

ICRG. 2010. “International Country Risk Guide.” Accessed July 2010: http: //www.prsgroup.com/ICRG.aspx. 162 “IKEA Halts New Investment in Russia”. 2009. Bloomberg News.

“In the Trap of Populism”. 2009. Nezavisimaya Gazeta.

Iwasaki, Ichiro & Keiko Suganuma. 2005. “Regional Distribution of Foreign Direct Investment in Russia.” Post-Communist Economies 17(2):153–172.

Jensen, Nathan M. 2006. Nation-States and the Multinational Corporation: A Polit- ical Economy of Foreign Direct Investment. Princeton University Press.

Johnson, Chalmers A. 1982. MITI and the Japanese Miracle: the Growth of Industrial Policy, 1925-1975. Stanford University Press.

Kaufman, Herbert. 1956. “Emerging conﬂicts in the doctrines of public administration.” The American Political Science Review pp. 1057–1073.

Kaufmann, Daniel, Aart Kraay & Massimo Mastruzzi. 2009. “Governance matters VIII: aggregate and individual governance indicators, 1996-2008.” World .

Knack, Stephen & Philip Keefer. 1995. “Institutions and economic performance: cross-country tests using alternative institutional measures.” Economics & Pol- itics 7(3):207–227.

Kohli, Atul. 2010. “State Capacity for Development.” United Nations Development Programme.

Kurtz, Marcus & Andrew Shrank. 2007. “Growth and Governance: Models, Mea- sures, and Mechanism.” Journal of Politics 69(2):538–554.

Kydland, Finn E. & Edward C. Prescott. 1977. “Rules Rather than Discretion: The Inconsistency of Optimal Plans.” Journal of Political Economy 8(3):473–492.

Laﬀont, Jean-Jacques & Jean Tirole. 1988. “The Dynamics of Incentive Contracts.” Econometrica 56:1153–76.

Leﬀ, Nathaniel H. 1964. “Economic Development Through Bureaucratic Corrup- tion.” American Behavioral Scientist 8:8–14.

Levitsky, Steven & Maria Victoria Murillo. 2009. “Variation in Institutional Strength.” Annual Review of Political Science 12:115–133.

Levy, Brian & Pablo Spiller. 1994. “International trade and international investment in the product cycle.” Journal of Law, Economics, and Organizations 10(2):201– 246.

Lewis, David E. 2003. Presidents and the Politics of Agency Design: Political Insu- lation in the United States Bureaucracy, 1946-1997. Stanford University Press.

163 Li, Quan. 2006. “Democracy, Autocracy, and Tax Incentives to Foreign Direct Investors: A Cross-National Analysis.” Journal of Politics 68(1):62–74.

Li, Quan & Adam Resnick. 2003. “Reversal of Fortunes: Democracy, Property Rights and Foreign Direct Investment Inﬂows in Developing Countries.” International Organization 57(1):175–214.

Lipsky, Michael. 1980. Street-Level Bureaucracy. Russell Sage Foundation.

MacKuen, Michael B., Robert S. Erickson & James A. Stimson. 1992. “Peasants or Bankers? The American Electorate and the U.S. Economy.” American Political Science Review 86(3):597–611.

Malesky, Edmund & Krislert Samphantharak. 2008. “Predictable Corruption and Firm Investment: Evidence from a Natural Experiment and Survey of Cambo- dian Entrepreneurs.” Quarterly Journal of Political Science 3:227–267.

Mashaw, Jerry. 1999. Greed, chaos, and governance: using public choice to improve public law. Yale University Press.

Mauro, Paolo. 1995. “Corruption and Growth.” Quarterly Journal of Economics 110(3):681–712.

Maxﬁeld, Sylvia. 1998. Gatekeepers of Growth: The International Political Economy of Central Banking in Developing Countries. Princeton University Press.

McCubbins, Matthew, Roger Noll & Barry Weingast. 1987. “Administrative Pro- cedures as Instruments of Political Control.” Journal of Law, Economics, & Organization 3:243–77.

McCubbins, Matthew & Thomas Schwartz. 1984. “Congressional Oversight Over- looked: Police Patrols Versus Fire Alarms.” American Journal of Political Sci- ence 28(1):165–179.

McMann, Kelly & Nikolai Petrov. 2000. “A Survey of Democracy in Russia’s Regions.” Post-Soviet Geography and Economics 41(3):155–182.

Mereu, Francesca. 2008. “Crisis Could Take a Bite Out of Corruption.” The Moscow Times.

Miller, Gary J. 2000. “Above Politics: Credible Commitment and Eﬃciency in the Design of Public Agencies.” Journal of Public Administration Research and Theory 10(2):289–328.

Miller, Gary J. 2005. “The Political Evolution of Principal-Agent Models.” Annual Review of Political Science 8:203–25.

164 Moe, Terry. 1990. “Public Institutions: The Neglected Side of the Story.” Journal of Law, Organizations, and Economics 6:213–53.

Munck, Gerardo L. & Jay Verkuilen. 2002. “Conceptualizing and Measuring Democ- racy: Evaluating Alternative Indices.” Comparative Political Studies 35(1):5–34.

Murrell, Peter. 1993. “What is Shock Therapy and What Did it Do in Poland and Russia?” Post-Soviet Aﬀairs 9(2):111–140.

Nooruddin, Irfan. 2011. Coalition Politics and Economic Development: Credibility and the Strength of Weak Governments. Cambridge University Press.

North, Douglass. 1990. Institutions, Institutional Change, and Economic Perfor- mance. Cambridge University Press.

North, Douglass & Barry R. Weingast. 1989. “Constitutions and Commitment: The Evolution of Institutions Governing Public Choice in Seventeenth-Century England.” Journal of Economic History 49(4):803–832.

Oleinik, Anton. 2008. “Existing and Potential Constraints Limiting State Servants’ Opportunism: The Russia Case.” Journal of Communist Studies and Transition Politics 24(1):803–832.

Popov, Nikolai. 2006. “To Give and Take.” INDEM Foundation. http://www. indem.ru/en/publicat/Popov/Give&Take.htm.

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

Przeworski, Adam. 2003. States and Markets: A Primer In Political Economy. Cambridge University Press.

Putnam, Robert D., Robert Leonardi & Raﬀaella Y. Nanetti. 1993. Making Democ- racy Work: Civic Traditions in Modern Italy. Princeton University Press.

Pyle, William. 2009. “Organized Business, Political Competition, and Property Rights: Evidence from the Russian Federation.” Journal of Law, Economics, & Organization 27(1):2–31.

Rauch, James. 1995. “Bureaucracy, Infrastructure, and Economic Growth: Evi- dence from U.S. Cities During the Progressive Era.” American Economic Review 85(4):968–979.

Rauch, James & Peter Evans. 1999. “Bureaucracy and Growth: A Cross-National Analysis of the Eﬀects of ”Weberian” State Structures on Economic Growth.” American Sociological Review 64(5):748–765.

165 Remington, Thomas F. 2006. Democratization, Separation of Powers, and State Capacity. In The State After Communism: Governance in the New Russia, ed. Timothy J. Colton & Stephen Holmes. Rowman & Littlefield Publishers, Inc. Rogoff, Kenneth. 1985. “The Optimal Degree of Commitment to an Intermediate Monetary Target.” The Quarterly Journal of Economics 100(4):1169–89. Roland, Gerard. 2002. “The Political Economy of Transition.” Journal of Economic Perspectives 16(1):29–50. Sachs, Jeffrey. 1994. Poland’s Jump to the Market Economy. MIT Press. Satarov, Georgiy, Sergey Parkhomenko, Dina Krylova & Yuliya Rostovikova. 2007. “Business Without Corruption: An Action Guide.” Center for International Pri- vate Enterprise (CIPE). http://www.cipe.org/regional/eurasia/pdf/INDEM% 20OPORA%20brochure%20ENG%200407.pdf. Sharafutdinova, Gulnaz. 2007. “Why Was Democracy Lost in Russia’s Regions? Lessons from Nizhnii Novgorod.” Communist and Post-Communist Studies 40:363–382. Shipan, Charles R. 2004. “Regulatory Regimes, Agency Actions, and the Condi- tional Nature of Congressional Influence.” American Political Science Review 98(3):467–480. Shleifer, Andrei. 1998. “State Versus Private Ownership.” Journal of Economic Perspectives 12:133–150. Shleifer, Andrei. 2005. A Normal Country: Russia after Communism. Havard University Press. Shleifer, Andrei & Daniel Treisman. 2004. “A Normal Country.” Foreign Affairs March/April. Slinko, Irina, Evgeny Yakolev & Ekaterina Zhuravskaya. 2005. “Laws for Sale: Evidence from Russia.” American Law and Economics Review 7(1):284–318. Solomon, Peter. 2008. “Law in Public Administration: How Russia Differs.” Journal of Communist Studies and Transition Politics 24(1):115–35. Spence, Michael & Richard Zeckhauser. 1971. “Insurance, Information, and Individ- ual Action.” American Economic Review 61(2):380–87. Stigler, George. 1971. “The Theory of Economic Regulation.” Bell Journal of Economics and Management Science 2:3–21. Talapina, E. V. & V. N. Yuzhakov. 2007. “Guidelines for Initial Assessment of Laws for Corruption Risks.” Center for Strategic Development. http://www.cipe. org/regional/eurasia/pdf/CSRmethodolgy.pdf. 166 Treier, Shawn & Simon Jackman. 2008. “Democracy as a Latent Variable.” Ameri- can Journal of Political science 52(1):201–217.

Treisman, Daniel. 2007. “What Have We Learned About the Causes of Corrup- tion From Ten Years of Cross-National Empirical Research?” Annual Review of Political Science 10:211–44.

UNDP. 2007. Human Development Report 2006/2007 for the Russian Federation. United Nations Development Programme.

Varese, Federico. 2001. The Russian Maﬁa. Oxford Press.

Vernon, Raymond. 1971. Sovereignty at Bay. Basic Books.

Vishny, Robert & Andrei Shleifer. 1993. “Corruption.” Quarterly Journal of Eco- nomics 108(3):599–617.

Vogel, David & Robert Kagan. 2004. The Dynamics of Regulatory Change: How Globalization Aﬀects National Regulatory Policies. University of California Press.

Von Twickel, Nickolaus. 2008. “Signs of a Kremlin Fearful of Unrest.” The Moscow Times.

Wade, Robert. 1990. Governing the Market: Economic Theory and the Role of Government in East Asian Industrialization. Princeton University Press.

“Wage Arrears Cause Protests”. 2009. Broadcast. Ekho Moskvy.

Wei, Shang-Jin. 1997. “Why is Corruption So Much More Taxing than Tax? Arbi- trariness Kills.” Working Paper 6255. National Bureau of Economic Research.

Weingast, Barry. 1984. “The Congressional Bureaucratic System: a Principal-Agent Perspective (with Applications to the SEC).” Public Choice 44:147–88.

Weingast, Barry & Mark Moran. 1983. “Bureaucratic Discretion or Congressional Control? Regulatory Policymaking by the Federal Trade Commission.” Journal of Political Economy 91:765–800.

Williamson, John. 1994. In Search of a Manual for Technopols. In The Political Economy of Policy Reform, ed. John Williamson. Institute for International Economics.

Wilson, James Q. 1989. Bureaucracy. Basic Books.

World Bank. 2005. Judicial Systems in Transition Economies: Assessing the Past, Looking to the Future. World Bank and the International Finance Corporation.

167 Wright, Joseph. 2008. “Do Authoritarian Institutions Constrain? How Legisla- tures Impact Economic Growth and Investment.” American Journal of Political Science 52:322–43.

Yakolev, Evgeny & Ekaterina V. Zhuravskaya. 2008. “Reforms in Business Regula- tion: Evidence from Russia.” Available at SSRN: http://ssrn.com/abstract= 965838.

168 Appendix A

ADDITIONAL MATERIAL FOR CHAPTER 3

Firm Characteristics Responses Average number of employees 727 Median number of employee 125 Industrial firms 58% Retail and wholesale trading firms 15% Construction/transport/communications firms 29% Members of business organization 37% Average age of the manager (yrs) 47 Managers with college degree 90% Privatized firm 59% State-owned firm 12% De novo private firm 29% No competitors 7% Competition from foreign firms 7% Member of production association, trust, holding 24% Profit in preceding year 69% Note: Table recreated from Frye (2006).

Table A.1: Descriptive Statistics of Frye (2006) Survey Sample

169 Survey Item Wording (Frye 2006):

Investment Question (DV)

“Do you plan to make any large investment in the next twelve months for the

development of your ﬁrm (i.e., construction, reconstruction, capital renovation of the

building or surroundings, equipment updates, etc.)?”

• 4 = “yes”

• 3 = “likely yes”

• 2 = “likely no”

• 1 = “no”

Bureaucratic Discretion Question (IV)

“To what degree is independent decision-making, separate from other government bodies, characteristic of bureaucrats, administrators, and various inspectors in your region [territory, republic, city]?”

• 4 = “to a high degree”

• 3 = “most likely to a high degree”

• 2 = “most likely to a lesser degree”

• 1 = “completely uncharacteristic”

170 Plans to Invest 1 = no, 4 = yes Ord. Logit OLS

Bureaucratic Discretion -0.37*** -0.22*** 1 = no discretion, 4 = high discretion (0.12) (0.07) Policy Volatility 0.12 0.07 1 = no obstacle, 5 = very serious obstacle (0.10) (0.05) High Tax Rates -0.35*** -0.22*** 1 = no obstacle, 5 = very serious obstacle (0.10) (0.06) Regional Administration 0.26** 0.14** 1 = poor job, 5 = excellent job (0.12) (0.07) Regional Courts -0.09 -0.05 1 = poor job, 5 = excellent job (0.12) (0.07) Competitive Pressures 0.15** 0.08* 1 = no obstacle, 5 = very serious obstacle (0.08) (0.04) Labor Shortages -0.08 -0.04 1 = no obstacle, 5 = very serious obstacle (0.08) (0.04) Privatized Firm -0.22 -0.15 dummy, 1 = privatized, former SOE (0.23) (0.14) Annual Sales 0.35** 0.20** -1 = decreasing, 1 = increasing (0.17) (0.09) Firm Size 0.21*** 0.13*** number of employees (logged) (0.07) (0.04) Private Firm 0.29 0.21 dummy, 1 = private ownership (0.37) (0.22) Constant 1.48*** (0.49)

No. of Cases 418 418 Note: Survey data from Frye (2006). Coeﬃcient estimates from ordered logit and ordinary least squares regression as indicated. Robust standard errors in parentheses. Estimates for sector-level dummies suppressed out of space constraints, as are cutpoints in the ordered logit model. *p<0.10, ** p<0.05,*** p<0.01

Table A.2: Firm-Level Analyses: Robustness Check (Ordinal DV)

171 Firm Investment Robust Clustered Binary Extra dummy, 1 = ﬁrm plans to invest during coming year Std. Err. Std. Err. Discretion Controls

Bureaucratic Discretion -0.653*** -0.653*** -0.594*** 1 = no discretion, 4 = high discretion (0.148) (0.194) (0.175) Bureaucratic Discretion -0.843** dummy, 1 = high discretion (0.409) Frequent Changes to Laws 0.115 0.115 0.108 0.019 1 = no obstacle, 5 = very serious obstacle (0.116) (0.120) (0.111) (0.122) High Tax Rates -0.387*** -0.387*** -0.379*** -0.389*** 1 = no obstacle, 5 = very serious obstacle (0.121) (0.137) (0.121) (0.138) Regional Administration 0.506*** 0.506*** 0.423** 0.478** 1 = poor job, 5 = excellent job (0.186) (0.092) (0.192) (0.225) Regional Courts -0.162 -0.162 -0.178 0.039 1 = poor job, 5 = excellent job (0.142) (0.151) (0.145) (0.168) Regional Governor -0.323* -0.323*** -0.289 -0.343** 1 = poor job, 5 = excellent job (0.169) (0.114) (0.182) (0.208) Access to Finance -0.048 -0.048 -0.021 0.023 1 = no obstacle, 5 = very serious obstacle (0.084) (0.084) (0.081) (0.090) Labor Shortages -0.004 -0.004 -0.044 -0.003 1 = no obstacle, 5 = very serious obstacle (0.085) (0.071) (0.085) (0.096) Competitive Pressures 0.101 0.101** 0.101 0.109 1 = no obstacle, 5 = very serious obstacle (0.085) (0.051) (0.084) (0.093) Privatized Firm 0.096 0.096 0.011 0.089 dummy, 1 = privatized, former SOE (0.269) (0.284) (0.270) (0.295) Annual Sales 0.449** 0.449** 0.445** 0.390* -1 = decreasing, 1 = increasing (0.194) (0.194) (0.181) (0.200) Firm Size 0.301*** 0.301*** 0.294*** 0.313*** number of employees (logged) (0.083) (0.083) (0.083) (0.093) Private Firm 0.714 0.714* 0.769* 0.451 dummy, 1 = private ownership (0.454) (0.380) (0.434) (0.468) Bureaucratic Corruption 0.143 dummy, 1 = perceived as corrupt (0.278) Tax Agency Assessment -0.155 1 = poor job, 5 = excellent job (0.153) Bribes for Inspectors -0.066 1 = do not occur, 5 = severe problem (0.096) Constant -0.619 -0.619 -1.422 -0.22 (1.070) (0.929) (0.990) (1.252) Log-likelihood -241.154 -241.154 -244.312 -200.606 AIC 510.308 502.308 522.624 441.213 No. of Cases 403 403 403 331 Note: Survey data from Frye (2006). Model 1: Coefficient estimates from logistic regression; robust standard errors in parentheses. Model 2: Coefficient estimates from logistic regression; standard errors clustered by region in parentheses. Models 3-4: Coefficients from multilevel logistic regression with random coefficient for the discretion variable and random intercepts at the region level. Unit-specific estimates suppressed out of space constraints. * p<0.10, ** p<0.05, *** p<0.01

Table A.3: Firm-Level Analyses: Alternate Model Speciﬁcations 172 Description of Interview Sample

Composition

The interviewees in my sample fall into two broad categories: business representatives and policy experts. I discuss each category separately.

Business interviews — Interviewees within the business community fit one of two profiles. First, I spoke with directors, executives, and other representatives of upper management in companies operating within Russia. Within this group, respondents represented firms ranging in size from small enterprises with only three employees up to representatives of international companies employing thousands in Russia alone.

Although I spoke with representatives from foreign companies such as Coca-Cola and similar companies, the majority of respondents represented domestic ﬁrms.

The second sub-group of business representatives includes the leaders (general directors, chairs, presidents) of Russian business associations. These associations range in size from as low as twenty members up into the thousands. With the exception of two Moscow-speciﬁc organizations, all other associations operated at a national level and had active members in multiple regions of the federation. Not surprisingly, the overlap between the two business groups is considerable: nearly 70% of the association representatives in my sample either currently run or did at one time run a business endeavor in addition to their duties as head of their business organization.

Firms and business associations in the sample come from a wide swath of economic sectors, including: medical technology, food processing, agriculture, nanotechnology, banking, retail, industrial manufacturing, construction, and ﬁnance.

Expert interviews — Respondents in the “policy expert” category possess unusual breadth and depth of knowledge about regional economic development, the Russian legal system, or administrative reforms. All publish or advise others reguarly on their

173 area of expertise; for example, many of these respondents are quoted regularly in major newspapers, both foreign and domestic. Roughly half the experts lead research teams in NGOs and think tanks, including Transparency International, INDEM, and the Moscow Carnegie Center. The other half come from research institutions, such as the Centre for Economic and Financial Research at the New Economic School in

Moscow.

Recruitment

I recruited interviewees from business by sending email, faxes, or phone calls to organizations in Moscow that were listed in published business directories or online at associations’ websites. For policy experts, I contacted individuals in Russia that had published in academic outlets or been quoted in the media on the topics of investment, regional economic development, or business regulation. After gaining several initial interviews through my own eﬀorts, I also adopted a “snowballing” approach to recruitment. Roughly half my interviews were recruited through cold calls; the other half came through referrals and introductions from other interviewees.

174 Appendix B

ADDITIONAL MATERIAL FOR CHAPTER 4

175 Variable N Median Mean Std. Dev. Min/Max

Openness 11 3.00 3.27 0.91 2/5 score for openness of political arena Elections 11 3.00 3.00 0.89 2/5 score for free and fair elections Pluralism 11 3.00 3.18 0.09 2/5 score for presence of stable, competitive parties Political Competitiveness 11 9.00 9.46 2.5 6/14 additive index Gross Regional Product 11 3.63 3.68 0.46 3.25/4.92 regional GDP per capita (logged) Civil 11 3.00 3.00 0.89 2/5 score for strength of civil society Proportional Representation 11 1.00 0.55 0.52 0/1 presence of PR rule for legislative seats Regional Bureaucracy Size 11 4.36 4.36 0.78 2.18/6.24 number of employees per capita (logged) Population 11 7.76 7.82 0.78 3.53/9.25 in thousands (logged) Transportation Infrastructure 11 5.21 5.08 0.86 3.3/6.35 km. of railways per km2 (logged)

Note: Region-level political variables come from the Moscow Carnegie Center’s Regional Monitoring Project. Data on GDP per capita, the number of regional bureaucrats per capita, population, and railway density come from annual Rosstat publications. These measures represent the average values for regions across the three years (2002-2004) immediately preceding respondents’ participation in the survey. Regions are those sampled in the Frye (2006) data: Sverdlovsk, Khabarovsk Krai, Moscow, Nizhniy Novgorod, Novgorod, Omsk, Smolensk, Tula, Voronezh, Rostov, and the Republic of Bashkortostan.

Table B.1: Summary Statistics for Region-Level Variables

176 Firm Investment No Random Extra dummy, 1 = ﬁrm plans to invest during coming year Eﬀects Controls

Bureaucratic Discretion -0.645*** -0.647*** 1 = no discretion, 4 = high discretion (0.161) (0.160) Frequent Changes to Laws 0.115 0.093 1 = no obstacle, 5 = very serious obstacle (0.116) (0.112) High Tax Rates -0.385*** -0.396*** 1 = no obstacle, 5 = very serious obstacle (0.121) (0.123) Regional Administration 0.508*** 0.534*** 1 = poor job, 5 = excellent job (0.190) (0.200) Regional Courts -0.149 -0.174 1 = poor job, 5 = excellent job (0.147) (0.149) Regional Governor -0.372** -0.385** 1 = poor job, 5 = excellent job (0.174) (0.192) Access to Finance -0.027 -0.021 1 = no obstacle, 5 = very serious obstacle (0.086) (0.083) Labor Shortages -0.046 -0.042 1 = no obstacle, 5 = very serious obstacle (0.087) (0.088) Competitive Pressures 0.089 0.108 1 = no obstacle, 5 = very serious obstacle (0.086) (0.086) Privatized Firm 0.078 0.059 dummy, 1 = privatized, former SOE (0.276) (0.276) Annual Sales 0.459** 0.446** -1 = decreasing, 1 = increasing (0.190) (0.183) Firm Size 0.306*** 0.324*** number of employees (logged) (0.082) (0.084) Private Firm 0.774 0.783* dummy, 1 = private ownership (0.474) (0.444) Constant -2.149 10.782 (1.475) (10.425)

Regional Bureaucracy Size -2.480 number of employees per capita (logged) (2.242) Population -1.101 in thousands (logged) (0.693) Transportation Infrastructure -0.166 2 km. of railways per km (logged) (0.237) GDP per capita 0.440* 0.449 in constant 2000 rubles per 1000 persons (logged) (0.262) (0.403) Political Political Competition -0.185 -0.089 index, -3 = low pol. competition, 5 = high pol. competition (0.121) (0.142) Competition × Discretion 0.144** 0.132* interaction (0.066) (0.068) Log-likelihood -235.432 -233.068 AIC 504.864 512.136 No. of Cases 403 403 Note: Survey data from Frye (2006). Region-level political variables come from the Moscow Carnegie Center’s Regional Monitoring Project. Data on GDP per capita from annual Rosstat publications. Column 1: Coefficients from logistic regression with robust standard errors in parentheses. Column 2: Coefficients from multilevel logistic regression with random coefficient for the discretion variable and random intercepts at the region level; standard errors in parentheses. Additional region-level data taken from Rosstat. Out of space concerns, unit-specific effects not reported. * p<0.10, ** p<0.05, *** p<0.01

Table B.2: Robustness Check: (Alternate Model Speciﬁcations)

177 Firm Investment dummy, 1 = ﬁrm plans to invest during coming year Mean Std. Dev Lower Upper Bureaucratic Discretion -1.324 0.536 -2.267 -0.501 1 = no discretion, 4 = high discretion Policy Volatility 0.107 0.111 -0.067 0.296 1 = no obstacle, 5 = very serious obstacle High Tax Rates -0.383 0.111 -0.558 -0.198 1 = no obstacle, 5 = very serious obstacle Regional Administration 0.259 0.150 0.013 0.509 1 = poor job, 5 = excellent job Regional Courts -0.196 0.140 -0.424 0.038 1 = poor job, 5 = excellent job Competitive Pressures 0.106 0.083 -0.032 0.240 1 = no obstacle, 5 = very serious obstacle Labor Shortages -0.044 0.093 -0.188 0.112 1 = no obstacle, 5 = very serious obstacle Privatized Firm -0.056 0.273 -0.504 0.404 dummy, 1 = privatized, former SOE Annual Sales 0.412 0.179 0.124 0.705 -1 = decreasing, 1 = increasing Firm Size 0.269 0.079 0.133 0.399 number of employees (logged) Private Firm 0.829 0.456 0.098 1.613 dummy, 1 = private ownership Constant -1.346 1.181 -3.112 0.637

GDP per capita 0.221 0.352 -0.406 0.721 in constant 2000 rubles per 1000 persons (logged) Political Competition -0.031 0.114 -0.211 0.146 index, 6 = low pol. competition, 14 = high pol. competition Competition × Discretion 0.079 0.053 -0.002 0.173 interaction

Observations 418

Note: Firm-level survey data from Frye (2006). Region-level political variables come from the Moscow Carnegie Center’s Regional Monitoring Project. Data on GDP per capita from annual Rosstat publications. Coefficient estimates from hierarchical Bayesian logistic regression. Lower and Upper indicate 90% Bayesian credible intervals for each estimate. Analysis using two MCMC chains at 20,000 iterations in WinBUGS through R. From the initial 40,000 samples, 6,000 samples remain after throwing out the first 5,000 of each chain and “thinning” to keep every fifth sample.

Table B.3: Robustness Check: Bayesian Hierarchical Analysis with Diﬀuse Priors

178 Firm Investment dummy, 1 = ﬁrm plans to invest during coming year

Bureaucratic Discretion -0.404*** -0.693*** -0.707*** -0.605*** 1 = no discretion, 4 = high discretion (0.135) (0.172) (0.183) (0.165) Frequent Changes to Laws 0.044 0.081 0.082 1 = no obstacle, 5 = very serious obstacle (0.116) (0.112) (0.112) High Tax Rates -0.385*** -0.410*** -0.410*** 1 = no obstacle, 5 = very serious obstacle (0.132) (0.124) (0.123) Regional Administration 0.571*** 0.546*** 0.558*** 1 = poor job, 5 = excellent job (0.217) (0.205) (0.207) Regional Courts -0.126 -0.177 -0.172 1 = poor job, 5 = excellent job (0.159) (0.149) (0.149) Regional Governor 0.008 -0.439** -0.398** -0.400** 1 = poor job, 5 = excellent job (0.103) (0.207) (0.194) (0.195) Access to Finance -0.086 0.009 -0.01 -0.011 1 = no obstacle, 5 = very serious obstacle (0.067) (0.087) (0.083) (0.083) Labor Shortages -0.022 -0.047 -0.041 1 = no obstacle, 5 = very serious obstacle (0.093) (0.087) (0.086) Competitive Pressures 0.071 0.113 0.113 1 = no obstacle, 5 = very serious obstacle (0.092) (0.085) (0.085) Privatized Firm 0.156 -0.055 -0.013 dummy, 1 = privatized, former SOE (0.324) (0.281) (0.280) Annual Sales 0.478*** 0.580*** 0.474*** 0.461** -1 = decreasing, 1 = increasing (0.165) (0.195) (0.183) (0.183) Firm Size 0.332*** 0.427*** 0.305*** 0.305*** number of employees (logged) (0.072) (0.108) (0.085) (0.084) Private Firm 0.670* 0.655 0.64 dummy, 1 = private ownership (0.342) (0.456) (0.456) Constant -3.177*** -3.637* -1.734 -1.957 (1.111) (1.937) (1.561) (1.471) GDP per capita 0.371 0.574** 0.486 0.477* in constant 2000 rubles per 1000 persons (logged) (0.228) (0.274) (0.308) (0.258) Political Competition – Index -0.111 -0.217 mean-centered index, -3 = uncompetitive, 5 = highly competitive (0.115) (0.142) Political Competition – Dichotomous Coding -1.236* dummy, 0 = uncompetitive, 1 = highly competitive (0.679) Political Competition – Pluralism -0.576 -1 = no pluralism, 2 = high pluralism (0.377) Political Competition × Bureaucratic Discretion 0.081 0.144* 0.739* 0.346* interaction (0.062) (0.077) (0.339) (0.200) Sector & Legal Form Dummies No Yes No No Log-likelihood -280.397 -218.121 -227.462 -228.376 AIC 586.795 510.243 494.924 496.752 No. of Cases 447 380 380 380 Note: Survey data from Frye (2006). Region-level political variables come from the Moscow Carnegie Center’s Regional Monitoring Project. Data on GDP per capita from annual Rosstat publications. Coefficients from multilevel logistic regression with random coefficient for the discretion variable and random intercepts at the region level; standard errors in parentheses. Out of space concerns, unit-specific effects not reported. * p<0.10, ** p<0.05, *** p<0.01

Explanation: The Rep. of Bashkortostan is a clear outlier in terms of high restrictions on political competition, widespread perceptions of high bureaucratic discretion, and low investment. This table shows that, even under a variety of speciﬁcations, the key ﬁndings of the conditional theory continue to hold after dropping Bashkortostan from the analyses.

Table B.4: Robustness Check: Dropping Outlying Region

179 Firm Investment dummy, 1 = ﬁrm plans to invest during coming year Pluralism Openness Elections

Bureaucratic Discretion -0.679*** -0.687*** -0.580*** 1 = no discretion, 4 = high discretion (0.161) (0.164) (0.151) Frequent Changes to Laws 0.121 0.119 0.104 1 = no obstacle, 5 = very serious obstacle (0.112) (0.112) (0.111) High Tax Rates -0.389*** -0.395*** -0.375*** 1 = no obstacle, 5 = very serious obstacle (0.122) (0.122) (0.122) Regional Administration 0.520*** 0.508*** 0.509*** 1 = poor job, 5 = excellent job (0.197) (0.196) (0.194) Regional Courts -0.156 -0.144 -0.141 1 = poor job, 5 = excellent job (0.148) (0.148) (0.148) Regional Governor -0.390** -0.381* -0.351* 1 = poor job, 5 = excellent job (0.185) (0.184) (0.185) Access to Finance -0.026 -0.035 -0.024 1 = no obstacle, 5 = very serious obstacle (0.083) (0.083) (0.083) Labor Shortages -0.053 -0.043 -0.03 1 = no obstacle, 5 = very serious obstacle (0.087) (0.086) (0.086) Competitive Pressures 0.09 0.091 0.093 1 = no obstacle, 5 = very serious obstacle (0.085) (0.085) (0.085) Privatized Firm 0.093 0.088 0.075 dummy, 1 = privatized, former SOE (0.275) (0.275) (0.273) Annual Sales 0.462** 0.452** 0.467* -1 = decreasing, 1 = increasing (0.183) (0.183) (0.182) Firm Size 0.308*** 0.307*** 0.303*** number of employees (logged) (0.083) (0.083) (0.083) Private Firm 0.762 0.773* 0.768* dummy, 1 = private ownership (0.442) (0.442) (0.439) Constant -2.044 -1.982 -2.540* (1.451) (1.451) (1.422)

GDP per capita 0.443* 0.422* 0.479 in constant 2000 rubles per 1000 persons (logged) (0.258) (0.255) (0.247) Index Component -0.674* -0.494 -0.293 mean-centered index, -3 = uncompetitive, 5 = highly competitive (0.364) (0.352) (0.351) Index Component × Bureaucratic Discretion 0.479** 0.390** 0.274 interaction (0.189) (0.181) (0.180) Log-likelihood -234.822 -235.514 -236.901 AIC 509.643 511.027 513.802 No. of Cases 403 403 403 Note: Survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center and economic data from Rosstat; Coefficients from multilevel logistic regression with random coefficient for the discretion variable and random intercepts at the region level; standard errors in parentheses. Out of space concerns, random effects not reported. * p<0.10, ** p<0.05, *** p<0.01

Table B.5: Multilevel Analyses: Disaggregating Political Competition Index

180 ● ●●●● ● ● ●● ●●●●●●● ● ●●●● ● ● ●● ●●●●● ●● ●●●●●●●●● ● ●● ● ●●●● ●●●●●●● ●● ●● ● ● ● ●●●●●● ●●●●●●● ●● ●●●●●●●● ●● ● 1.0 ● ● ●●●●●●●●●●● ●●●●●● ●●●●● ●● ●●●●●●●● ●●●●●●● ●●●●● ● ● ● ●●●●●●●● ●●●●●●●● ●●● ● ● Low Discretion●●●●●●●● ●●●● ● ●● ●● 0.9 High Discretion

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

Predicted Probability of Firm Investment ● ●● ●● ● ●● ● ● ●● ● ●●●● ●●●●●●●●●●●● ●●●●●●●●●● ● ●● ●●●●●●●● ●●●●●●●●●●●● ●● ●●●●●● ●●● ●●●●●●●●● ●●●●●●●●●●● ● ●●●●●●● ● ●●● ● ●●●●●●●●●●● ●●●●●●●●●●●● ●●● ●●● ●● ●●●●●●●●●● ● ●●●●● ● ●● 0.0 ●●●●●●●●●● ●●●●●●●●● ●●●●●● ● ●●●●●● ●●●●●●●●●●● ●●●● ●● ● ●●● ● ●●● ●●●●●● ●●●●●●●●●● ●●●●●●

Low Medium High Openness of Regional Politics

Note: Firm-level survey data from Frye (2006), region-level democracy scores for the eleven sampled regions come from the Moscow Carnegie Center and economic data from Rosstat. Bold lines represent the predicted probability that a hypothetical ﬁrm would invest, given the level of regional political competition; discretion is manipulated from minimum (1) to maximum (4) while holding all variables at their median value. Small dashed lines represent 90% conﬁdence intervals obtained via simulation in R.

Figure B.1: Predicted Probability Plots: Index Component – Openness

181 ●● ●●● ●●●●●●●●●● ● ●●● ●●● ●● ●●●●●●●●●● ● ● ● ● ● ●●● ●●●● ●●●●●●●●● ● ● ●●● ●●●●●●●● ●●●●●●● ●●● ●●● ● ●● ●● ●●● ●●●●●● ● ● ● 1.0 ●●● ●●●●●●●●● ● ● ● ●●●●●●● ●●●●●●●●●●●●● ●●●● ●● ●● ● ●●●● Low Discretion●●●●●●●●●● ● ●●● ●●●● 0.9 High Discretion

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

Predicted Probability of Firm Investment ● ●●●●●●●●● ●●●●●●●●●●● ● ●● ●●●● ●●●●●●●●●●●● ●●●●●●●●●●●● ● ● ●● ● ●●●● ●●●●●●●● ●●●●●●●●●●● ● ● ● ●● ●●●●●● ●● ●●●●●●●●●●●●● ●●● ● ●● ●●●●●●●●●●● ●●●●●●●●●●●●●● ● ●● ● 0.0 ●●●●● ●● ●●●●●●●●●●●●●● ●●● ● ●●●●●●●●● ●●●●●●●●●●●● ●● ●● ● ●● ●●●● ●●●● ●●●●●●●●●● ● ●● ●●●●●●

Low Medium High Quality of Regional Elections

Figure B.2: Predicted Probability Plots: Index Component – Elections

182 ● ● ● ●● ● ●●●●●●●●●● ●●●●●●●● ● ● ● ● ●●●●●●●●● ●●● ● ● ● ●●●●●●●●● ● ● ●●●●●● ●●●●● ●●●●●●●● ● ● ●● ●● ● ● ●● ●●● ●●●●●●●● ●●●● ●● 1.0 ●● ● ●●●●●●●● ● ●●●●● ● ●● ●● ● ●●●●●●●● ●●●●●●●● ●● ●● ●● ● Low Discretion●●●●●●●●● ●●●● ● ●● 0.9 High Discretion

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

Predicted Probability of Firm Investment ● ●●●● ●●●●● ●●● ●●●● ● ●● ● ●●●●●● ●●●●●●●●●●●● ●●● ● ● ●● ●●●●●●●●● ●●●●●●●●●●● ●● ● ●● ●●●● ●● ● ● ●●●●●●●●●●●● ● ●●●●● ●● ● ●●●●●● ●●●●●●●●●●● ●●●●●●● ●● 0.0 ●●●●●● ●●●●●●●●●●●●●● ●●●●●●● ● ● ●●●●●●● ●●●●●●●●●●●●● ●●●●●● ●●● ● ●●●●●● ●●●●●●●●●●●●● ●●●● ●●

Low Medium High Regional Political Pluralism

Figure B.3: Predicted Probability Plots: Index Component – Pluralism

183 Dichotomous Coding of Discretion

● ●● ●●● ●●●●● ●●● ●●● ●● ●● ●● ●● ●●●● ●● ●●● ●●● ●●●● ●●● ●●● ●●●● ● ● ●●●● ●●●●● ● ●●●● ●● ●●●● ● ● ●●● ● ●●●●● ● ●●●● ● ● ●● ●●●● 1.0 ●● ●● ●●●● ●● ●● ●●● ● ●●● ●●● ●●●● ● ●●● ●● ●●●● ●●● Low Discretion●●● ●●●● ●● ●● ●● ●●●●● 0.9 High Discretion

0.8

0.7

0.6

0.5

0.4

0.3

0.2

Predicted Probability of Firm Investment 0.1 ● ● ●● ●● ● ● ●●●●● ●●● ●●●● ●●●●● ●● ●●●●● ●● ●●●●● ●●●● ●●●●● ●●● ●●●● ●●●● ● ●●●●● ●● ●●●●● ●●●● ● ●●● ●●●●● ●●●●● ●●●●● ●●● ● ●●●● ●●●●● ●●●●● ●●●●●● ●● ● ●●● ●● 0.0 ●●● ●●● ●●●●●● ● ●●● ●●●● ●● ●●●●● ●● ●●●●●● ●●● ● ●●● ●● ●●●●●● ●●●● ●●●●●●● ●●●● ●●● ● ●●●●

Low Medium High

Regional Political Competition

Figure B.4: Predicted Probability Plots: Dichotomous Measure of Bureaucratic Dis- cretion

184 ●●●●● ●●●●● ●● ● ●● ●●● ●●● ● ●● ●●●●●●●●● ● ●● ● ● ●●● ●●●●●●● ● ●● ●● ● ●● ● ●●● ●●●●●● ●●●● ●●●● ●● ●●●●●●●●●●● ●● ● ●●● ●● ●●● ● ●●●● ●● ● ●● ●●●●●● ●●● ●● ●● ● ●● ● ● 1.0 ●●●●●●● ●●●● ● ● ● ●●●●● ● ●●●●●● ●●●● ●●● ●●● ● ●●●●●● ●●●● ●●●●●●●●●● ●●● Low Discretion●● ●● ●● ●●● High Discretion 0.9

0.8

0.7 ● 0.6 ●

0.5 ● 0.4

0.3

0.2 ● 0.1 ●●●●● ●●●●●●●●●●●●● ●● ●●●●●● ● ● ●●●●●●●●●●●● ●●●●● ●●●● ●●●● ●●●●●●●●● ●● ●●●●● ●●●●●●●● ●●● ● ● ●●●●●●●● ● ●● ●●●●●●●●●● ●●●●●●●● ● ●●●● ●● ●● ●●●●●●● ● ●●●●●●●● ●●●● ●●●●● ●●●●●● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●● ●●●●●●●● 0.0 ●●●●●● ●●●● ●●●●● ●● ●●● ● ●● ● Predicted Probability of Firms' Intent to Invest ●●● ●●●●●●●● ●●●●●● ●●● ● ● ●●● ●●●●●●●●●● ●●●●●●●● ●● ● ●●●●●●●●● Low High

Regional Political Competition

Figure B.5: Predicted Probability Plots: Dichotomous Measure of Political Compe- tition

185 Appendix C

ADDITIONAL MATERIAL FOR CHAPTER 5

186 Membership Benefit: Membership Benefit: Lobbying Setting Rules 100 100 Low Competition High Competition 80 ● 80

● ● 60 ● 60 ● ● 40 ● 40 ● ● ● ● 20 20 ●

Percent of Responding Firms Percent 0 0

Full Low High Full Low High sample discretion discretion sample discretion discretion

Lobbies Governor Former SOE Manager to Influence Laws 100 100 Low Competition High Competition 80 80 ●

● ● 60 60 ● ● ●

40 40

20 ● 20 ● ● ● ● Percent of Responding Firms Percent ● 0 0

Full Low High Full Low High sample discretion discretion sample discretion discretion

Note: Firm-level survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center. Bands represent 90% conﬁdence intervals.

Figure C.1: Diﬀerence between High- and Low-Competition Regions: Comparing Across Subsets

187 Regional City Tax Administrators Administrators Administrators Good Low Competition High Competition ● ● ● ● ● ● ● ● ● ● ● ● ● ● Satisfactory ● ● ● ●

Poor

Full Low High Full Low High Full Low High sample Discretion Discretion sample Discretion Discretion sample Discretion Discretion

Note: Firm-level survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center. Bands represent 90% conﬁdence intervals.

Figure C.2: Diﬀerence between High- and Low-Competition Regions: Comparing Across Subsets

188 Appendix D

ADDITIONAL TESTS: MISSING MECHANISMS,

MEASUREMENT, AND VALIDITY

Introduction

This appendix addresses specific concerns surrounding the survey data that I have used in the dissertation’s empirical analyses. In this appendix, I address three separate issues. First, the argument’s micro-logic asserts that bureaucratic discretion impedes investment by creating regulatory uncertainty, yet policy application’s predictability is never directly observed in quantitative analyses. Accordingly, I provide evidence from additional enterprise surveys on the hypothesized link between discretion and uncertainty. In the second section, I conduct additional analyses to mitigate concerns about the perception-based measure of bureaucratic discretion and bolster confidence about the internal validity of my findings. Finally, the last section questions the external validity of the discretion and investment measures, examining whether the perceptions in the survey correspond with patterns of observed data in

Russia’s regions.

Bureaucratic Discretion & Regulatory Uncertainty

In Chapter 3, I use qualitative evidence from ﬁeld interviews with business actors and policy experts to examine the mechanism by which bureaucratic discretion is 189 argued to reduce incentives to invest, namely by undermining the predictability of the regulatory environment. This claim goes untested in the quantitative survey data, however, because the Frye survey lacks an item asking ﬁrm managers about the predictability of policy implementation. In this section, I use additional survey data, both sub-national and cross-national, to bolster the evidence supporting this step in the micro-logic that bureaucratic discretion corresponds with greater uncertainty about the application of regulatory laws.

The Frye survey has a major benefit in that it asks firm managers about traits of specific political institutions in greater detail and on a broader range than similar surveys, such as the Business Environment and Enterprise Performance Survey

(BEEPS) run by the World Bank and European Bank of Reconstruction and De- velopment. When inquiring about political matters, the BEEPS and related World

Bank Enterprise Surveys (WBES) ask a set of very general governance questions.

Crucially, these rival surveys lack the two questions that are central to this study: they lack a targeted question about bureaucratic discretion and have no prospective questions asking about investment plans. Consequently, I do not use these surveys as primary sources of data for the dissertation research.

The BEEPS/WBES instruments do, however, have a useful question relating to policy uncertainty that the Frye survey does not have. This question asks firms to agree or disagree with the statement: “Interpretations of the laws and regulations affecting my firm are consistent and predictable.” Conditional on finding a plausible proxy for bureaucratic discretion within these surveys, this policy uncertainty question allows me to test the proposed link between bureaucratic discretion and heightened uncertainty about the policy environment.1 For a proxy of discretion at

1The 2005 BEEPS data includes enterprises from fourteen regions of Russia, ﬁve of which overlap with the Frye survey: Voronezh, Moscow, Nizhniy Novgorod, Rostov, and Sverdlovsk. From the WBES surveys, I look at all 101 countries surveyed between 2002-2006.

190 5.0

4.5 ●

●

● ● ● ● 4.0 ● ● ●

● ●

3.5 ●

3.0 ● ● ''Are Laws Applied Unpredictably?'' Applied Unpredictably?'' Laws ''Are (region means; higher = more unpredictable) (region means; higher = more unpredictable) 2.5

5 10 15 20 Time Spent with Officials (standard deviation within regions)

Note: Firm-level survey data are Russian firms from the 2005 Business Environment and Enterprise Performance Survey (BEEPS). The x-axis is bureaucratic discretion, as measured by the within- region variance to the following question: “What percent of senior management’s time over the last 12 months was spent in dealing with public officials about the application and interpretation of laws and regulations and to get or to maintain access to public services?” The y-axis is a measure of policy uncertainty, country means on a 6-point scale agreeing or disagreeing with the statement: “Interpretations of the laws and regulations affecting my firm are consistent and predictable.”

Figure D.1: Policy Uncertainty & Variance in Time with Oﬃcials (Russian BEEPS Data)

the aggregate level, I use the standard deviation within the same political unit to an experienced-based question asking about the time spent dealing with government

191 officials over the application and interpretation of laws.2 To the extent that within- region or within-country variation to these questions reflects officials’ discretion in the interpreting and applying the very same legal codes, then this measure should be correlated with greater unpredictability in the application and interpretation of laws.

Figures D.1 and D.2 show that this measure of discretion is positively correlated with higher policy uncertainty, just as I argued within Chapter 2. Together with the qualitative evidence from my ﬁeld interviews in Chapter 3, these quantitative results at the sub-national and cross-national level help to establish the link between bureaucratic discretion on one hand and increasingly unpredictable application of policy on the other.

Internal Validity & Perceptions of Discretion

Respondents’ ability to answer questions knowledgeably is a potential constraint for survey-based research (Iarossi 2006). The survey question that I use to measure bureaucratic discretion asks enterprise managers about the independence that regional bureaucrats have in making decisions separate from other government bodies. One potential danger is that respondents who do not know the answer will simply guess, and their responses will be influenced by factors that might also covary with firms’ intent to invest. As a result, correlations between firm managers’ answers to the discretion question and firm investments plans would not arise through the theory’s causal pathway, but instead would simply reflect investment’s relationship with other

2There is slight variation in how the question was worded across samples. Within the BEEPS data, the question was: “ What percent of senior management’s time over the last 12 months was spent in dealing with public oﬃcials about the application and interpretation of laws and regulations and to get or to maintain access to public services?” For countries in the WBES, the question was: “In a typical week, what percentage of senior management’s time is spent in dealing with requirements imposed by government regulations [e.g. taxes, customs, labor regulations, licensing and registration] including dealings with oﬃcials, completing forms, etc.?”

192 5.0

●

4.5 ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ●● 4.0 ● ● ● ● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 3.5 ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● 3.0 ● ● ● ● ''Are Laws Applied Unpredictably?'' Applied Unpredictably?'' Laws ''Are

(country means; higher = more unpredictable) 2.5 ●

0 10 20 30 Time Spent with Officials (standard deviation within countries)

Note: Firm-level survey data from World Bank Enterprise Surveys (WBES) from various years, 2002- 2006. Triangles represent OECD countries; dots are non-OECD countries. The x-axis is bureaucratic discretion, as measured by the within-country variance to the following question: “In a typical week, what percentage of senior management’s time is spent in dealing with requirements imposed by government regulations [e.g. taxes, customs, labor regulations, licensing and registration] including dealings with officials, completing forms, etc.?” The y-axis is a measure of policy uncertainty, country means on a 6-point scale agreeing or disagreeing with the statement: “Interpretations of the laws and regulations affecting my firm are consistent and predictable.”

Figure D.2: Policy Uncertainty & Variance in Time with Oﬃcials (WBES Data)

confounding factors, such as respondents’ underlying optimism or general attitude towards their location the underlying optimism of the respondent.

In this section, I identify a set of potential confounding factors that could bias my ﬁndings by inﬂuencing respondents’ perceptions of bureaucratic discretion. These factors include respondents’ lack of political knowledge, lack of external reference for comparison, and their underlying optimism. I address each concern separately, investigating whether the dissertation’s main empirical results continue to hold once the empirical models control for measures that are associated with each of these

193 alternate scenarios. Although such tests cannot provide deﬁnitive proof of the survey measures’ validity, they can shed light on whether or not the measure is hampered by conditions most likely to undermine internal validity.

Lack of Political Knowledge

The survey question that I use to measure bureaucratic discretion assumes implicitly that enterprise managers know something about the workings of their regions’ political and government oﬃces. Without this knowledge, respondents’ perceptions about regional bureaucracies are unlikely to be grounded in actual levels of discretion. If a lack of knowledge is driving the results in the body of the dissertation, then controlling for measures associated with lack of political knowledge should diminish any spurious relationships caused by respondents’ inability to answer the question knowledgeably.

I investigate the sensitivity of the main ﬁndings to controlling for measures that might reasonably correlate with respondents’ political knowledge. As mentioned earlier, the survey contains a set of six questions that ask respondents about political leaders’ views on economic competition, political competition, due process and pre- sumed innocence, and cooperation with the US and the international community.

Based on the assumption that missing values on these questions demonstrate a lack of political knowledge, I code the number of missing values that each respondent has across these six questions for their regional governor. The initial findings are robust to including this information in a number of ways: as a count for the total number of political questions that respondents skip, as a dummy variable for skipping all questions about the governor, or as a dummy variable for respondents that skip any of the questions. As a further check for whether results are influenced by respondents’ lack of political knowledge, I drop all respondents who skipped more than one out of the six political questions identified above. Table D.1 displays the results from these analyses. 194 Firm Investment No. Missed Any Skipped Dropping dummy, 1 = firm plans to invest during coming year Political Skipped All Low Questions Questions Questions Knowledge Bureaucratic Discretion -0.654*** -0.644*** -0.659*** -0.801*** 1 = no discretion, 4 = high discretion (0.158) (0.157) (0.158) (0.194) Frequent Changes to Laws 0.116 0.117 0.111 0.227* 1 = no obstacle, 5 = very serious obstacle (0.112) (0.112) (0.112) (0.137) High Tax Rates -0.383*** -0.385*** -0.383*** -0.571*** 1 = no obstacle, 5 = very serious obstacle (0.122) (0.121) (0.122) (0.151) Regional Administration 0.510*** 0.508*** 0.522*** 0.436* 1 = poor job, 5 = excellent job (0.197) (0.196) (0.197) (0.264) Regional Courts -0.161 -0.151 -0.165 -0.245 1 = poor job, 5 = excellent job (0.148) (0.148) (0.149) (0.180) Regional Governor -0.378** -0.373** -0.377** -0.389 1 = poor job, 5 = excellent job (0.187) (0.186) (0.186) (0.248) Access to Finance -0.023 -0.028 -0.02 -0.051 1 = no obstacle, 5 = very serious obstacle (0.083) (0.083) (0.083) (0.099) Labor Shortages -0.056 -0.048 -0.057 -0.108 1 = no obstacle, 5 = very serious obstacle (0.087) (0.087) (0.087) (0.108) Competitive Pressures 0.104 0.092 0.103 0.158 1 = no obstacle, 5 = very serious obstacle (0.086) (0.085) (0.086) (0.104) Privatized Firm 0.12 0.093 0.113 0.07 dummy, 1 = privatized, former SOE (0.277) (0.279) (0.276) (0.320) Annual Sales 0.469** 0.460** 0.472** 0.766*** -1 = decreasing, 1 = increasing (0.184) (0.183) (0.184) (0.231) Firm Size 0.311*** 0.307*** 0.310*** 0.325*** number of employees (logged) (0.083) (0.083) (0.084) (0.098) Private Firm 0.719 0.759 0.726* 0.639 dummy, 1 = private ownership (0.442) (0.443) (0.441) (0.516) Lack of Political Knowledge -0.071 -0.079 -0.503 non-response to political questions (0.055) (0.245) (0.351) Constant -2.046 -2.126 -2.052 -1.295 (1.440) (1.438) (1.438) (1.736)

GDP per capita 0.441* 0.441* 0.437* 0.527* in constant 2000 rubles per 1000 persons (logged) (0.253) (0.252) (0.253) (0.303) Regional Political Competition -0.181 -0.183 -0.178 -0.251 index, -3 = uncompetitive, 5 = highly competitive (0.128) (0.128) (0.128) (0.154) Competition × Discretion 0.140** 0.142** 0.139** 0.142* interaction (0.066) (0.066) (0.066) (0.078) Log-likelihood -234.612 -235.38 -234.383 -166.558 AIC 511.224 512.759 510.765 373.117 No. of Cases 403 403 403 297 Note: Survey data from Frye (2006). Region-level political data from the Moscow Carnegie Center; data on GDP per capita from Rosstat. Lack of political knowledge measured by non-response to six questions about the regional executive’s view: Column 1 is a count of the number of non-responses (0-6), Columns 2-3 are dichotomous indicators, Column 4 drops respondents missing more than one political question. Coefficients from multilevel logistic regression with random coefficient for the discretion variable and random intercepts at the region level; standard errors in parentheses. Out of space concerns, unit-specific effects not reported. * p<0.10, ** p<0.05, *** p<0.01

Table D.1: Controlling for Lack of Political Knowledge 195 Firm managers might learn about politics and bureaucratic affairs through contact with the government. As such, I re-estimate the baseline model using firms’ reported experiences with government as an alternative way to control for managers’ knowledge of bureaucratic discretion. Specifically, I add control variables indicating whether

firms report to having lobbied at any level of government, for whether or not they have received government subsidies in the last two years, and for having had a dispute with government agencies in the recent past. A positive response to any of these questions would potentially indicate that a firm manager has had close contact with government bodies and experiences that would shed light on how much decision- making discretion regional bureaucrats enjoy. Controlling for these experience-based indicators of political knowledge does not change the findings; the first column in

Table D.2 displays the results of these tests.

Alternatively, it is possible that understanding the nuances behind bureaucratic operations comes gradually, and that not all respondents have sufficient on-the-job or life experience to answer the discretion question properly. Consequently, I examine the results’ sensitivity to adding controls for years that respondents have lived in the city where the firm is located, years in their position within the company, education level, age, and whether or not managers had previously worked in a different sector or for a state-owned enterprise. If respondents’ answers to the bureaucratic discretion question are correlated with these proxies for personal perspective, we would expect the results to attenuate. Table D.2’s second column shows that (once again) the main empirical results hold in the presence of these controls.

In summary, using multiple tests and controlling for a large number of diverse factors that plausibly correlate with respondents’ familiarity with regional politics

196 Firm Investment Experience w/ Experience w/ dummy, 1 = ﬁrm plans to invest during coming year Government Job & Location

Bureaucratic Discretion -0.636*** -0.633*** 1 = no discretion, 4 = high discretion (0.164) (0.161) ...... control variables suppressed ...... Lobbying Experience 0.354 dummy, 1 = lobbied at any level of government (0.240) Chances of Receiving Government Bailout 0.037 1 = never, 4 = deﬁnitely (0.147) Past Court Dispute with Government 0.367 dummy, 1 = has used courts (0.255) Years in City -0.049 years lived in ﬁrm’s city (logged) (0.172) Years In Position -0.218 years in current position (logged) (0.184) Education 0.392 1 = secondary, 4 = PhD (0.260) Age -0.005 age of respondent (0.015) Outside Experience 0.291 dummy, 1 = experience in another sector (0.245) Past Experience at State-Owned Enterprise 0.318 dummy, 1 = management worked for SOE (0.287) Constant -1.904 -2.643 (1.482) (1.733)

GDP per capita 0.430* 0.467* in constant 2000 rubles per 1000 persons (logged) (0.258) (0.256) Regional Political Competition -0.179 -0.223* mean-centered index, -3 = uncompetitive, 5 = highly competitive (0.130) (0.134) Political Competition × Bureaucratic Discretion 0.139** 0.167** interaction (0.067) (0.070)

Log-likelihood -229.86 -231.017 AIC 505.721 514.034 No. of Cases 397 403 Note: Survey data from Frye (2006). Region-level political variables come from the Moscow Carnegie Center’s Regional Monitoring Project. Data on GDP per capita from annual Rosstat publications. Coefficients from multilevel logistic regression with random coefficient for the discretion variable and random intercepts at the region level; standard errors in parentheses. The model also includes additional control variables contained (for specification, see Model 6 in the body of the paper), but they, along with unit-specific effects, are not reported to conserve space. * p<0.10, ** p<0.05, *** p<0.01

Table D.2: Controlling for Experience with Government & Job

197 and bureaucratic affairs, I find no evidence that respondents’ political knowledge or lack thereof is driving the findings reported in this dissertation.3

3These findings corroborate an observation I made while conducting field interviews with firm directors and business associations in Russia. In the course of my interviews, I found that many business actors are demonstrably savvy about their political environment and have a solid understanding of how the legal rules that affect their business interests are made. 198 Lack of External Reference

In addition to worrying about respondents’ knowledge of government aﬀairs, there exists another concerns regarding respondents’ ability judge whether or not bureaucrats make decisions independent of other government bodies. Respondents might not understand the possible range of independence available to bureaucrats, particularly if they have never done business out of their speciﬁc region. To gauge whether

firms’ perceptions of bureaucratic discretion are biased by their lack of external reference, I re-estimate the baseline model controlling for measures that could potentially correlate with firms’ inexperience with affairs outside the region.4

Membership in business associations is one pathway by which firm managers might learn about variation in bureaucratic discretion across regions; thus, I add a statistical control for membership in such an organization. I control for exporting status, based on the logic that exporters conduct business outside their home region, and these interactions may give them additional information about the potential range of bureaucratic discretion. In the hopes of accounting for respondents that may have lived in other regions, I also code an indicator for firm managers that have lived for less than ten years in their firm’s city. I also include a measure for firms that report that their main rivals are firms from other regions, on the off-chance that they may have learned something about other regions through interactions with their competitors.

Finally, in another attempt to measure firms’ operations in other regions, I include an ordinal measure for the percentage of firm sales that come from their home region. If inexperience outside the region is driving firms’ perceptions of bureaucratic discretion in a way that biases results, controlling for these factors should alter the

4A design-based strategy for dealing with this problem would be to use an anchoring vignette that would provide respondents with a sense of the possible range of discretion. In theory, such priming creates explicit reference points for respondents to then use in their own evaluations of regional bureaucracy . Given that this strategy is unavailable once the survey data have been collected, I pursue a second-best approach here.

199 dissertation’s findings. The first column in Table D.3 shows that adding these controls to the statistical analyses does not change the substantive or statistical findings.

As an additional check, I re-estimate the baseline model using the subset of firms that are most likely to have had experiences that would provide a useful reference for the scale of bureaucratic discretion. Using the percentage of sales within the home region as a proxy for firm managers’ inexperience with regulatory arrangements outside their own region, I keep only those firms that make 50% or more of their sales within their home region. If the inexperienced firms’ perceptions were confounding the results in one direction or another, then we would expect the statistical models to produce different results once those respondents have been dropped. However, the second column in Table D.3 displays no meaningful change in the main results. The main findings of the dissertation do not appear to be driven by inexperienced firms’ inability to judge bureaucrats’ independent decision-making.

Respondent Optimism

Previous sections have found no evidence that the dissertation’s empirical results are an artifact of respondents’ inability to answer the survey question about bureaucratic discretion accurately. One major reason for concern is that, in the absence of real knowledge, ﬁrm managers’ responses might simply reﬂect general feelings of optimism

(or pessimism). Since such feelings probably also correlate with investment plans, there is a danger that my ﬁndings could be spurious.

In anticipation of these concern, my original models already control for respondents’ attitudes towards regional political institutions, including a rating of the governor, economic courts, and regional administration. To the degree that these institutions have any “halo effects” on firm managers’ perceptions of bureaucrats and/or their decisions to invest, the statistical controls should help account for this kind of attitudinal spillover. In a similar fashion, other controls in the models should be 200 Firm Investment Experience Dropping Firms dummy, 1 = firm plans to invest during coming year Outside Region w/o External Sales

Bureaucratic Discretion -0.692*** -0.804** 1 = no discretion, 4 = high discretion (0.164) (0.346) Frequent Changes to Laws 0.127 -0.208 1 = no obstacle, 5 = very serious obstacle (0.115) (0.228) High Tax Rates -0.335*** -0.238 1 = no obstacle, 5 = very serious obstacle (0.128) (0.244) Regional Administration 0.546*** 1.187*** 1 = poor job, 5 = excellent job (0.201) (0.412) Regional Courts -0.108 0.169 1 = poor job, 5 = excellent job (0.153) (0.303) Regional Governor -0.405** -0.883** 1 = poor job, 5 = excellent job (0.191) (0.366) Access to Finance 0.006 -0.097 1 = no obstacle, 5 = very serious obstacle (0.086) (0.155) Labor Shortages -0.087 -0.046 1 = no obstacle, 5 = very serious obstacle (0.091) (0.164) Competitive Pressures 0.096 -0.171 1 = no obstacle, 5 = very serious obstacle (0.088) (0.158) Privatized Firm 0.069 -0.721 dummy, 1 = privatized, former SOE (0.284) (0.586) Annual Sales 0.550*** 0.724** -1 = decreasing, 1 = increasing (0.192) (0.367) Firm Size 0.265*** 0.286 number of employees (logged) (0.096) (0.171) Private Firm 0.606 1.814 dummy, 1 = private ownership (0.449) (0.936) Exporting Experience -0.387 -1.581*** dummy, 1 = exporter (0.315) (0.509) In-Region Sales 0.240** 1 = no in-region sales, 4 = 100% of sales in-region (0.114) Business Association 0.389 dummy, 1 = membership in business association (0.245) Recent Move-In -0.002 dummy, 1 = lived in city < 10 yrs. (0.482) Constant -1.949 -3.833 (1.546) (3.417)

GDP per capita 0.331 1.509** in constant 2000 rubles per 1000 persons (logged) (0.267) (0.602) Regional Political Competition -0.169 -0.430 mean-centered index, -3 = uncompetitive, 5 = highly competitive (0.131) (0.322) Political Competition × Bureaucratic Discretion 0.130* 0.326 interaction (0.067) (0.180) Log-likelihood -221.744 -73.133 AIC 491.488 188.265 No. of Cases 383 153 Note: Survey data from Frye (2006). Region-level political variables come from the Moscow Carnegie Center’s Regional Monitoring Project. Data on GDP per capita from annual Rosstat publications. Coefficients from multilevel logistic regression with random coefficient for the discretion variable and random intercepts at the region level; standard errors in parentheses. Out of space concerns, unit-specific effects not reported. Column 2 drops all firms with more than 50% of sales originating in home region. * p<0.10, ** p<0.05, *** p<0.01

Table D.3: Controlling for Experience Outside the Region

201 picking up on economic factors that would be related to ﬁrms’ overall optimism regarding their own economic position, including control variables for problems with labor, problems with economic competition, and past sales performance.

On top of this relatively high baseline, I re-estimate models that include additional controls for respondents’ optimism. To account for respondents’ optimism regarding their own firm’s economic position, I include two measures: firms’ assessment of their economic fortunes over the last three years, and firms’ predictions about their economic fortunes for the coming three years. The survey contains related questions capturing firm managers’ sentiments towards Russia’s economic health, asking firms about whether the country’s economic performance improved in the recent past and whether they expect it to do so in the immediate future. While these latter questions do not speak directly to the regional environment, they do control for optimism about the national economy that was previously unaccounted for in the analyses.

Table D.4 shows that neither set of additional controls for respondents’ optimism change the statistical results in any meaningful way. Assuming that respondents’ optimism been modeled appropriately in the original analyses or captured by the robustness checks, I ﬁnd no evidence that respondent optimism is driving the dissertation’s main ﬁndings.

202 Firm Investment Optimism About Optimism About dummy, 1 = ﬁrm plans to invest during coming year Economy Own Firm

Bureaucratic Discretion -0.662*** -0.641*** 1 = no discretion, 4 = high discretion (0.171) (0.173) Frequent Changes to Laws 0.132 0.079 1 = no obstacle, 5 = very serious obstacle (0.117) (0.120) High Tax Rates -0.373*** -0.392*** 1 = no obstacle, 5 = very serious obstacle (0.130) (0.132) Regional Administration 0.538** 0.577*** 1 = poor job, 5 = excellent job (0.213) (0.216) Regional Courts -0.176 -0.214 1 = poor job, 5 = excellent job (0.156) (0.164) Regional Governor -0.378* -0.410** 1 = poor job, 5 = excellent job (0.204) (0.203) Access to Finance -0.036 -0.032 1 = no obstacle, 5 = very serious obstacle (0.088) (0.092) Labor Shortages -0.037 -0.053 1 = no obstacle, 5 = very serious obstacle (0.093) (0.095) Competitive Pressures 0.105 0.068 1 = no obstacle, 5 = very serious obstacle (0.092) (0.093) Privatized Firm 0.02 -0.044 dummy, 1 = privatized, former SOE (0.299) (0.304) Annual Sales 0.496** 0.410* -1 = decreasing, 1 = increasing (0.196) (0.229) Firm Size 0.313*** 0.313*** number of employees (logged) (0.087) (0.087) Private Firm 0.708 0.708 dummy, 1 = private ownership (0.464) (0.467) Economy’s Past Performance (last 2 yrs. ) 0.317 1 = improved greatly, 4 = deteriorated greatly (0.255) Economy’s Future Performance (next 2 yrs. ) -0.100 1 = will improve greatly, 4 = will deteriorate greatly (0.296) Firm’s Past Financial Position (last 2 yrs. ) -0.050 1 = improved greatly, 4 = deteriorated greatly (0.258) Firm’s Future Financial Position (next 2 yrs. ) 0.357 1 = will improve greatly, 4 = will deteriorate greatly (0.287) Constant -2.703* -2.299 (1.578) (1.754)

GDP per capita 0.419 0.387 in constant 2000 rubles per 1000 persons (logged) (0.274) (0.269) Regional Political Competition -0.229* -0.212 mean-centered index, -3 = uncompetitive, 5 = highly competitive (0.135) (0.135) Political Competition × Bureaucratic Discretion 0.153** 0.149** interaction (0.069) (0.068) Log-likelihood -205.828 -201.983 AIC 455.657 447.965 No. of Cases 352 346 Note: Survey data from Frye (2006). Region-level political variables come from the Moscow Carnegie Center’s Regional Monitoring Project. Data on GDP per capita from annual Rosstat publications. Coefficients from multilevel logistic regression with random coefficient for the discretion variable and random intercepts at the region level; standard errors in parentheses. Out of space concerns, unit-specific effects not reported. * p<0.10, ** p<0.05, *** p<0.01

Table D.4: Controlling for Respondent Optimism

203 External Validity

Validating Perceptions of Discretion

Regional Diﬀerences: ANOVA test

One preliminary test of the discretion measure’s external validity involves investi-

gating the degree to which responses vary by region. If respondents’ perceptions do

reﬂect actual levels of bureaucratic discretion and this discretion varies across regions,

then patterns in the data should show that respondents in diﬀerent regions answer

the question diﬀerently. If the data do not reveal diﬀerences in aggregate percep-

tions across regions, then it would be harder to claim that responses to the perceived

discretion question reﬂect actual levels of discretion among regional bureaucrats.

Sum of Squares df Mean Square F-stat p-value Between Groups 50.824 10 5.082 7.140 0.000 Within Groups 402.176 565 0.712 Total 453.000 575 Note: Survey data from Frye (2006). Results of the one-way ANOVA on this bureaucratic discretion measure indicates that the diﬀerence in means across these 11 regions are statistically distinct. More formally, the test rejects the null hypothesis that means across groups are statistically indistinguishable.

Table D.5: Testing Regional Diﬀerences in Perceptions of Discretion: ANOVA

A one-way ANOVA on this measure indicates that diﬀerences across these 11

regions are statistically distinct. More formally, the ANOVA test rejects the null

hypothesis that variances across groups are the same: df = 10, F = 7.140, and p <

0.000. This suggests strongly that, even though we do observe variation in perceptions within regions, perceptions within region do cluster suﬃciently that we can identify

204 diﬀerences in regional aggregates. These results bolster our conﬁdence in the validity

of the survey question. In the next section, I will show that these same regional

aggregates also correlate with “hard” measures of the regulatory environment, giving

us greater conﬁdence about the measure’s external validity.

An External Measure of Discretion

The previous section’s ANOVA results indicate that respondents’ perceptions hang

together in statistically distinct patterns across regions. For the standpoint of external

validity, what we would really like to know is whether those patterns reﬂect underlying

levels of discretion that are unique to each region. Is there evidence that these patterns

reﬂect real diﬀerences in the level of bureaucratic discretion across regions?

In this section, I use an original, external measure of bureaucratic discretion to

validate the survey-based measure of perceived discretion. Speciﬁcally, as a measure

of the statutory constraints that restrict bureaucrats’ freedom to interpret and apply

laws in a discretionary manner, I code the number of legal documents issued by the

governor and regional administration over the three years prior to the survey, 2002-

2004.5 The logic behind this measures is that bureaucrats’ discretion in regulating economic activity should relate inversely to the extensiveness with which the body of legal code speciﬁes procedural behavior and decision-making rules.6 Thus, as the

5These data were collected from Konsultant-Plus, a Moscow-based legal services company that archives Russian legislation. The counted documents are legally-binding documents entitled postanovleniya. While I have translated this as ‘directive,’ alternate translations might also include ‘resolution’ or ‘statement.’ They cover a wide range of issues, including important economic policies such as taxes and safety standards.

6While prominent research has used a similar argument for focusing on the speciﬁcity of individual pieces of legislation (Huber & Shipan 2002), a similar dynamic holds for the completeness of the body of law within a particular policy environment. Given extensive and oft-updated guidelines, the parameters for individual interpretation or arbitrary application are narrower; in contrast, policy application is likely to remain much more open to interpretation in regions characterized by little law-making activity.

205 number of directives from regional executives grows, the increasing body of detailed guidelines should, on average, place greater constraints regulatory bureaucrats’ ability to make decisions in interpreting and applying laws that are independent of other government bodies.

● 60

● 30 ● ●

● ●

Percent of Firms Perceiving Percent 20 High Bureaucratic Discretion High Bureaucratic ● ● ● ● 10 ●

0 500 1000 1500 2000 2500 Number of Executive Directives (2002−2004)

Note: Survey data from Frye (2006); data on volume of legal documents by regional executive collected by author from Konsultant-Plus. Regression line drops the Bashkortostan, the inﬂuential outlier at the very top of the graph. Including Bashkortostan only strengthens the correlation.

Figure D.3: Perceptions of Discretion versus Regulatory Executive Directives

Figure D.3 plots the regional averages for the survey’s discretion question on the number of legal documents issued by the regional executive. As anticipated, the data reveal an inverse relationship between aggregated perceptions of discretion 206 and the volume of legislative constraints on executive agents. In regions with a high volume of executive directives, average perceptions of bureaucratic discretion are indeed lower. Comparing the perception-based survey measure of bureaucratic discretion against a defensible and independent measure of bureaucratic discretion, I

ﬁnd evidence supporting the survey question’s external validity.

Validating Intent to Invest

Like the survey measure of bureaucratic discretion, the analyses’ dependent variable is also attitudinal. Firm managers’ report about whether or not they plan on investing in the next twelve months, but we do not observe actual investment. For various reasons, actual investment patterns might deviate from reported plans to invest.7

Thus, from the standpoint of external validity, we might wonder how closely reported investment matches up with actual investment patterns.

The anonymous nature of the survey makes precludes validating the measure by following up with ﬁrms to ﬁnd out whether or not they actually invested. Instead,

I examine aggregate data on private investment from the eleven sampled regions to see if higher average reported investment from the survey correspond with regional data on private investment taken from government economic statistics. A positive relationship between these two measures would indicate that reported investment does track along with actual investment patterns, increasing our conﬁdence in the measure’s external validity. A negative relationship or no relationship would suggest that the reported investment measure within the data has weak external validity.

Figure D.4 plots the percentage of surveyed ﬁrms in a given region that express an

7For an analogous problem, see the U.S. voting literature on reported voting versus actual voting.

207 50000 ● 10.5 ●

40000 10.0 ●

● ● 30000 9.5 ●

9.0 ● 20000 ● ● ● ● ● ● 8.5

10000 ● ● ●

8.0 ● ● ● ● ● ● Regional Private Investment (mil. of rubles) Investment Regional Private 0

0 10 20 30 40 50 60 logged (mil. of rubles) Investment, Regional Private 0 10 20 30 40 50 60

Percent of Firms with Intent to Invest Percent of Firms with Intent to Invest

Note: Survey data from Frye (2006); regional data on private investment for the year 2005 taken from Rosstat.

Figure D.4: Intentions to Invest and Regional Private Investment (2005)

intent to invest against actual observed values of regional private investment.8 Look-

ing at Figure D.4, we see a positive relationship between the percentage of surveyed

ﬁrms intending to invest and observed levels of private investment within regions for

the year directly following the survey. These ﬁndings increase our conﬁdence that

patterns within the survey data are representative of larger dynamics in the economy.

Summary

In this appendix, I have conducted extra analyses and robustness checks to address

speciﬁc concerns about the empirical analyses’ use of survey data. In addition to the

8These data come from 2005 in order to capture the majority of the twelve month period directly following the Frye survey’s completion. Using 2006 investment data, the plots look very similar.

208 qualitative evidence from my field interviews in Russia, I have shown that additional survey data demonstrates a link between measures of bureaucratic discretion and increased unpredictability in the application and interpretation of laws. I have also delved deeper into the vulnerabilities of my perception-based measure of bureaucratic discretion to previously unmodeled factors that could potentially confound my findings. The relationship between the reported perceptions of bureaucratic discretion and firms’ investment plans do not appear to be directly driven by respondents’ optimism about about regional political institutions, optimism about their own firm’s economic position, lack of political knowledge, inexperience with government, inexperience with other regions, or respondent’s demographic information. Taken together, such results provide encouraging support for the study’s internal validity. Further- more, I find empirical evidence that, once aggregated, perceptions of bureaucratic discretion show distinct differences across regions. This suggests that, although individual respondents within regions may have differing perceptions, their perceptions do hang together in discernible patterns that suggest a latent level of discretion that is unique to their region. After conducting further analyses, I find that regional aggregate perceptions of bureaucratic discretion are inversely related to the number of legal documents passed by regional executives; on average, firms’ perceive bureaucrats to have less independent decision-making ability as the volume of legal code within a region grows and places greater constraints on agents’ behavior. This suggests strongly that respondents’ perceptions of bureaucratic discretion appear to have some basis in actual regulatory conditions within the regions.

209