Is Corruption Curable?

Is Corruption Curable? KHALID SEKK SEKKAT SEK KAT

Is Corruption Curable?

Khalid Sekkat

Is Corruption Curable?

Khalid Sekkat Centre Emile Bernheim University of Brussels Brussels, Belgium

ISBN 978-3-319-98517-6 ISBN 978-3-319-98518-3  (eBook) https://doi.org/10.1007/978-3-319-98518-3 Library of Congress Control Number: 2018951566 © The Editor(s) (if applicable) and The Author(s) 2018 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Cover credit: YAY Media AS/Alamy Stock Photo This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

To Houda, my wife and best friend

Contents

Part I  Corruption, Extent, Causes and Consequences 1

Definition, Amount, and Coverage 5

2

Measurement Issues 39

3 Causes 71 4 Consequences 119 Part II  Anti-corruption Strategies: The Role of the State 5 Democracy 165 6

Electoral Rules 179

7 Decentralization 189 8 Regulation 203 vii

viii     Contents

9 Justice 211 10 Specialized Anti-corruption Agencies 231 11 Incentives and the Corruption Market 241 12 International Cooperation 265 Part III Anti-corruption Strategies: The Role of Civil Society 13 Civil Society and the Media 283 14 Civil Society and the Specific Role of ICT 299 15 Civil Society and the Role of Education 311 Conclusion 337 Index 341

Abbreviations

ACA ARA AU BEC BEEPS BHPS BPI CCI CIS CPI CPIB CPS CRF CSO DoJ DPA DR EBRD EDI EIU ESS

Anti-Corruption Agencies Autonomous Revenue Authorities African Union British Election Studies Business Environment and Enterprise Performance Survey British Household Panel Study Bribe Payers Index Control of Corruption Index Commonwealth of Independent States Corruption Perceptions Index Corrupt Practices Investigation Bureau Current Population Survey Clean Report of Finding Civil Society Organization Department of Justice Deferred Prosecution Agreement Discrepancy Report European Bank for Reconstruction and Development E-government Development Index Economist Intelligence Unit European Social Survey ix

x     Abbreviations

EU FCPA FERA FH GRC GSS GVS HF IACA IACAC ICAC ICRG ICT ICVS IMF IOC KICAC NES NFE NLSY NPA OAS OECD OPEN PERC PETS PSI QSDS RTIA SEC UNCAC UNCTAD WB WCR WEF WES WGI WJP WTO WVS

European Union Foreign Corrupt Practices Act Federal Emergency Relief Administration Freedom House Global Competitiveness Report General Social Surveys General Value Survey Heritage Foundation International Anti-Corruption Academy Inter-American Convention Against Corruption Independent Commission Against Corruption International Country Risk Guide Information and Communication Technology International Crime Victimization Surveys International Monetary Fund International Olympic Committee Korea Independent Commission Against Corruption National Election Studies Non-Formal Education National Longitudinal Survey of Youth Non-Prosecution Agreement Organization of American States Organization of Economic Cooperation and Development Online Procedures Enhancement Political and Economic Risk Consultancy Public Expenditure Tracking Surveys Pre-Shipment Inspection Quantitative Service Delivery Surveys Right to Information Act Securities and Exchange Commission United Nations Convention Against Corruption United Nations Conference on Trade and Development World Bank World Competitiveness Report World Economic Forum World Enterprise Surveys World Governance Indicators World Justice Project World Trade Organization World Values Survey

List of Tables

Chapter 2 Table 1 Comparison of country rankings based on perception indicators in 2010 49 Table 2 Comparison of country rankings: Experience versus perception 51 Table 3 Change in country ranking over time 61

xi

Introduction

Notwithstanding some dissenting voices, which will be discussed later in the book, corruption is now widely recognized as a major “disease” which threatens not only economic development but also the very foundations of societies. Numerous publications have examined the causes and consequences of corruption but almost none has offered a deep analysis of possible cures. This book aims to fill that gap. It discusses the solutions that have been adopted in different countries and at the international level to curb corruption. The objective is to explore possible effective means to reduce, as much as possible, the prevalence of corruption. In doing so, our hope is not to suggest a universally effective way to eradicate corruption. After all, the problem can be traced back to before 300 BC and affects almost all human activities. Rather, our purpose is more modest. By analyzing the different solutions that have been adopted around the world and the reasons for their success or failure, we seek to provide national and international policymakers with guidance on possible effective anti-corruption strategies. Such strategies can consist in adopting existing successful solutions, correcting unsuccessful ones, or developing other solutions which take into account the reasons why previous attempts failed. xiii

xiv     Introduction

The book is written with the vision that the different components of society are interrelated and that the well-functioning of each one is necessary to the well-functioning of the group as a whole. This is referred to as the “O-ring” theory of economic development, after the explosion of the Space Shuttle Challenger, which was caused by the failure of a small and relatively cheap component called an O-ring. This example highlights the complementarity between the components of a system and the importance of each of them (even the smallest or cheapest) in the sound functioning of the whole structure. To better illustrate this vision, let us consider the following observations. Autocratic regimes are associated with a high prevalence of corruption while democratic regimes are much less affected. However, democracy can also foster corruption because, for instance, election campaigns require funding, and more competitive elections may make political parties and candidates vulnerable to pressure from donors. For this reason, many countries have adopted laws to control campaign financing. The extent to which these laws are effective in curbing political corruption depends on the quality of the judicial system. If the rules are violated, opponents can call for justice. If the justice system is corrupt or under political influence, the appeal will not be effective. In such cases, opponents can alert the media with the objective of making citizens protest and ensuring that the law is upheld. But alerting the media will deliver results only if the media are not also corrupt or under influence. It is also important that citizens are educated enough to access, process, and use information in an efficient way in order to achieve the target of reducing corruption, which illustrates the importance of education in the field of anti-corruption. While this simple example is sufficient to illustrate the complementarity of different factors in the fight against corruption, the book offers a deeper discussion of other mechanisms which link the effectiveness of various strategies. The book is divided into three main parts. The first is devoted to the extent, causes, consequences, and persistence of corruption. It shows that no activity in society is immune to corruption, that the motives of corruption are complex and diverse, that the phenomenon has serious economic, social, and political consequences, and that corruption is persistent. Such persistence reveals the failure of anti-corruption strategies. It also hampers

Introduction     xv

their success. The following two parts of the book focus on the cure for corruption. Part II examines the institutional dimension, including the role of democracy, political parties, justice, decentralization, regulation, and international cooperation. The third and final part discusses the role of civil society, including citizens, the media, NGOs, social networks, and education. These institutional and societal dimensions together constitute the main pillars of an effective anti-corruption strategy. It is shown that addressing the problem of corruption through one dimension only is ineffective. Each of the three parts is composed of a number of chapters. Each chapter starts with a conceptual discussion of the issue at hand, explaining the logic and reasons underlying the issue in an informal way. While the explanation draws on formal economic studies (including theoretical and empirical models), it is non-technical in order to allow the reader to understand the problem without specific knowledge of mathematics or econometrics. The conceptual part is followed by an empirical discussion providing real-life illustrations as well as the results of more formal and comprehensive analyzes. The empirical discussion is also based on rigorous academic studies but presented in an informal style for the same reason as above. This book is intended for practitioners in national and international organizations, NGOs, students, and academics. In recent decades, concerns about corruption have increased so dramatically that nowadays almost every international organization has an anti-corruption department or program. A non-exhaustive list would include: • The United Nations (United Nations Convention Against Corruption, UNCAC). • The Council of Europe (Group of States against Corruption, GRECO). • The World Bank (Global Governance Practice). • The OECD (OECD Convention Against Corruption and OECD Convention on Combating Bribery of Foreign Public Officials in International Business Transactions). • The European Union (Convention on the Protection of the European Communities’ Financial Interests). • The Organization of American States (Inter-American Convention Against Corruption).

xvi     Introduction

In addition to the international organizations, a number of NGOs are working on corruption issues. Examples include: • Transparency International based in Geneva. • Freedom House based in Washington. • World Economic Forum based in Geneva. At the national level, a large number of countries have set up independent Anti-Corruption Agencies. A non-exhaustive list includes Albania, Argentina, Austria, Australia, Brunei, Brazil, Bulgaria, China, Colombia, Ecuador, El Salvador, Hong Kong, India, Indonesia, Korea, Malaysia, Mexico, New Zealand, Paraguay, Peru, Philippines, Sierra Leone, Singapore, South Africa, Thailand, and the USA. Finally, and probably more significantly, an International AntiCorruption Academy (IACA) was recently established. Based in Laxenburg, Austria, the IACA is dedicated to overcoming current shortcomings in knowledge and practice in the field of anti-corruption and seeks to empower professionals for the challenges of tomorrow. Students are important future players in society. Besides improving their exam marks, they can benefit from this book by gaining worthwhile information for their future jobs. The book will also be valuable to researchers working on corruption, by suggesting issues that remain underexamined but deserve further investigation. Last but not least, the book offers useful information and references to any teachers of development economics, institutional economics, or macroeconomics.

Part I Corruption, Extent, Causes and Consequences

Corruption is an old, widespread, and multifaceted phenomenon. In an early survey on corruption, Bardhan (1997)1 cites a treatise on public administration dating back to the fourth century BC: “Just as it is impossible not to taste the honey (or the poison) that finds itself at the tip of the tongue, so it is impossible for a government servant not to eat up, at least, a bit of the king’s revenue. Just as fish moving under water cannot possibly be found out either as drinking or not drinking water, so government servants employed in the government work cannot be found out (while) taking money (for themselves)” (Bardhan 1997, p. 1320). Mishra (2006) notes that references to corruption can be found in even more ancient sources, such as the Code of Hammurabi, King of Babylon (twenty-second century BC), or the Eddict of Harmhab, King of Egypt (fourteenth century BC). Maennig (2005) reports cases of corruption dating back to 388 BC. One case is attributed to the athlete Eupolos of Thessalia, who successfully bribed three of his competitors in the first combat tournament at the Olympic Games. Another early case 1Other

very interesting surveys are Aidt (2003) and Jain (2001).

2     Part I: Corruption, Extent, Causes and Consequences

of corruption also concerns sports. In twelfth-century BC, Damonikos of Elis, father of the Olympic wrestler Polyktor, attempted to bribe the father of the opponent to persuade his son to concede victory in the Olympic wrestling competition to Polyktor. Looking closer to our own era, Mishra (2006) observes that corruption emerged in China under the Ming dynasty in the fourteenth century and continued spreading during the Qing dynasty through the nineteenth century. There were many attempts to curb corruption, including large-scale salary reforms, but the problem persisted. Similarly, corrupt practices were widespread during the seventeenth century in Florence despite the presence of fairly repressive anti-corruption laws. Corruption is not only old but highly widespread. Almost no country or human activity around the world escapes the problem. In developing countries, it is almost impossible to find someone who has never been directly or indirectly (through parents or other relatives) asked for some kind of “inappropriate” payment to get something done. Although in developed countries this kind of corruption is very rare, it is equally difficult to find any country where no politician or high-ranking official has ever been convicted of taking some money from a firm or other “benefactor”. Looking at the recent release of the three most commonly used indicators of the prevalence of corruption, namely Transparency International’s Corruption Perceptions Index, the indicators provided by the International Country Risk Guide, and the Worldwide Governance Indicators, no country, developed or developing, attains the best possible score. Moreover, contrary to common belief, corruption is not limited to bureaucrats and politicians. Various studies document the prevalence of corruption in health, education, justice, law enforcement, water, sports, and even unions, and the media. Corruption is not rare even during humanitarian emergencies. According to Transparency International, corruption hampered relief efforts after the 2004 Southeast Asian tsunami. The multifaceted aspects of corruption include its form, amount, and organization. This has important implications for understanding and fighting the phenomenon. Forms of corruption include bribery, nepotism, theft of state assets, and diversion of state revenues. Amounts range from small

Part I: Corruption, Extent, Causes and Consequences      3

payments to low-ranking officials in exchange for services that citizens are entitled to in any case (i.e., “petty” corruption) to large payments to high-ranking officials and ministers in exchange for influencing or changing regulations and other state rules (i.e., “grand” corruption). While the former is the most visible and exasperating to ordinary citizens, the latter is less visible but has potentially higher adverse economic impacts. From the organization angle, corruption can be isolated or collusive, centralized or decentralized, and may or may not involve intermediaries. The above features show that winning the battle against corruption, while possible, is not easy. Mobilizing people to fight corruption is further complicated by the lack of consensus that such practices always have a negative impact. Some writers have suggested that corruption can “grease the wheels” of business, and thus, its economic impact might be positive. Corruption can be beneficial in a second-best world because of the distortions caused by ill-functioning institutions. In 1965, Leys (1965) titled his pioneering paper “What is the Problem about Corruption?” Bardhan (1997), meanwhile, recalls episodes in the history of Europe and the USA when corruption fostered economic development by allowing entrepreneurs to grow. Furthermore, Beck and Maher (1986) and Lien (1986) argue that corruption may raise efficiency. However, other researchers have argued that even if corruption can induce ex post positive effects, the perspective of getting money may cause bureaucrats to create ex ante conditions that support demand for corruption through new regulations, complexity, and unjustified delays. Moreover, the “grease the wheels” effect may become ineffective or too expensive when the administration is made up of a succession of decision centers. In this case, civil servants at each stage may have some form of veto power or ability to slow down the process (Shleifer and Vishny 1993). Bardhan (1997) reports a discussion with an Indian high official who points out that while a civil servant might not be able to move a file faster, he or she could always stop it. Against this background, the objective of Part I of this book is to discuss in depth the issues presented above in order to enable us to tackle efficiently and clearly the analysis in the rest of the book. Chapter 1 discusses the definition, amount, and overage of corruption around the

4     Part I: Corruption, Extent, Causes and Consequences

world. Chapter 2 focuses on the measurement of corruption. Chapter 3 examines the causes, while Chapter 4 deals with the consequences.

References Aidt, T. S. (2003). Economic Analysis of Corruption: A Survey. The Economic Journal, 113(491), F632–F652. Bardhan, P. (1997). Corruption and Development: A Review of Issues. Journal of Economic Literature, 35(3), 1320–1346. Beck, P. J., and Maher, M. W. (1986). A Comparison of Bribery and Bidding in Thin Markets. Economics Letters, 20(1), 1–5. Jain, A. K. (2001). Corruption: A Review. Journal of Economic Surveys, 15(1), 71–121. Leys, C. (1965). What is the Problem About Corruption? The Journal of Modern African Studies, 3(2), 215–230. Lien, D. H. D. (1986). A Note on Competitive Bribery Games. Economics Letters, 22(4), 337–341. Maennig, W. (2005). Corruption in International Sports and Sport Management: Forms, Tendencies, Extent and Countermeasures. European Sport Management Quartely, 5(2), 187–225. Mishra, A. (2006). Persistence of Corruption: Some Theoretical Perspectives. World Development, 34(2), 349–358. Shleifer, A., and Vishny, R. W. (1993). Corruption. Quarterly Journal of Economics, 108(3), 599–617.

1 Definition, Amount, and Coverage

This chapter introduces the issue of corruption and sets the stage for the rest of the book. It starts by discussing the various definitions of corruption in order to highlight the complexity of the phenomenon. It then discusses the amount of corruption at the macroeconomic level and in specific activities. The subsequent section of the chapter presents the extent of the phenomenon in selected sectors. In addition to examining certain key activities (Medical and Water sectors) which have serious implications for the poor, the focus is on activities (Politics, Bureaucracy, Justice, Media, and Education) covered in Parts 2 and 3 of the book, which are concerned with effective means to curb corruption.

1 Definition The literature contains many definitions of corruption. For instance, Rose-Ackerman (2002) sees corruption as “an illegal payment to a public agent to obtain a benefit that may or may not be deserved in the absence of payoffs”. Shleifer and Vishny (1993) refer to “the sale by government officials of government property for personal gain”. © The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_1

5

6     K. Sekkat

For Treisman (2000), it is “the misuse of public office for private gains”. For Transparency International (TI), an NGO which attempts to expose corruption, it is “the abuse of public office for private gain”. For the World Bank (WB), corruption is “the abuse of public office for private gain”. The Organization for Economic Co-operation and Development (OECD) proposes that it is “the promise of giving of any undue payment or other advantages whether directly or through intermediaries to, or for the benefit of, a public official to influence the official to act or refrain from acting in the performance of his or her official duties in order to obtain or retain business”. Finally, the European Bank for Reconstruction and Development (EBRD) states that “corrupt practices mean the bribery of public officials or other persons to gain improper commercial advantage”. However, none of these definitions applies to all forms, types, and degrees of corruption, or covers all acts which can be considered as corruption. In particular, most definitions relate corruption to the behavior of a public official, point to an illegal act, emphasize the payment of bribes, and assume some direct or indirect benefits to one or both parties. These are limitations that make it impossible to capture the whole scope of corruption. To be fair, some organizations have extended their definition to encompass more aspects of corruption. For instance, the United Nations Office on Drugs and Crime emphasizes that corruption can occur in both the public and private domains, while the WB has examined several cases of corruption among private corporations (Hodgson and Jiang 2007). In order to highlight the limitations of the common definitions of corruption, we will examine below some examples of serious cases of corruption which are not covered by the above definitions.

1.1 Public Officials Seeing corruption as limited to public officials may lead the wrong strategy to be adopted. Nobel Prize Gary Becker is quoted in Business Week as declaring: “if we abolish the state, we abolish corruption” (Hodgson and Jiang 2007, p. 1047). A less-extreme recommendation is to reduce

1  Definition, Amount, and Coverage     7

the size of the state in order to limit corruption. However, this is not necessarily an effective solution. For instance, in many former communist countries where the economic power of the state was reduced drastically, corruption did not decrease but increased dramatically after 1991. Moreover, some of the least-corrupt countries, such as Denmark, Finland, the Netherlands, Norway, and Sweden, have large public sectors as measured by the share of public spending in GDP. Finally, corruption is not confined to the public sector. The following provides examples of major corruption in the non-public sector. Scandals concerning large US corporations such as Enron, Xerox, and WorldCom illustrate cases of abuse of private office for private gains. Elias (2005) reports that in 2001 US authorities reported 260 fraud investigations. The majority of these investigations involved accounting practices known as “earnings management”, which paint a financial picture of the company that does not match reality. These practices resulted in the bankruptcies of Enron and WorldCom and were clearly detrimental to all stakeholders. Earnings management practices are committed by the most senior financial executives and, often, external auditing firms. Arthur Andersen, an accounting firm with a long-standing reputation for high ethical standards, was accused of cooking the books for Enron (Swanson and Frederick 2003). Corruption has also been found in sports, including the bribing of players, competitors, and organizers. Maennig (2005) reports several examples, such as the awarding of the 2002 Winter Olympics to Salt Lake City (USA), the gold medal decision in the 2002 Olympics figure skating competition in favor of the Russian skating pair (and against the Canadians), and the verdict in the finals of the Olympic boxing competition in Seoul in 1988. In Germany, a referee fixed soccer matches in the Bundesliga Second Division, for which he received money and goods to a value of at least €50,000 from three Croatian gamblers. In 2012, journalists disguised as London businessmen pretended to be willing to buy votes for London’s 2012 Olympic bid. An International Olympic Committee (IOC) member was caught on a hidden camera agreeing to the deal. These examples illustrate the scale of corruption in the non-public sector.

8     K. Sekkat

Another illustration concerns corruption in unions. According to Thieblot (2006), union corruption spans a broad spectrum, including embezzlement, bribes, and misuse of pension plan schemes. Holders of senior union positions might not only benefit from inflated bonuses and compensation, but also collect bribes in exchange for labor peace or job access, and engage in other corrupt practices. At one extreme, a corrupt union officer might merely turn a blind eye to unsafe working conditions in exchange for bribes. At the other extreme, a union leader might engage in criminal collaboration, as in the well-known case of Jimmy Hoffa in the USA (Coombs and Cebula 2011). However, there is little academic literature on union corruption except in the USA, and, to be fair, apart from some notorious older cases, there is no clear consensus about the pervasiveness of corruption in unions (Horowitz 2004). Freeman and Medoff (1984), examining criminal activity in labor unions between 1969 and 1978, suggest that union corruption is small scale in nature and that the incidence is low. In contrast, Thieblot (2006, p. 531) documents 1238 instances of corruption in American labor unions, distributed among 137 different organizations, between the middle of 1998 and the end of 2005.

1.2 Illegality Narrowing the focus of corruption to illegal activities might bias the estimate of its impact and overlook some forms that are as harmful to society as the illegal ones. First, in many developing countries, gift-exchange is a social norm in business transactions (Bardhan 1997, p. 1130). A bribe is distinct from a gift in the sense that the former implies reciprocity while the latter does not, although the distinction is sometimes difficult to make. Ignoring this results in considering one country to be more corrupt than another purely because of its social norms. Second, the rules which define a corrupt action differ across countries (Banerjee et al. 2012). The same act may be classified as corruption in one setting, but not in another. For example, in the USA and India, the law allows citizens to obtain passports faster by paying a fee. This is not considered as corruption in these countries but would be in

1  Definition, Amount, and Coverage     9

others where no such provision exists in the law. Third and more importantly, the legality criterion disregards practices such as state capture that have potentially more severe economic consequences than “illegal corruption”. State capture occurs when specific public authorities and business interests become so interdependent that economic decision processes are distorted. The most common example is one where a politician has close “connections” to the private sector, and both sides exploit the ties for mutual benefit. The parties exchange favors over time through the allocation of specific legislation or procurement contracts and political campaign funding, making it very difficult to prove that corruption has occurred (Kaufmann and Vicente 2011). Capture goes back to Montesquieu and Marx, who argued that governments may favor special interest groups. Taking a more disaggregated view of government which distinguishes regulatory agencies from political executives and recognizes the multi-principal nature of governments, the authors showed the distorting role that could be played by the various intermediaries needed to implement industrial policies. Campos and Giovannoni (2007) show that capture and corruption are substitutes and that capture is a much more effective instrument for political influence than corruption, even in poorer, less-developed countries. Hellman et al. (2003) present the results of a survey on how firms in transition economies are able to shape the rules of the game to their own advantage, at considerable social cost. Their paper represents the first major attempt to provide sound empirical measures of various forms of grand corruption, such as state capture as well as corruption in public procurement, and administrative corruption (petty forms of corruption). The results indicate that state capture generates considerable performance gains for the firm, while petty corruption is associated with slower firm-level growth rates. This suggests that while the rents generated by state capture are shared by firms and the state, the rents from petty corruption are largely retained by public officials. Hellman and Schankerman (2000) study the relationship between state capture and the EBRD index of economic reform as well as the relationship between state capture and the quality of governance. They find that state capture has a powerfully negative impact on the quality of governance in transition economies. High-capture countries exhibit heavier taxes and

10     K. Sekkat

regulation, greater corruption, poorer macroeconomic management and less-effective law and order. At the same time, state capture is strongly negatively associated with progress on economic reform.

1.3 Payments and Benefits Considering corrupt acts as involving only money or directly benefiting one or both parties also leaves aside important subsets of corrupt acts (Tanzi 1998). The abuse of power can involve no direct payment but only highly indirect benefits to one or both parties. This is the case, for instance, where a wealthy father makes a donation to a university with the “hope” of having his child enrolled in that university in the future. The dean does not put the money in his/her own pocket, and the father does not get any direct benefit from the donation. In the same vein, in many countries, huge amounts of private money finance the activities of political parties, but neither the president of the party nor the donor receives any direct financial gains. For instance, former German Chancellor Helmut Kohl was accused of accepting secret donations for his party. Although admitting to “some mistakes”, Kohl claimed that what he did was in the interest of the party. The investigation was completed in July 2002, and no evidence of Kohl’s personal enrichment was found. There was certainly a breach of the law, and Kohl was ordered to pay a fine of DM 300,000. His party, the CDU, was fined €21 million and suffered electoral losses (Rose-Ackerman 2006). Inside private corporations, one or several subgroups of agents can develop inside the firm and follow objectives that are different from the global strategy of the organization. These groups may bias the information that is communicated to other subgroups, managers, or stockholders. A subgroup’s strategy may, for instance, tend to favor internal supply, even if it is less competitive than the market. This leads to the maintenance of underperforming projects or divisions. This strategy can sometimes be justified by economic considerations: Because of the presence of fixed sunk costs, the activity of a division can be maintained as long as the price on the market is higher than the variable average cost. However, such a strategy may also result from coalitions inside

1  Definition, Amount, and Coverage     11

the company. The head of a division is always reluctant to lose her/his division (or her/his job) even if it performs poorly with respect to the market. He/she will try to find allies among peers to maintain the status quo as long as possible: “I buy from your division; you support my project proposal or job promotion and so forth” (Williamson 1975, p. 120). Finally, there are important cases where individuals act corruptly for moral reasons. This is sometimes called “noble cause corruption”. A good example was when Oskar Schindler bribed Nazi officials to prevent over a thousand Jews from being sent to concentration camps (Hodgson and Jiang 2007).

2 Amount The amount of corruption is very difficult to assess. For a long time, due to a lack of systematic investigations and data collection, the magnitude of corruption had to be assessed using anecdotal or case-study evidence. Since the 1980s, a number of public national and international institutions as well as NGOs have devoted remarkable efforts to helping document the scale of corruption. This section focuses on monetary considerations. Other aspects are discussed in subsequent chapters. Before turning to more rigorous estimates, here are some examples which illustrate the magnitude of the problem. Let us start with some well-publicized cases of corruption. A conservative estimate suggests that the former President of Zaire, Mobutu Sese Seko, looted his country’s treasury of some $5 billion (Svensson 2005), an amount that was equal to the country’s entire external debt at the time he was thrown out of power in 1997. The funds allegedly embezzled by former Indonesian president Mohamed Suharto and the former president of the Philippines Ferdinand Marcos are estimated to be two and seven times higher, respectively, than those misappropriated by Mobutu (Transparency International 2004). Another example comes from Pakistan, where the gold trade was formerly unregulated and smuggling was common. Shortly after Benazir Bhutto returned as prime minister in 1993, a Pakistani gold trader in Dubai proposed a

12     K. Sekkat

deal: In return for the exclusive right to import gold, he would help the government regularize trade and make some further private payments. In 1994, a payment of US$10 million to Ms. Bhutto’s husband was arranged. In November 1994, Pakistan’s Commerce Ministry wrote to the gold trader, informing him that he had been granted a license to be the country’s sole authorized gold importer (Rose-Ackerman 2006). In the early 1990s, the Goldenberg firm received as much as $1 billion from the Kenyan government as part of an export compensation scheme for fictitious exports of commodities. In Angola, a country where three quarters of the population live below the absolute poverty line ($1 a day) and where one in three children dies before the age of five, nearly $1 billion of oil revenues ($77 per capita) disappeared from the state coffers in 2001 alone, or about three times the value of the humanitarian aid received by Angola in 2001 (Svensson 2005). We turn now to less publicized but highly significant situations. In South India, an irrigation engineer may pay bribes of up to 14 times his annual salary in order to obtain two-year tenure at a particular location (Bardhan 1997). This illustrates the size of bribes the engineer is expecting from this position. In Cambodia, health practitioners interviewed for the Global Corruption Report 2006 estimated that, even before the health budget left central government, more than 5% was lost to corruption (Transparency International 2006). Bandiera et al. (2009) focus on Italy, which is often seen as one of the most corrupt countries in Europe. Examining the cost of public procurement, they find that different branches of government pay very different prices for exactly the same product. The difference can exceed 50%. Overall, they estimate that the government could save up to 2% of GDP if most purchase officers paid the same price as the most economical officers. Le Monde of March 17, 1995, reported that the bribes paid abroad by French companies in 1994 amounted to around FF10 billion. According to World Business of March 4, 1996, bribes paid abroad by German companies exceeded US$3 billion a year. Moreover, it seems that around 15% of the total money spent on weapons acquisitions goes toward “commissions” for different parties (Tanzi 1998). Svensson (2005), based on survey of Ugandan firms, reports that 80% of these firms said they needed to pay bribes. The 20% of firms reporting that they

1  Definition, Amount, and Coverage     13

had not paid bribes also minimize their contact with the public sector, meaning that avoiding graft comes at a cost. One field where the issue of corruption is especially critical is public procurement and service delivery programs. Reinikka and Svensson (2004), focusing on a Ugandan public education program that offered each student a grant to cover primary school expenditures, find that over the period 1991–1995, schools received only 13% of central government spending on the program. Most schools received nothing. The bulk of the grants were captured by local government officials and politicians. Olken (2005, 2006), using a similar methodology, finds that in a large anti-poverty program in Indonesia, 29% of funds allocated to a road-building project and 18% of rice subsidies were stolen. Di Tella and Schargrodsky (2003) compare prices paid for basic homogeneous inputs at public hospitals in the city of Buenos Aires. They show that prices paid fell by 15% during the first nine months of a government attack on corruption in 1996–1997. These figures are further supported when confronted with the scale of corruption in public procurement around the world. According to Kaufmann (2005), a conservative approach estimates annual worldwide bribery related to public procurement at about US$1000 billion, or equal to Mexico’s GDP (in 2000, PPP current). These figures include bribes between firms and public officials or politicians in the industrialized world, between multinational corporations from industrial countries and the public sector in emerging economies, and bribery within emerging economies. Finally, the estimate does not take account of the significant losses to a country’s investment, private sector development, and economic growth.

3 Coverage In terms of coverage, we will consider selected fields where corruption is present. It is not possible to present an exhaustive overview of all the fields affected by corruption. In addition to examining certain key activities (Medical and Water sectors) which have serious implications for the poor, we focus on activities (Politics, Bureaucracy, Justice, Media,

14     K. Sekkat

and Education) covered in Parts 2 and 3 and relating to the core concern of this book: Are there effective means to curb corruption?

4 Politics As discussed above, the distinction between bureaucratic corruption and political corruption is important, especially in that they affect developed and developing countries differently. In developing countries, the scope of political competition and checks and balances is very limited, which leaves room for a lot of political corruption. These countries also have a high degree of bureaucratic corruption. In developed countries, democracy is the rule and there is an active political opposition. Politicians face competition almost all the time, while bureaucrats in charge of specialized agencies or administrative departments face much less competition. Paradoxically, in these countries, bureaucratic corruption is very limited while political corruption is widespread (Bardhan 2006). In discussing political corruption, the literature distinguishes, in general, corruption while in office and corruption while running for office (i.e., corruption related to elections). For the moment, we will disregard situations where a politician is running for a first or an additional office. Political corruption while in office is not very different from other types of corruption. It is the result of rent-seeking, the desire to consolidate power, and so on. There are many examples besides the well-publicized cases discussed above (Mobutu Sese Seko, Mohamed Suharto, Ferdinand Marcos, and Benazir Bhutto). For instance, in Peru, Vladimiro Montesinos, President Alberto Fujimori’s spy chief, repeatedly bribed congressmen to switch to Fujimori‘s party to ensure its majority in Congress. In addition, large bribes enabled Montesinos to control most of the media and influence the judiciary (Hunt 2005). Ironically, Fujimori is recognized for bringing down petty corruption. His administration reduced the role of government in the economy not only on efficiency grounds but also on the argument that reducing the role of government would curb opportunities for corruption. In Turkey, the effect of the earthquake which resulted in thousands of deaths in 2004 would have been much less severe, according to the government,

1  Definition, Amount, and Coverage     15

if contractors had not been able to pay bribes to build homes with substandard materials (Kinzer 1999). Lambert-Mogiliansky et al. (2007) illustrate political corruption based on bankruptcy proceedings following the enactment of Russia’s 1998 bankruptcy law. They showed that firms located in a region with a politically powerful governor were significantly less likely to be liquidated under the new law. Political corruption while running for office aims to help a candidate or party to win elections, which are among the most important features of democracy. Parties engaging in elections incur substantial expenses in campaigns as they seek to present their programs and mobilize voters. A part of such expenses is in general covered by the state in accordance with specific rules. However, public funds are rarely sufficient, and parties frequently have recourse to private contributions, the bulk of which comes from business. While regulations exist in many countries for such contributions, numerous infringements taking the form of corruption have been reported. During the high-profile corruption trial involving Elf, an oil company accused of misappropriating hundreds of millions of dollars, former Elf executive Alfred Sirven told a court in Paris that much of the $50 million he withdrew in cash between 1990 and 1996 was used to fund French political parties and foreign leaders (BBC).1 Lambert and Kosenok (2006) cite the testimony of J. C. Mery, a Paris City Hall official, who admitted that for ten years (1985–1994) he organized and arbitrated collusion in the allocation of most construction and maintenance contracts for the City Hall. In exchange, firms paid bribes that were used to finance political parties. In Germany, the CDU party and former Chancellor Helmut Kohl were fined for receiving illegal campaign funding. Among Nordic countries, which consistently rank among the least corrupt in international comparisons, Swedish and Norwegian managers of state-owned companies have been found to be involved in bribe-taking. The issue of campaign financing has spawned a voluminous literature. One interesting question is why business contributes to political

1BBC

News (2003): Elf Funds “Went to French Parties” (http://news.bbc.co.uk/2/hi/ Europe/2926335.stm). Accessed 15 June 2017.

16     K. Sekkat

parties. According to McMenamin (2012), motivations differ across countries and depend, among other things, on differences in political economies and party systems. He groups different motivations into two broad categories: pragmatic and ideological. The pragmatic motivation seeks private gains from the political system while the ideological motivation expresses a preference for government based on a particular set of values and assumptions. A comparison of three countries which are similar in terms of turnover and under transparent and permissive regulation of finance (Australia, Canada, and Germany) shows interesting contrasts. In Canada, until the ban on corporate donations in 2004, money tended to speak pragmatically. A large number of firms expected some benefits in exchange for a donation to a party. In Germany, money tends to speak ideologically. A small number of companies contribute to a given party’s campaign as an expression of a political preference. In Australia, pragmatism dominates, but there is also an ideological preference for the right. This mix of motivations is combined with a high contribution rate. According to McMenamin (2012), these patterns are associated with fundamental differences in political economies and party systems. Pragmatic Canada and Australia are liberal market economies, while Germany is a coordinated market economy. Canada’s two traditional principal parties were almost ideologically indistinguishable, while Australia’s parties compete on a left–right basis.

4.1 Bureaucracy Bureaucratic corruption refers, in general, to a situation where a government official (“the bureaucrat”) exploits his/her authority over certain tasks to get some type of benefit. This kind of corruption is almost unknown in developed countries while it is an everyday problem in developing ones. Economists consider that excessive government interventions and regulations are major determinants of such corruption. Such interventions and regulations give the bureaucrat a non-negligible degree of discretion over the execution of his/her tasks. Typical situations of bureaucratic corruption concern a firm having to pay import duties, clearance of regulatory or licensing requirements, tax on returns,

1  Definition, Amount, and Coverage     17

registration of new firms, compliance with workers’ safety, and so forth. They also include citizens’ applications for government loans, driving licenses, passports, drinking water, and more. The degree of discretion a bureaucrat enjoys over the execution of his/her tasks comes from the complexity of regulations and, sometimes, the strangeness of legislation, which can make it almost impossible for ordinary people to understand and comply with all requirements. The Council of Europe (2009) reports examples of fire inspection regulations, standards, and rules which have been in place in Ukraine since Soviet times. In interviews, entrepreneurs mentioned that: “When visiting, the fire safety inspector demands installation of 20,000 UAH fire alarm systems, then control panel, then he requires connection to the control panel and so on and so forth. After a series of negotiations, we come to the conclusion that in fact, nobody needs this; instead what they need is a certain amount split on a quarterly basis or paid regularly on a monthly basis, which turns out to be far less expensive than following all fire inspection standards”. The entrepreneurs felt that the government officials intentionally used such complicated language in their interactions with business in order to suggest that it was impossible to work without violating the rules (Bose 2010). The problem of bureaucratic corruption is further magnified when citizens engaging in a project must obtain the approval of each one of a set of bureaucrats, often in a prescribed order. Each bureaucrat may demand a bribe as a condition of approval. One high official in New Delhi is reported to have told a friend: “If you want me to move a file faster, I am not sure if I can help you; but if you want me to stop a file I can do it immediately” (Bardhan 1997, p. 1324). This ability to “stop a file” at multiple points may end up deterring profitable projects. Although each of the bribes may be small, the total amount of the bribes may turn out to be excessive. Moreover, officials have, in general, early access to important information on all types of government programs that are valuable for outsiders. Thus, private individuals and firms may pay to obtain such information or to obtain it sooner than their competitors. This is a frequent issue in tenders such as those used in public procurement or privatizations. For instance, privatizations in Argentina allegedly favored

18     K. Sekkat

those with inside information and connections, while those in Thailand supposedly involved kickbacks (Rose-Ackerman 1997). Finally, citizens must often provide various kinds of information and documentation to prove that they deserve to receive certain publicly funded services, such as medical assistance or academic merit scholarships. With a well-designed and efficient bureaucratic process, such requirements can be relatively easily managed by citizens. With an ill-designed and inefficient bureaucratic process, it becomes very difficult to comply with such requirements. In such a case, corruption might be the only way for citizens to get the desired service (Leff 1964 and Huntingdon 1968). This is known as the “grease the wheels” approach, to which we will return in Chapter 4. The literature documents numerous examples of bureaucratic corruption. In Ukraine, respondents to the question “Where did you yourself or members of your family have to give bribes, make ‘charity donations’ or gifts within the past 12 months”, mentioned the State Traffic Police in first position with 6.1%, followed by public agencies to obtain certificates (4.1%) and permits (2.7%). When the respondents were entrepreneurs, the answers were truly surprising: to obtain various permits from public agencies (18.2%), during vehicle registration or inspection processes (18.0%), to obtain certificates (17.7%), at tax inspection agencies (17.2%), and in connection with business activities during audits (15.5%) (Council of Europe 2009). In Spain, the prosecutor general declared during a hearing in the Spanish Congress in November 2009 that his office was investigating almost 750 cases of government corruption, with more than 800 public officials involved. According to the information on judicial proceedings, these cases concerned all political parties with parliamentary representation and were related to land classification and construction permits granted by regional and local government officials (Villoria et al. 2013). According to Rose-Ackerman (1997), in Thailand, 20–40% of infrastructure project funds between 1960 and 1990, although disbursed, did not go to the projects. Other examples of bureaucratic corruption are related to governments’ social activities, such as subsidizing certain goods and services which are sold by the authorities at below-market prices. In general, these subsidies are only for targeting poor people. Documents proving entitlement

1  Definition, Amount, and Coverage     19

are provided by the administration, with some officials supplying undue proof in return for money. Moreover, dual prices often exist and consist of a low state price and a higher free-market price. In some countries, including China, a similar system exists for many raw materials which are sold on two separate markets: one using state-subsidized prices and the other using free-market prices. Here again, corruption is an efficient means to gain unfair access to the subsided price. Similar behavior is reported when the supply of credit and the rate of interest are controlled by the state. Interviews with business people in Eastern Europe and Russia indicate that payoffs are frequently needed to obtain credit (Rose-Ackerman 1997).

4.2 Justice Enforcement of the rule of law is essential in a civilized society. This task is assigned to the judiciary system. As pointed out by Hay et al. (1996), the rule of law means that people use the legal system to structure their activities (including economic activities) and resolve disputes. This entails learning what the legal rules are and structuring activities according to these rules. It also includes obtaining redress from those who break the rules by appealing to qualified institutions, such as the courts or the police. In the economic field, laws are expected to define and protect property rights, to set rules for trading these rights, to define rules for entering and exiting the market, and to promote fair competition and behavior. In this way, the justice system contributes to protecting the freedoms, peace, and equality that foster social and economic development (Abdulkarim 2012). A well-functioning and effective judiciary system is crucial to the fight against corruption (Schultz 2009). Under such a system, the investigation, exposure, and punishment of corruption cannot be blocked by powerful politicians or businesses. Similarly, assets accumulated through corruption by leaders of developing countries can be seized and returned to the country of origin, which is, in general, in crucial need of the funds. Without a well-functioning judiciary, it is much harder to trace, freeze, seize, and confiscate the assets in question, at the expense of the country’s development. Accordingly, an ill-functioning judiciary

20     K. Sekkat

system and especially a corrupt system not only prevent the abovementioned benefits from materializing but also hamper efforts to reduce corruption in other fields. For instance, a cross-country study (Herzfeld and Weiss 2003) showed that the relative attractiveness of corruption to bureaucrats depends on the effectiveness of the legal system and, more specifically, on the probability of being detected and punished. The drivers of an ill-functioning judiciary are interference and pressure from executive or powerful economic interests, lack of citizen voice, and insufficient financial and human resources (Schultz 2009). The resulting corruption in courts is perceived as a major problem worldwide. The 2009 Corruption Barometer published by TI showed that nearly half of the respondents across the world considered the judiciary to be corrupt (Transparency International 2007). The 2006 Barometer, based on a poll of 59,661 people in 62 countries, showed that in onethird of these countries, more than 10% of respondents who had interacted with the judicial system reported paying bribes to obtain a “fair” outcome in a court case. In Nigeria, surveys show that corruption facilitates the destruction of evidence and speedier hearings (Schultz 2009). In Jordan, the dominant concern is that judges’ rulings may be influenced by family or tribal affiliations. In 2005, a household survey by TI in Bangladesh found that two-thirds of respondents who had used the courts in the preceding year paid an average bribe of around US$108 per case, or 25% of their annual income. The survey also revealed that 39% of those who paid bribes to the judiciary said that they had paid through lawyers, who transmitted a portion to magistrates or judges. Public prosecutors reportedly extracted bribes from 4% of respondents (Transparency International 2007). In India, while the upper judiciary is relatively clean, the Center for Media Studies, based on a countrywide survey conducted in 2005, found that bribes seemed to be frequently solicited in the lower judiciary in order to get things done, with an estimated US$580 million paid in bribes in a 12-month period. In Indonesia, major enforcement operations against illegal timber smuggling have resulted in few convictions, and those who were imprisoned were released after a short time. Advocates argue that corruption of the police, judges, and prison officials undermines good-faith efforts to rein in the country’s timber tycoons (Schultz 2009).

1  Definition, Amount, and Coverage     21

4.3 Media One important role of free media is to convey to citizens information about rules, regulations, and services they are entitled to receive. From this perspective, the media can monitor and reinforce the transparency and accountability of private and public institutions. Many corruption issues have been brought to the public’s attention first through the media. In other words, the press serves as a watchdog and a partner to society in ensuring transparency and accountability and fighting crime in general and corruption in particular. However, the media itself is not exempt from corruption and potentially represents a useful asset for bribers. For instance, recorded bribe transactions involving Peru’s former secret-police chief, Vladimiro Montesinos, show that he paid television-channel owners 100 times in bribes what he paid to judges and politicians (Svensson 2005). This suggests that Montesinos saw news media as the strongest check on the government’s power. Media corruption has different nicknames around the world. The terms “brown envelope” or “red envelope” are used in some African and Asian countries, but more recently the term “ATM journalism” has become popular. The term refers to the switch to transferring bribes into journalists’ bank accounts electronically (Ristow 2010). Other terms such as “extortion journalism” or “sitting allowance” are also used (Adeyemi 2013). Generally, the brown envelope is a monetary bribe handed out to a person to put pressure on him or her to do what the bribe-giver wants. It is any form of gratification which a journalist may receive to cover an event or influence his/her judgment. The event could be a press conference, an interview of any sort, a workshop, or an impromptu or organized briefing. Many examples highlight the scale of corruption in the media. Some of these examples concern smaller payments, sometimes called “facilitation payments”, which are common in Africa, China, and elsewhere. They consist of US$5, US$10, or US$20 inside a brown or red envelope. Although supposedly provided to cover travel costs to some sort of press event, in fact they are distributed to all journalists, whether they have traveled or not. A Chinese journalist reports that when this happened for the first time (the envelope contained about $44),

22     K. Sekkat

she asked for the reason and was told that this was the “normal procedure” (Adeyemi 2013). Ristow (2010) reports different situations of this kind. In Ghana, a reporter goes to a press conference and inside her press packet finds a brown envelope containing the equivalent of US$20. Not surprised, she slips it into her purse before heading back to the office to write up the event. In Russia, a public relations agency sends out a false press release about a fictitious company. Thirteen publications believe the lure and agree to run the release just like a story but only against payment ranging from about $125 to nearly $2000. In South Africa, a journalist admitted that he and several others had set up a media relations firm which served as an “air force” to battle a politician’s rivals. The journalist added that he received around $700–$1400 per paper. However, he could not write negative reports about the politician or his allies. Besides such direct bribes, other common ways of influencing the media are ownership or advertising expenses. Such expenses target favorable coverage by newspapers, radio, or television of the briber’s activity. Such information bias is, however, different from that which emerges when, for instance, a newspaper publicly acknowledges being affiliated with a given party or sharing a similar ideology. This was the case in the USA in the second half of the nineteenth century, when most American newspapers were partisan (Petrova 2011), and in present-day France (e.g., l’Humanité is openly supporting the French Communist Party). When a given media source professes objectivity but delivers biased information, suspicions of corruption or at least capture became very serious. A well-known example of the use of ownership is Italy (Durante and Knight 2012), where a single politician with an identified ideology, Silvio Berlusconi, owns the main private television network, and where the public television corporation is traditionally controlled by the ruling political coalition. Durante and Knight (2012) investigated news content and viewership of the six top national television channels before and after the change in government in 2001, which shifted the government from a center-left coalition to Berlusconi’s center-right coalition. The findings show that Berlusconi’s private network provided more speaking time to the right during the period in which the right was in

1  Definition, Amount, and Coverage     23

power than to the left during the period in which the left was in power. Moreover, the public network shifted to the right, relative to the private network, following the change in control of the public network from the left to the right. In a similar vein, Djankov et al. (2003) examine whether there is a relationship between ownership patterns of different media (newspapers, television, and radio) and potential information bias. In a sample of 97 countries, they identify two dominant forms of media ownership: ownership by the state and ownership by concentrated private owners, namely controlling families. They find that government ownership of the media is greater in countries that are poorer and have greater overall state ownership in the economy, lower levels of school enrolment, and more autocratic regimes. More importantly, their analysis casts doubt on the notion that state ownership of the media serves benevolent ends. Moreover, greater state ownership of the media (especially the press) is associated with poor performance in terms of freedom of the press, political and economic freedom, and health status. Turning to advertising expenses, the problem is particularly worrying when a large share of advertising revenues comes from a limited number of institutions. Petrova (2011) uses data on American newspapers in 1880–1885 to test whether there is a positive relationship between the growth of the advertising market and the growth of independent media. During that period, the majority of newspapers were affiliated with a major political party, but some were independent. The main difference between partisan newspapers (newspapers affiliated with the Democratic or Republican Party) and independent newspapers was the extent of control a party had over the paper’s news coverage. Editors of independent newspapers could print whatever they wanted, while editors of partisan newspapers were restricted in their decisions. The results showed that with higher sources of advertising revenues, newspapers were more likely to be independent from political parties. Di Tella and Franceschelli (2011) examine the extent to which the four main newspapers in Argentina report government corruption on their front page during the period 1998–2007 and the correlation between the publication of such information and the amount of government advertising each newspaper received. The correlation is negative and significant, implying that a one standard deviation increase in

24     K. Sekkat

monthly government advertising (0.26 million 2000 pesos) is associated with a reduction in coverage of government corruption scandals of 25% of a standard deviation in the measure of monthly front-page coverage. Reuter and Zitzewitz (2006) test advertising bias within the financial media in the USA. Specifically, the paper examines whether a given journal’s recommendations about a mutual fund is related to the amount of advertising the journal has received from that fund. The analysis showed a positive correlation between a fund’s lagged advertising expenditures and the probability that this fund will be recommended in each of the finance publications in the sample (Money Magazine, Kiplinger’s Personal Finance, and Smart Money  ). Further investigations lead to the conclusion that the relationship is causal, in other words that finance publications bias their recommendations (either consciously or subconsciously) to favor advertisers. In contrast, the authors found no such relationship between advertising and content in other national newspapers (The New York Times and the Wall Street Journal ).

4.4 Education Education has positive impacts on many facets of society. Through its contribution to human capital building, it is one of the most important drivers of growth and development. Education also puts citizens in a better position to voice opinions and monitor politicians. This is another channel through which education can foster the development of democracy and trust in the country’s institutions. Moreover, educational institutions play an important role in professional certification, which should ensure good health and safety by allowing only competent doctors, teachers, and other professionals to graduate and perform their crucial tasks. However, corruption in education undermines such expected positive impacts. Rumyantseva (2005) reports that in the Commonwealth of Independent States (CIS), corruption in higher education completely annihilated employers’ and the general public’s trust in many of the country’s colleges and universities. Many job advertisements in Russia and Ukraine explicitly require graduates from certain universities.

1  Definition, Amount, and Coverage     25

Many forms of corruption in education are also present in other sectors (e.g., procurement, hiring, and misuse of public funds). The scale of and the means for fighting these forms of corruption are not so different in education from in other sectors (Rumyantseva 2005). However, other forms of corruption are specific to education and deserve particular attention because of the different and additional damage they create for society. Beside the costs of resource misallocation and extortion of citizens’ funds, corruption in education may impact the quality of human capital, which is one of the most precious assets in a society. As a consequence, in this section we will focus on education-specific corruption. Education-specific corruption involves students, teachers, researchers, and administrators. It has direct effects on students’ values, beliefs, and life chances. It includes, among other things, payment of bribes to get higher grades, purchase of diplomas, and unfair admission to certain universities. Besides the direct approach, by which a student offers money to a professor in order to get a good grade even though he or she does not know the subject, there are other ways to pay for grades. For example, a faculty member might sell a term paper to a student, or a professor might give a low grade to a student who knows the subject and recommend private tutoring, so giving an implicit guarantee that the student will pass regardless of how much he/she has learned. Finally, corruption to secure unfair admission can be direct or indirect. For instance, a prospective student’s parents could make a donation to the university in order to guarantee their child’s future admission. All these forms of corruption are encountered in developed and developing countries. Johnson (2008) provides many examples from around the world. In China, children are supposed to attend schools closest to their homes. However, parents can put their children in other schools that are further away by giving “gifts” to these schools. Such donations for school admissions are also common in India. In Russia, the cost of a bribe to get into a top Moscow university can be more than the cost of tuition for all five years of studies (Bennet 2001). In fact, out of 900 million rubles spent on education a year, half goes in bribes. In Kazakhstan’s universities, future doctors are graded according to the amount of money they give professors or their family networks. Turning

26     K. Sekkat

to more-developed countries, the most common corrupt practice in the USA consists in wealthy Americans buying their children’s admission to the most prestigious colleges and universities by making donations to those institutions. Such donations are sometimes very generous and include endowed scholarships, study centers, and buildings. According to Johnson (2008), in Great Britain, officials at Oxford‘s Pembroke College supposedly said that admission might be possible in exchange for a donation of approximately £300,000. Fake diplomas are another field of corruption in higher education that deserves special attention. This phenomenon is increasingly attracting the attention of the international community, including the media, governments, and universities. While there is no consensus on how to characterize a fake degree, we follow Grolleau et al. (2008) in considering that fake diplomas include both copies from authentic institutions and degrees that can either be bought or earned with little work from a somewhat legal institution that sells “degrees”. To illustrate the seriousness of the issue, imagine that the surgeon working on your body has no accredited diploma, that the house you just bought was constructed under the supervision of an architect with a fake diploma, or that lawyer representing you before a court has no law credentials. Although the history of fake diplomas can be traced back to the time of the first academic degrees, the extent of the problem is becoming so large that firms are now hiring specialized private services to verify college degrees and other educational qualifications. In January 2005, two experts in the area argued that the fake degree business is a billion-dollar industry that has sold more than a million fake college diplomas (Grolleau et al. 2008). The recent growth is impressive, with Cohen and Winch (2011) reporting an increase in 839 (48%) between 2010 and 2011 in the total number of bodies delivering fake diplomas around the world. While the increase concerns all regions, North America and Europe continue to host the largest number of such bodies, with 1095 in North America (up from 892, an increase in 23%) and 603 in Europe (up 31% from the previous year’s total). Besides the industry described above, fake diplomas might also come from well-established and accredited institutions. According to Transparency International (2013), in Niger selling diplomas is the

1  Definition, Amount, and Coverage     27

easiest source of income for some public officials. Today, the price of a BEPC (junior high school diploma) is roughly equivalent to an average teacher’s monthly salary (US$175), while a BAC (final high school diploma) can be bought for twice that. In 2006, 20 teachers were arrested in a single investigation for cheating during the baccalauréat exams. Ten were later found guilty of corruption and dismissed from their posts. Transparency International (2013) also reports the case of a former professor of the Institute of the History of Medicine at the University of Würzburg (Germany) who was suspected of having supervised and supported dozens of inadequate doctoral theses prior to his retirement in 2005. In general, these theses numbered around 35 pages and contained only meager meaningful research achievements. The professor was additionally suspected of accepting donations from doctoral students for his nonprofit societies. He had already paid a moderate fine for accepting €6000 (US$7390) from a consultant who connected him with physicians seeking doctoral degrees.

4.5 Medical Like education, medical services play an important role in society. Good health is a crucial determinant not only of individual well-being but also of economic and social development. Moreover, spending on health care represents a non-negligible share of national spending, amounting to approximately US$3 trillion and ranging from 5% of GDP in low-income countries to more than 15% in OECD countries (Holmberg and Rothstein 2011). The substantial resources spent in health sectors offer ample opportunities for abuse and illicit gain. As a result, corruption also undermines service delivery in the health sector (Hussmann 2011b). In the USA, fraud and abuse in health care are estimated to cost US$11.9 billion–US$23.2 billion per year. In Colombia, overpayments for seven specific medications in 32 public hospitals have been estimated at more than US$2 million per year. This amount would have paid for health insurance coverage for 24,000 people. The 2006 report by Transparency International shows that corruption and similar practices affect many different areas of the health sector, such as hospital

28     K. Sekkat

administration, under-the-table payments to doctors, counterfeit drugs, and over-billing of insurance companies. In rural Tanzania, the majority of deaths caused by malaria can be attributed to corrupt practices in the form of drug stealing, provider absenteeism, stolen equipment, and very low levels of diagnostic efforts (Holmberg and Rothstein 2011). As with education, corruption in health services also takes forms similar to those in other sectors. However, a number of specific features make the health sector particularly vulnerable to corruption (Hussmann 2011a). Inherent uncertainty about outcomes, asymmetry of information between patients and physicians, and large numbers of participants are examples of specific features that hinder transparency and accountability and create opportunities for corruption. The importance of these specific features is reflected in two parts of the health care process: at the point of health service delivery and in the pharmaceutical and equipment supply chain (Transparency International 2006). In service delivery, these specific features hamper the ability of patients to judge the decisions made on their behalf or assess the correctness of a bill. Similarly, insurance auditors have a hard time assessing whether billing is correct and whether the services provided were necessary. This can result in parties extorting or accepting under-the-table payments for services that are supposed to be provided free of charge and soliciting payments in exchange for special privileges or treatment (Transparency International 2006). Examples of corruption in health services exist all over the world. In Morocco, after being examined in a public hospital, a woman got a prescription for pills that she could not afford. Although she was entitled to receive the pills for free from the public hospital, a hospital employee told her that the pills were not available but that, for 20 or 30 dirhams, it would be possible to find “free medication” somewhere in the hospital (Transparency International 2006). In India, immediately after giving birth, a woman asked to hold her baby on her chest. However, a nurse took the infant away, and an attendant told the mother that the customary price to hold her own child directly after giving birth was US$12 for a boy and US$7 for a girl (Holmberg and Rothstein 2011). US$12 was a substantial amount of money for the family since the husband was working for less than US$1 a day. In Cambodia, health

1  Definition, Amount, and Coverage     29

workers have reported that it costs up to US$100,000 to get the position of director at the provincial or national offices of the Health Ministry. A low-level public servant job in the health sector may go for US$3000. These sums represent a large investment since government employee salaries generally average US$40 per month (Transparency International 2006). At the level of the pharmaceutical and equipment supply chain, products can be diverted or stolen at various points in the distribution system, officials may demand “fees” for approving products or for setting prices, doctors may adapt prescriptions to favor certain drug manufacturers, and experts can ask for or accept bribes to influence hiring decisions and decisions on licensing, accreditation, or certification of facilities. Even more worrying is the fact that counterfeit or other forms of substandard medicines may be allowed to freely circulate in the country. According to Hussmann (2011a), in the USA, up to 15% of all drugs sold are fake, and in some countries, this figure can rise to 50%. In Chad, the country’s regions received only a third of the centrally allocated health resources, while in Cambodia, 5–10% of the health budget is already lost at the central level. In Tanzania, local or district councils divert up to 41% of centrally disbursed funds (Transparency International 2006). We deliberately end this section by illustrating the point that corruption in the health sector can be a matter of life and death, especially for poor people in developing countries. Estimates suggest that in 2001 alone, 192,000 people died in China because of fake drugs (Transparency International 2006). An IMF study across 71 countries showed that countries with higher corruption also have higher infant mortality rates, even after adjusting for income, female education, health spending, and urbanization (Gupta et al. 2014).

4.6 Water Like health, water is an essential factor in life and socioeconomic well-being. In developing countries, about 80% of health problems can be linked to inadequate water and sanitation. Nearly 1.8 million

30     K. Sekkat

children die every year, and around 443 million school days are lost by children who suffer from water-related sicknesses. In some African regions, people need to walk more than ten kilometers to gather water in the dry season. It is also estimated that around 5% of GDP in Africa is lost due to illness and death caused by dirty water and poor sanitation. In El Salvador, Jamaica, and Nicaragua, the poorest households spend more than 10% of their income on water, while those in rich nations pay only a third of this share. These consequences of water availability are exacerbated by corruption and chiefly affect the poor and women (Transparency International 2008). Corruption in the water sector can be encountered at different levels of the supply chain, including connection to a water network, customer billing, and construction. It involves petty and grand corruption and affects water availability as well as allocation between different users and uses (safe drinking water, irrigation processes, water for energy use). All of these can constitute serious impediments to food security, poverty reduction, equity, and environmental protection. The poor are the main victims of water corruption. In some developing countries, corruption is estimated to raise the price for connecting a household to a water network by as much as 30%. A utility connection in Manila is equal to about three months of income for the poorest 20%, while the equivalent figure is six months in Kenya and more than a year in Uganda. Globally, two-thirds of the 1.2 billion people who do not have access to safe drinking water live with less than US$2 a day, and of the more than 2.6 billion people who lack basic sanitation, a half fall below that same poverty line. But water problems are not just an important cause of economic poverty; they are also a consequence of it. There is a bidirectional causal relationship between poverty and the lack of water. A company in charge of water distribution might be reluctant, especially if the company is private, to expand its network to low-income areas instead of rich ones. The reason is that, for a similar cost, the company will have to deal with poor customers who struggle to pay bills, rather than wealthy families. Besides the issue of poverty, corruption in the water sector has dramatic effects on the environment. A very limited number of companies properly treat sewage before discharging water into rivers. The vast

1  Definition, Amount, and Coverage     31

majority of them ignore guidelines to install cleaner technologies on the grounds that it would cost too much. Corruption is a significant facilitator in these cases. As a result, in China, for example, aquifers in 90% of cities are polluted, more than 75% of river water flowing through urban areas is unsuitable for drinking or fishing, and 30% of river water throughout the country is regarded as unsuitable for agricultural or industrial use. Moreover, about 700 million people drink water contaminated with animal and human waste. Water pollution has caused illness in 190 million Chinese people and is responsible for an estimated 60,000 premature deaths every year. Environmental degradation and pollution are believed to reduce China’s GDP by 8–12% annually. While developed countries are less affected by petty corruption in the water sector, they have to deal with major issues in terms of grand corruption. In both groups of countries, the award of contracts for water facilities is a fertile field for corruption (Transparency International 2008). Corruption can take the form of cash payments or other types of benefits, ranging from gourmet dinners and sumptuous holiday trips to luxury apartments, all for the purpose of securing the award of water and sanitation contracts or influencing their design. In Milan, an executive of a private water company was imprisoned in 2001 for planning to bribe local politicians with around US$2.9 million to win a US$145 million wastewater treatment contract. The city council president was also convicted and jailed. In Chicago, the head of the water department was found guilty of extorting campaign contributions from subcontractors. Under Japan’s “dango” system, bidders for public works projects decide among themselves who will win contracts. However, all members of the group submit arranged bids to public agencies to maintain the illusion of competition. In Lesotho, the chief executive of the Lesotho Highlands Development Agency was found guilty of accepting more than US$6 million in bribes from multinational companies to secure tenders for a project worth US$8 billion. In 2002, he was sentenced to 18 years in prison. Multinationals from the UK, France, Germany, Italy, Canada, and other countries have also been prosecuted for seeking to influence the tendering procedure. Finally, a survey in South Asia suggests that contractors frequently pay between 1 and 6% of contract values in bribes to win contracts. During construction, additional bribes

32     K. Sekkat

can increase the costs to companies by up to another 11% of the contract value. In addition to securing the award of contracts, corruption also helps to conceal low-quality work and non-delivery of goods.

5 Conclusion Corruption is a very complex phenomenon. It is old, which means that it is persistent and very hard to eradicate. Corruption is also highly widespread, affecting almost every country and every human activity. This implies that fighting it may require the cooperation of a large number of bodies and institutions. Finally, corruption is multifaceted. It takes different forms, including bribery, nepotism, theft of state assets, and diversion of state revenues. The amounts involved range from small payments to low-ranking officials in exchange for services that citizens are entitled to in any case (i.e., “petty” corruption) to large payments to high-ranking officials and ministers in exchange for influence over or changes to regulations and other state rules (i.e., “grand” corruption). However, it is misleading to associate corruption purely with the public sector, as this category of participants does not have a monopoly on corruption. Although public sector participants represent a large share of the persons and entities involved, the literature provides a wealth of examples of corrupt acts outside the public sector featuring private firms, sports, trade unions, the media, and NGOs. Furthermore, corruption is not always an illegal activity. Depending on history, customs, and values, the same practice may be considered as illegal in one country, but not in another. The economic impact of corruption is very difficult to assess, but some anecdotes and estimates are illuminating. In South India, an irrigation engineer may pay bribes of up to 14 times his annual salary in order to obtain two-year tenure at a particular location. This is because of the size of the bribes the engineer expects from this position. In Cambodia, health practitioners estimate that even before the health budget leaves central government, more than 5% is lost to corruption. At the global level, a conservative estimate puts bribes between firms and public officials or politicians at about US$1000 billion annually,

1  Definition, Amount, and Coverage     33

which is equivalent to Mexico’s GDP (in 2000, PPP current). These figures do not take account of the significant losses to a country’s investment, private sector development, and economic growth.

References Abdulkarim, Y. (2012). Socio–Economic Effects of Judicial Corruption in Nigeria. International Journal of Humanities and Social Science Invention, 1(1), 31–36. Adeyemi, A. (2013). Nigerian Media and Corrupt Practices: The Need for Paradigm Shift. European Scientific Journal, 9(1), 119–136. Bandiera, O., Prat, A., & Valletti, T. (2009). Active and Passive Waste in Government Spending: Evidence from a Policy Experiment. American Economic Review, 99(4), 1278–1308. Banerjee, A., Mullainathan, S., & Hanna, R. (2012). Corruption (National Bureau of Economic Research, WP 17968). Bardhan, P. (1997). Corruption and Development: A Review of Issues. Journal of Economic Literature, 35(3), 1320–1346. Bardhan, P. (2006). The Economist’s Approach to the Problem of Corruption. World Development, 34(2), 341–348. BBC News. (2003). Elf Funds ‘Went to French Parties. http://News.Bbc. Co.Uk/2/Hi/Europe/2926335.Stm. Accessed 15 June 2017. Bennet, N. (2001). Corruption in Education Systems in Developing Countries: What is it Doing to the Young? In 10th International AntiCorruption Conference, Prague, Czech Republic. Bose, G. (2010). Aspects of Bureaucratic Corruption. Australian School of Business Research Paper ECON 14. Campos, N. F., & Giovannoni, F. (2007). Lobbying, Corruption and Political Influence. Public Choice, 131(1), 1–21. Cohen, E. B., & Winch, R. (2011). Diploma and Accreditation Mills: New Trends in Credential Abuse. Bedford: Verifile Accredibase. Coombs, C. K., & Cebula, R. (2011). The Impact of Union Corruption on Union Membership. Industrial Relations: A Journal of Economy and Society, 50(1), 131–148. Council of Europe. (2009). Corruption Risks in the Fields of Administrative Services and Control-Supervision Activities of Public Administration

34     K. Sekkat

in Ukraine. http://Pravo.Org.Ua/Files/Book_UPAC_Eng_FINAL.Pdf. Accessed 15 June 2017. Di Tella, R., & Franceschelli, I. (2011). Government Advertising and Media Coverage of Corruption Scandals. American Economic Journal: Applied Economics, 3(4), 119–151. Di Tella, R., & Schargrodsky, E. (2003). The Role of Wages and Auditing During a Crackdown on Corruption in the City of Buenos Aires. Journal of Law and Economics, 46(1), 269–292. Djankov, S., Mcliesh, C., Nenova, T., & Shleifer, A. (2003). Who Owns the Media? The Journal of Law and Economics, 46(2), 341–382. Durante, R., & Knight, B. (2012). Partisan Control, Media Bias and Viewer Responses: Evidence from Berlusconi’s Italy. Journal of the European Economic Association, 10(3), 451–481. Elias, R. Z. (2005). The Effect of Corporate Ethical Values on Accountants’ Perceptions of Social Responsibility. Journal of Applied Business Research (JABR), 21(4), 1–10. Freeman, R., & Medoff, J. L. (1984). What Do Unions Do? New York: Basic Books. Grolleau, G., Lakhal, T., & Mzoughi, N. (2008). An Introduction to the Economics of Fake Degrees. Journal of Economic Issues, 42(3), 673–693. Gupta, S., Davoodi, H., & Tiongson, E. (2014). Corruption and the Provision of Health Care and Education Services. Annals of Economics and Finance, 15(2), 759–798. Hay, J. R., Shleifer, A., & Vishny, R. W. (1996). Toward a Theory of Legal Reform. European Economic Review, 40(3–5), 559–567. Hellman, J., & Schankerman, M. (2000). Intervention, Corruption and Capture: The Nexus Between Enterprises and the State. Economics of Transition, 8(3), 545–576. Hellman, J. S., Jones, G., & Kaufmann, D. (2003). Seize the State, Seize the Day: State Capture and Influence in Transition Economies. Journal of Comparative Economics, 31(4), 751–773. Herzfeld, T., & Weiss, C. (2003). Corruption and Legal (In) Effectiveness: An Empirical Investigation. European Journal of Political Economy, 19(3), 621–632. Hogdson, G. M., & Jiang, S. (2007). The Economics of Corruption and the Corruption of Economics: An Institutionalist Perspective. Journal of Economic Issues, 41(4), 1043–1061.

1  Definition, Amount, and Coverage     35

Holmberg, S., & Rothstein, B. (2011). Dying of Corruption. Health Economics, Policy and Law, 6(04), 529–547. Horowitz, C. F. (2004). Union Corruption in America: Still a Growth Industry. National Institute for Labor Relations Research. Hunt, J. (2005). Why Are Some Public Officials More Corrupt Than Others? (CEPR Discussion Papers, No. 5252). Huntington, S. P. (1968). Political Order in Changing Societies. New Haven: Yale University Press. Hussmann, K. (2011a). Vulnerabilities to Corruption in the Health Sector: Perspectives from Latin American Sub-Systems for the Poor (With a Special Focus on the Sub-National Level), United Nations Development Programme—UNDP Regional Centre Panama. Hussmann, K. (2011b). Addressing Corruption in the Health Sector. Securing Equitable Access to Healthcare for Everyone. U4, (U4). Johnson, V. R. (2008). Corruption in Education: A Global Legal Challenge. Santa Clara Law Review, 48(1), 1–77. Kaufmann D. (2005). “Six Questions on the Cost of Corruption with World Bank Institute Global Governance Director Daniel Kaufmann”, Governance and Anti-Corruption Website. http://go.worldbank.org/ KQH743GKF1. Accessed 5 May 2018. Kaufmann, D., & Vicente, P. C. (2011). Legal Corruption. Economics and Politics, 23(2), 195–219. Kinzer, S. (1999). The Turkish Quake’s Secret Accomplice: Corruption. New York Times, 29, 3. Lambert, A., & Kosenok, G. (2006). Public Markets Tailored for the Cartel. PSE, Paris-Jourdan Sciences Economiques: Mimeo. Lambert-Mogiliansky, A., Sonin, K., & Zhuravskaya, E. (2007). Are Russian Commercial Courts Biased? Evidence from a Bankruptcy Law Transplant. Journal of Comparative Economics, 35(2), 254–277. Leff, N. H. (1964). Economic Development through Bureaucratic Corruption. American Behavioral Scientist, 8(3), 8–14. Maennig, W. (2005). Corruption in International Sports and Sport Management: Forms, Tendencies, Extent and Countermeasures. European Sport Management Quartely, 5(2), 187–225. Mcmenamin, I. (2012). If Money Talks, What Does it Say? Varieties of Capitalism and Business Financing of Parties. World Politics, 64(01), 1–38. Olken, B. A. (2005). Monitoring Corruption: Evidence from a Field Experiment in Indonesia. National Bureau of Economic Research, WP11753.

36     K. Sekkat

Olken, B. A. (2006). Corruption and the Costs of Redistribution: Micro Evidence from Indonesia. Journal of Public Economics, 90(4), 853–870. Petrova, M. (2011). Newspapers and Parties: How Advertising Revenues Created an Independent Press. American Political Science Review, 105(4), 790–808. Reinikka, R., & Svensson, J. (2004). Local Capture: Evidence from a Central Government Transfer Program in Uganda. Quarterly Journal of Economics, 119(2), 679–705. Reuter, J., & Zitzewitz, E. (2006). Do Ads Influence Editors? Advertising and Bias in the Financial Media. Quarterly Journal of Economics, 121(1), 197–227. Ristow, B. (2010). Cash for Coverage: Bribery of Journalists Around the World. Report of the Center for International Media Assistance official Website: https://www.cima.ned.org/wp-content/uploads/2015/02/CIMABribery_of_Journalists-Report.pdf. Accessed 9 May 2018. Rose-Ackerman, S. (1997). The Political Economy of Corruption. In K. A. Elliott (Ed.), Corruption and the Global Economy (pp. 31–60). Washington, DC: Peterson Institute. Rose-Ackerman, S. (2002). When Is Corruption Harmful? In M. Johnston (Ed.), Political Corruption: Concepts and Contexts (pp. 353–371). New Brunswick and London: Transaction Publishers. Rose-Ackerman, S. (Ed.). (2006). International Handbook on the Economics of Corruption. Cheltenham: Edward Elgar Publishing. Rumyantseva, N. L. (2005). Taxonomy of Corruption in Higher Education. Peabody Journal of Education, 80(1), 81–92. Schultz, J. (2009). The UNCAC and Judicial Corruption: Requirements and Avenues for Reform. U4 Brief (18). Shleifer, A., & Vishny, R. W. (1993). Corruption. Quarterly Journal of Economics, 108(3), 599–617. Svensson, J. (2005). Eight Questions About Corruption. Journal of Economic Perspectives, 19(3), 19–42. Swanson, D. L., & Frederick, W. C. (2003). Are Business Schools Silent Partners in Corporate Crime. Journal of Corporate Citizenship, 9(1), 24–27. Tanzi, V. (1998). Corruption Around the World: Causes, Consequences, Scope and Cures. IMF Staff Papers, 45(4), 559–594. Thieblot, A. J. (2006). Perspectives on Union Corruption: Lessons from the Databases. Journal of Labor Research, 27(4), 513–536. Transparency International. (2004). Global Corruption Report 2004.

1  Definition, Amount, and Coverage     37

Transparency International. (2006). Global Corruption Report 2006. Transparency International. (2007). Global Corruption Report 2007. Transparency International. (2008). Global Corruption Report 2008. Transparency International. (2013). Global Corruption Report 2013. Treisman, D. (2000). The Causes of Corruption: A Cross-National Study. Journal of Public Economics, 76(3), 399–457. Villoria, M., Van Ryzin, G. G., & Lavena, C. F. (2013). Social and Political Consequences of Administrative Corruption: A Study of Public Perceptions in Spain. Public Administration Review, 73(1), 85–94. Williamson, O. E. (1975). Markets and Hierarchies. New York: Free Press.

2 Measurement Issues

Talking about corruption, its prevalence, causes, consequences and differences across countries implies the ability to measure the phenomenon. For a long time, due to a lack of systematic investigations and data collection, the magnitude of corruption had to be assessed using anecdotal or case-study evidence. Since the 1980s, a number of national and international institutions as well as NGOs have devoted remarkable efforts to helping document the scale of corruption. Broadly speaking, there are three groups of approaches to assessing corruption. One group is based on the perception of different experts about the prevalence of corruption in a country or a sector. The second group relies on actual acts of corruption and draws either on justice records (or other potential sources) regarding corruption convictions or on individuals’ experiences with corruption. The last group uses Public Expenditure Tracking Surveys and Quantitative Service Delivery Surveys to reveal potential corruption. The following analysis compares the different approaches to measuring corruption as well as their results. It also tackles an important issue for our purpose (i.e., fighting corruption), namely the degree of persistence of corruption. The aim is to get an idea of the

© The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_2

39

40     K. Sekkat

approximate time needed before citizens notice a potential improvement on the corruption front.

1 Measurement Issues There is a rich body of literature dealing with corruption, its prevalence, causes, consequences, and differences across countries. The preliminary step to any of these analyses is to build an adequate measure of the phenomenon. Such measures are important for a variety of participants, including international investors, international and national institutions, official development, and aid agencies and researchers. Numerous corruption indicators are now provided by various organizations such as Transparency International (Corruption Perceptions Index, CPI, and Bribe Payers Index, BPI), the European Bank for Reconstruction and Development (EBRD), the World Bank (Business Environment and Enterprise Performance Survey, BEEPS, and Worldwide Governance Indicators Database, WGI), the International Budget Partnership (Open Budget Survey), the Business International Corporation, now a part of the Economist Group (corruption or questionable payments in business transactions), Gallup International (Voices of the People Survey), the PRS Group (International Country Risk Guide, ICRG), the World Economic Forum (Global Competitiveness Report, GCR), and the United Nations (International Crime Victims Surveys, ICVS) (see Arndt and Oman 2006; Malito 2014 for further discussion).1 This increased attention to corruption measures and to the quality of governance in general in developing—and emerging-market economies is explained by a combination of reasons: (i) The spectacular increase in international investment in developing countries, (ii) The end of the Cold War, (iii) The failure of development policy reforms in the 1980s and 1990s, and (iv) Increased

1With a few exceptions (e.g., CPI, BEEPS, and WGI), these ratings are produced by private risk-rating agencies and accessing them costs several thousand dollars (Mauro 1995).

2  Measurement Issues     41

awareness of the importance of politics in economic development and policy reform (Arndt and Oman 2006). Corruption indicators can be classified according to various dimensions: Whether they are conducted by means of interviews with experts or with persons (including firms); whether they are first-hand or based on a combination of other indicators; and whether they are based on perception or actual experience. An example of an indicator using pools of experts who assess the level of corruption prevailing in a country is the Business International Corporation indicator. In contrast, the indexes provided in the World Economic Forum’s GCR or the United Nations’ ICVS are based on surveys of persons (including firms). The Control of Corruption Index, which is part of the WB’s World Governance Indicators, and the ICRG index are composites based on other indexes and aggregated following different methodologies. Perception-based indexes include TI’s Corruption Perceptions Index, while those based on experience include the ICVS index. While all these indexes are based on surveys of opinions or experiences, another measure, mainly used by studies on the USA, is based on the number of convictions for the abuse of public office in a US state. The data come from the Public Integrity Section of the US Department of Justice and pertain only to convictions that result from federal prosecutions. Users of these indicators are very diverse and include international investors, aid agencies, development analysts, academics, and governments. For investors, these indicators help to assess the risk of investing in or lending to a given country. Very often, business is confronted with corrupt behavior. The most common such behavior takes the form of demands for special payments and bribes connected with regulations such as import and export licenses, exchange controls, tax assessments, police protection, or loans. Such corruption makes it difficult to conduct business effectively and in some cases may entail the withdrawal or cancelation of an investment project. As far as official national and multilateral aid agencies are concerned, these indicators are increasingly and extensively considered as useful to guide official development assistance allocation decisions. Many bilateral and multilateral donors use the rankings provided by these aggregate indicators to determine the benefits of providing aid, and how much, to a given country. The reason for

42     K. Sekkat

this is that corruption distorts the economic and financial environment and reduces the efficiency of the allocated aid (Arndt and Oman 2006; Razafindrakoto and Roubaud 2010).

2 Selected Indicators This section compares the three most widely used indicators of perception of corruption: the CPI, ICRG, and WGI. While the CPI and the WGI are available free of charge, the ICRG has to be bought from the PRS Group. Since most of the debate about the merits of different indicators focuses on perceptions versus reality, we also consider an indicator based on experience and drawn from the World Enterprise Surveys (WES), which are also available free of charge. Conducted by the World Bank in various countries, the WES ask, among other things, questions about managers’ experiences with corruption. The following presentation is brief, and the interested reader should consult the respective organizations’ websites for more detailed information.

2.1 Corruption Perceptions Index The CPI is produced by Transparency International (TI), which is an NGO with a secretariat in Berlin. TI was established in the early 1990s at the initiative of a few individuals who decided to take a stand against corruption. Now TI is present in more than 100 countries. It works with partners in government, business, and civil society to put effective measures in place to tackle corruption. TI is funded by various governmental agencies, international foundations, and corporations. It claims to be politically non-partisan and places great importance on independence. As a part of its activity, TI publishes the CPI, which assesses expert perceptions of the level of corruption in the public sector across countries, the Global Corruption Barometer, which is concerned with attitudes toward and experiences of corruption among the general public, and the Bribe Payers Index, which focuses on the propensity of a country’s firms to bribe abroad.

2  Measurement Issues     43

The most widely used indicator in academic research is the CPI, which ranks countries in terms of the degree to which corruption, defined as the abuse of public office for private gain, is perceived to exist among public officials and politicians. It is a composite index, a poll of polls, drawing on data from experts (including those living in the countries being evaluated) and business surveys carried out by a variety of independent and reputable institutions, such as Freedom House (FH), Gallup International (GI), the Economist Intelligence Unit (EIU), the International Crime Victims Survey (ICVS), the World Bank, and the World Economic Forum (WEF). The surveys used in compiling the CPI ask questions relating to the misuse of public power for private benefit including bribery, kickbacks, and embezzlement. The index is computed yearly as an average of other indexes. A standardization procedure converts all the data sources to a scale of 0–100 where 0 indicates the highest level of perceived corruption and 100 the lowest level of perceived corruption. For a country to be included in the CPI, a minimum of three reliable sources of corruption-related data is required. Accordingly, inclusion in the index is not an indication of the existence of corruption but of the availability of data. The country with the lowest CPI score is the one where corruption is perceived to be greatest among those included in the list. However, one of the drawbacks of the CPI is that year-to-year differences in country scores do not only result from changing perceptions of the country’s performance but also from changes in sample and methodology (Lambsdorff 1999). For this reason, a time comparison of the index is not recommended. Moreover, the reliability of the CPI differs across countries. The scores of countries with a high number of sources and small differences in evaluations are more reliable.

2.2 International Country Risk Guide The ICRG is produced by the PRS Group, Inc., a private firm based in Syracuse, USA. Founded in 1979, the PRS Group is a global leader in political risk assessment and country risk forecasts. It produces a number of products at regular intervals throughout the year. These

44     K. Sekkat

publications and data are used extensively by investors and businesses, colleges and universities, private equity groups, and the main multilateral institutions. ICRG staff collect political information and financial and economic data and convert them into risk points for each individual risk component (political, financial, and economic risk). The political risk assessments are made on the basis of a subjective analysis of the available information, while the financial and economic risk assessments are made solely on the basis of objective data. In addition, the ICRG provides an overall score for each country. The political subcategory of the ICRG is the one that is relevant to our purpose since it covers corruption. In fact, the political subcategory is the sum of 12 components covering both political and social attributes. The components and their score intervals (in parentheses) are Government Stability (0–12), Socioeconomic Conditions (0–12), Investment Profile (0–12), Internal Conflict (0–12), External Conflict (0–12), Corruption (0–6), Military in Politics (0–6), Religious Tensions (0–6), Law and Order (0–6), Ethnic Tensions (0–6), Democratic Accountability (0–6), and Bureaucracy Quality (0–4). The highest value is 100 and the lowest is zero. Higher values mean less risk. The corruption component is an assessment of corruption within the political system. It takes account of the most common form of corruption encountered directly by business, which is financial corruption in the form of demands for special payments and bribes connected with import and export licenses, exchange controls, tax assessments, police protection, or loans. However, the measure is more concerned with actual or potential corruption in the form of excessive patronage, nepotism, job reservations, “favor-for-favors”, secret party funding and suspiciously close ties between politics and business.

2.3 Worldwide Governance Indicators The WGI are produced as a part of the WB’s “Governance Matters” project. The indicators measure perceptions of six dimensions of governance in most of the countries in the world. The six dimensions are

2  Measurement Issues     45

Voice and Accountability, Political Stability and Absence of Violence, Government Effectiveness, Regulatory Quality, Rule of Law, and Control of Corruption. The indicators are based on a statistical compilation of responses on the quality of governance given by a large number of enterprises, citizens, and expert surveys reported by a number of institutes, think tanks, non-governmental organizations, and international organizations. Respondents are from both industrial and developing countries. The governance indicators cover over 200 countries and territories since 1996. Each indicator varies between −2.5 and +2.5, with higher values meaning better conditions. The Control of Corruption Index (CCI) assesses perceptions of the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as state capture by elites and private interests.

2.4 World Enterprise Surveys The WES began in 2002 and have been conducted by different units within the World Bank (previous names include PICS and Investment Climate Surveys). Since 2006, most data collection efforts have been centralized within the Enterprise Analysis Unit. Centralizing survey implementation has resulted in a unified set of core survey questions and consistent application of survey methodology across countries. The WES is a firm-level survey of a representative sample of an economy’s private sector. The surveys cover a broad range of information including firm performance and a set of business environment variables such as access to finance, corruption, infrastructure, crime, and competition. The WES team collects these data from face-to-face interviews with top managers and business owners in over 155,000 companies in 148 developing economies. The enterprise surveys implemented in Eastern Europe and Central Asia are also known as Business Environment and Enterprise Performance Surveys (BEEPS) and are jointly conducted by the WB and the EBRD. Examples of questions asked in the framework of the WES and related to corruption are: (i) When visited by an inspector, has a gift or an informal payment been

46     K. Sekkat

requested? (ii) What percentage of the value of a contract with the government should your firm pay to officials to secure contract? and (iii) What percentage of an import transaction should your firm pay to customs officials?

2.5 Criticisms The common indicators of corruption have been widely criticized. Some criticisms concern the appropriateness of measuring an unobservable phenomenon and pertain to the whole group of indicators. Others concern specific indicators. In what follows, we focus on the criticisms pertaining to the whole group of indicators. Criticisms of specific indicators, being largely concerned with the distinction between perception and experience, are discussed separately (Sect. 3). Regarding the indicators as a group, one fundamental criticism concerns the relationship between the definition of corruption and its measurement (Andersson and Heywood 2009). Definitions determine what should be captured when measuring corruption and therefore condition the manner in which the methodology for collecting the information is set up. Just as there is no one definition of corruption, so there is no one unique way of constructing the indicators. Accordingly, the validity of a given measure is only relevant for the type of corruption covered by the definition. What is more, almost all indicators are based on expert opinions and/or surveys of citizens and firms. According to Bertrand and Mullainathan (2001), responses to survey questions are in many instances subjective. In addition to the subjectivity in interpreting the same facts, there are different issues connected with the way the surveys are conducted. For instance, survey responses may be affected by the ordering of questions. Whether question X is preceded by question Y or vice versa can substantially affect the answers. One reason for this effect is that people attempt to provide answers consistent with the ones they have already given in the survey. Another effect is related to the wording of questions. A classic example concerns the responses to the two following questions: “Do you think the United States should

2  Measurement Issues     47

forbid public speeches against democracy?” and “Do you think that the United States should allow public speeches against democracy?” Experiments show that while more than half of respondents answered yes to the first question (public speeches against democracy should be forbidden), three-quarters answered no to the second (public speeches against democracy should not be allowed). Finally, and more fundamentally, respondents may make little intellectual effort in answering the questions. They may not try to recall all the relevant information or to completely read all the alternative responses. Hence, the order of the possible responses in the list matters because respondents may simply pick the first or last available alternatives in the list. Turning now to experts, their opinion is, in general, based on information collected through publications or the media, which may bias their judgment For instance, the media may report on cases of corruption precisely because an active anti-corruption policy has been put in place. Experts might then consider that corruption is high or rising when just the opposite is true (Razafindrakoto and Roubaud 2010). Experts can also influence one another or use the same source, which greatly reduces the gains from using multiple experts. Moreover, corruption is a complex and multifaceted phenomenon, which raises the issue of weighting its different dimensions. For instance, in a country where bureaucratic (petty) corruption is low and political (grand) corruption is high, how should scores be weighted in comparison with a country where the reverse holds? Another issue concerns the fact that the indicator for a country is, in general, composite and based on different sources. Not only might aggregating different data sources be problematic, but the different sources might not cover all the countries under consideration, which implies that alternative sources of information need to be found. If missing data for some indicators force experts to derive information from other sources, the final outcomes might not be reliable unless the necessary efforts are devoted to examining whether the information from the different sources can be adequately aggregated (Razafindrakoto and Roubaud 2010). In defense of the existing corruption indicators, their supporters point to the fact that the estimated effect of these indicators on various economic variables corresponds to the theoretical expectations,

48     K. Sekkat

which constitutes indirect proof of their relevance. If the indicators only captured noise, they should not be significantly associated with real phenomena (Razafindrakoto and Roubaud 2010). Bertrand and Mullainathan (2001) provide a more careful treatment of the concerns about the measurement of corruption. They recall first that if the measurement error is white noise and uncorrelated with the characteristics of the individual, there should be no problem. However, the above discussion suggests that the measurement error is unlikely to be white noise. First, the mean of the error term will not necessarily be zero within a survey. For example, the fact that a survey uses “forbid” rather than “allow” in a question will affect answers. Second, many of the findings in the literature suggest that the error term will be correlated with observable and unobservable characteristics of individuals. The authors therefore conducted various experiments which led to two important insights. First, if the measurement error is small enough, subjective measures may be helpful as independent variables in predicting outcomes, with the caveat that the coefficients must be interpreted with caution. Second, subjective variables cannot reasonably be used as dependent variables, given that the measurement error likely correlates in a causal way with the explanatory variables In the context of the debate about measurement, Svensson (2005) argues that other approaches can lead to a better assessment of corruption. Public Expenditure Tracking Surveys (PETS) and Quantitative Service Delivery Surveys (QSDS) are among these approaches. These surveys permit measurement of corruption at the level of individual agents, such as schools, health clinics, or firms. They also permit the study of mechanisms responsible for corruption, including capture of public funds and bribery. They therefore seem useful for locating and quantifying political and bureaucratic capture, leakage of funds, and other problems. Typical PETS consist of surveys of frontline providers (schools and clinics and their staff) and local government (politicians and public officials) complemented by central government financial and other data. Such surveys track the flow of resources through the different levels in order to determine how much of the originally allocated resources reach each level. QSDS consist of systematic quantitative data collection on finances, inputs, outputs, pricing, quality, oversight, and

2  Measurement Issues     49

other aspects of service provision. They can be applied to government, private for-profit and not-for-profit providers. However, PETS and QSDS require considerable effort, cost, and time compared with some of their alternatives, especially surveying perceptions of users. These approaches have been used in Uganda, Ghana, Peru, Tanzania, Zambia, and Bangladesh.

2.6 Comparative Analysis of the Indicators In this section, we compare country corruption rankings to see the similarities and differences across indicators. We start by comparing the three perception indicators for the year 2010. For each indicator, countries are classified into quartiles according to their scores, from the most corrupt to the cleanest. Then, we cross the classifications of each pair of indicators, which gives three panels in Table 1. We keep only the countries which are present in the sources under comparison. The number in the diagonal gives the number of countries that are classified in the same quartile by both sources. For instance, in the first panel of the table, 24 countries, that is 18% of the total number of countries, are classified in the first quartile (most corrupt) based on both the CPI and Table 1  Comparison of country rankings based on perception indicators in 2010 1st quartile

2nd quartile

3rd quartile

4th quartile

1st quartile 2nd quartile 3rd quartile 4th quartile

24 (18%) 8 (6%) 2 (1%) 1 (1%)

8 (6%) 16 (12%) 10 (7%) 3 (2%)

1 (1%) 9 (7%) 19 (14%) 3 (2%)

0 (0%) 0 (0%) 2 (1%) 29 (21%)

1st quartile 2nd quartile 3rd quartile 4th quartile

31 (23%) 4 (3%) 0 (0%) 0 (0%)

2 (1%) 29 (21%) 6 (4%) 0 (0%)

0 (0%) 0 (0%) 27 (20%) 5 (4%)

0 (0%) 0 (0%) 0 (0%) 31 (23%)

1st quartile 2nd quartile 3rd quartile 4th quartile

23 (17%) 7 (5%) 2 (1%) 1 (1%)

9 (7%) 14 (10%) 8 (6%) 2 (1%)

1 (1%) 11 (8%) 20 (15%) 1 (1%)

(0%) 1 (1%) 3 (2%) 32 (24%)

CPI ICRG

CPI WGI

WGI ICRG

50     K. Sekkat

ICRG. The next cell down in the panel shows that eight countries (6% of the total) are classified in the second quartile by the CPI but rank in the first quartile according to the ICRG. The first panel in the table shows that based on the CPI, 35 countries are classified in the first quartile, 37 in the second quartile, 32 in the third, and 31 in the fourth. In percentage terms, this gives 26, 27, 24, and 23%, respectively. According to the ICRG, the distribution is 33 countries for each of the first three quartiles and 36 for the fourth. In percentage terms, this gives 24 and 26%. Although only indicative, these results suggest that a CPI-based classification leads to a slightly more “severe” judgment about countries’ corruption than one based on the ICRG. According to the CPI, the most corrupt countries account for 26% of the total, while based on the ICRG the share is 24%. The shares of the least corrupt countries are 23 and 26%, respectively. In the second panel, the distribution based on the CPI is, of course, the same as in the first panel. Surprisingly, the distribution based on the WGI is very similar to the one based on the ICRG in the first panel. The third panel confirms the fact that the ICRG—and WGI-based results are similar. It therefore seems that considering either the ICRG or the WGI as a second indicator in addition to the CPI can be informative, but that adding a third indicator does not bring any new information. The first panel in Table 1 also shows that the classification of 88 countries across quartiles stays the same irrespective of the classification base considered (CPI or ICRG). This represents 65% of the total number of countries in the sample. The second panel shows that the classification of 118 countries across quartiles stays the same irrespective of the classification base considered (CPI or WGI). This represents 87% of the total number of countries in the sample. When taken together with the discussion about the distribution based on each index separately, it seems that using both the CPI and the ICRG (rather than the WGI) can offer a more balanced view about a country’s corruption. Having compared the perception indexes, we turn now to their comparison with an index based on experience. This is the index constructed based on responses to the WES, which focuses on firms in developing countries only. While the period of the responses corresponds to the 2000s, the year of response differs across countries, as do

2  Measurement Issues     51 Table 2  Comparison of country rankings: Experience versus perception 1st quartile

2nd quartile

3rd quartile

4th quartile

1st quartile 2nd quartile 3rd quartile 4th quartile

12 (14%) 6 (7%) 2 (2%) 0 (0%)

7 (8%) 9 (11%) 3 (4%) 5 (6%)

1 (1%) 5 (6%) 11 (13%) 3 (4%)

0 (0%) 2 (2%) 4 (5%) 14 (17%)

1st quartile 2nd quartile 3rd quartile 4th quartile

7 (8%) 9 (11%) 2 (2%) 0 (0%)

8 (10%) 7 (8%) 10 (12%) 5 (6%)

1 (1%) 5 (6%) 5 (6%) 7 (8%)

4 (5%) 1 (1%) 3 (4%) 10 (12%)

CPI WES

ICRG WES

country scores. Table 2 shows that the distribution of countries across quartiles based on the CPI is similar to the distribution based on the WES. In contrast, the distribution based on the ICRG is different from the one based on the WES. Moreover, according to the WES, the share of firms in the fourth quartile (least corrupt) is much higher than the share derived from the ICRG. Conversely, according to the WES, the share of firms in the second quartile (corrupt) is much lower than the share derived from the ICRG. With the relevant caveats, a country is more likely to be classified as corrupt based on the ICRG than based on the WES. The first panel in Table 2 shows that the classification of 46 countries across quartiles stays the same irrespective of the classification base considered (CPI or WES). This represents 55% of the total number of countries in the sample. In contrast, the number of countries which are in the same quartile according to both the ICRG and the WES is 29, or 35% of the total. Overall, the results suggest non-negligible differences in country corruption rankings between perception-based and experience-based measures. Ahmad (2001) provides a more rigorous analysis of the potential differences across perception-based measures of corruption. The methodologies used include absolute rank correlation, rank correlation over time, rank correlation across groups of countries and regressions. The regression analyses aim at examining whether corruption indexes can be explained by the same set of variables, which is another method of comparison.

52     K. Sekkat

For correlations, four sources of indicators are considered: (i) World Competitiveness Report (WCR) in 1990, 1992, 1994 and 1996, (ii) TI from 1995 to 1998, (iii) the ICRG from 1982 to 1995, and (iv) WGI in 1996. The results show that the correlation coefficients between ICRG95 and WCR96 (0.80) and between WCR96 and WGI96 (0.82) are higher than the correlation between ICRG95 and WGI96 (0.56). The high correlation between WCR96 and ICRG95 can be attributed to the fact that both indexes focused on firms or businesses engaged in foreign activities. Similarly, the correlation between WCR96 and WGI96 is high possibly because these indexes represent internal viewpoints about corruption. In contrast, the correlation coefficient between WGI96 and ICRG95 is only 0.56. According to the author, one plausible reason for this low correlation is that the ICRG95 corruption indexes include foreign firms with external viewpoints about corruption, while WGI96 focuses on internal viewpoints and local business firms. However, our above comparison of the ICRG and WGI for the year 2010 shows that the distribution of countries across quartiles based on the WGI is very similar to the ICRG-based distribution. This is not in line with the author’s explanation and questions the consistency across classifications over time. The subsequent analysis in the paper examines the correlation of each index over time but not the change in the correlation between indexes over time. To examine whether a given index ranks countries consistently over time, rank correlation coefficients of various corruption indexes from the same sources but for different time periods are calculated. It appears that the ICRG, WCR, and TI indexes are highly correlated between any two consecutive surveys. Although the value of the correlation coefficient decreases as the time span between the two indexes expands, corruption rankings seem persistent over time. For instance, the minimum value of the correlation coefficient between ICRG82 and ICRG95 is 0.68, which is high. Turning to regression analyses, the dependent variables, i.e., corruption indexes, are taken from TI (1996–1998), the ICRG (1995), the WCR (1996), and WGI (1996). The set of explanatory variables includes an index of bureaucratic competition, newspaper circulation, urbanization, average years of schooling, an index of political liberty, the

2  Measurement Issues     53

share of government consumption in GDP, and an index of government regulation. The index of regulation is a sum of seven indexes assessing: (i) regulations for starting a business and new operations, (ii) price controls, (iii) regulations on foreign trade, (iv) labor regulations, (v) foreign currency regulations, (vi) tax regulations, and (vii) safety or environmental regulations. A separate estimation on the same set of explanatory variables is conducted for each index/year. However, the regression results must be considered with caution because these results may be influenced by the small sample size. There are only 20 countries due to data availability. The regressions lead to several interesting results. First, in all regressions, the independent variables explain more than 50% of variations in corruption. Second, with one exception, the relationship between the independent variables and the corruption index is consistent across all sources. Third, the coefficients of the measure of government regulations, bureaucratic competition, average years of schooling, and the index of political liberty have the expected signs in all cases, while the coefficients of urbanization and government size have the wrong signs in all cases. Overall, the various perception indicators are very similar.

3 Perception Versus Experience 3.1 Reasons for the Divergence An important line of criticism levelled at all corruption indicators concerns the distinction between measures based on perception of corruption and those based on experience of corruption. This matters because perceptions do not measure corruption itself but only opinions about its incidence. Subjective assessments of corruption may be critically influenced by other factors such as level of development or religion. Moreover, there is a risk that subjective measures of corruption may be biased through a band effect, which relates to the fact that respondents’ perceptions tend to follow the common perceptions of corruption in a given country. It is also likely that one element contributing

54     K. Sekkat

to perceptions of corruption in different countries is the previous ranking, thus introducing a certain endogeneity and persistence to the index (Andersson and Heywood 2009). The resulting discrepancy between perception and experience is especially important because the “distance” between opinions and experiences varies haphazardly from country to country (Weber 2007, p. 6). This makes cross-country comparisons very hazardous. In response to these critics, it can be argued that even biased perception is a relevant variable for investor decisions and government reform efforts. Moreover, measures based on experiences such as convictions might also be biased because gathering evidence is not easy, with the result that a number of corrupt transactions escape justice. Also, in a non-democratic regime, many allegations and convictions of corruption might be motivated by a desire to eliminate opposing voices. Likewise, basing the measure of corruption on experience is not necessarily the best solution because the respondent might, for various reasons, over- or under-report, for instance to avoid reprisals. It is fair to acknowledge that a perfect measure of corruption does not exist because, as with many illegal activities, the phenomenon is non-observable. However, this does not mean that available measures of corruption are not valuable. Given their importance both for investors and for countries seeking to improve the quality of their governance, continuing to improve the accuracy and transparency of the existing indicators seems to be a promising way forward. However, as recommended by Razafindrakoto and Roubaud (2010), the process of constructing indicators should make a clear separation between perception and experience. The perception and reality of a phenomenon are two distinct components, interlinked in a complex manner, and as such interactions between them should be treated with caution. For instance, perceptions can influence behavior in significant ways: If we believe that all around us people are engaging in corrupt behavior, that may make us more likely to adopt such practices ourselves. According to Razafindrakoto and Roubaud (2010), sources based on objective measurements should be excluded from the construction of global perception indicators. However, objective indicators based on real experiences

2  Measurement Issues     55

of the phenomenon, and not just on the idea formed of it, should be developed in parallel.

3.2 Evidence of Divergence Razafindrakoto and Roubaud (2010) conducted an informal comparison of victimization—and perception-based measures of corruption. Victimization is measured based on International Crime Victims Surveys (ICVS). First implemented in 1987, the ICVS include 55 countries, with samples of between 1000 and 2000 respondents per country. In 1996, a question on bribe victimization was included in the surveys. The question mainly concerns petty corruption and focuses on the experience of having been asked to pay a bribe or having observed a bribe being paid. It covers bribes to police and other public officials in the workplace, courts, schools, and public health/hospital systems. A similar question is asked about the perception of corruption in the same institutions. The results show that increases in corruption victimization go with significant increases in the perception of corruption. On average, however, 60–70% of respondents do not have any direct personal corruption experiences within the 12 months prior to the survey, which suggests that this finding should not be overstated. Respondents who have not been personally affected by corruption in the year prior to the survey generally view corruption as being pervasive in their countries. Moreover, except in the case of Paraguay, the experience of respondents who had been victimized two or more times in the previous year increases the level of perception by ten points on a 100-point scale, which is small. In sum, estimations of corruption on the basis of perception do not fit with experience. Dilyan and Ujhelyi (2014) examine a similar question (deviations of perceptions from experience) by regressing the corruption perception indexes on measures of experience of corruption and other country characteristics. To measure experience of corruption, the study used two sources. The first source is the 1996 and 2000 rounds of the International Crime Victims Survey (ICVS) administered in a total of

56     K. Sekkat

58 countries. The second is the World Business Environment Survey carried out by the WB and the EBRD in 1999–2000. The latter focuses on firms’ experience of corruption. Regarding perception, the three most widely used cross-country indexes are used. These are the WGI Control of Corruption Index, the CPI, and the ICRG corruption index. Control variables include dummies for British legal origin and for countries that were never colonized, the fraction of Protestants in the population, an ethnic fractionalization index, and a democracy index. Overall, the results indicate a disconnection between perception and experience. A country level of corruption perception is determined mainly by economic, cultural, and institutional factors rather than corruption experience. GDP, Protestantism, a democratic past and a common law system reduce perceived corruption, while natural resource endowments increase it. Moreover, perceptions exhibit diminishing sensitivity to experience. Once control variables are added, the point estimate of the effect of experience drops dramatically. In the most extensive specification, experience has a positive but insignificant effect on perceptions. When significant, the effect of experience is not always robust to outliers. Looking at the effect of different types of corruption experience (police, workplace, etc.), it appears that perception indexes are biased measures of a specific type of corruption experience. The results also show that the relationship between perception and corruption is nonlinear and depends on the level of experience and on the population. Moreover, the findings also suggest the presence of psychological biases in forming perceptions. More educated respondents report more widespread corruption for a given level of experience. Age has a nonlinear impact on perceptions, with a positive effect for younger people that steadily declines and becomes negative around the age of 50. Finally, higher income has a significant positive impact on perceived corruption, while individuals from larger cities perceive lower corruption. Razafindrakoto and Roubaud (2010) commissioned two types of surveys, which were conducted simultaneously in eight African countries, including Benin, Burkina Faso, Côte d’Ivoire, Madagascar, Mali, Niger, Senegal, and Togo, from 2001 to 2003. The objective was to examine the issue of perception versus experience pertaining to corruption. The first type of survey covered a sample of over 35,000 residents

2  Measurement Issues     57

and took an objective measure of the frequency and characteristics of petty bureaucratic corruption. The second type, called a mirror survey, reported the opinions of 350 experts on the same subjects in the same countries. The surveys also collected information on subjects including poverty, governance and democracy. In the first type of survey, respondents were asked whether they had been personally affected by corruption in the year preceding the survey and, if so, on what occasion (type of transaction and service concerned) and the total sum paid over the year. Corruption was defined as an illicit payment (in money or in kind) to a government official. The mirror surveys involved experts from the South and the North, who were asked two sets of questions. The first set concerned the expert’s own opinion on certain governance issues. The second set focused on what the experts thought interviewed residents answered on average. Comparing the two types of surveys showed that experts systematically overestimated the level of corruption suffered by citizens. While on average across countries 13% of the population said that they had been direct victims of corruption, the experts estimated this rate at 52%, i.e., a ratio of one to four. The ratio of overestimation varied from 2.5 in Burkina Faso to over 6 in Benin, Niger and Togo. Moreover, there was no correlation between the measure in the mirror survey and the one based on the population’s experience. The correlation coefficient was actually negative (−0.13), although nonsignificant. Interestingly, the mirror survey findings are correlated with the indicators published in the international databases. The Control of Corruption Index (CCI) in 2002 has a correlation coefficient of around 0.67 with the level of corruption found by the mirror survey. In contrast, the correlation between the actual rate of corruption and the CCI is not significant. Expert perceptions appear to be highly associated with the perception indicators published in the international databases, while actual experiences are not. The analysis then turns to determining the explanatory factors for the expert opinions to understand the disconnection between reality and experience. Using the individual mirror survey data, a first model explains the experts’ opinion in terms of the actual extent of corruption and a number of variables representing the experts’ origins, gender,

58     K. Sekkat

claims to know the country well and claims to know the subject well. The second model is similar to the first except that the actual extent of corruption is replaced by the CCI. The regressions confirm that expert assessments of the extent of corruption do not tie in with the actual level of corruption. However, they are significantly linked with the CCI. Two factors could be put forward to explain this link. Either the experts (or some of them) know of the CCI and are directly influenced by it, or expert opinions are conditioned by a common core of factors (e.g., the country’s overall image in terms of the quality of democracy and economic governance). Adding other explanatory variables, the authors find evidence of ideological biases too: the experts most in favor of free-market principles or who consider that the country has not sufficiently adopted its precepts are more often wrong and tend to unfairly penalize that country. Finally, the experts appear to base their judgments on a consistent, but erroneous, implicit cultural model of the way in which Africa operates. They tend to hugely overestimate the population’s level of tolerance of corrupt practices and underestimate the importance the population attaches to matters of good governance Olken (2009) complements the preceding analysis by focusing on another continent. The paper compares Indonesian villagers’ reported perceptions of corruption in a road-building project in their village with a more objective measure of missing expenditures. The study covers 477 villages in the two most populous provinces of Indonesia: East Java and Central Java. The data were collected between September 2003 and August 2004. The objective measure of corruption is constructed as follows. A team of engineers and surveyors excavated, after the roads were completed, core samples in each road to estimate the quantity of materials used, surveyed local suppliers to estimate prices, and interviewed villagers to determine the wages paid on the project. These data were used to construct an estimate of the actual cost of each road. This estimate was compared to what the village reported it spent on the project. The difference between the two amounts is used as a measure of actual corruption. To obtain data on villagers’ perceptions of corruption, a household survey was conducted asking villagers about the likelihood of corruption in the road project.

2  Measurement Issues     59

Using an ordered probit model, the respondents’ answers to the question about perceptions of corruption in the road project are explained in terms of the measure of actual corruption and a set of dummies for the household sampling methods, the type of questionnaire distributed, and the experimental treatments. The results show that while villagers’ reported perceptions of the likelihood of corruption are correlated with actual corruption, the sensitivity of perceptions to actual corruption is low. Increasing the actual corruption measure by 10% is associated with just a 0.8% increase in the probability of a villager believing that there is any corruption in the project. However, residents are not able to distinguish between general levels of corruption in the village and corruption in the particular road project. The inability of villagers to detect corruption in a specific road project is in line with the argument that officials are strategic in how they hide corruption. Monitoring the construction of a specific road project requires specialist auditors who can detect multiple types and complex ways of cheating and corruption. It also appears that reported perceptions are biased. Individual characteristics such as education and gender are systematically correlated with reported perceptions of corruption in the road project. Moreover, village characteristics also bias perceptions. Ethnic heterogeneity is associated with higher levels of reported corruption perceptions while increased levels of participation in social activities are associated with lower levels of reported corruption perceptions. Interestingly, ethnic heterogeneity is associated with lower levels of actual corruption while participation in social activities is not correlated with actual corruption. Echoing Bertrand and Mullainathan (2001), these last findings suggest that perceptions cannot reasonably be used as dependent variables in empirical analyses.

4 Persistence of Corruption We documented in Chapter 1 the fact that corruption is pervasive throughout the world. In one form or another, and to a lesser or greater extent, it exists in all societies, at all stages of development and in all types of activities. This is the reason why the fight against corruption ranks so highly on the agendas of governments and international

60     K. Sekkat

organizations. Another aspect of corruption which motivates further concerns about the phenomenon is its apparent persistence (Seldadyo and De Haan 2011). Once entrenched, corruption seems difficult to eliminate despite an understanding of its causes and consequences and specific attention on the part of policymakers. Some countries appear to get stuck in a bad equilibrium characterized by highly pervasive and persistent corruption while others witness a persistently low prevalence of corruption. These observations raise questions about the degree of persistence of corruption and the reason for such persistence (Seldadyo and De Haan 2011). Apart from its scientific appeal, an understanding of the economic, political, and historical factors underlying the persistence of corruption is of practical importance since it concerns the capacity of countries to improve their situation through well-designed institutional reforms. If the persistence of corruption reflects a frozen phenomenon, some countries would be predestined to remain poor and corrupt. Attempts at improving their situation would consequently prove futile. Even if corruption is not frozen but takes a very long time to change, governments would have few incentives to enter into costly reforms to eradicate corruption. Rulers have a limited time horizon, and there is no reason to expect them to implement policies that would produce effects after they have left office, especially if those policies also result in short-term costs. Moreover, uncertainty about the distribution of the gains and losses of reforms may cause individuals to favor the status quo. Specifically, individuals who would support the reform ex post, that is, when its effects have materialized, might oppose it ex ante in case they have to share its cost. The degree of persistence of corruption or, alternatively, the length of the period between the implementation of reforms and the materialization of their impacts, affects the political sustainability of reforms and the viability of the government that considers implementing them. Finally, the persistence of corruption may reflect the coexistence around the world of countries that are trapped in a bad equilibrium characterized by high pervasiveness of corruption alongside others with a persistently low prevalence of corruption. This suggests the existence of groups of countries featuring similar levels of prevalence within groups and different levels of prevalence between groups, i.e., two or more

2  Measurement Issues     61

economic equilibria around the world. This section deals with the issue of corruption persistence. It will first present available evidence about the phenomenon and then examine the reasons behind it.

4.1 Evidence of Persistence As a first illustration of the possible persistence of corruption, Table 3 compares the change in rankings of a large number of countries between two sufficiently distant points in time. The exercise is conducted using the three widely used indexes of perception of corruption. For each indicator and year, countries are classified into four quartiles. We then compute the number of countries which have changed quartile between the points in time for the same indicator. For instance, the table shows that based on the CPI, eight countries (19% of the total CPI sample) moved to a better quartile, that is, became “less corrupt”. The results should, however, be taken only as indicative given the criticisms discussed above and the fact that TI warns against year-to-year comparisons of the CPI. The table shows that the rankings of 75% of the countries in the CPI did not change between 1995 and 2010. The percentages are 44% in the ICRG sample and 61% in the WGI sample. Measured persistence thus appears to be markedly highest with the CPI and lowest with the ICRG. Ahmad (2001) examines a similar question using the indicators of corruption from the WCR for the years 1994–1996, TI for the years 1997–1998, and the ICRG for the years 1982–1995. The author categorizes countries into three groups: clean, partly corrupt, and corrupt. The focus is on the changes in corruption rankings. Table 3  Change in country ranking over time

Better quartile No change Worse quartile Total

CPI 1995–2010

ICRG 1995–2010

WGI 1996–2010

6 (15%) 31 (75%) 4 (10%) 41 (100%)

36 (28%) 57 (44%) 35 (27%) 128 (100%)

35 (19%) 113 (61%) 36 (20%) 184 (100%)

62     K. Sekkat

The results based on the WCR show that only eight countries (18% of the total WCR sample) out of 43 moved categories. Five countries improved their rankings while three did worse. Using the CPI, only four (8%) countries out of 52 changed categories. Three improved and one went down. Finally, using the ICRG, 34 (39%) countries out of 87 changed categories (27 improved and seven deteriorated). These results are not very different from the ones in Table 3. The above illustrations suggest that there is little change in countries’ rankings across quartiles. This points to the existence of at least two equilibria: one with high pervasiveness of corruption and another with low pervasiveness of corruption. In statistical terms, this pattern is manifested in the fact that the distribution of corruption across countries is multimodal. Herzfeld and Weiss (2007) investigate the existence of such multimodal distribution. They used Kernel density estimation techniques to analyze the cross-country distribution of corruption. Looking at the change in the distribution over time makes it possible to examine whether a particular distribution with, say, two groups of countries (bimodal: e.g., corrupt and clean countries) is gradually changing toward one with more or fewer classes of countries (unimodal distribution). The study uses data on corruption spanning the period from the mid-1990s to the early 2000s and drawn from four different sources: The Institute of Management Development (IMD), the World Economic Forum (WEF), the World Bank Control of Corruption Index (CCI), and Transparency International (CPI). The unimodality hypothesis is rejected for all indicators except the CCI. This suggests the existence of at least two groups of corrupt countries with different levels of prevalence of corruption. The change in the distributions over time suggests high persistence. Countries initially classified within the high corruption group are still classified within the same group at the end of the observation period. It seems that substantial changes in the economic, political, and cultural environment of countries in the “corruption club” are required before a significant decline in corruption can be expected. In sum, the estimated densities exhibit twin peaks, which gives empirical support to models predicting multiple equilibria.

2  Measurement Issues     63

While the previous studies support the hypothesis of a high persistence of corruption, Seldadyo and De Haan (2011) argue that this persistence might be purely because of the short-time horizon considered and that over a longer time horizon the hypothesis of corruption persistence would no longer hold. Using ICRG data for the period 1984–2008 and 101 countries, the authors apply different methodological approaches to investigate the issue of corruption persistence. The approaches consist of correlation, tests of convergence, panel unit root tests, regressions, and nonparametric tests. A first look at the data suggests that during the first half of the observation period, 40% of the countries in the sample reduced their level of corruption while another 40% had a stable level of corruption. In contrast, during the second half of the period approximately 75% of the countries (including various OECD countries) saw their level of corruption increase while only 15% of the countries decreased their level of corruption. The computed correlation between the level of corruption in 1984 and corruption in subsequent years is high but declining as the time lag increases. Over an interval of lags from 1 to 15 years, the correlation coefficients gradually decrease from more than 0.98 to 0.61. For higher lags, the correlation continues decreasing. Likewise, the R2 of regressions of a given year level of corruption on a constant and the level in 2004 decreases from more than 0.96 to only 0.36 as the time lag increases. Moreover, different tests of convergence among corruption groups, that is, whether corrupt countries improve their score more than clean countries, do not reject the convergence hypothesis. Panel unit root tests led to the same conclusion. Finally, an analysis of the change in the distribution of the level of corruption over time shows it to be marked by a slow shift in modality of the distribution from bimodal to unimodal. All the results in Seldadyo and De Haan (2011) support a long-term change in corruption and reject the persistence hypothesis, which contradicts the findings by Herzfeld and Weiss (2007). Comparing the two papers, however, leads to skepticism about the robustness of the findings of Seldadyo and De Haan (2011). Their analysis is based on only one indicator of corruption, the ICRG, while the one by Herzfeld and

64     K. Sekkat

Weiss (2007) is based on four different indexes (IMD, WEF, CCI, and CPI), and the results seem robust across indexes. Moreover, the results in Table 3, although indicative, suggest that the ICRG is the index which exhibits the least persistence among the studied indexes. Finally, from a policy point of view the “long term” (around 15 years) might be too long for policymakers and citizens to benefit from anti-corruption efforts which, as indicated above, would reduce government’s incentives to enter into potentially costly reforms.

4.2 Explanations of Persistence: Conceptual Discussion and Empirical Evidence Although the available evidence seems mixed, our discussion suggests that corruption is likely to be persistent. Understanding the reasons for such persistence is of high importance to the fight against corruption. Various explanations have been put forward and are comprehensively discussed in Bardhan (1997). For instance, liberal economists point the finger at the regulatory state, which creates opportunities for corruption through its overburdened and complex system of rules, permits, and licenses. Accordingly, corruption will persist as long as such a system exists and the difference in the prevalence of corruption across countries will reflect the difference in regulatory systems. Sociologists, meanwhile, emphasize the difference in social norms. An act which is seen as corruption in one society may well be considered as socially normal in another. Corruption persistence and the difference across countries regarding the prevalence of corruption are argued to be merely manifestations of the difference in norms. Finally, the explanations provided by economists largely rely in one way or another on the systemic dimension of corruption. Equilibria with high persistence of corruption and others with low persistence are both possible depending on whether the expected gain from corruption is reliant on the number of other people expected to be corrupt. Various models have been proposed to rigorously derive equilibria with persistent corruption. Cadot (1987) proposes a model of corrupt hierarchy where both the ordinary civil servant and the superior

2  Measurement Issues     65

officer are corrupt. With a corrupt superior officer, the probability that the ordinary civil servant will be sacked diminishes since corruption at one level feeds on the others. A similar idea can be found in Hillman and Katz (1987), who consider a hierarchical structure where an ordinary customs official is obliged to pay a part of his take of bribes to a superior. Tirole (1996) argues that the incentive for an individual to be corrupt depends on the collective reputation of the group to which he/she belongs. The existence of a large number of corrupt individuals impacts the collective reputation of the group and makes it beneficial for an individual to be corrupt too. In contrast, when the group as a whole has a good reputation, it pays for each agent to be honest as well. Lui (1986) focuses on the costs of auditing and shows that when corruption becomes more widespread in an economy, it is harder to audit corrupt officials effectively. Hence, the economy will remain highly corrupt. Conversely, if most officials do not accept bribes, it will be easier to discover those who do, and the corruption equilibrium will be lower (Herzfeld and Weiss 2007). Other studies put forward explanations not linked to the systemic dimension of corruption. Acemoglu (1995) considers individual choices between two activities: productive entrepreneurship and unproductive rent-seeking (corruption). A higher proportion of corrupt individuals reduces the return both to entrepreneurship and corruption. If the relative return to entrepreneurship falls faster, a multiplicity of equilibria may arise: As the number of individuals choosing to be corrupt increases (declines), the relative returns to entrepreneurship will decline (increase) and corruption becomes even more (less) attractive. In a similar vein, Damania et al. (2005) examine cases where a firm seeking to evade regulations bribes officials and politicians to encourage them to resist legal reforms which help to fight corruption. The authors show that political instability can weaken the institutions that are necessary to monitor and enforce compliance. Corruption therefore becomes more pervasive. Jain (2001) focuses on the transition from a low corruption regime to a high corruption regime. Consider a country with an ineffective legal system and which is affected by an unexpected external shock that increases corruption. The political elite, finding the increased income from corruption irresistible, will attempt to further undermine

66     K. Sekkat

the effectiveness of the legal and judicial systems through reduced resources. Such a decrease in resources will make it difficult for the legal system to combat corruption, thus enabling corruption to spread even more. Blackburn et al. (2006) also examine the transition across regimes (from a low to high corruption regime) but focus on the level of economic development. They construct a dynamic general equilibrium model of growth where the level of economic development and bureaucratic corruption are determined jointly. The incentives for corruption depend on aggregate economic activity, which, in turn, depends on the incidence of corruption. The model produces multiple development regimes, transition between which may or may not occur. Herzfeld and Weiss (2003) test for the role of the potential bidirectional relationship between corruption and the legal system as a driver of corruption persistence, a hypothesis suggested by Jain (2001). The exercise consists in estimating a two-equation system where corruption and legal effectiveness influence each other, using 2SLS to take account of endogeneity. Three different measures of corruption are used: the ICRG, the CPI, and data published by the Institute for Management Development (IMD) in the World Competitiveness Yearbook. The legal effectiveness measure is the indicator of law and order taken also from the ICRG. Control variables include gross domestic product, the index of political rights from Freedom House, the share of imports in GDP, the share of Protestants in the total population and dummies for a common law system, and ethnolinguistic division. The estimates of the different econometric models lead to unambiguous results. Legal effectiveness has a highly significant and negative impact on corruption. Conversely, corruption has a highly significant impact on legal effectiveness, suggesting that a higher level of corruption significantly reduces the acceptance of established institutions and undermines the quality of the judicial system. This is in accordance with the hypothesis by Jain (2001). Simulation experiments based on the econometric results indicate that a 10% exogenous increase in corruption combined with a 10% reduction in legal effectiveness will lead to a final effect of a 13% increase in corruption and a 21% decline in legal effectiveness.

2  Measurement Issues     67

Damania et al. (2005) draw three testable hypotheses from the theoretical model discussed above: (i) political instability reduces judicial efficiency, (ii) political instability indirectly increases corruption via its impact on judicial efficiency, and (iii) political instability reduces the level of compliance with regulations. These hypotheses are tested in the case of environmental compliance. The empirical exercise estimates a system of four equations on a cross-section sample of around 80 countries in the 1990s. Given the simultaneity of political stability, judicial efficiency, corruption, and compliance variables, an instrumental variable (2SLS) approach is used to estimate the equations. The dependent variables are political stability (from the WGI), judicial efficiency (from the WGI), corruption (from TI and WGI), and compliance (from the WEF). Each dependent variable also enters as an explanatory variable in the other equations. Control variables include GDP, democracy, the degree of racial tension, ethnolinguistic fractionalization, history of wars and civil wars, a dummy for the prevailing legal system, and trade openness. The empirical results provide strong support for the null hypothesis. Political instability appears to create institutional structures under which corruption is more pervasive and harder to eradicate. Political instability thus creates an environment in which corruption persists. Political instability also leads to lower levels of compliance with existing regulations due to its effect on judicial efficiency and corruption.

5 Conclusion Assessing the magnitude of corruption was long handicapped by the lack of systematic data collection. Since the 1980s, various national and international institutions have been filling this gap. Three approaches are taken to the assessment of corruption. One is based on the perception of experts, citizens, or business of the prevalence of corruption in a country or a sector. The second relies on actual acts of corruption and draws either on justice records or on individual experiences. The last approach uses Public Expenditure Tracking Surveys and Quantitative

68     K. Sekkat

Service Delivery Surveys. The results of the various assessment approaches lead, however, to different groupings of countries. Almost all the indicators used to assess the extent of corruption show a high degree of persistence. The explanations provided by economists largely rely in one way or another on the systemic dimension of corruption. Equilibria with high persistence of corruption and others with low persistence are both possible depending on whether the expected gain from corruption is reliant on the number of other people expected to be corrupt.

References Acemoglu, D. (1995). Reward Structures and the Allocation of Talent. European Economic Review, 39, 17–33. Ahmad, N. (2001). Corruption Perception Indices: A Comparative Analysis. Pakistan Development Review, 40(4), 813–830. Andersson, S., & Heywood, P. M. (2009). The Politics of Perception: Use and Abuse of Transparency International’s Approach to Measuring Corruption. Political Studies, 57(4), 746–767. Arndt, Ch., & Oman Ch. (2006). Uses and Abuses of Governance Indicators. OECD Development Centre. Bardhan, P. (1997). Corruption and Development: A Review of Issues. Journal of Economic Literature, 35(3), 1320–1346. Bertrand, M., & Mullainathan, S. (2001). Do People Mean What they Say? Implications for Subjective Survey Data. American Economic Review, 91(2), 67–72. Blackburn, K., Bose, N., & Haque, M. E. (2006). The Incidence and Persistence of Corruption in Economic Development. Journal of Economic Dynamics and Control, 30(12), 2447–2467. Cadot, O. (1987). Corruption as a Gamble. Journal of Public Economics, 33(2), 223–244. Damania, R., Fredriksson, P. G., & Mani, M. (2004). The Persistence of Corruption and Regulatory Compliance Failures: Theory and Evidence. Public Choice, 121(3), 363–390. Dilyan, D., & Ujhelyi, G. (2014). What Do Corruption Indices Measure? Economics and Politics, 26(2), 309–331.

2  Measurement Issues     69

Herzfeld, T., & Weiss, C. (2003). Corruption and Legal (In) Effectiveness: An Empirical Investigation. European Journal of Political Economy, 19(3), 621–632. Herzfeld, T., & Weiss, C. (2007). Corruption Clubs: Empirical Evidence from Kernel Density Estimates. Applied Economics, 39(12), 1565–1572. Hillman, A. L., & Katz, E. (1987). Hierarchical Structure and the Social Costs of Bribes and Transfers. Journal of Public Economics, 34(2), 129–142. Jain, A. K. (2001). Corruption: A Review. Journal of Economic Surveys, 15(1), 71–121. Lambsdorff, J. G. (1999). The Transparency International Corruption Perceptions Index 1999: Framework Document. Berlin: Transparency International. www.Transparency.De. Lui, F. T. (1986). A Dynamic Model of Corruption Deterrence. Journal of Public Economics, 31(2), 215–236. Malito, D. V. (2014). Measuring Corruption Indicators and Indices. Robert Schuman Centre for Advanced Studies Research Paper 13. Mauro, P. (1995). Corruption and Growth. Quarterly Journal of Economics, 110(3), 681–712. Olken, B. A. (2009). Corruption Perceptions vs. Corruption Reality. Journal of Public Economics, 93(7), 950–964. Razafindrakoto, M., & Roubaud, F. (2010). Are International Databases on Corruption Reliable? A Comparison of Expert Opinion Surveys and Household Surveys in Sub-Saharan Africa. World Development, 38(8), 1057–1069. Seldadyo, H., & De Haan, J. (2011). Is Corruption Really Persistent? Pacific Economic Review, 16(2), 192–206. Svensson, J. (2005). Eight Questions About Corruption. Journal of Economic Perspectives, 19(3), 19–42. Tirole, J. (1996). A Theory of Collective Reputations (With Applications to the Persistence of Corruption and to Firm Quality). Review of Economic Studies, 63(1), 1–22. Weber, A. C. (2007). How Much Do Perceptions of Corruption Really Tell Us? Economics Discussion Papers 2007–19. http://www.EconomicsEjournal.Org/Economics/Discussionpapers. Accessed 02 April 2018.

3 Causes

Examining possible cures for corruption implies knowing its causes. These causes explain why some people or countries are more corrupt than others. Based on the literature, the causes of corruption can be related to the geography or history of a country, its political and institutional systems, the characteristics of its population and its economic structure, and the corruption system itself. Some geographic and historical variables, such as natural resources’ abundance, corruption among neighbors, and colonial history, can significantly affect the prevalence of corruption. Democratic institutions limit corruption through increased competition for political mandates and possible voting out of corrupt leaders. Highly regulated economies are prone to the development of corruption because of the rent that can be secured through both the setting and implementation of rules. Population characteristics such as generalized trust, religion, and acceptance of hierarchy are often associated with a low prevalence of corruption. Finally, corruption can feed itself through its systemic nature. If a large number of officials of a department are corrupt, it becomes almost impossible for a single agent to remain honest. He/she runs the risk of facing the hostility of colleagues or even superiors by being seen as an obstacle to © The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_3

71

72     K. Sekkat

“doing business”. Furthermore, the delivery of, say, a license may need the approval of several civil servants. While one of them might speed up the process, others might slow it down unless they receive bribes.

1 Conceptual Analysis This section discusses how the various characteristics of citizens and their environment can cause corruption (Lambsdorff 2006). It considers, in turn, the features of a country (geography and history), institutions (democracy, functioning of democracy, decentralization), society and citizens (culture, values, and gender), the economy (regulatory quality, economic competition, and economic development), and the corruption system itself (systemic corruption).

1.1 Characteristics of the Country: Geography and History The geography of a country has important implications for the emergence and persistence of corrupt practices. On the one hand, countries covering a vast area might face more corruption because of the difficulty in implementing effective monitoring of public officials (potential bribe-takers), especially in isolated locations. On the other hand, employees in small areas face the threat of exposure and stigmatization, which might act as a deterrent to corruption. Furthermore, there might be a greater chance of corrupt practices being caught or exposed in areas with a high population density. Countries that are rich in certain natural resources are typical candidates for rent-seeking activities. Large endowments of valuable raw materials (fuels, minerals, and metals) offer greater potential gain to officials who allocate rights to exploit such resources (Kolstad and Søreide 2009). Those who seek such rights may devote to them large amounts of time, skill, and money that could otherwise be used in productive activities. In many resource-rich economies, skilled agents typically seek positions as oil bureaucrats or lobbyists rather than starting

3 Causes     73

a business in another field. In fact, rent-seeking consists in splitting up an existing cake instead of making it bigger. Each agent ignores the fact that a larger share for him/her means a reduced share for others, which is socially sub-optimal. Another problem is that abundant natural resources create incentives for rulers to pay off political supporters. This is a typical means used by many autocratic rulers to stay in power while continuing economic policies that are detrimental to their country (or no economic policy at all). Moreover, such public funds could be spent in more development-friendly ways. Finally, and very importantly, rent-seeking may be the subject of violent confrontations, which is destructive to fragile states. History, another important country characteristic, can have significant implications for corruption. History shapes a country’s cultural values (including religion), geography, and legal system. Cultural values make the characterization of an activity as corrupt different across countries. Bribe-giving and bribe-taking might, in some circumstances, be socially acceptable in one country and unacceptable in another country under similar circumstances. The exposure of corrupt acts depends on the likelihood of being caught, which in turn depends on the legal framework of the country. This framework is in general inherited from former colonial powers, although it may also include specific religious features and customs. Thus, countries with different colonial traditions can be expected to have different legal systems. Common wisdom is that former British colonies tend to be less corrupt. One possible explanation is that British rules respect the rule of law and procedural propriety. According to Treisman (2000), the reason for this is that common law systems (found mostly in Britain and its former colonies) developed to defend parliament and property owners against attempts by the sovereign to regulate and expropriate them. The greater protections of property against the state embodied in common law systems improve various aspects of government performance, including reducing corruption. The willingness of judges to follow procedures even when the results threaten senior figures clearly increases the chance that official corruption will be exposed. In contrast, civil law systems (found mostly in continental Europe and its former colonies), which developed more

74     K. Sekkat

as an instrument used by the sovereign to build the state and control economic life, are less conducive to such behavior.

1.2 Characteristics of Institutions: Democracy, Functioning of Democracy, and Decentralization Although corruption can take place in the private sector (see Chapter 1), it is traditionally associated with the public sector and the control that some officials have over certain decisions. Without checks, the official can abuse this control. Such checks depend on the political regime of the country and especially on the degree of political competition. In an autocratic regime, a large part of public life is under the control of the leader(s) and high-ranking officials who are subject to very few checks. Under such a regime, the tastes and wishes of a small coalition (i.e., the leaders and their companions) determine the conduct of public affairs with the aim of satisfying the coalition’s own interests. In contrast, rulers in a democracy are constrained by the ability of citizens to vote and participate in setting the rules governing their lives. The existence of competition for power between different political groups makes the position of rulers contestable. Rulers can be “dismissed” and replaced by another political group. Such a threat means that a democratic ruler will seek broader public support than an autocratic ruler will through, among other things, well-functioning institutions. Accordingly, democracy is likely to limit corruption through increased competition for political mandates. Moreover, the deeper democratic practices are rooted in a country’s tradition, the more effective the impact of political competition will be on reducing corruption. While democracy can reduce corruption, there are many cases of corrupt practices in democracies. Democratic regimes can exhibit a non-negligible degree of corruption because election campaigns require funding, which makes political parties and candidates vulnerable to pressure from funders. Moreover, institutions in charge of enforcing government accountability are often appointed or funded by the government itself. This can reduce their incentive and capacity to expose

3 Causes     75

government corruption. Other features of democracies can make some of them more corrupt than others. These include electoral rules and the design of democratic institutions such as majority or proportional systems, presidentialism, parliamentarism, federalism, and bicameralism. Decentralization is another feature of democratic systems which has generated a rich discussion in relation to corruption. Briefly defined, decentralization refers to the transfer of certain strands of central government competencies to local or regional public bodies. It can be political, fiscal, or administrative. The legal framework of decentralization does not assign it a specific task with respect to corruption. However, decentralization can lead to reduced corruption by bringing government closer to the people. It can also result in local government capture by strong local players, which increases corruption. For this reason, the change in the social and political environment that comes with decentralization is expected to affect corrupt behaviors. The literature distinguishes two types of possible effects of decentralization on corruption. One concerns the impact of decentralization on accountability and democracy. The other is based on its effect in terms of interjurisdictional competition. Decentralization brings decision-makers closer to the real situation in the territory and can lead to more informed decisions concerning taxation and expenditure. It also makes citizens better informed about local conditions. This allows them to better evaluate the performance of officials and to decide who to appoint or fire. However, decentralization creates a risk of easy capture of local government by local elites. Interjurisdictional competition refers to the risk that different local governments may compete with one another to attract investment. In theory, investors choose locations with disciplined local government officials who care about establishing market-friendly local laws and regulations and fighting bribery. However, too much competition for capital can induce negative externalities across jurisdictions. A local government may seek to attract investors away from other regions by offering them opportunities to evade central government taxes and regulations.

76     K. Sekkat

1.3 Characteristics of the Society and Citizens: Culture, Values, and Gender Sociologists are especially interested in the cultural causes of corruption. These causes are linked, inter alia, to religion, customs, and habits and encompass different features, such as acceptance that power is distributed unequally among members of a group, intolerance of uncertainty and ambiguity, the strength of ties between members of a group, membership of social networks or associations, levels of interpersonal trust, and norms of mutual aid and reciprocity. These factors also contribute to the formation of different sorts of social capital, which determine the degree to which individuals favor collective actions. A group with high social capital is likely to have effective civic institutions, to prosper and to succeed in maintaining law and order. Psychologists assert, however, that while a society’s cultural values can have impacts on corruption, individual personality traits should also be taken into account. These include extraversion, openness to experience, agreeableness, and conscientiousness. Research in psychology has explored the relationship between these factors and counterproductive work behaviors, including corruption, and has found strong relationships between the two. Moreover, involvement in corrupt acts can also be explained by rationalization tactics used by individuals committing unethical or fraudulent acts. Rationalization is a set of mental strategies that allow employees (and others around them) to view their corrupt acts as acceptable. Other studies have pointed to gender issues as a determinant of the prevalence of corruption in a country (Lambsdorff 2006). Justification is provided on two grounds. First, women as individuals are intrinsically less corrupt than men. Since this issue is beyond the scope of our analysis, we will not elaborate further on it. Note, however, that casual observations suggest that a better mix of sexes, as opposed to male dominance, is associated with lower corruption. Second, corruption might be related to the social evolution that goes with gender equality. Maledominated networks might encourage corruption when they are set up to advance specific interests at the expense of the rest of the society. Improved women’s rights might go with more transparency. This

3 Causes     77

is because parliamentary debates involve both sexes and bureaucratic decisions are communicated across sexual boundaries. Increased transparency is key to the reduction of corruption. However, the opposite direction of causality may also hold. A society with a lower prevalence of corruption may impose more restrictions on male-dominated networks and provide women with legal recourse and improved access to higher positions.

1.4 Characteristics of the Economy: Regulatory Quality, Economic Competition, and Economic Development The state intervenes in a number of fields through regulation in order to correct for market failures. These failures include abuse of dominant position, limited technological spillover, information asymmetry, and coordination among agents and other externalities. Externalities take place when firms (or individuals) do not take account of the costs or benefits their activities create for third parties. In these cases, the activities in question will be pursued either too intensely or insufficiently, which is sub-optimal from the point of view of the society as a whole. For instance, a manufacturing plant may discharge dangerous chemicals into the air or water, causing harm to the neighborhood. To address this risk, the government sets rules for emissions or imposes the use of specific technologies. While there is a consensus on the view that regulation may help protect the “public interest”, there are cases where regulation may harm such interest. These include situations of state capture. In such situations, regulation is issued in response to demand from interest groups which seek to maximize the gains for their members at the expense of the rest of the society. Regulation becomes a good whose allocation follows the laws of supply and demand, which opens the doors to corruption. The abuse of a dominant position results from weakness of competition on product markets. Theoretically, the effect of market competition on corruption is ambiguous (Diaby and Sylwester 2015). High

78     K. Sekkat

competition among firms drives firm and industry profits to zero, thereby reducing a firm’s ability to pay bribes. Less competition means firms enjoy high rents. Thus, bureaucrats with control rights over these firms (e.g., tax inspectors or regulators) will seek to get a part of the rent for themselves through corruption. Taking a dynamic perspective, however, high competition could lead to more corruption and low competition to less corruption. Under high competition, firms could use bribes to gain advantages over their competitors. Under low competition (i.e., high rents), the public could rewrite bureaucrats’ contracts and spend resources to control them more effectively. While it is well documented that corruption hampers economic development (see the next chapter), economic development is also considered to reduce corruption. With a high level of development, more transactions per unit of time occur, which increases the opportunity cost of time and encourages citizens to request faster procedures. This puts pressure on administrations to become more transparent and efficient. With a low level of development, the level of production is low and each transaction can take a lot of time. Furthermore, economic development goes with high levels of human and social capital, both of which act as deterrents to corruption. Finally, in some developing countries, bargaining is seen as an enjoyable way to use time and to distribute the gains from a transaction. Accordingly, the level of corruption may depend on the level of income rather than causing it.

1.5 Characteristics of the Corruption System: Systemic Corruption There is a growing consensus among corruption specialists that a part of the explanation for the failure of anti-corruption reforms in many countries is the systemic nature of the phenomenon. Many anti-corruption reforms are based on the principal–agent model, which takes the existence of non-corruptible principles for granted. In reality, rather than reporting and punishing corrupt behavior, political leaders, as well as citizens, seem to at least passively maintain the corrupt system (Persson et al. 2013). Corruption is thus more akin to a collective action

3 Causes     79

problem. Within a corrupt administration, it is almost impossible for a single official to remain honest. The official must not only renounce the gain from corruption but also face hostility from colleagues because he/ she diverts citizens to his/her desk or is suspected of reporting misbehavior. Furthermore, the delivery of, say, a license may involve different civil servants. While one of them might speed up the process, others might slow it down unless they receive bribes. These factors facilitate the emergence of a system of bribe-sharing that is very often organized by the top level of the hierarchy, which is supposed to control corruption. Those at the top may even exert pressure down the chain to make lower levels participate in the corrupt system. Moreover, such a system attracts a number of intermediaries who offer easier access to the bureaucracy in exchange for payment. The better this corrupt system works, the greater the incentive will be for officials and intermediaries to work together to perpetuate the system and get the best out of it. These objectives can be achieved by increasing the time and trouble imposed on citizens which, in turn, creates more opportunities for corruption. Besides inducing people to enter the corruption market, the factor described above favors the persistence of corruption, which is another reason to be concerned about the phenomenon. Observation shows that once pervasive, corruption is difficult to eliminate despite an understanding of its causes and consequences and specific attention on the part of policymakers. This causes some countries to get stuck in a bad equilibrium characterized by highly pervasive and persistent corruption, while others experience a persistently low prevalence of corruption (Seldadyo and De Haan 2011).

2 Empirical Evidence We start this section by reviewing empirical studies that have examined the contribution of each of the above-discussed causes to determining the difference in corruption score across countries. We then turn to studies dealing with specific groups of causes. The set of possible causes considered by Treisman (2000) includes whether the country has a British legal tradition, the share of

80     K. Sekkat

Protestants in the population, the share of primary product exports (including energy) in total exports, ethnolinguistic fractionalization, per capita GDP, the number of years the country has been democratic over the period 1950–1995, and whether the country has a federal or unitary system, the share of imports in GDP, an indicator of state intervention in the economy through regulations, government wages, and turnover of power between parties. The sample consisted of between 52 and 84 countries and covered the period 1996–1998. The dependent variables are the indexes of perceived corruption published by Transparency International and by Business International. To isolate causality from correlation, the analysis started by running a series of nested regressions, beginning with the most plausibly exogenous variables and progressively including groups of variables according to how slowly or quickly they are likely to change. This makes it possible to identify the additional contribution of each factor. The results show strong and robust support for the role of five causes. First, countries with Protestant traditions and those with more developed economies have lower perceived corruption. Evidence in this paper suggests that causation runs from economic development to lower corruption as well as from corruption to slower development. Second, countries with a history of British rule are less corrupt. Third, federal states are more “corrupt” than unitary ones. Fourth, while the current degree of democracy is not significant, a long period of exposure to democracy is. Finally, openness to trade seems able to reduce corruption. Pellegrini and Gerlagh (2008) can be seen as a robustness check of Treisman (2000). They address a similar question but use a larger number of countries (more than 100), a different index of corruption (WB Control of Corruption Index), and a different year of observation (2004). The set of possible causes is also large and includes the share of Protestants in the population, ethnolinguistic fractionalization, the share of fuels and minerals in exports, GDP per capita, state and local government expenses divided by central government (decentralization), the average (over the years 1994–2003) of the institutional democracy scores from the Polity IV dataset, daily newspaper circulation per ten persons, the share of imports in GDP, the percentage of veto players

3 Causes     81

that changed every year (instability), the average government wage as a percentage of GDP per capita, a dummy variable for former British colonies, a dummy variable for countries with the common law system in their commercial code, and government intervention in the economy. The last of these is a composite index based on government consumption as a percentage of GDP, government ownership of businesses and industries, the share of government revenues from state-owned enterprises, and economic output produced by the government. The analysis does not show a consistent effect from the common law system, a British colonial past, decentralization and ethnolinguistic fractionalization. In contrast, the presence of Protestants in the population is found to be associated with lower corruption. Richer countries are also found to be less corrupt. A long exposure (30 years) to uninterrupted democracy is associated with lower corruption. Finally, political instability tends to raise corruption, while the diffusion of newspapers is associated with lower corruption levels. Lippitt (2013) offers another robustness check of Treisman (2000). The difference here is that the measure of corruption (the dependent variable) is based on violations of the US Foreign Corrupt Practices Act (FCPA). The anti-bribery provision (see Chapter 12 for more details) of the act prohibits any direct or indirect payments or gifts to foreign officials with the aim of bypassing the rules for fair transactions. The study uses data on FCPA violation frequency during the years 2000–2011, which gives a total of 324 incidents of bribery. A violation frequency is assigned to the country where the bribe payment occurred. If the payment occurred in a group of countries, a violation is assigned to each of them. Specifically, if a single entity has been prosecuted for multiple incidents of bribery occurring in different countries, then each of the incidents of bribery is counted separately and assigned to the country where it occurred. Prosecutions against individuals are not included, and actions that were dismissed are omitted. The incidence of FCPA violations in each country is explained in terms of the share of fuel exports in total merchandise exports, the share of manufacturing exports in merchandise exports, government expenditure as a percentage of GDP, agricultural raw material exports as a percentage of merchandise exports, indicators of the ease of doing business,

82     K. Sekkat

initial (i.e. in 2000) democracy level, initial corruption score, initial GDP per capita, initial level of US investment, corruption growth over the observation period, and the number of years a country was in the WTO. The results show a positive and significant relationship between fuel exports and the number of FCPA violations, which is consistent with the “natural resource curse” literature. Government size is negatively and significantly correlated with FCPA violation frequency. The finding that countries with larger governments are less likely to be subject to FCPA enforcement has two possible explanations. Larger governments increase accountability, which deters corruption. Many developing countries tend to have ill-functioning institutions and smaller governments, while developed countries, particularly in Europe, have large governments and well-functioning institutions. Countries with larger manufacturing sectors tend to have more FCPA violations. Finally, the initial level of corruption is positively correlated with FCPA violations. This, unsurprisingly, means that an established “tradition” of corruption increases the number of FCPA violation and enforcement actions. The variables for agricultural exports, business regulatory environment, initial democracy level, initial level of GDP per capita, initial level of US foreign direct investment, and WTO membership were not significant.

2.1 Characteristics of the Country: Geography and History The analysis by Becker et al. (2016) focuses on the relationship between history and corruption. The authors offer an interesting and reliable approach using information going back to the Habsburg Empire. In contrast to other empires in Eastern Europe, the Habsburg Empire is historically known as multiethnic with a relatively well-functioning, honest, and respected bureaucracy. Such well-respected administration increased citizens’ trust in local public services and reduced corruption. The authors argued that these cultural norms still explain the functioning of communities today.

3 Causes     83

To test the validity of their expectation, the authors used the 2006 Life in Transition Survey (LiTS), which gives measures of trust and corruption in 17 Eastern European countries. Drawing on a number of historical sources, they coded the location of each observation in the LiTS dataset according to its former affiliation with the Habsburg Empire. They therefore worked out a specification which made it possible to compare individuals living in locations that used to be territory of the Habsburg Empire with those who did not. Both groups of populations live in locations within 200 kilometers of the long-gone Habsburg border. The border cuts straight through five countries today: Montenegro, Poland, Romania, Serbia, and Ukraine. The analysis suggests that the Habsburg Empire still exerts effects on cultural norms and on interactions of citizens with their state institutions today. Comparing individuals living on different sides of the border, those who live on former Habsburg territory have higher levels of trust in courts and police. They are also less likely to pay bribes for these local public services. It seems that the institutional heritage influences not only preferences and unilateral decisions but also bilateral bargaining situations in citizen–state interactions. As explained in Sect. 1 of this chapter, among the geographic characteristics of a country natural resource endowments have important implications for corruption. Montinola and Jackman (2002) focus on this aspect, considering two indicators of corruption: Business International covering the period 1980–1983 and 66 countries, and Transparency International covering the period 1988–1992 and 51 countries. The indicators are explained in terms of a dummy variable that equals 1 for OPEC states and 0 otherwise. Control variables include freedom of group opposition, political rights, effectiveness of the legislative body, public sector size (government share of GDP), real GDP per capita, and region dummies. The results show that OPEC membership is systematically and positively associated with corruption. Government size is negatively associated with corruption, although the effect declines with increasing government size. Finally, economic development also reduces corruption, but the magnitude of this effect declines with increasing per capita GDP.

84     K. Sekkat

Goel and Nelson (2010), using data for about 100 countries over three time periods (1995–1997, 1998–2000, and 2001–2003), examine the influence of both history and geography on corruption. The dependent variable is the TI corruption index. It is explained in terms of five measures of a country’s geography (land area, degree of urbanization, extent of digital networking, natural resource endowments, and first-order administrative divisions) and two history dummies to distinguish between newly independent or transition nations and nations that have been independent for a substantial period of time (a first dummy which equals 1 if the country became independent after 1950 and 0 otherwise, and a second dummy which equals 1 if the country became independent before 1900 and 0 otherwise). Control variables include indicators of government intervention. The latter consist of general government consumption as a percentage of GDP and a synthetic index for other government interventions in the economy (top marginal tax rates for individuals and corporations, monetary policy, the degree of regulation on foreign investment, wage, and price controls). The findings are that countries with higher degrees of urbanization have lower corruption. Corrupt practices might be easier to detect and stigmatize in areas with high population density, which might act as a deterrent to temptation. The other geographic indicators do not significantly impact the level of corruption. Both history dummies are negatively associated with corruption, which is surprising. Both newly independent countries and very old independent countries (independent before 1900) are more corrupt than the rest. According to the authors, a possible explanation is that young countries may have greater corruption due to relatively underdeveloped institutions. As countries become older, institutions tend to develop well and corruption tends to decrease, but government efforts to combat corruption might be countered over time. Arezki and Brückner (2011) rely on a more precise measure of oil rent to investigate the relationship between resource endowment and corruption. Using a panel of 30 oil-exporting countries over the period 1992–2005, they estimate a model which explains a country’s yearly change in corruption score in terms of the yearly change in the country’s oil rents and different fixed effects. The measure of corruption is

3 Causes     85

based on the ICRG index, and the proxy for oil rents is the oil export unit value. Control variables are the first difference in non-oil GDP, the first difference in oil production and lagged corruption. The authors find that an increase in oil rents significantly increases corruption. The point estimate of the coefficient of oil rents is 0.460 and is statistically significant at the 5% level. This point estimate implies that a one standard deviation increase in the unit export value of oil increases corruption by about 0.32 standard deviations. Vicente (2010) uses an event to study the relationship between natural resource endowment and corruption. The event is the oil discovery announcements in Sao Tome and Principe (STP) which occurred in the late nineties. A comparison is made between STP and Cape Verde (CV). Crucially, oil has never been found in CV and is said to be very unlikely to exist there or to have viable exploration in this territory. CV and STP are remarkably similar in different respects. STP is the second smallest country in sub-Saharan Africa with 155,000 inhabitants in 2006 and is composed of two main islands. CV has 519,000 residents (2006) and comprises nine inhabited islands. Both countries were Portuguese colonies and gained independence in the context of the last wave of African decolonization in 1975. Accordingly, the identification strategy in this study is based on a comparison of STP and CV before and after the discovery announcements. If natural resources endowment increases corruption, corruption should have increased in STP relative to CV after the announcements. The data used for the study come from household surveys conducted by a team recruited and trained by the author. In STP, the survey was submitted to 841 households in different areas between April and May 2004. In CV, the survey was submitted to 1066 households also in different areas of the country but during the period between December 2005 and February 2006. In each country, various questions were asked about corruption in different fields, such as vote buying, courts, customs, allocation of scholarships for higher education abroad, health care, police, and the allocation of state jobs. The questions pertained to the situation at different points in time to reflect the change before and after the oil discovery. The questionnaire also included a number of demographic questions.

86     K. Sekkat

The regression explains corruption in terms of a country dummy (STP and CV), a time dummy (pre-oil with value 0 and post-oil with value 1), and an interaction term between the two dummies. The coefficient of the interaction term reveals whether the difference in corruption between STP and CV changed after the oil discovery. Control variables reflect various demographic and attitudinal variables. The results confirm that corruption increased in STP after the oil discovery announcements of the late nineties. This was most visible in vote buying (a direct mechanism for holding onto political power), education (for holding future “elite” status), and customs (a channel that facilitates the consumption of imported goods, potentially directly funded by the new resources). Other services and allocations seemed to witness less clear increases in corruption. So far, in line with the majority of research on the causes of corruption, we have focused on intrinsic country characteristics, disregarding the possibility of contagion from neighboring countries. However, corruption may spread across national borders for different reasons. First, increasing cross-border business activities may allow corrupt behaviors to spread due to learning and peer-group effects. Secondly, corruption could also propagate because of increased cross-border activity by organized criminal groups. Third, countries in the same region tend to show similar levels of corruption because they have similar individual characteristics or similar institutional environments. Finally, indirect spillovers could also occur because attendance of the same regional universities or the same conferences may configure similar attitudes to corruption (Attila 2008). Márquez et al. (2011) examine whether corruption spreads across neighboring countries using spatial econometric techniques and a cross section of 171 economies. The average value of the CPI over the 2000s is used as a dependent variable. The main explanatory variable is a weighted average of corruption (based on the distance between countries) in neighboring countries. Control variables include average GDP growth, fuel exporters, urban population (as a percentage of total population), legal origin, and the WB governance indicators for political stability and voice and accountability.

3 Causes     87

The results show that the level of corruption in a country is not influenced by corruption of its neighbors. No significant spillovers or contagion effects are found once the relevant determinants of corruption are controlled for. Corruption is, therefore, not contagious. Similar individual characteristics or similar institutional environments may explain similar levels of corruption in neighboring countries, but corruption does not spread across them. The rest of the results suggest that corruption is explained by several economic and institutional variables, such as the level of economic development, the degree of urbanization, the legal origin of the country, and some governance variables. The level of GDP per capita and the degree of urban population are positively related to the absence of corruption, whereas dependence on natural resources and socialist and French legal origins seems to be associated with higher levels of corruption. Finally, it is found that political stability deters corruption. Correa et al. (2016) argue that the finding that corruption does not spread across neighboring countries is due to the nature of the econometric exercise, that is, cross section estimation, used by Márquez et al. (2011). Such econometric exercise does not make it possible to adequately control for country-specific heterogeneity relating to cultural, geographic, and historical peculiarities. The authors therefore used a panel data estimation approach covering 123 countries over the period 1995–2012, which made it possible to control for heterogeneity. The Freedom from Corruption Index provided by the Heritage Foundation is used as the dependent variable. The main explanatory variable is corruption levels in neighboring countries. It is constructed as the average corruption score of all neighboring countries weighted by the length of the common land border. Control variables are GDP per capita, government size, urbanization, press freedom, imports as a fraction of GDP, the duration of primary education, and country fixed effects, which aim to control for the time-invariant peculiarities of each country. Time fixed effects are also included. The analysis provides evidence for substantial spillover effects of corruption across countries. A one standard deviation increase in the Freedom from Corruption Index in neighboring countries increases the domestic index by 3.13 points on a scale of 0 to 100. This is equivalent

88     K. Sekkat

to the difference in corruption levels between El Salvador (36) and Italy (39) in 2012. This implies that particularly corrupt neighbors may make it more difficult for a country to get rid of its own corrupt tendencies. These findings can also have a more optimistic interpretation. It is possible that the fight against corruption within one country may spill over into neighboring countries and produce positive externalities. However, the identified relationship between neighborhood corruption and own corruption emerges only for countries exhibiting a GDP per capita above US$1230 in constant 2005 US$, and the magnitude increases for richer economies. Accordingly, the benefits might mainly hold for richer countries with well-developed institutions.

2.2 Characteristics of Institutions: Democracy, Functioning of Democracy, and Decentralization The empirical evidence pertaining to this section is extensively discussed in Chapters 5–7. To avoid too much overlap, while still giving a flavor of the most important empirical findings, the presentation here is limited to few studies. Regarding democracy, Montinola and Jackman (2002) examine the impact of competition in the political arenas on corruption. Two indicators of corruption are considered: Business International for the period 1980–1983 and 66 countries, and Transparency International for the period 1988–1992 and 51 countries. These indicators are explained in terms of political competition, which is assessed using the average of three subjective indicators and an indicator of voter turnout. The subjective indicators reflect (i) freedom of group opposition, (ii) political rights, and (iii) effectiveness of the legislative body. Other variables which enter the regression are the share of the public sector in GDP, a dummy variable that equals 1 for OPEC states, real GDP per capita, and region dummies. The authors find that political competition matters but that there is a threshold in this relationship. Below the threshold, corruption is a little higher in countries with a higher level of political competition than in those with a lower level of political competition. Above the threshold,

3 Causes     89

higher levels of political competition are associated with considerably less corruption. In other words, corruption is likely to be slightly lower in dictatorships than in countries that have partially democratized. However, with more complete democratization (reflected in the nature of elections and the effective power of elected legislators), countries experience much lower levels of corruption. This result suggests that where political competition is limited, such as in Russia, Nepal, or many Latin American countries, substantial corruption is likely even with relatively free and fair elections. An important mechanism by which democratic regimes succeed in curbing corruption is by voting out a corrupt incumbent party. Krause and Méndez (2009) examine the extent to which such mechanism holds. This is formalized through the probability that voters will retract their support for political candidates who they think are corrupt. Changes in the degree of corruption are based on a comparison of the CPI between the current and the last election. The dependent variable is the change in the share of votes received by the incumbent party as compared with the previous election, and the main independent variable is the perceived change in the degree of corruption. Control variables include a measure of absolute government support, a measure of the length of the incumbent’s tenure (in years), the ideological classification of the incumbent, the number of parliamentary seats controlled by the incumbent party, and the degree of fractionalization in the government. The sample includes 35 countries and covers the years between 1995 and 2007, which gives 107 elections. There are 14 European countries, three North American countries, nine Central and South American countries, six African countries, and three Asian countries. In the sample, 19 countries have a parliamentary system and 16 have a presidential system. The results suggest that corruption in public office is effectively punished by voters. Furthermore, the findings support the idea that both the political system and the length of democratic experience are important determinants of voter reactions to corruption. While voters in countries with parliamentary systems or with relatively low levels of democracy react negatively to an increase in corruption, no perceptible effect of this kind is found in countries with mature democracies, and

90     K. Sekkat

the evidence is inconclusive in the case of countries with presidential systems. Consistent with the studies discussed above, the relationship between democracy and corruption depends on a number of factors, including some features of democracy. Pellegata (2013) explores these factors, using the 2000 WB Control of Corruption Index as the dependent variable. A cross-sectional analysis is conducted on a sample of more than 100 countries. The measure of the current level of democracy is the Polity IV index, and the features of democracy are the age of democracy and total democratic years. The age of democracy is the number of consecutive years a country had remained democratic as of the year 2000. Total democratic years are the sum of the overall number of years a country was democratic, not including interruptions, between 1800 (or the year of independence) and 2000. Control variables include GDP per capita in 2000, imports as a percentage of GDP in 2000, the degree of state intervention in the market as measured by the Fraser Institute’s Index of Economic Freedom, and three dummy variables for, respectively, federal countries, Protestant countries, and former British colonies. Finally, the author allows the relationship between corruption and democracy to be nonlinear (quadratic and cubic). The investigation confirms that the level of democracy affects corruption in a nonlinear way. The two variables are significantly related in a quadratic manner. Neither the linear nor the cubic relationship is statistically significant. The signs of the coefficients associated with democracy and its square indicate that there is a U-shaped relationship between the current level of democracy and corruption control. In other words, an increase in the level of democracy starts having a significant and positive impact on corruption control only after a certain threshold. Regimes with “intermediate” levels of democracy tend not only to be more corrupt than consolidated democracies but also more corrupt than closed dictatorships. The analysis also confirms that the age of democracy and total democratic years are both significantly and positively related to corruption control.

3 Causes     91

2.3 Characteristics of the Society and Citizens: Culture, Values, and Gender Miller (2006) focuses on the relationship between a society’s values and corruption by comparing the values of Czechs, Slovaks, Bulgarians, and Ukrainians toward corrupt practices. The study was based on over 7300 interviews: 6000 with members of the public and 1300 with officials in the concerned countries. Interviews covered five categories of public services: health, education, welfare, police, and a mix of others (court, passport, and customs officials). Concerning values, interviewees were asked whether the use of money, presents, favors, or contacts to influence officials was (i) bad for the country and for those involved, (ii) bad for the country but unavoidable for people, or (iii) preferable because when you need a favor from an official you can get it. The results show that a majority of citizens consider such usage to be bad (between 58 and 69%). Between 25 and 34% consider it bad but unavoidable, while 7–12% see it as preferable. Czech citizens appear to be the most “ethical”, while the three other countries exhibit similar values. Although the figures are different, a similar response pattern is found for officials To examine whether the pictures that emerged from the questions about values differed from actual practices, another set of questions was posed. Citizens were asked: “If you had an important problem and an official asked you directly for money to solve it, would you (i) pay if you could afford it or (ii) refuse to pay even if you could afford it?” Officials were asked whether they would accept either a “small gift” or a “big gift” if it were offered by a client “for solving their problem”. The questions about practices show that, except in the Czech Republic, more citizens would pay than refuse (only a third). Regarding officials, on average, 47% said they would accept “a small gift” and 17% said that they would accept “big gift”. Across the five broad categories of officials, police staff were among the most reluctant to accept small gifts but not particularly reluctant to take big gifts. Health employees were

92     K. Sekkat

particularly willing to take both large and small gifts. Again, the Czech Republic appeared to be the most “ethical”. Overall, both citizens and officials explicitly condemn the use of bribes, but many confess to giving or taking them. This is not necessarily because their values are irrelevant but possibly because they have to contend with external pressures. Citizens respond to extortion, and officials respond to temptation. It seems that external pressures have more impact than internal values. Seleim and Bontis (2009) investigate the relationship between cultural values, practices, and corruption in a larger set of countries than the preceding paper. They use the Global Leadership and Organizational Behavior Effectiveness (GLOBE) dataset which is produced by the University of Pennsylvania and covered 62 countries in 2004. It is based on the responses of 17,300 middle managers from 951 organizations in the food-processing, financial services, and telecommunications services industries. The dataset covers nine dimensions of cultural values and practices: (i) uncertainty avoidance, (ii) orientation toward the future, (iii) institutional collectivism (individuals are encouraged to integrate into groups within organizations and society), (iv) humanist orientation, (v) orientation toward performance, (vi) individual collectivism (the strength of ties within small groups), (vii) attitude toward power, authority, and status, (viii) gender egalitarianism, and (ix) assertiveness. Together with GDP per capita and a human development index, these nine dimensions are used to explain the CPI. The findings differ across the various dimensions and, more interestingly, between values and practices. The tests suggest that seven out of nine dimensions of practices are associated with corruption. In contrast, cultural value dimensions are not associated with corruption with the exception of uncertainty avoidance values, which appear to significantly increase levels of corruption. These results highlight the importance of distinguishing between values and practices in understanding corruption. Uncertainty avoidance practices reduce the level of corruption, meaning that in societies that are characterized by high levels of uncertainty avoidance practices, people are more likely to be respectful of formal rules. Higher levels of future orientation practices reduce the level of

3 Causes     93

corruption. This might result from the association between rewards and future orientation behaviors in terms of long-term success. Institutional collectivism practices are associated with low levels of corruption, suggesting that high integration into groups or society reinforces ethical standards and anti-corruption behaviors. High levels of performance orientation practices are associated with lower levels of corruption, indicating that achievement-oriented cultural practices and performance excellence reduce corruption. One possible weakness in the above studies is that they use data pertaining to the values and practices in the country of residence of the respondent instead of the values and practices in the respondent’s country of origin. For a number of foreign residents, it is likely that the values of the country of origin rather than those of the country of residence will influence attitudes toward corruption. To address this issue, Barr and Serra (2010) focus on one country of residence, the UK, and different countries of origin.1 They use an experimental method to investigate whether cultural values at origin affect attitudes toward corruption. Specifically, they ask whether individuals who grow up in societies in which corruption is prevalent are more likely to be corrupt than those who grow up in societies where corruption is rare. The analysis is based on an experiment using two specially designed bribery games conducted in 2005 and 2007. The participants were Oxford University students originating from 40 different countries including a mix of high- and low-corruption countries. The game made it possible to create three dependent variables: whether the individual has offered a bribe, accepted a bribe, or engaged in bribery, irrespective of his/her role. These variables are explained in terms of the CPI of the country of origin of the individual and dummies for age, gender, whether the person is a graduate student or not, and an interaction term between the CPI and the graduate dummy.

1Countries

represented in the sample: Argentina, Australia, Bangladesh, Barbados, Belarus, Canada, China, Germany, Greece, Hong Kong, India, Italy, Kazakhstan, Malawi, Malaysia, Mauritius, New Zealand, Norway, Peru, The Philippines, Poland, Portugal, Romania, Russia, Singapore, Slovenia, South Africa, South Korea, Sweden, Switzerland, UK, Ukraine, USA, and Zimbabwe.

94     K. Sekkat

The results confirm that cultural values internalized during childhood play a determining role in individuals’ decisions about bribery later in life. However, the results are different for undergraduate and graduate students. The results for undergraduates support the hypothesis that individuals who grow up in societies in which corruption is prevalent are more likely to act corruptly than individuals who grow up in societies where corruption is rare. However, the hypothesis is not supported for graduates. The authors explored two possible explanations for the contrast between undergraduates and graduates: socialization (students adopt the host country’s values and practices after a certain time of residence) and self-selection (only “honest” students continue to the graduate level). The results support the socialization explanation. Foreign students assimilate more of the culture of their host country the more time they spend there. Given the importance of religion in shaping cultural values in many societies, some research has focused on the relationship between religion and corruption. Before proceeding, it should be kept in mind that the term religion is somewhat misused in almost all studies. As pointed out by North et al. (2013), in the absence of a thorough comparative theological investigation of the major world religions and their various views on economic, moral, and cooperative behavior, no one is in a position to assert which religions are more or less conducive to corruption. It is advisable to consider the following studies as an exploration of whether the difference in the share of the population professing a given religion is related to the prevalence of corruption. Mensah (2014) seeks to disentangle the effects of religion and other cultural values on corruption using a sample of 62 countries over the period 2000–2010. The cultural variables are drawn from the GLOBE database presented above and are used to identify the following 12 population groups: Anglo-Saxon, Nordic European, German European, Latin European, Eastern European, Latin American, sub-Saharan African, Middle Eastern, Confucian Asia, Southeast Asian, Caribbean, and Pacific Islander. The data sources for religion are the Pew Foundation, the CIA World Factbook, Wikipedia.org, and country-specific Web sites. These sources make it possible to compute the share of the population that professes one of the following religions:

3 Causes     95

Protestant Christian, non-Protestant Christian, Islamic, Buddhist, and Hindu. Three measures of corruption are used as dependent variables. These are the Control of Corruption Index, the Corruption Perceptions Index, and the Heritage Foundation’s Freedom from Corruption Index. Each dependent variable is explained in terms of religion and cultural cluster variables. Control variables include indicators of perceived political legitimacy and overall government effectiveness as well as per capita GDP, central government spending, literacy and infant mortality rates, and the percentage of raw fuels and minerals in exports. Both cultural and religious differences appear to be related to corruption, even after controlling for other economic and political factors. In particular, Protestantism, Buddhism, and Hinduism are associated with less corruption than non-Protestant Christian, Islam, and other religion/no religion. On the cultural side, the Anglo-Saxon cultural tradition is associated with less corruption than other European groups, although the results for German and Nordic cultures depend on the specification. All the non-European cultural clusters are associated with significantly higher corruption, but, here too, the results depend on the specification. Marquette (2012) argues that the evidence for a causal relationship between religion and corruption is not convincing for three reasons. First, the methodology in many studies identifies correlation not causality. Second, the datasets used vary highly across the literature, and their aggregation at the country level does not provide any insight into individuals’ attitudes toward corruption. Third, religion may have some impact on corruption at the country level but might have very little impact on a person’s actual corrupt tendencies because corruption is so systemic that being the only “clean” one often makes little sense. The analysis is based on the findings from a collaborative research project on religion and attitudes toward corruption in India and Nigeria. The research consisted of semi-structured interviews with 240 participants in total and focused on people’s views about religion and corruption. In India, the research centered on Sikhism and Hinduism in a major city in northern Punjab (Amritsar), Chandigarh (the capital of northern Punjab), and Hyderabad (the capital of Andhra Pradesh in southern India). In Nigeria, the research centered on Islam, Christianity

96     K. Sekkat

and African Traditional Religion in Kano (Northwest), Abuja (the Federal Capital, North central), Owerri (Southeast), and Ibadan (Southwest). Interview responses confirm that religion may have some impact on attitudes toward corruption, but is likely to have very little impact on actual corrupt behavior. This is because corruption is seen as being so widespread and so built into the system that being honest often makes little sense. Respondents point to a “selective moral disengagement” which justifies their own attitudes and behavior toward corruption. Corruption appears to be a collective action problem, rather than a problem of personal values or ethics. While Marquette (2012), by working at the individual level, addresses one of the criticisms (data aggregation) of the studies of the relationship between religion and corruption, North et al. (2013) complement the analysis by considering a large collection of countries and controlling for several variables including various ethnic religions and the rule of law. They also collected information on two sufficiently distant years: 1900 and 2000. The exercise consists of estimation regressions of the Control of Corruption Index in 2004 on religion variables in 1900 and 2000, which addresses the causality criticism. In all, 11 religious groups were considered: African ethnoreligion, Asian ethnoreligion, Buddhist, Catholic, Hindu, Islam, Jewish, Orthodox, Pacific Island ethnoreligion, Protestant, Unaffiliated Christian, and Nonreligious. In the regression, the religion variables are introduced as dummies. For each country, the largest religious group’s dummy variable is coded as 1 and all other religious groups’ dummy variables as 0. To measure corruption, the Control of Corruption Index is used. Overall, the findings are that corruption levels in 2004 are lower in countries that are Protestant or have an Asian ethnoreligion. However, the effects of religions groups are very different depending on whether the religious status corresponds to 1900 or 2000. A number of statistically significant results are obtained using the religious status in 1900, but very few using the religious status in 2000. The authors explain that the results based on the religious status in the earlier year (1900) better characterize the long-standing religious heritage of a nation, because

3 Causes     97

many countries saw important changes during the twentieth century. They also found that changes in the sample’s composition in terms of countries and changes in control variables had important effects on the estimates. Thus, disagreements in the literature over the effects of religious (and other) factors on corruption may only reflect differences in sample composition rather than differences in the actual effect of the variables. We have already discussed the possibility of contagion effects from neighboring countries in terms of corruption. The evidence provided supports the existence of such spillover across countries. Dong et al. (2012) address a similar question but at the individual level. Specifically, they investigate whether the tendency of a person to be corrupt depends on the corruption level of other individuals in the society. Their analysis uses the European Values Survey (EVS) 1999/2000 and the World Values Survey (WVS). The EVS is a European-wide investigation of sociocultural and political change. The WVS is a worldwide dataset that investigates sociocultural and political change. The WVS was first carried out in 1981–1983, with subsequent surveys being carried out in 1990–1993, 1995–1997, and 1999–2001. With the EVS, the analysis is based on the responses to several questions. The first question is whether it is always justified, never justified, or somewhere in between for someone to accept a bribe in the course of their duties. The responses are scaled from 1 (always justified) to 10 (never justified). The second question asks how many compatriots, according to the respondent, accept a bribe in the course of their duties. The responses are scaled from 4 (almost all) to 1 (almost none). An ordered Probit model is used to examine whether the justifiability of corruption is influenced by the perception of the prevalence of corruption. Accordingly, the dependent variable is the justifiability of corruption (first question), and the main explanatory variable is the prevalence of corruption (second question). Control variables include education level, political interest, religion, risk attitudes, the economic situation, urbanization, and employment and marital status. The results show that the higher is the perceived corruption of other persons, the higher is the justifiability of corruption. The relevant coefficient is always statistically significant at the 1% level, and the size of

98     K. Sekkat

the effect is substantial: If perceived corruption rises by one unit, the percentage of persons reporting that corruption is never justified falls by between 3.8 and 5.1 points. Looking at the other variables, the results reveal that political interest is negatively correlated with the justifiability of corruption. A one-unit increase in the political interest scale increases the probability of stating that taking bribes is never justified by around 1.5 points. There is also a negative correlation between education, age, women, and religiosity on the one hand, and justifiability on the other hand. With the WVS, the response to a similar question to that of the European Survey is used as dependent variable, namely the justifiability of corruption. As for perceived corruption, the question is slightly different. The respondent is asked how widespread bribe-taking and corruption are in the country. The possible responses are: Almost no public official takes bribes, few public officials do, or most public officials do. The regressions contain similar control variables as before. In all the specifications, the perceived level of corruption is statistically significant. If perceived corruption decreases by one unit, the percentage of persons reporting that corruption is never justified increases by between 0.6 and 3.5 points. The control variables show a similar configuration to that of the EVS estimates. To further check the robustness of their findings, the authors depart from the cross section used in the two previous sets of estimations to use panel data. The dependent variable is also different. It is the ICRG index of corruption over the period 1986–2003. The new estimation method consists in testing whether the present level of corruption is related to its past levels, which makes it possible to examine the dynamics of conditional corruption. The implicit assumption is that past, rather than present, experiences teach bureaucrats whether cheating is more or less pervasive in the economy and therefore affect their attitudes toward corruption. Thus, the regression explains the present level of corruption in terms of its past level and several control variables, such as law and order, democratic accountability, globalization, GDP per capita, and population. The coefficient of lagged corruption is highly significant and implies that the past level of corruption has a strong positive relationship with

3 Causes     99

the present level of corruption. Law and order and democratic accountability are also statistically significant. In sum, the micro-evidence and macro-evidence in the paper coincide in revealing that an individual’s own willingness to be corrupt depends on the corruption level of other individuals in a society. Lee and Guven (2013) investigate another issue, namely the role of gender. This is in line with a considerable body of work that has emerged over the past couple of decades and that has found systematic differences in behavioral characteristics across gender (e.g., Eckel and Grossman 1998). The study investigates whether the gender dimension affects the results of the above analysis. More specifically, it focuses on the impact of risk tolerance and gender on corruption. The study uses data from the European Social Survey (ESS), which covered 26 nations and 47,537 persons during the period 2004–2006. The ESS includes, among others, questions on family, work and well-being, health, and economic morality. The second round of the survey includes three key questions about corruption. The first question asks whether interviewees were asked for a bribe in the last five years. The second asks whether they have offered a bribe themselves in the last five years. The third question is on bribe justification and asks: “How wrong is a public official asking someone for a favor or bribe in return for their services?” The dependent variables are based on the responses to the three questions. They are coded as dummies where ever asked for a bribe = 1 and never = 0, ever offered a bribe =1 and never = 0, and bribe justification = 1 if seriously wrong and = 0 otherwise. The explanatory variables of interest are risk attitude and gender. Control variables relate to personal, demographic, and lifestyle characteristics of individuals, such as age, years of schooling, marriage, immigration, ethnic minority, trust, and religiosity. The analysis shows that men are associated with a lower probability of seeing that bribery is wrong. However, this opinion is not associated with offering a bribe. It is only associated with being asked for a bribe. The predicted probability of viewing bribery as being unjustified for those with no exposure to bribery experiences is 0.7088, but this decreases to 0.5971 when a person has been asked for a bribe and offered a bribe before. For both males and females, risk takers are

100     K. Sekkat

significantly more likely to offer a bribe and to be asked for a bribe. They are also less likely to view bribery as being wrong. Dollar et al. (2001) address a similar question using cross-country data on 100 countries for the years 1985, 1990, and 1995. Their study examines the relationship between female participation in parliament and the level of perceived corruption as measured by the ICRG index. The measure of female participation in parliament comes from a survey by the Inter-Parliamentary Union. Control variables include GDP, GDP squared, Gastil’s Civil Liberties Index, population, average years of schooling, openness to trade, and ethnic fractionalization. The estimated effect of female participation in parliament is significantly positive at 1% and equals 3.53. This implies that a one standard deviation increase in female participation in parliament leads to a 20% decline in the standard deviation of corruption. Thus, the presence of female parliamentarians seems to have a significant and negative effect on corruption. Swamy et al. (2001) complement the preceding analysis by examining whether corruption is less severe not only where women hold a larger share of parliamentary seats but also where they occupy more senior positions in the government. The analysis examines both the difference in gender judgement of bribe-taking and the difference in the practice of bribe-taking. To address the first question, WVS data covering 18 countries in 1981 and 43 countries in 1990–1991 are used. To examine the second question, the CPI is used. From the WVS, responses to the following question—“How do you see accepting a bribe in the course of one’s duties?”—are used. Responses are scaled from 1 to 10, where 1 indicates that the behavior can never be justified and 10 indicates that it can always be justified. These responses are used to construct the dependent variable as a dummy taking the value 1 if the respondent says that bribery is never justified and 0 otherwise. The explanatory variable of interest is a gender dummy. Control variables include schooling, marital status, religiosity, and age. The results reveal that the percentage of women who think that bribe-taking is never justifiable is much higher than the percentage of men. The coefficient of the gender dummy is statistically significant and

3 Causes     101

has the expected sign. The marginal effect corresponding to this coefficient suggests that, all else being equal, a man’s likelihood of responding that accepting a bribe is never justified is 4.3 percentage points less than the likelihood for a woman. Moreover, the gender differential in the attitude to corruption seems to be a worldwide phenomenon. The authors wonder whether the result that women disapprove of corruption more than men is driven solely by the fact that they are less likely than men to be employed and hence have less opportunity to benefit from corruption. To test this explanation, the CPI is regressed on three measures of women’s involvement in politics. The measures are the proportion of legislators in the national parliament who are female, the proportion of ministers and high-level government bureaucrats who are women, and women’s share of the labor force. The tests control for several variables, such as GDP per capita, the average years of education completed by adults, the percentage of people in the population who are Catholic and the percentage of people in the population who are Muslim, whether the country has ever been a colony, and the existence of democratic political institutions. The estimated coefficient of women’s share of parliamentary seats is highly significant. It implies that a one standard deviation increase in women’s share of parliamentary seats is associated with a decrease in corruption of slightly more than one-fifth of a standard deviation. The estimates using the share of top ministerial or bureaucratic positions held by women also give a highly significant coefficient, with a magnitude nearly identical to that for women in parliament. When the explanatory variable of interest is women’s share of the labor force, the coefficient is also highly significant, with a magnitude very similar to those obtained with the other measures. Overall, both exercises suggest not only that women are less tolerant of bribe-taking but also that corruption is less severe where women hold a larger share of parliamentary seats and senior positions in the government bureaucracy. Sung (2003) questions the above results concerning gender on the grounds that the observed association between gender and corruption could be spurious because liberal democracies are associated with both gender equality and better governance. Hence, this relationship might

102     K. Sekkat

be a by-product of the joint association with liberal democracies. To examine whether the argument is well founded, the author tests, among other things, whether (i) liberal democracy and female participation in government are positively correlated, (ii) liberal democracy is negatively associated with corruption, and (iii) when the strengths of liberal democratic institutions are held constant, the negative relationship between female participation and corruption disappears. The corruption measure is based on the CPI of 1999. Three female participation measures are used: the proportion of women among ministerial officials, the proportion of women among sub-ministerial officials, and the proportion of women among parliamentarians. Measures of democracy are drawn from the Fraser Institute and concern the rule of law, press freedom, and the democratic character of elections in 1999. Control variables are GDP per capita, the proportion of population below the poverty line and illiteracy. Regarding the first test, it appears that female participation in government is positively correlated with liberal democracies. The bivariate coefficient of correlation with the rule of the law ranges from 0.145 for women in sub-ministerial positions to 0.515 for women in parliament. The measure of women in ministerial positions is the most consistently and strongly associated with the rule of law (0.359), press freedom (0.489), and democratic elections (0.242). The proportion of women occupying sub-ministerial positions exhibits the lowest correlation with the same indicators of liberal democracy (0.145, 0.275, and 0.149, respectively). The results of the second test show that liberal democracy is negatively associated with corruption. The three measures of a liberal democracy are all very strongly correlated with lower levels of corruption. Turning to the third test, the results strongly support the argument that the negative correlation between gender variables and corruption is essentially mediated by the degree of liberal democratization. The regression coefficient pertaining to the influence of women in ministerial positions on corruption ceases to be statistically significant after the three liberal democracy variables are introduced into the equation. The same patterns of spuriousness hold for the impact on corruption of women in sub-ministerial positions and women occupying parliament

3 Causes     103

seats. When liberal democratic institutions are controlled for, the influence of gender drops irrespective of the gender indicator. In contrast, when the level of female participation in government is controlled for, the negative relationship between liberal democracy and corruption remains significant. Among the three control variables included in the multivariate equations, GDP per capita stands out as the most influential correlate of corruption. The very powerful negative association between GDP and corruption, across the different models, suggests that high economic performance is in essence incompatible with poor public governance. The relationship between poverty and corruption is statistically significant and positive. For illiteracy, the negative association with corruption is marginal. Using a different approach, namely experimental, Alatas et al. (2009) lend support to Sung (2003)’s idea that the observed association between gender and corruption is more likely to reflect the political and cultural context than any intrinsic difference between men and women. The experiment is conducted in four countries, of which two are consistently ranked among the least corrupt in the world (Australia and Singapore) and two are consistently ranked among the most corrupt (India and Indonesia). The experiments were run at the University of Melbourne, the Delhi School of Economics, the University of Indonesia in Jakarta, and the National University of Singapore. They involved third-year undergraduate or postgraduate students and spanned several sessions, each consisting of at least 30 subjects. A total of 1326 subjects, of which 596 (45%) were men, participated in the experiments. The number of participants in Australia, India, Indonesia, and Singapore was 642, 309, 180, and 195, respectively. The experiments were based on a three-person sequential-move game and sought to examine (i) the incentive to engage in a corrupt act and (ii) the incentive to punish a corrupt act even at a cost. The first player in the game is a firm and is given the option to initiate a corrupt act by offering a bribe to a government official. The second player is the official who can either accept or reject the bribe. If the bribe is accepted, both the firm and the official are economically better off at the expense of the third player, namely the citizen. This player can respond to the corrupt

104     K. Sekkat

act by punishing both the firm and the official. While the punishment is costly to the citizen, it imposes a much larger monetary cost on the firm and the official. Information about the bribe amount, the subject’s age, gender, income, education stream, employment history, and frequency of exposure to corruption was collected. Male and female subjects participated in the three roles. The results of the game are analyzed based on t-tests for differences in the means of the behavior of participants and multivariate regression analysis, where binary Probit models are estimated for the bribe, acceptance and punishment rates, and ordinary least square models for the bribe and punishment amounts. Gender is the explanatory variable of interest, and control variables are field of study (whether it is economics) and the percentage of each Australian subject’s life that has been spent outside of Australia. The t-tests show that overall the male participants have a higher propensity to offer bribes than the female participants, but that there are no statistically significant gender differences in other behaviors. When the data are split by individual countries, the observed difference in bribe rates appears to be driven by Australia. In Australia, 91.6% of male participants offered bribes compared with 80.4% of female participants. In none of the other countries are significant gender differences observed in the propensities to offer bribes. Further, in Australia, male participants also had higher acceptance rates and lower punishment rates than female participants. The bribe was accepted 92.1% of the time when it was offered to a male participant in Australia, while it was accepted 80% of the time when it was offered to a female participant. Australian male participants in the role of the citizen chose to punish 49.2% of the time, while Australian female participants chose to punish 62.6% of the time. In India, Indonesia, and Singapore, no significant differences in the behavior of the male and female participants in the three roles were found. It seems, therefore, that the gender differences reported in the previous studies may not be as universal as stated but rather culture specific. Connelly and Ones (2008) complement the cultural explanations of corruption by focusing on individual intrinsic characteristics. Their

3 Causes     105

exercise investigates whether, besides societal values, individual intrinsic characteristics determine the difference in the degree of corruption. The authors draw on the psychology literature, which has identified five main factors which underpin the personality of an individual. These are neuroticism, extraversion, openness to experience, agreeableness, and conscientiousness. Neuroticism describes individuals’ tendencies to be depressed, anxious, emotionally erratic, and lacking self-esteem. Extraversion is composed of traits of sociability, dominance, and activity. Openness to experience describes individuals’ tendencies to be interested in learning new ideas and culture. Agreeableness describes individuals’ tendencies to be kind and polite. Conscientiousness refers to the taste for achievements, attentiveness, reliability, and neatness. Considerable research has explored the relationship between these five personality factors and counterproductive work behaviors (CWBs), a domain of behaviors including corruption. In order to disentangle the effects of individual and societal characteristics on corruption, the cultural value scores discussed above are combined with the NEO Personality Inventory (NEO-PI-R). The NEO-PI-R is one of the most commonly used measures of the five traits discussed above and has been extensively used throughout the world. The sample size differs across countries: from 112 in Hong Kong to 3730 in Germany, with an average size of 986. The CPI is regressed on individual (country’s mean across individuals’ personality profile) and societal characteristics and control variables (GDP per capita and religion). The results show that many cultural dimensions are strongly correlated with national personality means. Countries whose citizens are higher in neuroticism have cultures that are higher in uncertainty avoidance. Nations with high scores in extraversion and openness tend to be high in individualism and lower in power distance, suggesting that such cultures likely have greater accessibility to experience ideas and socializing with others. Nations high in agreeableness and conscientiousness are high in power distance. More importantly, corruption appears related to national personality. Countries that are high on conscientiousness and low on openness to experience tend to be more corrupt. In addition,

106     K. Sekkat

neuroticism tends to be a strong predictor of national corruption. However, several cultural dimensions also seem related to corruption. Countries that are collectivistic, high in power distance, and high in uncertainty avoidance tend to be more corrupt. It appears that both culture and personality are related to corruption with an adjusted-R2 of 0.88. However, cultural dimensions provide more information about corruption than national levels of personality. National personality adds a modest level to national culture to explain corruption. The increase in the adjusted-R2 from adding culture to personality is 0.29 while when personality is added to culture the increase is only 0.09. Finally, countries with less wealth and lower proportions of Protestants tend to be more corrupt. Countries with higher proportions of Catholics and Muslims tend to be slightly more corrupt. However, the relationships between corruption and the percentage of Catholics and Muslims dissipate after wealth is controlled for. Anand et al. (2004) argue, however, that the above individual characteristics should be complemented by another which facilitates participation in corrupt acts: rationalization. This is linked to an interesting result we gathered from other analyses (e.g., Dong et al. 2012), which showed that the justifiability of corruption is not exogenous but depends on a number of factors such as, for instance, the prevalence of corruption among other people. The results show that the higher the perceived corruption of other people is, the higher the justifiability of corruption becomes. This specific trait helps a person to accept being corrupt. And indeed, real life shows that corruption often involves people who are far from having the prototypical image of a criminal. They are habitually upstanding community members, givers to charity, and caring parents who, in general, condemn corruption. One possible explanation for this puzzling observation is the rationalization tactics used by individuals committing corrupt acts. Rationalization is defined as a mental strategy that enables individuals to view their corrupt acts as justified. The strategy is used to neutralize any regrets or negative feelings that result from participating in unethical acts. It includes the following six tactics:

3 Causes     107

1. Denial of responsibility: An individual engaged in corrupt behavior perceives that he/she has no other choice than to participate in such activities. 2. Denial of injury: The person is convinced that no one is harmed by the actions; hence, the actions are not really bad. 3. Denial of victim: The person counters any blame for the actions by arguing that the violated party deserved whatever happened. 4. Social weighting: The person rejects criticism and argues that others are worse than he/she is. 5. Appeal to higher loyalties: The person argues that the violation of norms is due to the attempt to realize a higher-order value. 6. Metaphor of the ledger: The person rationalizes deviant behavior because of accrued dedication (time and effort) in his/her job.

2.4 Characteristics of the Economy: Regulatory Quality, Economic Competition, and Economic Development In the papers reviewed so far, GDP per capita has often been used as a control variable. Almost all these papers find a negative correlation between this variable and corruption, meaning that the more developed a country is, the lower corruption tends to be. However, such an association does not mean that development (higher per capita GDP) causes lower corruption. To address this question, Gundlach and Paldam (2009) use instrumental variable estimation methods where the average CPI over the period 1995–2006 is explained in terms of GDP per capita in constant dollars and a number of control variables, such as dummies for French and English origins of commercial and company laws, the share of Protestants in the population, the share of Roman Catholics in the population, and the number of suicides per 100,000 inhabitants. The explanatory variable of interest is GDP per capita of 98 countries. Using its contemporary level in the regression may give biased estimates because such a level may itself depend on contemporary corruption. To avoid this risk, GDP per capita is instrumented using prehistoric

108     K. Sekkat

measures of countries’ biology (the number of domesticable big mammals and the number of domesticable wild grasses) and geography (climatic conditions, latitude, relative East–West orientation, and size of the landmass to which a country belongs). The main result using the IV method is that the long-run causality is entirely from income to corruption. Accordingly, it seems that there is a corruption transition: As countries get rich, corruption vanishes. The estimated magnitude of such long-run effect implies that the difference between the 10th percentile (6.61) and the 90th percentile (9.93) of income results in a 4.95 corruption-point difference. This is sizeable since the CPI varies between 0 and 10. The findings are also that the share of Protestants decreases corruption while the share of Catholics increases corruption. Similarly, English origin of commercial and company laws is associated with lower corruption, while French origin is associated with higher corruption. Finally, the suicide rate is significantly and positively correlated with the degree of corruption. Another characteristic of the economy that is suggested as causing corruption is the degree of product market competition. In the conceptual part, it was explained, however, that the sign of the effect is ambiguous. Low competition means that firms enjoy high rents. Bureaucrats may therefore have incentives to engage in malfeasant behavior in order to get a part of the rent. However, high rents also imply that the legislator would have strong incentives to write bureaucrats’ contracts in a way that results in high control and less corruption. Ades and Di Tella (1999) offer one of the first empirical studies to examine the relationship between rents and market structure on the one hand and corruption on the other. Measures of corruption are drawn from Business International Corporation (52 countries over the period 1980–1983) and the World Competitiveness Report (31 countries for the years 1989 and 1990). Since direct measures of rents and market structure are not readily available, three proxies are used with the 1980s sample and five proxies with the 1990s sample. The three proxies are the share of imports in GDP, fuel and mineral exports, and trade distance. The last of these is the average distance to the capitals of the world’s 20 largest exporters weighted by values of bilateral exports. The five proxies include the above three variables plus the extent to which the

3 Causes     109

market is dominated by a limited number of enterprises and the effectiveness of antitrust laws, both from the World Competitiveness Report. Control variables are real per capita GDP, schooling and Gastil’s Index of Political Rights. Irrespective of the types of regression (cross section or panel with country and time fixed effects), control variables, and measures of corruption, the findings are that countries where firms enjoy higher rents tend to have higher corruption levels. In particular, corruption is higher in countries where domestic firms are protected from foreign competition with economies dominated by a small number of firms, or where antitrust regulations are not effective in preventing anti-competitive practices. The size of the effect is large: Almost a third of the corruption gap between Italy and Austria can be explained by Italy‘s lower exposure to foreign competition. Alexeev and Song (2013) use more accurate measures of corruption and market structure to examine their relationship. The data are collected at the firm level and come from the WB’s Productivity and the Investment Climate Private Enterprise Survey. They are based on responses to a questionnaire administered to several thousand firms, mostly in developing and transitional countries, between 2001 and 2005. The empirical strategy consists in regressing the responses related to corruption on measures of competition and control variables. Corruption is measured by the percentage of annual sales paid in bribes, which is based on the answer to the following question: “To get things done, what percent of annual sales would gifts or informal payments to public officials cost a typical firm like yours?” The survey also includes questions enabling the computation of several measures of the intensity of competition, such as the number of competitors, markup over firm’s costs, the extent of customer reaction to a hypothetical price increase, national market shares, and industry concentration ratios. Controls include firm characteristics that are likely to be exogenous to corruption, as well as country and year fixed effects. The results show no evidence of a negative relationship between competition and corruption. On the contrary, a significant and positive association between the intensity of competition and corruption emerges, implying that higher competition goes with higher

110     K. Sekkat

corruption. The effect is substantial: A one standard deviation increase in the markup measure decreases the paid bribe by about 0.8 percentage points. However, this positive relationship does not always hold strongly. The result that higher competition, which goes with lower rent, is associated with higher corruption might appear odd. According to the authors, the explanation lies in the distinction between extortive (or collusive) and cost-reducing corruption. Extortive corruption seeks rent-sharing without any counterpart: no rent–no corruption. Costreducing corruption implies that something will be received in return. The question in the survey (“To get things done…”) seems to reflect cost-reducing rather than extortive corruption. Since in more competitive environments firms are more sensitive to cost reduction than firms having high market power, the result is that higher competition goes with higher corruption. Note that Diaby and Sylwester (2015) examine similar question but focus on post-communist countries. They also find that greater market competition increases the amount of bribes paid.

2.5 Characteristics of the Corruption System: Systemic Corruption Systemic corruption refers to corruption which is sustained by implicit or explicit agreements among civil servants and/or politicians. Compared with the extensive empirical literature on other causes of corruption, there are very few empirical studies focusing on corruption’s systemic nature. There are, however, many specific examples of this phenomenon. Khan (2008) reports that, under the rule of Mobutu, the extent of systemic corruption in Zaire (now Democratic Republic of Congo) was estimated in the 1970s at 60% of the annual government operating budget. Zaire under Mobutu was the example par excellence of what some authors name “kleptocracy” to differentiate it from simple corruption. Gong (2002) provides different examples from China. In Lianyungang (Jiangshu Province), a smuggling case implicated at least 35 officials in different government agencies. Each of them took a share in a smuggling operation involving more than seventy luxury cars. This led to a loss to the state of RMB22 million

3 Causes     111

in duties. More surprising is that the bureau chief of the city’s Public Security Department secured parking spaces for the cars, while customs officers provided the smugglers with official documents. In Yuanjiang (Hunan province), investigations revealed that 61% of the cotton the government procured from peasants at the state price was re-sold to the market at a much higher price. The number of government officials involved in such activities was around 63 and included the wife of the mayor, the head of the city’s Price-Control Bureau, the director and the deputy director of the City Financial Commission, and the chief manager of the Textile Corporation. Finally, in Tai An (Shandong province) almost all the leading cadres in the city government were found guilty of corruption. The case in question involved a person who bribed the city’s police chief in order to obtain a license for his smuggled cars. The police chief shared bribes with his boss, the director of the Public Security Bureau, who further bribed the deputy mayor in charge of the city’s legal affairs. The process continued, involving the party leaders of the city, the deputy party secretary, and secretary. The secretary alone received over RMB 600,000 in bribes. Ades and Di Tella (1997) report cases from South Korea, where a defense procurement program representing one-third of government spending over the 1970s and 1980s (around US$9637 million) was the subject of an investigation that ended in 1993 with a former defense minister being convicted of accepting a US$370,000 bribe for arranging contracts. A second former defense minister was also convicted of accepting a kickback on a submarine contract. During 1993, the investigations led to no fewer than 39 generals being sacked, reprimanded or thrown in jail. Persson et al. (2013) shed some light on what makes such systemic corruption work. They report the results of interviews conducted in Kenya and Uganda which confirm that the failure of anti-corruption reforms in these countries came from ignoring the systemic nature of corruption. The interviews pointed to a number of costs to acting fairly. The majority of interviewees assert that acting fairly is meaningless since this will not make any difference anyway. Bribers do not care about paying because they want to access something that cannot be obtained without corruption. Bribe-takers argue that if they do not take the bribe, it will be taken by somebody else. Further, many of the

112     K. Sekkat

interviewees believe that acting honestly is a waste of time which can bring considerable trouble. Public officials who refuse to enrich themselves are regarded as stupid. Moreover, if you have an office but do not use it to help your family, you will suffer stigmatization and social exclusion. All these factors mean that the price for acting honestly is too high. The costs of acting fairly can go far beyond the inconveniences discussed above. They can include losing one’s job or even life. John Githongo, Kenya’s former Permanent Secretary in Charge of Governance and Ethics under President Mwai Kibaki, was told that the Kenyan intelligence would “put something in [his] tea” if he revealed what he knew about the political elite’s involvement in corruption. In 2006, Githongo eventually had to flee the country after realizing that even the president and his men had turned their backs on him (Persson et al. 2013). Ogungbamila (2014) conducts a similar study to the one above but in Nigeria. The analysis is based on the responses of 536 employees (298 males and 238 females) from public organizations in six states of southwestern Nigeria. All respondents had job tenure during the period of the study (2001–2013). They had to fill in a questionnaire covering their personal data, the frequency of reporting of corrupt acts per year during the period of the study, and the factors that affected their willingness to disclose corrupt acts during the same period. Respondents had to choose one or more of the following factors affecting their willingness to report: (i) perceived inability of reporting to bring about the desired change in the behavior of wrongdoers, (ii) fear of attack from wrongdoers, (iii) fear of being ostracized, (iv) government’s insensitivity to the trouble of the citizens, (v) lack of integrity in government’s anti-corruption crusades, (vi) lack of trust in AntiCorruption Agencies, (vii) perceived inefficiency in the court process, and (viii) the stress associated with being a witness. The results indicate that 2013 witnessed the highest frequency of reporting followed by 2012. Year 2002 witnessed the lowest frequency of reporting. There were no gender differences regarding the frequency of reporting from 2001 to 2012 and regarding the effect of social and psychological factors. In contrast, there were significant differences in

3 Causes     113

the roles of the various social and psychological factors. The factors cited were, by decreasing frequency, the perceived inelegance of whistleblowing (74.2%), the feeling that corruption had no direct victims (71.2%), corrupt persons are too powerful to be prosecuted (69.7%), the fear of being ostracized (69.1%), perceived inefficiency of the court process (63.8%), reporting would not bring the desired change in behavior (63.2%) and the perceived stress associated with being a witness to corrupt acts (61.5%). Tavits (2010) further illustrates what makes systemic corruption work by investigating whether public officials and citizens are more likely to engage in corruption when they perceive that corrupt behavior is widespread among their peers. The analysis is based on two nationally representative datasets from Estonia and refers to the year 2004. One dataset covers citizens (788 persons) and the other public officials (791 persons). In the case of citizens, the question concerns direct involvement in corruption. The respondents were asked: “Have you ever paid a bribe or offered a gift to a public official in order to influence the provision of a public service?” A dummy variable taking the value 1 if the response is yes and 0 otherwise was constructed and used subsequently as a dependent variable. For public officials, the central question is of the type: “Imagine that you must decide whether or not to give financial support for certain projects. One of the applicants for financial support offers you a trip to a summer resort in case you decide in favor of his/her project. Would you decide in favor of the project?” Here, too, a dummy variable is constructed and takes the value 1 if the response is yes and 0 otherwise. Since the focus here is on how involvement in corruption depends on the perceived pervasiveness of corruption, respondents were asked to give their opinion on a scale of 1 to 4 (1 = not at all common, 4 = very common) about the following behaviors: (i) A driver offers a police officer a good or a service from his or her firm at a discount price in order to avoid a speeding ticket, (ii) An entrepreneur offers the headmaster of an elite public school a trip to a summer resort for admitting his or her son to the school, (iii) A public official uses a government provided car for private purposes, (iv) A civil servant offers, for a fee,

114     K. Sekkat

lectures in the area of his or her work-related expertise, (v) An entrepreneur calls up a public official who he or she knows from previous personal contacts and asks to fast-track the processing of his or her file, (vi) An entrepreneur offers a public official personal favors in return for a public contract, (vii) A public official buys goods on behalf of his or her institution from a company owned by his or her relative, and (viii) A patient is moved up on a waiting list for surgery because his or her brother is friend of the surgeon. The explanatory variable of interest is the perceived pervasiveness of corruption and is computed as the average of all responses. Control variables include generalized trust, trust in government, salary level and satisfaction with the workplace. The results support the idea that perceived pervasiveness of corruption is a significant predictor of the corruptibility of a public official. For a unit increase in the perceived pervasiveness of corruption score, the probabilities of agreeing with a corrupt deal increase by a factor of 3.5. Put another way, somebody who thinks that corrupt activities are very common is about ten times more likely to be corruptible than somebody who thinks that corrupt activities are not at all common. Turning to the general public, both perceived pervasiveness and acceptability of corruption have the expected and statistically significant effect on the likelihood of having paid a bribe. The effects remain significant even when controlling for extortion. When all other variables are at their average value, the predicted probability of having paid a bribe is 0.10 for a respondent who has never been asked to pay. This probability increases to 0.16 for a respondent who has been asked once, 0.23 for someone who has been asked twice, and 0.91 when the variable is at its maximum value. In sum, the empirical models demonstrate that both public officials and citizens are influenced by their perception of what is acceptable and commonplace.

3 Conclusion The causes of corruption are numerous and go from the characteristics of a country (geography, history, political and institutional systems, centralization, and regulation) to the characteristics of its population

3 Causes     115

(customs, trust, and religion). The organization of corruption itself plays an important role in perpetuating corrupt situations. Analysis of the evidence shows the following regularities. Developed economies are less corrupt. Federal states are more “corrupt” than unitary ones. Democracy and elections reduce corruption but only to some extent. Corruption in public office is effectively punished by voters, but a partisan bias makes voters more tolerant of corruption within their party. More importantly, democratic culture and history are stronger deterrents of corruption than democratic status by itself. Openness to trade seems able to reduce corruption. Political instability tends to raise corruption. Natural resources’ abundance is positively and significantly associated with the extent of corruption. Finally, the systemic nature of corruption is highly responsible for the persistence of corruption. The evidence is mixed concerning the spillover of corrupt practices across countries or individuals, as it is concerning religion, gender, and history. The level of corruption in a country is not influenced by corruption of its neighbors, and an individual propensity to corruption may or may not be higher if corruption among others is high. Religion may have some impact on attitudes toward corruption, but it has very little impact on actual corrupt behavior. Countries with Protestant traditions and those with a history of British rule are sometimes found to be less corrupt. There is no evidence of a negative relationship between market competition and corruption. Finally, some studies suggest that women are not only less tolerant of bribe-taking but also that corruption is less severe where women hold more political power. Other studies, however, support the notion that the association between gender and corruption could be spurious because liberal democracies are associated with both gender equality and better governance.

References Ades, A., & Di Tella, R. (1997). National champions and corruption: Some unpleasant interventionist arithmetic. The Economic Journal, 107(443), 1023–1042.

116     K. Sekkat

Ades, A., & Di Tella, R. (1999). Rents. Competition and Corruption. American Economic Review, 89(4), 982–993. Alatas, V., Cameron, L., Chaudhuri, A., Erkal, N., & Gangadharan, L. (2009). Gender, Culture and Corruption: Insights from an Experimental Analysis. Southern Economic Journal, 75(3), 663–680. Alexeev, M., & Song, Y. (2013). Corruption and Product Market Competition: An Empirical Investigation. Journal of Development Economics, 103(C), 154–166. Anand, V., Ashforth, B. E., & Joshi, M. (2004). Business as Usual: The Acceptance and Perpetuation of Corruption in Organizations. Academy of Management Executive, 18(2), 39–53. Arezki, R., & Brückner, M. (2011). Oil Rents, Corruption and State Stability: Evidence from Panel Data Regressions. European Economic Review, 55(7), 955–963. Attila, G., (2008). Is Corruption Contagious? An Econometric Analysis. Norwegian Institute of International Affairs (NUPI) (Working Paper 742). Department of International Economics. Barr, A., & Serra, D. (2010). Corruption and Culture: An Experimental Analysis. Journal of Public Economics, 94(11), 862–869. Becker, S. O., Boeckh, K., Hainz, C., & Woessmann, L. (2016). The Empire Is Dead, Long Live the Empire! Long-Run Persistence of Trust and Corruption in the Bureaucracy. Economic Journal, 126(590), 40–74. Connelly, B. S., & Ones, D. S. (2008). The Personality of Corruption a National-Level Analysis. Cross-Cultural Research, 42(4), 353–385. Correa, E. A., Jetter, M., & Agudelo, A. M. (2016). Corruption: Transcending Borders. Kyklos, 69(2), 183–207. Diaby, A., & Sylwester, K. (2015). Corruption and Market Competition: Evidence from Post-Communist Countries. World Development, 66, 487–499. Dollar, D., Fisman, R., & Gatti, R. (2001). Are Women Really the “Fairer” Sex? Corruption and Women in Government. Journal of Economic Behavior and Organization, 46(4), 423–429. Dong, B., Dulleck, U., & Torgler, B. (2012). Conditional Corruption. Journal of Economic Psychology, 33(3), 609–627. Eckel, C. C., & Grossman, P. J. (1998). Are Women Less Selfish Than Men? Evidence from Dictator Experiments. Economic Journal, 108(448), 726–735.

3 Causes     117

Goel, R. K., & Nelson, M. A. (2010). Causes of Corruption: History, Geography and Government. Journal of Policy Modeling, 32(4), 433–447. Gong, T. (2002). Dangerous Collusion: Corruption as a Collective Venture in Contemporary China. Communist and Post-Communist Studies, 35(1), 85–103. Gundlach, E., & Paldam, M. (2009). The Transition of Corruption: From Poverty to Honesty. Economics Letters, 103(3), 146–148. Khan, F. (2008). Understanding the Spread of Systemic Corruption in the Third World. American Review of Political Economy, 6(2), 16–39. Kolstad, I., & Søreide, T. (2009). Corruption in Natural Resource Management: Implications for Policy Makers. Resources Policy, 34(4), 214–226. Krause, S., & Méndez, F. (2009). Corruption and Elections: An Empirical Study for a Cross-Section of Countries. Economics and Politics, 21(2), 179–200. Lambsdorff, J. G. (2006). Causes and Consequences of Corruption: What Do We Know from a Cross-Section of Countries? In S. Rose-Ackerman (Ed.), International Handbook on the Economics of Corruption (pp. 3–51). Cheltenham: Edward Elgar. Lee, W. S., & Guven, C. (2013). Engaging in Corruption: The Influence of Cultural Values and Contagion Effects at the Microlevel. Journal of Economic Psychology, 39, 287–300. Lippitt, A. H. (2013). An Empirical Analysis of the Foreign Corrupt Practices Act. Virginia Law Review, 98, 1893–1930. Marquette, H. (2012). ‘Finding God’ or ‘Moral Disengagement’ in the Fight against Corruption in Developing Countries? Evidence from India and Nigeria. Public Administration and Development, 32(1), 11–26. Márquez, M. A., Salinas-Jiménez, J., & Salinas-Jiménez, M. D. M. (2011). Exploring Differences in Corruption: The Role of Neighboring Countries. Journal of Economic Policy Reform, 14(1), 11–19. Mensah, Y. M. (2014). An Analysis of the Effect of Culture and Religion on Perceived Corruption in a Global Context. Journal of Business Ethics, 121(2), 255–282. Miller, W. L. (2006). Corruption and Corruptibility. World Development, 34(2), 371–380. Montinola, G. R., & Jackman, R. W. (2002). Sources of Corruption: A CrossCountry Study. British Journal of Political Science, 32(01), 147–170.

118     K. Sekkat

North, C. M., Orman, W. H., & Gwin, C. R. (2013). Religion, Corruption and the Rule of Law. Journal of Money, Credit and Banking, 45(5), 757–779. Ogungbamila, B. (2014). Whistleblowing and Anti-corruption Crusade: Evidence from Nigeria. Canadian Social Science, 10(4), 145–154. Pellegata, A. (2013). Constraining Political Corruption: An Empirical Analysis of the Impact of Democracy. Democratization, 20(7), 1195–1218. Pellegrini, L., & Gerlagh, R. (2008). Causes of Corruption: A Survey of CrossCountry Analyses and Extended Results. Economics of Governance, 9(3), 245–263. Persson, A., Rothstein, B., & Teorell, J. (2013). Why Anticorruption Reforms Fail—Systemic Corruption as a Collective Action Problem. Governance, 26(3), 449–471. Seldadyo, H., & De Haan, J. (2011). Is Corruption Really Persistent? Pacific Economic Review, 16(2), 192–206. Seleim, A., & Bontis, N. (2009). The Relationship Between Culture and Corruption: A Cross-National Study. Journal of Intellectual Capital, 10(1), 165–184. Sung, H. E. (2003). Fairer Sex or Fairer System? Gender and Corruption Revisited. Social Forces, 82(2), 703–723. Swamy, A., Knack, S., Lee, Y., & Azfar, O. (2001). Gender and Corruption. Journal of Development Economics, 64(1), 25–55. Tavits, M. (2010). Why Do People Engage in Corruption? The Case of Estonia. Social Forces, 88(3), 1257–1279. Treisman, D. (2000). The Causes of Corruption: A Cross-National Study. Journal of Public Economics, 76(3), 399–457. Vicente, P. C. (2010). Does Oil Corrupt? Evidence from a Natural Experiment in West Africa. Journal of Development Economics, 92(1), 28–38.

4 Consequences

Besides the moral considerations, the rationale for the fight against corruption is built on its economic, social, and political effects. The positive or negative nature of these effects is controversial, at least among economists. Some authors argue that corruption may be beneficial, “greasing-the-wheels” when bureaucracy is inefficient or regulation is too burdensome. Other authors argue that while corruption can grease the wheels at the microeconomic level, it ends up “sanding the wheels” at the macroeconomic level through spillovers and externalities among economic activities. The economic effects of corruption concern growth, physical and human capital formation, productivity, infrastructure, international trade, and FDI. The non-economic effects of corruption concern the provision of health care and education services, safety and security, environment, electoral participation, and confidence in public institutions.

1 Conceptual Analysis Insofar as corruption involves only transfer payments from bribe payers to bureaucrats or politicians, it does not necessarily impose a net social cost (Ehrlich and Lui 1999). Accordingly, the costs discussed in © The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_4

119

120     K. Sekkat

this section go beyond such transfer of wealth between two parties to examine the impacts on important features of the country. As explained above, however, the literature is not unanimous about the generality of such costs. One branch of the literature suggests that corruption can be beneficial to the society when other institutional dimensions are not functioning well. Another branch refutes this idea. In this conceptual section, we will start by briefly presenting the justifications for the costs of corruption independently of the functioning of other institutional aspects. In other words, we start with the unconditional effects of corruption and distinguish between the economic and the non-economic costs. Then, we will compare the points of views concerning the “sanding-the-wheels” vs. the “greasing-the-wheels” roles of corruption, which consider the impact of corruption as conditional on the functioning of other institutional dimensions.

1.1 Unconditional Economic Effects The rationale behind the economic costs of corruption is the following: 1. First, corruption raises the costs of doing business and thus undermines the incentives for private entrepreneurship. In addition, corruption causes firms to waste time negotiating with corrupt officials, which further augments the transaction costs of doing business. Finally, the existence of corruption diverts resources from their productive allocation toward the defense of property rights. These costs may become so high that firms are obliged to go at least partly informal or renounce certain investments (Boehm and Joerges 2008). 2. Second, given its secret nature, corruption increases the risk of undertaking transactions and contracts. Since corruption is, in general, illegal, litigation around a corrupt deal cannot be brought before a court. In particular, if the investment is specific, politicians or civil servants may extort additional favors once the investment has been made by threatening to break the corrupt deal. Such insecurity may prevent transactions even though they offer, ex-ante, potential mutual gains (Méon and Weill 2010).

4 Consequences     121

3. Third, the prospect of corruption can motivate unjustified interference from politicians and interest groups. Besides the resulting increase in transaction costs and uncertainty, such interference can distort the decision-making process and the setting of rules and regulations. In fact, in many instances, public decisions are not simply determined by public interests but can be affected by campaign contributions, political advertising, and other ways of exerting political pressure (Aidt 2016). In order to illustrate more precisely the impacts of corruption on the economic sphere, we will start from the traditional decomposition of GDP growth according to growth in its supply and demand sides. The former includes productivity and physical and human capital. The latter can be split into domestic and external demands. Corruption negatively affects physical capital through reduced investment. This is the result of the increased cost of investment because firms have to take into account the costs of bribery when setting up a business and keeping it running (Boehm and Joerges 2008). This provides an incentive not only to accumulate less capital but also to use it less efficiently. Depending on which activity or location is most vulnerable to corruption, firms may limit the size of their base or misallocate investment across plants (Méon and Weill 2010). Firms will also have an incentive to accumulate generic instead of specific capital although the latter might more productive. This is because generic capital has an option value which makes it more easily reallocated to other purposes than specific capital. The prospect of corruption may also encourage the government to favor public over private enterprises, which reduces private investment although it is generally seen as more productive than investment by public enterprises (Restuccia and Rogerson 2008). As with domestic investment, corruption can be a major obstacle to foreign direct investment (FDI), which is generally considered as a major source of positive spillovers for domestic firms. The literature suggests similar negative effects of corruption on FDI as in the case of domestic investment, i.e., an impact on the return and the risk associated with investing. However, it emphasizes that the main institutional impediment to FDI is the excess risk that corruption generates.

122     K. Sekkat

Investment is not only subject to a risk of predation and holdup but also, and chiefly, to a risk of expropriation and nationalization. Just as with physical capital, corruption has a potentially important effect on human capital. As we will see below, corruption reduces GDP per capita and facilitates the misappropriation of public funds. An important consequence is that corruption is likely to create major difficulties for governments to provide basic services such as public education and health services (Dridi 2014). By reducing GDP per capita, corruption affects tax revenue and hence the volume of funds available for government spending. Moreover, corruption diverts public resources from human capital formation to less growth-enhancing activities. For instance, it will reduce the share of public spending on education and increase the share of military spending. Finally, the possibility of rents induced by the existence of corruption distorts individual decisions to invest in human capital. To benefit from rents caused by government intervention in the economy, people are likely to spend less time in education and focus more on accumulating political capital that enables them to access bureaucratic power. The existence of rents influences choices in the field of education. It may prompt students to choose certain types of studies (e.g., law) over others (e.g., engineering), although the latter may be more conducive to growth. Turning to productivity, as we discussed above, corruption diverts physical and human capital from productive to less productive sectors. In addition, corruption can also negatively affect international trade, hampering imports of certain goods (machinery or intermediate products) which are important vehicles for the diffusion of new ideas and technologies and, hence, growth. Moreover, corruption can inhibit managers’ incentives to improve firm performances. Corruption also forces firms to devote time and effort to protect their property. Another effect comes from the fact that corruption can facilitate the creation or the survival of monopolistic positions. It is extensively documented that such positions, by limiting competition, do not incite firms to improve their productive and dynamic efficiency. Productive efficiency implies that the output–input combination is brought to the optimal production frontier and firms produce at the lowest cost. Dynamic efficiency pushes firms to seek opportunities for further reducing costs, improving

4 Consequences     123

product quality, or introducing new technologies. This, in turn, fosters technological progress, which is a major source of growth. On the demand side, we will start with the domestic component. The most important part of domestic demand is private consumption, which depends directly and heavily on income. Accordingly, a first effect of corruption on domestic demand is straightforward. Corruption reduces income growth, which can be expected to directly reduce private consumption growth. Other effects are subtler and stem from the impact of corruption on fiscal policy. The presence of corruption also affects the optimal mix between consumption and income taxation (Alm and Barreto 2003). The optimal tax mix for a corrupt government relies more heavily on consumption taxes than on income taxes, while for a non-corrupt government the optimal mix relies more heavily on income taxes than on consumption taxes. Shifting the tax mix from income taxes toward consumption taxes would negatively affect private agents’ welfare and, hence, consumption. International trade through exports affects the demand for a given country’s products and services. As before, corruption increases the cost and uncertainty of transactions. However, the main impediment to international trade is the problem of contract enforcement, especially since international transactions involve traders in countries whose legal and political jurisdictions differ. The lack of contract enforcement may act as a tariff on risk-neutral traders and therefore reduce trade.

1.2 Unconditional Non-economic Effects While many studies have examined the effect of corruption on economic variables, only a few have investigated the non-economic effects, such as those related to poverty, infrastructure, trust, political participation, and regime legitimacy. The rationales behind such non-economic effects are diverse. First, corruption could make politicians give economic development of the country less priority than their own interests or those of their allies. Other areas might also find themselves lower on politicians’ priority lists, such as income inequality, access to health, enforcement of the rule of law, or environmental degradation

124     K. Sekkat

(Levin and Satarov 2000). Second, as in the case of education discussed above, corruption reduces the amount of revenue that can be collected through taxation and, hence, the resources available for public spending. Third, corruption distorts the structure of public spending in favor of costly projects, military equipment, government security services, and vote-buying. In addition, corrupt governments prefer, in general, to fund big projects (“white elephants”) because they are more visible to citizens and also because they offer more opportunities for corrupt deals (Boehm and Joerges 2008). In contrast, spending on operation and maintenance is given a lower priority, which results in a deterioration in the quality of infrastructure. Over time, some projects that were initially development-enhancing become unproductive. Fourth, corruption can undermine people’s trust in the political system, its institutions and its leaders, leading to the destruction of, or at least serious damage to, the country’s “social contract” (Uneke 2010). This contributes to the creation of an atmosphere of tension, dishonesty, weak or selective law enforcement, cynicism, and erosion of faith in the political and administrative system. More dangerously, corruption becomes commonplace and even a survival strategy (Boehm and Joerges 2008). The system can become caught in vicious circle where bad institutions favor corruption, which in turn further weakens the quality of institutions. Once trapped in such a circle, it is extremely difficult for a country to escape. The resulting social tensions and political stability may provide a pretext for military or foreign interventions and often create further handicaps for national economic development (Uneke 2010).

1.3 Conditional Effects: “Greasing” Versus “Sanding” the Wheels So far, we have discussed the potential negative impacts of corruption without explicitly taking account of other institutional features of the country. Some studies point to possible positive effects of corruption. The core of the debate lies in the combination of corruption with a low quality of other institutional features of the country. We refer to this controversy as the “grease-the-wheels” versus the “sand-the-wheels” debate.

4 Consequences     125

Staring with the “grease-the-wheels” hypothesis, a poorly functioning bureaucracy is considered to be the most prominent inefficiency that corruption can solve. Huntington (1968) states: “In terms of economic growth, the only thing worse than a society with a rigid, overcentralized, dishonest bureaucracy is one with a rigid, overcentralized, honest bureaucracy”. There are various aspects of an ill-functioning bureaucracy that can be compensated by corruption. One example is slowness. Using a formal economic model, Lui (1986) shows that corruption can efficiently lessen the time spent in queues. The reason is that bribes give bureaucrats an incentive to speed up the process, in an otherwise sluggish administration (see also Leys 1965). Furthermore, Huntington (1968) argues that corruption can help surmount tedious bureaucratic regulations and foster growth. According to him, such a phenomenon was observed in the 1870s and 1880s in the USA, where corruption by railroad, utility, and industrial corporations resulted in faster growth. Another consequence of an ill-functioning bureaucracy concerns the quality of civil servants. Leys (1965) and Bailey (1966) argue that corruption can amend a bureaucracy by improving the quality of its civil servants. If wages in government service are insufficient, the existence of perks may constitute a complement that may attract able civil servants who would have otherwise opted for another line of business. Finally, Beck and Maher (1986) and Lien (1986) suggest that corruption may encourage good decisions by officials. If bureaucrats do not have enough information or are not competent to take certain decisions, corruption can replicate the outcome of a competitive auction. The authors formally show that when attributing government procurement contracts, the ranking of bribes can replicate the ranking of firms by efficiency. Moreover, if some investment projects are dependent on the attribution of a license, corruption may be an efficient way of selecting such projects. Here again, corruption in the attribution of a government license is very similar to a competitive auction. The intuition (Leff 1964) is that licenses tend to be allocated to the more generous bribers, who may be more efficient. The capacity to offer a bribe is thus correlated with talent. Another inefficiency put forward as driving the “grease-the-wheels” hypothesis concerns regulation and policies. Bailey (1966), for instance,

126     K. Sekkat

argues that if bribes can help private agents to evade a public policy designed to solve a particular problem, they may thereby allow them to find an overlooked and better-suited solution. This may, in turn, allow an improvement in the policy’s outcome even in terms of the government’s objectives. Leff (1964) and Bailey (1966) also argue that graft may simply be a hedge against bad public policies. This is particularly true if institutions are biased against entrepreneurship, due for instance to an ideological bias. By simply impeding inefficient regulations, corruption may then limit their adverse effects. It may also result in an alteration of the policy in a way that is friendlier to growth. It has also been argued that graft may in some circumstances improve the quality of investments. This is the case (Leff 1964) when government spending is inefficient. If corruption is a means of tax evasion, it can reduce public tax revenue. Provided the bribers invest efficiently, the overall efficiency of investment will be improved. In addition to the quality of investments, some authors argue that corruption may also raise the level of investment. For instance, Leff (1964) asserts that corruption may constitute a hedge against other risks originating from the political system, such as expropriation or violence. If corruption helps to mitigate those risks, investment will turn out to be less risky and may accordingly increase. All the above-mentioned arguments share the presumption that corruption may positively contribute to growth and development because it compensates for the consequences of a defective bureaucracy and bad policies. One may nevertheless wonder whether corruption creates or reinforces other inefficiencies and whether bribers are always taking more efficient decisions than the public authorities. Although bribery may have benefits in a weak institutional environment, it may also impose additional costs in the same circumstances. The existence of such costs provides a rationale for the “sand-the-wheels” hypothesis. To discuss the “sand-the-wheels” hypothesis, we again start with an ill-functioning bureaucracy. The positive impact of corruption on slowness rests on the assumption that a civil servant can speed up an “exogenously” slow process. However, corrupt civil servants may cause delays that would not appear otherwise, just to get the opportunity to extract a bribe (Myrdal 1968). Moreover, the ability of civil servants to speed up

4 Consequences     127

the process can be very limited when the administration is made up of a succession of decision centers. In this case, civil servants at each stage can have some form of veto power or some capacity to slow down a project. Using industrial organization models, Shleifer and Vishny (1993) show that the cost of corruption can be very high when, say, to get an authorization for a project, many independent agents are involved rather than only one. Bardhan (1997) reports an Indian high official as declaring that while he could not be confident of moving a file faster, he could immediately stop it. The increased number of transactions due to graft may well offset the increased efficiency with which transactions are carried out (Jain 2001). Under these circumstances, one distortion is added to the others instead of compensating for them, which is precisely the meaning of the “sand-the-wheels” hypothesis. At an aggregate level, the impact of corruption on the quality of civil servants is questionable. Kurer (1993) argues that corrupt officials have an incentive to create other distortions in the economy to preserve their illegal source of income. For instance, a civil servant may have an incentive to ration the provision of a public service just to be able to decide to whom to allocate that service in exchange for a bribe. Similarly, a civil servant also has an incentive to limit the access of new civil servants (especially competent ones) to key positions in order to preserve the rent from corruption. While individual bribers can indeed improve their own situation through perks, nothing is gained from corruption at the aggregate level. The argument that corruption may encourage good decisions is also subject to doubt. There are reasons to believe that agents paying the highest bribes are not always able to improve efficiency. Rose-Ackerman (1997) argues that a firm may be able to pay the highest bribe simply because it compromises on the quality of the goods it will produce if it gets a license. Mankiw and Whinston (1986) show that entry on a market may be beneficial for a firm but detrimental to welfare. In these cases, entry is, in general, subject to an authorization. Although entry is detrimental to welfare, the firm may find it profitable to pay the bribe to get the authorization and enter the market. Finally, if the profitability of a license is uncertain, the winner of the auction may be the more optimistic party rather than the most efficient, a situation known as the

128     K. Sekkat

“winner’s curse”. In these cases, corruption is not the best way to award a license. Thus, even if the analogy between corruption and a competitive auction holds, there are situations where the winner does not enhance efficiency. Turning to the second category of institutional deficiencies, i.e., policy options by public authorities, the argument in favor of corruption can be counterbalanced in various respects. The argument according to which corruption may raise both the quantity and the quality of investment is questionable. There is evidence that this may not be true for public investment. Empirical evidence shows that higher corruption is associated with higher public investment (Tanzi and Davoodi 1998) and that this results in a diversion of public spending toward less efficient allocations (Mauro 1998). In other words, corruption results in a greater amount of public investment in unproductive sectors, which is unlikely to improve efficiency and result in faster growth. There are also grounds to doubt that corruption acts as a hedge against risk in a politically uncertain environment. This may be true only if corruption does not imply additional risk-taking. However, corruption is not a simple transaction. As it is illegal, the commitment to comply with the terms of the agreement may be very weak, which may lead to opportunism, especially on the part of the person being bribed. As Bardhan (1997) points out, the inherent uncertainty of corrupt agreements may simply make the efficiency-enhancing mechanisms ineffective. This presumption is supported by the results obtained by Campos et al. (1999) and Lambsdorff (2003b), who observe that the unpredictability of corruption has an impact on investment and capital inflows that is independent of the impact of the level of corruption. As a result, it is likely that corruption may increase the risks associated with a weak rule of law instead of compensating for them.

2 Evidence: Economic Effects The conceptual analysis has shown that the core of the “greasing-thewheels” vs. the “sanding-the-wheels” debate is not whether corruption always induces economic inefficiency. Instead, the concern is whether

4 Consequences     129

corruption increases or decreases efficiency when the quality of governance is low. In econometric terms, testing the “greasing-the-wheels” vs. the “sanding-the-wheels” hypotheses entails testing how the quality of governance affects the impact (coefficient) of corruption on efficiency. Accordingly, the usual set of explanatory variables in the growth regression, for instance, is complemented by a corruption index, a quality-of-governance index, and an interaction term defined as the corruption index multiplied by the quality-of-governance index. One strand of the empirical literature has focused on the average impact of corruption without making such effect conditional on the quality of governance, while another strand explicitly examines such conditional effect. In what follows, we will discuss each in turn.

2.1 Unconditional Effects The empirical literature has examined the average impact of corruption on different economic variables including growth, physical and human capital, productivity, international trade, and FDI. Mauro (1995) was one of the first papers to investigate the issue by focusing on growth and investment. The exercise consisted in regressing average per capita GDP growth over the period 1960–1985 on the corruption index and control variables. A similar regression is performed for the average ratio of investment over GDP between 1980 and 1985. The rationale for studying both growth and investment is to see whether corruption has only a direct effect on growth or both a direct and an indirect effect through investment. The corruption data are those collected by Business International (BI) for the period 1980–1983 and around 50 countries. Control variables include measures of institutional efficiency, political stability, bureaucratic efficiency, ethnolinguistic fractionalization, per capita GDP in 1960, education in 1960, and population growth. The results show that corruption lowers growth independently of its effect on investment (direct effect). It also depresses private investment, thereby reducing economic growth (indirect effect). The negative associations between corruption and investment and between corruption and

130     K. Sekkat

growth are significant, both in statistical and in economic terms. For instance, if Bangladesh were to reduce its corruption to Uruguay’s level, its investment ratio would rise by almost five percentage points and its yearly GDP growth rate would rise by over half a percentage point. Ali and Isse (2002) lend further support to the negative impact of corruption on economic growth. They address possible bias in the preceding estimations, namely the existence of a reverse causality between growth and corruption. The concern is that if growth affects corruption, the estimated impact of corruption on growth would be biased and even spurious. The analysis uses a sample of 57 countries and the average corruption score for the 1990s from Transparency International (TI) as a dependent variable. Besides economic growth, explanatory variables include education, judicial efficiency, the size of government, political and economic freedom, foreign aid, ethnicity, and the type of political regime. To address the possible bias, the authors use a 2SLS approach and ethnolinguistic fractionalization as an instrumental variable. They also use the Granger causality test, which determines the direction of the causal effect. Both the 2SLS and the Granger causality approaches reject the hypothesis that per capita GDP growth causes corruption. In contrast, corruption causes the GDP growth rate. Other results are that corruption is negatively and significantly correlated with the level of education, judicial efficiency, and economic freedom. It is positively and significantly correlated with foreign aid and the size of government. Mo (2001) complements Mauro (1995) by investigating other channels through which corruption affects growth. The channels under consideration include investment (ratio over GDP), human capital, and political instability. A preliminary investigation having shown that these three variables affect growth, the main analysis focuses on the impact of corruption on each of these variables. Accordingly, each of these variables is explained in terms of corruption and various control variables. The measure of corruption is obtained from TI, and the other explanatory variables are population, an index of political rights, and the ratio of public investment over GDP. All variables are averages over the period from 1970 to 1985 and cover 45 countries.

4 Consequences     131

The results of the ordinary least squares estimation suggest that a oneunit increase in the corruption index reduces the growth rate by 0.545 percentage points. The most important channel through which corruption affects economic growth is political instability, which accounts for about 53% of the total effect. The other channels include the level of human capital and the share of private investment in GDP. Using the 2SLS estimation generates qualitatively similar results but the total effect of corruption and the effects of the transmission channels are larger in absolute terms. Tanzi and Davoodi (1998) add to the literature by examining another channel of the impact of corruption on growth, namely public investment. Their empirical analysis distinguishes between the impacts on the quality and on the quantity of public investment. It uses the indexes of corruption from Business International (BI) and from the International Country Risk Guide (ICRG) for 68 countries averaged over the 1980–1983 period. The quantity of public investment is measured as a percentage of GDP. The quality of public investment is proxied using the following indicators: 1. Paved roads in good condition as a percentage of total paved roads 2. Electric power system losses as a percentage of total power output 3. Telecommunication faults per 100 mainlines 4. Water losses as a percentage of total water provision 5. Railway diesels in use as a percentage of total diesel inventory. Control variables are real per capita GDP, the ratio of government revenue to GDP, and the ratio of public investment to GDP. The results show that high corruption is associated with a high quantity of public investment but with low government revenue. More importantly, high corruption appears to be associated with low-quality infrastructure. These results are in line with our conceptual discussion. Corrupt governments prefer to fund big projects (“white elephants”) because they are more visible to citizens and also because they offer more opportunities for corrupt deals. In contrast, spending on operation and maintenance is given a lower priority, which results in a deterioration in the quality of infrastructure.

132     K. Sekkat

Habib and Zurawicki (2001) refine the analysis of the effect of corruption on investment by distinguishing domestic and Foreign Direct Investment (FDI). Based on a sample of 111 countries over 1994– 1998, domestic investment is calculated as the difference between FDI and total gross capital formation in the country. Measures of corruption are the Corruption Perceptions Index (CPI) and the ICRG’s index. Control variables are population, GDP per capita, annual GDP growth, exports plus imports as a percentage of GDP, inflation, and the political risk index from ICRG. The findings confirm the negative effects of corruption on investments and its two components. The degree of international openness and the political stability of the host market seem to moderate the influence of corruption. There is, however, an important difference between the two components of investment. The impact of corruption on domestic investment is substantially weaker than the impact on FDI. One explanation is that local businesses are more used to dealing with corruption and better equipped to handle it than foreigners. Foreign investors are at a relative disadvantage in this respect. The higher impact on FDI implies that corruption not only reduces productive capital but also limits technological progress since FDI is, generally, associated with substantial positive technological spillovers from foreign to domestic firms. Wei (2000a) supports these concerns about the impact of corruption on FDI. The analysis is based on bilateral stocks of FDI from 12 source countries to 45 host countries. FDI is explained in terms of three measures of corruption: the ICRG, TI, and BI indexes. Control variables include GDP, population, a dummy for linguistic ties, the adult literacy ratio, and host countries’ tax rates on foreign corporations. The results show that the host country’s tax rate and corruption both significantly deter FDI. The coefficients for the tax rate and corruption measures remain negative and statistically significant irrespective of the measure of corruption, the additional explanatory variables, and the estimation method. Using the estimated coefficients, the author finds that a one-unit increase in the corruption index is equivalent to a rise in the tax rate of 7.53 percentage points. For instance, an increase in corruption level from that of Singapore to that of Mexico has the same

4 Consequences     133

negative effect on inward foreign investment as raising the tax rate by over 50 percentage points. Shifting attention from the impact of corruption on factor accumulation to its impact on productivity as a source of growth, Lambsdorff (2003a) analyzes a sample of 69 countries in 2000. Productivity is measured by the ratio of GDP to capital stock, while corruption is measured using the 2001 CPI. Control variables are capital stock per capita, exports of fuels and minerals, the ratio of the deflators of investment and GDP, openness, secondary enrolment, and dummies for Africa and Asia. The author found that the absence of corruption is positively associated with productivity. A one-point increase in corruption on a scale of 0–10 lowers productivity by 2%. Improving the corruption score by six points, that is from Tanzania’s score to the UK’s score, increases productivity by more than 10%. In the sample, the capital stock is, on average, twice the value of GDP. The income level would thus rise by about 20%. Instead of the productivity level, Olson et al. (2000) investigate the impact on productivity growth. The sample includes 58 countries for the years 1960–1987. Average TFP growth over the period of observation is used as the dependent variable and explained in terms of the ICRG index of corruption for the year 1982. Control variables are the black market premium, the share of government consumption in GDP, GDP per capita in 1960, secondary enrolment in 1960, and regional dummies. The main finding is that the coefficient of corruption is significant and implies that corruption reduces the growth rate of productivity. For instance, if Haiti had had the same corruption as Hong Kong, its productivity would have grown by 1.49 percentage points faster per annum, which is a substantial increase. One major source of productivity growth is innovation, which can also be affected by corruption. Mahagaonkar (2008) uses African firm-level data to investigate this issue. The data come from the World Bank’s Productivity and the Investment Climate: Private Enterprise Survey, which took place between 2002 and 2004. The data of Benin, Madagascar, Mali, Mauritius, Tanzania, Zambia, and South Africa were included in the study, giving a total of 3477 firms. A distinction is made between four innovative activities: Product Innovation, Process

134     K. Sekkat

Innovation, Marketing Innovation, and Organizational Innovation. The dependent variables are dummies pertaining to each type of innovation and are based on the response to a question about whether the firm performs the type of innovation under consideration. Explanatory variables are corruption, reinvested profits, firm size, client technology, supplier technology, in-house technology, foreign ownership, financial access problems, and country effects. All variables are drawn from the survey responses. In particular, the corruption measure is the percentage of sales used to give gifts or informal payments to public officials to “get things done”. The analysis led to the conclusion that an increase in corruption negatively affects the likelihood of product innovation and the likelihood of organizational innovation. No significant effect of corruption on process innovation was found, while corruption appears to increase the likelihood of marketing innovation. Regarding the other explanatory variables, reinvested profits increase the likelihood of product innovation while problems with access to finance decrease such likelihood. As expected in many developing countries, the findings show that large firms are mainly responsible for the increase in the likelihood of product innovation. Similar results emerge for process innovation. We now turn to the demand side determinants of growth. As explained above, we will focus on the external component, namely trade. Anderson and Marcouiller (2002) investigate whether corrupt officials reduce trade. The measure of corruption concerns the importing country and is based on the World Economic Forum’s (WEF) 1997 Executive Survey. Participants in the WEF survey were asked to assign a score ranging from 1 (strongly disagree) to 7 (strongly agree) to the statement: Irregular additional payments are not common in business and official transactions. The dependent variable is the 1996 bilateral import expenditures of 48 countries. In addition to corruption, the explanatory variables include population, GDP, distance from capital city to capital city, unweighted average tariffs and dummy variables to capture sharing a common border, common language, or common membership of ASEAN, the EU, Mercosur, or NAFTA.

4 Consequences     135

The findings are that a 10% rise in a country’s corruption index leads to a 5% decrease in its import volumes. Costs associated with corruption appear to be a serious impediment to countries’ trade. For example, if the seven Latin American countries in the sample (Argentina, Brazil, Chile, Colombia, Mexico, Peru, and Venezuela) had similar corruption indexes to the mean of the members of the European Union, Latin American import volumes would be 30% higher. Interestingly, lowering Latin American tariffs to the levels applied by the USA would have a similar effect, increasing trade by 35%. A much larger increase (51%) in GDP would be necessary to generate a comparable increase in imports. While Anderson and Marcouiller (2002) focus on corruption in the importing country, Méon and Sekkat (2004) investigate the effects of corruption in the exporting countries. The analysis uses data on total manufactured exports from 40 countries over the 1990s. The ratio of manufactured exports to GDP is explained in terms of corruption, the real effective exchange rate, the average GDP growth rate of trading partners, and investment in the manufactured goods sector. The measures of corruption are the CPI and World Bank (WB) indexes. The estimated coefficients of corruption always have the expected signs and are significant. A reduction in the level of corruption results in an increase in manufactured exports. The analysis shows that the results are robust to different econometric approaches and institutional indicators. Further calculations show that if, say, Morocco’s corruption index improved to the level of the Swiss one, its manufactured exports ratio would increase by 18.45%. Morocco’s manufactured export ratio would then have been 12.32% instead of 10.4% in 1997, i.e., a ratio similar to Poland’s. Musila and Sigué (2010) consider corruption levels in exporting and importing countries simultaneously. The analysis is based on a gravity model using annual data over the period 1998–2007 for 47 African countries as importers and 180 exporting countries. The CPI is the measure of corruption. The additional explanatory variables are the traditional gravity model variables such as the GDP of trading partners, the distance between partners, and dummies for the existence of a common border, common language, common currencies, or a free trade area.

136     K. Sekkat

The estimated coefficients of corruption have negative signs and are statistically significant. Corruption in both importing and exporting countries negatively affects trade. This implies that the export and import trade of African countries would increase not only if African countries became cleaner but also if their trading partners do too. The estimates suggest that if a country with a CPI of 2.8 (African average) improved to 5.9 (level of Botswana), its exports to Africa would improve by at least 15%. Corruption not only reduces the volume of international trade, but also distorts its geographic distribution. In other words, countries that are more inclined to pay bribes trade more with corrupt countries than do countries that are less inclined to pay bribes. Lambsdorff (1998) investigates the validity of this statement using data on the 19 biggest exporting countries and the 87 biggest importing countries for the four-year period 1992–1995. In order to examine whether there is a tendency for some countries to export to corrupt markets, a separate equation for each exporting country is estimated. The level is determined by the 1996 CPI. The rest of the explanatory variables are traditional gravity model variables similar to those mentioned above. Testing whether the coefficients of the corruption index are equal for two given countries makes it possible to identify which countries are more inclined to export to corrupt countries. The results reject the null hypothesis of equal export behavior for many pairs of countries. For instance, the test for Malaysia and the UK implies that the UK is significantly more inclined than Malaysia to export more to corrupt countries. Similar results were found for Germany as compared to Sweden or South Korea in comparison with Australia. For the rest, it appears that Belgium, France, Italy, the Netherlands, and South Korea tend to export more to corrupt countries. Méon and Sekkat (2008) complement the above analysis, which looks at the geographic composition of trade, by examining the impact of corruption on the goods composition of trade. Specifically, they examine whether the impact of corruption differs for manufactured and non-manufactured exports. The distinction is important because all exports are not equivalent in terms of development and growth. In

4 Consequences     137

particular, the development economics literature suggests that manufactured exports are more conducive to growth than non-manufactured exports. Using panel data for 60 countries over 1990–2000, a separate equation is estimated using each type of exports as a dependent variable. The corruption indicator comes from the WB, and control variables are the real effective exchange rate and the GDP growth rates of the country’s partners. Two estimation methods are used: ordinary least squares and two-stage least squares. The results suggest that corruption chiefly hurts a country’s capacity to export manufactured goods. Accordingly, an improvement in institutional quality should result in an increase in manufactured exports. Exports of non-manufactured goods seem to be related to corruption in the opposite way. However, the results are sensitive to the estimation method.

2.2 Conditional Effects Almost all the studies that examine the effects of corruption conditional on certain factors (quality of bureaucracy, rule of law, or level of development) developed in the framework of the “greasing-the-wheels” vs. the “sanding-the-wheels” arguments discussed above. The debate concerns the extent to which corruption may be beneficial in a second-best world because of the distortions caused by the ill-functioning of other aspects of the economic environment. For instance, an inefficient bureaucracy constitutes an impediment to investment that some speed or grease money may help to circumvent. For the sake of clarity, we will start with microeconomic studies before turning to macroeconomics. One of the early empirical investigations of this issue is Kaufmann and Wei (1999). Using data from three World Enterprise Surveys, the authors examine the relationship between bribe payment and management time wasted with bureaucrats. The dependent variable is the time spent by managers negotiating with bureaucrats. It corresponds to the response to the question: “On a one-to-seven scale, whether the senior management of your company spends more or less than 30 percent of

138     K. Sekkat

its time dealing with government bureaucracy”. The main explanatory variable is corruption and corresponds to the response to the question: “On a one-to seven scale, rate the extent of irregular, additional payments connected with imports and exports permits, business licenses, exchange controls, tax assessments, police protection or loan applications”. Control variables include regulatory burden, regulatory discretion, firm size, foreign investor and country and sector dummies. The results show that the coefficient associated with corruption is positive and statistically significant. Paying bribes does not reduce the time spent with bureaucracy. On the contrary, firms that pay more bribes also spend more time negotiating with bureaucracy, which is inconsistent with the “greasing-the-wheels” hypothesis. Large or foreign-owned firms waste less time negotiating with government officials. De Rosa et al. (2015) use the 2009 Business Environment and Enterprise Performance Survey (BEEPS) data for the economies of Central and Eastern Europe and the CIS to investigate the relationship between corruption, red tape, and productivity. Specifically, the authors examine whether corruption reduces the negative effect of red tape on productivity. The dependent variable (productivity) is explained in terms of corruption, red tape, an interaction term for the two variables, and a set of firm (size, age, exporter, innovator, and foreign-owned), industry, and country characteristics. The extent of red tape is proxied by the percentage of time spent by senior management negotiating with officials in order to obtain a favorable interpretation of the regulations. Corruption is a dummy which equals 1 if the firm replies that it is frequent, usual, or always common to pay some irregular additional payment or gifts to “get things done”. The interaction term shows whether, when regulation is overly restrictive, corruption helps entrepreneurs to negotiate with bureaucrats and leads to a less negative impact on productivity. The results of the analysis show that the interaction term is not significant while corruption has a statistically significant negative effect on productivity. Accordingly, the results lend no support to the “greasing-the-wheels” argument. Corruption is not a second-best option to achieve higher productivity levels by helping firms to circumvent burdensome regulatory requirements.

4 Consequences     139

Dreher and Gassebner (2013) study another aspect of the “greasing-the-wheels” hypothesis by investigating whether the impact of regulations on entrepreneurship depends on corruption. In other words, they test whether corruption reduces the negative impact of regulations on entrepreneurship. The data are drawn from the Global Entrepreneurship Monitor (GEM), which contains survey-based annual data on early-stage entrepreneurial activity. The sample covers 43 countries for the period 2003–2005. The dependent variable is defined as the percentage of people in the population aged between 18 and 64 who have taken steps toward creating a new business in the past year. To measure corruption, the CPI and WB indexes are used. To proxy the severity of regulation, four variables, drawn from the WB doing business dataset, are combined. These are the number of procedures required to start a new business, the number of days required to start a new business, the costs of starting a new business, and the minimum capital required to start a new business. Again, the “sanding-the-wheels” versus the “greasing-the-wheels” hypotheses are tested using an interaction between the severity of regulation and corruption. Control variables include lagged GDP per capita, the square of lagged GDP per capita, and a dummy for former communist countries. Preliminary results show that some regulations matter for entrepreneurial activity. Specifically, the number of procedures required to start a business and the minimum capital requirements is detrimental to entrepreneurship. In contrast, the number of days required to start a new business and the out-of-pocket costs required to start a business do not appear to be obstacles to entrepreneurial activity. In the absence of any costs to starting a business, corruption has a negative effect on entrepreneurship. Likewise, in the absence of corruption, the costs of starting a business handicap entrepreneurship. Since the main focus here is on the “greasing-the-wheels” vs. the “sanding-the-wheels” hypotheses, the coefficients of the interaction term deserve special attention. These coefficients are significant and positive, implying that corruption reduces the impact of the costs of doing business on entrepreneurship. This fits with the “greasing-thewheels” hypothesis: Corruption facilitates firm entry in highly regulated economies.

140     K. Sekkat

While at the microeconomic level corruption might appear to have positive or negative impacts in the presence of certain other institutional imperfections, the question of the impact on the whole economy, that is, once all effects have operated, remains unanswered. Accordingly, the above analyses need to be complemented by macroeconomic studies which take account of all the effects on the economy. Wei (2000b) addresses this concern by examining whether corruption tends to mitigate the effects of tax rates and capital controls on aggregate bilateral FDI. The dependent variable is bilateral stocks of FDI in 1991 from 14 major source countries to 45 host countries. The list of source countries includes the seven largest foreign investors in the world.1 The explanatory variables of interest are the host country’s tax rates, corruption, capital control measures and interaction terms between corruption and tax rates, and capital controls, respectively. The author uses two capital control measures and two corruption measures. One measure of capital control comes from Business International (BI) while the other is a dummy based on the IMF’s Annual Report on Exchange Arrangements and Exchange Restrictions. Corruption measures are the BI and the TI indexes. Control variables include GDP, bilateral distance and indicators of restrictions on inward FDI, on joint ventures with domestic firms, on bids on public sector projects, and on corporate control rights. Disregarding the interaction terms, capital control, tax rates, and corruption have negative and statistically significant coefficients. The coefficient of the interaction term with corruption is not significant irrespective of the measure of corruption. This implies that there is no statistical support for the “greasing-the-wheels” argument. Regarding capital control, the coefficient of the interaction with corruption is positive and statistically significant with one measure of capital control (BI) and non-significant with the other (IMF). Thus, support for the view that corruption reduces the incidence of taxation on FDI is weak. In sum, the data do not support the “greasing-the-wheels” argument.

1The

USA, Japan, Germany, the UK France, Canada, and Italy.

4 Consequences     141

While De Rosa et al. (2015) examine at the microeconomic level the “greasing-the-wheels” vs. the “sanding-the-wheels” hypotheses from the perspective of productivity, Méon and Weill (2010) address a similar question at the macroeconomic level. The study analyzes the interaction between aggregate efficiency, corruption, and other dimensions of governance using a panel of 54 developed and developing countries over the period 1994–1997. There are three measures of corruption: the WB index, the CPI, and a corruption index used by Wei (2000a).2 The other institutional quality indicators are drawn from the WB and concern government effectiveness, lack of violence, regulatory burden, rule of law, and voice and accountability. Finally, efficiency is measured using the technical efficiency approach developed by Battese and Coelli (1995). An interaction term between corruption and the other quality-of-governance indicators is incorporated in the estimation. Disregarding the interaction terms, the results show that aggregate efficiency rises with the quality of governance as measured by the WB indicators. Control of Corruption Index leads to the same qualitative results. This implies that lower corruption is associated on average with greater efficiency. The coefficient of the interaction terms between corruption and other facets of governance is either positive or insignificant. The results provide some support for the “greasing-the-wheels” hypothesis but the support is weak since the conclusion depends on the other indexes of quality of governance. Another study of the “greasing-the-wheels” vs. the “sanding-thewheels” hypotheses at the macroeconomic level is Méon and Sekkat (2005). The authors examine the impact of corruption on growth and investment in the presence of weaknesses in other aspects of governance using a sample of around 70 countries between 1970 and 1998. The dependent variables are per capita GDP growth and the investment to GDP ratio. The corruption data and the other quality-of-governance indicators are the same as in Méon and Weill (2010) presented above. 2Wei’s

index is an extension of the corruption index published in the World Economic Forum’s Global Competitiveness Report 1997. To increase the coverage of his dataset, Wei (2000a) filled the gaps left by that first index with the information provided by the World Bank’s 1997 World Development Report.

142     K. Sekkat

Control variables are the initial per capita income, the initial level of schooling, the average population growth rate over the period, the average ratio of investment to GDP over the period, and the degree of openness of the economy. The relevant interaction terms are also introduced. Like previous studies, the findings show a negative effect of corruption on both growth and investment. Unlike previous studies, however, corruption has a negative impact on growth independently of its impact on investment. These impacts are, however, different depending on the quality of governance. They tend to worsen when indicators of the quality of governance deteriorate. Specifically, corruption slows growth down even more in countries suffering from weak rule of law and inefficient government, even when one controls for investment. Moreover, it is found that weak rule of law, inefficient government, and political violence also tend to worsen the negative impact of corruption on investment. Overall, the results strongly reject the “greasing-the-wheels” hypothesis in favor of the “sanding-the-wheels” one. They imply that reducing corruption would be more profitable in countries where other aspects of governance are poor, which stands in sharp contrast to the opinion of those who view corruption as a lubricant. While not dealing with the “greasing versus sanding-the-wheels” hypothesis per se, Aidt et al. (2008) give strong support to the idea that the relationship between corruption and growth is nonlinear. They develop a conceptual framework where corruption is treated as an endogenous variable which depends on the quality of political institutions. The model considers two distinct governance regimes. In one regime, institutions are of a sufficiently high quality to allow citizens to use the threat of ruler’s replacement to reduce corruption. In the other regime, institutions are deficient and citizens cannot use such threat. The empirical model is estimated using a sample of around 70 countries drawn from all continents. Two measures of corruption are used: the CPI from Transparency International and the CCI from the World Bank. The measure of the quality of political institutions is the “voice and accountability” index from the World Bank. Control variables include investment share, population growth, gross enrolment in primary education, the initial level of GDP, regional effects, and whether a

4 Consequences     143

country belongs to the common law tradition. The estimation method is the GMM with the index of ethnolinguistic fractionalization and the age of democracy as instruments. The results are that corruption has a substantial negative impact on growth in the regime with high-quality political institutions. Corruption has no impact on growth in the regime with low-quality institutions. A one-point reduction in corruption increases growth in the short run by 0.5–0.6 percentage points and by 0.37–0.39 in the long run for countries in the good institutions regime. However, the authors note that while their analysis confirms the importance of allowing for nonlinear effects in the relation between corruption and economic growth, the result is not a support for the “greasing the wheels” hypothesis. This hypothesis implies that corruption improves efficiency in the presence of weak institution while the paper finds that corruption has no impact on growth in the regime with weak institutions.

3 Evidence: Non-economic Effects Like other sciences, economics ultimately seeks to improve human well-being. By enhancing growth, the fight against corruption affects only one necessary condition of human well-being. Other dimensions such as poverty, health, education, or trust also need to be improved for human well-being to be enhanced further. Akçay (2006) studies the impact of corruption on the Human Development Index (HDI) as captured by the three following indicators: life expectancy at birth, educational attainment, and real GDP per capita in purchasing power parity (PPP). The indicators have equal weight and are scaled to vary between 0 and 1, with 0 indicating the lowest level of human development. The analysis is conducted for 63 countries in 1998. The control variables are the urbanization rate, economic freedom, and democracy. Corruption is assessed on the basis of the CPI, the International Country Risk Guide (ICRG) index, and the WB index. The empirical analysis suggests that more corrupt countries tend to have lower levels of human development irrespective of the measure of

144     K. Sekkat

corruption used. Using the CPI, for instance, shows that a one-point increase in corruption reduces human development by 0.013 points. With the ICRG and the WB indexes, a one-point increase in corruption reduces human development by 0.048 and 0.041 points, respectively. Instead of the impact of corruption on the HDI as discussed above, Li et al. (2000) examine the impact of corruption on income distribution. They explain the GINI coefficient of income distribution of 48 countries over the period 1982–1994 in terms of real per capita GDP growth, the ICRG corruption index, the lagged value of years of primary schooling, the financial development of the country, GDP, and the initial distribution of assets as measured by the initial land GINI coefficient. To allow for a possible nonlinear relationship between corruption and GINI, the former is introduced among the explanatory variables in level and squared form. The results show that corruption alone explains a large proportion of the GINI differential between the groups of developing and industrial countries. However, corruption and income inequality are not monotonically related. Corruption affects the GINI coefficient in an inverted U-shaped way. Inequality is low when levels of corruption are low. It starts increasing with corruption until a certain point, after which inequality decreases as corruption increases. In countries with more inequality in asset allocation, corruption also raises inequality. In Latin America, corruption has distinct effects: It has a greater impact on inequality relative to other continents. In particular, when government spending is higher, corruption is more harmful for growth. The above study is based on a sample which contains virtually no African countries. Yet, Africa is a continent with high corruption and high income inequality. Accordingly, it is clearly advisable to examine the same issue using African data. This is accomplished by GyimahBrempong (2002), who uses annual data for a sample of 21 African countries over the period 1993–1999. The measure of income inequality is the GINI coefficient, while the measure of corruption is the CPI. In addition to corruption, the GINI coefficient is explained in terms of the growth rate of income, the level of per capita income, government consumption, and education.

4 Consequences     145

The estimated coefficient of corruption is negative, relatively large and significant. The result is robust to estimation methodology. A oneunit increase in corruption is associated with an increase of between four and seven units in the GINI coefficient of income inequality. Alternatively, a standard deviation decrease in corruption is associated with a decrease of between 7.3 and 12.3 units in the GINI coefficient. In addition to this direct effect, corruption seems to be correlated with income inequality through other channels. For instance, the results show that an increased growth rate of per capita income decreases the GINI and that corruption has a large negative effect on economic growth. Taken together, the two results highlight one of the indirect channels alluded to. More importantly, a strand of the economic development literature suggests that income inequality negatively affects economic growth. Combined with the results of the study, this suggests the risk of the emergence of a vicious circle in which corruption increases inequality, which in turn decreases economic growth, which feeds back into income inequality. Overall, the combined effects suggest that corruption hurts the poor more than the rich in African countries. Corruption not only affects the weakest class of society through inequality and poverty but also through access to numerous public services such as health care and education services. To highlight these effects, Huang (2008) conducts an empirical examination of the relationship between corruption and educational outcomes, using a sample of 50 countries. The study focuses on two educational outcome measures, which are used as dependent variables. The first outcome concerns the quality of education and is assessed on the basis of an Educational Performance Index (EPI). The EPI is a composite index derived from each country’s average science and mathematics scores in the 2003 Trends in International Mathematics and Science Study (TIMSS). Educational quantity is measured by UNESCO’s School LifeExpectancy (SLE) indicator. The explanatory variable of interest is the CPI, and control variables are average size of households, the level of political and civil freedom, and the level of democracy in a country. Correlation analysis indicates that the CPI is significantly and negatively correlated with the EPI (−0.43) and the SLE (−0.75), implying that countries with higher CPI scores have lower EPI and SLE scores.

146     K. Sekkat

Regression analyses, which control for the effect of other variables, confirm the significant negative association between corruption and the EPI and SLE. On average, a one-point increase in the CPI decreases the EPI by 11.5% of its standard deviation. In the case of the SLE, a one-point increase in the CPI decreases the SLE by 25% of its standard deviation, or approximately eight months. These results here are in line with Mauro (1998) who, in examining the impact of corruption on government expenditure, finds a negative and significant relationship between corruption and government expenditure on education. The paper uses the ICRG index of corruption (1982–1995 average) to explain the composition of government expenditure (average 1970–1985). Control variables are GDP per capita 1980, the share of population aged between 5 and 20, and the BI Political Stability Index for 1980–1983. Jain (2002) examines the impact of corruption on an important dimension of human capital, namely the provision of health care. Such provision is proxied using three indicators as dependent variables. These are the rates for immunization, births attended by health staff, and child and infant mortality. The dependent variables are explained in terms of three corruption indexes (ICRG, CPI, and the WB) and per capita income, public health spending, average years of education, females aged 15 or older, the dependency ratio, and urbanization in a sample of 100 countries over the period 1985–1997 The empirical analysis shows that a high level of corruption has adverse consequences for child and infant mortality rates and for the percentage of low-birthweight babies in total births. In particular, child mortality rates in countries with high corruption are about one-third higher than in countries with low corruption; infant mortality rates and percentages of low-birthweight babies are almost twice as high. The results are robust to different estimation methods and controls. The preceding findings are confirmed in the study by Azfar and Gurgur (2008). They focus on the Philippines and use data from a survey undertaken in the spring of 2000 and covering 19 provinces and 80 municipalities from 11 regions. The results indicate a significant and negative effect of corruption on the quality of health services. Moreover, corruption affects health outcomes differently in rural versus urban

4 Consequences     147

areas. Rural areas suffer from longer wait times at public health clinics, late immunization of infants, and less satisfaction with public health services. Klomp and De Haan (2008) explore both the direct effect of corruption on the health of individuals and the possible indirect effect through income for instance. Examining 101 countries over the period 2000–2005, the analysis uses the ICRG indicators of corruption for 2003, bureaucratic quality, and law and order as well as Freedom of the World data from the Fraser Institute. A synthetic indicator of individual health and another for the health care sector are constructed for 2006 using information from the WHO and the World Bank. Control variables include GDP per capita in 1980, income inequality, primary school enrolment, secondary school enrolment, investment as a share of GDP, and the share of population living in rural areas. It appears that none of the good governance indicators has a direct effect on the health of individuals. They have, however, a positive effect on the health care sector. A 1% increase in the quality of governance leads to an increase of 0.28% in the quality of the health care sector which, in turn, increases the health of individuals by about 0.23%. Moreover, the quality of governance has another indirect impact on health via its positive impact on income. Through income, a 1% increase in the quality of governance leads indirectly to an increase in the health of individuals of about 3.54%. It follows, therefore, that the indirect effect of governance through income is the most important channel through which governance can help to improve health. Besides the effects on individuals’ health, education, or income, corruption impacts other, sometimes unsuspected, aspects of people’s life. These include increased vulnerability, unsafe behavior, deforestation, confidence in public institutions, regime legitimacy, and voter attendance Starting with vulnerability, Hunt (2007) uses a cross-country sample and another sample which focuses on Peru to investigate whether crime victims are more likely to bribe public officials than non-victims. The idea is that misfortune might increase a victim’s demand for public services and increase bribery because a desperate or vulnerable victim sees it as the only way to obtain reparation. Powerful victims can more easily use their large networks of “friends” or “relatives” to obtain reparation.

148     K. Sekkat

The analysis is based on the International Crime Victims Survey (ICVS, 1992–2001) and the Peruvian Household Survey (ENAHO). The ICVS, conducted for the United Nations Interregional Crime and Justice Research Institute, provides individual-level data from 37 transition, middle- or low-income countries. It asks the question: “In some countries, there is a problem of corruption among government or public officials. During the 1990s, has any government official, for instance a customs officer, a police officer or inspector in your country asked you, or expected you to pay a bribe for his or her services?” The ENAHO is conducted yearly by Peru’s national statistical agency (INEI). In 2002 and 2003, the ENAHO asked households having used one of 21 different types of officials or institutions whether an official asked for a bribe, whether the respondent felt obliged to make such a payment, whether he/she made such a payment voluntarily, or refused to make such a payment and the amount paid if he/she paid. A probit specification is used for each sample, where the dependent variable is whether a given individual in a given country has paid a bribe. The explanatory variable of interest is a dummy indicating whether the individual was a crime victim. Control variables are the respondent’s income quartile, city size, and indicators of the size of the household. The findings confirm that victims of crime are substantially more likely to bribe than others. Using the ICVS sample, crime victims are between 2.9 and 8.2 percentage points more likely to pay bribes than similar non-victims. In Peru, crime victims are 6.3 percentage points more likely to pay a bribe than similar non-victims. In contrast to the likelihood of paying a bribe, the amount of the bribe is independent of the fact of being a victim. These results hold even after controlling for a large number of individual and household characteristics and are confirmed in the case of other misfortunes. For instance, the bribery rate increases by 2.2 percentage points for a victim of job loss and by 3.8 percentage points for the death of the household’s income earner. The effects are strongest with the police and strong with the judiciary. Bertrand et al. (2007) investigate whether corruption produces unsafe drivers. They collect detailed micro-data on 822 driver’s license applicants in New Delhi. The analysis is structured around five

4 Consequences     149

questions: Do people pay bribes to get a license? Can corruption be used to speed up the process of getting a license? Do bad drivers use bribes to get a license? Do bureaucrats raise hurdles to extract bribes? Finally, how does corruption take place? Participants were randomly assigned to one of three groups: bonus, lesson, and comparison groups. Participants in the bonus group were offered a financial reward if they could obtain their license fast. Participants in the lesson group were offered free driving lessons. Finally, the comparison group was composed of applicants who were simply asked to participate in the survey. A surprise driving test was organized after participants had obtained their licenses. The authors found that those who want their license faster (e.g., the bonus group) get it 40% faster and at a 20% higher rate, implying that bureaucracy responds to individual needs. However, members of this group do not learn to drive safely. In fact, 69% of them were rated as “failures” on the surprise driving test. Accordingly, the bureaucracy does not seem concerned with the public interest. Members of the lesson group have superior driving skills. However, they are only slightly more likely to obtain a license than the comparison group and far less likely than the bonus group. Further investigation showed that bureaucrats arbitrarily fail drivers at a high rate during the driving exam irrespective of their ability to drive. As a result, individuals agree to bribe the bureaucrat in order to avoid taking the exam. In other words, bureaucrats create red tape to extract bribes, and this corruption undermines the very purpose of the driving regulation. Deforestation, a major global issue, is not immune to corruption. In its 2001 report, the UN Food and Agriculture Organization (FAO) stated that corruption is one of the main causes of deforestation. Corruption by forestry officials takes different forms: approval of illegal contracts with private enterprises, illegal sale of harvesting permits, under-declaring volumes cut in public forests, underpricing of wood in concessions, harvesting of protected trees, and smuggling of forest products across borders. Koyuncu and Yilmaz (2009) explore the issue of deforestation and corruption using the CPI, the ICRG index, and the Business Intelligence Index. The dependent variable is measured as the

150     K. Sekkat

percentage change in forest surface in each country. Control variables are rural population growth, permanent cropland (as a % of land area), and GDP. The sample is composed of 100 countries for the periods 1980–1990, 1990–1995, and 1990–2000. Irrespective of the indicator of corruption, a highly statistically significant positive association was found between the corruption level in a country and the deforestation rate. The findings are robust to the introduction of other determinants of deforestation and hold across all periods and with different estimation methods. Anderson and Tverdova (2003) shift the concern to the political arena, namely attitude toward government. They examine whether citizens in countries with higher levels of corruption express more negative evaluations of the performance of the political system and exhibit lower levels of trust in civil servants. They use individual-level data from the International Social Survey Program (ISSP) of 1996 for 16 developed and transition economies. The analysis distinguishes between citizens’ attitude toward a country’s government and attitude toward other institutions. The former is based on the responses to the question: “All in all, how well or badly do you think the system of democracy in your country works these days?” The possible answers were: (i) It works well and needs no changes; (ii) it works well, but needs some changes; (iii) it does not work well and needs a lot of changes; and (iv) it does not work well and needs to be completely changed. Attitudes toward other institutions are based on the responses to the question: “Most civil servants can be trusted to do what is best for the country”. Respondents could answer “strongly agree”, “agree”, “neither agree nor disagree”, “disagree”, and “strongly disagree”. The responses to the above question were used to construct two dependent variables, which are explained in terms of corruption as measured by the CPI and control variables. These variables include both country-level variables, such as macroeconomic performance, democratic longevity and the level of democracy, and individual-level variables such as political interest, electoral participation, socioeconomic status, employment situation and other demographic variables. To examine whether the results differ between citizens who support the

4 Consequences     151

government and others, an interaction term between corruption and support for the government is added to the equation. The results provide clear support for the hypothesis that individuals in countries with higher levels of corruption evaluate the performance of the political system more negatively and have less trust in civil servants. However, supporters of the government had less negative attitudes. A respondent in a country where corruption is absent scores 3.22 on the four-point scale measuring system performance evaluations. The score drops to 2.80 in a country scoring in the mid-range of the corruption measure, while in the most corrupt country the score drops to 2.61. Moreover, a respondent in a country where corruption is absent scores 4.26 on the five-point scale measuring trust in civil servants. The score drops to 3.33 in countries in the mid-range of the corruption scale and to 2.76 in the most corrupt countries. Clausen et al. (2011) extend the study of the relationship between corruption and trust to different types of institutions, not only civil servants. They use the 2008/2009 wave of the Gallup World Poll (GWP), which is a large cross-country household survey covering over 78,000 respondents in 103 countries. The poll asks respondents whether they have confidence in the military, the judicial system and courts, national government, health care or medical systems, financial institutions or banks, religious organizations, the media, and honesty of elections. The responses are combined to create an overall index of confidence which ranges from 0 (no confidence in any of the institutions) to 4 (confidence in all institutions). The index and its components are separately explained in terms of corruption experience and perception. Experience is based on the responses to the question: “Sometimes people have to give a bribe or present in order to solve their problems. In the last 12 months, were you, personally, faced with this kind of situation, or not (regardless of whether you gave a bribe/present)?” Perception is based on the response to the question: “Is corruption widespread throughout the government in this country, or not?” Control variables are respondent age, gender, marital status, education income, and whether the household has access to the Internet and television.

152     K. Sekkat

All estimates give a highly significant and negative correlation between corruption and confidence in institutions. The coefficients imply that a one standard deviation increase in corruption reduces confidence in institutions by between 0.2 and 0.3 points on a 0–4 scale. The result holds for both the perception and the experience measures and is causal. Older individuals seem to have a lower degree of confidence in institutions. Married respondents exhibit higher confidence than singles. Higher income and education as well as access to Internet and TV appear to reduce confidence although the effects are not robust. Finally, the analysis of the impacts of corruption on specific institutions gives similar results. Since the preceding study pointed to a specific role for education, Hakhverdian and Mayne (2012) investigate whether the negative effect of corruption on trust in institutions depends on education. They use the results of the European Social Survey (ESS) conducted in 2008– 2009 and focus on eight Eastern and Central European and 13 Western European countries. The dependent variable, institutional trust, is based on a combination of the responses of citizens to questions about their levels of trust in their country’s parliament, legal system, police, politicians, and political parties. The combination gives an overall index of institutional trust which ranges from 0 (“no trust at all”) to 10 (“completely trusting”). The two main explanatory variables are the levels of education and corruption. The former comes from the ESS results on education. The latter is the CPI. Control variables are age, gender, income, religion, religious attendance, paternal level of education, social trust, satisfaction with the present state of the national economy, GDP per capita, and level of unemployment. The results show that when the corruption score is at its maximum, the most educated persons are less politically trusting than the least educated. The gap between the two groups decreases as corruption becomes less prevalent. In a perfectly clean society, more educated people are estimated to be more trusting than less educated people. In sum, in relatively corrupt societies the most educated people are less trustful than the least educated people; in moderately clean societies, they are equally trusting; and in relatively clean societies, they are more trusting.

4 Consequences     153

Since corruption erodes trust in the political system, one can expect that it would also affect citizens’ participation in political life. The issue is particularly important because if corruption also dissuades citizens from participating in democratic elections, regime legitimacy problems may be exacerbated, increasing political instability and creating the threat of an autocratic reversal. Stockemer et al. (2013) investigate the link between corruption and voter behavior in a sample of around 100 democracies, including newly emerging or poor democracies over the period 1985–2007. Turnout is the dependent variable and is measured as the percentage of eligible adult citizens that cast a ballot in their country’s national legislative elections. The measure of corruption is the ICRG index. Control variables are compulsory voting laws, electoral system type, decisiveness of the election, and competitiveness of electoral races. Instrumental variable estimation is used to account for feedback loops or reversed causation. The findings are that corruption has a statistically significant and negative impact on voter turnout. For every one-point increase in a country’s corruption, turnout decreases by over six points, implying that very corrupt countries have 20–30 percentage points fewer citizens turning out at elections compared with countries with little corruption.

4 Conclusion There is controversy about whether corruption has positive or negative effects. Some authors argue that corruption may be beneficial by “greasing the wheels” when bureaucracy is inefficient or regulation is too burdensome, while others state that it is always detrimental and ends up “sanding the wheels”. There are two possible sources for such a divergence. The first one concerns the basis of the comparison. Specifically, does the comparison concern a country with inefficiency but no corruption and a country with both inefficiency and corruption, or does it concern a country free from inefficiency and corruption and a country with both of them? The source of divergence comes from whether the analysis is conducted in a partial or a general equilibrium. While corruption can “grease the wheels” from a partial equilibrium perspective,

154     K. Sekkat

it may ultimately “sand the wheels” from a general equilibrium perspective through spillovers and externalities among economic activities. Analyses of the effects of corruption have developed along two lines. One distinguishes economic and non-economic effects. The other contrasts conditional and unconditional effects. The economic effects concern growth, physical and human capital formation, productivity, infrastructure, international trade, and FDI. The non-economic effects of corruption concern the provision of health care and education services, safety and security, environment, electoral participation, and confidence in public institutions. The conditional effects relate to the debate around the “greasing-the-wheels” versus the “sanding-thewheels” hypotheses. The unconditional effects concern the “average” impact of corruption independently of the functioning of other institutional aspects. As far as the unconditional economic effects are concerned, empirical analyses show that corruption influences GDP directly and indirectly through investment, human capital, public spending, or political instability. Both the direct and the indirect influences are negative but their magnitude depends on the variable under consideration. For instance, the impact of corruption on domestic investment is substantially weaker than the impact on FDI. Such higher impact on FDI implies that corruption not only reduces productive capital but also limits technological progress since FDI is, generally, associated with substantial technological positive spillovers from foreign to domestic firms. Corruption also affects productivity and innovation, which are major sources of growth. Corruption not only affects imports and exports but also their composition. It was found, for instance, that countries which are more inclined to pay bribes would trade more with corrupt countries than do countries less inclined to bribe. Moreover, corruption affects manufactured exports more than non-manufactured exports. The distinction is important because the development economics literature suggests that manufactured exports are more conducive to growth than non-manufactured exports. The non-economic impacts of corruption include an increase in poverty and daily interactions with the administration. Corruption also hurts the poor more than the rich, especially in African countries. The

4 Consequences     155

findings also confirm the significant negative association between corruption and the provision of education and health services. Individuals in countries with higher levels of corruption are found to evaluate the performance of the political system more negatively and have less trust in civil servants. Related findings are that corruption has a statistically significant and negative impact on the number of voters. Turning to the conditional effects of corruption, the evidence is less consensual than for the unconditional effects. Corruption does not reduce the time spent negotiating with bureaucracy. On the contrary, firms that pay more bribes also spend more time dealing with burdensome regulations. This is inconsistent with the “greasing-the-wheels” hypothesis. Some evidence suggests that corruption reduces productivity, while other evidence finds the reverse, that is, corruption greases the wheels of productivity. Corruption also seems to grease the wheels of entrepreneurship. Finally, and less encouragingly, corruption slows growth even more in countries suffering from weak rule of law and inefficient government.

References Aidt, T. S. (2016). Rent Seeking and the Economics of Corruption. Constitutional Political Economy, 27(2), 142–157. Aidt, T., Dutta, J., & Sena, V. (2008). Governance Regimes, Corruption and Growth: Theory and Evidence. Journal of Comparative Economics, 36, 195–220. Akçay, S. (2006). Corruption and Human Development. Cato Journal, 26(1), 29–48. Ali, A. M., & Isse, H. S. (2002). Determinants of Economic Corruption: A Cross-Country Comparison. Cato Journal, 22(3), 449–467. Alm, J. R., & Barreto, R. A. (2003). Corruption, Optimal Taxation and Growth. Public Finance Review, 31(3), 207–240. Anderson, J. E., & Marcouiller, D. (2002). Insecurity and the Pattern of Trade: An Empirical Investigation. Review of Economics and Statistics, 84(2), 342–352.

156     K. Sekkat

Anderson, C. J., & Tverdova, Y. V. (2003). Corruption, Political Allegiances and Attitudes Toward Government in Contemporary Democracies. American Journal of Political Science, 47(1), 91–109. Azfar, O., & Gurgur, T. (2008). Does Corruption Affect Health Outcomes in the Philippines? Economics of Governance, 9(3), 197–244. Bailey, D. H. (1966). The Effects of Corruption in a Developing Nation. Western Political Quarterly, 19(4), 719–732. Bardhan, P. (1997). Corruption and Development: A Review of Issues. Journal of Economic Literature, 35(3), 1320–1346. Battese, G. E., & Coelli, T. J. (1995). A Model for Technical Inefficiency Effects in a Stochastic Frontier Production Function for Panel Data. Empirical Economics, 20(2), 325–332. Beck, P. J., & Maher, M. W. (1986). A Comparison of Bribery and Bidding in Thin Markets. Economics Letters, 20(1), 1–5. Bertrand, M., Djankov, S., Hanna, R., & Mullainathan, S. (2007). Obtaining a Driver’s License in India: An Experimental Approach to Studying Corruption. Quarterly Journal of Economics, 122(4), 1639–1676. Boehm, F., & Joerges J. (2008). Costs of Corruption: Everyone Pays—and the Poor More Than Others, Division State and Democracy Supporting the Implementation of the UN Convention Against Corruption (UNCAC). https://www.Giz.De/Fachexpertise/Downloads/Gtz2008-En-GermanUncac-Project-Costs-Of-Corruption.Pdf. Accessed 18 July 2017. Campos, J. E., Lien, D., & Pradhan, S. (1999). The Impact of Corruption on Investment: Predictability Matters. World Development, 27, 1059–1067. Clausen, B., Kraay, A., & Nyiri, Z. (2011). Corruption and Confidence in Public Institutions: Evidence from a Global Survey. World Bank Economic Review, 25(2), 212–249. De Rosa, D., Gooroochurn, N., & Görg, H. (2015). Corruption and Productivity: Firm-Level Evidence. Jahrbücher Für Nationalökonomie Und Statistik, 235(2), 115–138. Dreher, A., & Gassebner, M. (2013). Greasing the Wheels? The Impact of Regulations and Corruption on Firm Entry. Public Choice, 155(3–4), 413–432. Dridi, M. (2014). Corruption and Education: Empirical Evidence. International Journal of Economics and Financial Issues, 4(3), 476. Ehrlich, I., & Lui, F. T. (1999). Bureaucratic Corruption and Endogenous Economic Growth. Journal of Political Economy, 107(S6), S270–S293.

4 Consequences     157

Gyimah-Brempong, K. (2002). Corruption, Economic Growth and Income Inequality in Africa. Economics of Governance, 3(3), 183–209. Habib, M., & Zurawicki, L. (2001). Country-Level Investments and the Effect of Corruption—Some Empirical Evidence. International Business Review, 10(6), 687–700. Hakhverdian, A., & Mayne, Q. (2012). Institutional Trust, Education and Corruption: A Micro-Macro Interactive Approach. Journal of Politics, 74(3), 739–750. Huang, F. L. (2008). Corruption and Educational Outcomes: Two Steps Forward, One Step Back. International Journal of Education Policy and Leadership, 3(9), 1–10. Hunt, J. (2007). How Corruption Hits People When They Are Down. Journal of Development Economics, 84(2), 574–589. Huntington, S. P. (1968). Political Order in Changing Societies. New Haven: Yale University Press. Jain, A. K. (2001). Corruption: A Review. Journal of Economic Surveys, 15(1), 71–121. Jain, A. K. (Ed.). (2002). The Political Economy of Corruption (Vol. 2). London: Routledge. Kaufmann, D., & Wei, S. J. (1999). Does “Grease Money” Speed Up the Wheels of Commerce? (National Bureau of Economic Research, WP 7093). Klomp, J., & De Haan, J. (2008). Effects of Governance on Health: A CrossNational Analysis of 101 Countries. Kyklos, 61(4), 599–614. Koyuncu, C., & Yilmaz, R. (2009). The Impact of Corruption on Deforestation: A Cross-Country Evidence. Journal of Developing Areas, 42(2), 213–222. Kurer, O. (1993). Clientelism, Corruption and the Allocation of Resources. Public Choice, 77(2), 259–273. Lambsdorff, J. G. (1998). An Empirical Investigation of Bribery in International Trade. European Journal of Development Research, 10(1), 40–59. Lambsdorff, J. G. (2003a). How Corruption Affects Productivity. Kyklos, 56(4), 457–474. Lambsdorff, J. G. (2003b). How Corruption Affects Persistent Capital Flows. Economics of Governance, 4(3), 229–243. Leff, N. H. (1964). Economic Development Through Bureaucratic Corruption. American Behavioral Scientist, 8(3), 8–14.

158     K. Sekkat

Levin, M., & Satarov, G. (2000). Corruption and Institutions in Russia. European Journal of Political Economy, 16(1), 113–132. Leys, C. (1965). What Is the Problem about Corruption? The Journal of Modern African Studies, 3(2), 215–230. Li, H., Xu, L. C., & Zou, H. F. (2000). Corruption, Income Distribution and Growth. Economics and Politics, 12(2), 155–182. Lien, D. H. D. (1986). A Note on Competitive Bribery Games. Economics Letters, 22(4), 337–341. Lui, F. T. (1986). A Dynamic Model of Corruption Deterrence. Journal of Public Economics, 31(2), 215–237. Mahagaonkar, P. (2008). Corruption and Innovation: A Grease or Sand Relationship? Jena Economic Research Papers, 2008, 017. Mankiw, G., & Whinston, M. (1986). Free Entry and Social Inefficiency. Rand Journal of Economics, 17(1), 48–58. Mauro, P. (1995). Corruption and Growth. Quarterly Journal of Economics, 110(3), 681–712. Mauro, P. (1998). Corruption and the Composition of Government Expenditure. Journal of Public Economics, 69(2), 263–279. Méon, P. G., & Sekkat, Kh. (2004). Does the Quality of Institutions Limit the MENA’s Integration in the World Economy? The World Economy, 27(9), 1475–1498. Méon, P. G., & Sekkat, Kh. (2005). Does Corruption Grease or Sand the Wheels of Growth? Public Choice, 122(1), 69–97. Méon, P. G., & Sekkat, Kh. (2008). Institutional Quality and Trade: Which Institutions? Which Trade? Economic Inquiry, 46(2), 227–240. Méon, P. G., & Weill, L. (2010). Is Corruption an Efficient Grease? World Development, 38(3), 244–259. Mo, P. H. (2001). Corruption and Economic Growth. Journal of Comparative Economics, 29(1), 66–79. Musila, J. W., & Sigué, S. P. (2010). Corruption and International Trade: An Empirical Investigation of African Countries. World Economy, 33(1), 129–146. Myrdal, G. (1968). Asian Drama: An Enquiry into the Poverty of Nations. London: Alien Lane, The Penguin Press. Olson, M., Sarna, N., & Swamy, A. V. (2000). Governance and Growth: A Simple Hypothesis Explaining Cross-Country Differences in Productivity Growth. Public Choice, 102(3–4), 341–364.

4 Consequences     159

Restuccia, D., & Rogerson, R. (2008). Policy Distortions and Aggregate Productivity with Heterogeneous Establishments. Review of Economic Dynamics, 11(4), 707–720. Rose-Ackerman, S. (1997). The Political Economy of Corruption. In Elliott, K. A. (Ed.), Corruption and the Global Economy (pp. 31–60). Washington, DC: Peterson Institute. Shleifer, A., & Vishny, R. W. (1993). Corruption. Quarterly Journal of Economics, 108(3), 599–617. Stockemer, D., Lamontagne, B., & Scruggs, L. (2013). Bribes and Ballots: The Impact of Corruption on Voter Turnout in Democracies. International Political Science Review, 34(1), 74–90. Tanzi, V., & Davoodi, H. (1998). Corruption, Public Investment and Growth. In the Welfare State, Public Investment and Growth: Selected Papers from the 53rd Congress of the International Institute of Public Finance (Vol. 53, p. 41). Springer Science and Business Media. Uneke, O. (2010). Corruption in Africa South of the Sahara: Bureaucratic Facilitator or Handicap to Development? Journal of Pan African Studies, 3(6), 111–129. Wei, S. J. (2000a). How Taxing is Corruption on International Investors? Review of Economics and Statistics, 82(1), 1–11. Wei, S. J. (2000b). Does Corruption Relieve Foreign Investors of the Burden of Taxes and Capital Controls? In Hines, J. R. (Ed.), International Taxation and Multinational Activity (pp. 73–88). Chicago: University of Chicago Press.

Part II Anti-corruption Strategies: The Role of the State

The discussion in the previous chapters has shown that corruption constitutes a major obstacle to economic, social, and political development as well as to cohesion within countries. These considerations provide ample justification for a resolute fight against corruption. At the same time, the analysis has revealed the complexity of the phenomenon and its causes. It follows that a realistic strategy to tackle corruption should recognize and take account of such complexity and of the different channels of influence. Corruption involves those who demand certain deserved or undeserved actions, those who are willing to accomplish these acts in return for unjustifiable compensation, along with intermediaries and third parties, some of whom are interested in seeing corruption fought, while others are interested in seeing it perpetuated. The relative power of these participants and the outcome of their interactions depend on the cultural, political, social, institutional, and economic architecture of each country and on the functioning of this architecture. For instance, some blame the state which, through the monopoly it has on issuing laws, regulation or licensing, creates opportunities that lead to corruption. In contrast, others consider that the state by itself is not the main initiator

162     Part II: Anti-corruption Strategies: The Role of the State

of corruption and that the blame should instead be put on the prevailing incentive system as the main motive for corruption (Tanzi 1998). Notwithstanding these opposing views, the consensus is that corruption will be more or less widespread depending on the economic, political, social, and value-system features of the country (Bardhan 2006). Although corruption can take place in the private sector, it is traditionally associated with the public sector. It is then referred to as the use of the position held by an individual in the public sector for private gains. The reference to such position shows the need for checks to fight corruption and thus raises the question of the political regime. In an autocratic regime, a large part of public life is under the control of the leader(s) and high-ranking officials, with very few checks in place. Under such regimes, the tastes and the wishes of a small coalition (i.e., leaders and their companions) determine the conduct of public affairs with the aim of fulfilling the coalition’s own interests. This includes organizing the market for corruption. To support the coalition’s behavior, the regime generally tolerates the fact that low-ranking state servants use their public function for private gains. The absence of a strong and credible system of checks and balances makes the fight against such abuse almost impossible. Accordingly, a necessary, although not sufficient, condition for success in the fight against corruption is the existence of a system of checks and balances, to which we refer here as democracy. In contrast to autocracy, democracy allows citizens to voice their opinions and participate in setting the rules which govern their lives and play against corruption. However, even in democratic regimes, corruption is not absent. First, the rules which were approved by a large majority of citizens might not correspond to other citizens’ interests. This gives unhappy citizens an incentive to circumvent the rules even at some expense. Second, implementation of the rules is often delegated to public servants and might be subject to interpretation or arbitrariness. Controlling and monitoring these public servants can be very difficult and costly due to their dispersion across the country, lack of skilled monitors, or the existence of systemic coalitions. Third, information asymmetry and opportunism can allow elected elites to issue laws and regulations that are mainly targeted toward their own interest. This

Part II: Anti-corruption Strategies: The Role of the State      163

might facilitate capture and corruption and other uses of public office for private gains. It appears that under both democratic and autocratic regimes, the issuing and implementation of laws and regulations are at the heart of the debate. A simplistic suggestion would be to abolish or severely constrain the issuing of regulations. While it is true that senseless regulations can be found in any country around the world, many regulations have well-founded economic, political, or social justifications. Abolishing such regulations could result in worse outcomes for the country. What is needed are mechanisms to discourage unscrupulous participants. Such mechanisms may take the form of punishment for infringement, reward for compliance, contestable market power, and so on. Punishment raises a number of issues that are not easy to deal with. These include whom to punish (briber, recipient, intermediaries, or everyone), the character and the severity of the punishment, and the cost of trial, which includes collection of evidence, potential conviction of the defendant, and enforcement of the verdict. These tasks may involve a number of judges, lawyers, police representatives, and auditors. This is problematic in many developing countries, where such bodies are weak and often corrupt themselves. More importantly, Tanzi (1998) and Bardhan (2006) have argued that in these countries, the poor might be better off thanks to corruption. For instance, some of these countries have policies involving subsidies of food for the poor. Unsurprisingly, the distribution of food by civil servants is often impacted by corruption and leakage. However, without such subsidies, the poor will be left to the market, where the price may increase to exceed the total cost of the public subsidy plus the bribe. In this case, not punishing civil servants results in lower food costs for the poor. When punishment poses problems, other mechanisms can be effective and less costly in reducing corruption. These mechanisms include reward for compliance, contestable market power, and citizen monitoring. Reward may consist in offering higher wages to civil servants. This is supposed to reduce their incentives to ask for bribes. Badly paid civil servants often ask for bribes to secure decent living conditions. However, this solution might not work if these employees consider the increase in wages as top-up revenue instead of as a substitute

164     Part II: Anti-corruption Strategies: The Role of the State

for corruption. Alternatively, since the ability to request or offer bribes comes chiefly from the “monopoly” power a civil servant has on certain tasks, another solution consists in breaking up this power. This can be achieved either by making the task equally accomplishable by different employees or by rotating the civil servants in charge of the task. A variant of this solution consists in transferring some tasks to the private sector, which may improve service delivery and reduce corruption. However, if the private sector firm engaged in service delivery is left without control, the door will be reopened to corruption. Separate chapters will tackle the different solutions discussed above. Each chapter starts with a brief review of the arguments for a given solution before presenting in more detail the empirical findings regarding its effectiveness. We draw on laboratory experiments, case studies, and micro- and macro-econometric analyses. We consider the different sources of evidence as complementary rather than as substitutes and see them as highlighting different dimensions of the phenomenon. The remainder of this part is composed of eight chapters. Chapters 5–7 examine the relationship between corruption and, respectively, democracy, electoral rules, and decentralization. Chapters 8–10 discuss, respectively, the role of regulation, justice, and specialized AntiCorruption Agencies (ACA). Finally, Chapters 11 and 12 focus on changing the incentives of civil servants and international cooperation.

References Bardhan, P. (2006). The Economist’s Approach to the Problem of Corruption. World Development, 34(2), 341–348. Tanzi, V. (1998). Corruption Around the World: Causes, Consequences, Scope and Cures. IMF Staff Papers, 45(4), 559–594.

5 Democracy

This chapter discusses the effects of democracy on corruption. Autocratic regimes are in general associated with a high prevalence of corruption while democratic regimes seem much less affected by corruption. The reason why democracies are less corrupt is that elections increase the probability of corrupt officials being kicked out. Moreover, effective checks and balances under democracy increase the probability of corrupt acts being disclosed. Finally, transparency erodes the rent associated with being close to power. Tests of these arguments have been conducted, but the results are mixed. While there is a consensus about the anti-corruption effects of democracy, the empirical evidence suggests that the magnitude of the effects depends on various factors, such as the age and the degree of maturity of democracy as well as electoral rules (majority or proportional systems, presidential, parliamentary or federal approaches).

© The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_5

165

166     K. Sekkat

1 Expected Impacts of Democracy on Corruption As explained in the introduction, corruption is often associated with the power that a person acquires when holding a specific position in the public sector. The extent of such power depends both on the position of the person and on the political regime of the country. Autocratic regimes are typically personalized, and the ruler has very extensive power. Although in such regimes public opinion could still have an impact, divisions within countries along economic, social, or ethnic classes provide the leader with significant autonomy. In this context, the ruler can maintain considerable power by relying on narrow group of citizens to which we refer as the elite. Although some exceptions may exist (e.g., Singapore, South Korea, or Taiwan some years ago), the ruler and the elite are mainly concerned with their own interest rather than with broad-based wealth. Accordingly, the autocratic leader needs to satisfy only this narrow coalition. In contrast, a democratic leader needs a majority in a larger constituency and is therefore accountable to far more citizens. For this reason, a democratic leader has to obtain broad public satisfaction through, among other things, well-functioning institutions. Accountability and the potential role of well-functioning institutions in improving growth and equity make the democratic leader much more averse to corruption than the autocrat. The net impact of democracy on corruption is disputed even at the theoretical level. Democracy is expected to reduce corruption for several reasons. First, the election process increases the probability of corrupt acts being disclosed and of corrupt officials being punished. On the one hand, the opposition has an incentive to uncover and publicize corrupt activities by the incumbent. On the other hand, voters have an interest in not reelecting politicians that favor their own private interests over those of the electorate. Second, democracy creates a more open system of government. This implies less information asymmetry on how the system works, which promotes a decrease in the rent associated with being close to power. Third, effective checks and balances within government will constrain the ability of officials to deviate from impartial

5 Democracy     167

practices. Finally, democratic culture goes, in general, with a negative view of corruption in society. As a result, corrupt activities become repellant because they involve greater stigma. Democracy can also have a negative impact on corruption. First, election campaigns require funding, and more competitive elections may make political parties and candidates vulnerable to pressure from funders (Rose-Ackerman 1999). Second, even a rational and informed voter may choose to vote for a corrupt government for strategic reasons (Pani 2011). Third, the effect of a more open government is also ambiguous. Bac (2001) argues that transparency makes it easier to identify which official to bribe, and shows that this effect may dominate a corruption detection effect. Fourth, institutions in charge of enforcing government accountability are often appointed or funded by the government itself. This can reduce their incentive and capacity to expose government corruption. In some cases, these institutions have been used to persecute political opponents of the government, rather than to hold the government accountable. The effect of democracy on corruption also depends on the degree of maturity of a democracy. In its early stages, a democracy might suffer from weak checks and balances and insufficient transparency. In this context, rent-seekers can continue collecting bribes with a low risk of being under public scrutiny. As transparency and accountability become more effective, the cost of rent-seeking (including the probability of getting caught and punished) increases and, as a result, aggregate rents and corrupt activity are likely to decline. Following Rock (2009), this implies an inverted-U relationship between corruption and the durability of democracies.

2 Actual Impacts of Democracy on Corruption The above conceptual analysis suggests that democracy may reduce or increase corruption depending on a number of factors. In spite of the importance of the question, very few studies have examined the effect

168     K. Sekkat

of democracy per se on corruption. A potential reason for this limited number of investigations may be the common belief that democracy indisputably reduces corruption. Another reason is that both democracy and corruption are multifaceted processes and affect each other in numerous and sometimes conflicting manners. Meanwhile, as pointed out by Sung (2004), numerous corrupt practices re-emerged following political liberalization in Southeast Asia, Latin America, and the former Soviet republics. In fact, the few empirical analyses examining the impact of democracy per se on corruption provide mixed evidence. Other more numerous studies have investigated the factors behind such lack of consensus. In general, these studies point to the nonlinearity of the effect of democracy on corruption or to the dependence of this effect on other variables. Akarca and Tansel (2016) focus on how the government of Turkey dealt with the consequences of the two major earthquakes which struck northwestern Turkey in 1999 to examine how voters exposed and punished politicians. After the quakes, a lot of old buildings remained standing but many of the recently constructed ones collapsed due to numerous violations of construction rules and zoning codes. For many observers, these violations and the subsequent mismanagement related to relief were poisoned by corruption. To determine whether and how voters punished the elected members of the parties that were responsible for construction of the second-rate buildings and for the disappointing management of relief efforts, the paper uses the results of elections following the earthquakes. A vote equation is estimated for each of the country’s seven major political parties. The vote share of party i, in province j, in the election held in year t is explained in terms of the same share but in past elections, the provincial mean years of schooling of the population over age six, the proportion of the urban population in the province, the proportion of women in non-agricultural employment in the province, the provincial growth rate of per capita real GDP, and the number of residences and business places in the province having suffered heavy damage during the earthquakes. The years of observation are those of the two successive elections of 1999 and 2002. The results show a general shift in votes between the two elections from the ruling parties in favor of the

5 Democracy     169

non-ruling Islamic party. The shift was more pronounced in provinces which suffered heavy earthquake damage. In addition to the ruling parties since 1991, other parties which were in power when the second-rate buildings were built were held accountable by the electorate. This voter reaction seems to have been one of the factors that allowed the Islamic party (established only a year before) to capture a large majority of votes in the country. The preceding example may lead to a highly optimistic view regarding the impact of democracy on corruption. Without refuting the importance of democracy, various papers mitigate its impact on corruption. For instance, Paldam (2002) and Goel and Nelson (2005) emphasize the role of economic freedom in completing the impact of democracy on corruption. Iwasaki and Suzuki (2012) put forward the difference between old and young democracies. Bhattacharyya and Hodler (2010) and Henderson and Kuncoro (2011) investigate two under-examined issues, namely the role of natural resources and religion, respectively. Finally, other papers such as Rock (2009), Billger and Goel (2009), and Kolstad and Wiig (2016) focus on more technical but very important problems which might affect estimations of the impact of democracy on corruption. Paldam (2002) examines the impact of economic, institutional, and cultural factors in curbing corruption as measured by the 1999 corruption index from Transparency International (TI). The economic variables are the level and growth of real income per capita, the inflation rate, and the economic freedom index. The cultural variables use a set of dummies for “cultural areas”, while the institutional variable is the Gastil Democracy Index. The results show that more democracy is associated with less corruption. More generally, the transition of a country toward a wealthy liberal democracy goes with a dramatic reduction in the level of corruption. Goel and Nelson (2005) focus on the impact that economic and political freedoms have on the prevalence of corruption in 63 developing and developed countries in 2000. TI’s Corruption Perceptions Index (CPI) is used as the measure of corruption, the Heritage Foundation’s index is used as a measure of economic freedom, and democracy is measured by the Gastil Index of Civil Liberties and Political Rights in

170     K. Sekkat

the country. Other explanatory variables include educational attainment, per capita GNP, and the black market premium. The results show that less democratic countries are associated with higher corruption. It also appears that economic freedom may serve as a deterrent to corruption and that such effect is stronger than the effect of democracy. In other words, greater economic freedom complements democracy as a deterrent to corrupt activities. Among the different components of economic freedom, monetary policy seems to have a strong influence on the level of corrupt activity in a country. Countries with less regulated financial sectors and those with smaller black markets are likely to be less corrupt. The transition of many European countries from communist authoritarian regimes to more market-oriented democratic ones offered Iwasaki and Suzuki (2012) a useful way to test the relationship between corruption and democracy. Their analysis used panel data for the period 1996–2006 covering 202 countries throughout the world. It regressed the World Bank’s Control of Corruption Index (CCI) on a dummy variable assigning a value of 1 to transition countries and 0 otherwise and a set of control variables including marketization, rule of law, GDP per capita, the construction industry’s share of GDP, and the Protestant population share. The results show that the extent of corruption control in former socialist states decreased when compared with the trend for the whole world, even after considering the economic development level of the countries. However, the results also indicate that a noticeable difference in the degree of corruption arose among transitional countries themselves. The lack of a democracy effect found here might be due to the fact that such democracies are still “young”. As discussed in Chapter 3, Treisman (2000) finds that while the current degree of democracy is not significant in explaining corruption, a long period of exposure to democracy is. Bhattacharyya and Hodler (2010) study the interplay between democracy, natural resources, and corruption using panel data covering the period from 1980 to 2004 and 124 countries. The intuition behind the research question comes from a game between politicians and the people. Politicians may primarily care about social welfare and/or about the revenues they can generate through corrupt activities. Citizens care

5 Democracy     171

about their own interest and hence prefer politicians who act accordingly. Incumbent politicians therefore have an incentive to avoid corruption in order to improve their chances of remaining in power. However, such mechanisms will work only if politicians are accountable and citizens can effectively kick politicians out of office. As a result, the level of corruption in the presence of abundant natural resources will depend on the strength of democracy in the country. In empirical terms, this can be tested by introducing an interaction term for democracy and natural resources in the equation linking corruption to natural resources. If the story fits, the results should show a threshold level of democracy below which the effect of resource abundance on corruption is more corruption and above which the effect of resource abundance is less corruption. The ICRG corruption index is regressed on a measure of natural resource rents, the Polity 2 Democracy Index and the interaction of these two variables. The natural resource rent measure is the log per capita rent from natural resources including energy, minerals, and forestry. It is computed using the difference between the world price of a commodity and the average extraction costs (both expressed in current US dollars). Control variables include the log per capita income, legal origin dummies, ethnic fractionalization, and the black market premium. The estimates confirm that the relationship between resource rents and corruption depends on the degree of democracy. Resource rents are positively associated with corruption in countries that have a Polity 2 score of 8.5 or less for an index varying between −10 (lowest level of democracy) and 10 (highest level of democracy). These findings imply that resource-rich countries have a tendency to be corrupt because resource abundance encourages their governments to engage in rent-seeking. This tendency can be countered if governments are accountable to the people. However, given the necessary threshold, the effect is likely to appear in highly democratic countries such as Australia or Norway. Indonesia is a country with a tradition of corruption among local officials who harass and collect bribes from firms. In 1999, the first free elections took place and involved five major political parties, including the long-standing secular parties led by Muhammad Suharto and

172     K. Sekkat

Megawati Sukarnoputri, respectively. At the national level, these elections led initially to a coalition government between the Megawati party and the main Islamic party. In 2001, the national coalition collapsed and Megawati took over as president. At the local level, her secular party often allied with the other major secular party. Megawati held office until the end of the first election cycle in 2004. Both the 1991 elections and the change in coalitions after 2001 introduced local democratization and decentralization. Henderson and Kuncoro (2011) examine the impact of these changes on corruption and on the role of religious vs secular parties. The base of the study is a survey of firms carried out in early 2005 across all districts of Java. The survey focuses on bribe activities in previous years and is assumed to reflect the influence of the composition of the assemblies elected in 1999. The study also uses the results of another survey conducted in 2001 and covering one-third of the districts of Java. This second survey asked about corruption in 2001. The survey for 2004 covered 2707 firms, all in manufacturing in Java (97 districts). The 2001 survey was a random sample of 1808 enterprises spread over all economic sectors in 64 districts of Indonesia. In both surveys, a question specifically concerned the fraction of costs devoted to bribes. This is used as the dependent variable and explained in terms of the share of the two major secular parties in the assembly elected in 1999. Control variables include GDP per capita, changes in religiosity (measured as the ratio of Islamic to state elementary schools) and whether the owner of the firm is a Chinese–Indonesian (traditionally subject to more harassment). The results show that the introduction of local democracy is associated with decreased local corruption. However, as local shares of secular parties rise, the relative degree of corruption rises. The baseline results suggest that a secular party share increase of 10% raises the bribe ratio by 1.2. The effect is non-negligible given that the mean of the bribe ratio is 3.6. As the authors point out, the results should not be interpreted as an indication about religion per se but rather as a sign that the entry of new parties seeking political support is good for anti-corruption concerns. Many papers relevant for our purpose assume a linear relationship between democracy and corruption. If this assumption is not valid, the

5 Democracy     173

conclusions we can draw from these studies might be worthless. Nonvalidity of the assumption may also explain the high heterogeneity in impacts of democracy on corruption across countries. Rock (2009) argues that the relationship between corruption and democracy has the form of an inverted-U. As explained above, young democracies might suffer from weak checks and balances and insufficient transparency. These provide corrupt people with greater opportunities for collecting bribes without being under the effective threat of public investigation. As the institutions of transparency and accountability develop, corrupt activities decrease because the cost of corruption (including the probability of getting caught and punished) increases. If well-founded, this mechanism implies an inverted-U relationship between corruption and the maturity of democracies. Rock (2009) investigates whether such an inverted-U relationship exists using the World Bank’s (WB) CCI as a dependent variable. Three measures pertaining to democracy—Institutionalized democracy, Institutionalized autocracy, and Durability of democracy—are used and come from Polity V. Control variables include the WB Government Effectiveness and Rule of Law indexes, the log of real GDP per capita, and an ethnolinguistic fractionalization index. The panel dataset covers around 100 countries between 1996 and 2003. To test for the existence of the inverted-U relationship, the variables related to democracy are introduced and squared. The results show strong support for the inverted-U relationship between corruption and durability of democracy. The turning point in the relationship seems to occur between 4 and 15 years, which is a relatively young age. The conclusion is that some low-income countries might be able to reduce corruption relatively quickly after opting for democracy. Billger and Goel (2009) focus on a different type of nonlinearity than the preceding paper. While Rock (2009) investigates whether the impact on corruption depends on the level of the explanatory variable, Billger and Goel (2009) examine whether the impact on corruption varies with the level of the dependent variable itself, that is, corruption. The idea is that there may be subtle institutional differences between corrupt and “clean” countries that might lead to differences in the impact of the determinants and in the efficacy of anti-corruption

174     K. Sekkat

policies. The authors use quantile regression techniques, which make it possible to investigate whether the relationship between corruption and the explanatory variables varies over the distribution of the dependent variable. The data include cross-sectional observations on 99 countries from 2001 to 2003. The dependent variable is the CPI. Democracy is the sum of the Freedom House Political Rights and Civil Liberties Indexes. Control variables include economic freedom from the Heritage Foundation, real GDP per capita, general government final consumption expenditure (% of GDP), and the urban population (% of total). The results support the previous findings in the literature but also reveal differences across quantiles. Democracy significantly decreases corruption irrespective of the quantile. However, the magnitude of the effect depends on the quantile. The decrease in corruption due to democracy is much higher in the most corrupt nations than in the least corrupt. Greater economic prosperity (i.e., real GDP per capita) also reduces corruption in all cases. The effect of greater economic freedom is not statistically significant in any of the quantiles. A complementary analysis by Saha and Su (2012) using a very similar approach to Billger and Goel (2009) adds the finding that democracy may not be effective in combating corruption at a low level of economic freedom (for countries at any level of corruption). In contrast, democracy becomes effective in curing corruption for the most corrupt countries when the level of economic freedom is high. Another important technical issue that may affect the relationship between democracy and corruption is simultaneity. On the one hand, both democracy and corruption are likely to be affected by other variables that are hard to observe or quantify, such as culture. On the other hand, there may be a reverse causality by which corruption affects democracy. For instance, corruption may undermine the confidence of voters in the democratic system and hence trigger reversals. According to Kolstad and Wiig (2016), these technical issues might explain the conflicting findings with regard to the relationship between democracy and corruption. The authors address this challenge by using an instrumental variables approach. They instrument democracy using a dummy variable

5 Democracy     175

reflecting whether a country has been at war with a democracy in the period 1946–2009, while controlling for the extent to which countries have been at war in general. A dummy variable is used to instrument the level of democracy in 2008. The idea behind the instrument is that democracies rarely go to war against each other. Controlling for conflict in general, conflict with democracies should therefore be a valid instrument. As measures of the dependent variable, the authors use two different indexes: The World Bank’s CCI and Transparency International’s CPI. As measures of democracy, they use the Polity IV Democracy Index and the Freedom House Political Rights Index. Control variables include the log of GDP per capita, the legal origin of countries, labor participation rates, and the proportion of Catholics in the country. For variables other than those related to conflicts, data are taken from the year 2008. The results suggest that democracy may be more effective in reducing corruption than indicated by estimates not taking the endogeneity of democracy into account. Accordingly, there may be a substantial effect from improving democracy in developing countries, where the problem of corruption is most prevalent. The authors do not, however, address the issue of heterogeneity in impacts of democracy on corruption across countries.

3 Conclusion Although democracy contributes to reducing corruption, it cannot eradicate it. A country’s democratic status is not sufficient to deter corruption. A number of studies point to the role of other variables in determining the relationship between corruption and democracy. These include the extent of economic freedom, the age and maturity of democracies, natural resources, and culture. Others studies focus on more technical but very important problems which might affect the estimation of the impact of democracy on corruption. The evidence suggests that the current degree of democracy is not significant in explaining corruption, while a long period of exposure to democracy is. Moreover, economic freedom seems to be a stronger deterrent to corruption than democracy alone. Countries with less

176     K. Sekkat

regulation and those with smaller black markets are likely to be less corrupt. Resource-rich countries have a tendency to be corrupt, which implies that the degree of accountability and monitoring needed to counter this tendency is higher than in countries with no natural resources. Regarding the technical issues, it seems that the assumption of a linear relationship between corruption and democracy hinders the identification of links between the two variables. Some findings suggest that the relationship takes the form of an inverted-U. One reason is that young democracies might suffer from accountability and transparency issues, which favor corruption because of the low likelihood of being exposed. When democracies mature, transparency and accountability develop and corrupt activities decrease because of the higher probability of getting caught and punished. Another technical problem that may affect the link between democracy and corruption is endogeneity. Both democracy and corruption may be affected by other variables, which can lead to the identification of spurious relationships. Moreover, the relationship between corruption and democracy might be bidirectional. Corruption may undermine the confidence of voters in the democratic system and thus generate an autocratic reversal, which in turn creates opportunities for corruption.

References Akarca, A. T., & Tansel, A. (2016). Voter Reaction to Government Incompetence and Corruption Related to the 1999 Earthquakes in Turkey. Journal of Economic Studies, 43(2), 309–335. Bac, M. (2001). Corruption, Connections and Transparency: Does a Better Screen Imply a Better Scene? Public Choice, 107(1–2), 87–96. Bhattacharyya, S., & Hodler, R. (2010). Natural Resources. Democracy and Corruption. European Economic Review, 54(4), 608–621. Billger, Sh. M., & Goel, R. K. (2009). Do Existing Corruption Levels Matter in Controlling Corruption? Cross-Country Quantile Regression Estimates. Journal of Development Economics, 90(2), 299–305.

5 Democracy     177

Goel, R. K., & Nelson, M. A. (2005). Economic Freedom Versus Political Freedom: Cross-Country Influences on Corruption. Australian Economic Papers, 44(2), 121–133. Henderson, J. V., & Kuncoro, A. (2011). Corruption and Local Democratization in Indonesia: The Role of Islamic Parties. Journal of Development Economics, 94(2), 164–180. Iwasaki, I., & Suzuki, T. (2012). The Determinants of Corruption in Transition Economies. Economics Letters, 114(1), 54–60. Kolstad, I., & Wiig, A. (2016). Does Democracy Reduce Corruption? Democratization, 23(7), 1198–1215. Paldam, M. (2002). The Cross-Country Pattern of Corruption: Economics, Culture and the Seesaw Dynamics. European Journal of Political Economy, 18(2), 215–240. Pani, M. (2011). Hold Your Nose and Vote: Corruption and Public Decisions in a Representative Democracy. Public Choice, 148(1–2), 163–196. Rock, M. T. (2009). Corruption and Democracy. Journal of Development Studies, 45(1), 55–75. Rose-Ackerman, S. (1999). Political Corruption and Democracy. Conn. J. Int’l L., 14, 363–378. Saha, S., & Su, J. J. (2012). Investigating the Interaction Effect of Democracy and Economic Freedom on Corruption: A Cross-Country Quantile Regression Analysis. Economic Analysis & Policy, 42(3), 389–396. Sung, H. E. (2004). Democracy and Political Corruption: A Cross-National Comparison. Crime, Law and Social Change, 41(2), 179–193. Treisman, D. (2000). The Causes of Corruption: A Cross-National Study. Journal of Public Economics, 76(3), 399–457.

6 Electoral Rules

Under democratic regimes, the disciplinary role of elections may be inhibited because voters might reelect a corrupt government for strategic or partisan reasons. In addition, election campaigns require funding, which makes political parties and candidates vulnerable to pressure from contributors. Finally, institutions in charge of enforcing government accountability are often appointed or funded by the government itself, which hampers their capacity to expose government corruption. This chapter reviews the factors which limit the impact of democracy on corruption. The limitation of the disciplinary role of democracy on corruption depends on the functioning of two mechanisms. One is based on competition, while the other draws on career concerns. Competition makes it possible to contest the power of a politician. Career concerns place self-restraints on a politician. The following analysis is organized around these two mechanisms and explicitly discusses the financing of electoral campaigns under different democratic settings, the voting behavior of the different participants, and the role of opponents in exposing incumbents’ corrupt practices.

© The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_6

179

180     K. Sekkat

1 Expected Impacts of Electoral Rules on Corruption As discussed in the preceding chapter, while there are several reasons for democracy to reduce corruption, there are also arguments suggesting that democracy can increase corruption. Moreover, the empirical evidence is mixed. One possible reason for such mixed evidence is that the corruption indexes used do not make it possible to distinguish between petty and grand corruption. The former generally concerns small payments to low-ranking officials in exchange for services that citizens are entitled to in any case. The latter concerns large payments to high-ranking officials and ministers in exchange for influencing or changing regulations and other state rules. While petty corruption is the most visible and exasperating to ordinary citizens, grand corruption has potentially higher adverse economic impacts. In developing countries, it is almost impossible to find someone who has never been directly or indirectly (through parents or other relatives) asked for some kind of payment to get things done (petty corruption). In developed countries, this kind of corruption is very rare. However, it is equally difficult to find a developed country where no politician or high-ranking official has been convicted of taking some money from firms or other “benefactors”. Accordingly, in developed democracies, the main issue is with grand corruption. One of the most frequently cited drivers of corruption in this context is elections. Election campaigns require, in general, huge amounts of funds. This makes political parties and candidates vulnerable to pressure from funders (Rose-Ackerman 1999), especially if elections are competitive (Anduiza et al. 2013). Another possible reason why democracy does not reduce corruption is that, as observed by (Rose-Ackerman 2006), in some countries citizens continue voting for corrupt candidates. This may be due to the institutional system or to voters’ tastes. For instance, a party that uncovers a scandal might be rewarded at the polls for its vigilance and integrity. However, the incentive to reveal scandals also depends on the cost of exposing malfeasance, which may be high not only in terms of money but also in terms of future collaborations, as scandalmongers

6  Electoral Rules     181

may be punished by being excluded from future coalitions. This may give everyone an incentive to keep quiet in the face of scandals involving others. In addition, corrupt politicians might not be kicked out because of partisan attitudes toward corruption (Anduiza et al. 2013). For instance, a survey carried out in Spain, a country that has seen a non-negligible number of corruption scandals, showed that the same infringement is judged differently depending on whether the responsible politician is a member of the respondent’s party or not. In sum, in democratic systems too, politicians may engage in corruption irrespective of whether they are in office or seeking office. While the above discussion highlights some of the reasons for corruption, there are many other features of democracies which can make some of democracies more corrupt than others. According to Rose-Ackerman (2006), these relate to the design of democratic institutions, such as majority or proportional systems, presidentialism, parliamentarism, federalism, bicameralism, and so on. Majority systems mean that a single member of a party is elected by district, while proportional systems mean that more members potentially from different parties are elected per district. Under majority systems, voters elect a single representative from their district of residence. The proportional system includes two variants: closed list and open list. Under a closed-list system, party leaders rank candidates and voters only vote for parties. Candidates are elected following the order established by the party up to the limit of the party’s score. Under an open-list system, voters both select a party and rank candidates based on the party’s selection of candidates. Analyses of the effects of political institutions on corruption are mainly based on two strands of economic theory. These are competition, which makes it possible to contest the power of a politician, and career concerns, which place self-restraints on a politician. Theories based on political competition suggest that electoral rules that lower barriers to entry into political competition should be better at limiting corruption. In contrast, theories based on career concerns suggest that the fact that opportunities for corruption are concentrated among a handful of political players enables better monitoring by voters and should constrain corruption. According to Birch (2007), majority rule

182     K. Sekkat

should be more conducive to corruption than proportional rule, for three reasons. First, politicians have more to gain from individual efforts to corrupt elections due to the winner-take-all nature of majority systems. Second, many voters tend to associate some forms of corruption (special gifts, vote-buying) with individual candidates and other forms with the party. The candidate is therefore encouraged to tailor his/her practices to satisfy these expectations, even at the expense of the party. Third, because the number of votes that must be changed in order to alter the outcome of the election is typically smaller in majority than in proportional systems, the cost of malfeasance is lower in majority systems. However, other contributions raise doubts about these outcomes. For instance, Kunicova and Rose-Ackerman (2001) contend that closed-list proportional representation (PR) systems are more susceptible to corruption than open-list PR and majority systems. The reason seems to lie with the fact that majority systems (single representative by district) highly constrain politicians’ rent-seeking because they make representatives easily accountable to voters. Voters in such single-member districts are better able to observe their representative’s performance and potential misconduct. Voters can therefore identify and eject corrupt representatives at the time of elections. Conversely, in closed-list proportional systems, it is more difficult for voters to punish politicians’ malfeasance because they vote for parties, not individual politicians. The link between the performance of individual politicians and their reelection is weakened by the fact that the candidate’s position on the electoral list (and his/her chance of election) is decided by the party, not voters. The open-list proportional system lies between the two extremes. Representatives can be punished because voters vote for a particular representative on the list and can kick out corrupt ones. However, Chang (2005) argues that in open-list PR systems, politicians are more corrupt because of the need to finance campaigns. Under such systems, personal votes are critical for politicians to win election. This kind of system makes elections more uncertain for individual politicians, and this uncertainty about their chances of winning election may drive politicians to corruption in order to finance campaigns. To sum up, there is no consensus at the theoretical level. Any conclusion may come only from empirical analyses. These are reviewed below

6  Electoral Rules     183

and suggest a number of other factors which are in play, including whether the politician holds or is seeking office, the size of districts, and the degree and type of decentralization.

2 Actual Impacts of Electoral Rules on Corruption As an illustration of the role of democratic institutions, Rose-Ackerman (2006) contrasts the French presidential elections of 2002 with the case of Helmut Kohl in Germany. France has a presidential system where the government’s chief executive is directly elected in a run-off system. French voters’ decision-making differs greatly from that of the German parliamentary system, where the chief executive is a chancellor (prime minister) who is usually a leader of the strongest party. In France, just before the 2002 presidential election, the incumbent president, Jacques Chirac, and his party were involved in a series of corruption scandals involving inflated construction contracts, fictitious jobs, use of public funds for personal expenses, and vote-rigging in previous elections. However, no evidence of Chirac’s personal enrichment was found. Despite the abundant evidence, Chirac simply refused to admit that the scandals existed, which is unsurprising. The surprise came from the fact that the opposition Socialist Party did not seize upon this to push its own candidate. The other candidates’ parties did similarly. As a result, corruption did not become a major issue in the campaign, and Jacques Chirac got into the run-off while the Socialists failed to do so. In Germany, an investigation concerning Chancellor Helmut Kohl was completed in July 2002. He was accused of accepting secret donations for his party. Helmut Kohl admitted “some mistakes” and claimed that what he did was in the interest of the party. The investigation showed no evidence of Kohl’s personal enrichment but concluded that there were infringements. Because of the breach of the law, both Kohl and his party, the CDU, had to pay fines. In contrast to the French case, where the opposition was silent about the corruption scandal,

184     K. Sekkat

the CDU’s main rival, the Social Democratic Party (SPD), seized the opportunity to criticize the CDU. Voters responded forcefully by strengthening the SPD’s position in both houses of the German parliament. Leaders of the CDU ensured that Kohl would never again lead the CDU as candidate for chancellor in a federal election. Looking beyond this anecdote, the empirical literature has investigated more formally the impact of different democratic institutions on corruption. In general, it associates majority rule with single-member districts. The country is divided into multiple districts, each of which elects a single candidate. Proportional representation (PR) systems, meanwhile, tend to have large, often national, districts, electing the entire legislature in one or a few districts. Of course, there are differences among PR systems in terms of district magnitude, but the basic comparison is that under PR district magnitude is larger than under plurality rule. Kunicova and Rose-Ackerman (2005) test the hypothesis presented above that closed-list PR systems are more susceptible to corruption than open-list PR and majority systems. They used the CPI constructed by Transparency International (TI) and the CCI from the World Bank (WB). As far as electoral rules are concerned, they use the WB Database on Political Institutions. They construct a dummy for each of the three systems for a cross-section of 105 countries in 1997. Control variables include indicators of presidentialism, federalism, GDP per capita, press freedom, degree of political competition, and ethnolinguistic fractionalization. The results confirm the main expectations and in particular the existence of a relationship between electoral rules and corruption. Closed-list PR is more corrupt than open-list or majority systems. The geographic size of districts (rather than district magnitude) is a driving force that makes closed-list systems most corrupt. Finally, presidential proportional systems are more corrupt than their parliamentary counterparts. Chang (2005) examines the support for his hypothesis that openlist PR drives politicians to corruption. Specifically, the paper examines how the electoral pressure created by intra-party competition under open-list PR systems might drive individuals, especially those who are at risk of losing reelection, to resort to political corruption. The paper

6  Electoral Rules     185

focuses on the pre-1994 Italian legislatures and uses a dataset on charges of political corruption against Italian members of parliament. The unit of analysis is the individual legislator. The hypothesis to be tested is that a legislator is likely to be charged with political corruption if he/ she faces more uncertainty about reelection. The main explanatory variable is electoral uncertainty resulting from intra-party competition. It is measured using the ratio of victory in the previous election. The ratio is the number of preference votes an incumbent received divided by the number of preference votes for the lowest winner from the same party in that district. It is assumed that there is a direct inverse relationship between the ratio of victory and uncertainty (a legislator’s uncertainty decreases as the ratio increases). Control variables include the number of legislative terms the candidate has served, the number of corrupt competitors a candidate competes with in a given district, GDP per capita, and the size of government. The results exhibit a significantly positive coefficient for the electoral uncertainty variable, as expected. This means that a candidate is more likely to be corrupt if he/she has greater doubt about being elected. A candidate who is ranked lower on the final party list in the previous election has more incentive to “work harder” in making more budgetary amendments so as to attract new supporters for the next election. Overall, under electoral systems in which politicians are in critical need of personal votes, electoral uncertainty actually drives politicians to political corruption. Chang and Golden (2007) supplement the above paper by emphasizing the role of district magnitude (i.e., number of seats by district) under open-list systems in driving corruption. This is because under open-list systems, the incentives to amass resources increase with district magnitude. The opposite seems to hold under closed-list systems. Accordingly, the hypotheses to be tested are that corruption increases with district magnitude under open-list systems and decreases with district magnitude under closed-list systems. The first part of the paper uses data from the years 1996, 1997, and 1998 across 40-odd contemporary democratic nations. The CPI and WB measures of corruption are used. The explanatory variables include whether a country qualifies as a democracy. The information is based on

186     K. Sekkat

Freedom House and Polity IV data. The data on electoral systems and district magnitude come from the Database on Political Institutions. Control is made for the effective number of political parties, presidential versus parliamentary, and federal versus unitary systems. The results show that political corruption gets more severe as district magnitude increases under open-list systems. Conversely, under closedlist systems, political corruption gets less severe as district magnitude increases. Moreover, after a certain threshold of magnitude (around 15 according to the author), corruption becomes greater in open-list systems than in closed-list systems. Only in small districts (below 15 seats) are closed-list systems associated with more corruption, as commonly accepted. It follows that countries using open-list rules should keep district magnitudes small in order to discourage the intense intra-party political competition that can drive important political corruption. This finding is robust to the inclusion of various important control variables, including per capita wealth and the efficiency of the judiciary. The second part of the paper explores the relationship between district magnitude and political corruption in Italy for the years from 1948 to 1994. Over this period, Italy used an open-list system in which voters could decide to use as many as three preference votes for individual candidates on the party list of their choice. Corruption is measured based on charges of suspected (mainly political) corruption against Italian members of parliament. The evidence corroborates the hypothesis. Larger electoral districts are associated with higher levels of corruption. In addition to electoral rules (majority versus proportional), Persson et al. (2003) consider the respective effects of electoral competition and career concerns on corruption. Using a panel of about 80 democracies in the 1990s, they investigate whether the degree of electoral competition and career concerns explain differences in corruption across countries. Specifically, they assume that larger district magnitude and lower thresholds for representation reflect more electoral competition and lead to less corruption. Another assumption is that a larger share of representatives elected on an individual ballot instead of party lists is associated with career concerns and causes corruption. Finally, competition under a proportional system with a single nationwide district is assumed to be weaker than under a majority system with several single-member

6  Electoral Rules     187

districts. As a consequence, a combination of majority rule with small districts is associated with less corruption. The measure of democracy is based on the Gastil Indexes of Political Rights and Civil Liberties published by Freedom House. In addition, the age of a democracy, defined as the fraction of time of uninterrupted democratic rule going back in time, is also considered. The measures of corruption are the CPI and the WB and ICRG indexes. The existence of barriers to entry or low competition is reflected in the average magnitude of voting districts, defined as the number of districts divided by the number of seats in the lower house. To proxy the career-concern effect, two measures are used. One is the proportion of legislators in the lower house who are elected on an individual ballot by majority rule. The other is the proportion of legislators in the lower house elected individually or on open lists. Finally, a dummy is introduced to distinguish between majority and proportional rules, as well as a number of additional economic, social, cultural, historical, and geographic variables that correlate with the incidence of corruption. The findings are that smaller electoral districts tend to be associated with more corruption, which is in line with the barriers-to-entry or low-competition hypothesis. Voting for individuals instead of voting for party lists leads to less corruption, as predicted by the career-concern hypothesis.

3 Conclusion The disciplinary role of elections on corruption depends on the design of democratic institutions, such as majority or proportional systems, presidentialism, parliamentarism, federalism, bicameralism, and so on. For instance, closed-list proportional representation systems are more corrupt than open-list or majority systems. However, intra-party competition may favor corruption under open-list proportional systems. Voting for individuals instead of voting for party lists leads to less corruption because of candidates’ career concerns. Presidential proportional systems are more corrupt than their parliamentary counterparts.

188     K. Sekkat

References Anduiza, E., Gallego, A., & Muñoz, J. (2013). Turning a Blind Eye Experimental Evidence of Partisan Bias in Attitudes Toward Corruption. Comparative Political Studies, 46(12), 1664–1692. Birch, S. (2007). Electoral Systems and Electoral Misconduct. Comparative Political Studies, 40(12), 1533–1556. Chang, E. C. (2005). Electoral Incentives for Political Corruption Under Open-List Proportional Representation. Journal of Politics, 67(3), 716–730. Chang, E. C., & Golden, M. A. (2007). Electoral Systems, District Magnitude and Corruption. British Journal of Political Science, 37(1), 115–137. Kunicova, J., & Rose-Ackerman, S. (2001). Electoral Rules as Constraints on Corruption: The Risks of Closed-List Proportional Representation. New Haven, CT: Department of Political Sciences, Yale University. Kunicova, J., & Rose-Ackerman, S. (2005). Electoral Rules and Constitutional Structures as Constraints on Corruption. British Journal of Political Science, 35(4), 573–606. Persson, T., Tabellini, G., & Trebbi, F. (2003). Electoral Rules and Corruption. Journal of the European Economic Association, 1(4), 958–989. Rose-Ackerman, S. (1999). Political Corruption and Democracy. Connecticut Journal of International Law, 14, 363–378. Rose-Ackerman, S. (Ed.). (2006). International Handbook on the Economics of Corruption. Cheltenham: Edward Elgar.

7 Decentralization

Decentralization is often seen as a means of reducing corruption. This chapter examines the arguments supporting or contradicting this view and their empirical validity. Decentralization implies the transfer of some components of central government responsibilities to local or regional public bodies. It can be political, fiscal, or administrative. The most important channels through which decentralization can affect corruption are accountability, capture, and interjurisdictional competition. Accountability is based on the idea that citizens are better informed of local conditions, which allows them to better evaluate the performance of local government. Capture refers to situations where public authorities and some interest groups become so interdependent that economic decision processes are distorted. Finally, interjurisdictional competition refers to the possibility that different local governments compete with one another to attract investment. This might reduce corruption when investors choose locations which establish market-friendly local laws and regulations and fight bribes. The evidence shows that the impact of decentralization is, in general, corruption-reducing but that it depends on other factors. Decentralization reduces corruption in countries with good monitoring (e.g., freedom of the press) and high education levels. © The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_7

189

190     K. Sekkat

1 Expected Impacts of Decentralization on Corruption Electoral rules are not the only aspects of democratic systems which have spurred an economic debate about the impact on corruption. Decentralization is another feature of democratic systems which generates a rich discussion. Briefly defined, decentralization refers to the transfer of some strands of central government tasks to local or regional public bodies. The legal framework of decentralization does not assign it the task of fighting corruption. However, the change in the social and political environment that should come with decentralization is expected to affect corrupt behaviors. In particular, the close interaction between citizens and local authorities is an important channel through which the effect is expected to operate. Over the last two decades, decentralization has been at the center stage of policy experiments in many developing and transition countries of Latin America, Africa, and Asia. In the two largest countries of the world, China and India, decentralization has been considered as the major institutional framework in the context of their impressive industrial growth. International institutions such as the World Bank (WB) have put decentralization at the heart of their reform agendas (Bardhan 2002). One important difficulty when tackling the effects of decentralization concerns the differences in the definitions used by various authors. Political scientists are more interested in the political decentralization of government systems, economists in fiscal decentralization, and administrative scientists in the decentralization of administrative structures. Political decentralization refers to some degree of transfer of decision-making power to local officials who are elected by the local population. Political decentralization often requires constitutional reforms, development of pluralistic parties, strengthening of legislatures, and public participation in budgeting. Financial decentralization, on the other hand, means that local bodies have authority to make significant decisions regarding spending and taxation. To this end, they must have some degree of local authority to determine the level and

7 Decentralization     191

the nature of local expenditure (spending autonomy) and service delivery as well as some ability to impose fees and taxes (revenue collection autonomy). Finally, administrative decentralization involves shifting decision-making powers from central government officials to officials located outside the capital city. Moreover, a source of further ambiguity concerns the level of decentralization. The literature on federalism traditionally considers the region or province, one step below the national government, as the relevant level of decentralization. In contrast, many recent analyses of decentralization in developing countries focus on lower levels of government, at the municipal or village level. For our purpose, the main question concerns the relationship between decentralization and corruption: Does decentralization lead to more or less corruption? The responses that have been advanced in support of one or the other effect can be grouped into two sets of arguments (Bardhan 2002). One concerns the impact of decentralization on accountability and democracy. The other is based on the effect of decentralization in terms of interjurisdictional competition. The accountability argument considers that on the one hand, decentralization can lead to better-informed decisions concerning taxation or expenditure allocations. However, such decisions are made by agents whose interests may differ from those of citizens and lead to abuse of power. On the other hand, citizens are also better informed about local conditions, which allows them to better evaluate the performance of local government officials. Decentralization then becomes a means for citizens to evaluate the performance of officials and to decide whom to appoint and whom to fire. However, the literature warns against another risk of decentralization, namely capture of local government by elites. Capture refers to situations where public authorities and some interest groups become so interdependent that economic decision processes are distorted. Some argue that the risk of capture is higher at local than national level because media coverage of specific scandals is lower at the local level and local officials are to some extent isolated and more vulnerable to manipulation. Others consider the risk of capture higher at the national level because of the greater need for funds to finance campaigns and the greater difficulty for citizens to evaluate candidates on nationwide issues.

192     K. Sekkat

Interjurisdictional competition refers to the risk that different local governments may compete with one another to attract investment. In theory, investors choose locations with disciplined local government officials who care about establishing market-friendly local laws and regulations and fighting bribes. They avoid locations with poor governance, high levels of corruption, and low provision of necessary infrastructure. However, too much competition for capital can result in negative externalities across jurisdictions. A local government may seek to attract investors away from other regions by offering them opportunities to evade central government taxes and regulations. While the potential benefit may go to the local government, the cost is borne by other regions. Since each local government has the same incentive, the final outcome could be nil or negative for the country as a whole. The effects of decentralization and the related interjurisdictional competition will depend on which of these strategies is more effective at attracting private investors. This depends on numerous factors and can only be answered through empirical studies.

2 Actual Impacts of Decentralization on Corruption Fisman and Gatti (2002) offer an early cross-country study of the relationship between fiscal decentralization and corruption. The empirical model explains the degree of corruption in a country in terms of a measure of decentralization and a number of control variables. Controls include GDP per capita, civil liberties, the size of government, population, openness to trade, and ethnic fractionalization. The sample contains around 60 developing and developed countries and covers the period 1980–1998. The measure of corruption comes from the ICRG. Decentralization is measured as the share of sub-national (state and local) spending in total government spending (state, local, and central). The results show a very strong negative association between decentralization and corruption, meaning that fiscal decentralization is strongly and significantly associated with lower corruption. This association

7 Decentralization     193

is consistent and robust to controlling for a wide range of potential sources of omitted variable bias as well as endogeneity bias. Gerring and Thacker (2004) focus on the relationship between centralization1 and parliamentarism on one hand and corruption on the other. The study includes 125 countries for the period 1997–1998. Centralization is defined here as a political system where the national government is sovereign relative to its territorial units. Parliamentarism concerns systems in which the executive is chosen by, and accountable to, an elective body. It is expected that centralization and parliamentarism are inversely correlated with political corruption. Moreover, such a relationship is expected to be causal, which means that centralization and parliamentarism lower corruption. These expectations are based on the assumptions that fewer veto points and a more hierarchical arrangement of political institutions foster accountability and result in lower levels of corruption. The indicators of corruption are the WB and TI indexes. The indicator of centralization is based on the territorial pattern of government and on the division of power between the lower house and the upper house. The territorial pattern of government distinguishes non-federal, semi-federal, and federal government. The division of power is based on whether there is no upper house or a weak upper house and whether the upper house is not dominated by the lower house. Parliamentarism is based on dummies distinguishing presidential, semi-presidential, and parliamentary systems. Controls include per capita GDP, trade openness, value added in industry, value added in agriculture, agricultural labor as percentage of labor force, legal origin, religion, and democracy. The findings are that centralized and parliamentary forms of government help reduce the level of corruption. The causal effect works through openness, party competition, decision rules, and other factors. In other words, it is not just the number of “veto points” or “accountability” which matter. Rather, the effect of centralization and parliamentarism works through multiple channels to influence corruption outcomes. 1In

fact, the author uses the term unitarism instead of centralization.

194     K. Sekkat

Altunbas and Thornton (2012) investigate the relationship between the degree of fiscal decentralization and corruption using a dataset comprising 64 developed and developing economies for the period 1995– 2008. Specifically, the paper examines whether countries in which a larger share of fiscal revenues and expenditures (high degree) is located at the level of sub-national government are more or less corrupt. The indicator of corruption is the index of corruption in government produced by the ICRG. Fiscal decentralization is measured through four indicators: (i) sub-national government revenues (state and local), (ii) tax revenue, (iii) expenditure, and (iv) compensation of employees. Each indicator is taken at the local level as a percentage of general government revenues and expenditures. The analysis also uses interaction terms in order to distinguish whether the effect of the degree of decentralization on corruption depends on the existence of administrative sub-national autonomy, such as a federal constitution, direct election of the bottom tier of government or constitutional provisions allowing for some autonomy at the local level. Control variables are political accountability measured by an index of press freedom, duration of democracy, a dummy for presidential democracy, real GDP per capita, population, a dummy for major fuel exporters, and the degree of ethnic fractionalization. The results provide strong support for the view that countries in which a larger share of fiscal revenues and expenditures is located at the level of sub-national government are less corrupt. However, sub-national administrative decentralization appears to reduce the beneficial impact of fiscal decentralization on corruption, which suggests that fiscal decentralization is most effective in reducing corruption when the use of these resources is largely directed by central government. Goel and Nelson (2011) focus on the structure of local government. Local government structure is assessed both in terms of the scope of services offered and the degree of fragmentation. The measure of the scope of services distinguishes between multi-purpose (general-purpose) and single-purpose units of local government. The degree of fragmentation is the number of units of local government serving a state’s population. There are reasons to expect that both the scope and degree of fragmentation of local governments could influence the level of corruption in

7 Decentralization     195

a state. Local governments with responsibilities that are broad in scope have more potential gains from corruption, while it is harder for citizens to monitor them. In contrast, citizens residing in jurisdictions that provide specific services may be closely tied or related to elected officials and this may act as a corruption deterrent. The analysis is conducted for the USA. For each state, corruption is measured by the number of convictions for the abuse of public office per 100,000 people and focuses only on convictions that result from federal prosecutions. The data are averaged over three five-year time periods: 1993–1997, 1998–2002, and 2003–2007. Corruption is explained by the number of normalized (per 100,000 people) local government units in state, local government expenditures as a share of total state-local government expenditures, real GDP per capita, population, etc. A large (normalized) number of local government units in a state imply that each unit serves on average a small population. It is then associated with a more fragmented local government structure. Control variables include per capita GDP and population. The evidence shows that both the size and scope of local government can explain the number of public officials convicted of corruption in a state. In particular, more general-purpose governmental units serving a given population are associated with greater levels of corruption. The evidence regarding the effects of special-purpose government units is, however, mixed. Greater fiscal decentralization is associated with lower levels of corrupt activity. This is interpreted as the result of greater transparency and accountability of spending at the local level. The impact of the degree of press freedom on the relationship between decentralization and corruption is investigated by Lessmann and Markwardt (2010) using a cross section of 64 countries. Three measures of corruption are used as dependent variables: the ICRG, WB, and TI indexes. Two explanatory variables are of special interest: decentralization and freedom of the press. Decentralization is measured using different dummies capturing whether a federal constitution exists, the number of vertical government tiers, the share of local expenditures (or revenues) relative to total government expenditures (or revenues) and the share of sub-national government employment in total government employment. The indicator of press freedom is drawn from

196     K. Sekkat

Freedom House. To examine how the relationship between decentralization and corruption is affected by freedom of the press, an interaction term between decentralization and press freedom is introduced into the regression. Control variables include population, per capita GDP, the degree of openness, ethnic fractionalization, and the share of government expenditures in GDP. To avoid causality problems, the independent variables are averaged for the period 1980–1995, while the dependent variables are averages over 1996–2000. The results show no impact of decentralization on corruption in general. However, in countries with effective monitoring (free press), decentralization has a negative impact on corruption. In contrast, if monitoring is not effective, decentralization increases corruption. Albornoz and Cabrales (2013) test the hypothesis that the impact of decentralization depends on political competition. The authors postulate that decentralization is associated with lower levels of corruption only where political competition is sufficiently high. To implement the test, data for 110 countries covering a period between 1996 and 2007 are used. Corruption measures are drawn from the ICRG, TI and WB indexes. Decentralization is defined according to the extent to which sub-national levels of government make decisions about taxation and regulation. Political competition is measured by the WB Voice and Accountability Indicators. This index captures perceptions about how the country’s citizens are involved in selecting their government. Another indicator of political competition used in the study is the Freedom House Index of Political Freedom. The results support the notion that the relationship between fiscal decentralization and corruption is conditional on political competition: Decentralization is associated with lower levels of corruption where levels of political competition are sufficiently high. In contrast to the previous studies, Henderson and Kuncoro (2004) focus on a reverse causality between decentralization and corruption. The paper examines how decentralization can be used by local politicians to issue regulations that allow them to collect more bribes. It takes the example of Indonesian localities which are hampered by insufficient revenues to pay competitive salaries while funding demanded levels of public services. Effective local tax rates are limited at different levels

7 Decentralization     197

across localities by central government. In addition, intergovernmental transfers are limited. As a result, local governments rely on “any” means to fund their local expenditures. Regulation generates direct revenues in the form of taxes plus indirect revenues in the form of bribes. The latter is capitalized into wages paid by localities to public officials. Licensing is one of the two main sources of bribe activity. It also constitutes the most important source of local discretionary revenues. Firms must procure a variety of licenses, both to start and then to operate a business. The analysis is conducted at a level of locality equivalent to a US county. The dataset was collected in 2001–2002 by the University of Indonesia and covers 1808 firms in 64 (out of about 300) local government areas. Respondents were asked about the ratio of bribes and taxes to total costs. The empirical analysis estimates the effect of differential revenue sources on the variation in harassment measured as the number of licenses across localities. The results show that a large reduction in the main form of measurable regulation (licenses) is associated with better-funded localities. Localities having sufficient revenue sources do not need to rely on “red tape” and corruption to effectively compensate local officials and raise local revenues. It is also found that regulation declines with increased education of local officials. Fan et al. (2009) use the WB’s World Business Environment Survey to explore the relationship between decentralization and firms’ actual experiences with corruption. The survey is conducted with managers from more than 9000 firms in 1999–2000 in more than 80 countries. The two questions of interest for our purposes are: “Is it common for firms in your line of business to have to pay some irregular ‘additional payments’ to get things done?” and “On average, what percent of total annual sales do firms like yours typically pay in unofficial payments/ gifts to public officials?” Political and fiscal decentralization are measured based on various indicators: (i) the number of tiers of government in the country, (ii) the average land area of the lowest tier units, (iii) the share of sub-national tax revenues as a share of GDP, and (iv) the share of sub-national government personnel in total civilian government personnel. A tier is defined based on the conditions that the executive body at that level is funded from the public budget, has authority to administer a range of public services, and has a territorial jurisdiction.

198     K. Sekkat

Firm-level control variables indicate ownership (public, foreign, and private domestic). Country controls are GDP per capita, share of imports in GDP, a dummy for democracy, status as a former British colony, the share of minerals and fuels in manufacturing exports, and the proportion of Protestants in the population. The results show that in countries with a larger number of tiers, firms reported that bribery was both more frequent and costlier to firms. Other things being equal, in a country with six tiers of government (such as Uganda), the probability that firms reported “never” being expected to pay bribes was 0.32 lower than the same probability in a country with two tiers (such as Slovenia). The effects are strongest in developing countries. More tiers also appeared to be associated with more frequent bribery in relation to government contracts, connection to public utilities and customs. However, the effect of more tiers is particularly strong when it comes to obtaining business licenses and tax collection. As the number of tiers in a country increases, its effect cancels out other factors related to bribery, such as the size of firms or the country’s religious tradition. Interestingly, the responses seem to confirm previous findings that giving local governments a larger stake in locally generated income can reduce their bribe extraction. This effect holds for both developing and developed countries. The effect is, however, stronger among developed countries. A larger stake in locally generated income is also associated with less frequent bribery. Asthana (2008) presents an empirical analysis of the relationship between decentralization of water supply provision and corruption in India. Specifically, the paper compares the level of corruption between the utilities run by local governments, which are defined as decentralized agencies, and those run by state governments, which are defined as centralized. The hypothesis is that the level of corruption is higher among decentralized agencies. The study covers the rural and semi-urban areas of two large Indian states: Madhya Pradesh and Chhattisgarh. The data are obtained from a large survey covering small towns and villages with a population of more than 2000. Different measures of corruption are used and include the proportion of people who acknowledge paying bribes and the amount of

7 Decentralization     199

corruption related to water bills, repairs, and new connections. These constitute the dependent variables. The explanatory variables are education, income, and a dummy variable which takes a value of 1 for utilities run by local government. The results confirm that corruption in water supply agencies run by local government is higher than in the case of agencies run by regional government. Among respondents, 51% had paid bribes in the case of decentralized agencies and 41% in the case of centralized agencies. The difference is statistically significant. After controlling for other factors, the proportion of people who paid bribes in matters relating to bills is 9% higher in the case of decentralization as compared with centralization. The results for other water supply-related matters confirm that the level of corruption in water supply agencies run by local government is higher than in the case of agencies run by regional government. Regression results also show that educational level has a negative effect while income level has a positive effect on corruption. As summarized by Rose-Ackerman (2006), the effects of decentralization on corruption and government accountability are complex. This implies that decentralization by itself is not a panacea for problems of accountability and must be accompanied by institutional policies to prevent excessive capture of local government. These policies include fostering literacy, expanding information campaigns and monitoring by civic associations and media.

3 Conclusion Decentralization can affect corruption through accountability, capture, and interjurisdictional competition. Assuming that citizens are better informed about local conditions, decentralization is supposed to improve local government accountability. However, the reduced size of jurisdictions under decentralization as compared with centralization makes local authorities more vulnerable to the strategies of certain interest groups. Finally, in order to attract investment, local governments may compete with each other in terms of establishing market-friendly local laws and regulations and fighting bribery.

200     K. Sekkat

The empirical analyses suggest that fiscal decentralization is strongly and significantly associated with lower corruption. This seems to be a result of greater transparency and accountability on spending at the local level. However, administrative decentralization reduces the benefit from fiscal decentralization in terms of corruption. This suggests that fiscal decentralization is most effective in reducing corruption when the use of resources is largely directed by central government. Decentralization is also associated with lower levels of corruption under sufficiently high levels of political competition. A large reduction in the extent of measurable regulation associated with better-funded localities are effective anti-corruption tools. Local governments which receive sufficient revenues do not seek to use “red tape” and corruption to effectively compensate for a lack of resources. Likewise, giving local government a larger claim on locally generated income can reduce bribe extraction. Finally, the degree of literacy, citizen monitoring, and press freedom are crucial for decentralization to reduce corruption.

References Albornoz, F., & Cabrales, A. (2013). Decentralization, Political Competition and Corruption. Journal of Development Economics, 105(C), 103–111. Altunbas, Y., & Thornton, J. (2012). Fiscal Decentralization and Governance. Public Finance Review, 40(1), 66–85. Asthana, A. N. (2008). Decentralization and Corruption: Evidence from Drinking Water Sector. Public Administration and Development, 28(3), 181–189. Bardhan, P. (2002). Decentralization of Governance and Development. Journal of Economic Perspectives, 16(4), 185–205. Fan, C. S., Lin, C., & Treisman, D. (2009). Political Decentralization and Corruption: Evidence from Around the World. Journal of Public Economics, 93(1), 14–34. Fisman, R., & Gatti, R. (2002). Decentralization and Corruption: Evidence from US Federal Transfer Programs. Public Choice, 113(1), 25–35. Gerring, J., & Thacker, S. C. (2004). Political Institutions and Corruption: The Role of Unitarism and Parliamentarism. British Journal of Political Science, 34(2), 295–330.

7 Decentralization     201

Goel, R. K., & Nelson, M. A. (2011). Government Fragmentation Versus Fiscal Decentralization and Corruption. Public Choice, 148(3), 471–490. Henderson, J. V., & Kuncoro, A. (2004). Corruption in Indonesia (National Bureau of Economic Research, WP 10674). Lessmann, C., & Markwardt, G. (2010). One Size Fits All? Decentralization, Corruption and the Monitoring of Bureaucrats. World Development, 38(4), 631–646. Rose-Ackerman, S. (Ed.). (2006). International Handbook on the Economics of Corruption. Cheltenham: Edward Elgar.

8 Regulation

Regulation encompasses a number of instruments and covers various economic, social, and environmental fields. It concerns diverse matters such as entry into some markets (e.g., law, accountancy, and medicine), control of certain prices or the setting of rules governing the functioning of business, or individual activities. Regulation is seen either as seeking to meet the “public interest” or as a response to specific “interest groups”. The first view stipulates that regulation is a response to public demand to correct inefficient or inequitable market outcomes. The second view emphasizes the notion of “capture”, whereby regulation is designed to favor certain interest groups. In relation to corruption, regulation is at risk at two levels: setting and implementation. Regulationsetting is suspected of opening the door to corruption in order to get certain legislation proposed and passed. As regards implementation, bureaucrats enjoy, in general, a large degree of discretion in the interpretation, speed, and complexity of the process. This enables them to ask for bribes. The empirical evidence suggests that regulation is less conducive to corruption when the procedures for setting the rules are open and transparent and implementation is regularly audited.

© The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_8

203

204     K. Sekkat

1 Expected Impacts of Regulation on Corruption Regulation is, in general, justified on the grounds that some “market failures” must be corrected for. These failures concern imperfect competition, technological spillover, information asymmetry, coordination among agents, and other externalities. A classic way in which the market fails is when firms (or individuals) do not take account of the costs or benefits their activities imply for third parties, i.e., externalities. When this happens, the activities may be pursued too intensely or not enough, resulting in losses for third parties and society in general. For example, without regulation, a manufacturing plant may discharge dangerous chemicals into the air and water, causing harm to its neighbors. Governments respond to this risk by setting standards for emissions or the requirement to use specific technologies. Two main views have traditionally dominated the debate on regulation. One emphasizes the “public interest”, saying that regulation is a response to public demand to correct inefficient or inequitable market practices. The other favors the notion of “capture”. Under this view, regulation is supplied in response to the demands of interest groups which seek to maximize the gains for their members. Irrespective of the motivation, regulation is seen as responding to some demand. In 1971, George Stigler proposed a path-breaking approach which clarifies this dimension of regulation. While his theory seems at first glance to be a refined version of long-established views, it had the great merit of being well-grounded in strong economic theory. In particular, his theory considers that people pursue specific objectives and do so rationally. The coercive power of government can be used to fulfill the objective of particular individuals or groups. Hence, economic regulation can be viewed as a good whose allocation follows the laws of supply and demand (see Posner 1974). When it comes to the relationship between regulation and corruption, the problem is posed at two levels (Parker and Kirkpatrick 2012; Aidt and Dutta 2008), i.e., when rules are set and when they are implemented. As regards rule-setting, the theory presented above postulates

8 Regulation     205

that regulation is driven by special interest groups lobbying for legislative changes for their own personal gain. This opens the door to corruption to ensure that such legislation is proposed and passed. At the implementation level, bureaucrats enjoy, in general, a large degree of freedom in the interpretation, speed, and complexity of the process. Even if a request for permits, licenses or tax benefits fulfills the requirement of the law, bureaucrats can exploit their discretionary power and asymmetric information in order to extract bribes. Alternatively, a regulated firm could bribe the regulator to set a higher price or not to enforce a particular regulatory statute. Similarly, a firm could seek to bribe an auditor to hide the fact that the firm made a larger profit than it claimed. Once they became aware of such costs of regulation, many countries started devoting efforts to reducing red tape and the regulatory burden on business as well as to more carefully examining the merits of new and existing regulations. Some countries have made notable progress in improving regulatory practices, but other nations are lagging behind.

2 Actual Impacts of Regulation on Corruption At the empirical level, a first important question is whether regulation causes corruption. Djankov et al. (2002) tackle this question by examining whether countries with heavier regulation of entry have higher corruption. The authors construct a database describing the regulation of entry by start-up companies in 75 developed and developing countries. The focus is on the steps that an entrepreneur needs to take to begin operating legally. The data were collected using available written information from government publications, WB and USAID-sponsored studies, and the Internet. The information is completed and double-checked with the assistance of private companies or through direct contact with government agencies. Three measures of entry regulation are used: the number of procedures that firms must go through, the official time required to complete

206     K. Sekkat

the process, and its official cost, i.e., the cost of entry regulation based on all identifiable official expenses. The CPI for 1999 is explained in terms of the number of procedures, the time required to complete the process, and the cost of entry regulation. The results show that more regulation is significantly associated with more corruption. The effect is also large in economic terms. By examining separately the relationship between entry regulations and corruption in countries with above and below world median income, the results show that regulations actually have a stronger effect on corruption in the sub-sample of richer countries. Recanatini et al. (2005) investigate which features of public agencies influence corruption in eight Latin American and African countries. More specifically, the objective is to explain the diverse patterns of corruption across and within countries on the basis of specific features of the internal organization of each agency and other agency-level variables. To describe the organization of each agency, three continuous variables, each measuring a specific feature of the agency, are constructed. The first variable refers to auditing and provides information on whether decisions on personnel and budget management, as well as procurement, are subject to regular internal or external audits. The second variable concerns openness and shows whether the same set of decisions is publicly announced inside and outside the public agency and whether agencies’ financial status is regularly disclosed to the public. The third variable pertains to merit and assesses whether decisions on personnel management are based on professional experience, merit, performance, and the level of education. These indicators vary from 0 to 100. To differentiate the type of agency, a dummy variable which takes the value 1 if the agency is part of the judicial system is constructed. Another dummy variable equals 1 if the agency operates exclusively at the municipal level. Moreover, the analysis distinguishes between agencies with a popularly elected head, agencies where the head is politically appointed but cannot be removed at will by the political body, and those where the head can be dismissed by politicians (e.g., a central banker). Finally, to take account of the “demand” side of corruption,

8 Regulation     207

three types of typical customer of the agency are considered: families, firms, and foreigners. The preceding variables are used to explain the difference in corruption across agencies. The authors use measures of overall corruption as well as of specific types of corruption such as corruption in public procurement, corruption in personnel management, and legal and regulatory corruption. Control variables include average levels of education, average income, and large firms’ concentration. The results show that corruption varies greatly not only between countries but also within countries. The internal design of the agency is systematically associated with perceptions of corruption, both by insiders and by its customers. Corruption is lower when internal decisions on budget, procurement, and personnel are regularly audited, and when these same decisions are taken with open and transparent procedures. Corruption in personnel is also lower when such decisions are based on merit and clearly stated professional criteria. Interestingly, agencies whose head is popularly elected are systematically more corrupt and have worse internal organizations. The opposite holds for independent agencies whose head is appointed, but not removable by political bodies. Finally, corruption also seems to be influenced by demand-side factors and not just by its internal organization. Agencies that provide services to firms (rather than households) are more prone to corruption. Berg et al. (2012) are interested in how different dimensions of regulation affect corruption. The investigation is conducted for the telecom sector. The dataset includes 3731 firms from 26 transitional economies. The dependent variable is corruption and is constructed from responses to the following question contained in the World Business Environment Survey: “Do firms like yours typically need to make extra, unofficial payments to service providers to get connected to telephone?” The main explanatory variables are measures of the regulation systems, ownership of the telecom firm (public, partially public, or private), the level of competition in local telephone service, and the tariff level (including the installation fee and subscription). The authors distinguish two dimensions of regulation: regulatory substance, which includes tariff setting, quality of service standards, accountants’ ratio, and periodic review, and regulatory governance, which includes independence of the regulator,

208     K. Sekkat

clarity of responsibility, accountability, transparency, and participation. Control variables include the responding firm’s ownership, size, and sector. The results show that strong regulatory substance (the content of regulation) and regulatory governance reduce corruption and that competition and privatization reduce corruption. The effects of regulatory substance on corruption are stronger in countries with state-owned or partially state-owned telecoms, greater competition, and higher telecommunication fees. Bureaucratic quality exerts substitution effects to regulatory substance in deterring corruption. Dreher and Schneider (2010) focus on one determinant of corruption and on how corruption exerts a feedback effect on this determinant. The determinant is the existence of a shadow economy. Theoretical analyses are inconclusive about the linkages between corruption and the shadow economy. Too much regulation can prompt producers to go underground to escape the cost of complying with regulation, which may constrain the ability of tax officials, for instance, to ask for bribes and thus reduce corruption. Similarly, a highly corrupt tax administration may prompt entrepreneurs to go underground, which increases the size of the shadow economy. But producing in the shadow economy also exposes producers to inspectors and might increase corruption. Moreover, the linkage may differ depending on whether developing or developed countries are considered. In developed countries, corruption often consists in bribing officials to get government contracts. Since these contracts are then run in the official economy, corruption in these countries is associated with a smaller shadow economy. In developing countries, corruption often takes place in order to facilitate shadow economy activities. To get additional income from a shadow economy entrepreneur, it is natural for public officials to ask for bribes and thus benefit from the shadow market. Accordingly, the shadow economy reinforces corruption. The empirical analysis is based on a cross-section of 98 countries over the period 1999–2002. Data for the shadow economy are taken from Schneider (2004). To measure corruption, the ICRG, TI, and WB indexes are used together with an index based on a structural model of

8 Regulation     209

the causes. The latter is the fitted values of the regression of corruption on the rule of law, school enrolment, latitude, German legal origin, and age of democracy. Two equations are estimated. One explains corruption in terms of the shadow economy and control variables. The other explains the shadow economy in terms of corruption and control variables. Control variables are tax rates, government revenues, measures of regulation, and institutional quality. The results show no robust relationship between corruption and the size of the shadow economy when the traditional perception indexes are used. When an index based on a structural model is used, however, corruption and the shadow economy are complements in countries with low income, but not in high-income countries. In general, we must admit that we have no clear and robust pattern that confirms our hypotheses among the range of indicators and specifications employed.

3 Conclusion Regulations are issued in response either to public demand to correct inefficient or inequitable market outcomes or to demand from interest groups seeking gains for their members. Regulation creates opportunities for corruption when rules are set and when they are implemented. The empirical evidence shows that, in general, more regulation is significantly associated with more corruption. However, the effects vary greatly not only between countries but also within countries. They are much lower when decisions on budget, procurement, and personnel are regularly audited and are taken with open and transparent procedures. Corruption of people in charge of enforcing regulation is also lower when they are hired on the basis of merit and clearly stated professional criteria. Elected heads of independent regulation agencies are very often more corrupt and implement bad internal organizations. The opposite holds when the heads are appointed, but not removable by political bodies. Finally, agencies that provide services to firms are more prone to corruption than those providing services to households.

210     K. Sekkat

References Aidt, T. S., & Dutta, J. (2008). Policy Compromises: Corruption and Regulation in a Democracy. Economics and Politics, 20(3), 335–360. Berg, S. V., Jiang, L., & Lin, C. (2012). Regulation and Corporate Corruption: New Evidence from the Telecom Sector. Journal of Comparative Economics, 40(1), 22–43. Djankov, S., La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (2002). The Regulation of Entry. The Quarterly Journal of Economics, 117(1), 1–37. Dreher, A., & Schneider, F. (2010). Corruption and the Shadow Economy: An Empirical Analysis. Public Choice, 144(1), 215–238. Parker, D., & Kirkpatrick, C. (2012). Measuring Regulatory Performance. The Economic Impact of Regulatory Policy: A Literature Review of Quantitative Evidence. OECD. Expert Paper No. 3. Posner, R. A. (1974). Theories of Economic Regulation. Bell Journal of Economics, 5(2), 335–358. Recanatini, F., Prati, A., & Tabellini, G. (2005, November 3–4). Why Are Some Public Agencies Less Corrupt Than Others? Lessons for Institutional Reform from Survey Data. In Sixth Jacques Polak Annual Research Conference, IMF. https://pdfs.semanticscholar.org/ef94/d7d211859a6503dadb7bde58aff0acf3005a.pdf. Accessed 9 May 2018. Schneider, F. (2004). The Size of the Shadow Economies of 145 Countries All Over the World: First Results over the Period 1999 to 2003 (IZA Discussion Paper Series, No. 1431).

9 Justice

In the case of problems related to regulation, political competition, or decentralization, citizens can appeal to justice when the rules are not respected or if application of the rules harms their own interests. However, access to and the functioning of justice are not immune to corruption. Three main issues emerge concerning corruption in justice: access to justice, functioning of the judicial system, and punishment of corrupt acts. Access to justice concerns not only the cost for citizens to have their rights upheld but also the human and financial resources (often limited in poor countries) that the state should devote to this task. Suggested solutions to the access problem include the use, when possible, of alternative means of conflict resolution. A well-functioning judicial system requires staff, such as judges, prosecutors, and clerks, to be both independent and accountable. Here, the debate centers on the appointment process (e.g., election or nomination by the government or other bodies) and oversight. The issue of punishment concerns whom to punish: the briber, the recipient, or both? In other words, is paying a bribe worse than accepting a bribe, or vice versa? The problem is further complicated by the existence of other participants in corrupt transactions, such as intermediaries, “advisers”, and other facilitators. © The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_9

211

212     K. Sekkat

1 Expected Impacts of Justice on Corruption: Structural and Procedural Aspects In this section, we distinguish two important dimensions in the role of the judiciary in curbing corruption. The first dimension concerns corruption of the judicial system. The question concerns which structure and operating approach minimize the risk of corruption in the judiciary. The second dimension investigates which sentencing approach is most effective at deterring corruption. Access to justice is an important driver of development and equity. It refers to the ability of citizens to get legal information and legal services and resolve disputes. It includes access to court procedure, to legal aid, and to extra-legal mechanisms to resolve conflicts. Justice is an important instrument in the fight against corruption. However, access to justice is often problematic. It is not easy for claimants with limited resources or knowledge. Moreover, access to justice can be hampered by various obstacles, such as insufficient or poorly trained staff (e.g., judges and clerks) or inadequate infrastructure (e.g., courts and computer systems). These problems can be magnified by indiscriminate access to justice on the grounds of equity and fairness (Botero et al. 2003). Indiscriminate access can lead to excessive use of the courts, which, instead of increasing fairness, floods courts with many trivial cases and highly impacts their effectiveness. According to Transparency International (2007), as of February 2006, 33,635 cases were pending in the Indian Supreme Court, which has 26 judges, 3,341,040 cases were in the high courts, which have 670 judges, and 25,306,458 cases were pending in the 13,204 subordinate courts. This vast backlog leads to long adjournments and prompts people to pay to speed up the process. Estimates in 1999 suggested that “at the current rate of disposal it would take another 350 years for disposal of the pending cases even if no other cases were added”. While justice is in principle a strong instrument against corruption, judges and other judicial staff may be themselves corrupt either owing to rent-seeking or because of political pressures. According to Transparency International (2007), political interference is the most

9 Justice     213

problematic when it comes to the judiciary. Political interference can take the form of threats, intimidation or manipulation of judicial appointments, salaries, and conditions of service. It is difficult to make a harassment case before a court based on these behaviors because they are close to some recommendations seeking to make the administration function more efficiently, as we will see below. The key to preventing political pressure is constitutional and legal mechanisms that protect judicial staff. This protection ensures that courts, judges, and their judgments are independent. However, the design of these mechanisms depends on the whole legal system, which is part of a broader set of cultural rules and customs and varies considerably across countries. There are four predominant legal systems around the world: (i) common law, (ii) civil law, (iii) socialist law, and (iv) religious law. Moreover, a country’s legal system is often a mix of the four predominant systems, especially in developing countries. A way to insulate judicial staff from political pressure is to have them be elected. While most countries around the world use the appointment system for public officials, a large number of US states rely on direct elections for prosecutors and regulators. Judicial elections began in 1789 in Georgia. Mississippi adopted elections for all state judges in 1832. Since 1846, a large number of US states have chosen to hold elections. There are three main types of elections: partisan elections where candidates run for office with a partisan label, nonpartisan elections where candidates have no partisan label, and retention elections where the judge does not face a challenger on the ballot and wins re-election if a specified percentage of votes approve his or her retaining the seat. Retention elections are seen as better insulating judges from the typical pressures of contestable races (Transparency International 2007). An important risk is that too much protection insulates judicial staff from accountability. Judges and judicial staff have numerous means to manipulate the judicial process. They can allow or exclude evidence, set court dates to favor one party or another, inaccurately summarize court proceedings, distort witness testimony, and “lose” files. Moreover, judicial corruption spans all steps of a judicial procedure, from pre-trial to trial proceedings, settlement and ultimately enforcement of decisions. The appeals process is not exempt from opportunities for judicial

214     K. Sekkat

corruption either. Moreover, appeals tend to favor the party with large resources, meaning that poorer people may not be able to pursue their case, even if legitimate, beyond the first instance. In sum, reducing corruption in justice therefore requires a subtle balance between independence and accountability.

2 Actual Impacts of Justice on Corruption: Structural and Procedural Aspects Buscaglia (2001) examines the factors that explain corruption in the judiciary. The study is related to pilot programs in Argentina, Ecuador, and Venezuela. The programs contained policy prescriptions targeting reform of the judiciary and implemented in the three countries between 1993 and 1995. Annual surveys were conducted between 1991 and 1999; i.e., a period covering the years before and after the policy was implemented. The questions in all surveys sought to capture the frequency of different corrupt practices (fraud, embezzlement, court-related political clientelism, politically or financially motivated changes in rulings, politically or financially motivated changes of venue, speed money, and extortion) within a sample of 450 commercial cases in 27 pilot courts. The dependent variable is the year-to-year change in the frequency of corruption by court and country. There are six explanatory variables of interest: the number of procedural and administrative steps followed in each of the 450 cases, the times to disposition for each of these cases, the proportion of all administrative and jurisdictional tasks concentrated in the hands of each employee that is allocated through “informal” mechanisms, the growth of alternative dispute resolution channels (mediation, arbitration, conciliation), the weighted average real incomes of judges, clerks, and other court personnel, and, finally, a measure of the degree of use of court-related information technology. The results of the regression are similar across the three countries, and the coefficients are significant and show the expected signs except for average real compensation, which is non-significant. A higher

9 Justice     215

proportion of administrative and adjudication tasks allocated to court personnel in an informal manner increases corruption. Longer procedural times cause significant increases in corruption, which is in line with the idea of “speed money-related corruption”. An increase in the number of administrative and procedural steps also comes with significant increases in corruption. Information technology has a significant negative impact on corruption. The introduction and legalization of alternative dispute resolution cause a significant reduction in the perceived frequencies of corrupt practices. As a by-product of Buscaglia (2001), it appears that providing more resources to the judiciary reduces corruption. Alt and Lassen (2014) further highlight this dimension by investigating the effect of prosecutorial resources on corruption prosecutions and convictions. Panel data on corruption convictions in US states from 1977 to 2003 are used to examine how public resources available for the investigation and prosecution of offenders affect the extent of corruption. The dependent variable is corruption measured by the number of corruption convictions normalized by state population. The sample counts 21,000 cases observed between 1977 and 2003. This information is collected at the state level. The main explanatory variables of interest are resources. The two measures used are the number of general attorneys in US Attorney offices (prosecutorial resources) and the wage of public employees (a deterrence consideration). The resources of the Attorney office are the number of federal full-time-equivalent positions in the office by year and judicial district normalized by state population. Three variables measure wages: (i) average wage for state and local government, in current dollars, (ii) ratio of state and local government wage to the average non-government wage in the state, and (iii) average wage for state and local government, in constant dollars, adjusted for the difference in purchasing power. Controls include real income per capita in the state, the population share with at least high school education, inequality, size of the state’s government, and whether the legislature and executive are controlled by different parties (divided government). The results show that greater prosecutorial resources result in more convictions for corruption. Moreover, when prosecutorial resources are

216     K. Sekkat

allowed to depend on the intensity of the work, the estimated effect of resources on convictions is larger. Relative public sector pay has no consistent effect on corruption. A divided government appears to be associated with lower corruption, while term limits appear to be associated with higher corruption. Gordon (2009) focuses on another possible problem relating to the functioning of the judiciary, namely that of partisan judgments, prosecutions, and convictions. The author develops a theoretical model of the interaction between officials contemplating corruption and a prosecutor deciding whether to pursue cases against them. One testable prediction of the model is that biased prosecutors will be willing to file weaker cases against political opponents than against allies. The empirical analysis is based on two samples of state and local corruption prosecutions. The first consists of cases concluded during the Bush administration (from 2004 to 2006). This sample contains information on the prosecution outcomes of 222 defendants of which 204 resulted in guilty verdicts. In this sample, 84 individuals were clearly identified as Democrats or affiliated with Democrats, and 23 individuals were clearly identified as Republicans or affiliated with Republicans. The second sample consists of cases concluded during the Clinton administration (from 1998 to 2000). It contains 223 cases of which 210 resulting in guilty verdicts. In this sample, 49 could clearly be identified as Democrats, and 28 as Republicans. The above figures show a difference between the two samples. During the Bush era, the ratio of defendants is 3.65 Democrats to 1 Republican. During the Clinton era, the ratio of defendants is 1.75 Democrats to 1 Republican. These crude accounts cannot, however, be interpreted as evidence of the existence of partisan bias because of the many other variables in play but not taken into account. For this reason, the analysis used different techniques, among which regression, to address the purpose of the paper. A regression was conducted on each sample separately and used as its dependent variable the difference in sentences between Republican—and Democratic-affiliated defendants. The main explanatory variables included whether the defendant was a private citizen, whether he or she was an elected official, whether the

9 Justice     217

sentence was approved by a judge appointed by a Democratic president, and whether the sentence was part of a case with multiple defendants. Estimates from all three methods strongly support the same finding: There is a partisan bias. Such bias existed under both the Clinton and the Bush eras. However, the evidence also pointed to the disproportionate prosecution of Democrats under both Bush and Clinton. Determining the difference in the bias between the two administrations is more challenging. Having found that corruption prosecutions in the USA are likely to be impacted by partisan bias, the next question is whether independent prosecutors to a better job of deterring political corruption. This is the issue investigated by van Aaken et al. (2010) using a sample of up to 78 developed and developing countries for the years 1998–2006. One difficulty in conducting such investigation is the distinction between de jure and de facto prosecutor independence (PI). The former refers to legal or constitutional provisions that ensure independence while the latter focuses on independence in practice. Accordingly, the construction of de jure independence relies on legal documents while more steps are required to construct de facto independence. De jure PI is based on five criteria: (i) whether the prosecution agency is mentioned in the constitution, what the formal requirements are to become a prosecutor, etc., (ii) the procedure for appointing, promoting, transferring, and removal prosecutors from office, (iii) the right of government members to give positive/negative instructions to prosecutors, (iv) the ways to get prosecutions started, and (v) the degree of discretion that prosecutors enjoy in pursuing their cases. To assess de facto PI, a survey was conducted in the 78 countries and contained six questions of interest: (i) How frequently are prosecutors forced to retire against their will? (ii) How frequently are prosecutors removed from office against their will? (iii) How frequently do members of government change the legal foundations for prosecution? (iv) Has the income of prosecutors remained at least constant in real terms since 1960? (v) Has the budget of the prosecutorial offices remained at least constant since 1960? and (vi) How many cases are initiated by actors other than the state prosecutors? The responses to the six questions were combined with the number of politically motivated assassinations

218     K. Sekkat

counted in a country to create an index of de facto independence. The number of assassinations was drawn from Banks (2004). To test for the impact of the two measures of PI on corruption, an equation is estimated where the average CPI for the years 1998–2006 is explained in terms of de jure and de facto PI and a number of control variables such as real per capita income in 1990, trade openness, population, the share of Protestants in the population, and a dummy for former British colonies. The findings are that greater prosecutorial independence in fact (rather than just in law) leads to lower levels of perceived corruption. This result is quite robust across various specifications. Formal prosecutorial independence is not significantly correlated with corruption. Voigt and Gutmann (2015) extend the analysis of the determinants of judicial corruption and corruption in general to more characteristics of judicial organization. These characteristics include: (i) stability of judges’ salaries, (ii) obligation for judges to extensively justify their decisions, (iii) obligation to publish judgments, extended proof and dissenting opinions, (iv) average time needed to get one’s rights enforced, (v) separation of powers measured as the number of actual veto players in a political system, and (vi) degree of prosecutors’ monopoly power to prosecute crimes. Measures of corruption of the judiciary are drawn from the World Economic Forum, Transparency International, and the World Justice Project. Control variables are real per capita income in 1990, trade openness, population, the share of Protestants in the population, and a dummy for former British colonies. The results show that countries face less judicial corruption if judges’ salaries have not decreased in real terms. Serious requirements for legal justification are associated with less judicial corruption. The average time needed to get one’s rights enforced is not significantly correlated with corruption in the judiciary. More extended publication requirements are not associated with corruption. Finally, monopolization of the power to prosecute fosters judicial corruption. Turning to corruption at large, the dependent variable is the CPI for the year 2013. The main explanatory variables are judicial independence (JI) from Feld and Voigt (2003), prosecutorial independence from van Aaken et al. (2010), and judicial accountability from Voigt (2008).

9 Justice     219

Similar control variables to those mentioned above are introduced. The results show that the independence of the judiciary or prosecutors is associated with lower levels of corruption. Judicial accountability is associated with lower levels of corruption. However, judicial accountability and independence work as complements: Judicial accountability reduces corruption only when the judiciary’s independence (including prosecutors) is sufficiently high. More interestingly, the independence of judges and prosecutors may increase corruption at low levels of judicial accountability, but it is related to significantly lower corruption levels when judges are held sufficiently accountable. Although not directly focusing on corruption, Goelzhauser (2012) highlights another aspect of judicial performance (case disposition times) that may impact corruption. The theoretical argument builds on the notion that judges who face election have an incentive to produce efforts that help their re-election. These judges face, however, different types of pressure. Judges that face partisan elections often bear a significant amount of political pressure and are routinely voted out of office. Judges who face retention elections are relatively secure, which gives them different levels of independence but influences the degree of their accountability. To capture differences in judicial independence, the study uses a quasi-natural experiment in Kansas, where 17 judicial districts use non-competitive retention elections while 14 employ partisan elections. The dependent variable is the median disposition time of a case in days for each district. The independent variables include a dummy equal to 1 for counties that use partisan judicial elections, the total number of case filings per judge, urbanization, county ideology, the percentage of residents aged at least 25 and having a high school diploma, the percentage of residents who are black, and county household income as a percentage of the state’s median household income. The results do not show any systematic and consistent effect of less independence (partisan judges) on time disposal. Another study not directly linked to corruption but useful to understanding the impact of judicial organization is Hayoa and Voigt (2007). Specifically, the paper investigates why the judiciary is de facto independent in some legal systems and highly dependent on others. Recall

220     K. Sekkat

that van Aaken et al. (2010) discussed above showed that greater de facto prosecutorial independence leads to lower levels of perceived corruption in a sample of 46 developed and developing countries. The analysis explains the degree of de facto judicial independence (JI) in terms of the degree of de jure JI (Feld and Voigt 2003), trust, legal system, religious affiliations, and other variables. The degree of de facto JI varies between 0 and 1; greater values indicate a higher degree. The results show that de facto JI is robustly explained by de jure JI, which means that de facto JI can at least be partially modified by institutional choice. The degree of de facto JI is further determined by the amount of confidence that the citizens of a country have in their legal system as well as by their religious affiliations. The elasticity estimate for de jure JI is the highest compared to other variables, which confirms that the most effective way of increasing de facto JI is through the creation of formal laws.

3 Expected Impacts of Justice on Corruption: Sentencing Since corruption is either an infringement of existing rules and/or harmful to society, punishment would seem to be the fair response. However, punishment might be ineffective or unfair. This may be the case for different reasons, such as the number of persons involved, their relative power and incentives, the cost of law enforcement, and the issue of setting the “right” level of punishment (Engel et al. 2013). The number of persons involved in a corrupt activity starts at two but can attain very high levels. A basic transaction may involve only one briber and one recipient. Sometimes the transaction involves an intermediary. In other cases, the transaction involves multiple officials, which further increases the number of participants. The relative power of the various participants is not always the same. For instance, corruption might be coercive, and the briber may be unable to avoid it. These features of corrupt activities raise the initial question of whom to punish: the briber, the recipient, or both? In other words, is paying a

9 Justice     221

bribe worse than accepting a bribe or vice versa? In most jurisdictions, both are generally criminal offenses and incur parallel punishments. However, it would be fair to distinguish situations where the briber has no other choice than to bribe (extortion) and situations where he/ she chooses and even initiates the process. There are therefore grounds for treating the two situations asymmetrically. This is the case in some countries, such as Taiwan, where paying off an official is only a crime when the payment is made to obtain an illegal service. In all other cases, the payer is not subject to criminal sanctions. Under Romanian law, making a payoff is not a crime if the briber has been obliged in any way by the recipient. Moreover, a briber in this position can claim restitution of such payments. In other countries, the reverse is true. In Chile in the 1990s, payment of a bribe was a criminal offense, but accepting a bribe was not unless accompanied by other misconduct. Under American law, the maximum penalties are symmetric for those who make and those who accept corrupt payments. The offender can receive a maximum sentence of three times the monetary equivalent of the bribe, be imprisoned or both. He/she can be disqualified from holding any office of honor, trust, or profit (Engel et al. 2013). The debate about the symmetry of punishment gained further intensity after the suggestion by Basu (2011) concerning harassment bribes, i.e., payments that people have to make in order to get what they are legally entitled to. The author suggested that in all such cases the act of giving a bribe should be treated as a legitimate activity. In other words, the bribe-giver should have full immunity from any punitive action by the state. He argued that this would cause a sharp decline in the incidence of bribery. The reasoning is the following. Once the law is altered in this manner, the interests of the briber and the recipient will diverge. The briber will be willing to cooperate in getting the recipient caught. Knowing that this will happen, the recipient will be deterred from taking a bribe. Drèze (2011) criticizes the suggestion that legalizing bribe-giving will cause a sharp decline in the incidence of bribery. To show why the idea does not work, he points out that a briber has three options: do not bribe, bribe and report to the authorities, and bribe but do not report. The “legalization” proposal enhances the attractiveness of the second option vis-à-vis the first. However, it also makes the third

222     K. Sekkat

option more attractive if reporting to the authorities incurs even a small cost. Besides litigation costs, there is the possibility of harassment by the recipient’s colleagues, especially if the chance of getting justice is small. In this situation, the second option is no longer an option because of these costs. The real choice is between not paying a bribe and paying a bribe without reporting. The bribe will be paid anyway, and no sharp decline in the incidence of bribery will be observed. The above proposal is consistent with some practices around the world. For instance, in the USA, two similar procedures are used. The False Claims Act rewards those in the private sector who report irregularities in government contracts and protects them from reprisals (RoseAckerman 2010). Informers are also paid a share of the total penalties. The Whistleblower Protection Act protects informers inside government agencies from reprisal but does not give them a financial reward. Such systems of carrots and sticks, however, depend on the existence of a credible system of investigation and law enforcement that might punish the corrupt deal on its own. The use of undercover operations can be leveraged into a tool that encourages those offered or pressured for bribes to come forward. If they do not, they know that the corrupt offer may be a trap set by law enforcement authorities (Rose-Ackerman 2010). A further complication in the discussion about punishment for corruption comes from the fact the process may involve other players such as intermediaries, colleagues, and even heads of department (Hasker and Okten 2008). Intermediaries helping individuals and firms are common in developing countries and their use is a familiar feature of many corrupt systems. A person with “connections” can smooth the route through the bureaucracy against payments. Such payments are used both to bribe public officials and to compensate the agent. The intermediary is often either a former or off-duty official. With knowledge of the practices and the “rates” for each service, the intermediary can save time by eliminating the need for annoying extra visits to government offices. In complex systems, intermediaries can make things simpler and speed up the bureaucratic process. For a given procedure, individuals using intermediaries are better off than if intermediaries did not exist. Intermediaries grease the

9 Justice     223

wheels. However, the more effective this corrupt system turns out to be, the greater the incentive becomes for officials and middlemen to work together to perpetuate its existence and maximize its return. These objectives can be achieved by increasing the time and trouble imposed on citizens which, in turn, offers more opportunities for corruption. Intermediaries can thus improve access to the bureaucracy, but also strengthen incentives to create red tape (Rose-Ackerman 2010). As a result, various surveys of businessmen suggest that intermediaries (or middlemen hired by corporations and individuals) are a major cause of persistent and high corruption in the developing world (Hasker and Okten 2008). For example, Fjeldstad (2003, p. 172) explains how this phenomenon emerged as an unintended consequence of a Tanzanian government reform and anti-corruption campaign, which led to the firing of one-third of bureaucrats in the tax administration. Private businesses hired these fired bureaucrats to benefit from their knowledge and insider contacts. New corrupt networks soon emerged. Ironically, the solution introduced to tackle corruption caused the emergence of new corrupt networks. In addition to the problem of intermediaries, the fight is often complicated by the systemic nature of corruption. Within a corrupt administration, it is almost impossible for a single official to remain honest. The official must not only renounce the gain from corruption but may also face hostility from colleagues because he/she diverts citizens to his/her desk or is suspected of reporting misbehavior. Furthermore, the delivery of, say, a license may need the approval of different civil servants. While one of them might speed up the process, others might slow it down unless they receive bribes. Such a situation often creates a system of bribe-sharing that is generally organized by those at the top of the hierarchy. Of course, those at the top have an incentive to put pressure on those further down the chain to become part of the corrupt system. Even if the question of whom to punish is settled, the level and type of punishment still have to be determined (Rose-Ackerman 2010). Fines seem much less resource-intensive for the government than putting people in prison, and their level could be calibrated to produce equivalent deterrent effects for defendants. Clearly, overly low

224     K. Sekkat

levels of fines will have no deterrent effect on the briber or the recipient. However, an overly high level might not deter bribery, especially in developing countries, where officials are not well paid and fines exceed their family’s total wealth. Setting fines at an overly high level, by raising the costs to extortionists, can only result in demands for higher bribes. This will lower the incidence of corruption but increase the average level of bribes paid. One way to reinforce the deterrent effect of penalties is to increase the risk of being apprehended. However, this poses other problems as explained below. Apprehension and law enforcement incur specific costs for the justice system. Besides the costs common to the functioning of any administration, the functioning of justice involves additional costs related to gathering evidence, negotiating leniency, and rewarding informers. Deterrence of wrong-doing cannot be effective unless enforcement authorities are able to obtain relevant evidence to be used in potential negotiation or when determining the kind and level of punishment. Obtaining relevant evidence is challenging since, in general, only the parties to a corrupt activity know how the deal functions. The information an investigator can access depends on the probability that one or more of the participants has an incentive to report. Those who can provide such information might, however, be reluctant to do so, fearing future reprisal by the other member in the deal. To counteract such fears, the authorities must offer protection, which might be very costly. In sum, law enforcement may require substantial resources, which are lacking in many countries, especially developing ones (Rose-Ackerman 2010).

4 Actual Impacts of Justice on Corruption: Sentencing Engel et al. (2013) use a corruption game to see whether symmetric or asymmetric punishment is the most effective in deterring corruption. Asymmetry is defined in terms of the level of the fines imposed on the recipient and the briber. For this study, experiments were conducted in

9 Justice     225

Germany and China. Choosing two different locations is a robustness check which aims at showing whether the results reflect a generalizable effect or differences in national cultures or in the legal environment. In Germany, the experiments were run at the University of Bonn and in China at Shanghai Jiao Tong University. The summary of the game is as follows. An individual has to decide whether to bribe or not in order to get a favor from an official. If the decision is not to bribe, the game ends. If the decision is to bribe, an official is approached with the offer. The official can reject the offer, accept it and provide the favor, or accept the money without providing the favor. If the official rejects the offer, the game ends. If the official accepts the offer, money is transferred by the individual. The official has now two possibilities: grant the favor or not. Granting the favor implies the risk of being detected by the authorities. Granting no favor might encourage the briber to report to the authorities. The asymmetry comes here from whether in the event that the corrupt deal is discovered (by the briber being reported or detected by the authorities) the briber is fined less than the recipient. In this game, the briber is never fined more than the recipient. The results of the experiments show that under both symmetric and asymmetric punishments corruption takes place. However, fewer deals are implemented under symmetric punishments. These results hold for the experiments in both China and Germany, which suggests that the effect does not depend on a specific social, political, economic, or legal culture. If bribers are punished more leniently (asymmetric punishment), there is more corruption. Under such asymmetric punishment, bribers are less hesitant to approach a public official and offer a side payment in exchange for a favor. If the official breaks the deal, the briber can at a relatively small cost impose severe harm on the official. The experiment shows that bribers do indeed use this threat and that it is correctly anticipated by most officials who therefore provide the favor. We conducted a conceptual discussion above about the proposal by Basu (2011) to consider the act of bribing to avoid harassment as a legitimate activity. The proposal differs from the paper by Engel et al. (2013) discussed above in that the official always initiates the corrupt transaction and that, more importantly, punishment is asymmetric in

226     K. Sekkat

the sense that the briber risks no punishment. Abbink et al. (2014) examine the effectiveness of such asymmetric liability in combatting harassment bribes. The paper is based on the results of an experimental study in India involving 360 undergraduate and graduate students. The experiment went as follows. Participants are randomly assigned the roles of a citizen or an official. The citizen meets the official and the game starts. The official can decide to ask the citizen for a bribe or can decide not to ask for a bribe. An official who decides to ask for a bribe has to choose the amount. The citizen then has three options: refuse to pay the bribe, pay the bribe, or pay and report the corrupt act. Reporting the bribe makes it much more likely that the official will be caught and fined. If the citizen reports the bribe, but the authorities do not found sufficient evidence to fine the official, the citizen may incur a penalty. The experiment showed that compared with symmetric liability, granting the briber legal immunity increases the reporting of bribe demands and reduces demand for bribes. The study also found that a substantial minority of citizens refuse to pay the bribe despite the significant monetary cost of doing so. Moreover, it appeared that strict financial incentives do not necessarily drive reporting behavior. Nonmonetary factors can motivate reporting behavior as well. An analysis of officials’ behavior suggests that Basu’s proposal has limited ability to curb corruption when officials are able to retaliate against citizens who report bribe demands. The authors suggest that Basu’s proposal should be implemented along with complementary measures such as policies to rotate officials in different posts to mitigate the effectiveness of retaliation against citizens who report bribe demands. Moreover, to further protect citizens’ vulnerability, informers may need to be given protection, such as anonymity. Wu and Abbink (2013) investigate the effect of asymmetric reporting using a game scheme similar to Engel et al. (2013) but focus on rewards for self-reporting. In this game, there are three reporting scenarios: both the individual and the official may self-report, only the individual may self-report, and only the official may self-report. Accordingly, the asymmetry comes here from which party is permitted to self-report, instead of from the relative size of penalties applied to parties like in the

9 Justice     227

majority of other studies. All experimental sessions were conducted at Xiamen University in China with undergraduate students. In all, 230 subjects participated in the experiment. The idea of offering rewards is an interesting one. In reality, there are instances where the relation between the briber and the recipient is characterized by repeated interactions and others where it is not. Repeated interactions are likely to create trust between the parties. The trust that is established implies little or no perceived risk of being caught. Offering a reward for self-reporting might undermine such trust and pave the road toward less corruption. Offering rewards to combat misconduct is not an entirely new idea. For instance, Singapore rewards officials who refuse bribes and expose their briber. Programs such as the 1993 DoJ Corporate Leniency Policy in the USA and the 1996 EU Leniency Program are designed to combat cartel formation. The experiment showed that none of the asymmetric regimes outperforms the symmetric regime in terms of deterring corruption. When agents anticipate future interactions, symmetric reporting and “Only Official Reports” are only slightly effective in reducing bribery, but they are more effective than “Only Individual Reports”. When agents expect they will not be interacting with their partner in the future, all reward mechanisms are extremely effective in reducing the incidence of bribery. These results support the implementation of a reward mechanism to deter bribery, especially petty corruption, where a given citizen is unlikely to meet the same official in the future. The implications are less clear for grand corruption (e.g., a multi-million-dollar bribe to secure a very large government contract) or instances where the individual interacts with the same official on a repeated basis.

5 Conclusion Justice, which is central to the functioning of all societies, is also susceptible to corruption at three levels: access to justice, functioning of the judicial system, and punishment of corrupt acts. Access to justice involves important costs which many persons and countries cannot afford. For the judicial system to function well, staff, including judges,

228     K. Sekkat

prosecutors, and clerks, must be independent, accountable, fairly paid, and not overwhelmed. Finally, punishment can be symmetric or asymmetric, punishing both the briber and the recipient, or only one of them. Investigations show that a high number of administrative and adjudication tasks allocated to court staff as well as a high number of procedural steps can increase corruption. Long procedural times also cause increased corruption. In accordance with these findings, greater prosecutorial resources and their connection to the intensity of the work result in higher rates of convictions, especially for corruption. Greater prosecutorial independence in fact (rather than just in law) coupled with judicial accountability lead to lower levels of corruption. Strict requirements for legal justification are associated with less judicial corruption. More extended publication requirements are not associated with corruption. Surprisingly, wage increases have no consistent effect on corruption, but countries have seen a decrease in judicial salaries in real terms show higher corruption. Regarding punishment, laboratory experiments show that asymmetric punishment by means of which bribers are punished more leniently leads to more corruption. In fact, granting bribers legal immunity leads to increased offers of corruption. It also increases reporting of bribe demands and reduces demand for bribes. The implementation of reward mechanisms to deter bribery, especially petty corruption, is effective where a given citizen is unlikely to meet the same official in the future. The implications are less clear for grand corruption and when individuals interact frequently with the same official.

References Abbink, K., Dasgupta, U., Gangadharan, L., & Jain, T. (2014). Letting the Briber Go Free: An Experiment on Mitigating Harassment Bribes. Journal of Public Economics, 111, 17–28. Alt, J. E., & Lassen, D. D. (2014). Enforcement and Public Corruption: Evidence from the American States. Journal of Law Economics and Organization, 30(2), 306–338.

9 Justice     229

Banks, A. (2004). Banks’ Cross-National Time-Series Data Archive. Binghamton, NY: Databanks International. Basu, K. (2011). Why, for a Class of Bribes, the Act of Giving a Bribe Should Be Treated as Legal (Ministry of Finance Working Paper). Botero, J. C., La Porta, R., López-De-Silanes, F., Shleifer, A., & Volokh, A. (2003). Judicial Reform. World Bank Research Observer, 18(1), 61–88. Buscaglia, E. (2001). An Analysis of Judicial Corruption and Its Causes: An Objective Governing-Based Approach. International Review of Law and Economics, 21(2), 233–249. Drèze, J. (2011, April 23). The Bribing Game. Indian Express. http://econdse. org/wp-content/uploads/2012/09/JD-The-Bribing-Game-2011.pdf. Accessed 9 May 2018. Engel, C., Goerg, S., & Yu, G. (2013). Symmetric vs. Asymmetric Punishment Regimes for Bribery. Max Planck Institute for Research on Collective Goods (No. 2012_01). https://www.econstor.eu/bitstream/10419/57476/1/685077438.pdf. Accessed 9 May 2018. Feld, L. P., & Voigt, S. (2003). Economic Growth and Judicial Independence: Cross Country Evidence Using a New Set of Indicators. European Journal of Political Economy, 19(3), 497–527. Fjeldstad, O. (2003). Fighting Fiscal Corruption: Lessons from the Tanzania Revenue Authority. Public Administration and Development, 23(2), 165–175. Goelzhauser, G. (2012). Accountability and Judicial Performance: Evidence from Case Dispositions. Justice System Journal, 33(3), 249–261. Gordon, S. C. (2009). Assessing Partisan Bias in Federal Public Corruption Prosecutions. American Political Science Review, 103(4), 534–554. Hasker, K., & Okten, C. (2008). Intermediaries and Corruption. Journal of Economic Behavior and Organization, 67(1), 103–115. Hayo, B., & Voigt, S. (2007). Explaining De Facto Judicial Independence. International Review of Law and Economics, 27(3), 269–290. Transparency International. (2007). Global Corruption Report. Van Aaken, A., Feld, L. P., & Voigt, S. (2010). Do Independent Prosecutors Deter Political Corruption? An Empirical Evaluation Across Seventy-Eight Countries. American Law and Economics Review, 12(1), 204–244. Voigt, S. (2008). The Economic Effects of Judicial Accountability: CrossCountry Evidence. European Journal of Law and Economics, 25(2), 95–123.

230     K. Sekkat

Voigt, S., & Gutmann, J. (2015). On the Wrong Side of the Law—Causes and Consequences of a Corrupt Judiciary. International Review of Law and Economics, 43(C), 156–166. Wu, K., & Abbink, K. (2013). Reward Self-Reporting to Deter Corruption: An Experiment on Mitigating Collusive Bribery (No. 42–13). Monash University, Department of Economics. Rose-Ackerman, S. (2010). The Law and Economics of Bribery and Extortion. Annual Review of Law and Social Science, 6, 217–238.

10 Specialized Anti-corruption Agencies

Many countries have created agencies specialized in the fight against corruption both as an alternative to recourse to the judicial system and because the complexity of the corruption phenomenon is escalating. The fight against corruption requires specific skills in a variety of fields, including law, finance, economics, accounting, civil engineering, and social sciences. The main functions of such agencies are law enforcement, prevention, policy development, and coordination. A comparative review of different agencies around the world reveals that their effectiveness in curbing corruption differs highly across countries. The most successful examples are those of Singapore and Hong Kong. Such a review also shows that when the agencies do not deliver “success”, failure is not entirely attributable to them. A major source of failures is the relationship with governments and donors in terms of funding, independence, accountability, and transparency.

1 Expected Impacts of ACAs on Corruption Given how hard it is for the traditional judicial system to curb corruption by itself, as shown above, most countries have established AntiCorruption Agencies (ACAs). Operating under different names, such as © The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_10

231

232     K. Sekkat

corruption prevention bureaus, autonomous anti-corruption commissions, and investigation commissions, ACAs are separate, permanent specialized agencies established by governments for the specific purpose of combating corruption in their countries (Charron 2008; Quah 2009). One of their most important benefits is to send a powerful signal to citizens in the country that the government is committed to fighting corruption. A visible, credible, and independent ACA is crucial to success in the fight against corruption. This means that the government should demonstrate its commitment by granting the ACA sufficient legal powers so that it can investigate anyone regardless of status or position. Moreover, the government should provide adequate human and financial resources, so the ACA can fulfill its role. A comparative overview of different types of ACA reveals that their main functions are law enforcement, prevention, policy development, and coordination. The overview also makes it possible to distinguish two broad ACA models: the single-agency model (Hong Kong, Singapore, Argentina, Malaysia, and Tanzania) and the multiple-agency model (Britain, France, India, and Mexico). The first model has attracted most visibility and interest because it is, in general, multi-purpose and combines in one institution a multifaceted approach comprising prevention, investigation, and education. These features make it the closest to the model of an ACA. The multiple-agency approach is less ambitious since it only creates one or more additional units or agencies with specific anti-corruption responsibilities that either did not previously exist or were scattered among different departments. This approach prevents the set-up of a strong “lead” anti-corruption agency. As a consequence, it is seen as less threatening to some private or public interest groups (Klemenčič and Stusek 2008; Meagher 2005). Three main factors contributed to the spread of ACAs. First, the belief among many governments and international institutions that just strengthening legislation was not sufficient to effectively control corruption. Corruption had spread so widely that it even affected important institutions in charge of fighting corruption, such as the police or the judiciary, with the result that bribery offenses had ceased to be investigated or prosecuted in many countries. Moreover,

10  Specialized Anti-corruption Agencies     233

the multifaceted nature of corruption means that traditional public institutions cannot control it. The fight against corruption requires specific expertise, knowledge, and skills in a variety of fields, such as law, finance, economics, accounting, civil engineering, and social sciences. These skills, which are frequently dispersed across various institutions, must be concentrated in a specific agency in order to generate synergy between staff from different fields of knowledge and make their action more effective (Klemenčič and Stusek 2008; Charron 2008). Second, by the mid-1990s, the problem of corruption had become a major concern at the international level. A number of international organizations, including the UN, the OECD, and the EU, started designing and adopting instruments to fight corruption. Although these international legal instruments varied in many respects, they all aimed to promote specialization of law enforcement and prosecution bodies in the field of anti-corruption (Klemenčič and Stusek 2008). Third, the success they seemed to meet in fighting corruption in Singapore and Hong Kong made ACAs the example to follow.1 In Singapore, the Corrupt Practices Investigation Bureau (CPIB) was established in 1952. In Hong Kong, the Independent Commission Against Corruption (ICAC) was founded in 1974. Despite their similar origins, design, and success, the ICAC and CPIB take different approaches. The ICAC is very large, well-resourced, and strongly oriented toward transparency and civic partnership, including outreach and education. The CPIB is much smaller and does not seek to educate or mobilize the population in the fight against corruption. Like the ICAC, it has played a deterrent role by investigating a number of “big fish”, such as ministers, MPs, and senior directors in government agencies and companies (see Meagher [2005] for a critical review).

1ACAs

are frequently described as starting with the establishment of Singapore’s commission in the 1930s. In fact, a similar model began operating in New York City in the 1870s.

234     K. Sekkat

2 Actual Impacts of ACAs on Corruption Although not dealing with ACAs per se, Olken (2007) shows the role of public monitoring in fighting corruption. The paper is based on a randomized field experiment in over 600 Indonesian villages. The experiment concerns road construction projects which were part of a nationwide village-level infrastructure project. The data in the paper cover two of Indonesia’s most populous provinces, East Java, and Central Java and were collected between September 2003 and August 2004. Corruption in this context can take several forms which are, in general, based on collusion between the implementation teams, the village head, and suppliers to inflate costs. Accordingly, several mechanisms are put in place to ensure proper use of project funds. A first mechanism consists in village-level accountability meetings. Funds are released to the implementation team in three tranches: 40, 40, and 20%. In order to obtain the next tranche, the implementation team must present a report in an open village meeting explaining how all funds were used. Only after that meeting has approved, the report is the next tranche of funds released. Similar approaches relating to participation in future projects are imposed: A final cumulative accountability report must be presented at the end of the project and must be approved by a village meeting. A second mechanism consists of audits of selected projects by an independent government development audit agency. Each village-level project in this study has about a four-percent baseline chance of being audited. Auditors come to these selected villages and cross-check all the financial records looking for irregularities. They also inspect the physical infrastructure. The findings of the audit can potentially lead to criminal action. Often, however, officials found to have stolen funds are forced to publicly return the money. This leads to substantial social sanctions. The evidence from the study suggests that increasing the probability of audits substantially reduces funds stolen during the project. An increase in the probability that a village will be audited from 4 to 100% reduces missing money from 27.7 percentage points to 19.2 percentage

10  Specialized Anti-corruption Agencies     235

points. Increasing mass participation (individual citizens) in monitoring reduces missing expenditures only in specific cases, such as labor and materials. Finally, distributing anonymous comment forms to citizens reduces missing expenditures only if the distribution bypasses village officials who may have been involved in the project. Accordingly, mass participation in monitoring should be designed to prevent capture by local elites. Assessing the performance of Anti-Corruption Agencies is a challenging task. On the one hand, their missions are, in general, broadly defined and hence difficult to assess. On the other hand, even if the objectives are more concretely defined, the data on outputs and intermediate outcomes are highly imperfect. Keeping such caveats in mind Meagher (2005) tries to assess the accomplishments of ACAs in different countries.2 The author examines how good ACAs are at what they are supposed to do. Having collected information on some 30 ACAs, Meagher identifies six functions as commonly performed by the agencies. These are: receive and respond to complaints; intelligence, monitoring, and investigation; prosecutions and administrative orders; preventive research, analysis, and technical assistance; ethics policy guidance, compliance review, and scrutiny of asset declarations; and public information, education, and outreach. However, the available data mean that only a portion of these functions can be examined. The data come from ACA reports and different published papers. Moreover, the years of observations may differ. The analysis is conducted on an annual average basis. Subject to caveats, the author draws the following conclusions: ACAs in Hong Kong, Australia/New South Wales, Malaysia, and Singapore have been significantly more successful than other agencies. They are actually adding value in anti-corruption terms, and this contributes to the strong governance ratings of those countries. These results do not appear to be achieved through multi-agency cooperation in the absence of an ACA. The analysis broadly indicates that the ACAs in these

2Argentina,

Australia/New South Wales, Botswana, Ecuador, South Korea, Malaysia, the Philippines, Tanzania, Thailand, and Uganda.

236     K. Sekkat

countries have managed to cope with coordination, information, and leadership constraints that a multiple-agency approach might not have been able to overcome. It also appears that an ACA’s success depends largely on cooperative relationships with other elements of government. However, this is rarely the case, and such relationships sometimes break down where they have been established. As a result, ACAs are regularly frustrated by their inability to secure information, cooperation, prosecutions, and so on. In contrast to the previous paper, Quah (2009) focuses on the means and credibility of seven Asian ACAs.3 Means are measured using per capita expenditure and the staff to population ratio. This makes it possible to assess whether the ACAs are provided with adequate personnel and budget to perform their functions. Credibility is based on four indicators. These include whether the agency considers all complaints, public perceptions of the ACA’s professionalism, enforcement of anti-corruption laws, and the public image of the agency. However, only the first indicator (consideration of all complaints) was analyzed for all the ACAs under examination. The effects are considered using three indicators: Transparency International’s 2008 CPI score, the World Bank’s Control of Corruption Index (CCI), and Political and Economic Risk Consultancy’s (PERC) survey on corruption. These indicators reflect the effectiveness of the countries’ anti-corruption strategies which are implemented by their respective ACAs. In term of means, Macao’s agency is the best funded, with per capita expenditure of US$21.72, ahead of Hong Kong’s, which has a per capita expenditure of US$12.14. Singapore’s agency has the third-highest per capita expenditure (US$1.79). In terms of staff, Macao’s agency is first once again with a staff-population ratio of 1 per 4358 inhabitants, while Hong Kong’s has a staff-population ratio of 1 per 5863 inhabitants, and Singapore’s agency comes third with a staff-population ratio of 1 per 53,086 inhabitants. Thailand is fourth with a staff-population ratio of 1 per 69,481 inhabitants, and the Philippines is ranked fifth with a staff-population ratio of 1 per 85,057 inhabitants. India comes 3Singapore,

Hong Kong, Macao, India, South Korea, Thailand, and the Philippines.

10  Specialized Anti-corruption Agencies     237

sixth with a staff-population ratio of 1 per 229,505 inhabitants. The South Korea has the worst staff-population ratio, at 1 per 233,171 inhabitants, indicating that its agency is understaffed. Turning to credibility, the figures show that the proportion of complaints investigated increased from 78.3% in 2001 to 91.5% in 2004 in Hong Kong. In Singapore, the proportion decreased from 68.9% in 1999 to 47.6% in 2002. However, Macao showed a disappointing score with only 19% of complaints investigated during the period 2000– 2007. Regarding effectiveness, the ranking of the seven countries does not change much across the measures. Singapore and Hong Kong are ranked first and second, respectively, while the Philippines are always last. Doig et al. (2005) focus on the operating context and strategies of ACAs in five Sub-Saharan African countries. As in the preceding studies, here too the data used should be treated with caution. Out of the five countries, four (Uganda, Tanzania, Malawi, and Zambia) have a single agency, and one (Ghana) has two agencies: A Commission for Human Rights and an administrative justice body. The Commission for Human Rights is larger than the administrative justice body, but its main activities are not concerned with corruption. The ACA in Tanzania is by far the largest: It is nine times the size of Malawi’s agency. In Ghana and Uganda, the organizations are principally funded by their respective governments. The others largely rely on donor support. The dependency on funding by governments and donors suggests the ACAs cannot fully control their own strategies, staffing, and activities. In most cases, ACAs are dependent on other parts of government, usually the Attorney General’s office, for permission to prosecute. The analysis shows that, although investigation has long been seen as the primary purpose of African ACAs, the reality is that only a relatively small proportion of staff are involved in investigations. In Uganda, only about 10% of staff are involved in investigation while in the most specialized investigation agency (the administrative justice body in Ghana), the proportion is 50%. Moreover, all five countries have weak accountability, scrutiny, and monitoring arrangements. The anti-corruption architecture is ad hoc, poorly planned, and inadequately executed. Besides the issues of low staff skills (Tanzania) or the prevalence of a

238     K. Sekkat

culture of impunity (Uganda), funding seems to pose the most important problems. It often leads to donor intervention, donor-driven outputs, and uncertainty. One exception seems to be Ghana where the two agencies are fully funded by government. Donors add support at the margin but for activities that the agencies themselves have identified. Uncertainty about the possible unilateral decision of a major donor to stop funding the agency, as in Malawi in 2002, results in discontinuity in the development and promulgation of anti-corruption programs. The study concludes that, overall, the ACAs in question did not deliver “success”, but that they were not entirely responsible for this failure. A non-negligible part of the failure comes from the relationship with governments and donors.

3 Conclusion Because of the increasing complexity of the corruption phenomenon and congestion of the judicial system, many countries have created agencies specialized in the fight against corruption. The achievements of these agencies in curbing corruption differ highly across countries and depend on financial and human resources, independence from external pressure, accountability, and the scope and purpose of their assignments. The most successful examples are those of Singapore and Hong Kong. Available assessments of the effectiveness of different anti-corruption strategies show that, overall, the experience is not always a success. However, the failure is not entirely attributable to the agencies themselves, but comes mainly from the relationship with governments and donors.

References Charron, N. (2008, November 13–15). Mapping and Measuring the Impact of Anti-Corruption Agencies: A New Dataset for 18 Countries. Paper Presented at the New Public Management and the Quality of Government Conference, Göteborg, Sweden.

10  Specialized Anti-corruption Agencies     239

Doig, A., Watt, D., & Williams, R. (2005). Measuring ‘Success’ in Five African Anti-Corruption Commissions-The Cases of Ghana, Malawi, Tanzania, Uganda and Zambia. U4 Anti-Corruption Research Centre. https://www. Shareweb.Ch/Site/DDLGN/Documents/U4Report-2005-1.Pdf. Accessed 12 Sept 2008. Klemenčič, G., & Stusek, J. (2008). Specialised Anti-Corruption Institutions: Review of Models. Paris: OECD. Meagher, P. (2005). Anti-Corruption Agencies: Rhetoric Versus Reality. Journal of Policy Reform, 8(1), 69–103. Olken, B. A. (2007). Monitoring Corruption: Evidence from a Field Experiment in Indonesia. Journal of Political Economy, 115(2), 200–249. Quah, J. S. (2009). Benchmarking for Excellence: A Comparative Analysis of Seven Asian Anti-Corruption Agencies. Asia Pacific Journal of Public Administration, 31(2), 171–195.

11 Incentives and the Corruption Market

The analysis in the preceding chapters focused mainly on the “legal” responses to corruption. However, such responses have their own limits and might not be sufficient to achieve a significant reduction in corruption. Complementary approaches, generally associated with economics, have been put forward as tools to fight corruption. They are mainly based on changing the incentives and the “market structure” of institutional supply. They include wage fairness, staff rotation, meritocratic recruitment, and privatization. The evidence confirms the impact of wage increases on corruption reduction, but estimates suggest that relying only on this instrument to eradicate corruption necessitates a very high increase in wages. Moreover, wage increases should always be coupled with real monitoring. Regarding market structure, it appears that making the same service deliverable by different offices decreases corruption. Experiments also show that staff rotation and meritocratic recruitment can be effective in reducing corruption. Finally, privatization also appears to be a useful tool in curbing corruption, but requires autonomy and accountability.

© The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_11

241

242     K. Sekkat

1 Expected Impacts on Corruption 1.1 Incentives and Wages One complementary approach in the fight against corruption is to change the incentive schemes for civil servants. This consists, in general, in running public administrations with well-paid officials.1 Economists recommend a similar approach, which they call the “efficiency wage” method (Bardhan 2006). The wage paid is above the market-clearing level, which is supposed to enhance efficiency. In the context of corruption, higher wages imply higher costs for engaging in wrongdoing. A bureaucrat is expected to be reluctant to put a well-paying job at risk by taking part in corrupt acts. Conversely, underpaid bureaucrats have a greater incentive to solicit bribes as they can easily find an equivalent wage in the private sector. The generous public pension systems frequently seen around the world also act as a deterrent to corruption because pensions could be lost in the case of conviction for corrupt behavior. However, if a bribe is high and the risk of being caught is low, increasing officials’ wages might not help (Svensson 2005). Moreover, higher wages strengthen an official’s bargaining power and can thus lead to a higher bribe level if the official and the briber bargain over the amount of the bribe. Another issue with increasing wages concerns the size of the increase. Small changes are likely to have little effect on bribe demand. A large increase in public wages may simply shift corruption to the stage of selecting candidates for a civil servant position.

1.2 Public Service Delivery and the Bureaucrat’s Monopoly An alternative to changing wages is to change the “market structure” of institutional supply (Bardhan 2006). Quite often corruption occurs

1In imperial China under the Ching dynasty, district magistrates were paid an extra allowance called yang-lien yin, or “money to nourish honesty”.

11  Incentives and the Corruption Market     243

because bureaucrats have a “monopoly power” over the requested service. One solution is therefore to introduce more competition among officials in the supply of public services. If a citizen who is eligible for a given benefit can only obtain it from one office or one civil servant, he/ she has no choice other than to pay the bribe to get the desired service. If multiple bureaucrats can deliver the service, the citizen has more bargaining power because he/she can move to another bureaucrat to obtain the service for a lower bribe or no bribe. However, this is conditional on the absence of systemic corruption and collusion among bureaucrats. Staff rotation is another solution, since corruption is based on trust and reciprocity between the briber and the recipient (Abbink 2004). Repeated interactions between the two parties create a very favorable environment for bribery. With staff rotation, the potential briber has no previous experience with the new public official and hence will have difficulty in predicting the official’s reaction to the bribe offer. Similar uncertainty holds for the new official vis-à-vis the citizen. However, staff rotation can be costly. Long-term relationships between public officials and citizens have the benefit of relieving officials of the need to frequently adapt to new routines, cases, and rules (Abbink 2004). Moreover, the mechanism will only work if the delivery of a given service depends on one official only. If multiple officials must approve delivery, the rotation of one official at a time might be ineffective. The rotation of a high number of officials at the same time would be very costly to the state (Svensson 2005).

1.3 Shifting Service Provision to the Private Sector Instead of searching for an efficient incentive scheme targeting public officials, another solution to fight corruption consists in shifting service provision to the private sector. In practice, private firms have taken over basic service provision in parts of India, tax collection in Uganda, transportation in Mexico City, and parts of customs inspections in over 50 developing countries. However, shifting service provision to the private sector assumes that in this respect the private sector is able to do better than the public authorities. The assumption is not

244     K. Sekkat

necessarily well-founded because the functioning of the private sector can be impacted by market failures. Such failures cover a wide range of issues, such as information, incentives, and externalities. From the information angle, transferring the supply of public services to private firms can improve welfare only if the private sector can access and process the information more effectively than the authorities. In terms of incentives, the functioning of the private sector should prevent the risk of corruption by private employees. Finally, the private sector must take into account the effects of externalities better than the public sector does. In order terms, the potential net benefit from shifting service provision to the private sector depends on the relative importance of “government” versus “market” failures. Moreover, even if privatization can improve service delivery and reduce corruption, private actors in charge of such delivery must also be monitored and audited and, in turn, punished for any misbehavior. Otherwise, firms once thought to solve the corruption problem may ultimately promote corruption. Private actors may ultimately become more corrupt and parasitic than the government bureaucrats they replace. Honest and competent supervision of private sector activity is necessary and should accompany the shift of public services provision. If such control is not possible, it may be better to keep the service inside the public sector.

2 Actual Impacts on Corruption 2.1 Incentives and Wages The empirical analysis of incentives focuses on three variables: career, salaries, and depoliticization (Dahlström et al. 2012). Career considerations include meritocratic recruitment of bureaucracy through competitive formal examinations, special laws for public employment and career stability. This is supposed to create an “esprit de corps” which promotes socialization of certain values, strong ties among the members of the “corps” and isolation from external influences. Fair salaries are designed to ensure that officials do not engage in corrupt behavior

11  Incentives and the Corruption Market     245

to complement their salaries, while increasing the opportunity costs of being convicted for corrupt activities. Finally, depoliticization leads to a separation of interests. Politician and bureaucrats have separate interests because they answer to different chains of accountability. Rauch and Evans (2000) examine the effect of the components (competitive salaries, internal promotion, career stability, and meritocratic recruitment) of the three abovementioned variables on corruption. The analysis is based on data on these components for 35 less developed countries. The data were collected through a survey and concern the responses to a questionnaire which gives an assessment of the components over the period 1970–1990 as a whole. The collected information is used to explain the degree of corruption as measured by the ICRG index. Control variables include GDP per capita at the beginning of the time period and the level of education. The results show that meritocratic recruitment is the element of bureaucratic structure that is most important for decreasing corruption. Wages, internal promotion, and career stability are at best of secondary importance. Van Rijckeghem and Weder (2001) focus on the effect of civil service salaries on corruption in a sample of 31 developing and low-income OECD countries over the period 1982–1994. The dependent variable is the degree of corruption as provided by the ICRG. The explanatory variable of interest is the ratio of government to manufacturing wages. Important control variables include the probability of detection (measured by the quality of the bureaucracy and the rule of law from the ICRG) and the Index of Political Rights and Civil Liberties drawn from Freedom House. Other control variables are per capita GDP and secondary school enrolment. The findings point to a statistically and economically significant relationship between relative wages and corruption in the long run only. However, the relationship implies that if one relies only on wages to eradicate corruption, the wage increase must be very high. Dahlström et al. (2012) consider simultaneously the three variables presented at the beginning of this section: career, salaries, and depoliticization. To construct these variables, data were collected in 52 countries through a country-expert survey completed by 520 public administration experts. The studied countries are of high or middle

246     K. Sekkat

income.2 This contrasts with the Rauch and Evans (2000) sample, which focused on developing countries. The dependent variable is the World Bank’s corruption index for 2008. Control variables are GDP per capita, level of education, degree of ethnolinguistic fractionalization, political fragmentation, and the proportion of Protestants in the population. The results show that, in line with Rauch and Evans (2000), meritocratic recruitment reduces corruption, even when controlling for a large set of alternative explanations, while public employees’ competitive salaries, career stability, and internal promotion do not have a significant impact. Le et al. (2013) use a similar approach to Dahlström et al. (2012) to study the relationship between government wages and corruption, but the dataset is larger, covering 113 countries over the period 1989–2010. The ICRG corruption index is used as the dependent variable and is explained in terms of the relative wages of government staff (computed as in Van Rijckeghem and Weder 2001) and other control variables. In particular, the income level in the country is introduced separately and interacted with government staff wages. The reason for introducing the interaction term is to examine whether the impact of government wages on corruption is different between rich and poor countries. A large number of other political, economic, and institutional factors are also introduced. The findings show that there is an impact of government wages on corruption, but that this impact depends on the level of per capita income. When income per capita is relatively low, higher government wages reduce corruption. For instance, if the average government wage relative to the average wage in the manufacturing industry increases from 100% to 200%, corruption decreases by about one point (out of a maximum of six) when the income level is between $1000 and $2000 (in 2012 prices). However, this impact shrinks as the level of income increases. In sum, higher government wages may not be an 2They include Western European, North American, and post-communist Eastern European countries along with seven other countries: India, Brazil, South Africa, Japan, South Korea, Mexico, and Turkey. The last four are members of the Organization for Economic Co-operation and Development (OECD).

11  Incentives and the Corruption Market     247

efficient policy tool to reduce corruption in upper-middle- or high-income countries with a relatively high level of corruption, such as Greece and Italy. In low-income countries, increasing wages may substantially reduce corruption. Foltz and Opoku-Agyemang (2015) took advantage of an ambitious public sector reform in Ghana in 2010 to investigate the impact of a wage increase on petty corruption. The reform consisted in doubling police officer salaries and increasing the enforcement of anti-corruption laws. The analysis is based on the results of surveys conducted between 2006 and 2012 and concerning long-haul trucks traveling back and forth from Ouagadougou, the capital of Burkina Faso, to the port town of Tema, Ghana. Seven types of stops are considered: customs, forestry, gendarmerie, health, police, and unions. Each stop for an individual driver represents a data point. The questions in the surveys concern: (i) the time an official used to ask for a bribe, (ii) the amount of bribes paid at each stop, (iii) the number of stops where no bribe is paid, and (iv) the total amounts paid on a road. The responses to these questions are used as dependent variables. The data also include information on road, time, country characteristics, country of origin of the driver and truck, truck type (tanker, container, and general purpose), truck value, and the driver’s education level. The analysis showed that policemen who received the salary increase allocate more effort to collecting bribes in terms of the time spent asking for bribes, the value of bribes they took and the total amount that truckers had to pay on the road. It appears therefore that higher civil service salaries encourage civil servants to demand higher bribes. Moreover, the enforcement of anti-corruption laws, which is supposed to go up with salary increases, remains lax. Accordingly, merely raising salaries without changing the intensity of enforcement does not produce a drop in corruption. Duflo et al. (2012) use a randomized experiment to test whether financial incentives together with monitoring can reduce teacher absence and increase learning in Indian Non-Formal Education (NFEs) centers. These centers are run by non-governmental organizations and local government. The focus is on centers managed by one of India’s leading development non-profit organizations, Seva Mandir, which runs

248     K. Sekkat

about 150 NFEs in the tribal villages of Udaipur, Rajasthan. Udaipur is a sparsely populated and hard-to-access region, which makes it difficult to regularly monitor the NFEs. Accordingly, absenteeism is high. In September 2003, Seva Mandir implemented an external monitoring and incentive program on an experimental basis. The program covered 120 schools, of which 60 were randomly selected as the treatment group and the remaining 60 as the comparison group. In the treatment schools, each teacher received a camera. One of the students was tasked with taking a photo of the teacher and the other students at the start and at the end of each school day. The cameras had a tamper-proof date and time function which made it possible to precisely track each school’s openings and closings. Rolls were collected every two months at regularly scheduled teacher meetings, and payments were distributed every two months. At the start of the program, participating teachers’ monthly base salary was $23 for at least 20 days of work per month. In the treatment schools, teachers received a bonus of $1.15 for each additional day they attended in excess of the 20 days. In contrast, they were fined $1.15 for each of the 20 days they skipped work. In the comparison schools, teachers were paid a flat rate of $25 and were reminded that regular attendance was required and that they could, in principle, be dismissed for poor attendance. The experiment resulted in average teacher absenteeism of 42% in the comparison schools and 21% in the treatment schools. The students in treatment schools benefited from about 30% more instruction time. As a consequence, the program had a significant impact on test scores: 0.17 standard deviations higher in treatment than in comparison schools after one year. Children were also much more likely to be admitted to government schools. A similar experiment was conducted by the government of Rajasthan to combat nurse absenteeism. The program worked well in the early months of its implementation, achieving about a 50% reduction in absenteeism. After a few months, however, the government, while maintaining the monitoring, started granting a large number of “exemptions”. Absenteeism went up. It appears therefore that monitoring can be effective but should be coupled with real incentives.

11  Incentives and the Corruption Market     249

To further highlight the interplay between incentives and monitoring in reducing corruption, Barr et al. (2009) use an economic experiment focusing on the provision of public services by agents. The experiment consists of a game involving three categories of players: community members, a public service provider, and a monitor. The game is implemented in several rounds. The basic principles of the game are the following. Each round, the community members, a public service provider, and a monitor are assigned their roles randomly. The service provider and the monitor receive a predetermined salary at the end of each game. The public service provider receives some items to distribute to community members. He or she can decide to cheat (keep some items) or not. The monitor puts effort into discovering whether the service provider is cheating. If the monitor discovers that the provider has cheated, the latter receives his or her salary and is excluded from being the service provider in the next rounds; i.e., he or she loses future income. If the provider is not convicted of cheating, he or she receives the predetermined salary and can be selected randomly for the next round. The whole game is repeated under different rules in order to test certain hypotheses. First, instead of being randomly selected, the monitor is elected by the community in each round. The objective is to see whether community members re-elect monitors who put more effort into discovering cheats. Second, the degree of difficulty of discovering a cheating provider is changed. This makes it possible to assess how the ease of detection by monitors affects the probability of cheating by the service provider. Finally, two levels (high and low) of the service provider’s wage are considered. This makes it possible to test whether service providers receiving a higher wage perform better. The results show that when the monitor is randomly selected, monitors put no effort into detecting cheating service providers, who keep all the valuable tiles. However, when monitors are elected, service providers perform better and elected monitors put greater effort into monitoring. Moreover, service providers perform better when observability is higher and community members re-elect monitors who put more effort into exposing expropriation. However, the findings provide only weak evidence that public servants who receive a higher wage expropriate less.

250     K. Sekkat

Increasing the wage by 200% leads to a less than 30% reduction in resource expropriation. Finally, monitors put more effort into exposing underperformance by public servants who receive a higher wage. Azfar and Nelson (2007) extend the game proposed by Barr et al. (2009) to cases where both the officials and the monitor are elected. In this way, the analysis highlights another important aspect of the fight against corruption, that is, election of the executive. The difference between the players in the game proposed by Barr et al. (2009) and those in the paper in question is as follows. Community members are now voters, the service provider is the executive, and the monitor is the attorney. The executive is determined by popular vote. Corruption is defined as the number of valuable goods that an executive steals. Accountability is influenced by the costs and probabilities of being caught. The probability of being caught is, in turn, dependent on the ease of detecting corruption (transparency) and on the incentives faced by the law enforcement officer (separation of powers). Like in Barr et al. (2009), three changes are used to test the effects of accountability on corruption. First, the difficulty of hiding corruption is low, moderate, or high. Second, the wages of the executive and the attorney general are either high or low. Finally, the attorney general is either appointed by the executive or selected in a separate simultaneous election. The analysis shows that voters rarely re-elect chief executives found to be corrupt and that they reward presidents who had good luck by re-electing them. Directly elected law enforcement officers work more vigilantly at exposing corruption than those that are appointed. In particular, it appears that elected attorneys general collude less often with the executive. Increasing both government wages and the ease of detecting corruption reduces corruption. In this paper, the ease of detecting corruption is related to increasing transparency, which can be achieved through different means. These include improving accounting and audit systems, checking on bank accounts, regular declaration of assets of public sector employees, using reports from the media and public, and providing incentives for officials to report bribes. While the above studies are based on laboratory experiments, Di Tella and Schargrodsky (2003) present evidence based on a natural experiment. They take advantage of a crackdown on corruption that

11  Incentives and the Corruption Market     251

occurred in the city of Buenos Aires, Argentina, in 1996–1997. After a change of government, the new authorities implemented a policy of monitoring input prices paid by hospitals. Each hospital has one procurement officer. From October 1996 to December 1997 (the monitoring period), the government circulated a list showing the input prices paid by each hospital. The list highlighted the hospitals that paid the lowest and the highest prices for each product. The authors distinguished three periods: the first three months prior to the introduction of the monitoring policy, the first nine months after the introduction of the policy, and the last seven months of the observation period. During the first period auditing was low, during the second period auditing was at its maximum level, while during the third period auditing intensity declined relative to the second period. Using regression techniques, the authors examine the effect of procurement officer wages on the prices paid by the hospitals at different audit levels. The dependent variable is the price of the input bought by a given hospital during a given period. The explanatory variables include the wage of the procurement officer of the same hospital in the period, dummies pertaining to the strength of auditing and other control variables. The estimation results show no clear effect of wages on input prices. In contrast, monitoring has a clearer effect. Prices decreased by 14.6% in Period 2 relative to their original levels, but recovered by 5 percentage points in Period 3. Taken on their own, prices during Period 3 were still 9.7% lower than in the first. The results also show that the immediate effect of monitoring (Period 2) is stronger than its longer-term effect (Period 3). It appears that audit intensity is more effective than wages in deterring corruption. Borcan et al. (2014) present the results of another natural experiment based on an unexpected 25% wage cut applied in 2010 to all Romanian public sector employees, including public education staff. The analysis focused on a corruptible risk exam taking place shortly after the announcement of the wage cut. Since private schools were not affected by the policy, the corruption measure is based on a comparison of the changes in exam outcomes from 2009 to 2010 between public and private schools. The reasoning is the following. Before 2010,

252     K. Sekkat

exam outcomes are assumed to be inflated, for both public and private schools. Additionally, it is assumed that the incentives and level of corruption intensity for private schools should stay constant. The substantial wage loss for public school staff might prompt teachers to compensate for their forgone income by increasing the prevalence of corruption. The estimates show that the wage cut caused a disproportionate increase in average grades and passing rates in public high schools relative to private ones.

2.2 Public Service Delivery and the Bureaucrat’s Monopoly As explained previously, another solution to curbing corruption may be to break the monopoly enjoyed by bureaucrats. Different countries have taken this approach. For instance, the German federal government has adopted staff rotation for sensitive areas such as public procurement, a field where corruption is common (Abbink 2004). In Nepal, where traders were offered the possibility of using several points to pass through customs, they flocked to entry points where bribe levels were lowest. In the USA, a citizen can get a passport from almost any post office. In India, however, people can only go to one passport office, where officials have “monopoly power”, which they can exert to extract bribes. In another case, drug-related corruption in the New York Police Department was curbed by the involvement of officers from different agencies with overlapping jurisdictions (Bardhan 2006). Besides these examples, there are specific studies of the effect on corruption of competition among officials. Ryvkin and Serra (2015) use a laboratory experiment to examine the effectiveness of such competition among officials in the provision of the same good or service. Their focus was on extortion corruption; i.e., bribe demands for the provision of services that clients are entitled to receive. The experiment involved citizens applying for a license from one of many available offices. Officials decide whether or not to demand a bribe and the amount of the bribe simultaneously at the beginning of the period. In the first instance, citizens engage in a search whereby they visit an office at no cost and

11  Incentives and the Corruption Market     253

discover whether the official requires a bribe and the size of the bribe. Citizens then choose whether to pay the requested bribe or visit a different office by paying a fixed search cost. Citizens can get the license from any of the previously visited offices at no additional search cost. By changing the number of available offices and the size of the search cost, the authors seek to isolate the impact of increased competition among officials on both bribe demands and bribe payments. The analysis suggests that increasing the number of offices in charge of providing the same license may decrease extortion corruption depending on the size of search costs. If search costs are high, increasing the number of offices has no effect on bribe demands. A reduction in search costs while keeping the number of offices fixed unambiguously lowers bribes. Search costs can be reduced by improving infrastructures, such as roads or public transportation, or by promoting information-sharing about the size of bribes demanded by different officials. Sequeira and Djankov (2010) use a natural experiment to examine the effectiveness of competition among officials as an anti-corruption policy. The focus is on corruption in ports in Southern Africa. Specifically, two competing transport corridors connect South Africa‘s mining, agricultural, and industrial hubs to the ports of Durban in South Africa and Maputo in Mozambique. Some South African firms face the choice of using one of the two ports, especially since 2004 when the barriers for freight transit between South Africa and the port of Maputo were significantly reduced. However, freight travels long distances (around 588 km) between centers of production or consumption and ports. The choice of which port to use is therefore not trivial. The two countries are similar in many respects, including the level of red tape and the number of documents to process the clearing of goods through their ports. The two ports are also similar in terms of overland transport costs, cargo-handling technologies, and logistics services for standard cargo. They differ, however, in the levels of expected corruption. The port bureaucracies of Maputo and Durban differ in three important ways. First, in Durban all clearance documentation is processed online. This sets the level of direct interaction between clearing agents and customs agents at a minimum. In contrast, this level of interaction

254     K. Sekkat

is high in Maputo because all clearance documentation must be submitted in-person by the clearing agent. As a result, the number of opportunities for corrupt behavior is expected to be higher in Maputo. Second, while in Maputo port operators are privately managed, in Durban this is only the case for bulk cargo terminals. Container terminals are still under public control. Private management in Maputo and in the bulk terminals in Durban are expected to lead to less corruption, while publicly managed container terminals in Durban are expected to generate high bribes. Third, the two ports have different policies for customs official management. Customs in Maputo have a policy of rotating agents across different ports and terminals. Customs officials in Durban are subject to little or almost no rotation. The authors collected three datasets: (1) measures of transport costs on both the Maputo and Durban corridors, (2) measures of the level and frequency of bribe payments at each port, and (3) firm route choices and other financial information. The empirical analysis is based on a binomial probability model for the choice between the two corridors by each firm. The explanatory variables are firm location, the level of urgency of the shipments and the characteristics of the cargo that make it more or less vulnerable to paying a bribe in Maputo or in Durban. The findings show that if a South African firm ships goods that are subject to a high tariff classification in Mozambique, the probability of choosing Maputo declines by approximately 22–23%. Even when accounting for distance, the perishability and the urgency of the shipment as well as the expected bribe are a strong predictor of the choice of port. As an example, 46% of South African firms located in regions in which overland costs to the port of Maputo are 57% lower still go the long way round to Durban to avoid higher bribe payments. Of these, 75% are shipping perishable cargo, and 74% are shipping urgent cargo. A firm located in the town of Nelspruit, the capital of a province in northeastern South Africa, is 171 km from Maputo and 992 km from Durban. If this firm ships a high tariff good, it is 22% more likely to ship through Durban than through Maputo in spite of the 210% increase in overall costs. Firms that re-route to the least corrupt port incur an additional 8% increase in yearly transport costs in comparison

11  Incentives and the Corruption Market     255

with firms shipping cargos less vulnerable to corruption. Interestingly, the diversion costs of corruption for each individual firm are eight times higher than the actual bribes collected by customs officials in Maputo. This suggests a very high aversion among firms to paying bribes. Abbink (2004) examines the impact of staff rotation on corruption using a laboratory experiment. The experiment involved 18 persons split into two categories: a potential briber (typically a firm) and a public official. At the beginning of the experiment, the category of each participant is randomly drawn and remains unchanged throughout the experiment. The experiment consists of 30 rounds. In each round, pairs of players are matched randomly. Thus, the players do not know with whom they will play in a particular round. This rule is a way to capture the impact of staff rotation as compared with a game where the pairs are kept the same for all 30 rounds. Each round goes as follows. The firm decides whether and how much to transfer to the public official. If it decides not to transfer, both players go directly to the service delivery stage. If it does transfer, the public official accepts or rejects the bribe. If the public official rejects the bribe, both players go directly to the service delivery stage. If the public official accepts the bribe, the amount offered is deducted from the firm’s account. The amount is then multiplied by a factor of three before being credited to the official’s account. The multiplier reflects a difference in marginal utility: The same amount of money can be expected to mean much less to a large firm than to a public official with a small income. However, a lottery is played out to determine whether the players are considered to be caught or not. If not caught, the players go to the service delivery stage. If caught, both players are disqualified from the experiment and receive no money. This represents the consequences arising from discovery of corrupt activities, namely drastic fines and job loss. At the service delivery stage, the public official has two choices: i) option X (the “honest” option), which is, apart from eventual bribes, slightly preferable (as manipulating a decision requires effort to justify the choice before superiors) or ii) option Y (the “manipulating” option), which is much more favorable to the briber and entails a bribe that more than compensates for the effort required to manipulate the

256     K. Sekkat

decision. Moreover, option Y harms the public because, in this case, each of the other participants in the experiment suffers a deduction in his or her gains. The experiment shows that rotation is effective in reducing corruption. On average, bribes are reduced by almost one-half and, more importantly, the average frequency of inefficient decisions caused by bribery decreases even more strongly. The effect observed in the experiment is due to a lower tendency of firms to pay bribes as well as to a lower propensity of public officials to be influenced by them in favor of the briber. Moreover, the fact that bribers cannot reciprocate favorable decisions by paying bribes in later cases results in a significantly lower tendency to pay higher bribes after firms experience an advantageous decision.

2.3 Shifting Service Provision to the Private Sector Frequently cited examples of shifting public service provision to the private sector concern Autonomous Revenue Authorities (ARAs) and Pre-Shipment Inspection (PSI). The system is expected to have two main benefits: reduce corrupt practices and enhance tax collection. There are various opportunities for corruption in tax collection: undervaluation or under-declaration of goods, intervention by politicians to exempt supporters from taxes, or use of the tax administration to harass political opponents through audits (Martini 2014). ARAs were inspired by a radical program of public sector reform in the 1980s implemented by the UK, the USA, Australia, and New Zealand. Countries in Africa and Latin America followed with the establishment of semi-autonomous revenue authorities.3 The ARAs replaced the old model of tax collection, which was dispersed among a number of departments, did not work in a coordinated fashion and offered several

3Examples of countries that adopted ARAs are Malaysia, Singapore, Kenya, Malawi, Rwanda, South Africa, Tanzania, Uganda, Zambia, Bolivia, Guatemala, Guyana, and Mexico (Taliercio 2004).

11  Incentives and the Corruption Market     257

opportunities for rent-seeking and corruption. As a result, the amount of collected taxes was very low. The new systems aim to provide autonomy from political interference and efficiency in tax collection. ARAs are supposed to have independent management boards, although they are not as autonomous as a central bank, and an operational budget independent from the regular annual budgeting process. In order to attract more qualified and motivated staff and reduce incentives for corruption, ARAs enjoy more flexibility in hiring, paying, and managing staff. The results so far have been mixed. Studies from a number of countries in Latin America and Africa show that the reforms appeared to be successful in the initial years. However, in many cases, the successes were not sustained (see Fjeldstad and Moore (2009) for a deeper analysis). Pre-Shipment Inspection (PSI) programs have similar objectives to the ARAs but concern customs. First introduced in Zaire in 1963, PSI has been adopted by many countries worldwide (Anson et al. 2006). In total, over 50 developing countries have implemented PSI programs, at least for a given period of time (Rose-Ackerman 2006). Interestingly, countries with lower per capita GDP and more bureaucratic corruption are those which are likely to adopt PSI (Yang 2008). PSI is conducted by one or more firms hired by the government to inspect incoming shipments. Such inspections are typically initiated and supervised by the country’s finance ministry. In general, inspecting firms do not collect the import duties, which remains the responsibility of customs officials in the shipment’s destination country. Import procedures under PSI vary greatly, but the typical approach is roughly the following (Anson et al. 2006). The trader operating in the port of shipment must first provide the local PSI company with a detailed description of the shipment, which is then inspected. Upon inspection, the PSI company issues a Report of Findings. The latter falls into two categories: A Clean Report of Finding (CRF), when the PSI company confirms the trader’s declaration, or a Discrepancy Report (DR), when it raises the trader’s declared value. The CRF or the DR serves as the basis for determination of the applicable import-tax regime (tariff line, special regimes, exemptions, etc.) and is sent to the destination port’s customs. On the basis of these documents (CRF/DR and

258     K. Sekkat

customs documents), the PSI company calculates all taxes and duties, which are paid by the importer to a bank and transferred to the customs’ account at the Central Bank and then finally to the Treasury. For these duties, the PSI company adds a fee paid by the importer, typically about 1% with a minimum amount. Customs also sometimes perform independent inspections in addition to PSI. Practices vary widely across countries, with “second-inspection” rates ranging from 5% for some countries to 100% for others, such as Nigeria. As explained above, the privatization process is aimed at providing autonomy from political interference and improving efficiency in tax collection. However, the private bodies in charge of delivering public services should have autonomy in practice and not only on paper. Autonomy is the factor that enables politicians to make the commitment that the tax administration will be more effective, fair, and competent. The taxpayer needs to see that tax collection is effective, fair, and transparent. To this end, the ARA’s budget should be, as far as possible, a function of revenues collected, tax administrators should be trained professionals operating in a meritocratic organization, and the ARA superintendent should be free from political interference to pursue his mission. Taliercio (2004) focuses on the autonomy of the revenue authority in Latin America. The analysis is based on a survey of 200 randomly selected large corporate taxpayers and professional tax consultants carried out over an 11-month period in 1998–1999 in Bolivia, Mexico, Peru, and Venezuela. The responses served to construct indexes of ARA organizational autonomy, performance, and political commitment to reform. The indexes are then analyzed econometrically. The dependent variable is taxpayers’ perception of a credible commitment by the state. The independent variables include an indicator of the general stability and performance of government, tax policy complexity, insulation of agency management from politics, professionalism of personnel, funding levels, incentives, and funding mechanism. The results support the hypothesis that autonomy matters. Autonomy has an independent effect on perceptions of the political commitment to reform, even controlling for other variables, such as specific service quality, the wider institutional context, and fixed

11  Incentives and the Corruption Market     259

country effects. This indicates that autonomy is valuable in itself, and not only as a means to deliver better services. Moreover, higher autonomy ratings seem to signal the government’s commitment to reform. They also go with better organizational performance and better perceptions of administrative fairness. In other words, when taxpayers perceive the management of the tax agency as independent of political influence, they also perceive the tax administration as fairer. Sarr (2016) assesses the performance of 20 developing countries’ ARAs in terms of tax collection over the period 1980–2010. Only countries that had created such agencies before 2000 were included in the sample. The outcome variable is government revenue excluding grants and is drawn from the World Bank World Development Indicators. Revenue is receipts from taxes, social contributions, and other revenues such as fines, fees, rent, and income from property or sales. The determinants of revenue are per capita GDP, value added of the agriculture sector, the natural resource share of GDP, the debtto-GDP ratio, the degree of international trade, the investment rate, the money supply, foreign aid, the political regime, the ICRG quality-of-government index, the percentage of rural population, the existence of armed conflicts, the ethnic fractionalization rate, a proxy for the level of education and health, legal origin, and the country’s geographic localization. The results show that the creation of revenue agencies does not always produce the expected outcomes. Out of 20 countries analyzed, only five (Argentina, Bolivia, Guyana, Malawi, and South Africa) seem to sustainably outperform a traditional finance ministry in terms of revenue collection. In five other countries (Colombia, Guatemala, Rwanda, Uganda, and Zimbabwe), performances are mixed. Finally, in six countries (Kenya, Mexico, Peru, Tanzania, Venezuela, and Zambia), creating the agencies led to worse revenue collection. In this last set of countries, government revenue would have been higher if it had been kept under the finance ministry. Such contrasting outcomes suggest that establishing a revenue authority should not be viewed as a panacea. The difference in the performance of these revenue authorities suggests that other factors, such as the quality of staff and the degree of independence of the agency from the political authorities, might be crucial to success.

260     K. Sekkat

Turning to PSI programs, Yang (2008) analyzes the relationship between their implementation and import duty collections for the years 1980–2000. The author considers that this is an indirect way to assess the effect on corruption because the impact of PSI on import duty collections is due in large part to reductions in customs corruption. The sample contains 1372 observations from 104 countries, 19 of which are observed before and after the start of the PSI programs. The remaining countries serve as controls. The main question in the paper concerns the impact of PSI on collected import duties. This is the dependent variable and is explained in terms of the existence and age of countries’ PSI programs, other tax revenues (excluding import duties) and average import tariffs. The findings are that PSIs lead to large increases in import duties: 15–30% during the five years after implementation. The programs seem cost-effective: The improvements in import duty collections during the first five years represent 2.6 times program costs. Such improvement is the result of reduced falsification of import documentation, including declines in undervaluation and misreporting of goods classifications. Moreover, PSI programs seem to become more effective over time, since old programs tend to have larger impacts than the most recent ones. While the above paper suggests that PSI programs are, on average across countries, cost-effective, micro studies suggest that the conclusion may differ depending on the country. For instance, Anson et al. (2006) focus on the impact of PSI on under-invoicing in three countries (Argentina, Indonesia, and the Philippines). The empirical analysis is based on panels of imports to the three countries from the EU. Since tariff evasion, like all forms of fraud, cannot be measured directly, the empirical exercise is based on a comparison of the records of source and destination customs. Importers attempting to evade tariffs will under-declare the value of shipments at destination customs, not at origin. Accordingly, discrepancies between source and destination import values data will reflect, in addition to CIF/FOB differences and measurement errors, the extent of deliberate under-declaration. Using a proxy for CIF/FOB differences between source and destination, the authors investigate the extent of under-declaration in pre- and post-PSI years, respectively. The econometric results suggest that the introduction

11  Incentives and the Corruption Market     261

of PSI services increased under-invoicing in Argentina and Indonesia and reduced it in the Philippines. However, as mentioned by the authors, the Philippines case is not necessarily a proof of success for PSI. The econometric analysis cannot control for import deflection to dutyfree zones, which became a vehicle for tariff evasion in the Philippines after the introduction of PSI. Echoing this concern, Rose-Ackerman (2006) analyzes the microeconomic impact of PSI on tariff evasion in the Philippines and Colombia. In particular, the paper examines whether increased enforcement would lead criminal activity to be displaced to alternative lawbreaking methods. The discussion suggests that to be successful, anti-corruption reforms should be “broad” in the sense of encompassing a wide range of possible alternative methods of committing the illegal activity of interest. Otherwise, displacement to alternative methods can inhibit the original goals of the reform. Starting with the Philippines, the customs procedure was the following. Before 1990, shipments valued under US$5000 were exempt from PSI. A common method of avoiding the inspection was therefore to split shipments into pieces so that each one could be valued below that level. In the first half of 1990, the government reduced the value threshold first to $2500 and then to $500. Note that another characteristic pertaining to the Philippines is that only shipments from a subset of countries were subject to PSI in the first place. Focusing on imports with declared values between $2500 and $5000, a comparison of imports from countries subject to PSI and from those which are not shows a decline in the fraction of imports from countries subject to PSI. Moreover, during the same period, shipments via export processing zones from the countries subject to PSI increased. Imports from countries subject to PSI seem to have been encouraged to take advantage of the exemption for export processing zone shipments from these countries. Increased enforcement reduced the original method of duty avoidance (valuation under the old minimum value threshold), but led to substantial displacement to an alternative duty-avoidance method (shipping via duty-exempt export processing zones). Overall, estimates suggest that the minimum value threshold reductions led to significant

262     K. Sekkat

net tariff revenue losses (net of PSI fees) of around $36.8 million for the Philippine government. In Colombia, when the government implemented its PSI program, it required PSI only for a defined subset of products (“PSI products”). The fact that the Colombian PSI program required inspections for only a subset of products enabled importers to evade import duties by misclassifying imports in non-PSI product categories. To measure how duty avoidance operated after PSI was introduced, a ratio comparing Colombia’s reported imports of a given product to other countries’ reported exports of the same product to Colombia is constructed. The results show substantial displacement of PSI products through misclassification to enable duty avoidance.

3 Conclusion Because the “legal” responses to corruption discussed in previous chapters have their own limits and might not be sufficient to significantly reduce corruption, complementary approaches have been suggested. These approaches propose changing the incentives and the “market structure” for the supply of institutional services. They include wage fairness, staff rotation, meritocratic recruitment, and privatization. Assessments of the effectiveness of these proposals suggest that meritocratic recruitment is the element of bureaucratic structure that is most important for decreasing corruption. Wages, internal promotion, and career stability are at best of secondary importance. Although the relationship between wages and corruption is statistically significant, it seems economically infeasible. For wages to eradicate corruption, their increase needs to be very large. Breaking the monopoly enjoyed by bureaucrats by increasing the number of offices in charge of providing the same services decreases corruption. Staff rotation is very effective in reducing corruption, making it possible to decrease the average amount of corruption by almost one-half. More importantly, the average frequency of inefficient decisions caused by bribery decreases by even more. The shift of public service provision to the private sector appears to be successful in the initial years of privatization. However, in many

11  Incentives and the Corruption Market     263

cases, the success is not sustained. Crucial to success are independence from the political authorities, monitoring, and staff quality.

References Abbink, K. (2004). Staff Rotation as an Anti-Corruption Policy: An Experimental Study. European Journal of Political Economy, 20(4), 887–906. Anson, J., Cadot, O., & Olarreaga, M. (2006). Tariff Evasion and Customs Corruption: Does Pre-Shipment Inspection Help? The BE Journal of Economic Analysis and Policy, 5(1), 1–26. Azfar, O., & Nelson, W. R. (2007). Transparency, Wages and the Separation of Powers: An Experimental Analysis of Corruption. Public Choice, 130(3), 471–493. Bardhan, P. (2006). The Economist’s Approach to the Problem of Corruption. World Development, 34(2), 341–348. Barr, A., Lindelow, M., & Serneels, P. (2009). Corruption in Public Service Delivery: An Experimental Analysis. Journal of Economic Behavior & Organization, 72(1), 225–239. Borcan, O., Lindahl, M., & Mitrut, A. (2014). The Impact of an Unexpected Wage Cut on Corruption: Evidence from a “Xeroxed” Exam. Journal of Public Economics, 120, 32–47. Dahlström, C., Lapuente, V., & Teorell, J. (2012). The Merit of Meritocratization: Politics, Bureaucracy and the Institutional Deterrents of Corruption. Political Research Quarterly, 65(3), 656–668. Di Tella, R., & Schargrodsky, E. (2003). The Role of Wages and Auditing During a Crackdown on Corruption in the City of Buenos Aires. Journal of Law and Economics, 46(1), 269–292. Duflo, E., Hanna, R., & Rya, S. P. (2012). Incentives Work: Getting Teachers to Come to School. American Economic Review, 102(4), 1241–1278. Fjeldstad, O. H., & Moore, M. (2009). Revenue Authorities and Public Authority in Sub-Saharan Africa. Journal of Modern African Studies, 47(1), 1–18. Foltz, J. D., & Opoku-Agyemang, K. A. (2015). Do Higher Salaries Lower Petty Corruption? A Policy Experiment on West Africa’s Highways. International Growth Centre (IGC). London, UK.

264     K. Sekkat

Le, V. H., De Haan, J., & Dietzenbacher, E. (2013). Do Higher Government Wages Reduce Corruption? Evidence Based on a Novel Dataset (Cesifo Working Paper No. 4254). Martini, M. (2014). Approaches to Curbing Corruption in Tax Administration in Africa. Transparency International, U4 Anti-Corruption Resource Center, No. 11. Rauch, J. E., & Evans, P. B. (2000). Bureaucratic Structure and Bureaucratic Performance in Less Developed Countries. Journal of Public Economics, 75(1), 49–71. Rose-Ackerman, S. (Ed.). (2006). International Handbook on the Economics of Corruption. Cheltenham: Edward Elgar. Ryvkin, D., & Serra, D. (2015). Is More Competition Always Better? An Experimental Study of Extortionary Corruption (WP 2015_10_01). Florida State University, Department of Economics. Sarr, B. (2016). Assessing Revenue Authority Performance in Developing Countries: A Synthetic Control Approach. International Journal of Public Administration, 39(2), 146–156. Sequeira, S., & Djankov, S. (2010). An Empirical Study of Corruption in Ports (MPRA_Paper_21791). Svensson, J. (2005). Eight Questions About Corruption. Journal of Economic Perspectives, 19(3), 19–42. Taliercio, R. R. (2004). Administrative Reform as Credible Commitment: The Impact of Autonomy on Revenue Authority Performance in Latin America. World Development, 32(2), 213–232. Van Rijckeghem, C., & Weder, B. (2001). Bureaucratic Corruption and the Rate of Temptation: Do Wages in the Civil Service Affect Corruption and By How Much? Journal of Development Economics, 65(2), 307–331. Yang, D. (2008). Integrity for Hire: An Analysis of a Widespread Customs Reform. Journal of Law and Economics, 51(1), 25–57.

12 International Cooperation

This chapter examines the role of international initiatives and cooperation in curbing corruption. The analysis is limited to the three most important initiatives, which are the US Foreign Corrupt Practices Act, the OECD Convention Against Corruption, and the United Nations Convention Against Corruption. The chapter also examines the validity of fears among the international business community that these initiatives put such businesses at a significant competitive disadvantage. The empirical results do not lend strong support to these fears. In contrast, the findings suggest that international programs against bribery abroad can be effective in making investors more sensitive to host country corruption. However, recovering assets stolen by former leaders (Suharto, Mobutu, Marcos, and others) seems much more difficult and calls for better international legal cooperation and harmonization.

1 Instruments of International Cooperation 1.1 Foreign Corrupt Practices Act (FCPA) The FCPA was passed by the US Congress in December 1977 and has been amended twice. The 1988 amendment was a response to © The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_12

265

266     K. Sekkat

complaints by US corporations, which found the original act too vague and wide in scope. The FCPA was again amended in 1998 to extend application to “any person” over which the US Department of Justice (DoJ) has jurisdiction. The FCPA adopts a twin approach to counter bribery: accounting and penalization. The accounting approach requires all corporations that have stock registered with the Securities and Exchange Commission (SEC) to keep accurate books and accounts of all transactions and to adopt a system of internal accounting controls. The penalization approach prohibits US corporations and their agents from making payments to foreign officials in return for their influence to help the corporation. The SEC is responsible for enforcing the accounting requirements, while the SEC and the DoJ share responsibility for enforcing the anti-bribery provisions. In August 2012, the SEC adopted strict transparency rules regulating payments by oil and gas companies and firms in extraction mining industries. While the previous treatments of corruption focused on the demand side (e.g., public officials who accept or require bribe payments), the FCPA also adopted a supply-side approach (i.e., foreign corporations offering to pay bribes). There are significant penalties for violating the FCPA. A person found guilty of violating this act may be fined up to $100,000 and/or receive a five-year prison sentence. A corporation may be fined up to $2,000,000 for violating the anti-bribery provisions of the FCPA. There are also provisions with serious commercial consequences, such as suspension or denial of export licenses or loss of the privilege of doing business with the government (La Roche and Flanigan 2011). Until 2004, the DoJ had two choices when a business organization was the subject of an FCPA investigation. It could either charge the entity with an FCPA violation or not. In 2004, the DoJ started using two additional instruments as alternative resolution vehicles: NonProsecution Agreements (NPAs) and Deferred Prosecution Agreements (DPAs). NPAs permit non-prosecution in exchange for cooperation. They concern cases where a corporation’s timely cooperation appears necessary to the public interest and where other means of obtaining the desired cooperation are unavailable or not effective. Under a DPA,

12  International Cooperation     267

the DoJ agrees to defer prosecution of the alleged crime (usually for an eighteen-month to three-year period). In exchange, the company acknowledges responsibility for the alleged conduct and agrees to implement a set of compliance actions (Koehler 2015). The use of alternative resolution vehicles to resolve FCPA inquiries was motivated by the need to avoid crippling business organizations of critical importance to the economy. In the 2000s, the DoJ therefore moved from its historical binary option approach to resolving corporate criminal liability to a more flexible one. Although it was frequently claimed that alternative resolution vehicles were not the norm in corporate investigations, between 2006 and 2007 there were 12 corporate criminal FCPA enforcement actions and 100% of the enforcement actions involved either an NPA or DPA. Moreover, since the DoJ first used the alternative resolution vehicles in December 2004, there have been 84 criminal FCPA enforcement actions against business organizations, and 70 of these enforcement actions (approximately 85%) have involved an alternative resolution vehicle. In contrast, available data show that between 1977 and 2004 (the year when alternative resolution mechanisms were first used), there were 24 FCPA enforcement actions against business organizations and 20 of these enforcement actions (83%) involved criminal prosecutions of company employees (Koehler 2015). Since its inception, the FCPA has been criticized for placing US firms at a competitive disadvantage based on the fact that few countries forbid bribery of foreign officials. To compound the problem, many countries permit tax deductions for corrupt payments made in connection with foreign trade. In this context, US firms feared losing overseas business opportunities, particularly in emerging markets where corruption is rampant. In an effort to address this concern, the USA started putting diplomatic pressure on its international partners to prohibit bribery of foreign officials and politicians. As a consequence, several international treaties were ratified that seek to criminalize bribery and eliminate the tax deductibility of corrupt payments made to foreign public officials (La Roche and Flanigan 2011).

268     K. Sekkat

1.2 United Nations Convention Against Corruption (UNCAC) In 2003, the UN General Assembly adopted the first global response to anti-corruption, the United Nations Convention Against Corruption (UNCAC). The convention entered into force in 2005 with 140 signatories. Each new member to the UN or a new regional integration organization becomes de facto a party to the convention. This new affiliation enters into force on the 30th day after the date of deposit by such state or organization of the relevant information. The main additions of the UNCAC concern international monitoring, cooperation, prosecution, and criminalization and the fight against laundering. In particular, the convention covers: • Prevention: It includes preventive policies, such as the establishment of anti-corruption bodies and enhanced transparency in the financing of election campaigns and political parties. • Criminalization: It requires countries to establish criminal and other offenses to cover a wide range of acts of corruption. • International cooperation: Countries agree to cooperate with one another in every aspect of the fight against corruption, including prevention, investigation, and prosecution. • Asset recovery: Countries agree on asset recovery which is a particularly important issue for many developing countries, where high-level corruption has plundered the national wealth, hampering the reconstruction and rehabilitation of societies.

1.3 Other Anti-Corruption Initiatives Significant rules have been established by supranational organizations including the Organization of American States (OAS), the Organization for Economic Co-operation and Development (OECD), and the World Bank (WB) in order to establish moral business conduct abroad. In 1997, the WB adopted a comprehensive strategy aimed at preventing corruption associated with the projects that it finances. As part of its

12  International Cooperation     269

anti-corruption strategy, the WB also decided to help countries around the world in fighting corruption and to support international efforts with a similar target. Meanwhile, in 1997, ministers of the OECD countries along with five other countries (Argentina, Brazil, Bulgaria, Chile, and Slovakia) signed the Convention on Combating Bribery of Foreign Public Officials in International Business Transactions (known as the OECD Convention). This convention came into effect in 1999, and its provisions criminalize bribery and seek to eliminate tax deductions for corrupt payments made to foreign public officials (La Roche and Flanigan 2011). At the regional level, the OAS adopted in 1996 the Inter-American Convention Against Corruption (IACAC), which was the first pan-regional anti-corruption treaty. Unlike other conventions, the IACAC was not modeled on the US FCPA. The IACAC signatory countries agreed on specific preventative and remedial measures to eliminate corruption. They also implemented mechanisms to monitor compliance with the convention: Each country must report on its progress in implementation, evaluate adequacy and assess achievements. In 1995, the EU started with modest anti-corruption instruments which mainly concerned the misuse of EU funds. Gradually, the EU broadened its focus and adopted a comprehensive two-year process for reviewing Member States’ anti-corruption achievements. Last but not least, in 2003 the African Union Convention on Preventing and Combating Corruption (AU Convention) was adopted and entered into force around three years later.

2 The Debate on the Costs and Benefits of These Instruments Many US commentators, politicians, and business leaders have argued that the FCPA harms US companies. These businesses, they claim, are put at a significant competitive disadvantage because they are in competition on overseas markets with foreign counterparts which are not subject to the FCPA and are thus free to pay bribes. Some evidence supports

270     K. Sekkat

the claim that the FCPA has caused US companies operating abroad to lose a substantial amount of business or export opportunities to foreign competitors. One early study by the US Department of Commerce reported 21 embassies as saying that the FCPA was an export disincentive in the countries in which they were located. In 1981, a General Accounting Office study looked at 250 US firms and found that 30% reported losses of foreign business as a result of the act. More recently, a 1995 Central Intelligence Agency report estimated that over 1994– 1995, the USA lost $36 billion in international business to bribe-paying international competitors. The most high-profile bribery-related proceedings concerned Siemens AG and three of its affiliates, which were suspected of paying substantial bribes to officials in countries including Argentina, Bangladesh, Iraq, and Venezuela. To settle the allegations, the firms paid penalties totaling over $1.7 billion to the authorities in the USA and Germany and to the World Bank (Davis 2009). However, other arguments mitigate the above discussed negative impact of international anti-corruption programs on business. First, by forcing corporations to comply with anti-bribery laws, these programs may push businesses to improve their competitiveness through more “healthy” methods. Thus, in order to win contracts without paying bribes, they might focus on the quality of their goods and services. Such an increase in quality would also generate benefits on non-corrupt markets. Second, international anti-corruption programs can help companies by providing an excuse not to pay bribes. Some observers have suggested that the FCPA allows companies to save face while refusing to pay bribes and thus insulates them from the costs of bribery. For instance, Colgate-Palmolive cited the FCPA and its prohibition of bribery in response to demands for bribes from Chinese officials regarding the construction of a factory. The factory ultimately opened in 1992 without any bribes being paid. Finally, the provisions under international anti-corruption programs have allowed a number of countries whose head of state stole national assets to recover some or all of these assets, although this is not an easy task. For instance, Switzerland returned approximately US$700 million misappropriated by Ferdinand Marcos to the Philippines. However, other cases, such as those of Mobutu and Jean-Claude Duvalier, show that asset-recovery proceedings

12  International Cooperation     271

can be hampered by political influence and a poorly functioning judiciary in the country of origin (Transparency International 2007).

3 Actual Impacts of International Cooperation Empirical analyses of international anti-corruption programs focus mainly on the FCPA and two if its consequences: (i) whether, by putting firms at a disadvantage abroad, the programs have reduced international trade and investment and (ii) whether the programs have been successful in reducing corruption. Regarding the first issue, Graham (1984) investigates whether the FCPA resulted in a reduction in the share of imports of foreign countries from the USA. The analysis is based on the results of a report by the US Commerce Department in 1980 regarding export promotion and export disincentives. Responses were received from 51 embassies, representing countries accounting for 80% of total US exports in 1979. Based on the responses, two classifications are applied to the 51 countries. The first classification takes the form of a dichotomous variable separating countries where the FCPA was not mentioned as an export disincentive from those where the FCPA was mentioned as an export disincentive. The second classification takes the form of an ordinal variable which evaluates the impact of the FCPA on a four-point scale: major, significant, minor, or no impact. The analysis examines the relationship between each of these two variables and the changes in the share of imports from the USA between 1977–1978, 1978–1979, 1979–1980, 1977–1979, and 1977–1980. The results lend no support to the “competitive disadvantage” hypothesis. No statistically significant differences between US trade performances in the various groups (embassy responses) are discovered. Instead of US exports, Graham and Stroup (2016) focus on bilateral fixed capital flows from January 1, 1990 to January 1, 2010. The analysis is based on regression techniques where the dependent variable is the number of acquisitions announced by US acquirers in year t of firms headquartered in a foreign country j. The explanatory variable of interest is a dummy which takes a value of unity in all years following

272     K. Sekkat

the announcement of an FCPA enforcement action against the acquisition announced by a US company of a firm headquartered in country j. Control variables include the distance between Washington, DC and country j ’s capital, whether the countries share a common language, share a physical border, are members of the World Trade Organization and country j ’s GDP. The findings show that anti-bribery enforcement in a country is followed by a 40% reduction in foreign fixed capital investment by US companies in that country. One reason behind the reduction in foreign investment in corrupt countries is that FCPA enforcement increases the cost of doing business for affected firms. This increase stems from accounting controls and compliance programs imposed on firms. Even firms which are not contemplating bribery may incur an increased cost of doing business due to the cost of compliance. To examine this question, Lippitt (2013) conducts an empirical analysis of the relationship between FCPA violations and the growth of US foreign investment flows. If the FCPA has harmed US investment abroad, one would expect a negative correlation between US foreign investment growth and FCPA enforcement. Accordingly, the dependent variable is the growth rate of US foreign direct investment flows from 2000 to 2011. The explanatory variable of interest is the frequency of prosecuted FCPA violations during the same period. Several control variables were used. The results do not show a statistically significant correlation between the growth of US foreign direct investment and the frequency of FCPA violations in a given country. Thus, these results do not provide support for the hypothesis that US foreign direct investment was negatively affected by the FCPA. This conclusion is not in line with Graham and Stroup (2016). One reason for the divergence might lie in the measure of foreign investment. Graham and Stroup (2016) use the number of acquisitions announced by US acquirers while Lippitt (2013) considers the amount in US$ of foreign investment. In addition to the FCPA, Cuervo-Cazurra (2008) considers the impact on bilateral fixed capital flows of the OECD Anti-Bribery Convention. The study covered 103 economies between 1996 and 2002. Bilateral investment data are used to construct the dependent variable and come from the United Nations Conference on Trade and

12  International Cooperation     273

Development (UNCTAD). The main explanatory variables are the Transparency International (TI) and the WB measures of corruption. A large number of control variables are considered including GDP, population, geographic characteristics, languages, and colonial histories. To test the effectiveness of the FCPA, a distinction is made between the USA as the country of origin of the investor and other countries of origin. The logic behind the test is that if the FCPA has been effective, the investment entering a given country from the USA should be lower after 1977 (date of FCPA adoption) than investment from other countries. A similar idea is applied to investment originating from countries which have signed the OECD Anti-Bribery Convention and which have domestic laws against bribery abroad. The results regarding the effectiveness of the OECD Anti-Bribery Convention show that investors from countries that implemented the convention become more sensitive to host country corruption and reduce their investments in corrupt countries. Similar results hold for the FCPA. US investors are more sensitive to host country corruption than other investors. In sum, the results support the idea that international programs against bribery abroad can be effective in making investors more sensitive to host country corruption Using firm-level data, Trzcinski (2012) mitigates, however, the effectiveness of the FCPA. The paper examines the reaction of companies that have been subject to FCPA enforcement. The considered reactions concern whether a firm withdraws from or stays in a country if it has received a sanction because of an alleged FCPA violation. The dataset is compiled from press releases announcing the resolution of given matters and court records on complaints and litigation. It includes all companies that went through enforcement proceedings related to anti-corruption laws and conventions between 2000 and 2010. The sample is not limited to US companies because the jurisdiction of the FCPA has been expanded to include foreign companies that are US issuers1 and non-US persons when the conduct takes place while in the territory of the USA. 1An issuer is a company that is listed on a national securities exchange in the USA (either stock or American Depository Receipts) or the company’s stock trades in the over-the-counter market in the USA, and the company is required to file SEC reports (Trzcinski 2012).

274     K. Sekkat

The dependent variable measures the response and takes the value 1 if the company continues operating in the country and 0 if it divests. The explanatory variables include the host country’s GDP and natural resources abundance as well as proxies for the probability of detection of corrupt activities by FCPA regulators and the sanction related to being caught. The results show that companies involved in FCPA enforcement actions do not generally divest from the countries where they have been sanctioned for alleged violations. In over 70% of cases, the company continues to do business in the country in question. The findings also reveal that companies headquartered in countries with the cleanest corporate cultures continue to operate in implicated countries. However, the marginal effects of home country culture exhibit diminishing incremental value. Comparing a company from a country with a low bribery index score of 7 (e.g., Mexico) to another from a country with a more moderate level of 7.5 (e.g., Taiwan) shows a 15% increase in the probability of continuing operation in an implicated country. However, the difference between a company from France (with an index score of 8) and a company from Canada (with an index score of 8.5) results in an increase in probability of only 4.5%. In contrast to the above papers, Lippitt (2013) directly examines whether the FCPA has decreased corruption. The logic behind the question is that the substantial fines imposed on firms convicted of FCPA violations should discourage companies from entering corrupt deals. This implies a negative correlation between FCPA enforcement and corruption growth. The data used range from 63 to 84 US partner countries. The growth rate of corruption from 2000 to 2011 is the dependent variable. The measure of corruption is derived from TI’s annually published Corruption Perceptions Index (CPI). The growth rate of corruption is explained in terms of FCPA violation frequency during the same period. Over this period, there were a total of 324 bribery suspicions prosecuted under the FCPA. Violation frequency is assigned to the country where the bribe payment occurred. Only actions against legal entities (not individuals) are included in the analysis. Actions against a subsidiary are also included but separately counted. Actions that were dismissed are omitted. Control variables include the host country GDP, a measure of democracy at the

12  International Cooperation     275

beginning of the period, and the host country corruption score at the beginning of the period. The results show a significant negative correlation between prosecuted FCPA violations and corruption growth. This implies that the FCPA might have helped to decrease the perception of corruption in foreign countries. The results also suggest that GDP growth is negatively correlated with corruption growth. Democracy is significantly and negatively correlated with corruption growth, which suggests that democratic countries tend to be less corrupt. Like the previous paper, Samanta and Sanyal (2016) investigate directly the impact of international programs on the propensity of firms to give bribes. However, the focus here is on the OECD Convention on Combating Bribery. The sample uses two groups of countries: those which have signed the OECD Convention and those which have not. TI’s Bribe Payers Index (BPI), which indicates the perceived propensity of firms based in a particular country to give bribes to foreign government officials, is used. The main analysis is based on non-parametric tests such as the Kolmogorov-Smirnov test and the empirical distribution of the two groups of countries. The tests compare the BPI distributions of the two groups of countries. The results of both types of tests show that these two groups of countries differ significantly in their BPI scores. The results support the hypothesis that firms from countries that are signatories of the OECD Convention are less likely to give bribes than firms based in other countries. In other words, the OECD Convention has had the intended effect of reducing bribe-giving by firms from member countries when conducting international business

4 Conclusion To be confident of retrieving money which they have obtained through illegal acts, politicians, rulers, and high-ranking civil servants put this money outside their country, while at the same time establishing highly sophisticated systems to make it very hard to track down the embezzled money. These systems often overlap with other illegal activities such as the financing of terrorist acts, the covering up of narcotics trafficking,

276     K. Sekkat

and the embezzlement of development aid. This situation means that international cooperation is increasingly required. One of the first initiatives in this respect was the adoption in 1977 by the US Congress of the Foreign Corrupt Practices Act (FCPA), which prohibits transnational bribery. Since then, the international approach to corruption has gained broader coverage and now involves various international institutions. Many business leaders have, however, argued that the FCPA puts US firms at a significant competitive disadvantage because of competition from foreign counterparts which are not subject to the FCPA and thus free to pay bribes. Empirical analyses of international anti-corruption programs lend no strong support to the “competitive disadvantage” hypothesis. No statistically significant decline in the performance of US firms abroad is discovered. The association between the enforcement of these international laws and corruption is also found to be significantly negative. The FCPA and other programs seem to have helped to decrease corruption in foreign countries.

References Cuervo-Cazurra, A. (2008). The Effectiveness of Laws Against Bribery Abroad. Journal of International Business Studies, 39(4), 634–651. Davis, K. E. (2009). Does The Globalization of Anti-Corruption Law Help Developing Countries? (New York University Law and Economics Working Papers. Paper 203). Graham, J. L. (1984). The Foreign Corrupt Practices Act: A New Perspective. Journal of International Business Studies, 15(3), 107–121. Graham, B., & Stroup, C. (2016). Does Anti-bribery Enforcement Deter Foreign Investment? Applied Economics Letters, 23(1), 63–67. Koehler, M. (2015). Measuring the Impact of Non-Prosecution and Deferred Prosecution Agreements on Foreign Corrupt Practices Act Enforcement. UC Davis L. Rev., 49, 497–741. La Roche, C. R., & Flanigan, M. A. (2011). International Initiatives to Eliminate Corruption: Has Bribery Declined? International Business and Economics Research Journal, 3(9), 15–20. Lippitt, A. H. (2013). An Empirical Analysis of the Foreign Corrupt Practices Act. Virginia Law Review, 98, 1893–1930.

12  International Cooperation     277

Samanta, S., & Sanyal, R. (2016). The Effect of the OECD Convention in Reducing Bribery in International Business. Global Business and Management Research, 8(1), 68. Transparency International. (2007). Global Corruption Report 2007. Trzcinski, L. M. (2012). The Impact of the Foreign Corrupt Practices Act on Emerging Markets: Company Decision-Making in a Regulated World. New York University Journal of International Law and Politics, 45, 1201–1285.

Part III Anti-corruption Strategies: The Role of Civil Society

So far, we have examined how different institutional or market changes can help to reduce corruption. The analysis has brought us to the conclusion that whatever system is suggested to curb corruption, monitoring and control are necessary. For instance, while the move from an autocratic regime to a democratic one removes one of the prime causes of corruption, the extent of its effect depends on the democratic rules. The establishment, monitoring, and application of these rules can substantially reduce, although not eliminate, grand corruption but might be ineffective against petty corruption. The latter might persist even in democratic countries and is often the most exasperating to ordinary citizens. The most illustrative example is India, the biggest democracy in the world and at the same time one of the most corrupt. Because even an honest ruler cannot control the actions of each civil servant, additional monitoring mechanisms appear to be necessary. For monitoring and control to be successful in curbing corruption, a certain number of conditions need to be fulfilled. First, the monitor (individual, firm, or other organization) should have high standards of ethical and civic values to ensure that it is trusted by citizens. Otherwise, a “monitor to monitor the monitor” is needed. These

280     Part III: Anti-corruption Strategies: The Role of Civil Society

qualities are, however, not easily observable and an alternative is to choose a monitor whose own interest is served by achieving the objective of the mission. Citizens as a group meet this requirement best. Second, effective monitoring entails the ability to access, process, and efficiently use information, especially since a number of fields and dimensions related to corruption are very complex. Such access and efficient use depend heavily on the quality of the information system in the country. Independent media play a crucial role in supplying reliable information and promoting transparency and publicity. They can also be very useful in exposing misbehavior. Third, monitors must have a certain level of education not only to exploit the information but also to strengthen their civic and ethical stance. Education participates in making citizens aware of the harm caused by non-civic attitudes such as corruption. It can make people intolerant of such misbehavior even if they are not directly affected. These three conditions are interlinked and meeting one helps to support the others to provide a potentially powerful anti-corruption tool. For instance, empirical evidence shows that a strong civil society has a substantial anti-corruption impact only in countries with high press freedom. It has no significant impact on corruption in countries with low press freedom (Themudo 2013). Other evidence suggests that a strong civil society can be a powerful complement to the media as a corruption deterrent. A strong civil society increases the pressure for information disclosure and accountability and provides important support to the media in their task of exposing corruption (Shim and Eom 2009). Accordingly, countries concerned with the question of corruption should seek to improve the awareness of their citizens about corruption issues. To this end, education is a very effective tool. It is commonly associated with the acquisition of reading and writing skills, which are prerequisites for citizens’ efficient use of the information provided by the media. In addition, education equips citizens with tools to participate effectively in a representative democracy. In other words, a strong civil society owes a lot to education. However, having well-educated people is not enough if there are obstacles to communication and exchange through media or ICT. To reap the civic and political benefits of education, citizens must be allowed to access government

Part III: Anti-corruption Strategies: The Role of Civil Society      281

information and engage in social interaction, through which they can communicate their opinions, suggest possible solutions, expose misbehavior, and promote ethical behavior. In sum, citizens should be involved in the fight against corruption because they are the main beneficiaries of such action in terms of both money and social cohesion. To effectively accomplish their monitoring tasks, citizens need access to information and the ability to hold those who misbehave accountable. If the activities of the government are publicized, citizens can hold high officials as well as civil servants accountable for misconduct (Rose-Ackerman and Truex 2012).

References Rose-Ackerman, S., and Truex, R. (2012). Corruption and Policy Reform (Yale Law and Economics Research Paper No. 444). Shim, D. C., and Eom, T. H. (2009). Anticorruption Effects of Information Communication and Technology (ICT) and Social Capital. International Review of Administrative Sciences, 75(1), 99–116. Themudo, N. S. (2013). Reassessing the Impact of Civil Society: Nonprofit Sector, Press Freedom and Corruption. Governance, 26(1), 63–89.

13 Civil Society and the Media

The mass media can play a crucial role in supplying citizens with reliable information, which promotes transparency and publicity. They can also be very useful in exposing misbehavior. Various research has extensively documented the efficacy of information and the media as weapons against corruption. Studies inspired by the experience of radio have shown that the emergence of this media strengthened the influence of information by increasing the speed of transmission and reaching a larger population. Similar phenomena have been documented for television and the Internet. The latter provides, within the more general framework of Information and Communication Technology (ICT), a useful tool for faster and better communication, retrieval of data, and utilization of information. The expected positive impacts listed above are, however, not guaranteed, because the media themselves may be corrupt, while private or government groups may exert pressure and make threats. The historical record of corruption in the media suggests, however, that these problems do not persist for a long time. Empirical analyses confirm that the media can help in reducing corruption, but not alone. In countries with low levels of education, the media need to be accompanied by policies © The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_13

283

284     K. Sekkat

aiming at strengthening the capacity of citizens to access, process, and use information.

1 Expected Impacts of the Media on Corruption Effective monitoring by citizens requires access to information and the ability to hold those who misbehave accountable. In other words, increased transparency and bottom-up accountability are both needed. If the activities of the government are publicized, citizens can hold high officials as well as civil servants accountable if they notice wrongdoing. Citizens can initiate complaints, engage in social exclusion, cast votes, or conduct denunciation campaigns. However, the information providing the basis for citizens’ actions must be reliable. Rumors are not enough to initiate and coordinate actions. The mass media are good candidates for supplying such reliable information and hence for promoting transparency and publicity (Court et al. 2003). The media must be strong and independent from government and wealthy private interests to help fight corruption (Brunetti and Weder 2003). For instance, Reinikka and Svensson (2004), focusing on leakage of an education grant in Uganda, show that the detailed publication by the government of education funding processes in local newspapers allowed citizens and schoolmasters to better monitor the release and use of government funds. In particular, communities with better access to newspapers experienced lower leakage rates. Thus, launching newspaper campaigns can substantially reduce corruption, leakage, and associated embezzlement. An experimental study in Brazil (Winters and Weitz-Shapiro 2014) showed that respondents who were informed of a corrupt act expressed a desire to punish the guilty politician regardless of his/her level of performance. de Figueirido et al. (2013) reveal that knowledge of corruption can reduce the vote share by 2.6%. Again in Brazil, Ferraz and Finan (2011) look at a federal government initiative to randomly audit municipalities on the use of received funds. The results of the audits were publicized in local newspapers and radio programs.

13  Civil Society and the Media     285

The study showed that the release of audit results negatively affected the electoral performance of corrupt incumbent mayors. Finally, Peisakhin and Pinto (2010) found that India’s Right to Information Act (RTIA) allowed citizens to access a public service more effectively without having to resort to petty corruption. The role of the media in fighting corruption has to be nuanced, however (Ferraz and Finan 2011). Although greater information provision is good, the capacity of the public to digest information and act accordingly should not be overestimated. Most citizens may be incapable of understanding the subtleties of the issue under consideration or of devoting enough time and effort to make the best of the available information. Moreover, transparency is only effective if citizens have the ability to report infringements which are followed by punishment (Adsera et al. 2003). Besides these problems, the provision of information itself is not without risk. Private and government groups which are reluctant to see their corrupt activities publicized might exert pressure on information providers and threaten them with punishment. This might discourage the provision of information unless informers are credibly protected in one way or another. Moreover, such threats are not always illegal and may be based on well-established judicial rules. While the jailing of journalists is now rare in many developed societies, it tends to persist in a number of developing ones. Another method which is more frequent even in developed societies is to use defamation laws, which can result in massive fines or damages, although means of correcting or getting compensation for the potential harm exist. Such fines can result in bankruptcy of the journalist or publishing company for a single “error”. An alternative means of intimidation concerns publication licenses or restrictions, which can undermine the willingness to investigate. The external obstacles discussed above, all of which can impact the effectiveness of the media in combating corruption, are not the only factors at play. There are also internal factors which can have a similar influence. Of prime importance is the risk that journalists might be corrupt themselves. Chapter 1 gives many examples of such corruption. However, the historical record of corruption in some developed countries suggests that the problem might not persist for a long time. For

286     K. Sekkat

instance, Bignon and Flandreau (2011) study the phenomenon of “badmouthing”—or the threat of printing negative information for extortion purposes—in the pre-1914 French capital market. At that time, many journals were using badmouthing techniques to get payments from issuers of bonds and other assets. Faced with this “credible” threat, many issuers were tempted to pay to prevent damaging rumors from circulating. Badmouthing is easy to do because journals do not need to make explicit allegations. They need only indicate, suggest, or let it be understood that something is wrong and let the rumor spread. To control the damage from badmouthing, corporations, banks, and governments can organize themselves to form a counterweight to make better information available to the public. Moreover, since the noise generated by badmouthing is potentially harmful not only to investors but also to borrowers, both groups seek press coverage from more prestigious and reliable sources of information. The “good” press enjoys a windfall while the other perishes. Gentzkow et al. (2006) draw on the remarkable evolution in the media and corruption in USA between 1870 and 1920 to document a similar pattern. Today, the USA is commonly considered to be among the least corrupt nations in the world. In the past, however, America encountered many political scandals and widespread corruption comparable to the situation in the most corrupt nations nowadays (Glaeser and Goldin 2004). In 1870, the US press was partisan and insincere. By 1920, most newspapers had become less partisan and more prone to reporting the facts without bias. The reason for this shift seems to lie with the increasing financial returns earned from selling newspapers rather than from efforts to satisfy politicians in search of patronage and other privileges.

2 Actual Impacts of the Media on Corruption The previous section argued that, to be successful, efforts by citizens to curb corruption need effective and powerful media. Themudo (2013) examines the extent to which the impact of civil society depends on the press being free and generating sufficient public pressure on the administration. The study uses two samples of data on civil society

13  Civil Society and the Media     287

and corruption: a cross-national sample of 40 countries with data on non-profit sector size and a longitudinal sample of 118 countries with data on civil liberties. The dependent variable, corruption, is measured using TI’s Corruption Perceptions Index (CPI). It is explained in terms of the strength of civil society, freedom of the press, and an interaction term between the two. The interaction term is used to test whether the impact of the strength of civil society on corruption depends on the degree of press freedom. The strength of civil society was measured using three different indicators: the share of non-profit sector employment (paid and full-time equivalent volunteer labor) in total employment, Freedom House’s Civil Liberties Indicator and the same indicator but recalculated to exclude the freedom of expression score. Press freedom is measured using Freedom House’s Freedom of the Press Index. Control variables include income per capita, the degree of state intervention in the economy, openness to trade, political institutions (the level of democracy, federalism, independence of the judiciary), and historical and cultural variables. The analysis showed that, irrespective of its measure, a strong civil society is associated with lower levels of corruption across countries and over time. However, such effect is strong only in countries with high press freedom. There is no significant impact on corruption in countries with low press freedom. These findings suggest that the impact of civil society on corruption is significantly dependent on its ability to generate public pressure through a free press. Two channels through which press freedom and civil society can effectively counter corruption are transparency (disclosing information about government and other institutions) and publicity (making the information known to citizens). Lindstedt and Naurin (2006) sought to test the validity of this expectation in reality. Using a sample of 111 countries over the period 2001–2004, they study the relationship between corruption, transparency, and publicity. Corruption measures are drawn from three institutions, namely the World Bank (WB), TI, and the International Country Risk Guide (ICRG). These measures are explained in terms of political transparency (from the WB), freedom of the press (from Freedom House and Reporters Without Borders),

288     K. Sekkat

newspaper circulation, and the number of radio receivers per capita (both from the WB). Other control variables such as democracy, education, economic development (measured by GDP per capita), and British colonial heritage are also used. The results confirm that transparency can reduce corruption but not by itself. Making information available does not prevent corruption if the conditions for publicity are weak. Furthermore, such publicity must be independent from the government. Freedom of information and other transparency laws implemented by the government are less effective against corruption than a free press. The results also point to the importance of education in fighting corruption. In countries with low levels of education and media reach, transparency needs to be accompanied by policies aiming at strengthening the capacity of citizens to access and process information. Brunettia and Weder (2003) bring additional insights to the importance of the media. Using average corruption scores between 1994 and 1998 in 128 countries, they examine the importance of a free press as a means of controlling corruption. The measures of corruption come from the ICRG, the WB, and TI. The measure of press freedom is drawn from Freedom House. Control variables are the quality of the bureaucracy, respect for the law by citizens (taken from the ICRG), the level of per capita GDP, educational attainment, and the degree of ethnolinguistic diversity (all from the WB). The results show a strong association between the level of press freedom and the level of corruption across countries, suggesting that an independent press may represent an important weapon against corruption. This result is robust to different checks. To give a flavor of the economic significance of the impacts, the authors compute the improvement in corruption that countries can expect from more press freedom. They show that an improvement of one standard deviation in press freedom could reduce the corruption score by between 0.4 and 0.9 points (on a scale of 0–6). Similarly, they consider how much an improvement in press freedom to the level of Norway (the country with the freest press) would affect the corruption index for countries with particularly repressive practices. They find that such improvement in Indonesia would bring corruption down to the level of Singapore.

13  Civil Society and the Media     289

For the Russian Federation, it would bring the corruption level down to that of the Slovak Republic, while corruption in Nigeria would fall to the level of Belgium. While the above studies focus on institutional aspects affecting the role of the media, i.e., press freedom, Adsera (2003) examines a more practical dimension. The paper uses media circulation for a cross-section of between 155 and 173 countries over the period 1997–1998. A similar analysis is conducted for US states. Starting with the cross-country analysis, the dependent variable is the WB corruption index. It is explained in terms of the circulation of daily newspapers per person in 1995, which also comes from the WB. The level of democracy in 1994 (taken from the Polity III database), the percentages of the population of the country that belong to the three most widespread religions (Catholicism, Islam, and Protestantism), ethnic fractionalization, and per capita real income (both from the WB) are introduced as control variables. An interactive term between the level of democracy and the circulation of newspapers is also incorporated into the regression to gauge the extent to which the impact of newspaper readership on political accountability depends on political freedom. Massive levels of readership without political liberties, like in the former Soviet Union, clearly did not make the politburo accountable to the public. The results show that political control of public officials depends on two key factors. The first is the existence of free and regular elections, which allow citizens to discipline politicians. The second, and equally important, is the degree of freedom of information, which enables citizens to curb the opportunities for politicians to engage in corruption and mismanagement. The presence of a well-informed electorate in a democratic setting explains between one-half and two-thirds of the variance in levels of governmental performance and corruption. This result is robust to numerous robustness checks. The authors conduct a similar analysis for the USA. The measure of political corruption here is different, however. For each state, it is the number of public officials who have been convicted for violating laws against public corruption. To eliminate random variations in yearly data, the total number of convictions for two separate periods,

290     K. Sekkat

1977–1987 and 1986–1995, is used. The data come from the US Department of Justice. Here, the presence of a well-informed electorate accounts for the impressive cleanliness of American states in the Plains as well as for the much higher level of federal prosecutions of public officials in the South of the USA. Costa (2012) focuses on the impact of Freedom of Information (FOI) laws on corruption. The first FOI-type law was enacted by Sweden in 1766. Finland was the next to adopt, in 1951, but a wave of adoptions was launched following the USA’s enactment of a similar law in 1966. As of September 2013, at least 95 countries had nationwide laws establishing the right of, and procedures to facilitate, access to records held by government bodies. While governments may have had different reasons for adopting such laws, anti-corruption concerns rank among the most important (Banisar 2006). This paper explores whether corruption perceptions are reduced following adoption of an FOI law. It uses a panel of up to 128 countries from 1984 to 2007. Among these countries, 40 enacted FOI legislation during the period, and 12 already had an FOI law during the period for which data were available. The dependent variables are the ICRG and TI corruption perception indexes. The main explanatory variables are a dummy which equals 1 starting on the year an FOI law entered into force in a country and the degree of press freedom from Freedom House. Control variables are the Polity 2 Democracy Index and GDP per capita. The results indicate that perceptions of corruption rise following the adoption of FOI laws. More importantly, this increase in perceived corruption takes place in the first five years after adoption. In addition, perceptions of corruption rise in countries with a free press. The cross-country studies presented above are usefully complemented by those focusing on one country. Cross-country studies reveal regularities, or averages, in the relationship under examination, while single-country studies uncover specific country characteristics that affect the “average” relationship. Focusing on India, Transparency International (2008) highlights the importance of access to information. In 2005, the government of India issued the Right to Information Act (RTI Act), which established the right of citizens to access

13  Civil Society and the Media     291

information and the practical procedures for doing so. It also instituted a Central Information Commission and State Information Commissions. The objective was to promote transparency and accountability in the working of public authorities. The RTI Act has been used several times by citizen groups to request information from the authorities. For instance, in Keolari, a small village in the central state of Madhya Pradesh, a local elected politician was building a home in December 2006 when he erected a wall around a well that his father had donated to the community nine years earlier. This well was one of only two sources of potable water available to the village’s 2500 residents. After the politician refused requests to access the well, residents filed complaints and got local newspapers to write about the problem. With the help of some NGOs, the citizens used the RTI Act to ask the authorities to provide copies of the gift deed signed by the father as well as information on any public money spent to maintain the well. Based on this information and despite the inevitable bureaucratic complications, the politician’s wall was declared illegal, and he was ordered to demolish it within a week. Since then, citizens have been able to draw water from the well. Instead of access to information, Stromberg (2004) focuses on the circulation of information. Specifically, the paper uses a natural experiment to examine how access to radios in the USA since the early twentieth century has affected corruption. The focus on the introduction of new media technologies is very useful. Such introduction causes dramatic changes in people’s access to mass media as well as large geographic variations in the share of the population with access to the new media. Radio was introduced as a mass medium in the early 1920s and expanded rapidly to reach a household penetration rate of around 80 percent by 1940. This brought about major changes in voter access to mass media because radio was the first broadcast media with characteristics very different from print media. During the period of radio’s introduction, the USA also witnessed rapid changes in economic policy-making. In the middle of radio’s expansion period, the New Deal was launched. The paper investigates the extent to which radio influenced the distribution of funds in a New Deal program which provided unemployment

292     K. Sekkat

relief and was implemented over the period 1933–1935. This program in question was the largest of the early New Deal programs and was administered by the Federal Emergency Relief Administration (FERA). The allocation of the relief funds was, however, the responsibility of the governor. The empirical analysis investigates whether governors spent more money in areas where a large share of the population had a radio. The analysis was conducted for up to 2500 US counties over the period 1920–1940. The dependent variable is the per capita amount of the state relief budget allocated by the governor to a given county. The main explanatory variables are the share of households with radios in the county, the share of illiterate people in the county, and school enrolment in the county. Controls include voter density, population density, unemployment rate, farm value, and ethnic composition. The findings confirm that governors allocated more relief funds to areas where a larger share of the population had radios. The effects are both statistically and economically significant. A one-percentage-point increase in the share of households with radios increases per capita relief spending by 0.6%. Similarly, a one standard deviation increase raises spending by nine percent. Finally, a move from the lowest to the mean share of households with radios increases spending by 60%. An important by-product of the study concerns the effect of illiteracy. Fewer relief funds are allocated to areas with a large share of illiterate people. The reason is that illiterate people are less likely to be informed about who is responsible for cuts in the amount allocated to their county. The effect of illiteracy is economically significant. A one-percentage-point increase in the illiteracy rate makes governors cut spending by more than 1%. A similar analysis in spirit is conducted by Besley and Burgess (2002) to examine the impact of mass media in inciting local government to meet citizens’ needs across Indian states. The newspaper industry in India publishes in English, Hindi, and in many other languages, such as Assamese, Bengali, Gujarati, and Urdu. In contrast to the majority of other developing countries, the Indian press is relatively free and independent, which provides an opportunity to learn about government responses to local conditions. The authors focus on how the media

13  Civil Society and the Media     293

affect the public distribution of food and state government expenditures on calamity relief across the 16 major Indian states over the period 1958–1992. The Famine Relief Codes put in place after 1880 govern public distribution of food and calamity relief in India. Elected state governments are responsible for looking at signs of crises, such as large drops in food production and increases in food prices, and for counteracting the possible damage by increasing the public distribution of food and setting up public works programs and relief centers. Per capita public food distribution and calamity relief expenditure are explained in terms of the circulation of all newspapers. The breakdown of newspapers by language of circulation is also considered. Control variables include drought and flood, food grain production, political competition and population density. The analysis shows that public food distribution and calamity relief expenditure increase with electoral accountability and newspaper circulation. A similar question to that of Besley and Burgess (2002) is investigated by Ferraz and Finan (2008) for Brazil. In 2003, the federal government started randomly selecting municipal governments to be audited regarding their use of federal funds. After completion of the audit, the results are communicated to the municipalities, federal prosecutors, media, and posted on the Internet. The authors study the effects of such disclosure of local governmental practices on the electoral outcomes of incumbents in 373 Brazilian municipalities. Specifically, the analysis compares electoral outcomes of the 2004 municipal elections for mayors eligible for re-election. The authors seek to assess the effects of publicizing the audits in terms of the type of information disclosed and the presence of the local media. To this end, they construct an indicator based on the number of times irregularities appear in the audits and define this sum as a measure of corruption. The irregularities concern fraud in procurements, diversion of public funds, or over-invoicing. The indicator of corruption is explained in terms of newspaper circulation and other control variables such as mayoral characteristics, socioeconomic characteristics of the municipality, the proportion of mayors who ran for re-election in 2004, the number of parties in 2000, per capita income, and income inequality.

294     K. Sekkat

The results show that the publicity of information concerning corruption had a significant effect on incumbents’ electoral performance. It reduced the incumbent’s likelihood of re-election by seven percentage points (or 17%) in municipalities where at least two violations were reported. Moreover, corrupt politicians were punished relatively more in places where local radio stations were present and divulged the findings of the audit reports. Finally, while divulgation by local radio negatively affected outcomes for corrupt mayors, it promoted non-corrupt incumbents by markedly increasing their likelihood of re-election. Transparency International (2006) focuses on the impact of the Mexican Transparency and Access to Public Information Law introduced in June 2002. This law proved helpful to civil organizations fighting corruption. For example, in 2003 Mexico’s Congress voted for programs that were intended to promote women’s health and in particular HIV/AIDS programs. Out of a total budget of US$56.5 million, senior officials from Congress and the Health Ministry decided that US$2.8 million should be allocated to a private organization, Provida, and targeted toward HIV/AIDS public campaigns. Six Mexican Civil Society Organizations (CSOs) felt concerned about the reasons for the budget’s alteration. They arranged for the budget to be investigated to see how the US$2.8 million was spent and found evidence of misuse and corruption, including collusion between Provida and some of its suppliers, purchases of luxury pens, clothing and groceries, together with numerous other fiscal inconsistencies. The CSOs revealed their findings at a press conference in June 2004 at which they launched a campaign demanding transparency and accountability. In April 2005, after many twists and turns, the officials involved were dismissed, and Provida’s legal representative was fined and forbidden from occupying public office. Although the US$2.8 million was not returned and no fine was paid, the cause for greater and effective transparency was strongly promoted thanks to the CSOs’ actions. Reinikka and Svensson (2005) investigate the impact of a newspaper campaign which provided schools and parents with information about local officials’ handling of a large education grant program in Uganda. The source of this newspaper campaign was a Public Expenditure Tracking Survey (PETS) conducted for the government to gauge to

13  Civil Society and the Media     295

what extent public resources reached schools. The surveys showed that in the mid-1990s a school received, on average, only around 20% of the resources released by the central government for the program. Most schools received nothing. The lion’s share of the grants was captured by local government officials (and politicians) in charge of disbursing the grant to the schools. Moreover, schools in poorer communities suffered significantly more from capture. As a result, the government initiated a newspaper campaign (in the national newspapers and their local language editions) giving data on monthly transfers of grants to districts. To assess the effects of improved access to public information on capture of funds, Reinikka and Svensson (2005) use two waves of the PETS: the original 1996 survey and another conducted in 2002. The original survey was conducted over 250 schools, randomly drawn from 18 districts. The 2002 survey revisited these original schools plus an additional 170 schools from 9 of the original 18 districts. The measures of the effectiveness of the press campaign (the dependent variables) are school-specific and measure the extent of capture, enrolment, and scores. The extent of capture is measured as the share of the total grants disbursed by the central government that was effectively received by a school. A low value indicates that the school suffers more from local capture. The explanatory variable of interest measures head teachers’ knowledge about the grant program, which comes mainly from their exposure to the newspaper campaign. Control variables include the municipality’s average income and fixed effects. The results show that following the start of the campaign, schools with more informed head teachers experienced significantly larger reductions in local capture. The effect is economically large: A one standard deviation increase in head teacher information results in a 1.1 standard deviation increase in the amount that reaches the school. Regarding enrolment, a one standard deviation increase in the share of spending reaching the school results in a 0.66 standard deviation increase in school enrolment. Finally, in a previously “high capture” district but where head teachers have been exposed to the newspaper campaign, students scored on average 0.42 standard deviations better than those in districts where head teachers were less exposed to the newspaper campaign.

296     K. Sekkat

3 Conclusion In various parts of the preceding chapters, we have pointed to the importance of transparency, accountability, and control in fighting corruption. The media are very helpful in accomplishing these tasks, especially since a number of fields and dimensions related to corruption are very complex. The studies show a strong association between the level of press freedom and the level of corruption, suggesting that an independent press may represent an important weapon against corruption. The speed with which information circulates is also crucial. Historical studies suggest that the introduction of radios, television, and the Internet have strengthened the influence of information by increasing the speed of transmission and reaching a broader population. However, information and publicity must be independent from the government. Finally, the results also show that education and the media strongly complement each other in fighting corruption. In countries with low levels of education, citizens are not able to efficiently access, process, and use information.

References Adsera, A., Boix, C., & Payne, M. (2003). Are You Being Served? Political Accountability and Quality of Government. Journal of Law Economics and Organization, 19(2), 445–490. Banisar, D. (2006). Freedom of Information Around the World 2006: A Global Survey of Access to Government Information Laws. London: Privacy International. Besley, T., & Burgess, R. (2002). The Political Economy of Government Responsiveness: Theory and Evidence from India. Quarterly Journal of Economics, 117(4), 1415–1451. Bignon, V., & Flandreau, M. (2011). The Economics of Badmouthing: Libel Law and the Underworld of the Financial Press in France Before World War I. Journal of Economic History, 71(3), 616–653. Brunetti, A., & Weder, B. (2003). A Free Press Is Bad News for Corruption. Journal of Public Economics, 87(7), 1801–1824. Costa, S. (2012). Do Freedom of Information Laws Decrease Corruption? Journal of Law Economics and Organization, 29(6), 1317–1343.

13  Civil Society and the Media     297

Court, J., Hyden, G., & Mease, K. (2003). The Judiciary and Governance in 16 Developing Countries (World Governance Survey Discussion Paper No. 9). Tokyo: United Nations University. de Figueiredo, M. F. P., Hidalgo, F. D., & Kasahara, Y. (2013). When Do Voters Punish Corrupt Politicians? Experimental Evidence from Brazil. Department of Political Science. Berkeley: University of California. Ferraz, C., & Finan, F. (2008). Exposing Corrupt Politicians: The Effects of Brazil’s Publicly Released Audits on Electoral Outcomes. Quarterly Journal of Economics, 123(2), 703–745. Ferraz, C., & Finan, F. (2011). Electoral Accountability and Corruption: Evidence from the Audits of Local Governments. American Economic Review, 101(4), 1274–1311. Gentzkow, M., Glaeser, E. L., & Goldin, C. (2006). The Rise of the Fourth Estate. How Newspapers Became Informative and Why it Mattered. In E. L. Glaeser & C. Goldin (Eds.), Corruption and Reform: Lessons from America’s Economic History (pp. 187–230). Chicago: University of Chicago Press. Glaeser, E. L., & Goldin, C. (2004). Corruption and Reform: An Introduction. WP 10775. National Bureau of Economic Research. Lindstedt, C., & Naurin, D. (2006). Transparency Against Corruption-A Cross-Country Analysis (pp. 9–13). In IPSA 20th World Congress. Fukuoka, Japan. https://Pdfs.Semanticscholar.Org/2c20/ F9805f55256e78640f8733314f3aef8c64fe.Pdf. Accessed 2 April 2018. Peisakhin, L., & Pinto, P. (2010). Is Transparency an Effective Anti-Corruption Strategy? Evidence from a Field Experiment in India. Regulation & Governance, 4(3), 261–280. Reinikka, R., & Svensson, J. (2004). Local Capture: Evidence from a Central Government Transfer Program in Uganda. Quarterly Journal of Economics, 119(2), 679–705. Reinikka, R., & Svensson, J. (2005). Fighting Corruption to Improve Schooling: Evidence from a Newspaper Campaign in Uganda. Journal of the European Economic Association, 3(2–3), 259–267. Strömberg, D. (2004). Radio’s Impact on Public Spending. Quarterly Journal of Economics, 119(1), 189–221. Themudo, N. S. (2013). Reassessing the Impact of Civil Society: Nonprofit Sector, Press Freedom and Corruption. Governance, 26(1), 63–89. Transparency International. (2006). Global Corruption Report 2006. Transparency International. (2008). Global Corruption Report 2008. Winters, M. S., & Weitz-Shapiro, R. (2014). Political Corruption and Partisan Engagement: Evidence from Brazil. Journal of Politics in Latin America, 7(1), 45–81.

14 Civil Society and the Specific Role of ICT

The efficacy of information and the media as weapons against corruption having been documented, this chapter discusses the complementary role of Information and Communication Technology (ICT). Including ICT can be more effective in combating corruption than relying on traditional media alone. ICT enables citizens to access many government services online, which weakens the role of bureaucrats as intermediaries between the government and the public. Moreover, citizens can use the “social web” to communicate their opinions, suggest possible solutions, expose misbehavior, and promote ethical behavior over long distances very quickly. Very importantly, ICT greatly facilitates investigation and asset tracking. Empirical findings reveal that ICT does indeed have a significant role in reducing corruption. Interestingly, e-government and Internet penetration do a better job of explaining variations in corruption among countries than bureaucratic quality and law enforcement.

© The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_14

299

300     K. Sekkat

1 Expected Impacts of ICT on Corruption Spreading information about officials’ misbehavior increases the risk of detection and makes the corrupt behavior less attractive. Over time, these features have spurred governments around the world to rely on ICT not only to provide better service to citizens but also as a means of fighting corruption. This evolution led to what is now commonly referred to as “e-government”, which emerged with the development of government Web sites in the late 1990s. E-government refers to a government’s use of ICT to enhance access to and delivery of information and services to citizens, business partners, employees, other agencies, and other government entities. E-government is also expected to help build better relationships between the government and the public by making interactions with citizens smoother, easier, and more efficient (Layne and Lee 2001). There are several reasons why including ICT can be more effective in fighting corruption than relying only on traditional media (SPIDER 2010). First, by allowing citizens access to government services online, ICT removes the role of bureaucrats as intermediaries between the government and the public, thus limiting interactions between potentially corrupt officials and the public. Second, online systems require standardized rules and procedures. This reduces bureaucratic discretion and increases transparency as compared with the arbitrariness of civil servants dealing with the public on a case-by-case basis. Third, the information held by government agencies, or individual civil servants, on electronic platforms is, in general, automatically provided in forms defined by laws and protocols linking different databases. Automation and protocols highly limit the ability of an individual to exert influence by manipulating or censoring information. Fourth, ICT has markedly fostered social interaction over long distances through a “social web” where users can communicate their opinions, suggest possible solutions, expose misbehavior, and promote ethical behavior. ICT can also be very useful in investigation and asset tracking. Fifth, besides these practical advantages, ICT has a first-order impact on economic growth. Since economic growth is frequently reported to lower corruption, this

14  Civil Society and the Specific Role of ICT     301

technology is likely to help in reducing corruption. Sixth, since rapid technological change encourages, in general, investment in human capital and since human capital accumulation is often hostile to corruption, ICT can help to reduce corruption through this channel (Andersen et al. 2011).

2 Actual Impacts of ICT on Corruption One of the most successful experiences with ICT as a weapon against corruption is the Online Procedures Enhancement (OPEN) system of the Seoul Metropolitan Government (SMG). The system was suggested and supported by Mayor Kun Koh as a means to fight corruption within the SMG (Kim et al. 2009). This followed a series of corruption scandals that had affected the SMG. Developed by the SMG and launched on April 15, 1999, OPEN aimed at achieving transparency in the civil administration and preventing unnecessary delays and unfair handling of civil affairs on the part of civil servants. Before the introduction of OPEN, citizens had to spend a long time waiting for city offices to deal with their affairs. The introduction of the system and the display of information in standardized forms allowed citizens to save a lot of time by being able to browse details online at home. OPEN also provides application status notifications through short message services (SMS) and email and opens every detailed procedure of all services, helping to increase trust between civil officers and citizens. Eight years after the introduction of OPEN, more than 6.7 million citizens had visited the site and more than 2.9 million document registrations had been recorded. The percentage of delayed data input had decreased from 15 to 2% and the services had been extended to other tasks. Meanwhile, the SMG’s anti-corruption index climbed from 64 in 1999 to 84.9 in 2006. Finally, an integrity assessment performed by the Korea Independent Commission Against Corruption (KICAC) confirmed this success. Citizens, public administration specialists, and government employees voted the OPEN system Seoul’s Most Valuable Policy in 1999 and 2000. As a result, the OPEN system

302     K. Sekkat

gained recognition from international organizations such as the UN, the OECD, and the World Bank in addition to reputable media outlets like Business Week and Time. Several countries, including Vietnam, Indonesia, Bulgaria, Japan, China, Egypt, and Nepal drew on the OPEN experience when developing their own anti-corruption strategies. The system was also recognized by the central government of Korea as a success and included in the “Saeol”, an e-government system which has been implemented and is being used nationwide. Mistry and Jalal (2012) examine and compare the impact of e-government on corruption in developed and developing countries. Specifically, the paper investigates two questions. The first concerns the existence of an impact of e-government on corruption. The second focuses on whether this potential impact differs across countries. In order to address the two questions, an equation explains the change in corruption (between 2003 and 2010) in terms of the change in the E-government Development Index (EDI) (between 2003 and 2010), the country’s level of development, and an interaction term between e-government and the country’s level of development. The interaction makes it possible to test whether the impact of e-government on corruption is higher or lower in developed countries. The measure of corruption is the Corruption Perceptions Index, while e-government is based on the UN Readiness Index, later renamed the EDI. Control variables include per capita real GDP, the black market premium, bureaucracy, and civil rights. The results support the hypothesis that e-government decreases corruption. A 1% increase in the e-government index results in a 1.17% decrease in corruption. Moreover, developing countries benefit most from increased use of e-government Andersen (2009) uses IV estimation methods to examine whether the association between e-government and corruption can be given a causal interpretation. The paper focuses on 149 non-OECD countries observed in 1996 and in 2006. The measure of corruption is the WB Control of Corruption Index. The e-government variable is drawn from West (2006). Control variables include country fixed effects, real GDP per capita, and press freedom. The results show that e-government has an important impact on corruption: An increase from the

14  Civil Society and the Specific Role of ICT     303

10th percentile to the 90th percentile in the e-government distribution is associated with a movement from the 10th percentile to the 23rd percentile in the “control of corruption” distribution. Since the Internet is the main vehicle used to operate e-government, Andersen et al. (2011) investigate the causal impact of Internet on corruption using IV estimation methods. The study considers the impact across 48 contiguous US states and across 105 countries. For the USA, corruption convictions are used as dependent variables. The change in this variable between 1991 and 2006 is explained in terms of the change in Internet use over the same period. Internet use is measured by the percentage of households with Internet access and comes from the Current Population Survey (CPS). Control variables include initial state population, the growth of state population, and the initial level of corruption. In the cross-country analysis, the change in the ICRG index of corruption between 1991 and 2006 is the dependent variable. It is explained in terms of the change between 1991 and 2006 in the number of Internet users per 100 people, which is drawn from the International Telecommunication Union. The results are statistically significant, suggesting that the Internet has helped to suppress corruption since its emergence. Although the effect is stronger for the USA, the cross-country analysis suggests that the US state-level results generalize to an international setting. Elbahnasawy (2014) uses a different test of causality from that of Andersen et al. (2011) in order to examine the impact of e-government and Internet adoption on corruption. The paper implements Granger causality tests on a sample of 160 countries over the period 1995–2009. The measure of e-government comes from the United Nations. The measure of corruption is the TI Corruption Perceptions Index (CPI). Internet adoption is measured by the number of Internet users per 100 people, which is obtained from the WB. The set of control variables includes GDP per capita, openness to international trade, law enforcement, and the Property Rights (PR) Index from the Heritage Foundation. The results of the Granger causality tests suggest that a higher e-government index reduces corruption, but not the other way round. The results of panel Granger causality tests indicate that causality between

304     K. Sekkat

Internet adoption and corruption is bidirectional. Interestingly, the driving force behind the favorable impact of e-government on curbing corruption is the telecommunication infrastructure. This suggests that e-government and Internet adoption should be considered as complements, rather than substitutes, in anti-corruption programs. In addition, e-government reinforces the influence of law enforcement on corruption reduction. Lio et al. (2011) also conduct Granger causality tests between Internet adoption and corruption reduction but add to the analysis by quantifying the effect. Using a panel of 70 countries over the period 1998–2005, the paper defines Internet adoption as the number of Internet users per 100 inhabitants and the level of corruption by the CPI. The results of the Granger causality tests reveal that the causality between Internet adoption and corruption is bidirectional, which is in line with the preceding study. One lesson is that the estimated effects of Internet adoption on corruption reduction may be overestimated if the bidirectional causality is not taken into account. To quantify the effect of Internet adoption on corruption, the authors employ the dynamic GMM model. This model makes it possible to take into account the reverse causality running from corruption to Internet adoption by using an instrumental variables approach and exploiting the dynamic properties of the sample. The model explains the CPI in terms of Internet adoption, GDP per capita, and the level of education in the country. The estimation results show a statistically significant effect of Internet adoption on corruption reduction. The economic effect is, however, not substantial, which confirms the importance of taking into account the bidirectional causality between Internet and corruption. On average, an increase of ten in the number of Internet users per 100 inhabitants improves the CPI by about 0.05 points on a ten-point scale, which is not a substantial amount. Assuming that greater corruption awareness acts as a corruption deterrent, Goel et al. (2012) investigate the effect of Internet awareness about corruption on the prevalence and perceptions of corruption. The originality of this paper lies in the measure of awareness about corruption, which is based on a search for keywords “corruption”, “bribery”, and “country name” through the two leading Internet

14  Civil Society and the Specific Role of ICT     305

search engines (Google and Yahoo). The approach adopted in the paper clearly improves on the other measure of Internet diffusion, because an Internet user can surf the Web without paying any particular attention to corruption. The searches were conducted for 150 countries over two separate time periods: December 2009 and January 2011. The averages of the Google and Yahoo results are normalized using the country’s population to produce a country score of corruption awareness. Three well-known indexes of corruption (CPI, the World Bank index, and the World Business Environment Survey index) are used as dependent variables. Each of these indexes is explained in terms of the country’s search score, real GDP per capita, government size, degree of economic freedom, democratic accountability, bureaucratic quality, and law and order. The results show that corruption awareness is associated with lower corruption. A one standard deviation increase in a country’s Internet score decreases corruption as measured by the CPI is 0.18. The corresponding impact on corruption as measured by the WB is 0.07. For the World Business Environment Survey index, the same increase in corruption awareness is far more significant, resulting in a change of over three points in this corruption index, which ranges from one to six. Shim and Eom (2009) focus on the role of social capital in complementing the role of ICT as a corruption deterrent. Broadly defined, social capital is comprised of social organizations and networks, norms and social trust which facilitate citizens’ coordination and cooperation for mutual benefit. Combining ICT and social capital can be a major factor in corruption reduction. On the one hand, ICT not only reduces unnecessary human intervention in government work but also speeds up information circulation and facilitates interactions among citizens. On the other hand, a high level of social capital makes citizens more actively involved in political decision-making and increases the likelihood that corrupt behavior by public employees will be exposed to a densely connected public. The study uses the CPI as a measure of corruption. This is explained in terms of the E-Government Readiness and E-Participation Indexes (both from the UN) and Internet penetration from the World Bank. The measure of the other explanatory variable of interest (social capital)

306     K. Sekkat

is based on the World Values Survey (WVS). This is an international research project that measures the values of people from around the world. The survey has been conducted in successive waves since 1981 and asks citizens around the world several questions about social values in their countries. The responses to the two following questions from the 1999 and 2004 waves are used: 1. In general, do you think that most people can be trusted? 2. Do you find the following justifiable: (i) claiming government benefits, (ii) avoiding a fare on public transport, (iii) cheating on taxes, or (iv) accepting bribes? The results reveal that both ICT and social capital have significant and independent roles in reducing a country’s corruption, even after controlling for traditional anti-corruption indicators. Interestingly, e-government and Internet penetration do a better job of explaining variations in corruption among countries than bureaucratic quality and law enforcement. The results concerning the role of social capital in complementing the effect of ICT on corruption are inconclusive. While the majority of studies on the impact of ICT on corruption focus on the Internet, this is not the only component of ICT that can play against corruption. Unfortunately, formal studies directly linking non-Internet ICT components and corruption are almost inexistent. However, indirect inferences can be made. Since the preceding paper (Shim and Eom 2009) showed that social capital can be effective in fighting corruption, studies revealing a link between other ICT components and social capital might support the existence of a relationship between such components and corruption. Campbell and Kwak (2010) examine the impact of mobile phones and their embedded applications on civic and political engagement, which are two dimensions of social capital (See Chapter 15). The data come from a national mail survey that was conducted in the USA immediately following the 2006 mid-term congressional elections. A large number of people were contacted and asked questions about their personal characteristics, civic engagement, political participation, along with the intensity and purpose (recreation, sociability,

14  Civil Society and the Specific Role of ICT     307

information, and opinion exchanges) of mobile phone use. Civic engagement concerns involvement in volunteer and community activities (doing volunteer work, working on a community project, contributing money to a social group or cause, going to a community or neighborhood meeting, and working on behalf of a social group or cause). Political participation includes attending political gatherings (meeting, rally, or speech), circulating a petition for a candidate or issue and contacting a public official or a political party. These two variables are explained in terms of intensity and purpose of mobile phone use and control variables (age, gender, education, household income, and political interest). The results reveal a positive relationship between mobile phone use for information exchange and civic and political involvement. However, individuals who feel more comfortable with mobile phones and who use them for exchanging information about public affairs tend to be more civically and politically engaged than those who are less comfortable with the technology. Cheng et al. (2015) conduct a similar study to the one above but focused on China. Using the results of a survey of university students in three Chinese cities (Guangzhou, Shenzhen, and Zhuhai), the authors examine the relationship between mobile phone uses and civic engagement. The survey used a questionnaire pertaining to different reasons for mobile phone use and different types of civic engagement. The authors consider four reasons for using a mobile phone: technological accessibility, information exchange, social interaction, and recreation (e.g., relaxation, fun, fashion, and status). Civic engagement includes a range of activities such as individual voluntarism, organizational involvement, and electoral participation. Both the online and the offline dimensions of civic engagement are considered. Each of the civic engagement variables is explained in terms of the different reasons for mobile phone use and various control variables such as gender, age, and education level. The results show that accessibility, information exchange, and social interaction are important determinants of individuals’ engagement in civic activities. Surprisingly, fashion and status were also found to be significant determinants of civic engagement. According to the authors,

308     K. Sekkat

a possible explanation is that catching up on the latest social trends stimulates individuals’ attention to and interaction on current events. In this sense, fashion/status is not only a type of recreational gratification, but also reflects the need for social connection. Neither entertainment nor affection was a significant predictor of civic engagement.

3 Conclusion Information and Communication Technology (ICT) and e-government are increasingly being adopted by many countries as a weapon against corruption. They give access to many government services online and therefore markedly reduce the importance of the role played by bureaucrats as intermediaries between the government and the public, so decreasing the opportunities for corruption. ICT is also very helpful in investigating and tracking “lost” assets. Moreover, ICT complements the role of the media as a tool against corruption. Information, communication, and interactions can be exchanged over long distances in a short time. Thus, through the “social web”, citizens who live far apart can quickly communicate their opinions, suggest possible solutions, expose misbehavior, and promote ethical behavior. Assessments of various initiatives around the world converge in considering that e-government is a clear success. E-government decreases corruption. Developing countries are found to benefit most from increased use of e-government. Unsurprisingly, the driving force behind the favorable impact of e-government on curbing corruption is the telecommunication infrastructure. E-government and Internet adoption seem to be complements, rather than substitutes, in anti-corruption programs. Interestingly, e-government and Internet penetration do a better job of explaining variations in corruption among countries than bureaucratic quality and law enforcement. Another component of ICT, mobile telephony, has also been found to be very useful for information exchange and civic and political involvement. However, the benefits in terms of civic and political engagement mainly concern individuals who feel comfortable with using mobile phones.

14  Civil Society and the Specific Role of ICT     309

References Andersen, T. B. (2009). E-government as an Anti-Corruption Strategy. Information Economics and Policy, 21(3), 201–210. Andersen, T. B., Bentzen, J., Dalgaard, C. J., & Selaya, P. (2011). Does the Internet Reduce Corruption? Evidence from US States and Across Countries. World Bank Economic Review, 25(3), 387–417. Campbell, S. W., & Kwak, N. (2010). Mobile Communication and Civic Life: Linking Patterns of Use to Civic and Political Engagement. Journal of Communication, 60(3), 536–555. Cheng, Y., Liang, J., & Leung, L. (2015). Social Network Service Use on Mobile Devices: An Examination of Gratifications, Civic Attitudes and Civic Engagement in China. New Media and Society, 17(7), 1096–1116. Elbahnasawy, N. G. (2014). E-Government, Internet Adoption and Corruption: An Empirical Investigation. World Development, 57, 114–126. Goel, R. K., Nelson, M. A., & Naretta, M. A. (2012). The Internet as an Indicator of Corruption Awareness. European Journal of Political Economy, 28(1), 64–75. Kim, S., Kim, H. J., & Lee, H. (2009). An Institutional Analysis of an E-Government System for Anti-Corruption: The Case of OPEN. Government Information Quarterly, 26(1), 42–50. Layne, K., & Lee, J. (2001). Developing Fully Functional E-Government: A Four Stage Model. Government Information Quarterly, 18(2), 122–136. Lio, M. C., Liu, M. C., & Ou, Y. P. (2011). Can the Internet Reduce Corruption? A Cross-Country Study Based on Dynamic Panel Data Models. Government Information Quarterly, 28(1), 47–53. Mistry, J. J., & Jalal, A. (2012). An Empirical Analysis of the Relationship between E-Government and Corruption. International Journal of Digital Accounting Research, 12(18), 145–176. Shim, D. C., & Eom, T. H. (2009). Anticorruption Effects of Information Communication and Technology (ICT) and Social Capital. International Review of Administrative Sciences, 75(1), 99–116. SPIDER (2010). Increasing Transparency and Fighting Corruption through ICT: Empowering People and Communities. SPIDER ICT4D Series No. 3. West, D. M. (2006). Global E-Government 2006, Brown University. http:// www.Insidepolitics.Org/Egovt06int.Pdf. Accessed 02/04/2018.

15 Civil Society and the Role of Education

Although ICT is spreading very quickly around the world, its use in administrative and political processes is not accessible to everybody, especially in developing countries where infrastructure is poor and a large portion of the population is illiterate. Education is crucial in order to address these shortcomings. In addition to enabling the effective use of ICT in the fight against corruption, education can affect citizens’ attitudes toward corruption in important ways. Schools play a prime role in the formation of citizenship, providing the first non-familial context in which individuals’ moral and thinking capacities are developed. Students at school learn about basic norms and responsibilities, find out how democracy functions, and are encouraged to participate in social activities which foster trust and reciprocity. Moreover, by enrolling students from different socioeconomic backgrounds in the same system, schools lead to the creation of mutual respect. Research results support the idea that education has an impact on corruption. There is a strong correlation between the average number of years of schooling and corruption. Moreover, additional schooling, both at the secondary and at the post-secondary levels, has large and statistically significant

© The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3_15

311

312     K. Sekkat

effects on voter participation, the frequency of newspaper readership, and the degree of support for controversial free speech.

1 Expected Impacts of Education on Corruption 1.1 Social Capital and Corruption The paper by Shim and Eom (2009) that we discussed in the preceding chapter shows that social capital acts as a deterrent to corruption. Broadly defined, social capital consists of social organizations and networks, norms and social trust which facilitate citizens’ coordination and cooperation for mutual benefit. Although it has become very popular in economics during recent decades, the term does not embody a new concept for sociologists. The idea that involvement and participation in groups can have positive consequences for both individuals and communities is a notion dating back to Durkheim’s emphasis on the role of “group spirit” as an antidote to anomie and self-destruction and to Marx’s distinction between an atomized class-in-itself and a mobilized and effective class-for-itself (Portes 1998). The main features of social capital are membership of social networks or associations, high levels of interpersonal trust, and norms of mutual aid and reciprocity (Lochner et al. 1999). The level of social capital in a society is, therefore, measured by indicators such as the density of membership in voluntary associations of all kinds, the extent of interpersonal trust between citizens, their perceptions of the value of mutual aid and connections (neighborhood, work, family and friends), tolerance of diversity, civic engagement, and active citizenship (Lochner et al. 1999). Empirical analyses dealing with social capital are based on surveys of households and communities which collect answers to questions useful for the construction of the relevant indicators (Woolcock and Narayan 2000). However, such empirical evidence is still scarce. Knack and Keefer (1997), using indicators of trust and civic norms from the World Values Survey, find that in a sample of 29 market economies,

15  Civil Society and the Role of Education     313

“social capital” matters to measurable economic performance. Trust and civic norms are stronger in nations with higher and more equal incomes, institutions that restrain predatory actions of chief executives, and better-educated and ethnically homogeneous populations. Narayan and Pritchett (1999) construct a measure of “social capital” in rural Tanzania based on the extent and characteristics of individuals’ associational activity and trust in various institutions and in other individuals. They show that “social capital” is both capital, in that it raises incomes, and social, in that household outcomes depend on village social capital not just on household social capital. The magnitude of the effect of social capital on incomes is impressively large: a one standard deviation increase in village social capital increases household expenditures per person (a proxy for income) by at least 20–30%. This impact is as large as an equivalent increase in non-farming assets or a threefold increase in the level of education. Beugelsdijk and Van Schaik (2005), using a cross-section of 54 European regions, examine whether regional differences in economic growth are related to social capital, in the form of the generalized trust and associational activity. They show that growth differentials in European regions are positively related to social capital measured by associational activity. The simple existence of a network does not stimulate regional economic growth but active involvement in networks does. It appears therefore that social capital acts as a resource for individuals and facilitates collective action. A community well endowed with social capital is likely to have effective civic institutions, to prosper and to succeed in maintaining law and order (Woolcock and Narayan 2000). Family, friends, and associates constitute an important asset which can be called upon in a crisis, enjoyed for its own sake, or used for material gain. Since what is true for individuals also holds for groups, communities endowed with diverse social networks and civic associations will be in a strong position to confront poverty and vulnerability and to take advantage of new opportunities. Moreover, social capital reinforces the call for information disclosure policies and cultivates citizens’ demand for accountability.

314     K. Sekkat

1.2 Education and Social Capital It follows from the above discussion that as the level of social capital in a country increases, the occurrence of corruption should become less likely. Accordingly, countries concerned with the issue of corruption should seek to improve the quantity and quality (more trust, more participation, and so on) of their social capital. To this end, education can be a very effective tool. Education is commonly associated with the acquisition of reading and writing skills, which are prerequisites for citizens’ efficient use of the information provided by the media and ICT. It can therefore make a marked contribution to the fight against corruption. However, the impact of education is not limited to the acquisition of such “technical” skills and touches on the building of a country’s social capital. Accordingly, education plays a prime role in the formation of citizenship. Education can both enhance the quality of participation by individual citizens and foster a spirit of participation among large groups of citizens (Milligan et al. 2004). Regarding the quality of participation, education equips citizens with tools for effective participation in a representative democracy. The skills acquired thanks to education reduce the costs and complexity associated with participation. Education also encourages citizens to defend civic values and helps them assess the behavior of the institutions governing the country. Accordingly, education increases citizens’ ability to select able leaders, understand the issues upon which they will vote, act as a check on potential abuses by the government, monitor bureaucratic proficiency, and expose corrupt leaders (Brand 2010). In terms of fostering a spirit of participation, schools are one of the main drivers of socialization. They motivate students to recognize and adhere to collective interests that override individual preferences, to show civic behavior, to choose mutual respect and trust, and to foster reciprocity. In fact, most colleges and universities encourage students to participate in various forms of voluntary service. Empirical evidence shows that participation during college positively affects citizens’ interest in and knowledge of political issues, their involvement in the political process, and the effectiveness of their political participation (Huang

15  Civil Society and the Role of Education     315

et al. 2009). Moreover, by enrolling students from different socioeconomic backgrounds in the same system, schools foster acceptance of difference in races, religions, and opinions, which further develops mutual respect. The next section is devoted to a discussion of the empirical studies tackling the nexus between corruption, social capital, and education. It starts with the link between social capital and corruption, continues with the relationship between education and social capital, and ends with studies that directly examine the link between education and corruption.

2 Actual Impacts of Education on Corruption 2.1 Social Capital and Corruption Grießhaber and Geys (2012) analyze the relationship between one component of social capital (civic engagement) and corruption using a sample of 20 European democracies and the Corruption Perceptions Index (CPI) of Transparency International (TI) for the year 2003 as the central dependent variable. The main independent variables incorporate information about citizens’ involvement in formal social networks. The information is extracted from the 2002/2003 round of the European Social Survey (ESS), which asks respondents, among other things, whether over the preceding 12 months they were members of, participated in, donated money to or did voluntary work for one or more of 12 types of association (sports/outdoor activity, culture/hobby, trade unions, professional, consumer, humanitarian/human rights, environment/peace/animal rights, religious, political, education/teachers/ parents, social club, other). A distinction between exclusive and inclusive social capital is made, based on the difference between the networks or organizations in which the respondent participates. Exclusive social capital is associated with a strong in-group/out-group distinction and inward orientation. It focuses predominantly on members’ personal interests. Inclusive social capital has a broader societal focus that spans

316     K. Sekkat

across the boundaries of the organization. GDP per capita is used as a control variable. The results show that the level of perceived corruption in a country is negatively associated with a society’s degree of civic engagement. However, not all types of associations share the same positive effect. The distinction between inclusive and exclusive networks reveals that only involvement in inclusive networks is associated with lower corruption, while involvement in exclusive networks shows the reverse tendency. Uslaner (2005) focuses on the causal links between one component of social capital, namely trust, and corruption. The analysis considers 23 developed and developing countries and uses the 2SLS estimation method, which makes it possible to determine the cause(s), the consequences(s), and the potential existence of a reciprocal causation. Trust and corruption are the dependent variables. Corruption is measured by the CPI for the year 1998. It is explained in terms of generalized trust, external tariffs, protection of property rights, democratic freedoms in a country, and a measure of how strongly people believe that the devil exists. Trust data come from the World Values Survey and pertain to year 1995. They concern the response to the question: “Generally speaking, do you believe that most people can be trusted, or can’t you be too careful in dealing with people?” Trust is explained in terms of corruption, the percentage of Catholics in the population, and the level of economic inequality. Two versions of the system are estimated: one uses variables in levels and the other uses variables in first differences. The estimates using variables in levels suggest that trust causes corruption. According to the corruption regression, more than half of the total distance between the most corrupt and the least corrupt countries in this sample is caused by trust. In the sample, the total distance between the most corrupt and the least corrupt countries amounts to 8.1 points, and moving from the least to the most trusting society reduces corruption by 4.278 points. Belief in the devil has the same general effect, changing corruption by 4.298 points. The other variables have weaker effects. Moving from the least to the freest country reduces corruption by 2.376 points. Countries with the greatest protection of property rights have substantially less corruption (a reduction of 3.75 points). Turning to the impact on trust, the estimation shows that the

15  Civil Society and the Role of Education     317

biggest impact comes from corruption. The impact of corruption is substantially larger than that of trust on corruption. Societies with more economic inequality and with larger shares of Catholics are less trusting. The estimates using variables in first differences suggest a more important impact of trust on corruption. Countries that become more trusting become less corrupt, and countries that become less trusting show an increase in corruption. However, the analysis shows no causal effect of corruption on trust. As societies become less corrupt, they do not become more trusting. Hence, it appears that the most robust causal link is from trust to corruption, not the other way around. Hollard and Sene (2016) further highlight the role of trust by testing the causal impact of self-reported trust on access to basic health facilities in 16 sub-Saharan African countries. The barriers to access to basic health facilities include corruption, costs, supply of medicine, doctor absenteeism, long waiting times, dirty facilities, and lack of attention. The data come from the Afrobarometer on health-center quality for the year 2005. Each of the indicators of barriers to access to basic health facilities is aggregated at the district level and explained in terms of trust and control variables. Two indicators of trust are used: generalized trust and trust in neighbors. The generalized trust indicator is measured based on the General Value Survey (GVS) question: “Generally speaking, would you say that most people can be trusted, or that you cannot be too careful in dealing with people?” The district level of trust is given by the percentage of individuals who answer “Most people can be trusted”. The indicator of trust in neighbors is based on the question: “How much do you trust your neighbors?” Respondents choose between four possible answers: (i) not at all, (ii) just a little, (iii) somewhat, and (iv) a lot. Control variables include ethnic fractionalization, median age, proportion of men, education, and proportions of religious groups. Instrumental variable estimation is used to single out causality. The results reveal a positive and significant effect of trust on healthcenter quality. The coefficients are both highly significant and large. A one standard deviation rise in trust leads to a fall of 0.221 standard deviations in doctor absenteeism, 0.307 standard deviations in waiting times, 0.301 standard deviations in bribes, and 0.330 standard deviations in problems with poor facilities. Moreover, trust in neighbors

318     K. Sekkat

seems to play a larger role than generalized trust regarding the ability to produce local health-center quality. Li and Wu (2010) take a different perspective to examine the role of trust in relation to corruption. They examine whether the impact of corruption on economic growth depends on the level of trust. In other words, does corruption tend to reduce economic growth more in countries with a relatively low level of trust than in countries with a relatively high level of trust? To test for this possibility, the authors conduct a quantitative test using a pooled dataset of 65 countries in two time periods: 1994–1999 and 1999–2004. The dependent variable is economic growth and the two main variables of interest are corruption and trust. The estimated equation incorporates an interaction term between corruption and trust to test the hypothesis that the effect of corruption on growth depends on the level of trust. The measure of corruption is the CPI. The measure of trust comes from the World Values Survey based on a question in the survey which asks: “Generally speaking, would you say that most people can be trusted or that you cannot be too careful in dealing with people?” The percentage of people answering “yes” over the total sample is used as a measure of trust. Control variables include years of schooling of the total population aged over 15 from Barro and Lee (2001) and Freedom House’s classification of the political system of a country into “Not Free”, “Partially Free”, and “Free”. Econometric tests support the hypothesis that trust mitigates the negative effect of corruption on economic growth. Corruption has a negative effect on economic growth, trust shows a positive effect on economic growth, and the interaction term has the expected sign. The interaction term between corruption and trust suggests that the negative effect of corruption on economic growth is reduced by a higher level of trust. López and Santos (2014) argue that trust can take different forms with different impacts on corruption. They therefore split trust into two categories. One category concerns universalistic or generalized trust, which refers to the attitude toward society as a whole and favors the functioning of the economy and society by bridging relationships with members of different communities (ethnic, socioeconomic, and so on).

15  Civil Society and the Role of Education     319

The other category concerns particularistic trust, which is made up of family ties, relations with fellow members of the same family, or specific groups with strong intra-group ties. Universalistic trust is expected to constitute positive social capital that is negatively linked to corruption, while particularistic trust constitutes negative social capital favoring corruption. Consequently, societies featuring family trust and strong intragroup ties are more likely to develop networks that facilitate corruption and even its social acceptability. This is especially the case when high particularistic trust and very low general trust coexist. Econometric tests of the predictions are applied to a cross-section of 60 countries for the year 2001. The TI corruption index is explained in terms of trust indicators and a Global Freedom Index. Indicators of universalistic or generalized trust and of particularistic trust are derived from the World Values Survey. The analysis shows a clear link between generalized trust and lack of corruption. Conversely, particularistic trust is related to the presence of corruption.

2.2 Education and Social Capital Dee (2004) seeks to identify possible causal effects of additional schooling on civic behavior. The paper uses the pooled 1972–2000 General Social Surveys (GSS), which are nationwide surveys conducted on a broad range of attitudes and behavior. The surveys concern persons who lived in the USA at age 16 and were 14 years old between 1914 and 1978. In each survey, respondents were asked questions about their educational attainment and whether they voted in the last presidential election. In most years, GSS respondents were also asked how often they read newspapers, their group memberships (e.g., fraternal and community service, political, school service, youth, church), and their attitudes toward free speech for particular groups (homophobic, racist, etc.). Control variables include gender, year of birth, and the quality of public schools. The econometric analysis shows that the number of years of completed schooling (main explanatory variable) influences each of the measures of civic attitudes (dependent variable). The results suggest that

320     K. Sekkat

additional schooling, both at the secondary and at the post-secondary levels, has large and statistically significant effects on voter participation. Additional secondary schooling significantly increases the frequency of newspaper readership as well as the degree of support for controversial free speech. Milligan et al. (2004) also investigate whether education provides social benefits through enhanced civic participation. The authors focus on the USA and the UK and test whether civic participation as measured by the probability of voting is affected by the degree of educational attainment. The data for the USA come from National Election Studies (NES) spanning the period 1948–2000 and are combined with demographic information on the respondent and a wide variety of questions about political affiliations, voting behavior, knowledge, and attitudes. For the UK, the data come from the British Election Studies (BES), which collect information pertaining to general elections, and from the Eurobarometer survey assembled by the European Commission and designed to track opinions and attitudes among European citizens. The empirical analysis uses responses to the question “Did you vote in the last elections?” as the dependent variable. The explanatory variable of interest is a dummy equal to 1 if the respondent has a high school education or more and zero otherwise. Control variables include demographic information on the respondent and information about her/his political affiliations, voting behavior, knowledge, and attitudes. The analysis reveals a strong and robust relationship between education and voting in the USA, but not in the UK. However, education increases citizens’ attention to public affairs and politics in both countries. More educated citizens appear to have more information on candidates and campaigns. Overall, these results lend support to the notion that education can have social externalities through the production of a better polity. Persson (2014) argues that most of the results regarding the relationship between education and social capital might reveal correlation instead of causality. The idea is that education is not a direct cause of political participation but only a proxy for the real cause. The relationship occurs due to self-selection, i.e., the kind of people who seek higher education are also more likely to participate in politics regardless

15  Civil Society and the Role of Education     321

of their level of education. The author uses matching techniques to assess whether there is a direct causal effect of education on political participation. The data employed concern everyone born during one week in April 1970 in the UK and include a rich set of variables measuring various factors pertaining to childhood and adolescence, intellectual ability, and family socioeconomic status. The dependent variables reflect the political participation profile of the individual and include voting in the 2001 election, signing a petition, contacting a member of parliament (MP), attending a public meeting or rally, or participating in a demonstration during the last 12 months. Each of these participation indicators is analyzed separately and controlled for the effects of intellectual ability and family socioeconomic status, among other things. Overall, the results do not support the hypothesis that education has a significant effect on political participation, which is similar to the findings by Milligan et al. (2004) for the UK. Accordingly, both studies seem to suggest that the education effect is context dependent and that direct causal effects could develop in some contexts but not in others. As noted by Persson 2014, it is not clear that applying his approach to the USA would have contradicted the results of Milligan et al. (2004), that is, education does have an effect on political participation. Britain is a society in which the link between social class origin and adult position is strong. As a result, while the absence of an impact of education on participation in the UK seems to be robust, it cannot be generalized to other contexts. Egerton (2002) mitigates the above findings about the missing link between education and social capital in the UK by showing that the effect depends on the type of civic participation and social group under consideration. The paper examines the effects of tertiary education on the social and civic engagement of young people, using the British Household Panel Study (BHPS), which collects data annually on memberships and activity in a variety of organizations such as political parties, trade unions, environmental groups, school associations, residents’ groups, religious groups, and sports clubs. The youth in question were observed in the year before entering tertiary education (age 17/18) and on completing their tertiary education (age 22/23). Membership and

322     K. Sekkat

activity are explored using 1991 data and coded such that 0 means no membership, 1 means membership, and 2 means active membership. These scores are used as dependent variables in a logistic regression which explains the likelihood of activity in a given organization at age 22 or at the end of tertiary education by previous activity and education level. Five levels of post-secondary education are considered: graduate, sub-degree, A-level, NVQ3, and below NVQ3.1 Distinction is made according to gender, social class of origin, and father’s occupation. Social classes include manual workers, self-employed, clerical, professional, ancillary, and managerial. The results are the following. No important differences are found in the participation rate in sports and social clubs before and after tertiary education. However, differences exist before and after tertiary education regarding participation in civic and religious organizations. Higher education seems to have an effect on the probability of involvement in civic organizations. The children of professionals are more likely to be active in civic organizations compared with the children of manual workers, while the same disparity does not exist for the children of managers compared with the children of manual workers. The analysis by Brand (2010) offers a similar mitigation regarding the effect of education on social capital, but focuses on the USA. The focal point of the analysis is whether schooling succeeds in promoting civic participation among underprivileged groups. To this end, the effects of college on civic participation are allowed to differ between disadvantaged and advantaged participants. The data are drawn from the US National Longitudinal Survey of Youth (NLSY) 1979. Besides demographic and school achievement information, the NLSY gives information about civic participation. The empirical exercise consists in estimating the effects of college completion on civic participation. Civic participation is measured through two dichotomous indicators which are constructed from the responses to the following two questions: (i) “Have you performed any unpaid volunteer work in the past

1Level 3 National Vocational Qualification (NVQ3): competence-based vocational qualification equivalent to an A-level school-leaving qualification.

15  Civil Society and the Role of Education     323

12 months for civic community or youth groups?” and (ii) “Have you performed any unpaid volunteer work in the past 12 months for charitable organizations or social welfare groups?” Civic participation is allowed to differ between disadvantaged and advantaged participants. The separation between the two groups is based on the likelihood of college completion, which is assumed to depend on gender, race, ethnicity, family background, friends’ plans, and parents’ encouragement. The analysis confirms the results of the literature on the determinants of college completion. Black and Hispanic students are significantly less likely to complete college. High socioeconomic background, cognitive ability, academic achievement in high school, friends’ plans for college, and parents’ encouragement to attend college also strongly predict college completion. Regarding civic participation, college completion has the largest impact on volunteering among disadvantaged individuals. The college effect generally decreases as the likelihood of college completion increases. Charron and Rothstein (2016) examine the impact of education and institutional quality on trust in individuals. The tests are based on the results of a survey conducted by the authors in 2010 and in 2013 in 24 European countries. The 2010 survey consisted of about 34,000 citizen interviews, while the 2013 survey covered over 85,000 individuals. The respondents were sampled by regions in European countries, which made it possible to test for sub-national variations within countries. The surveys focused on citizen perceptions and experiences regarding the quality of their regional institutions and also included questions about social trust. The measure of generalized trust is the dependent variable and is based on questions similar to those mentioned above. It is coded as a dummy taking the value 1 if the answer to the question is “Most people can be trusted” and 0 if the answer is “Can’t be too careful”. The two independent variables of interest are education and institutional quality. Education consists of four ordered categories: less than secondary, secondary degree, tertiary degree (university or college), and post-tertiary degree (such as an MA, Ph.D., or JD). Institutional quality indicators include “control of corruption”, “government effectiveness”,

324     K. Sekkat

“rule of law”, and “voice and accountability” taken from the World Governance Indicators. The regression also includes an interaction between education and institutional quality. Control variables are age, income, number of working females, population, and ethnolinguistic fractionalization. The findings show that the effect of education on generalized trust is highly conditional on country levels of institutional quality. At low levels of institutional quality, the results show no statistical differences in the impact across education levels. The difference in the impact on trust across various levels of education increases significantly with better institutional quality. The conditional effect of institutional quality is consistent at both the regional and the country levels but it is noticeably stronger at the country level at all levels of education. The empirical evidence is highly robust to alternative model specifications, data sources, and the removal of outliers. Huang et al. (2009) conduct a meta-analysis of 142 studies on social trust and 268 on social participation with the objective of assessing the effect of education on social trust and social participation. The results support the positive impact of education on the generation of social capital. The mean effect of the studies is 0.046 for social trust and 0.055 for social participation per year of schooling. Thus, one standard deviation of years of schooling accounts for 12–17% of the standard deviation in social trust and social participation.

2.3 Education and Corruption Uslaner and Rothstein (2016) offer a historical perspective on the relationship between education and corruption. Specifically, the paper examines the link between levels of mass education in 1870 and corruption levels in 2010 for 78 countries. Two interesting observations are behind this exercise. First, the persistence of high levels of corruption in many countries suggests that corruption has very old roots. Second, while the mean education level across countries increased by sixfold between 1870 and 2010, the ranking of countries remained almost unchanged. Countries with the highest levels of education at the start of

15  Civil Society and the Role of Education     325

the period were also those at the top 140 years later. Although countries can “catch up”, the weight of history seems heavy. Using an instrumental variable estimation method, the 2010 level of corruption drawn from TI is explained in terms of levels of education in 1870, the mean school year change between 1870 and 2010, GDP per capita in 2000, the Polity IV Democracy Index, a Gini index for 2004, and the Press Freedom Index for 2002. The results show a strong correlation between the mean number of years of schooling in 1870 and the level of corruption in 2010. Moving from the lowest mean years of education (0.01 for four African nations) to the highest (6.07 in Switzerland) increases the CPI by 7.0 points on a ten-point scale. The effect of the level of education in 1870 explains the level of corruption in 2010 much more than GDP per capita. Combining the results of various regressions, the authors also find that press freedom can help combat corruption, but the power of the press depends upon a literate public. Dutta and Mukherjee (2016) examine the combined effect of education attainment and durable democratic systems on corruption using cross-national panel data for the period 1986–2009. The data cover 92 countries and are averaged over five-year periods. The dependent variable, corruption, is taken from the ICRG. The independent variables of interest are durability of a democratic system and measures of educational attainment. Durability is computed based on the Polity IV database. Educational attainment comes from the World Development Indicators database and reflects the percentage of people aged 15 and above who can understand, read, and write a short, simple statement on their everyday life. Control variables also come from the World Development Indicators and include GDP per capita, the degree of trade openness, Internet users (per 200 people), mobile users (per 100 people), life expectancy rate, mortality rates, and poverty. The potential endogeneity concerns are addressed using system GMM. The findings indicate that a stable democratic regime and an educated populace act as complements in combating corruption. Literacy favors corruption at low levels of durability but as a democracy grows more durable, literacy reduces corruption. In contrast, the impact of

326     K. Sekkat

durability alone is nonsignificant irrespective of the levels of literacy, which contradicts Treisman (2000). Cheunga and Chan (2008) examine the interplay between education and national culture in reducing corruption. As discussed above, by giving citizens the ability to learn, comment, and organize, education can help fight corruption. Cultural values which develop social trust and encourage coordination and cooperation for mutual benefit can be efficient tools against corruption. The two factors are not independent, however. Education has an important influence on values and beliefs, teaching students about what is legal and illegal for example. It thus gives students the knowledge and experience to assume ethical responsibilities and maintain value correctness. To examine this nexus, the study explains the TI corruption index in 56 countries for the years 2002 and 2005 in terms of enrolment in tertiary education and Hofstede’s four cultural indexes. The latter include the extent to which a society accepts human inequality, whether the individual is the main focus of people’s life, the extent to which a society treats men better than women, and finally, tolerance for uncertainty and ambiguity within a society. As the control variables, GDP per capita and political rights are included. From the regression, the coefficient of GDP per capita and enrolment in tertiary education are highly significant and have positive slopes. This means that these variables reduce corruption in the two time periods 2002 and 2005. Moreover, they predict around 80% of the total variance of the CPI scores. In comparison, cultural variables explain only a modest portion of the CPI scores. However, cultural variables appear as strong determinants of GDP per capita and enrolment in tertiary education in the two time periods. Hence, cultural dimension variables affect the CPI scores indirectly. In sum, policies to fight corruption should also incorporate instruments that encourage higher enrolment in tertiary education and economic development. Glaeser and Saks (2006), focusing on US states, examine whether voters with more education and income are more willing and able to monitor public employees and to take action when these employees violate the law. Corruption data come from the Justice Department and concern the number of federal, state, and local public officials convicted

15  Civil Society and the Role of Education     327

of corruption between 1976 and 2002. Dividing these numbers by average state population gives the state conviction rate per capita, which is explained in terms of education measured as the share of the population aged 25 and up with four or more years of college education. Control variables include state population, the share of urbanized population, and the fraction of workers employed in government. An instrumental variable estimation method is used to deal with potential endogeneity. The results show that the impact of education on the rate of corruption convictions is strong and much more robust than the impact of income. As the share of highly educated people increases by one standard deviation (2.2 percentage points), the corruption conviction rate decreases by about half of a standard deviation (0.064 percentage points). While recognizing education as a tool in the fight against corruption, Asongu and Nwachukwu (2016) investigate whether the move from one level of education to the next entails an additional reduction in corruption. The paper also examines whether there is a synergy effect in the sense that the effect of the combined three levels of duration is greater than the sum of the impacts from moving from one level to another. The study uses a panel of 53 African countries for the period 1996–2010. Corruption is measured on the basis of the TI and World Bank (WB) indexes. The explanatory variables of interest are primary, secondary, and tertiary school enrolment. To get the synergy indicator, a principal component analysis is employed over the three levels of education. Control variables include GDP growth, inflation, and openness to trade. To deal with potential endogeneity, the system GMM estimation method is adopted. The main policy implications derived from the study are as follows: (i) Education is a powerful tool in the fight against corruption; (ii) there is evidence of an incremental effect in the transition from secondary to tertiary education; and (iii) there is evidence of a “synergy effect” because the impact of combining the three levels of school is greater than the sum of the individual effects of the three educational levels. The various mechanisms discussed so far have shown education to have positive effects in curbing corruption. One interpretation of this

328     K. Sekkat

effect is that education makes citizens less tolerant of corrupt acts. This is the question addressed by Truex (2011) on the basis of a survey of Kathmandu residents. The survey was developed to isolate attitudinal differences across seven dimensions of corruption. It was administered in Kathmandu in June 2009 to 853 Nepali citizens. Participants were asked a series of 13 questions that contained short descriptions of behaviors. For each question, respondents gave their opinion about the behavior on a scale of one to five, with one being “very acceptable” and five being “very unacceptable”. Lower scores indicate greater acceptance of the behavior. A preliminary analysis of the survey results shows that, in general, behavior other than large-scale bribery (grand corruption) proved more socially acceptable. There is also substantial variance among Kathmandu residents in attitudes toward different types of corrupt behavior. Respondents are more tolerant of small-scale corruption (petty corruption), especially when citizens are seeking access to services that they are entitled to anyway. Favoritism also emerges as more acceptable, although attitudes depend on the specific context. The responses suggest that respondents make a distinction between public and private behavior. In contrast, they make no significant distinctions between gift and cash bribes, briber and recipient, and nepotism involving politicians and bureaucrats. To examine the reason for the differences in tolerance toward corruption, an OLS regression is performed using the respondent’s response to the questions as dependent variable and education as the explanatory variable of interest. Education is coded into five categories pertaining to respondents’ highest level of completed schooling: no formal schooling, elementary school, middle school, high school, and university. Control variables include a number of personal variables such as gender, age, and region of origin. In all specifications, respondents with higher levels of education show less tolerance. This attitude is continuous across levels of education: University graduates are less tolerant than high school graduates, high school graduates are less tolerant than middle school graduates, and so forth. The results are robust to the inclusion of the personal covariates. Overall, there is evidence that education does indeed have positive

15  Civil Society and the Role of Education     329

benefits in terms of reducing tolerance of corrupt behavior, and that these effects continue at higher levels of schooling. Since corruption is a form of cheating, Magnus et al. (2002) further highlight the possible impact of education on corruption by examining student opinions about cheating in Russia, the Netherlands, Israel, and the USA. The basis of the study is a survey conducted in 1997 in the four countries at three different levels of education. The sample contained 885 students from the four countries: 92 high school students, 554 university undergraduates, and 239 postgraduates. Out of the 885 students, 506 were from Russia, 112 from the USA, 247 from the Netherlands, and 20 from Israel. Each student was asked to consider the following situation: Student A cheats while taking an exam, C reports to the departmental office that A copied the answers from student B with the consent of student B. Each respondent has to give his/her opinion about each of A, B, and C on a five-point scale: strongly negative (–2), negative (–1), neutral (0), positive (+1), or strongly positive (+2). The authors applied an ordered-response model to identify differences in attitudes toward cheating across countries and education levels. The test showed that students of the same educational level in different countries had very different attitudes toward A, B, and C. However, students within the same country had the same attitude toward A and B irrespective of their educational level. Students within the same country did not have the same attitude toward A and C. To directly link the analysis to corruption issues, the authors use the responses to the interviews to construct a synthetic indicator by country and education level of tolerance to cheating. The index is a weighted average of the attitude toward A, B, and C and varies between 0 and 10. The higher the index, the lower the tolerance of cheating. Tests show that the index explains most of the variation in the data. Computing the correlation between TI’s CPI and the synthetic index of tolerance of cheating shows very high correlation. From disapproving of corrupt acts to reporting them takes additional courage and greater intolerance of corruption. Botero et al. (2013) explore whether educated people are more likely to complain about and report officials’ misconduct. They also investigate whether

330     K. Sekkat

such complaints and reports increase as the education level in a country rises. The ultimate impact on corruption comes from the fact that a public official choosing to break the rules must trade off the risk of being disciplined, which increases with the number of complaining citizens, against the benefits of misconduct. Of course, the effectiveness of this decentralized process also depends on the prevailing political arrangements, such as democracy or dictatorship, and the quality of institutions. Accordingly, control for these factors should be carefully implemented. The analysis is based on cross-sectional data from the World Justice Project (WJP) surveys conducted in 88 countries in 2009, 2011, and 2012 and comprising 1000 respondents drawn from the three largest cities in each country. The questions relevant to the analysis are worded as follows: (i) “During the last year, did you submit any complaint about the services provided by the different government agencies in your country?”; (ii) “In the last 3 years, have you or someone in your household, been subjected to physical abuse by the police or the military?”; and (iii) “Did you or anyone else report the crime to the police or other authority?” The WJP data also provide information on education, income, trust, gender, and asset ownership. These make it possible to construct two indicators of the education level: college and high/ middle school. WJP data are supplemented by information from the International Crime Victims Survey (ICVS) and TI’s Global Corruption Barometer. The econometric analysis examines the relationship between education and the reporting of officials’ misconduct or crime. The dependent variable is a dummy equal to zero or one depending on whether reporting occurred. The main findings confirm that, within countries, more educated people complain more, both about officials’ misconduct and crime in general. This relationship is stronger in developing countries. A higher country-average assessment that a policeman violating the law will be punished is associated with higher probability of making a complaint. Interestingly, this relationship is particularly strong in autocracies, suggesting that voting might not be the only important tool against corruption. Finally, these results seem to be linked to

15  Civil Society and the Role of Education     331

education per se and not to other factors correlated with education such as income, trust, social status, or gender. In the chapter on justice, we discussed the issue of partisan bias, through which voters continue to elect an incumbent convicted of corruption for partisan reasons (Gordon 2009). Winters and Weitz-Shapiro (2014) examine the effect of education on such bias. The basis of the analysis is a survey conducted in 2010 in Brazil with 2000 respondents. The survey presented respondents with variants of the following scenario: Imagine a person named Gabriel, who is a person like you, living in a neighborhood like yours, but in a different city in Brazil. The mayor of Gabriel’s city is running for reelection in October. He is a member of the party X. In Gabriel’s city, it is well known that the mayor never takes bribes when giving out government contracts. The mayor has completed few public works projects during his term in office. In this city, the election for mayor is expected to be very close.

The variants of the scenario consist in replacing the words in bold and italics by one or more of the following: The party Y, frequently takes bribes and many public works projects. The variants are assigned to respondents randomly. Some respondents learn that it is well known that the mayor takes bribes while others learn that it is well known that the mayor does not take bribes. The experimental scenario was followed immediately by a question that asked the respondent the likelihood (on a four-point scale) that Gabriel would vote for the named mayor. Respondents were also asked to select the party to which they felt closest from a list of 27 parties, as well as their education level. To examine how corruption information affects partisanship, the authors examine differences in the distribution of expressed partisan identities across corrupt and non-corrupt conditions. The randomized design makes it possible to attribute changes in this distribution to the corruption effect. The results show that among less educated respondents, the majority of people did not update their partisan identification after becoming aware of corruption in conjunction with a particular party. Among the highly educated group, sympathy for the party

332     K. Sekkat

associated with corruption decreased while sympathy for other parties increased. Specifically, the share of highly educated respondents who sympathized with the “corrupt” party dropped from 25 to 20%. In contrast, the share of the same set of respondents declaring a partisan identity with a party other than the “corrupt” one increased by 15 percentage points (from 32 to 47%).

3 Conclusion Education plays a major role in enhancing the fight against corruption in many respects. First, the use of ICT in administrative and political processes is obviously very inaccessible to the illiterate population. Second, education contributes to building a population’s social capital, broadly defined as a set of norms, organizations, and networks which creates hostility toward corruption. Third, education favors civic engagement and encourages people to go beyond disapproving of corrupt acts to reporting and exposing them. Observations show that countries that are more endowed with social capital are less corrupt. More interestingly, countries where social capital has been depleted become more corrupt. The level of corruption in a country is also found to be negatively associated with a society’s degree of civic engagement. Moreover, as the number of years of completed schooling increases, civic engagement intensifies. Additional years of schooling have large and statistically significant effects on the frequency of newspaper readership, the degree of support for controversial free speech and voter participation. Furthermore, educated citizens have more information on candidates and campaigns. Historical investigations over a very long period (1870–2010) confirm the above findings. The results show a strong negative correlation between the average number of years of schooling in a country in 1870 and the level of corruption in 2010. Moreover, a stable democratic regime and an educated population act as complements in combating corruption. Even within countries, more educated people are found to complain more, both about officials’ misconduct and crime in general.

15  Civil Society and the Role of Education     333

In other words, more educated persons are not afraid to complain. This relationship is stronger in developing countries. In sum, anti-corruption policies should incorporate instruments that encourage higher enrolment in tertiary education.

References Asongu, S. A., & Nwachukwu, J. C. (2016). The Mobile Phone in the Diffusion of Knowledge for Institutional Quality in Sub-Saharan Africa. World Development, 86, 133–147. Barro, R. J., & Lee, J. W. (2001). International Data on Educational Attainment: Updates and Implications. Oxford Economic Papers, 53(3), 541–563. Beugelsdijk, S., & Van Schaik, T. (2005). Social Capital and Growth in European Regions: An Empirical Test. European Journal of Political Economy, 21(2), 301–324. Botero, J., Ponce, A., & Shleifer, A. (2013). Education, Complaints and Accountability. Journal of Law and Economics, 56(4), 959–996. Brand, J. E. (2010). Civic Returns to Higher Education: A Note on Heterogeneous Effects. Social Forces, 89(2), 417–433. Charron, N., & Rothstein, B. (2016). Does Education Lead to Higher Generalized Trust? The Importance of Quality of Government. International Journal of Educational Development, 50(C), 59–73. Cheung, H. Y., & Chan, A. W. (2008). Corruption Across Countries: Impacts from Education and Cultural Dimensions. Social Science Journal, 45(2), 223–239. Dee, T. S. (2004). Are There Civic Returns to Education? Journal of Public Economics, 88(9), 1697–1720. Dutta, N., & Mukherjee, D. (2016). Do Literacy and a Mature Democratic Regime Cure Corruption? Journal of Economic Development, 41(2), 1–26. Egerton, M. (2002). Higher Education and Civic Engagement. British Journal of Sociology, 53(4), 603–620. Glaeser, E. L., & Saks, R. E. (2006). Corruption in America. Journal of Public Economics, 90(6), 1053–1072. Gordon, S. C. (2009). Assessing Partisan Bias in Federal Public Corruption Prosecutions. American Political Science Review, 103(04), 534–554.

334     K. Sekkat

Grießhaber, N., & Geys, B. (2012). Civic Engagement and Corruption in 20 European Democracies. European Societies, 14(1), 57–81. Hollard, G., & Sene, O. (2016). Social Capital and Access to Primary Health Care in Developing Countries: Evidence from Sub-Saharan Africa. Journal of Health Economics, 45, 1–11. Huang, J., Van Den Brink, H. M., & Groot, W. (2009). A Meta-Analysis of the Effect of Education on Social Capital. Economics of Education Review, 28(4), 454–464. Knack, S., & Keefer, P. (1997). Does Social Capital Have an Economic Payoff? A Cross-Country Investigation. Quarterly Journal of Economics, 112(4), 1251–1288. Li, S., & Wu, J. (2010). Why Some Countries Thrive Despite Corruption: The Role of Trust in the Corruption-Efficiency Relationship. Review of International Political Economy, 17(1), 129–154. Lochner, K., Kawachi, I., & Kennedy, B. P. (1999). Social Capital: A Guide to Its Measurement. Health and Place, 5(4), 259–270. López, J. A. P., & Santos, J. M. S. (2014). Does Corruption Have Social Roots? The Role of Culture and Social Capital. Journal of Business Ethics, 122(4), 697–708. Magnus, J. R., Polterovich, V. M., Danilov, D. L., & Savvateev, A. V. (2002). Tolerance of Cheating: An Analysis Across Countries. Journal of Economic Education, 33(2), 125–135. Milligan, K., Moretti, E., & Oreopoulos, P. (2004). Does Education Improve Citizenship? Evidence from the United States and the United Kingdom. Journal of Public Economics, 88(9), 1667–1695. Narayan, D., & Pritchett, L. (1999). Cents and Sociability: Household Income and Social Capital in Rural Tanzania. Economic Development and Cultural Change, 47(4), 871–897. Persson, M. (2014). Testing the Relationship Between Education and Political Participation Using the 1970 British Cohort Study. Political Behavior, 36(4), 877–897. Portes, A. (1998). Social Capital: Its Origins and Applications in Modern Sociology. Annual Review of Sociology, 24(1), 1–24. Shim, D. C., & Eom, T. H. (2009). Anticorruption Effects of Information Communication and Technology (ICT) and Social Capital. International Review of Administrative Sciences, 75(1), 99–116. Treisman, D. (2000). The Causes of Corruption: A Cross-National Study. Journal of Public Economics, 76(3), 399–457.

15  Civil Society and the Role of Education     335

Truex, R. (2011). Corruption, Attitudes and Education: Survey Evidence from Nepal. World Development, 39(7), 1133–1142. Uslaner, E. M. (2005). Trust and Corruption. In J. G. Lambsdorff, M. Taube, & M. Schramm (Eds.), The New Institutional Economics of Corruption (pp. 76–92). London: Routledge. Uslaner, E. M., & Rothstein, B. (2016). The Historical Roots of Corruption: State Building, Economic Inequality and Mass Education. Comparative Politics, 48(2), 227–248. Winters, M. S., & Weitz-Shapiro, R. (2014). Political Corruption and Partisan Engagement: Evidence from Brazil. Journal of Politics in Latin America, 7(1), 45–81. Woolcock, M., & Narayan, D. (2000). Social Capital: Implications for Development Theory, Research and Policy. World Bank Research Observer, 15(2), 225–249.

Conclusion

Corruption is old, widespread, and multifaceted. The phenomenon can be traced back to before 300 B.C. Corruption is prevalent in diverse fields such as health, education, justice, sports, unions, and the media. Corruption ranges from transfers of relatively small payments to low-ranking officials (“petty” corruption) to large payments to high-ranking civil servants or politicians (“grand” corruption). It also involves nepotism, favoritism, and theft of state assets. The large majority of the empirical evidence shows that corruption imposes important costs on society, although some economists and political scientists argue that it “greases the wheels” of business in some circumstances. The amount and costs of corruption are unknown since the phenomenon is, by definition, secret. Available assessments suggest, however, that the total amount of corruption is more than US$1 trillion per year. This does not include the negative effects of corruption on investment, human capital, and productivity, each of which reduces GDP by almost 5%. Corruption also has non-economic costs in terms of quality of infrastructure, trust in the state and other citizens, political participation, and regime legitimacy. Corruption has international ramifications. Very often, corrupt politicians, rulers, or high-ranking civil servants put the proceeds of their © The Editor(s) (if applicable) and The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3

337

338     Conclusion

illegal activities in specific foreign countries, with the result that corruption is increasingly associated with the financing of international terrorism and narcotics traffic. Accordingly, the remedy for corruption needs to be transnational. The USA was the first to adopt such transnational treatment in 1977. Its Foreign Corrupt Practices Act (FCPA) punishes corrupt acts outside the USA by any firm linked to the US economy. Since then, similar initiatives have been developed by, for instance, the UN, the OECD, and the WTO. The outcomes of FCPA enforcement are highly encouraging. Anti-corruption strategies adopted around the world can be split into two complementary categories. One focuses on institutional reforms while the other emphasizes the role of civil society. Starting with the top institutional level, the fact that those autocratic regimes are, in general, more corrupt than democracies suggest that moving toward democracy is a prerequisite for the fight against corruption. However, corruption exists in democracies as well for instance because of the need to fund electoral campaigns. Accordingly, complementary mechanisms such as punishments, rewards, and challenges to officials’ decision-making powers have been suggested as ways to keep officials away from corruption. However, implementing such mechanisms is not an easy task. Punishment poses the question of whether the briber, the recipient, or both should be punished. This question is fueling a passionate debate in India, which is both the biggest democracy and one of the most corrupt countries in the world. Punishment also entails costs relating to the collection of evidence, potential conviction of defendants, and enforcement of judgements. Many developing countries lack the financial and human resources to accomplish such complex tasks. Furthermore, the justice system may be itself corrupt or under influence. As an alternative, many countries have set up, with the help of donors, independent Anti-Corruption Agencies (ACAs). Except in Singapore and Hong Kong, ACAs have mostly been failures owing to ineffective independence, insufficient resources and poor monitoring. Rewards generally consist in increasing salaries. In all the countries that adopted this solution, corruption did not decrease because the problem of monitoring remained unresolved. Very often, officials see

Conclusion     339

the increase in wages as top-up revenue instead of as a substitute for corruption. The “monopoly” power of civil servants can be broken up by making specific tasks accomplishable by different administrations, rotating employees in charge of certain tasks or transferring some tasks to the private sector. With the exception of approaches based on having different administrations deliver a public service, these solutions have proven very costly in many cases because agents had to adapt to new routines, cases, and rules. However, when private firms engaged in service delivery are left to themselves, they start promoting corruption instead of curbing it. Overall, the evidence points to inadequate monitoring and the risk of capture by governments and donors as the main reasons for such lack of success. In line with most of the recent researches, the above discussion shows that whatever anti-corruption strategy is adopted, monitoring and control are crucial. This puts civil society on the front line. Because even an “honest” ruler cannot control the actions of each civil servant, collective monitoring by civil society is more effective. Moreover, corruption is increasingly systemic, which explains the failure of many anti-corruption projects. It is almost impossible for a single official to remain honest without running the risk of encountering the hostility of colleagues or superiors, being ostracized or even losing his or her life. For collective monitoring and control to be successful, monitors should be able to access, process, and efficiently use the relevant information. Observations show that a strong civil society has a substantial anti-corruption impact only in countries with high pressfreedom. At the same time, a strong civil society can powerfully complement the media through ICT-mediated social fora, which raise the pressure for information disclosure and accountability. However, the effectiveness of these mechanisms depends on the education level of citizens. Education is commonly associated with the acquisition of knowledge, which is a necessary step toward citizens’ efficient use of the information provided by the media. In addition, education equips citizens with tools for effective participation in democracy. Finally, education fosters citizens’ adoption of civic values.

340     Conclusion

In sum, the fight against corruption, although not easy, is not a lost battle. Some countries, including those in Scandinavia and North America, were among the most corrupt in the world two centuries ago, but are now cited among the “cleanest”. More recently, the anti-corruption strategies of Singapore and Hong Kong have been cited as examples of great success. The historical record of these countries confirms the crucial role played by collective action, information, and education.

Index

A

Absenteeism 28, 248, 317 Accountability 21, 28, 44, 45, 74, 75, 82, 86, 98, 99, 141, 142, 166, 167, 173, 176, 179, 189, 191, 193–196, 199, 200, 208, 213, 214, 218, 219, 228, 231, 234, 237, 238, 241, 245, 250, 284, 289, 291, 293, 294, 296, 305, 313, 324, 339 Accreditation 29 Adjournment 212 Advertisements 24 Africa 21, 22, 30, 58, 85, 93, 133, 136, 144, 190, 246, 253, 254, 256, 257, 259 African Union (AU) 269 Afrobarometer 317 Aid 12, 40–42, 76, 130, 212, 259, 276, 312 Albania xvi

Alert xiv Allegation 54, 270, 286 Amendment 265 America 26, 144, 168, 190, 256– 258, 286 Angola 12 Anonymity 226 Anti-bribery 81, 266, 270, 272, 273 Anti-corruption 47, 64, 78, 93, 111, 112, 165, 172, 173, 200, 223, 231–233, 235–238, 247, 253, 261, 268–271, 273, 276, 290, 301, 302, 304, 306, 308, 333, 338–340 Anti-Corruption Agencies (ACA) xvi, 232, 235–237 Appeal xiv, 60, 107, 211 Appoint 75, 191 Aquifers 31

© The Editor(s) (if applicable) and The Author(s) 2018 K. Sekkat, Is Corruption Curable?, https://doi.org/10.1007/978-3-319-98518-3

341

342     Index

Argentina xvi, 17, 23, 93, 135, 214, 232, 235, 251, 259–261, 269, 270 Asia 31, 45, 94, 133, 168, 190 Assassinations 217, 218 Asset 21, 144, 235, 268, 299, 300, 313, 330 Asset-recovery 270 Asymmetric 205, 224–228 Attorney 215, 250 Attorney-General 215, 237, 250 Audit 65, 234, 250, 251, 284, 285, 293, 294 Australia xvi, 16, 93, 103, 104, 136, 171, 235, 256 Austria xvi, 109 Authoritarian 170 Autonomous Revenue Authorities (ARA) 258 Autonomy 166, 191, 194, 241, 257–259 B

Badmouthing 286 Bangladesh 20, 49, 93, 130, 270 Bankruptcy 15, 285 Banks 151, 218, 286 Barr 93, 249, 250 Barriers-to-entry 187 Bengali 292 Benin 56, 57, 133 Berlin 42 Berlusconi, Silvio 22 Bhutto, Benazir 11, 14 Bicameralism 75, 181, 187 Bolivia 256, 258, 259 Bonn 225

Brazil xvi, 135, 246, 269, 284, 293, 331 Breach 10, 183 Break 19, 120, 236, 252, 330 Bribe 8, 15, 17, 20, 21, 25, 31, 42, 55, 72, 73, 79, 81, 93, 97–101, 103, 104, 110, 111, 113–115, 119, 125–127, 137, 147–149, 151, 154, 167, 172, 197, 198, 200, 205, 211, 221–223, 225–228, 242, 243, 247, 252–255, 266, 270, 274, 275 Bribe Payers Index (BPI) 40, 42, 275 Briber 22, 211, 220, 221, 224– 228, 242, 243, 255, 256, 328, 338 Bribery 6, 13, 32, 43, 48, 75, 81, 93, 94, 99, 100, 121, 126, 147, 148, 198, 199, 221, 222, 224, 227, 228, 232, 243, 256, 262, 265–267, 269, 270, 272–276, 304, 328 Britain 26, 73, 232, 321 British Election Studies (BES) 320 British Household Panel Study (BHPS) 321 Buenos Aires 13, 251 Bulgaria xvi, 269, 302 Bureaucracy 5, 13, 16, 44, 79, 82, 101, 119, 125, 126, 137, 138, 149, 153, 155, 222, 223, 244, 245, 288, 302 Bureaucrat 16, 17, 149, 242, 243, 252 Burkina Faso 56, 57, 247 Bush 216, 217

Index     343

Business Environment and Enterprise Performance Survey (BEEPS) 40, 45, 138 C

Cambodia 12, 28, 29, 32 Canada 16, 31, 93, 140, 274 Candidate 15, 182–185, 307 Cape Verde (CV) 85, 86 Capture 6, 9, 10, 22, 45, 48, 75, 77, 134, 169, 189, 191, 199, 203, 204, 214, 219, 235, 255, 295, 339 Career-concern 187 Cargo 253, 254 Catholic 96, 101 Causality 77, 80, 95, 96, 108, 130, 174, 196, 303, 304, 317, 320 Cell 50 Censoring 300 Centralization 114, 193, 199 Chad 29 Chancellor 10, 15, 183, 184 Charge 14, 28, 30, 42, 74, 111, 112, 167, 179, 209, 232, 244, 253, 258, 262, 266, 295, 339 Cheating 27, 59, 98, 249, 306, 329 Check 21, 80, 81, 98, 225, 234, 314 Chemicals 77, 204 Chicago 31 Chile 135, 221, 269 China xvi, 2, 19, 21, 25, 29, 31, 93, 110, 190, 225, 227, 242, 302, 307 Chirac, Jacques 183 Church 319

Citizen 20, 83, 103, 104, 200, 216, 226–228, 243, 252, 291, 323 City 7, 13, 15, 31, 95, 111, 134, 148, 191, 233, 243, 251, 301, 331 Civic 76, 199, 233, 306–308, 312–316, 319–323, 332, 339 Civil xv, 3, 42, 64, 65, 67, 72, 73, 79, 100, 110, 113, 120, 125– 127, 145, 150, 151, 155, 169, 174, 187, 192, 213, 223, 231, 233, 242, 243, 245, 247, 275, 284, 286, 287, 294, 300–302, 337–339 Civil Society Organization (CSO) 294 Clean Report of Finding (CRF) 257 Clientelism 214 Clinton 216, 217 Closed-list 181, 182, 184–187 Coercive 204, 220 Colgate-Palmolive 270 Collusion 15, 234, 243, 294 Colombia xvi, 27, 135, 259, 261, 262 Colonial 71, 73, 81, 273, 288 Commitment 128, 232, 258, 259 Commonwealth of Independent States (CIS) 24, 138 Communist 7, 22, 110, 139, 170, 246 Competition 2, 7, 14, 19, 31, 45, 52, 53, 71, 72, 74, 75, 77, 78, 88, 89, 107–110, 115, 122, 179, 181, 184–187, 189, 191–193, 196, 199, 200, 204,

344     Index

207, 208, 211, 243, 252, 253, 269, 276, 293 Composite 43, 47, 81, 145 Computer 212 Conduct 41, 74, 267, 268, 273, 284, 289, 304, 307, 318, 324 Confiscating 19 Conflict 44, 175, 211 Congress 14, 18, 265, 276, 294 Constituency 166 Contaminate 31 Control xiv, 14, 17, 23, 41, 45, 56, 66, 67, 74, 78, 79, 83–87, 89, 90, 95, 97–105, 107–109, 111, 114, 129– 133, 137–143, 145–148, 150–153, 170–175, 184– 186, 192, 194–196, 198, 203, 207–209, 218, 219, 232, 233, 237, 244–246, 251, 254, 261, 272–274, 286–290, 293, 295, 296, 302, 303, 307, 316–320, 323–328, 330, 339 Control of Corruption Index (CCI) 45, 56–58, 62, 64, 80, 90, 95, 96, 142, 170, 173, 175, 184, 236, 302 Conviction 242, 327, 338 Corporation 22, 40, 41, 108, 111, 266 Corridor 253, 254 Corrupt xiv, 2, 6–8, 10, 12, 15, 20, 26, 28, 32, 41, 49–51, 54, 58, 60–65, 68, 71–76, 78–81, 84, 86, 88–91, 93, 95–97, 99, 103, 105–107, 112–115, 120, 123, 124, 126–128, 131, 134,

136, 143, 151–154, 165–168, 170, 171, 173, 174, 176, 179–182, 184, 185, 187, 190, 194, 195, 207–209, 211, 212, 214, 215, 220–227, 242, 244, 245, 250, 254–256, 267, 269, 270, 272–275, 283–286, 294, 300, 305, 314, 316, 317, 328, 329, 331, 332, 337, 338, 340 Corruption xiii–xvi, 5–32, 39–68, 71–115, 119–155, 165–176, 179–187, 189–200, 203–209, 211–220, 222–228, 231–234, 236–238, 241–247, 249–257, 260, 262, 265–269, 271, 273–276, 283–291, 293, 294, 296, 299–306, 308, 311, 312, 314–319, 323–332, 337–340 Corruption Perceptions Index (CPI) 2, 40–43, 49–51, 56, 61, 62, 64, 66, 86, 89, 92, 93, 95, 100–102, 105, 107, 108, 132, 133, 135, 136, 139, 141–146, 149, 150, 152, 169, 174, 175, 184, 185, 187, 206, 218, 236, 274, 287, 302–305, 315, 316, 318, 325, 326, 329 Corrupt Practices Investigation Bureau (CPIB) 233 Cost-effective 260 Côte d’Ivoire 56 Counterfeit 28, 29 Court 15, 20, 26, 91, 112, 113, 120, 212–215, 228, 273, 284 Crackdown 250 Credential 26 Crime 6, 21, 40, 43, 45, 55, 147, 148, 221, 267, 330, 332

Index     345

Criminalize 267, 269 Croatian 7 Culture 72, 76, 91, 94, 104–106, 115, 167, 174, 175, 225, 238, 274, 315, 326 Current Population Survey (CPS) 303 Custom 32, 65, 85, 111, 243, 252, 254, 257, 258 D

Dean 10 Death 29, 30, 148 Decentralization xv, 72, 74, 75, 80, 81, 88, 172, 183, 189–192, 194–200, 211 de facto 217–220, 268 Defamation 285 Deferred Prosecution Agreement (DPA) 266, 267 Degree 14, 16, 17, 26, 39, 43, 60, 67, 68, 74, 76, 80, 84, 87, 89, 90, 102, 105, 108, 132, 142, 152, 165, 167, 170–172, 175, 176, 183, 184, 186, 190, 192, 194–196, 200, 203, 205, 214, 217–220, 245, 246, 249, 259, 287–290, 305, 312, 316, 320, 322, 323, 325, 332 de jure 217, 218, 220 Delhi 17, 103, 148 Democracy xiv, 14, 15, 24, 47, 56–58, 67, 72, 74, 75, 80–82, 88–90, 102, 103, 115, 143, 145, 150, 165–176, 179, 180, 185, 187, 191, 193, 194, 198,

209, 274, 275, 287–290, 311, 314, 325, 330, 338, 339 Democrats 216, 217 Denmark 7 Denunciation 284 Department of Justice (DoJ) 41, 227, 266, 267, 290 Detect 59, 84 Deter 132, 175, 224, 227, 228 Diploma 26, 27, 219 Discharge 77, 204 Disclosure 293, 313, 339 Discrepancies 54, 260 Discrepancy Report (DR) 257 Disease xiii Diversion 2, 32, 128, 255, 293 Donation 10, 16, 25, 26 Donor 10, 237, 238 Drive 149, 182, 184, 186, 226 Drug 28, 29, 252 Dubai 11 Duty 222, 260–262 Duvalier, J.C. 270 E

Earthquake 14, 169 Economist Intelligence Unit (EIU) 43 Ecuador xvi, 214, 235 Educate 233 Education xiv, xv, 5, 13, 14, 24–29, 59, 85–87, 91, 97, 98, 101, 104, 119, 122, 124, 129, 130, 142–147, 151, 152, 154, 155, 189, 197, 199, 206, 207, 215, 232, 233, 235, 245–247, 251, 259, 283, 284, 288, 294, 296,

346     Index

304, 307, 311–315, 317, 319–333, 337, 339, 340 Effect 3, 14, 46, 47, 53, 56, 66, 67, 75, 77, 81, 83, 89, 97, 98, 100, 101, 108–110, 112, 114, 122, 123, 129–135, 138, 139, 142, 145–147, 152, 167, 168, 170–172, 174, 175, 187, 190, 191, 193, 194, 197–199, 206, 208, 215, 216, 219, 224–226, 228, 242, 245, 251–253, 256, 258, 260, 269, 275, 287, 292, 294, 295, 303, 304, 306, 313, 316–318, 321–325, 327, 328, 331 Effective xiii–xv, 5, 7, 9, 10, 14, 19, 42, 72, 74, 76, 89, 109, 135, 137, 165–167, 173–175, 186, 192, 194, 196, 200, 212, 220, 223, 224, 227, 228, 233, 241, 248, 251, 256, 258, 260, 262, 265, 266, 273, 284–286, 288, 294, 299, 300, 306, 311–314, 339 E-government 299, 300, 302–306, 308 E-government Development Index (EDI) 302 Egypt 1, 302 Election xiv, 74, 89, 153, 166–168, 172, 179, 180, 182–185, 194, 211, 213, 219, 250, 268, 293, 294, 319–321, 331 Electronic 300 Embezzlement 8, 43, 214, 276, 284 Enforce 65, 205 Entrenched 60 Entrepreneur 113, 114, 205, 208

Entry 127, 139, 172, 181, 187, 203, 205, 206, 252 Envelope 21, 22 Environment 30, 42, 45, 56, 62, 67, 72, 75, 82, 119, 126, 128, 137, 154, 190, 197, 207, 225, 243, 305, 315 E-participation 305 Equity 30, 44, 166, 212 Ethnic 44, 56, 59, 82, 96, 99, 100, 166, 171, 192, 194, 196, 259, 289, 292, 317, 318 Ethnolinguistic 66, 67, 80, 81, 129, 130, 143, 173, 184, 246, 288, 324 Eurobarometer 320 Europe xv, 3, 12, 17–19, 26, 45, 73, 82, 138 European Bank for Reconstruction and Development (EBRD) 6, 9, 40, 45, 56 European Social Survey (ESS) 99, 152, 315 European Union (EU) xv, 134, 135, 227, 233, 260, 269 Exam xvi, 149, 251, 252, 329 Expenditure 75, 81, 146, 174, 191, 194, 236, 293, 294 Export 12, 41, 44, 85, 135–137, 261, 266, 270, 271 Externalities 75, 77, 88, 119, 154, 192, 204, 244, 320 Extortion 21, 25, 92, 114, 214, 221, 252, 253, 286 F

Faculty 25

Index     347

Fake 26, 29 Falsification 260 Federal Emergency Relief Administration (FERA) 292 Federalism 75, 181, 184, 187, 191, 287 FF 12 Fine 10, 27, 226, 294 Finland 7, 290 Florence 2 Food 30, 92, 293 Foreign Corrupt Practices Act (FCPA) 81, 82, 265–267, 269–276, 338 Foreign Direct Investment (FDI) 82, 119, 121, 129, 132, 140, 154, 272 Forestry 149, 171, 247 Foundation 87, 94, 95, 169, 174, 303 France 22, 31, 136, 140, 183, 232, 274 Freedom 23, 83, 87, 88, 90, 95, 102, 130, 143, 145, 147, 169, 170, 174, 175, 184, 189, 194–196, 200, 205, 287–290, 296, 302, 305, 319, 325, 339 Freedom House (FH) xvi, 43, 66, 174, 175, 185, 187, 195, 196, 245, 287, 288, 290, 318 Freedom of Information (FOI) 290 Free-market 58 Fujimori, Alberto 14 G

Gallup 40, 151 Gastil 100, 109, 169, 187

Gender 57, 59, 72, 76, 91–93, 99–104, 112, 115, 151, 152, 307, 319, 322, 323, 328, 330, 331 General Social Surveys (GSS) 319 General Value Survey (GVS) 317 Geography 71–73, 82, 84, 108, 114 Georgia 213 Germany 7, 15, 16, 27, 31, 93, 105, 136, 140, 183, 225, 270 Ghana 22, 49, 237, 238, 247 Gift 8, 45, 91, 113, 291, 328 Global Competitiveness Report (GRC) 40, 41 Governance xv, 2, 9, 40, 44, 45, 54, 57, 58, 86, 87, 101, 103, 112, 115, 129, 141, 142, 147, 192, 207, 208, 235 Graft 13, 126, 127 Grease the wheels 18, 119, 153, 155, 222 Greece 93, 247 Growth 9, 13, 23, 24, 26, 33, 66, 82, 86, 119, 121–123, 125, 126, 128–137, 141–145, 150, 154, 155, 166, 168, 169, 190, 214, 272, 274, 275, 300, 303, 313, 318, 327 Guatemala 256, 259 Guyana 256, 259 H

Harassment 172, 197, 213, 221, 222, 225, 226 Hierarchy 64, 71, 79, 223 Hindi 292 Hispanic 323

348     Index

History 3, 26, 27, 32, 67, 71–73, 80, 82, 84, 104, 114, 115, 325 Hong Kong xvi, 93, 105, 133, 231–233, 235–238, 338, 340 Hospital 27, 28, 55, 251 I

Ideology 22, 219 Illegal 5, 6, 8, 9, 15, 20, 32, 54, 120, 127, 128, 149, 221, 261, 275, 285, 291, 326, 338 Illiterate 292, 311, 332 Import 12, 16, 41, 44, 46, 134–136, 257, 260–262 Import-tax 257 Independent Commission Against Corruption (ICAC) 233 India xvi, 8, 12, 20, 25, 28, 32, 93, 95, 103, 104, 190, 198, 226, 232, 236, 243, 246, 247, 252, 285, 290, 292, 293, 338 Indices 41, 43, 50, 52–56, 61, 64, 80, 81, 129–131, 141–143, 151, 169, 171, 173, 194, 196, 208, 218, 246, 258, 259, 274, 288, 290, 302, 326, 329 Indonesian 11, 58, 172, 196, 234 Inflate 234 Informal xv, 45, 55, 109, 120, 134, 214, 215 Information and Communication Technology (ICT) 283, 299–301, 305, 306, 308, 311, 314, 332, 339 Infrastructure 18, 45, 119, 123, 124, 131, 154, 192, 212, 234, 304, 308, 311, 337

Inspection 17, 18, 257, 258, 261 Integrity 41, 112, 180, 301 Inter-American Convention Against Corruption (IACAC) 269 Inter-governmental 197 Inter-jurisdictional 75, 189, 191, 192 International Anti-Corruption Academy (IACA) xvi International Country Risk Guide (ICRG) 2, 40–44, 49–52, 56, 61–64, 66, 84, 98, 100, 131– 133, 143, 144, 146, 147, 149, 153, 171, 187, 192, 194–196, 208, 245, 246, 259, 287, 288, 290, 303, 325 International Crime Victimization Surveys (ICVS) 40, 41, 43, 55, 148, 330 International Monetary Fund (IMF) 29, 140 International Olympic Committee (IOC) 7 Internet 151, 152, 205, 283, 293, 296, 299, 303–306, 308, 325 Intimidation 213, 285 Invoicing 260, 261 Iraq 270 Irrigation 12, 30, 32 Islam 95, 96, 289 Israel 329 Italy 12, 22, 31, 88, 93, 109, 136, 140, 186, 247 J

Jamaica 30 Japan 31, 140, 246, 302

Index     349

Java 58, 172, 234 Jews 11 Jordan 20 Judge 28, 213, 217, 219 Judicial Independence (JI) 218–220 Jurisdiction 197, 266, 273 Justice xiv, xv, 2, 5, 13, 19, 39, 54, 67, 148, 211, 212, 214, 220, 222, 224, 227, 237, 326, 331, 337, 338 K

Kathmandu 328 Kazakhstan 25, 93 Kenya 30, 111, 112, 256, 259 Kickbacks 18, 43 Kohl, Helmut 10, 15, 183, 184 Korea xvi, 93, 111, 136, 166, 235–237, 246, 302 Korea Independent Commission Against Corruption (KICAC) 301 Kun Koh 301 L

Language 17, 134, 135, 272, 293, 295 Latin 89, 94, 135, 144, 168, 190, 206, 256–258 Laundering 268 Law xiv, 2, 8–10, 15, 19, 26, 44, 45, 56, 66, 73, 76, 81, 96, 98, 99, 102, 122–124, 128, 137, 141–143, 147, 155, 170, 173, 183, 203, 205, 209, 213, 218, 220–222, 224, 228, 231–233,

245, 250, 288, 290, 294, 299, 303–306, 308, 313, 324, 326, 330 Lax 247 Laxenburg xvi Leader 8, 43, 74, 166, 183 Legitimacy 95, 123, 147, 153, 337 Leniency 224, 227 Lesotho 31 Leverage 222 l’Humanité 22 Literacy 95, 132, 199, 200, 325, 326 Literate 325 London 7 Long-haul 247 Luxury 31, 110, 294 M

Macao 236, 237 Madhya Pradesh 198, 291 Majority 7, 14, 23, 28, 31, 75, 86, 91, 111, 165, 166, 169, 181, 182, 184, 186, 187, 227, 292, 306, 331, 337 Malaria 28 Malawi 93, 237, 238, 256, 259 Malaysia xvi, 93, 136, 232, 235, 256 Mali 56, 133 Manila 30 Maputo 253–255 Marcos, F. 11, 14, 270 Marx 9, 312 Mayor 111, 301, 331 Measure 24, 39–41, 44, 46, 53–55, 57–59, 66, 81, 84, 89, 90, 96, 100, 102, 110, 130, 132, 134, 135, 139, 140, 142–144,

350     Index

151, 153, 169, 171, 187, 192, 194, 208, 214, 215, 251, 262, 272, 274, 287–289, 293, 295, 302–305, 313, 316, 318, 323 Measurement 4, 40, 46, 48, 260 Medical 5, 13, 18, 27, 151 Member of parliament (MP) 321 Meritocratic 241, 244–246, 258, 262 Mery, J.C. 15 Mexico xvi, 13, 33, 132, 135, 232, 243, 246, 256, 258, 259, 274, 294 Milan 31 Military 44, 122, 124, 151, 330 Mineral 108 Ming 2 Mississippi 213 Mobile 306–308, 325 Mobutu Sese Seko 11, 14 Montesinos, Vladimir 14, 21 Montesquieu 9 Morocco 28, 135 Moscow 25 Mozambique 253, 254 Multi-agency 235 Multifaceted 1, 2, 32, 47, 168, 232, 233, 337 Multilateral 41, 44 Multinational 13, 31 Multi-purpose 194, 232 N

National Election Studies (NES) 320 National Longitudinal Survey of Youth (NLSY) 322

Nazi 11 Neighbors 71, 87, 88, 115, 204, 317 Nepal 89, 252, 302 Nepotism 2, 32, 44, 328, 337 Netherlands 7, 136, 329 Network 22, 23, 30, 313 Newspaper 22, 23, 52, 80, 284, 288, 289, 292–295, 312, 320, 332 Nicaragua 30 Niger 26, 56, 57 Nigeria 20, 95, 112, 258, 289 Non-Formal Education (NFE) 247 Non-Prosecution Agreement (NPA) 267 Nordic 15, 94, 95 Norway 7, 93, 171, 288 Nurse 28, 248 O

Olympic 1, 2, 7 Oman 40–42 Online Procedures Enhancement (OPEN) 301, 302 Organization of American States (OAS) xv, 268, 269 Organization of Economic Cooperation and Development (OECD) xv, 6, 27, 63, 233, 245, 246, 265, 268, 269, 272, 273, 275, 302, 338 O-ring xiv Ostracize 339 Ouagadougou 247 Over-billing 28 Over-invoicing 293 Oxford 26, 93

Index     351 P

Pakistan 11, 12 Paraguay xvi, 55 Paris 15 Parliament 73, 100–102, 152, 184–186 Parliamentarism 75, 181, 187, 193 Parties xiv, xv, 6, 9, 10, 12, 15, 16, 18, 23, 28, 74, 77, 80, 120, 152, 167–169, 171, 172, 179–183, 186, 190, 204, 215, 224, 226, 227, 243, 268, 293, 321, 331, 332 Partisan 22, 23, 42, 115, 179, 181, 213, 216, 217, 219, 286, 331, 332 Perception 39, 41, 42, 46, 49–51, 53–57, 61, 67, 97, 114, 151, 152, 209, 258, 275, 290 Persistence xiv, 39, 54, 59–64, 66, 68, 72, 79, 115, 324 Peru xvi, 14, 21, 49, 93, 135, 147, 148, 258, 259 Pervasive 55, 59, 60, 65, 67, 79, 98 Petty 3, 9, 14, 30–32, 45, 47, 55, 57, 180, 227, 228, 247, 285, 328, 337 Pharmaceutical 28, 29 Philippine 262 Political and Economic Risk Consultancy (PERC) 236 Politics 5, 13, 14, 41, 44, 101, 258, 320 Polity 80, 90, 173, 175, 186, 320, 325 Polity 2 171, 290

Poll 20, 43, 151 Pollution 31 Poor 5, 13, 18, 23, 29, 30, 60, 103, 142, 145, 153, 154, 192, 211, 246, 248, 311, 317, 338 Port 247, 253, 254, 257 Poverty 12, 13, 30, 57, 102, 103, 123, 143, 145, 154, 313, 325 Pre-shipment inspection (PSI) 256–258, 260–262 Presidential 89, 90, 165, 183, 184, 186, 187, 193, 194, 319 Presidentialism 75, 181, 184, 187 Press 21–23, 87, 102, 184, 189, 194–196, 200, 273, 286–290, 292, 294–296, 302, 325, 339 Privatization 208, 241, 244, 258, 262 Procurement 9, 12, 13, 17, 25, 111, 125, 206, 207, 209, 251, 252 Propensity 42, 104, 115, 256, 275 Property 5, 19, 73, 120, 122, 259, 303, 316 Property-rights 19, 120, 303, 316 Prosecute 218, 237 Protestantism 56, 95, 289 Psychological 56, 112, 113 Public Expenditure Tracking Surveys (PETS) 39, 48, 49, 67, 294, 295 Q

Quantitative Service Delivery Surveys (QSDS) 39, 48, 49, 67

352     Index R

Radio 22, 23, 283, 284, 288, 291, 292, 294 Rajasthan 248 Rank 15, 50–52, 181, 290 Red Tape 138, 149, 197, 200, 205, 223, 253 Re-elect 179, 249, 250 Regulation xv, 10, 16, 53, 77, 84, 114, 119, 125, 138, 139, 149, 153, 176, 196, 197, 200, 203–209, 211 Religion 53, 71, 73, 76, 94–97, 105, 115, 152, 169, 172, 193 Religiosity 98–100, 172 Rent-seeking 14, 65, 72, 73, 167, 171, 182, 212, 257 Republican 23, 216 Resource-rich 72, 171, 176 Reuter 24 Right to Information Act (RTIA) 285, 290 Roads 58, 131, 253 Romanian 221, 251 Russia 15, 19, 22, 24, 25, 89, 93, 329 Rwanda 256, 259 S

Saeol 302 Salary 2, 12, 27, 32, 114, 247–249 Salvador xvi, 30, 88 Sand the wheels 154 Sao Tome and Principe (STP) 85, 86 Scandalmonger 180 Scandinavian 340 Secular 171, 172

Securities and Exchange Commission (SEC) 266, 273 Senegal 56 Sentencing 212, 220, 224 Seoul 7, 301 Seva Mandir 247, 248 Shanghai 225 Shenzhen 307 Shipment 254, 257 Siemens 270 Singapore xvi, 93, 103, 104, 132, 166, 227, 231–233, 235–238, 256, 288, 338, 340 Slovak 289 Slovenia 93, 198 Smuggling 11, 20, 110, 149 Sociability 105, 306 Social capital 76, 78, 305, 306, 312–316, 319–322, 324, 332 Socialist 87, 170, 183, 213 Soviet 17, 168, 289 Spain 18, 181 SPIDER 300 Spillover 77, 87, 97, 115, 204 Staffing 237 Sub-Saharan 85, 94, 237, 317 Suharto, Mohamed 11, 14, 171, 265 Sweden 7, 93, 136, 290 Switzerland 93, 270, 325 Syracuse 43 Systemic 64, 65, 68, 71, 72, 78, 95, 110, 111, 113, 115, 223, 243, 339 T

Taiwan 166, 221, 274

Index     353

Tanzania 28, 29, 49, 133, 232, 235, 237, 256, 259, 313 Tariff 123, 207, 254, 257, 260–262 Tax 16, 18, 41, 44, 53, 78, 84, 122, 123, 126, 132, 133, 138, 140, 194, 196–198, 205, 208, 209, 223, 243, 256–260, 267, 269 Telecommunication 131, 208, 303, 304, 308 Telephone 207 Television 21–23, 151, 283, 296 Terrorism 338 Testimony 15, 213 Thailand xvi, 18, 235, 236 Togo 56, 57 Transnational 276, 338 Transparency xvi, 2, 6, 11, 12, 20, 21, 26–31, 40, 42, 54, 62, 76, 77, 80, 83, 88, 142, 165, 167, 169, 173, 175, 176, 184, 195, 200, 208, 212, 213, 218, 231, 233, 236, 250, 266, 268, 271, 273, 283–285, 287, 288, 290, 291, 294, 296, 300, 301, 315 Trust 24, 71, 76, 82, 83, 99, 112, 114, 115, 123, 124, 143, 150–153, 155, 220, 221, 227, 243, 301, 305, 311–314, 316–319, 323, 324, 326, 330, 331, 337 Tsunami 2 Turkey 14, 168, 246 U

Udaipur 248

Uganda 30, 49, 111, 198, 235, 237, 238, 243, 256, 259, 284, 294 Ukraine 17, 18, 24, 83, 93 Under-the-table 28 Unions 2, 8, 32, 247, 315, 321, 337 United Nations Conference on Trade and Development (UNCTAD) 272 United Nations Convention Against Corruption (UNCAC) xv, 265, 268 University 10, 25, 27, 92, 93, 103, 197, 225, 227, 307, 323, 328, 329 Unsafe 8, 147, 148 V

Venezuela 135, 214, 258, 259, 270 Victimization 55 Vietnam 302 Voices xiii, 40, 54 W

Wage 81, 84, 215, 228, 241, 242, 245–247, 249–252, 262 Washington xvi, 272 Watchdog 21 Water 1, 2, 5, 13, 17, 29–31, 77, 131, 198, 199, 204, 291 Websites 42, 94, 300 Whistleblower 222 Women 30, 76, 77, 98, 100–103, 115, 168, 294, 326 World Bank (WB) xv, 6, 40–45, 56, 62, 80, 86, 90, 109, 133, 135, 137, 139, 141–144, 146, 147,

354     Index

170, 173, 175, 184, 185, 187, 190, 193, 195–197, 205, 208, 236, 246, 259, 268–270, 273, 287–289, 302, 303, 305, 327 World Competitiveness Report (WCR) 52, 61, 62, 108, 109 World Economic Forum (WEF) xvi, 40, 41, 43, 62, 64, 67, 134, 141, 218 World Enterprise Surveys (WES) 42, 45, 50, 51, 137 World Governance Indicators (WGI) 40–42, 44, 49, 50, 52, 56, 61, 67, 324 World Justice Project (WJP) 218, 330 World Trade Organization (WTO) 82, 272, 338 World Values Survey (WVS) 97, 98, 100, 306, 316, 318, 319 Wrestler 2

X

Xerox 7 Xiamen 227 Y

Yahoo 305 Z

Zaire 11, 110, 257 Zambia 49, 133, 237, 256, 259 Zealand 93 Zhuhai 307 Zimbabwe 93, 259