The Political Economy of Corruption [1 ed.] 9781003142300

Citation preview

The Political Economy of Corruption

Corruption, commonly defined as the misuse of public office for private gains, is multifaceted, multidimensional and ubiquitous. This edited collection, featuring contributions from leading scholars in the field of corruption, goes beyond the standard enforcement framework wherein individuals only compare the expected costs and benefits of a corrupt act. These chapters explore the political-​cultural contexts, legal and regulatory process and, above all, moral and psychological factors in attempts to understand and explain corruption. The book explores a broad canvas where gender, technology, culture and institutional structures influence attitudes towards corruption. Design and implementation of anti-​corruption strategies benefit from suitable identification of these factors contributing to the prevalence and persistence of corruption. Combining theoretical and empirical studies with evidence from experiments as well as case studies, the book provides crucial state of the art in corruption research in a highly accessible manner. This book serves as a vital reference to students and scholars in economics, politics and development studies. Additionally, policymakers and development practitioners can use the insights from this book in successful design and implementation of anti-​ corruption policies. Chandan Kumar Jha is Associate Professor of Finance at the Madden School of Business, Le Moyne College. He holds a Ph.D. and an M.S. in economics from Louisiana State University. Some of the topics he has worked on, or is currently working on, include the evolution of gender norms, corruption, finance and economic development, crime, entrepreneurship, international trade and socioeconomic inequality such as race and gender. Ajit Mishra is a development economist based in Bath, UK. After graduating from the Delhi School of Economics in 1993, he has been engaged in research and lecturing at various universities, including the University of Bath, University of Dundee, Delhi School of Economics, Indira Gandhi Institute of Development Research and Ashoka University. He has also served as Director of the Institute of Economic Growth, Delhi. His key research areas include the

study of corruption and governance, informal sector, inequality, vulnerability and poverty. Sudipta Sarangi is Professor and Head of Department of Economics at Virginia Tech. Prior to joining Virginia Tech, he has been a distinguished professor of business administration at Louisiana State University and a program director at the National Science Foundation. His research interests range from network theory, experimental and behavioral economics to development economics. He is a research associate of GATE, University of Lyon–​St. Etienne and the Lima School of Economics.

Routledge Frontiers of Political Economy

Modern Money and the Rise and Fall of Capitalist Finance The Institutionalization of Trusts, Personae, and Indebtedness Jongchul Kim Innovation, Complexity and Economic Evolution From Theory to Policy Pier Paolo Saviotti Economic Growth and Inequality The Economist’s Dilemma Laurent Dobuzinskis Wellbeing, Nature and Moral Values in Economics How Modern Economic Analysis Faces the Challenges Ahead Heinz Welsch Why Are Presidential Regimes Bad for the Economy? Understanding the Link between Forms of Government and Economic Outcomes Richard McManus and Gulcin Ozkan Critical Theory and Economics Philosophical Notes on Contemporary Inequality Robin Maialeh Corporate Financialization An Economic Sociology Perspective Marcelo José do Carmo, Mário Sacomano Neto and Julio Cesar Donadone The Political Economy of Corruption Edited by Chandan Kumar Jha, Ajit Mishra and Sudipta Sarangi For more information about this series, please visit: www.routle​dge.com/​ Routle​dge-​Fronti​ers-​of-​Politi​cal-​Econ​omy/​book-​ser​ies/​SE0​345

The Political Economy of Corruption Edited by Chandan Kumar Jha, Ajit Mishra and Sudipta Sarangi

First published 2023 by Routledge 4 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 605 Third Avenue, New York, NY 10158 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2023 selection and editorial matter, Chandan Kumar Jha, Ajit Mishra and Sudipta Sarangi individual chapters, the contributors The right of Chandan Kumar Jha, Ajit Mishra and Sudipta Sarangi to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-​in-​Publication Data A catalogue record for this book is available from the British Library ISBN: 978-​0-​367-​69563-​7 (hbk) ISBN: 978-​0-​367-​69564-​4 (pbk) ISBN: 978-​1-​003-​14230-​0 (ebk) DOI: 10.4324/​9781003142300 Typeset in Bembo by Newgen Publishing UK

For our children Anika, Aditya and Amrita

Contents

Notes on Contributors 1 The Political Economy of Corruption: Some New Perspectives

xi

1

C H A N DAN K U M AR JHA, AJI T MI SHRA AND SUDIP TA SA R A N GI

2 The Political Economy of Corruption: On the Link between Corruption Control and Cronyism

6

K AU S H I K BA SU

3 Corruption, Institutional Trust and Legitimacy: A Vicious Circle

15

A M AD O U B O LY AND RO B E RT GI LLAND E RS

4 Legal Systems and Corruption

31

RU B E N KO R STE N AND AND RE W SAMUE L

5 Corruption and Optimal Enforcement

58

A JI T M I S H R A

6 A Theory of Joint Evolution of Corruption and Growth

71

N I LOY B O SE , R I CHARD COTHRE N AND NAZAN IN SEDAGHAT KISH

7 Corruption and the Financial Sector: An Examination of the Literature

87

A RU SH A C O O R AY

8 Stopping the Rot I: A Review of Models and Experimental Methods of Corruption Experiments R I TW I K BA N E R JE E , UTTE E YO DASGUPTA AND SATA RUPA M IT R A

100

x Contents

9 Stopping the Rot II: Consequences, Causes and Policy Lessons from the Recent Experiments on Corruption

115

RI TW I K BA NE RJE E , UTTE E YO DASGUPTA AND SATA RU PA M IT R A

10 The Past, Present and Future of Research on Gender and Corruption

139

JU STI N E SARE Y AND N. VALD E S

11 The Culture-​Corruption Hypothesis Revisited: Organizational Culture, Corruption and Worker Preferences

153

SH E H E RYA R BANURI

12 How Advances in Information and Communication Technologies Impact Corruption?

172

C H AN DAN K UMAR JHA AND SUD I PTA SARANGI

13 Corruption in Europe: The Underestimated Devil and the Role of the European Union

189

I N A K U B B E AND STOYAN PANOV

14 Tackling Corruption: Practical Perspectives

216

C H AN DAN K UMAR JHA AND NE E LE SH K UMAR SA H

Index

233

Contributors

Ritwik Banerjee is Associate Professor of the Economics area at Indian Institute of Management Bangalore, India. His primary research area is at the intersection between behavioral and development economics, and he extensively uses experimental methods for his research. Some of the topics he has investigated or is currently investigating are corruption, inequality, education and discrimination. Sheheryar Banuri is a behavioral economist and an expert on motivation and incentives, behavior and public policy. He is currently Associate Professor in the School of Economics at the University of East Anglia. His policy work has provided guidance to the governments of Indonesia, the Philippines and Burkina Faso. His work has been published in leading academic journals in multiple fields. He has authored a number of policy briefs and is co-​author of the World Bank’s World Development Report 2015: Mind, Society, and Behavior and the author of Good Decisions for Strange Situations and The Decisive Mind. Kaushik Basu is Professor of Economics and Carl Marks Professor of International Studies at Cornell University. From 2012 to 2016 he was Chief Economist of the World Bank, and from 2017 to 2021 he served as President of the International Economic Association. In 2021, he was awarded the Humboldt Research Prize. Amadou Boly is Special Assistant to the Chief Economist and Vice-​President at the African Development Bank. Before joining the bank, he worked at UNU-​WIDER in Finland and UNIDO in Austria. Boly holds a Ph.D. in economics from the University of Montreal (Canada), an M.Sc. in international business from the University of Groningen (Netherlands) and a maîtrise in economics and management from the University of Ouagadougou (Burkina Faso). Niloy Bose is Professor of Economics at Virginia Tech. He previously held faculty positions at the University of Wisconsin, Milwaukee, and the University of Manchester, England. He also served as the Economics Program director at the US National Science Foundation. His main research interests are macroeconomics and monetary economics, economic development and

xii  Notes on Contributors growth theory, political economy and the theories of private information. His publications include articles in academic journals and chapters in edited volumes, many of which address issues related to corruption and property rights and their effects on the misallocation of resources. Arusha Cooray is Professor of Finance at the College of Business, Law and Governance, James Cook University, and Head of the Accounting and Finance Academic Group. She completed her Ph.D. at the University of New South Wales, Australia. Arusha holds followships at the United Nations University (UNU-​WIDER), the Centre for Poverty Analysis (Sri Lanka), the Centre for Applied Macroeconomic Analysis (ANU) and the Centre for Development Economics and Sustainability (Monash University). Richard Cothren is a long-​ time faculty member of the Department of Economics at Virginia Tech. His research interests include economic growth and monetary policy. Utteeyo Dasgupta is Associate Professor of Economics at Fordham University. He is a behavioral economist broadly interested in the positive, normative and strategic aspects of decision-​making. He primarily uses experimental methods in his research. His research work has been published in the Review of Economics and Statistics, the Journal of Public Economics and other peer-​ reviewed journals. He is affiliated with the Institute for the Study of Labor (IZA) as a research fellow and with the Global Labor Organization as a fellow. Justin Esarey is Associate Professor of Politics and International Affairs at Wake Forest University. He received his Ph.D. in political science from Florida State University in 2008 following a BA in political science and a BS in economics from Bowling Green State University in 2002. His area of specialization is political methodology, especially hypothesis testing and the scientific ecosystem. His current substantive projects study the relationship between corruption and female participation in government. Esarey is currently co-​ editor of PS: Political Science and Politics and the principal investigator of the International Methods Colloquium project. Robert Gillanders is Associate Professor of Economics at Dublin City University (DCU) and co-​ founder and co-​ director of the DCU Anti-​ corruption Research Centre (DCU-​ARC). Robert has published extensively on the causes and consequences of corruption and also on issues relating to enterprise and the business environment. He is co-​principal investigator of the Irish Research Council–​funded project “Corruption, Gender, and Sustainable Development” (2021–​24) and serves on the Irish government’s advisory council against economic crime and corruption. From 2019 to 2022, he served as Ireland’s local research correspondent on corruption as part of an EU-​wide network.

Notes on Contributors  xiii Chandan Kumar Jha is Associate Professor of Finance at the Madden School of Business, Le Moyne College. He holds a Ph.D. and an M.S. in economics from Louisiana State University. Some of the topics he has worked on or is currently working on include the evolution of gender norms, corruption, finance and economic development, crime, entrepreneurship, international trade and socioeconomic inequality, such as race and gender. Ruben Korsten is a Ph.D. researcher and lecturer at the Center of Advanced Studies in Law and Economics at Ghent University, Belgium. He graduated law “cum honore” and obtained his master degree in private law at the University of Amsterdam (the Netherlands). Ruben Korsten also holds master degrees in law and economics from the Indira Gandhi Institute of Development Research, Mumbai (India), Hamburg University (Germany) and the University of Bologna (Italy). He was a visiting scholar at the Pontifical Xavierian University (Bogota, Colombia) and Loyola University (Baltimore, MD). His research focuses on general law and economics, anti-​ trust and self-​reporting. Ina Kubbe is a post-​ doc researcher at the School of Political Science, Government and International Relations in the University of Tel Aviv, where she mainly researches and teaches on corruption, migration, gender politics and conflict resolution. Ina is also Professor at the International Anti-​ Corruption Academy (IACA) in Austria. She specializes in social science methodology and comparative research on empirical democracy, corruption and governance research, with a special focus on Europe and the Middle East. Ina has published several books, special issues and articles in the field and is also one of the founding members of the Interdisciplinary Corruption Research Network (ICRN) as well as Chair of the ECPR Standing Group on “(Anti)Corruption and Integrity.” Ajit Mishra is a development economist based in Bath, UK. After graduating from the Delhi School of Economics in 1993, he has been engaged in research and lecturing at various universities, including the University of Bath, University of Dundee, Delhi School of Economics, Indira Gandhi Institute of Development Research and Ashoka University. He has also served as Director of the Institute of Economic Growth, Delhi. His key research areas include the study of corruption and governance, informal sector, inequality, vulnerability and poverty. Satarupa Mitra is a doctoral student in economics and social sciences at the Indian Institute of Management Bangalore, India. Primarily, her interest lies in behavioral development economics. Her life experience in India has deeply inspired her to investigate problems related to various development issues. Her current research focuses on the experimental inquiry of risk preferences, exponential growth bias, corruption and mental health. As an academic member, she aspires to pursue research at the intersection between

xiv  Notes on Contributors development and behavioral economics and contribute to teaching basic and advanced field courses in economics. Stoyan Panov is Lecturer of international law and EU law at University College Freiburg. He has published in the areas of judicial anticorruption reforms, EU’s response to challenges to the rule of law and corruption, populism and the rule of law, human rights protections in preventive seizure and confiscation of assets and property, among other topics. He was a fellow of the 2019–​20 re:constitution “Exchange and Analysis on Democracy and the Rule of Law in Europe” project on EU’s multilayered response to the backsliding of democracy and the rule of law in Central and Eastern Europe. Neelesh Kumar Sah is an Indian bureaucrat from the Indian Audit and Accounts Service. He is currently Joint Secretary of Climate Change in the Ministry of Environment, Forest and Climate Change, Government of India. He has wide experience in evaluating systems and conduct of audits. He served as Technical Specialist (Fraud and Corruption Detection) with UNDP during 2009–​12. He has been involved in setting up Information Technology Audit and Data Analytics capacities in the Supreme Audit Institution of India apart from setting up various IT systems. He has been conferred with the highest civil service award in India –​the Prime Minister’s Award for Excellence in Public Administration for IT Audit Initiative (2006–​07). Andrew Samuel is Professor of Economics at Loyola University Maryland. His research lies at the nexus of law and economics, regulation and industrial organization and is centered around two thrusts: (1) Do market forces and regulatory enforcement reinforce or crowd each other out? (2) What is the optimal type of monitoring or enforcement of regulations within a given market environment? He uses game theory, micro theory and some empirical methods to answer these questions. Sudipta Sarangi is Professor and Head of Economics Department at Virginia Tech. Prior to joining Virginia Tech, he had been a distinguished professor of business administration at Louisiana State University and a program director at the National Science Foundation. His research interests range from network theory, experimental and behavioral economics to development economics. He is a research associate of GATE, University of Lyon–​St. Etienne and the Lima School of Economics. Nazanin Sedaghatkish is an advanced Ph.D. student in the Department of Economics at Virginia Tech. Before joining Virginia Tech, she received a master degree in economics from Sharif University of Technology in Iran. Her main research interests include applied microeconomics, causal inference and machine learning. Her research has also explored theoretical linkages between the quality of property rights, efficiency of financial markets and the level of economic prosperity.

newgenprepdf

Notes on Contributors  xv N. Valdes is an analyst working for Deloitte Consulting for their government public services practice. She received a BA in politics and international affairs from Wake Forest University in 2020. Her written work largely focuses on gender, corruption and gender quotas.

1 The Political Economy of Corruption Some New Perspectives Chandan Kumar Jha, Ajit Mishra and Sudipta Sarangi

The study of corruption in economics has been dominated by the Beckerian crime enforcement model where financial incentives determine the extent of illegal/​corrupt behavior by agents.Whether it is bureaucratic corruption where officers take bribes to grant favors, or political corruption where elected leaders engage in various forms of corruption, agents always consider the expected costs and benefits before engaging in the corrupt activities. Several decades of anticorruption strategies and their lack of adequate success suggest that the incentive parameters are only a part of the whole story.This collection of essays broadens the canvas and offers different insights into the study of corruption. Corruption, commonly defined as the misuse of public office for private gains, is multifaceted, multidimensional and ubiquitous in several countries. Hence a proper understanding of corruption requires us to look at the political-​ cultural contexts, legal and regulatory process, the nature of technology and governance and, above all, moral and psychological factors impacting behavior. As we outline below, the essays in the volume share this broad objective.The first section of essays (Chapters 2–​5) consists of theoretical models that go beyond the Beckerian perspective on corruption. The next two chapters (Chapters 6 and 7) deal with macro-​financial aspects of corruption, both theoretically and empirically. The third section of the book (Chapters 8–​ 12) consists of five essays that dig into new areas of corruption research. They explore gender and corruption, how behavioral and cultural issues impact corruption and how technological developments that have occurred in the last two decades can tackle corruption. To complete this collection, we touch upon some themes that do not get adequate attention in the literature—​corruption in the developed world (specifically Europe) and a practitioner’s perspective in Chapters 13 and 14. In many cases, it is not just corruption that is the problem, but the entire anticorruption apparatus can be the source of the problem as well. The persistence and pervasiveness of corruption are partly due to the fact that anticorruption efforts often lead to corruption changing its form. Kaushik Basu (Chapter 2) explains how corruption can turn into cronyism when corruption control can be influenced by the political leadership in power. Using a simple model, he shows how a political leader who comes to power with a promise to DOI: 10.4324/9781003142300-1

2  Chandan Kumar Jha, Ajit Mishra and Sudipta Sarangi curb corruption by increasing the crackdown can target only the opposition and leave their own supporters untouched. Hence, despite the seemingly more intense anticorruption drive, total corruption may not go down. The importance of an autonomous body to lead the anticorruption drive is crucial in stopping corruption control from becoming an instrument of oppression. Amadou Boly and Robert Gillanders (Chapter 3) discuss how corruption persists and can even be stubborn despite significant anticorruption efforts and offers an interesting insight. On the one hand, anticorruption instruments and policy directions can be controlled by predatory actors who benefit from corruption. On the other hand, corruption-​induced loss of trust in institutions leads to greater reliance on corruption by the citizens. They draw attention to the role of “legitimacy,” which is a property of a law that enhances compliance by citizens even when the probability of sanction is very low. For example, scholars often point toward tax compliance by individuals when the detection probability of evasion is extremely low. It can be because of social and moral norms, but in their model, it can be argued that compliance is high because tax policy and its administration are viewed as legitimate. In the corruption context, with low levels of trust in institutions and anticorruption efforts, legitimacy gets eroded leading to noncompliance. Hence, societies find themselves in a low-​ legitimacy high-​corruption trap. The next two chapters examine the legal and regulatory system and its relationship with corruption. Ruben Korsten and Andrew Samuel (Chapter 4) examine how two dominant legal systems—​the civil law system and the common law system—​affect corruption in different ways. While it has been suggested that corruption may be less prevalent in common law countries, this chapter offers a formal model of the legal system and bribery to analyze this causal link. Interestingly, the authors do find that common law countries are more likely to exhibit corruption, but the relationship is not straightforward. It depends on the nature and strength of the bureaucracy, its extractive power and the corruptibility of the judges. Compared to civil law, we are likely to see greater judicial involvement in common law countries, and corruption in the judiciary will remain an additional consideration. Ajit Mishra (Chapter 5) considers the nature of regulatory policies when different forms of corruption, namely collusion and extortion, are present. The chapter uses a simple model of regulatory enforcement to see how optimal regulatory policies are affected. Collusion, where the citizen or firm and the enforcement officer collude, leads to a dilution of enforcement and compliance incentives. Extortion also has a similar dilution of compliance incentives, as those needing to comply can be subject to possible harassment and extortion. He argues that, in some situations, the regulator may optimally choose to overenforce when faced with corrupt officers. Overenforcement refers to a case of preventing an agent from choosing an action even when its private benefits from the action exceed social costs. This can be viewed as a form of “stricter enforcement” compared to what is considered efficient in the absence of corruption.

The Political Economy of Corruption  3 The persistence of corruption issue, considered in earlier chapters, can also be viewed in terms of the two-​way relationship between corruption and development. The growth-​retarding effects of corruption have been studied extensively, but the reverse impact of lower growth on corruption is also important. Nilay Bose, Richard Cothren and Nazanin Sedaghatkish (Chapter 6) offer a simple model to explain why corruption is persistent at lower levels of development and why widespread corruption leads to underdevelopment. Building on some previous work, this proposed model explores various aspects of the corruption–​development relationship. A key channel through which corruption impacts growth is financial sector development. Arusha Cooray (Chapter 7) reviews the literature on corruption and financial sector development, finding overwhelming support for the view that corruption damages the development of the financial sector. She highlights the role of improved governance and rule of law in raising efficiency. It is no surprise that strong institutions reduce these negative impacts of corruption. But then (a fact already highlighted by other chapters in this volume), institutions themselves are undermined by pervasive corruption in the system. The next five chapters in Section 3 focus on how corruption is shaped by various behavioral, cultural and technological factors. Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra (Chapters 8 and 9) focus on the behavioral experimental literature and provide an in-​depth review of this relatively recent literature. The authors discuss the pros and cons of using experiments in corruption literature. The experimental literature has shown that many assumptions underlying the classical theories of corruption might not hold and hence the anticorruption policies based on these models might not work. The authors discuss various experimental studies that offer guidance on designing anticorruption policies.The experiments allow researchers to explore questions that cannot be answered using the traditional empirical literature by allowing them to manipulate the experimental environment. The authors review a number of recent studies that offer experimental evidence on the causes and consequences of corruption. It is however important to be cautious in generalizing the results from the lab experiments. The external validity of experimental studies might be questionable as culture—​a crucial determinant of corruption—​has important bearings on corruption, and cultures vary remarkably across societies. Authors conclude that laboratory experiments indeed have the potential to make further contributions to the corruption literature. Justin Esarey and N. Valdes (Chapter 10) focus on the relationship between gender and corruption. The existing research documents a negative correlation between gender representation in various economic and political positions and corruption. Numerous mechanisms like gender differences in risk aversion, opportunistic behavior, access to corruption networks, the know-​ how of participating in corrupt behaviors, preferences for different policies and consequences of engaging in corrupt practices have been used to explain this finding. In this chapter, the authors draw our attention to the fact that despite the sizable literature on gender and corruption, our understanding of causal

4  Chandan Kumar Jha, Ajit Mishra and Sudipta Sarangi mechanisms underlying this association is very limited. They argue that multiple causal mechanisms sometimes work in the same direction making a strong relationship between gender and corruption, while in opposite directions at other times, weakening the association. Sheheryar Banuri (Chapter 11) reviews the literature on organizational culture and corruption with a focus on the experimental literature. An established literature documents that prosocial tendencies and moral costs play crucial roles in determining corrupt practices of workers.The workers opting into the public sector have been shown to have distinct preferences, and monetary incentives can change the types of workers attracted to the public sector. The question then is: Can corruption among public employees be reduced by raising wages? He notes that more research is needed to answer this and many other pertinent questions. For instance, can social intervention be used to change corrupt behavior among public officials? While a great deal is known about corruption in the public sector, we know very little about the culture of corruption in the wider population. So, this chapter tries to answer whether the corrupt practices by public officials in controlled environments can be extrapolated to the wider population. Advances in information and communication technologies (ICTs) have become an important anticorruption force in recent years. Chandan Kumar Jha and Sudipta Sarangi (Chapter 12) provide a review of literature on ICTs and corruption. Their review provides strong support in favor of the corruption-​ reducing effects of recent advances in ICTs. Various types of ICTs such as mobile phones, the internet and social media and their impact on corruption are discussed separately in this chapter. Moreover, advances in ICTs allow for the provision of various public services electronically—​a process known as e-​ government. The authors discuss how ICTs and the e-​government promote transparency in the provision of government services, enhance the accountability of government officials and eliminate the need for interaction between government officials and citizens looking for a government service.The authors caution against the increased clampdown on internet freedom and spying on citizens by governments around the world, including countries considered to be the champions of democracy like the United States and the United Kingdom. They emphasize the crucial role that the civil society has to play to ensure that ICTs remain an effective tool against corruption. In Chapter 13 (Section 4), Ina Kubbe and Stoyan Panov argue that corruption in Europe is underestimated and considered to be homogenous. This perception results in the lack of context-​specific approaches to deal with corruption. They find that both corruption perception and anticorruption norms in the region have been worsening in recent years. They discuss many anticorruption initiatives taken by the European Union (EU). While many of these developments are well placed, a number of key areas need additional attention.The authors also note that the implementation of the EU legal framework remains a concern in many member states.

The Political Economy of Corruption  5 While there has been a lot of academic research on topics related to corruption, the last chapter (Chapter 14) explores corruption from the perspective of a government attempting to deal with the practical issues involved in coming up with procedures for reducing corruption. Chandan Kumar Jha and Neelesh Kumar Sah describe the processes involved in the provision of government services and how these processes give rise to opportunities for corruption. To lower corruption in the provision of government services, it is imperative that these processes are simplified and streamlined to promote transparency, reduce discretion and increase accountability of the officials involved at each stage of the process. However, many challenges prop up in doing so, including the resistance from within the system. The role of technology in fighting corruption is highlighted again as the authors discuss a case study from India to show how the use of technology has led to a decline in corruption in tax filings. The authors specifically note that while implementing changes to procedures involves the provision of public services to reduce the system’s vulnerability to corruption, it is crucial to seek feedback from government officials. These officials possess the best knowledge of what loopholes might arise in the updated system and how they can be abused for rent extraction. Our objective in this collection of essays has been to advance the anticorruption debate by incorporating recent developments and insights. We study how incentives matter for all parties involved in corruption, including the parties that are responsible for ensuring that anticorruption measures work. We also explore various cultural, behavioral and technological factors that determine the prevalence and persistence of corruption. The book can be used by a wide range of scholars interested in understanding new developments in the corruption literature—​from advanced undergraduate students, graduate students to researchers in the disciplines like economics, sociology and political science. We also believe that the book will serve as a guide to policymakers and practitioners and help to keep the conversation on anticorruption research going. Happy reading!

2 The Political Economy of Corruption On the Link between Corruption Control and Cronyism Kaushik Basu

Corruption Control: Promise and Practice History has many examples of individuals with mal-​intent, who through persuasion, politics and guile manage to come to power and then do what they wanted to do—​loot and plunder, allowing corruption to flourish. Interestingly, history is also replete with examples of leaders who come to power, concerned about the level of corruption in society and with the aim of controlling or even eradicating corruption but end up not just doing little but exacerbating the problem, with the nation getting mired in even more corruption and cronyism. There is no way to be sure of the original intention of a leader. No one knows what led Vladimir Putin to seek power, but to Russians, Putin, in his early days, with his reputation for professionalism, from the time he was adviser to the democratically elected mayor of St. Petersburg, brought hope that he would end the erratic rule of Boris Yeltsin and the rampant corruption, such as the one in cahoots with the Swiss construction company, Mabetex, in which Yeltsin’s family members were engaged in. Likewise, Daniel Ortega brought hope to Nicaraguans by his grit and determination to fight and overthrow the openly corrupt government run by Anastasio Somoza Debayle.The glimmer of hope however got blighted, in both these cases, as corruption became rampant, as did cronyism. Indonesia’s Suharto came to power dislodging the old order, supported by students, and giving hope for a just and civilized society and “social justice for all.” His thirty-​one-​year rule ended with Indonesia ranked as the most corrupt nation in the world by Transparency International. In all these cases, the nation is usually left with widespread corruption, disproportionate amount of business and work captured by a few cronies of the leader, and the nation mired in odious debt. This is such a widespread phenomenon, of a leader coming to power full of good intention to end corruption and then ending up propagating more of the same, that we need to understand why this happens. As it turns out, there is good reason why this happens, and the aim of this chapter is to explain this. DOI: 10.4324/9781003142300-2

The Political Economy of Corruption  7 Here are the basic premises of my argument. First, note that when corruption is widespread it is easy to get away being corrupt since there is only that much resource available with the government to detect corruption and then arrest the corrupt. Secondly, in countries with a lot of complex laws and rules, it is natural to see widespread violation of the law, including indulgence in corrupt activities. Some of this happens because people do not even know that they are violating the law, since so many of the laws are dormant, gathering dust, with neither citizen nor the enforcer of the law even being aware of them. The ubiquity of corruption gives rise to unexpected political power in the hands of the oligarch at the top or the political leader, which the leader may not be fully aware of before coming to the seat of power. Once they become leader, they soon realize they can basically have anyone arrested for violating the law, since there is such an embarrassment of riches to choose from, since virtually everybody violates some law. It was not hard for Putin’s government to find reasons to incarcerate Alexei Navalny, though the real reason for the arrest was transparently obvious to all observers. Next note that, in this setting, it will be in the interest of the political leader to direct the limited capacity to stop violation of the law and corruption to those who oppose the government or criticize the leader. To do otherwise and arrest one’s friends and supporters would be to erode one’s own power base. On the other hand, to arrest the corrupt but confine this to those who oppose the government will show determination to curb corruption (you are arresting the number of corrupt people you had promised to arrest at the time of your election) and, at the same time, shore up your political strength, having more and more of the opposition behind bars. Hence, corruption control and the enforcement of the law, even when it begins with a serious intention, can easily end up as an instrument for silencing the opposition, critical media and voices of dissent in society. Conversely, this provides comfort and shelter to the leader’s friends and supporters. In short, corruption continues to flourish, and worse, may morph into cronyism since those who are close to the leader feel safe and can violate the law more egregiously. An interesting by-​product of this analysis is that even when leaders carry out their pre-​election promise of incarcerating more corrupt people than may have been happening till then, the total amount of corruption in the nation may rise. This counterintuitive result needs a little formal analysis to demonstrate. This is done in the next section. What I construct in the chapter is a theoretical model, but my aim is to contribute ideas for the design of policy to control corruption. I have argued elsewhere (Basu, 2016, ­chapter 8) that to control corruption we need determination and courage, but we also need analysis and theory.We have commonly neglected the latter and for that reason have failed widely in our effort to control corruption. I do not actually go into policy design in this chapter, but the chapter may be viewed as an effort to contribute ingredients for the design and control of corruption.

8  Kaushik Basu The argument in my chapter suggests the need for an autonomous body to control corruption, as has been tried in some countries such as Indonesia. But, of course, there are risks with autonomous bodies that can become loose cannons and go out of control. One of the reasons for constructing a formal model is that it can give us insights into starting up a discussion and analysis for designing more effective policy.

How Corruption Control May Exacerbate Corruption: A Model Consider a society with a set N = {1, 2,…, n} of individuals. I am making the standard neoclassical assumption that all individuals are potentially corrupt, that is, they will choose to indulge in corruption if the net return from this is positive. I want to, however, put on record my belief that in reality there are people—probably lots of them—who are innately not corrupt, that is, they reject corrupt behavior (or at least certain kinds of corrupt behavior) simply because their moral compass does not allow this, and not because by doing cost-benefit analysis they conclude that corruption does not pay.1 This alerts us to the fact that nurturing a moral compass can play an important in controlling corruption, though this is a topic neglected by mainstream economics. This is not a theme pursued in the present paper but should not be forgotten when devising policy. With this caveat in mind, let me focus on the set N of ‘corruptible’ people. Since we are focusing on the set N, we can use Becker’s (1968) logic in explaining human behavior. Let B(i) denote the benefit that person i ∈ N gets from an act of corruption. I shall follow the convention that if i > j, then B(i) < B(j). Excepting for the fact that I am ruling out tie-​breakers (purely for simplicity of analysis), this is not an assumption but simply a convention for naming individuals. The person who gets the highest benefit from the corrupt act is named person 1, the person who gets the least benefit from the corrupt act is named person n, and the names occur monotonically for those in between. A possible B(i) function is illustrated in Figure 2.1. Let us suppose that currently there is a limit to the government’s vigilance capacity and it can investigate at most m (m) people, but she will do this by an unusual method. This will unlikely be made by a public order, but it will be made clear through nods and winks. She will not investigating any of her supporters and investigating only the opponents. Clearly then each of the opponents faces a 2m ’ probability of being investigated for corruption. n It is obvious from Figure 2.2 that corruption among the opponents will fall to t 2* and the corruption among the cronies will rise to t1* . Depending on the curvature of the b(i) function, it is entirely possible that t1* + t 2* >

t t + = t (3) 2 2

In brief, going for greater vigilance on corruption, but, at the same time, directing this entirely at those who oppose you, while, at the same time, protecting your friends and supporters, can make for a heady brew, giving you crony capitalism and overall higher corruption all at the same time. Nations around the world have stumbled into this. The higher corruption and crony capitalism can be the outcome of a good leader, who comes to office with genuine interest in curbing corruption and drifts into this cronyism-​ corruption trap.

12  Kaushik Basu

Corruption and Policy What does one do to control corruption? This is a big topic; this has concerned rich and poor countries, even though the nature and extent of corruption varies widely across nations. In a lot of emerging economies, corruption affects the everyday life of ordinary people in ways that one does not see in most advanced economies, where corruption takes the form of fewer but larger scams involving arms and other deals. Among activists the end of corruption is often treated as purely a matter of grit and determination. If you are serious about rooting out corruption, you can. In reality, taking corruption out of the economy is difficult and an intricate matter of science, analysis and design. As the model above shows, there are situations where catching and punishing more corrupt people can exacerbate rather than limit corruption. Further, what I do not go into here but is important to note is the problem of collateral damage (Basu, 2016). Removing corruption from society is akin to the problem of taking a tumor out from the human body. One has to do this without damaging healthy tissues around the tumor, or else the cure can be worse than the disease. Likewise, corruption has to be taken out of the economy without damaging legitimate activities. Many nations have suffered by using blunt instruments like demonetizing the nation’s currency, which can do a disproportionate amount of collateral damage to the economy. Corruption is a big challenge for policymakers and researchers working in the area of law and economics. Not surprisingly, a lot has been written on this, from providing early conceptual foundations to analyze the problem of corruption and its control (Rose-​Akerman, 1975; Bardhan 1997) to taking on contemporary concerns, using tools of game theory and controlled laboratory tests, which help us understand the challenges arising from increasing globalization and its attendant problems involving cross-​ country conflict and intercountry finance, and from the rise of digital technology and social media.2 One important regulatory idea that emerges from this chapter in the context of the problem illustrated in the previous section is the need for an autonomous body that is outside the control of politicians, like a nation’s judiciary or the central bank in many countries, and can monitor, investigate and contain corruption. For a good leader, who comes to power with the genuine interest of weeding out corruption, the wise decision is to create such an autonomous body, empowered to control corruption and punish the guilty with the help of the nation’s judiciary and beyond the reach of political leaders, including the good leader himself or herself. This is because, as the model above shows, if politicians have the power to decide who to investigate, they will not arrest their friends, and as a result, the shelter provided to friends and cronies can make the problem worse. So it is in the interest of the new leader keen to control corruption to take the power to control corruption out of his or her own hand and vest it in an autonomous authority.

The Political Economy of Corruption  13 A good example of what this can do comes from Indonesia. In 2002, Indonesia tried to implement the kind of policy I am suggesting here. Indonesia, bruised by a history of devastating corruption, set up a new entity to control corruption. The Corruption Eradication Commission, which has the acronym KPK, is an empowered, autonomous body with the responsibility of investigating and controlling corruption. After Suharto fell from power in 1998, having pilfered 35 billion dollars and leaving the nation ravaged by corruption, the country faced a challenge like few other nations had faced (Blank, 2019). It is arguable that the creation of KPK played a major role in cleansing quite a bit of this legacy. From being the world’s most corrupt nation in 1995, it became 89th out of 180 nations ranked by Transparency International. As Blank (2019) argues, KPK played a major role in this improvement. It is arguable that this would not have been possible if KPK was kept under the control of the government. In that case, no matter what the original intention of the government’s leaders, this would soon become an instrument for silencing opposition and lead to more cronyism and possibly even more corruption. Let me close the chapter by returning to what was pointed out earlier, namely, the need to step beyond the model discussed here and pointing out what is anathema for mainstream economics, namely the need to educate individuals to carry their own moral compass in their heads and to desist from being corrupt. I began this chapter by recognizing that societies do have people who are not corrupt. These are people who have it in them hardwired not to indulge in certain kinds of activities. While mainstream economics does not recognize this, because of the axiom that individuals are uncompromisingly selfish, it is arguable that in reality the success of human society depends in large measure on us carrying certain kinds of norms in our head. What Ken Arrow pointed out in his 1978 essay is right. If we did not have a basic commitment to fairness and trust, the “world of total self-​interest would not survive for ten minutes.” Just as the invisible hand of the market relies on these implicit assumptions about cultural norms, ignored by economists but critically important for the market to function effectively, the control of corruption does rely a lot on the norms we individuals carry in our heads. The time has come to study, understand and nurture the appropriate norms.

Acknowledgment I am grateful to Ajit Mishra for discussion on this topic and related ideas over many years, and to Sudipta Sarangi for helpful comments and suggestions on this chapter. I would also like to thank Hannah Kim and Haokun Sun for valuable research assistance.

Notes 1 The clear contrast is with Becker (1968), who assumes all human beings are rational in the sense of being amoral. In Basu (1983), I had argued that not only are many

14  Kaushik Basu human beings programed not to be corrupt or at least not to be corrupt in certain ways but that economic life is made possible by this kind of hardwiring in human beings. Subsequently, it has been shown that there are ways of getting around the argument I had produced in that paper by demonstrating that pure rationality and self-​interest can lead to similar behavior (Myerson, 2004; Larry Samuelson and Stacchetti, 2017), for instance, by allowing for the war of attrition to break out when people violate certain norms. While admitting that this is valid, I remain convinced that human beings do have moral hardwiring in their heads. 2 See, for instance, Basu, Bhattacharya and Mishra (1992), Mishra (2006), Choi and Posner (2007), Luddington, Gulati and Brophy (2010), Abbink et al. (2014), Oak (2015), Jha and Sarangi (2017).

References Abbink, K., Dasgupta, U., Gangadharan, L. and Jain, T. (2014), “Letting the Briber Go Free: An Experiment on Mitigating Harassment Bribes,” Journal of Public Economics, vol. 111. 17–​28. Arrow, K. (1978), “A Cautious Case for Socialism,” Dissent, Fall. www.diss​entm​agaz​ine. org/​arti​cle/​a-​cauti​ous-​case-​for-​social​ism Bardhan, P. (1997), “Corruption and Development: A Review of Issues,” Journal of Economic Literature, vol. 35. 1320–​1346. Basu, K. (1983), “On Why We Do Not Try to Walk Off without Paying after a Taxi Ride,” Economic and Political Weekly, vol. 18. 2011–​2012. Basu, K. (2016), An Economist in the Real World:The Art of Policymaking in India, Cambridge, MA: MIT Press. Basu, K., Bhattacharya, S. and Mishra, A. (1992), “Notes on Bribery and the Control of Corruption,” Journal of Public Economics, vol. 48. 349–​359. Becker, G. (1968), “Crime and Punishment: An Economic Approach,” Journal of Political Economy, vol. 76. 169–​217. Blank, J. (2019), “How the (Once) Most Corrupt Country in the World Got Clean(er),” The Atlantic, May 2. Choi, A. and Posner, E. (2007), “A Critique of the Odious Debt Doctrine,” Law and Contemporary Problems, vol. 70. 33–​51. Jha, C. and Sarangi, S. (2017), “Does Social Media Reduce Corruption,” Information Economics and Policy, vol. 39. 60–​71. Ludington, S., Gulati, M. and Brophy, A. (2010), “Applied Legal History: Demystifying the Doctrine of Odious Debt,” Theoretical Inquiries in Law, vol. 11. 247–​281. Mishra, A. (2006), “Corruption, Hierarchies and Bureaucratic Structures,” in S. Rose-​Akerman (ed.), International Handbook on the Economics of Corruption, Cheltenham: Edward Elgar, 189–​215. Myerson, R. (2004), “Justice, Institutions, and Multiple Equilibria,” Chicago Journal of International Law, vol. 5. 91–​107. Oak, M. (2015), “Legalization of Bribe-​Giving When Bribe Type Is Endogenous,” Journal of Public Economic Theory, vol. 17. 580–​604. Rose-​Akerman, S. (1975), “The Economics of Corruption,” Journal of Public Economics, vol. 4. 187–​203. Samuelson, L. and Stacchetti, E. (2017), “Even Up: Maintaining Relationships,” Journal of Economic Theory, vol. 169. 170–​217.

3 Corruption, Institutional Trust and Legitimacy A Vicious Circle Amadou Boly and Robert Gillanders

Introduction Corruption remains endemic in much of the world with very few countries on course to meet the 2030 corruption target enshrined in the United Nations’ Sustainable Development Goals. In part, this is explicable by entrenched elites and their simultaneous responsibility and reluctance to monitor and punish corruption. In part, this is because ‘anticorruption strategies are adopted and implemented in cooperation with the very predators who control the government and, in some cases, the anticorruption instruments themselves’ (Mungiu-​ Pippidi, 2006: 87). Another reason for the persistence of corruption as a leading barrier to economic and social development is its self-​reinforcing nature. Stephenson (2020) outlines several ways in which a corruption trap can arise. These include a fall in the chance of an individual being caught and punished as fixed law enforcement resources are stretched thin as corruption spreads; outright capture of the justice system; increased confidence in the complicity of others in corrupt acts; a collapse in both internal ‘shame costs’ and external stigma; selection into roles that offer opportunities for corruption; and a smaller tax and resource base. Stephenson also points to a corruption-​ induced collapse in trust and increase in cynicism as a fertile ground for increased corruption. By undermining expectations of fairness and efficiency, corruption can create a vicious circle of mistrust and bribery. As Cho and Kirwin (2007: 6) put it, ‘people who have low expectations regarding the efficiency and impartiality of government try to look for an alternative to obtain access to public resources’. Cho and Kirwin test and confirm this bidirectional link using survey data from sub-​ Saharan Africa. In this chapter we draw together results from several related literatures and argue that taken together they point to a related mechanism that can explain why corruption is so persistent in so many parts of the world. Even if elites or reformers are willing to put laws in place, the legacy of corruption and mistrust may render them ineffective. We begin by discussing the literature that has shown that corruption has a significant, in both economic and statistical terms, effect on citizens’ trust in the institutions of the state. Predatory and DOI: 10.4324/9781003142300-3

16  Amadou Boly and Robert Gillanders rent-​seeking agents, even at a local level, generate a dim view of the trustworthiness of the state and even of other types of governance institutions. The role of contagion and peer effects in corruption is then discussed, showing that corruption can have socially undesirable consequences in terms of objectionable behaviours such as tax evasion. We then link these literatures to the emerging empirical work in economics and political science on the concept of legitimacy. Legitimacy, a property of law or policy that enhances compliance even when the probability of punishment is low, is also undermined by corruption. Therefore, even when existing elites or reformers put meaningful anti-​corruption laws in place, their efficacy is reduced by the legacy or perception of corruption, resulting in a vicious cycle of corruption. While we do not present a complete solution for this problem, our discussion points to the necessity of fighting corruption at the highest levels of authority in order to build legitimate and accountable states, noting that, to paraphrase a well-​known proverb, ‘actions speak louder than policies’.

The Importance of Trust Trust is a powerful force in economies and societies, particularly those in which the legal enforcement of contracts is prohibitively expensive or uncertain. Incomplete and asymmetric information and legal uncertainly can combine to tip the scales of rational decision-​making away from activities that would otherwise be pursued. Trust also makes it easier for citizens to solve collective action problems (Coleman, 1990; Putnam, 1993). Zak and Knack (2001) demonstrate the power of trust as an economic lubricant. Using data from the World Values Surveys, they show that trust is a significant predictor of growth and investment. While the small sample size available does require us to treat such findings with caution, a host of studies point to mechanisms that can drive such an observed relationship. For example, a number of studies have found that trust shapes investment behaviour (Guiso, Sapienza & Zingales, 2008; Georgarakos & Pasini, 2011). Tu and Bulte (2010) find that more trusting people in rural China are more likely to participate in the formal labour market. Trust also influences the way that society is organized and the economy is regulated. Trust has been found to be associated with support for deregulation and the level of economic regulation (Heinemann & Tanz, 2008; Aghion, Algan, Cahuc & Shleifer, 2010; Pinotti, 2012; Leibrecht & Pitlik, 2015). By increasing the scope for cooperation, trust also shapes support for redistributive policies and increases their effectiveness (Bergh & Bjørnskov, 2014; Daniele & Geys, 2015). It is hard to escape the conclusion that trust shapes both the outcomes of the economy in the aggregate and the ways in which people choose to live their lives and interact with others. Thus, given all we know about the costs of corruption, the scope for corruption to lead to a low-​trust, high-​ corruption vicious cycle has clear implications for economic and human development.

Corruption, Institutional Trust and Legitimacy  17

Corruption and Trust Many studies, drawing on survey and experimental data from around the world, have found that corruption corrodes trust in people and institutions. Corruption is almost always a violation of the law and, by definition, entails an abuse of entrusted power. Both collusive and harassment corruption transform the distribution of a society’s resources, and often justice, into a transactional and predatory system. While the transactional nature of such a system may require a form of honour or trust amongst thieves to function, observational and experimental studies point to corruption (both the experience and the perception) as being a significant inhibitor of trust between people in a society (Seligson, 2002; Uslaner, 2005; Banerjee, 2016). Our interest lies with the large body of work that has concluded that corruption, or the perception thereof, reduces public trust in the agents and institutions of a state. Importantly, this corrosive effect of corruption on institutional trust has been found to hold in different contexts in different parts of the world. An early contribution to this literature was provided by Seligson (2002). Using survey data from Latin America, Seligson shows that an experience of corruption predicts lower levels of perceived ‘regime legitimacy’, an index made up of perceptions of judicial fairness and trust in the police as well as respect, pride and support for the country’s political system and institutions. Many of the important studies in this literature have drawn on Latin American data. For example, Morris and Klesner (2010) find a similar relationship in Mexican survey data, though in this case the focus is on individual perceptions of corruption and the outcome variables explicitly capture trust in a range of public institutions. Blanco and Ruiz (2013) draw on Columbian data to exploit the opportunity to use both an experience-​based metric of corruption and explicit trust questions and find that being asked for bribe predicts lower satisfaction with democracy and lower trust in public institutions While there has been something of a focus on Latin America, these relationships are evident in data from other parts of the world. Anderson and Tverdova (2003) draw on survey data for a wide range of countries, including long established democracies and rich economies. They conclude that people in more corrupt countries, as per the Corruption Perceptions Index, are less trusting of civil servants. We also have evidence of these effects in Europe, sub-​Saharan Africa and East Asia. Ares and Hernández (2017) exploit the overlap in the timing of a corruption scandal in Spain and fieldwork for the European Social Survey to show using a natural experiment that the revelations led to a fall in trust in politicians. Drawing on Afrobarometer data, Lavallée, Razafindrakoto and Roubaud (2008) find that both experienced and perceived corruption reduces trust in public institutions. A similar result was found by Chang and Chu (2006) in the context of established and emerging Asian democracies using data from the East Asia Barometer, though only in terms of perceptions of corruption.

18  Amadou Boly and Robert Gillanders This brief overview of the literature makes it clear that the claim that corruption undermines trust has substantial empirical support and that the result is not contingent on particular cultural, institutional or historical features being present. The fact that corruption has an effect on trust in classes of actors and on institutions in general and not just an effect on particular actors is important for our idea of a vicious cycle. By destroying trust in the institutions of the state and in the type of people who serve as agents of the state, corruption, we will argue, destroys the perceived legitimacy of the state and the efficacy of many of the state’s (benign) interventions. As to why corruption spills over in this way, the literature on peer and contagion effects in corruption and similar norm violations is instructive (Gatti et al., 2003; Moxnes & Van der Heijden, 2003; Fisman & Miguel, 2007; Dong et al., 2012; d’Adda et al., 2017; Boly et al., 2019). The 2015 World Development Report also points to the powerful imitative effect of corruption (World Bank, 2014). In countries where corruption has become the norm, ‘acting corruptly may become automatic thinking for officials’ (p. 60). If citizens understand that norm violations are contagious in this way, individual experiences of corruption or perceptions as to the honesty of a particular official may spill over to undermine trust in the political or bureaucratic class as a group. Citizens may also use their experience as a rule of thumb when considering future interactions with the state and its representatives. As the 2015 World Development Report puts it, ‘without realizing it, people tend to fill in information gaps based on default assumptions consistent with their mental models’ (World Bank, 2014: 69). The results of Chong et al. (2015) further support the idea that corruption tars all politicians with the same brush.They report on the effects of a field experiment carried out in Mexico that provided information on incumbent corruption ahead of local elections. While voter turnout and support for corrupt incumbents fell, the same was also true for their opponents. Corruption and mistrust go hand in hand. As we will next argue, this has profound implications for the functioning and stability of society.

Trust, Corruption and Compliance The standard model of compliance in economics sees people considering what they stand to gain and lose and acting to maximize expected utility (Becker, 1968). Compliance can be incentivized by increasing the chance of getting caught or the punishment if a violation is detected. While such policies have been shown to be effective in terms of curbing, for example, corruption (Abbink et al., 2002; Olken, 2007), the standard model does not fully capture the factors shaping compliance behaviour. Evidence abounds that corruption and a lack of trust shape compliance behaviour. Of particular salience at the time of writing –​as the world struggles to contain the COVID-​19 pandemic and begins to think about how to avert the next global pandemic –​is the literature which has pointed to a lack of trust as

Corruption, Institutional Trust and Legitimacy  19 a key driver of compliance with public health policies and recommendations, up to and including vaccine uptake. Institutional trust was a driver of compliance with public health advice in the Ebola outbreaks that affected several African countries in the years prior to 2020. Blair et al. (2017) find that Liberian citizens expressing lower trust in government during the 2014/​15 outbreak adopted fewer of the suggested precautionary measures and were less compliant with public health requirements such as social distancing and the safe burial of victims. Importantly, low trust did not predict less understanding of the nature of Ebola and how it spread. Similar results were found by Vinck et al. (2019) in relation to a later outbreak in the Democratic Republic of Congo. Mistrust predicted less avoidance behaviour and reduced willingness to accept an Ebola vaccine. As was the case with the relationship between corruption and trust, it is important to note that this relationship between mistrust and compliance is discernible outside of sub-​Saharan Africa and the particular case of the Ebola virus. Similar relationships have been found in data from the United States (Ronnerstrand, 2014; Jung et al., 2013), Sweden (Ronnerstrand, 2013), Taiwan (Chuang et al., 2015) and Japan (Nawa & Fujiwara, 2019). Dincer and Gillanders (2021) find that more corrupt US states had lower levels of compliance with COVID-​19 shelter in place orders and argue that this arises due to the negative effect of corruption on institutional trust and legitimacy. There is also evidence that corruption reduces tax compliance, which is critical for public good provision, particularly in developing countries. The link between corruption and taxation can be understood through the fiscal exchange theory framework, which suggests that tax compliance is higher when citizens are offered valuable public goods –​such as representation, public services, security and infrastructure –​in exchange for tax payments (see Buchanan, 1976; Alm, Jackson & McKee, 1993). Consequently, citizens are likely to evade taxes more when there is a possibility that the promised services will not be delivered, for example, due to corruption (Besley & Persson, 2014). Levi and Sacks (2009) show using survey data from sub-​Saharan Africa that perceptions of government effectiveness and procedural justice, two factors that are generally viewed as incompatible with high levels of corruption, are associated with attitudes to tax compliance. Ali et al. (2014) find evidence that corruption weakens attitudes towards tax compliance in the case of South Africa and perhaps Uganda, though not in Kenya or Tanzania. At least in part, these results can be understood through the lens of ‘tax morale’ –​broadly understood as non-​pecuniary factors (intrinsic motivation, reciprocity) that encourage voluntary compliance. Tax morale, like so many of the forces that lead to the smooth functioning of society, is undermined by corruption (Picur & Riahi-​Belkaoui, 2006; Jahnke & Weisser, 2019). Boly, Konte and Shimeles (2021) also examine the effect of the quality of governance (proxied by perceived corruption) on attitudes towards paying taxes using the Afrobarometer surveys for thirty-​six African countries.They find that perceived corruption in the President’s Office has a significant and negative effect on

20  Amadou Boly and Robert Gillanders reported attitude towards taxation, even after controlling for people’s experience of corruption, as captured by bribe payments made to government officials in the past twelve months. This result suggests that perceptions of corruption at the highest level of government matter for tax morale. Although subjective, perceptions of corruption tend to contain real information and matter in shaping behavioural responses. For example, Olken (2009) examined the accuracy of Indonesian villagers’ corruption perceptions to a more objective measure of ‘missing expenditures’ in a road-​building project in their village and found that villagers’ reported perceptions do contain real information about the level of missing expenditures in the project. Likewise, in a corruption laboratory experiment in Kenya, Banerjee et al. (2022) find that participants’ expectations about corruption were in line with what they ultimately experienced in terms of embezzlement, supporting the notion that participants possessed local knowledge as to how people in the role of public official would behave. In a policy-​oriented laboratory experiment on tax evasion and corruption, Banerjee, Boly and Gillanders (2022) find that fighting corruption (or eliminating it altogether) also has a significant and negative impact on the share of income evaded, suggesting spill-​over effects from anti-​corruption to tax evasion behaviour. They also find a clear ranking in terms of public good provision, whereby fighting corruption is more effective than fighting tax evasion, suggesting that when confronted with the twin evils of corruption and tax evasion in the context of public goods provision, allocating more resource to fighting the former may be a better policy choice. Beyond taxation, Sundström (2012) provides a similar finding in the context of regulatory compliance by small-​scale fishermen (96% were men) in South Africa. In a survey experiment, fishermen who held a dimmer view of the corruptibility of the officials charged with enforcing the regulations were less willing to comply. One way in which these results can be reconciled with the standard model is that a lack of trust in the state leads to differences in the perceived expected costs and benefits of following public advice or engaging with the state. Similarly, the tax evasion and tax morale results can be understood in terms of corruption making tax evasion more practicable, and embezzlement from public funds reducing the benefits of paying taxes. Another strand of thought posits that a lack of social capital can explain these results by aggravating the collective action problem (Olson, 1971; Sønderskov, 2009). However, as we will now argue, the concept of legitimacy offers an additional mechanism to understand these results –​one that can give rise to a vicious cycle of corruption, mistrust and low compliance.

The Legitimacy Trap in Anti-​corruption Reforms Legitimacy is a property of a law, law enforcement agent, policy or policymaker that enhances compliance even when the threat of sanction is low (Tyler, 2006). The idea of legitimacy as a potent force shaping compliance decisions can be

Corruption, Institutional Trust and Legitimacy  21 traced back to the writings of Max Weber (Weber, 1968). MacCormack (1981) holds that ‘legitimacy arises from the ideal of a society maintained through impersonal, efficient procedures’ (p. 424), and the evidence supports this contention. The perceived fairness and impartiality of rules and procedures shape behaviours, even when the result is unequal (Lind, 2001; Falk et al., 2003; Tyler, 2004; Bolton et al., 2005). The importance of legitimacy in shaping decision-​making is borne out by several interesting quantitative studies. Paternoster et al. (1997) find that reoffending was less likely for those arrested for domestic assault in a procedurally fair way. Kuperan and Sutinen (1998) study Malaysian fishermen and find that compliance with regulations was higher amongst those who perceived the requirements as legitimate. A legitimacy effect is also apparent in the finding of Chen (2016) that differential rates of executions explain the higher rates of absences for Irish soldiers serving in the UK armed forces during the First World War. More contemporary evidence is provided by Christensen and Laegreid (2020), who argue that the Norwegian fight against COVID-​19 was bolstered by a strong sense of legitimacy. This argument fits well with the finding of Dincer and Gillanders (2021) that more corrupt US states faced lower levels of compliance with social-​distancing directives. The notions of institutional trust and legitimacy have considerable overlap, as noted by Jackson and Gau (2016). However, our view is that the legitimacy and procedural justice literatures offer a clear mechanism that can explain, at least in part, the observed negative effects corruption and mistrust have on compliance. Corruption is a fundamentally unfair phenomenon in which power and resources can be secured by one party at the expense of another, or of society in general, in a process that is almost always illegal and contrary to established formal processes. Experiencing this first hand or holding the belief that such acts are commonplace undermines the legitimacy effect that would otherwise enhance compliance with legal requirements and government requests in a host of contexts, including, as we have seen above, the domains of death and taxes. In the context of anti-​corruption policy, however, the potential for a feedback loop exits in which corruption destroys legitimacy which in turn renders anti-​ corruption efforts ineffective. The literatures that we have drawn on thus provide us with the following insights: • • • •

Corruption undermines trust in the institutions and agents of the state. Beliefs and experiences of corruption in relation to a given individual or class of actors can spill over to reduce trust in the state more generally. A lack of institutional trust is associated with reduced levels of compliance with formal legal requirements and government requests across a wide range of state–​citizen interactions. This can be understood by the damaging effect corruption –​acts of which inevitably violate principles of procedural justice and fairness –​has on the perceived legitimacy of the state and its agents.

22  Amadou Boly and Robert Gillanders

Widespread Corruption

Low Trust/Legtimacy

Lack of Compliance with AntiCorrupion Laws and Policies

Figure 3.1 The legitimacy trap.

Taken together, this suggests that it is plausible that corruption, by compromising the perceived legitimacy of anti-​ corruption policymakers and enforcement agencies, can foster further corruption by reducing compliance with anti-​ corruption laws and policies. Importantly, given the results that suggest that corruption spills over to create mistrust in political and bureaucratic actors as a class, even well-​intentioned reforms and reformers may be met with lower compliance than the standard model of compliance would suggest. Less effective anti-​corruption policy will lead to more corruption, and therefore public trust and legitimacy remain low. This vicious cycle, or legitimacy trap, is illustrated in Figure 3.1. While we have discussed papers that point to legitimacy effects in various contexts, none of the studies cited thus far have provided evidence of a legitimacy effect in anti-​corruption policy. However, in this specific domain of anti-​ corruption policy, Boly, Gillanders and Miettinen (2019) provide evidence of a legitimacy effect. Drawing on a framed laboratory experiment conducted in Nairobi, Kenya, this study provides clear evidence that the behaviour of an anti-​corruption policymaker determines the efficacy of policy. The basic experimental game sees two participants (drawn from a subject pool of undergraduate students) randomly paired and assigned the role of public officials charged with investing money into real local charities. Both have the possibility of embezzling some of these funds before passing them on. Public Official A makes her decisions first, which are observed by the other member of the pair, Public Official B. In this design, legitimacy is operationalized by allowing ‘Public Official A’ who chooses the strength of the deterrence policy to be corrupt or not. While the control treatment has no mechanism through which corruption can be detected and punished, additional treatments introduce an anti-​ corruption policy based on deterrence in which corruption, if detected, results

ED

END

Control

(2 0%

) (3 0%

-2 5% )

% (5 % -1 5 M

Lo w

H ig h

(3 0%

)

) Ze ro

) (2 0%

-2 5%

) M

ed iu m

XND

Control

Control

(5 % -1 5%

XND

H ig h

XND XND

ed iu m

Control

Lo w

Ze ro

Share of Offical Bs who are Corrupt .7 .8 .9 1

Corruption, Institutional Trust and Legitimacy  23

Probability of Detection Honest A

Corrupt A

Figure 3.2 The legitimacy effect in anticorruption policymaking. Source: Authors’ calculations adapted from Boly et al. (2019).

in the complete loss of earnings for that round of play. In the first, the policy is endogenous and discretionary (ED). Public Official A chooses the strength of the policy (a 0–​30% chance of detection and punishment), but it only applies to Public Official B. In the second, the policy is endogenous and non-​discretionary (END) in that the policy selected by Public Official A applies to both officials. In the final, the policy is exogenously set (at 30%) and non-​discretionary (XND). Figure 3.2, adapted from Boly et al. (2019), shows clear evidence of a legitimacy effect in the END treatment. A stronger anti-​corruption policy only deters Public Official B from embezzlement if the policymaker has acted honestly herself. Indeed, the effect of the policy is much weaker if the source is a corrupt policymaker. The lack of an effect in the ED treatment, where the anti-​corruption policy did not apply to the ‘higher’ public official, speaks to the importance of perceived procedural fairness in creating a legitimacy effect. Indeed, the procedural fairness literature suggests that legitimacy stems from a shared perception between all relevant parties on the fairness of the procedures applied, even though outcomes may be unequal (Lind 2001; Tyler 2004). Although further research is needed, the implications of this legitimacy effect could also extend to other policymaking situations where the actions of the policymaker in a given domain run contrary to the objectives of their promoted policy. For example, a policymaker trying to curb tax evasion while putting his/​ her own wealth in a tax haven. To paraphrase a well-​known proverb: ‘Actions speak louder than policies’. Attempts to deter corruption through punishment institutions have typically failed in societies with high levels of corruption (see, e.g., Mungiu-​Pippidi,

24  Amadou Boly and Robert Gillanders 2017; Rothstein, 2011). We contend that a legitimacy trap in which corruption renders anti-​corruption policy less effective is a plausible mechanism to explain, at least in part, the dearth of sustained escapes from a high corruption equilibrium. Corruption begets corruption by reducing compliance with anti-​ corruption policy. As we argue in the next section, with reference to previous anti-​corruption campaigns, re-​establishing trust and legitimacy in authorities is an essential component for any successful anti-​corruption campaign.

Breaking Out of the Legitimacy Trap in Anti-​corruption Reforms To some extent, breaking out of the legitimacy trap requires top-​level leaders to show, through their own words and actions, a zero tolerance for corruption. Yet, looking through history, one could conclude that leaders who were intrinsically motivated to fight corruption are rare. One example is Thomas Sankara, former president of Burkina Faso (1983–​1987), whose rhetoric focused heavily on anti-​corruption principles, going so far as renaming the country Burkina Faso (land of honest/​incorruptible men) from Upper Volta (Hagberg, 2015). However, in 1987, Sankara’s regime was overthrown in a military coup by one of his companions, Blaise Compaoré, and corruption gradually become a widespread problem again in Burkina Faso (Ardigo, 2019). Such an outcome suggests that the emergence of a public-​minded leader is a necessary but not sufficient condition to break out of the legitimacy trap in anti-​corruption reforms, as public-​minded reformers could face strong opposition due to powerful interests that benefit from corruption. To possibly decrease resistance, leaders can also develop a strategy based on a gradual approach by directing anti-​corruption efforts to critical sectors (such as courts, customs, education) or to those where corruption is perceived as acute by the public. Singapore’s example demonstrates the long-​term aspect of successful anti-​ corruption reforms and the need for a broad base of support. While Thomas Sankara’s anti-​corruption efforts were led by a charismatic leader and lasted for about four years, the People’s Action Party, which successfully implemented Singapore’s anti-​corruption strategy (as part of the country’s strategy to achieve its development goals), has been ruling since 1959, making it one of the parties that have served longest and uninterruptedly among multiparty parliamentary democracies in the world. A consistent and lasting implementation of a comprehensive strategy that seeks to reduce both the opportunities and incentives for corruption partly explains Singapore’s success of becoming one of the least corrupt countries in the world (see, e.g., Quah, 2001). Indeed, moving a country from a ‘high corruption’ to a ‘low corruption’ equilibrium is a long-​ term social transformation that requires sufficient time and high level of public participation. While some level of national effort is a prerequisite in the fight against corruption, such efforts are likely to be insufficient and ineffective without international cooperation, particularly regarding corruption by leaders or top

Corruption, Institutional Trust and Legitimacy  25 officials. International cooperation is critical in discouraging corruption by top officials, by restricting illegal financial outflows by corrupt leaders seeking to place their assets abroad and by facilitating recovery of these assets. An illustrative example of the role played by foreign financial institutions is the case of former Nigerian president Sani Abacha (1993–​1998), who managed to deposit about USD 2.4 billion or more in bank accounts in Switzerland, Liechtenstein, Jersey, the UK and France. Nigeria has been trying but has not managed to fully recover the funds since 1999. International conditionalities that apply across many countries, particularly development aid, could also incentivize a national leader to fight corruption, or help justify anti-​corruption measures that might otherwise be challenging to implement (Klitgaard, 1998). In the international cooperation area, external or non-​state actors such as multilateral development organizations may, in some contexts, also be able to assist in breaking the cycle by being the guarantors of anti-​corruption policies and reforms, provided they are perceived as honest and objective brokers. However, this is not necessarily the case. For example, Breen and Gillanders (2015) find that ratings of the effectiveness of both the IMF and the World Bank are undermined by domestic corruption perception. To the extent that this will lower public support for interventions from these entities, it will limit the scope for these external actors to assist in breaking the cycle of corruption by rule-​setting or monitoring. Fortunately, further evidence suggests that, at least in developing and transition economies, trust in the UN increases with the level of domestic corruption (Torgler, 2008).

Conclusion The literatures that show that corruption undermines public trust in government and that compliance is weaker once legitimacy has been undermined combine to paint a grim picture. A legitimacy trap, whereby corruption begets corruption through a corrosive effect on compliance, is a plausible mechanism that can help to explain why corruption remains so hard to fight. Even well-​ intentioned and honest reformers may find their policies ineffective as they inherit a legacy of mistrust and perceived unfairness and illegitimacy. As Boly and Gillanders (2018) show in their experiment that policymakers can distort anti-​corruption institutions in a ‘weak’ institutional environment, we would also argue that to stop the cycle from starting, an honest top leadership should establish strong institutions while in power by strengthening the legislative and judicial power as well as by ensuring a free and vibrant media. Boly and Gillanders (2018) also offer some reason for optimism.While participants in this experiment did choose weaker deterrence policies when they would find themselves constrained by the instrument, even those who went on to embezzle themselves did not, on average, implement a zero deterrence policy. If leaders are on average possessed of these kinds of prosocial preferences, understanding how to strengthen them and credibly communicate this fact to citizens could go some way to breaking the legitimacy trap.

26  Amadou Boly and Robert Gillanders Given how difficult corruption has proved to be to curb, never mind eradicate, we believe that this idea of a legitimacy trap warrants further exploration. While we have no fully compelling solution for this problem, the literature strongly points to the importance of top-​level leaders showing a commitment to zero tolerance policy for corruption through their own words and actions. Yet, leadership can clearly be endogenous. Self-​selection and political processes can mean that, at least in some contexts with some frequency, those with a low threshold for participating in corruption, or an out-​and-​out preference for corruption, can find themselves in leadership positions and as the custodians of vested interests. In those contexts, serendipity is part of the process that would lead to the emergence of an uncorrupted and/​or public-​minded leader. As all cannot be left to serendipity, efforts to raise citizens’ awareness on the decisive role of leadership in curbing corruption (including through empirical research as in, e.g., Boly & Gillanders 2018; Boly, Gillanders & Miettinen, 2019; Boly, Konte & Shimeles, 2021) can create some pressure on leadership to change behaviour and commit to fighting corruption. More surveys and experimental work could usefully explore the strength and duration of the corruption compliance cycle and what can be done to break it. One path worth exploring is suggested by the arguments of Singer (2006). Singer argues that the legitimacy of a legal system requires three criteria to be satisfied: it must form its own worldview; it must negotiate its own role within society; and it must take responsibility for its own integrity. Clearly, in the context of corruption, each of these criteria is particularly challenging, but modelling and experimental tests of these concepts could shed further light on how legitimacy traps emerge and how societies can escape them.

References Abbink, K., Irlenbusch, B. & Renner, E. (2002). An experimental bribery game. Journal of Law, Economics, and Organization, 18(2), 428–​454. Aghion, P., Algan, Y., Cahuc, P. & Shleifer, A. (2010). Regulation and distrust. Quarterly Journal of Economics, 125(3), 1015–​1049. Ali, M., Fjeldstad, O. H., & Sjursen, I. H. (2014).To pay or not to pay? Citizens’ attitudes toward taxation in Kenya, Tanzania, Uganda, and South Africa. World development, 64, 828–​842. Alm, J., Jackson, B. R. & McKee, M. (1993). Fiscal exchange, collective decision institutions, and tax compliance. Journal of Economic Behavior & Organization, 22(3), 285–​303. Alt, J. E. & Lassen, D. D. (2012). Enforcement and public corruption: Evidence from the American states. Journal of Law, Economics, and Organization, 30(2), 306–​338. 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. Ardigo, I. A. (2019). Burkina Faso: Overview of corruption and anti-​ corruption. Transparency International. https://​knowl​edge​hub.trans​pare​ncy.org/​ass​ets/​uplo​ads/​ helpd​esk/​Coun​try-​prof​i le-​Burk​ina-​Faso-​2019​_​PR.pdf.

Corruption, Institutional Trust and Legitimacy  27 Ares, M. & Hernández, E. (2017). The corrosive effect of corruption on trust in politicians: Evidence from a natural experiment. Research & Politics, 4(2), 2053168017714185. Banerjee, R. (2016). Corruption, norm violation and decay in social capital. Journal of Public Economics, 137, 14–​27. Banerjee, R., Boly, A., & Gillanders, R. (2022). Anti-tax evasion, anti-corruption and public good provision: An experimental analysis of policy spillovers. Journal of Economic Behavior & Organization, 197, 179–194. Becker, G. (1968). Crime and punishment: An economic approach. Journal of Political Economy, 169, 176–​177. Bergh, A. & Bjørnskov, C. (2014). Trust, welfare states and income equality: Sorting out the causality. European Journal of Political Economy, 35, 183–​199. Besley, T. & Persson, T. (2014). Why do developing countries tax so little? Journal of Economic Perspectives, 28(4), 99–​120. Blair, R. A., Morse, B. S. & Tsai, L. L. (2017). Public health and public trust: Survey evidence from the Ebola virus disease epidemic in Liberia. Social Science & Medicine, 172, 89–​97. Blanco, L. & Ruiz, I. (2013). The impact of crime and insecurity on trust in democracy and institutions. American Economic Review, 103(3), 284–​288. Bolton, G. E., Brandts, J. and Ockenfels, A. (2005). Fair procedures: Evidence from games involving lotteries. Economic Journal, 115(506), 1054–​1076. Boly, A. & Gillanders, R. (2018). Anti-​corruption policy making, discretionary power and institutional quality: An experimental analysis. Journal of Economic Behavior & Organization, 152, 314–​327. Boly, A., Gillanders, R. & Miettinen, T. (2019). Deterrence, contagion, and legitimacy in anticorruption policy making: An experimental analysis. Journal of Legal Studies, 48(2), 275–​305. Boly, A., Konte, M. and Shimeles, A. (2021). Corruption perception and attitude towards taxation in Africa. Journal of African Economies, 30(Issue Supplement_​1), pp. i140–​i157. Breen, M. & Gillanders, R. (2015). Political trust, corruption, and ratings of the IMF and the World Bank. International Interactions, 41(2), 337–​364. Buchanan, J. M. (1976). Barro on the Ricardian equivalence theorem. Journal of Political Economy, 84(2), 337–​342. Chang, E. C. & Chu,Y. H. (2006). Corruption and trust: Exceptionalism in Asian democracies? Journal of Politics, 68(2), 259–​271. Chen, D. L. (2016). The deterrent effect of the death penalty? Evidence from British commutations during World War I. IAST Working Papers 16–​ 50. Institute for Advanced Study in Toulouse. Cho,W. & Kirwin, M. F. (2007). A vicious circle of corruption and mistrust in institutions in sub-​Saharan Africa: A micro-​level analysis. Afrobarometer Working Paper No. 71. Institute for Democracy in South Africa. Chong, A., De La O, A. L., Karlan, D. & Wantchekon, L. (2015). Does corruption information inspire the fight or quash the hope? A field experiment in Mexico on voter turnout, choice, and party identification. Journal of Politics, 77(1), 55–​71. Christensen,T. & Lægreid, P. (2020). Balancing governance capacity and legitimacy: How the Norwegian government handled the COVID-​19 crisis as a high performer. Public Administration Review, 80(5), 774–​779.

28  Amadou Boly and Robert Gillanders Chuang,Y. C., Huang,Y. L., Tseng, K. C.,Yen, C. H. & Yang, L. H. (2015). Social capital and health-​protective behavior intentions in an influenza pandemic. PloS One, 10(4), e0122970. Coleman, J. (1990). Foundations of SocialTheory. Cambridge, MA: Harvard University Press. d’Adda, G., Darai, D., Pavanini, N. and Weber, R. A. (2017). Do leaders affect ethical conduct? Journal of the European Economic Association, 15(6), 1177–​1213. Daniele, G. & Geys, B. (2015). Interpersonal trust and welfare state support. European Journal of Political Economy, 39, 1–​12. Dincer, O. & Gillanders, R. (2021). Shelter in place? Depends on the place: Corruption and social distancing in American states. Social Science & Medicine, 269, 113569. Dong, B., Dulleck, U. & Torgler, B. (2012). Conditional corruption. Journal of Economic Psychology, 33(3), 609–​627. Falk, A., Fehr, E. & Fischbacher, U. (2003). On the nature of fair behavior. Economic Inquiry, 41(1), 20–​26. Fisman, R. & Miguel, E. (2007). Corruption, norms, and legal enforcement: Evidence from diplomatic parking tickets. Journal of Political Economy, 115(6), 1020–​1048. Gatti, R., Paternostro, S. and Rigolini, J. (2003). Individual attitudes toward corruption: Do social effects matter? World Bank Policy Research Working Paper 3122. Washington, DC: World Bank. Georgarakos, D. & Pasini, G. (2011). Trust, sociability, and stock market participation. Review of Finance, 15(4), 693–​725. Guiso, L., Sapienza, P. & Zingales, L. (2008).Trusting the stock market. Journal of Finance, 63(6), 2557–​2600. Hagberg, S. 2015. ‘Thousands of New Sankaras’: Resistance and struggle in Burkina Faso. Africa Spectrum, 50(3), 109–​121. Heinemann, F. & Tanz, B. (2008). The impact of trust on reforms. Journal of Economic Policy Reform, 11(3), 173–​185. Jackson, J. & Gau, J. M. (2016). Carving up concepts? Differentiating between trust and legitimacy in public attitudes towards legal authority. In Interdisciplinary Perspectives on Trust (pp. 49–​69). Cham: Springer. Jahnke, B. & Weisser, R. A. (2019). How does petty corruption affect tax morale in sub-​ Saharan Africa? European Journal of Political Economy, 60, 101751. Jung, M., Lin, L. & Viswanath, K. (2013). Associations between health communication behaviors, neighborhood social capital, vaccine knowledge, and parents’ H1N1 vaccination of their children. Vaccine, 31(42), 4860–​4866. Klitgaard, R. (1998). International cooperation against corruption. Finance & Development, 35(1), 3–​6. Kuperan, K. & Sutinen, J. G. (1998). Blue water crime: Deterrence, legitimacy, and compliance in fisheries. Law and Society Review, 32(2), 309–​338. Lavallée, E., Razafindrakoto, M. & Roubaud, F. (2008). Corruption and trust in political institutions in sub-​Saharan Africa. Afrobarometer Working Paper No. 102. Leibrecht, M. & Pitlik, H. (2015). Social trust, institutional and political constraints on the executive and deregulation of markets. European Journal of Political Economy, 39, 249–​268. Levi, M. & Sacks, A. (2009). Legitimating beliefs: Sources and indicators. Regulation & Governance, 3(4), 311–​333. Lind, E. A. (2001). Fairness heuristic theory: Justice judgments as pivotal cognitions in organizational relations. Advances in Organizational Justice, 56(8), 56–​88.

Corruption, Institutional Trust and Legitimacy  29 MacCormack, C. P. (1981). Health care and the concept of legitimacy. Social Science & Medicine. Part B: Medical Anthropology, 15(3), 423–​428. Morris, S. D. & Klesner, J. L. (2010). Corruption and trust: Theoretical considerations and evidence from Mexico. Comparative Political Studies, 43(10), 1258–​1285. Moxnes, E. & Van der Heijden, E. (2003).The effect of leadership in a public bad experiment. Journal of Conflict Resolution, 47(6), 773–​795. Mungiu, A. (2006). Corruption: Diagnosis and treatment. Journal of democracy, 17(3), 86–​99. Mungiu-​Pippidi, A. (2017). The time has come for evidence-​based anticorruption. Nature Human Behaviour, 1(1), 1–​3. Nawa, N. & Fujiwara, T. (2019). Association between social capital and second dose of measles vaccination in Japan: Results from the A-​CHILD study. Vaccine, 37(6), 877–​881. Olken, B. A. (2007). Monitoring corruption: Evidence from a field experiment in Indonesia. Journal of Political Economy, 115(2), 200–​249. Olken, B. A. (2009). Corruption perceptions vs. corruption reality. Journal of Public Economics, 93(7–​8), 950–​964. Olson, M. (1971). The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge, MA: Harvard University Press. Paternoster, R., Brame, R., Bachman, R. & Sherman, L. W. (1997). Do fair procedures matter? The effect of procedural justice on spouse assault. Law and Society Review, 31(1), 163–​204. Picur, R. D. & Riahi-​Belkaoui, A. (2006).The impact of bureaucracy, corruption and tax compliance. Review of Accounting and Finance, 5(2), 174–​180. Pinotti, P. (2012). Trust, regulation and market failures. Review of Economics and Statistics, 94(3), 650–​658. Putnam, R. (1993). Making Democracy Work: Civic Tradition in Modern Italy. Princeton: Princeton University Press. Quah, J. S. (2001). Combating corruption in Singapore: What can be learned? Journal of Contingencies and Crisis Management, 9(1), 29–​35. Ronnerstrand, B. (2013). Social capital and immunization against the 2009 A(H1N1) pandemic in Sweden. Scandinavian Journal of Public Health, 41, 853–​859. Rönnerstrand, B. (2014). Social capital and immunization against the 2009 A (H1N1) pandemic in the American states. Public Health, 128(8), 709–​715. Rothstein, B. (2011). Anti-​corruption: The indirect ‘big bang’ approach. Review of International Political Economy, 18(2), 228–​250. Seligson, M. A. (2002). The impact of corruption on regime legitimacy: A comparative study of four Latin American countries. Journal of Politics, 64(2), 408–​433. Serra, D. (2006). Empirical determinants of corruption: A sensitivity analysis. Public Choice, 126(1), 225–​256. Singer, M. (2006). Legitimacy criteria for legal systems. King’s Law Journal, 17(2), 229–​253. Sønderskov, K. M. (2009). Different goods, different effects: Exploring the effects of generalized social trust in large-​N collective action. Public Choice, 140(1–​2), 145–​160. Stephenson, M. C. (2020). Corruption as a self-​reinforcing trap: Implications for reform strategy. World Bank Research Observer, 35(2), 192–​226. Sundström, A. (2012). Corruption and regulatory compliance: Experimental findings from South African small-​scale fisheries. Marine Policy, 36(6), 1255–​1264.

30  Amadou Boly and Robert Gillanders Torgler, B. (2008). Trust in international organizations: An empirical investigation focusing on the United Nations. Review of International Organizations, 3(1), 65–​93. Tu, Q. & Bulte, E. (2010).Trust, market participation and economic outcomes: Evidence from rural China. World Development, 38(8), 1179–​1190. Tyler, T. R. (2004). Enhancing police legitimacy. Annals of the American Academy of Political and Social Science, 593(1), 84–​99. Tyler, T. R. (2006). Psychological perspectives on legitimacy and legitimation. Annual Review of Psychology, 57, 375–​400. Uslaner, E. M. (2005).Trust and corruption. In J. G. Lambsdorf, M.Taube, & M. Schramm (Eds.), The new institutional economics of corruption (pp. 76–​92). Routledge: London and New York. Vinck, P., Pham, P. N., Bindu, K. K., Bedford, J. & Nilles, E. J. (2019). Institutional trust and misinformation in the response to the 2018–​19 Ebola outbreak in North Kivu, DR Congo: A population-​based survey. Lancet Infectious Diseases, 19(5), 529–​536. Weber, M. (1968). Economy and Society (ed. G. Roth and C.Wittich), Berkeley: University of California Press. World Bank. (2014). World Development Report 2015: Mind, Society, and Behavior. The World Bank. Zak, P. J. & Knack, S. (2001). Trust and growth. Economic Journal, 111(470), 295–​321.

4 Legal Systems and Corruption Ruben Korsten and Andrew Samuel

Introduction The relationship between corruption and “the law” is multifaceted because both corruption and the ability to fight it occur within the “shadow of the law.” Specifically, the presence of a formal legal system along with officials who enforce those laws is what makes corruption –​“[the] abuse of public office … for private gain” (Naidoo, 2013) –​feasible. In the absence of such laws and regulations, there would be little to no scope for corruption because regulators and officials would not have “control rights” and thereby the ability to accept bribes to avoid regulations, or kickbacks to grant favors. Since corruption occurs within the shadow of the law, arguably, it could be affected by the legal system from which those laws arise (see a.o.Treisman, 2000; Dreher, Kotsogiannis & McCorriston, 2009; La Porta, Lopez-​de Silanes, Shleifer & Vishny, 1999; Djankov, La Porta, Lopez-​de Silanes & Shleifer, 2002; La Porta, Lopez-​de Silanes & Shleifer, 2008). Broadly construed, there are two dominant legal systems worldwide: the (English) common law tradition that is found in anglophone North America, the United Kingdom (except for Scotland) and most of the Commonwealth, and the civil law tradition found across continental Europe, South America, (non-​Commonwealth) Asia and (francophone) Africa (David & Brierley, 1978). In a series of papers (a.o. La Porta, Lopez-​de Silanes, Shleifer & Vishny, 1997, 1998) summarized in the survey, La Porta et al. (2008) argue that when comparing common law versus civil law jurisdictions, the latter tend to be more corrupt and experience less economic growth than their common law counterparts. The association between a country’s legal origin and the level of corruption has been attributed to various differences between the two major legal traditions. Inter alia, it has been argued that the common law was designed to resolve disputes, often between the ruler and the people. In contrast, civil law systems were largely developed to implement the policies of the ruler. This difference is said to affect corruption in three ways. First, because it was developed to resolve disputes between the ruler and the ruled, the common law provides greater protection against exploitation by the state (which in turn translates into less corruption) (La Porta et al., 2008; Mahoney, 2001). Also, the willingness of DOI: 10.4324/9781003142300-4

32  Ruben Korsten and Andrew Samuel common law judges to follow procedure even when the results threaten hierarchy increases the chances that corruption will be exposed (in this regard, see Treisman, 2000). Second, because the civil law systems can be associated with a more interventionist government, officials have greater powers, which invites bureaucracy and corruption (La Porta et al., 1999). Third, since there is less (emphasis on) regulation in common law countries, there is also less scope for bribery (La Porta et al., 2008; Djankov et al., 2002). Although the literature on legal origins and corruption is well established, two critical issues remain unresolved. First, the mapping between common and civil law and more/​less corrupt regimes is not perfect.There are several civil law countries (e.g., Germany and Scandinavian countries) that rank low on almost any measure of corruption. Indeed, there are sufficiently many civil law countries with low levels of corruption that they hardly count as anomalies. The extant literature explains this by arguing that the type of civil law present in these countries differs from those with a French civil law legal origin. But, as we will argue subsequently, parsing the civil law tradition to explain such phenomena is somewhat arbitrary from the stand point of legal scholarship in this area. Second, to the extent that these patterns are observed empirically, the causal pathway linking legal origin to corruption is not well understood. In part, this is because, until recently, there has been very limited formal modeling of legal systems. In this chapter, we attempt to address these two deficiencies by building a formal model to link one type of corruption –​bribery –​to legal systems. Corruption can manifest itself in many ways. Thus, building a formal model to encapsulate every type of corruption would not be feasible. Hence, we focus on studying how the presence of bribery in enforcement settings may be linked to a country’s legal system. This extends prior work on the common law (Parameswaran, 2018; Parameswaran & Samuel, 2019) to encapsulate both common and civil law. Developing this theoretical framework allows us to begin addressing the two shortcomings of previous literature. First, the model allows us to clearly identify causal links between a legal system and corruption. Second, using the model, we show that it is indeed possible for corruption to be lower in civil law countries, depending on some key parameters such as the power that bureaucrats hold relative to citizens. Specifically, we show that if bribery is to be more frequent in civil law countries, then bureaucratic power needs to be lower in civil law countries than in common law countries. This, of course, goes against informal arguments that claim that higher bureaucratic power in civil law countries (relative to common law countries) explains why there is more corruption in those countries. Thus, the model suggests more caution when interpreting observed patterns in the data, specifically without understanding the possible underlying causal pathways more clearly. To understand how we model a legal system, it is first important to understand the key differences across legal systems. Thus, in the section Legal Origins and Corruption below we first discuss the differences between the common and civil law. Then, in the following sections we model these two systems formally

Legal Systems and Corruption  33 with and without bribery. Finally, we conclude by discussing how the model can be used to (re)interpret findings concerning bribery and legal origins.

Legal Origins and Corruption To understand the link between legal origin and corruption, we must first understand how it enters a system. Corruption is defined as the misuse of public office for private gain (Naidoo, 2013). Typically, the official misusing public office is a regulator who is tasked with enforcing some regulation, such as environmental laws against pollution.1 Instead of enforcing these laws, an official may accept gifts or bribes in exchange for being lenient with the firm. Thus, in order to understand how corruption fits into a legal system, we need to understand the (differences in the) role of regulators and regulation and how cases are decided across common and civil law countries.2 In almost any jurisdiction, laws are enforced jointly by a regulator and the courts. For example, an Antitrust Authority will typically file a case against a cartel, which will often end up in court. Even more mundane laws, such as for speeding, are enforced by the police and the courts because the accused can go to “traffic court” to dispute the ticket. Of course, in straightforward cases such as speeding, the outcome is so clear that the regulator (the police) can summarily enforce the law without recourse to the court.3 Whereas, more complex cases such as an antitrust case will often come before a judge, because the alleged violation is not easily determined. Regardless, laws are enforced within the shadow of their legal system. In civil law countries, the legislature codifies the law extensively, so that in general there are few scenarios in which the regulator will need to go to a judge. Civil law countries tend to have a system of rules that regulate actions in advance (for all future cases), as a result of which in most cases the regulator will be able to summarily determine whether an agent has violated the law. Indeed, even when a case comes before a judge, judges search the legislation for (a) rule(s) that govern(s) the dispute and apply it/​them to the particular facts of the case, focusing primarily on the (original) intent of the legislator.4 Thus, in a loose sense, because the law is so highly codified, most enforcement is closer to the traffic violation scenario described above, where the case can be summarily decided. In contrast, in common law systems, the law is far less codified. Legislation in common law systems is often incomplete, vague or ambiguous in the sense that it is typically not written in a way that regulates all possible actions in advance. Thus, regulators will often not be able to summarily decide a case and will need to seek the court’s remedy. When presented with a case, judges search to see if there is relevant precedent (from a previous legal case). This precedent leads to (ambiguous) legal principles, giving the courts the space to “narrow” their implementation (and interpretation) over time (solvitur ambulando, loosely translated from Latin: “it is solved by walking”; Cooper, 1950; see Zweigert & Kötz, 1998; Coffee Jr., 1998; Johnson, La Porta, Lopez-​de Silanes & Shleifer, 2000).5 Accordingly, courts play a crucial role in clarifying the rule or creating new law (via precedent).6

34  Ruben Korsten and Andrew Samuel Thus, to summarize the key distinction between these two systems: in civil law countries, the primary source of law is the legislation, which leaves less interpretation to the courts (David & Brierley, 1978; Dainow, 1966). Hence, either the regulator can enforce the law almost independently, or the court searches for the applicable rule and summarily decides the case. Whereas in common law jurisdictions the regulator will often need to seek clarification from the courts before enforcing the law. Historical differences in the origins of the two legal systems have led to these important differences. The civil law tradition originates in both Roman and canon law (private law) and natural law (public law) and builds on statute law and abstract rules as the primary source of law, with the purpose of providing solutions to all future cases (Zweigert & Kötz, 1998; David & Brierley, 1978; Dainow, 1966). In the twelfth and thirteenth centuries, the European universities and the Catholic Church rediscovered the Roman law and embraced it to lay the foundations for (private) secular laws in many European countries (David & Brierley, 1978). In the words of Thomas Aquinas, also seen as the “architect of the law’s secularization” (David & Brierley, 1978): He who resists the power, withstands the ordinance of God (Rom. xiii, 2). Now it is not legitimate to withstand the ordinance of God. Hence it is not legitimate either to withstand secular power … Christians are bound to obey the authorities inasmuch as they are from God; and they are not bound to obey inasmuch as the authority is not from God. (Aquinas, Phelan, Kenny & Eschmann, 2014) Thus, under this view, the ruler has the authority, but since the church is above the ruler in matters of doctrine and morality, he is obliged to adapt his laws accordingly. As such, the influence of the hierarchical Catholic Church on the law and the state is also believed to have conditioned social hierarchy in civil law countries (see Treisman, 2000; La Porta et al., 1999). In the seventeenth and eighteenth centuries, based on the “principles of reason” of the Enlightenment, liberal thinkers (such as Grotius, Hugo and Domat) started to lay the foundations for a public law that was to protect the individual rights and liberties of man. The Natural Law School entrusted the sovereign to formulate rules that conformed to reason and the new principles of “justice, liberty and dignity” (David & Brierley, 1978). This was the final step toward the “period of legislative law” (David & Brierley, 1978) or codification in Europe. The introduction of codes in the nineteenth century, combined with nationalistic sentiment, led to a legal positivist approach to law, that is, the will of the legislator, embodied in the national law (and nothing else), was the law and that this was the supreme interpretation of justice (David & Brierley, 1978). This civil law system, which originated in mainland Europe, has now been adopted in many other parts of the world, mostly imposed through conquest and colonization by European countries in the past, but also voluntarily (see a.o. Zweigert & Kötz, 1998; David & Brierley, 1978).The system can be divided

Legal Systems and Corruption  35 into various sub-​traditions, such as French, German and Scandinavian civil law. Proponents of the legal origins theory (La Porta et al., 2008) tend to emphasize the differences in these sub-​divisions in order to explain why many civil law countries (such as Germany) are the least corrupt. For example, they argue that the German civil law has more judicial lawmaking when compared to French civil law. They also argue that German civil law countries may have more efficient bureaucracies than French civil law countries, but without any substantial explanation for this, this statement seems ad hoc. Accordingly, many proponents of the legal origins theory (La Porta et al., 1997, 1998, 2008) argue that when comparing the common and the civil law, only the “French” civil law7 should be compared with the common law.8 We find no basis for this conclusion in the extensive comparative legal literature.9 Additionally, we find no grounds to treat the various civil law systems as substantially different from each other. Indeed, based on the extensive comparative legal literature, we find that in all civil law jurisdictions, the primary focus when trying to find a solution for a specific case will be on finding the appropriate rules in the legislation, searching –​primarily –​for the intention of the legislator (David & Brierley, 1978; Dainow, 1966); working syllogistically,10 resolving a concrete case under a general rule (Germani et al., 2007). Rather than being a system of norms, the legislation is usually seen as a framework of which the interpretation and application is a hermeneutical discipline aimed at a coherent interpretation of law based on an internal consistency of the system derived from language and value judgments (Gelter & Grechenig, 2014). The same goes for judges in civil law countries: judges do at times make use of the possibilities of interpreting the law creatively, but a true doctrine of precedent or stare decisis or judicial lawmaking is rejected and the power of judges goes no further than administering the law and interpreting the intent of the legislator (David & Brierley, 1978; Zweigert & Kötz, 1998; Dainow, 1966). For the historic reasons mentioned above, civil law countries can thus be associated with a high trust of government and a large activist government. Hence, all civil law traditions (German, French or others) were largely developed to implement the policies of the ruler (the church and the state). They are legislature-​based and focus on finding just (often in the eyes of the legislator) solutions to disputes. The common law developed in England,11 because aristocrats/​landowners and merchants wanted a system of law that provided strong protections for property and contract rights precisely against the Crown (and also the church) (Seng, 2001). In response, the king, as early as the twelfth and thirteenth centuries, established a system of king’s courts and judges that was meant to establish legal order and create uniform rules that were common throughout the whole country (hence the “common law”) (Dainow, 1966; Zweigert & Kötz, 1998). The source and essence of the law were therefore judicial decisions (Dainow, 1966); a strong legislative body did not exist until much later. It builds on a strong judiciary that is independent from the executive and the legislator and who shapes the law through legal precedent by solving legal cases.Treisman (2000) identifies a strong focus on the procedural aspects of the law and a more

36  Ruben Korsten and Andrew Samuel pragmatic approach with regard to policy (see also Seng, 2001). Inter alia, it can be argued that the common law was designed to resolve disputes –​often between the ruler and the people –​and that its intention is to limit the power of the state or even that, historically, the common law embodies an inherent distrust of government (Seng, 2001). Whereas in civil law jurisdictions’ views on the separation of powers it is believed that judges should only administer the law, the common law’s views led to a “system of checks and balances,” meaning that the judiciary checks on the legislator and the administration. The common law’s development comes primarily from the judge (Dainow, 1966; Zweigert & Kötz, 1998). Despite the influence of the English Enlightenment thinker Bentham (an advocate for the codification and reform of the common law), the common law was not affected by the Enlightenment ideal of a law of reason, nor was there a need to “unify” the law by codification (because the common law had already been unified from a very early stage on) (Zweigert & Kötz, 1998).12 Comparing the two systems,13 it can be said that civil law countries are more comfortable with a centralized and activist government (Mahoney, 2001), whereas common law countries are characterized by a distrust of government and the reliance on a strong judiciary that protects the individual freedoms of the citizens. Compared to common law countries, civil law countries generally have more legal formalism, lower judicial tenure14 and lower constitutional acceptance of case law (La Porta et al., 2008).The last point also illustrates another important difference between common and civil law, which is the acceptance in common law jurisdictions of the judge as a lawmaker or –​as La Porta et al. (2008) call it –​“an independent source of legal change separate from Parliament,” whereas in civil law countries it is believed that, due to the separation of powers, judges should only administer the law (see also Dainow, 1966).The attitude of the judge is also different; civil law judges tend to be active in truth-​finding (their attitude is inquisitorial; they question and advise parties), whereas common law judges are rather passive as it is believed that the truth should come to them (via the evidence presented; their attitude is adversarial). This of course fits with the idea that the civil law judge is a “civil servant,” whereas the common law judge is expected to take a more distant role (Zweigert & Kötz, 1998). Civil law countries also show more commitment to social rights than common law countries (so there is more labor regulation in civil law countries) (Ben-​Bassat & Dahan, 2008). Freedom of contract is balanced against social/​public considerations in the French and German civil law systems, unlike their common law counterpart (Ahlering & Deakin, 2007). Also, common law courts have the power to review administrative acts, whereas in civil law countries this is usually reserved to special administrative courts applying administrative law, largely at the discretion of the administration itself (David & Brierley, 1978). Despite these many differences, our model is an attempt to encapsulate the salient similarities and differences between the common and civil law. In doing so we focus on “the method of deciding cases” (see above with regard to the different judicial approaches) of the respective legal systems. We will now try to

Legal Systems and Corruption  37 give a short illustration of these different judicial decision-​making approaches related to the essence of common law and civil law systems by looking at the legal institution of environmental torts in France versus the United States. In France, the fast-​growing environmental regulation had made environmental law overly complex; the adoption of a Code de l’environnement [C. env.] [Environmental Code] (Fr.) (2001) was meant to ensure a more comprehensible implementation (OECD, 1996). The Code de l’environnement [C. env.] [Environmental Code] (Fr.) (2001) consists of eight books in total. Books 2, 3 and 4 deal with the essential elements of the environment (air, water, climate, fauna, flora, landscapes) and the rules regarding pollution. Book 5 deals with environmental damage. In addition to the Code de l’environnement [C. env.] [Environmental Code] (Fr.). (2001), there is an environmental charter enclosed in the constitution, which is considered to be a superior right (see also Huglo, 2020). The most important aspects of environmental litigation are assigned to the administrative courts; however, the civil courts handle disputes between private persons regarding environmental torts for nuisances or pollution. LOI n° 2016–​1087 du 8 août 2016 pour la reconqûete de la biodiversité, de la nature et des paysages [Law No. 2016–​1087 of 8 August 2016 on the recovery of biodiversity, nature and landscapes awarded damages for ecological damage] (Fr.) (2016) awards damages for ecological damage; the rules relating to this claim are found in Article 1246 et seq. of the Code Civil [C. civ.] [Civil Code] (Fr.) (1804). French law moreover offers a long list of regulation directly and indirectly related to environmental torts covering a wide range of categories of persons and situations.This shows that even in a relatively straightforward environmental tort case, judges have to draw on a wide range of regulation, mainly found in the Code de l’environnement [C. env.] [Environmental Code] (Fr.) (2001) (see also Jaluzot & Meiselles, 2009). The common law torts most commonly encountered in the environmental field are nuisance, trespass, negligence and strict liability, which have developed over many centuries via judicial decisions (since 1331; Loengard, 2012). Plaintiffs use negligence and strict liability to claim damages for personal injury from environmental pollution. For invasions of property, plaintiffs use trespass and nuisance actions to address environmental damages (Steinway & Botts, 2007). Since nuisance is the most frequently litigated common law action in environmental litigation in the United States (Steinway & Botts, 2007) and we only use this comparative law example to illustrate the differences between the common law and civil law approaches to environmental law, we will limit our analysis to a short elaboration of environmental law tort claims brought under nuisance in the United States. Based on US case law, nuisance can be defined as activity which arises from the unreasonable, unwarrantable or unlawful use by a person of his own property, working an obstruction or injury to the right of another or to the public, and producing such material annoyance, inconvenience, and discomfort that the law will presume resulting damage. (Steinway & Botts, 2007)

38  Ruben Korsten and Andrew Samuel As such, a nuisance arises whenever a person uses his property to cause material injury or annoyance to a reasonable neighbor (Steinway & Botts, 2007). Plaintiffs have used private nuisance actions to claim compensation and force polluters to stop their interference with their private property (Steinway & Botts, 2007). A state (and the public) may also bring a public nuisance action as an exercise of its police powers if there are damages, interference or inconvenience to the public. Some notable judicial decisions in this regard are Carter v. Chotiner (Cal. 1930), where the court stated that the pollution of water is a public nuisance; New Jersey v. NewYork (S. Ct. 1931), where it was decided that dumping garbage into the ocean or waters of the United States constituted a public nuisance; Selma Pressure Treating Co., Inc. v. Osmose Wood Preserving Co. of America (Cal. App. 1990), in which the court held that anyone who contributes to a nuisance is responsible for its discontinuance and compensations; and Village of Wilsonville v. SCA Services, Inc. (Ill. 1981), in which the court found that the prospect of future damages also constituted a nuisance and hence ordered a site clean-​up (see also Buck, 2007). Recapitulating, regulators in common law jurisdictions will sometimes need to seek clarification from the courts before enforcing the law, whereas within civil law jurisdictions regulators will always be able to independently determine whether or not the law has been violated by looking at the regulation. The common law judge directly creates rules of law, which is historically an important part of his function. The civil law judge, when interpreting the law, will focus on the policy of the legislator and the rationale of a rule of law (see Dainow, 1966; Cooper, 1950).15 In the example, the common law shows the evolution of the environmental torts through precedent and the crucial role of the court in clarifying/​shaping the law, whereas the civil law judge has to draw on the vast body of legislation and the words/​intent of the legislator. Meiners and Yandle (1992) argue that judge-​made law in the United States provides more environmental protection for water and the rest of the environment than the whole regulatory process. They argue that the common law judicial decisions produce more sensible principles than legislation, due to the special interests and lack of competitiveness involved with the latter. Another issue is the flexibility of the law: it has been argued that the common law judge-​ made law is more flexible than legislation, especially when courts have the task of filling in open norms (Posner, 2014; Germani et al., 2007).

The Model Consider an economy with a representative who chooses its quantity to maxi1 mize profits: π (q ) = q − q 2 . The production of q generates a negative exter2 nality (harm) whose total external cost is qθ so that the marginal social cost is θ > 0. The parameter θ is a random variable that is uniformly distributed between [l , u ] and is initially unknown to all players. Thus, although θ is a

Legal Systems and Corruption  39 random variable, the firm’s choice of q determines the level of external harm so that the level of harm is endogenous. We assume that l < u < 1 without loss of generality. In the absence of any regulation, a profit-​maximizing firm will choose qL = 1 where L represents the laissez faire level of production, which generates the maximal harm θ . Since this level of production is not socially optimal, there is potential for welfare-​enhancing regulation. If θ were known, then the first best level of q , which we define as qS , maximizes π (q ) − θq .That is, qS = 1 − θ < qL. However, since θ is unknown, regulating the firm’s output depends on the institutional context. The institutional structure that regulates the firm consists of three entities: the legislative branch (“the legislator”), the executive branch (“the regulator”) and the judiciary (“the court”). That is, regulation of the firm’s output is jointly conducted by both the legislator and the courts. The role of the court is to, based on legislation and/​or precedent, determine the statutory levels of production that are either prohibited or permitted given the court’s knowledge of θ . The regulator then takes this threshold as given and then enforces the “law” honestly (i.e., follows the law accurately) if incorruptible, or dishonestly if it can be bribed. The law itself is not homogeneous but depends on the legal origin or legal system, that is, whether it is a civil or common law system. In what follows in mathematical notation we will denote the common law with a subscript M and correspondingly the civil law with a subscript V . In common law jurisdictions, the court does not set a single threshold but rather determines a threshold of no liability below which a firm’s output is acceptable, and a threshold of strict liability above which its output is always sanctioned. In between is a region of ambiguity. In this region, the regulator may take the firm to the court, which investigates the case and determines the actual marginal cost θ , which in turn will determine the acceptable quantity. In a civil law system, the legislator sets a single threshold above which the firm’s quantity is sanctioned and below which it is permitted. Formally, in common law jurisdictions, the court sets the threshold of permissiveness λ ∈ (0,1) below which the firm’s quantity is permitted, and a restrictive threshold µ > λ above which the firm’s quantity is prohibited. In contrast, in civil law jurisdictions, the legislator sets a threshold τ ∈ (0,1). A firm that produces q ≥ τ is sanctioned, but not for any q < τ. This assumption follows the general idea that civil law systems tend to be rule-​based and designed to implement the policies of the ruler. Taking these legal thresholds or statutes as given, an incorruptible regulator then enforces them honestly. Specifically, the regulator observes the firm’s equilibrium quantity q * and sanctions any firm in accordance with the statutes and, by extension, the relevant legal system. In common law jurisdictions, a firm that chooses q * ≤ λ is not punished and a firm that chooses q * ≥ µ is sanctioned f . Cases in which firms choose q * ∈ ( λ, µ ) go to trial. At trial, the

40  Ruben Korsten and Andrew Samuel court investigates the matter (at some cost), and this investigation reveals the true value of θ (with certainty). If it turns out that q * θ , the firm is sanctioned; otherwise it is not. In contrast, in civil law jurisdictions, the regulator observes whether or not q * ≥ τ . If it is, the firm is sanctioned in accordance with the penalties specified in civil law jurisdictions. Given this setup, in common law jurisdictions, an efficiency-​ minded court will never choose λ < 1 − u since doing so would incentivize firms to choose a sub-​optimal level of output even if θ = u. Similarly, the court would never choose µ > 1 − l since that would incentivize the firm to produce too much, even if θ = l. Hence, a common law court will choose λ ≥ 1 − u and µ ≤ 1 − l. A court’s rules are said to be “broad” if λ > 1 − u or µ < 1 − l; otherwise they are narrow. As a benchmark, we only allow the court to announce narrow rulings. Hence, an efficiency-​oriented common law court will choose λ = 1 − u and µ = 1 − l. Civil law courts are also efficiency-​minded. However, they are restricted to only applying a single threshold. Accordingly, a civil law legislator will choose the ex ante level of harm as its threshold. Thus, the threshold in a civil law court is τ = 1 − E (θ) . It is important to recognize here that in this system once the legislature specifies the rules, the courts in some ways are merely functionaries that implement or help to enforce that rule. In contrast, common law courts can shape the law by learning the threshold over time. This distinction mirrors Tullock, Owens and Rowley’s (1997) understanding that civil law judges play a more “functionary” role within the legal system, in contrast to common law judges. The firm’s profits depend on the legal system in which it operates. A firm’s profits in a civil law jurisdiction are:  1 2 q − 2 q if q ≤ τ πV (q ) =  q − 1 q 2 − qE (θ) if q > τ  2 A firm’s profits in a common law jurisdiction will depend on the expected fine if it chooses a quantity in the ambiguous region. Specifically, if q ≤ λ = 1 − u, the firm is not liable and hence the regulator will not take the firm to trial. If q > µ = 1 − l , the firm is automatically held liable and the regulator sanctions the firm qE (θ) . If q ∈ ( µ, λ ), then if the regulator goes to trial, with probu

ability

∫ θ� dF (θ), the court discovers that the firm’s quantity is above 1 − θ , so

1− q

that it is fined qθ . Solving this, the expected penalty received (conditional on the firm choosing a value in the ambiguous region) is: I (q, l , u ) = q

u 2 − (1 − q )2 . 2 (u − l )

Legal Systems and Corruption  41 Here it is worth noting that I (q, l , u ) could also represent the expected damages awarded to a plaintiff. Using this expression, the firm’s profits in common law jurisdictions are:  1 2 q − 2 q if q ≤ λ   1 π M (q ) = q − q 2 − I (q, u, l ) if q ∈ ( λ, µ )  2  1 2 q − 2 q − qE (θ) if q > µ.  The regulator’s payoff in each regime is simply the expected fine. Thus, in the civil law system it is qE (θ) if q > τ or 0 otherwise. In the common law system, it is I (q, u, l ) if q ∈ ( λ, µ ) or qE (θ) if q ≥ µ . The timing of the game is as follows: • • •

•

Taking the legal environment as given, the firm first chooses quantity q. Given q and the costs of going to trial c P , the regulator then decides whether or not to go to trial. In common law jurisdictions, the court (upon investigation) determines the true θ and, therefore, whether or not q is inefficient. If it is above the socially efficient quantity, it is fined qθ . In civil law jurisdictions, the court simply imposes a fine fV = E (θ) q . The game ends and all penalties and so on are transferred.

Equilibrium Analysis We begin by analyzing the behavior of the regulator and the court given some q. Given this, it is individually rational for the regulator to go to trial only if I (q, l , u ) ≥ 0. 

(IR_​M)

Under a civil law system, the court can sanction a firm that chooses a quantity above τ = 1 − E (θ). This implies that its revenue from the fine is qE (θ ) if the firm chooses a quantity above τ. Accordingly, in civil law jurisdictions, it is individually rational for the regulator to go to trial only if qE (θ) ≥ 0. 

(IR_​V)

First, recognize that for any q ≥ λ , the regulator’s IR constraints are satisfied in both legal systems. Thus, going to trial is always a credible threat. Given this behavior of the court and the regulator, we now characterize the firm’s behavior.

42  Ruben Korsten and Andrew Samuel Proposition 1: Within each legal system the firm’s equilibrium behavior is as follows: •

In civil law jurisdictions, the firm’s equilibrium quantity is qV* = τ.

•

In common law jurisdictions, the firm’s equilibrium quantity is λ if u − l ≤ 1 − u qM* =  qA ∈ ( λ,1 − E (θ)) if u − l > 1 − u 1 where qA maximizes q − q 2 − I (q, l , u ) . 2

•

The equilibrium q is always higher in civil law jurisdictions than in common law jurisdictions.

Proof. The firm’s equilibrium quantity in civil law jurisdictions follows directly from the profit function.The equilibrium quantity in common law jurisdictions follows directly from Parameswaran (2018), which we outline partially here. For 1 ease of exposition, we denote π λ = λ − 1 / 2λ 2 and π A = qA − qA2 − I (qA , l , u ) . 2 1 Then it is straightforward to show that if u − l > 1 − u then q − q 2 − I (q, l , u ) 2 1 is strictly increasing at λ , so that the quantity that maximizes q − q 2 − I (q, l , u ) 2 1 is strictly greater than λ . If u − l < 1 − u , then q − q 2 − I (q, l , u ) is decreasing 2 1 at λ , so that λ is the constrained maximum of q − q 2 − I ( q, l, u ). Finally, it 2 1 2 is also straightforward to show that q − q − I (q, l , u ) is always decreasing at 2 1 − E (θ) so that the profit-​maximizing quantity is always less than qE. We refer the reader to Parameswaran (2018) for further details. The previous proposition shows that in the absence of bribery, the equilibrium quantity is always lower in common law jurisdictions than in civil law jurisdictions. If τ is the ex ante efficient quantity, then civil law systems will always yield a higher level of ex ante efficient outcomes, whereas in common

Legal Systems and Corruption  43 law jurisdictions there will always be under-​provision of q in an ex ante sense. This argument is consistent with Tullock et al. (1997), who for very different reasons argue that the common law is often less efficient than the civil law.

Model with Bribery Consider a case where there is further delegation in the sense that enforcement is done by a lower-​level functionary, an “official,” who must observe and report q if the quantity is too large. The functionary has limited discretion and can only sanction a firm that has unambiguously violated the regulatory standard. Thus, in civil law jurisdictions the official can sanction any firm whose quantity q > τ. Whereas in common law jurisdictions the official can sanction any firm that has chosen q > µ. Under both systems, the official is corruptible and may accept a bribe in exchange for not sanctioning the firm. We assume that corruption does not occur during the trial process in the common law and that when q ∈( µ, λ ), the official does not have authority to sanction the firm summarily, but must report it to her superiors in order to take the firm to trial. We believe that this assumption is reasonable for the following reasons.When q is in the ambiguous region, then there are two possible ways in which bribery can occur. First, bribery may occur on a grand scale wherein the regulator, the courts and the firm all collude.While such grand corruption may be possible, it would be much harder to coordinate. Thus, we believe that this possibility is highly unlikely. Alternatively, the regulator can attempt to extract a bribe from the firm. But, in this case, the firm that has chosen q need not yield to that threat since presumably it can also appeal to the court. In fact, any long-​ lived firm would likely prefer to have the law clarified since that would allow it to choose q = θ in the future. Thus, such threats would not be credible and unlikely.That said, in the conclusion we explore the implications of this type of corruption for our results. First, consider the case of a firm operating in a civil law jurisdiction. If the firm chooses to be honest, it chooses q = τ and receives a payoff of 1 τ − τ2 ≡ πτ . 2 Instead, if it chooses to be corrupt, it chooses q > τ and pays a bribe b. We assume that the official’s bargaining power is α ∈ (0,1) and that the official can observe q. Thus, the official demands a bribe b = αqE (θ) and reports that q ≤ τ. A firm that bribes, therefore, chooses a quantity to maximize: 1 q − q 2 − αqE (θ) ≡ πB (q ) . 2

44  Ruben Korsten and Andrew Samuel The solution to this maximization problem is qB ( α ) = 1 − αE (θ) . Thus, a firm’s decision in civil law jurisdictions is solved by the following maximization problem: max

q = {τ ,qB (α )}

{π , π ( q τ

B

B

(α ))} .

In common law jurisdictions, (a lower-​level) official is authorized to sanction the firm if q > µ. Instead, if q < λ the official may not sanction the firm, whereas if q ∈( µ, λ ) , the official has no discretion and must report the firm to her superiors, who take the firm to trial.16 Accordingly, a firm either decides to be honest and chooses a quantity < µ as characterized in proposition 1 , or chooses q ≥ µ and pays a bribe. We assume that the official’s bargaining power is α and that she demands a bribe b = αqE (θ) . Hence, a corrupt firm chooses q to maximize 1 q − q 2 − αqE (θ) , 2 subject to q ≥ µ. Clearly, conditional on being corrupt, a firm will either choose q = µ (when the above constraint is binding) or qB ( α ), which is identical to the expression in civil law jurisdictions. Indeed, note that the key difference between bribery in civil law and common law jurisdictions is that bribery occurs under stricter constraint in common law jurisdictions. Accordingly, to simplify notation, we will refer to the corruption profits as πB (qB ( α )) for both the common and

the civil law. Further, in a sense, this aspect of our model reflects the idea that civil law regulation would create more scope for bribery than the judge-​made common law (La Porta et al., 1998). However, as we will show, even if in this sense there is more scope for bribery in civil law jurisdictions, in equilibrium bribery need not be more likely to occur in civil law jurisdictions. In common law jurisdictions, a firm will either choose a quantity q = λ or qA depending on whether 1 − u > u − l and remain honest, or choose q = µ , or B ( α ) and be corrupt. Thus, in common law jurisdictions, a firm solves the following problem:

Legal Systems and Corruption  45 max

q = {λ ,qA ,µ ,qB (α )}

{π , π λ

A

}

, πB ( µ ) , πB (qB ( α )) .

Using the above framework, we now characterize the equilibrium decisions of the firms in the following proposition. Proposition 2: Bribery encourages firms to overproduce. Specifically, • •

In civil law jurisdictions, there exists an αV ∈ (0,1) , such that the firm chooses a quantity q > µ if α < αV ; otherwise it chooses to produce τ units. In common law jurisdictions, there exists an α M ∈ (0,1) such that the firm chooses a quantity greater than or equal to µ if α < α M ; otherwise it chooses to produce λ, or qA units.

Proof. The claim for the civil law follows directly from proposition 2. The claims concerning the common law follow from both proposition 2 and claim 2 as follows. First, if in civil law jurisdictions the firm does not bribe, it will choose q = τ and receive a payoff of π ( τ ) . If it bribes, its payoff is

πB (qB (α )) =

2 1 1 − αE (θ )) , ( 2

where it is clear that ∂πB (qB ( α )) ∂α

< 0,

with πB (qB ( α )) → 1 / 2 (the maximum possible payoff) as α → 0. Further, it is

strictly less than π τ as α → 1. Thus, there exists an αV ∈ (0,1) such that q = τ is profit-​maximizing if α > αV , and q = 1 − αE (θ) is maximizing if α < αV . Second, consider a firm in a common law jurisdiction. For such a firm we know from proposition 1 that, if it is honest, it chooses either qA (if u − l > 1 − u ) or λ (if u − l ≤ 1 − u ). Whereas, if it is corrupt, it chooses qM . It is straightforward to show that as α → 1, the profit-​maximizing quantity subject to q ≥ µ (in which bribery is possible) is q = µ . Thus, the firm’s profits here are 1 µ − µ 2 − µE ( θ ) , 2

46  Ruben Korsten and Andrew Samuel where the right hand side of the previous inequality is the payoff from corruption at the constrained optimum under corruption: q = µ. Next observe that at α = 1 , the unconstrained maximizer of the expression 1 q − q 2 − qE (θ) is at q = 1 − E (θ) . If u − l ≤ 1 − l , then a firm that does not 2 engage in corruption chooses q = λ . A straightforward calculation shows that 1 1 1 λ − λ 2 > qB (α ) − qB ( α )2 − qB (α ) E (θ) > µ − µ 2 − µE (θ) . 2 2 2 because qB ( α ) is the unconstrained maximizer of q − 1 / 2q 2 − αE (θ) (and also because profits are decreasing for all q ≥ λ ). By similar reasoning, if u − l > 1 − l , the honest firm chooses qA and receives the following payoff from honesty: 1 π A > µ − µ 2 − µE ( θ ) . 2 Thus, at α = 1 , the firm prefers to be honest. As α → 0 , the payoffs from corruption approach 1/2 , which is strictly greater than max {π A , π λ } . Hence, corruption is preferred at α = 0.

Finally, as noted earlier, πB (qB ( α )) is decreasing in α . Hence, there exists

an α M such that the firm’s profits are maximized by choosing q = 1 − αE (θ) or µ if α < α M ; otherwise it chooses either qA or λ according to proposition 1.

Figure 4.1 Bribery in both systems (l = .4, u = .7, α = .20 ) .

Legal Systems and Corruption  47

Figure 4.2 Bribery only in common law jurisdictions (l = .4, u = .7, α = .35) .

The ideas of proposition 1 are illustrated in the following figures, which present the payoffs under each regime with and without bribery. In both figures, πi (q ) is measured on the y axis for i = {V , B, M } . In Figure 4.1, we observe that in the absence of bribery, firms choose τ in civil law jurisdictions and λ in common law jurisdictions. With bribery, note that the payoffs are identical for any q ≥ µ . Further, the payoff at q = .8 (with bribery) is higher. Hence, the behavior is identical in both regimes in the sense that firms in both regimes choose the same payoff and bribe. In Figure 4.2, a firm in a civil law jurisdiction receives a higher payoff from remaining honest (at q = τ ), but a firm in a common law jurisdiction will prefer to bribe. Importantly, this arises because the civil law is more permissive, as it does not sanction all firms that choose a q > λ , whereas the common law does. Hence, firms find it more attractive to bribe in common law jurisdictions. The analysis therefore shows that although there is more scope for bribery in civil law jurisdictions, in equilibrium there may be less bribery in civil law jurisdictions than in common law jurisdictions. To fully analyze the implications of this result, we first need to understand the firm’s decisions in each regime.The firm’s decisions in civil law jurisdictions are fully characterized by proposition 2. However, the profit-​maximizing quantity in common law jurisdictions is more complicated. First, we analyze the conditions under which q ≥ µ is binding. A straightforward calculation shows that qB ( α ) = 1 − αE (θ) > µ ⇐⇒ α ≥

1− µ (1) E (θ )

48  Ruben Korsten and Andrew Samuel If this condition is satisfied, the firm chooses q = µ , conditional on choosing the dishonest quantity, or it chooses the honest quantity of qA or λ . Next, consider the case where condition (1) is violated so that the corrupt quantity is qB ( α ). In this case, corrupt profits are 1 πB (qB (α )) = (1 − αE (θ))2 . 2 Whereas honest profits are either π λ or π A (depending on whether 1 − u > u − l ). Then a straightforward calculation shows that profits are higher under bribery only if

α < α1 ≡

1 − 2max {π λ , π A } E (θ )

.

 1 − µ  If α < min α1 ,  , then the firm chooses the dishonest quantity.To study  E (θ )  behavior for other values of α , we need to consider two cases. If α1
α1 . E (θ)   Further, it will always choose to be honest for any α >

1− µ , because if E (θ )

honesty is preferred over the unconstrained profit-​maximizing quantity under bribery, it is also preferred to the constrained profit-​maximizing quantity under bribery. 1− µ  1− µ < α1 . In Next, we consider the case where α1 ∈  , α1  , because E (θ ) θ E  ( )  this case, the firm’s corrupt profits are

πB ( µ ) , because the constraint that q ≥ µ is not binding. Thus, the corrupt quantity is preferred if 1 µ − µ 2 − αµE (θ) > max {π A , π λ } . 2

Legal Systems and Corruption  49 Solving for α yields 1 2   µ − µ  − max {π A , π λ } 2 α< ≡ α2 . E (θ ) Whether corruption or honesty is preferred now depends on these threshold values of α. Thus, we first establish the following claims concerning the rela1− µ tionship between α1 , α 2 and . E (θ ) Claim 1: The threshold values of α possess the following relationships: (1) For any parameter values, α 2 ≤ α1 . (2) If max {π λ , π A } > 1 µ 2 , then α1 < 1 − µ . 2 E (θ ) 1 2 1− µ (3) If max {π λ , π A } < µ , then < α 2 < α1. 2 E (θ ) Proof: First note that the unconstrained dishonest profits are

πB (qB (α )) > πB (µ ) 1− µ , where the two are equal. Further, πB (qB ( α )) is convex E (θ ) (and strictly decreasing) in α . Since

except at α =

α1 : max {π λ , π A } = πB (qB (α )) whereas

α 2 : max {π λ , π A } = πB (µ ) it follows that α1 > α 2 , except at α = Next, note that 1 2 1 µ = (1 − αE (θ))2 |α = (1− µ)/E (θ) . 2 2

1− µ (where they are equal). E (θ )

50  Ruben Korsten and Andrew Samuel Hence, since α1 solves max {π λ , π A } = max {π λ , π A } >

1 2 µ 2

then

α1
α 2 >

1 2 µ 2

1− µ . E (θ )

1 2 µ . 2 Using the above insights we can now make the following claim regarding quantity under the common and civil law legal systems. For convenience, Figure 4.3 presents the case with max {π λ , π A } ≥

Figure 4.3 α 2 < α1
α . 2 A   Further, in this case, α M = α 2 . If

α1
α1 the firm chooses the honest quantity of either λ or qA for any  1− µ α ∈  α1 ,  . Of course, this then implies that it will also choose the honest  E (θ )  1− µ quantity for any α > since profits are decreasing in α. E (θ ) Next, consider the case where

α1 ≥

1− µ . E (θ )

1− µ  1− µ , α2  , the firm chooses q ( α ) . If α ∈   E (θ )  E (θ ) the firm chooses µ (and bribes). If α > α 2 , then the firm chooses the In this case, if α
max {π λ , π A } then αV < α1 , whereas if

π τ < max {π λ , π A } then αV < / < α 2 . Finally, recall that π τ > max {π A , π λ } . Hence if α1


1− µ , E (θ )

then αV may be above or below α M . The preceding results present the main arguments of this chapter. It shows that although there is more scope for bribery in civil law countries, in the sense that bribery occurs for any q > µ > τ , it is also more likely to occur for a larger range of α than in common law countries. Specifically, as we

Legal Systems and Corruption  53 showed above, αV is in many cases smaller than α M , so that the threshold below which bribery occurs is larger in common law systems. We elaborate on the implications of this in the conclusion. It is often argued that bureaucrats have more control rights and power in civil law than in common law countries. The higher levels of corruption in civil law countries have been attributed to these differences. However, our result shows that there is more bribery in civil law countries only if the bargaining power of its officials is lower than that of those in common law countries. Thus, the fact that bureaucrats may have more power in civil law countries (i.e., a higher α ) is not likely to be the explanation as to why there is more corruption in the form of bribery in those countries. Importantly, it also suggests that without including measures of α , empirical studies linking corruption levels or bribe levels to a country’s legal system are not robust. Next, the previous result also implies that the average bribe need not be smaller in a common law system. Consider an interpretation of our framework in which α is distributed over the unit interval by some density function f (.) . If αV < α1 , then the probability of bribery is F ( αV ) in civil law jurisdictions, and F ( α1 ) in common law jurisdictions. The conditional mean bribe will therefore be lower in civil law jurisdictions. Thus, not only the incidence but also the magnitude of the bribe will be lower in civil law jurisdictions.

Conclusion Legal origins theory tries to draw a causal pathway between a country’s legal origins and the quality of its government. It draws on observed correlations to argue that compared to civil law countries, those with common law systems have better quality of government, such as lower corruption and more efficiency (see a.o. La Porta et al., 1997, 1998, 1999, 2008;Treisman, 2000; Djankov et al., 2002; Mahoney, 2001; Ben-​Bassat & Dahan, 2008). It also observes that countries with a strong bureaucracy are also likely to see more corruption (Djankov et al., 2002; La Porta et al., 1999). Our framework suggests that interpreting these correlations as a causal pathway between a legal system and corruption is not straightforward. In general, bribery is less likely if a bureaucrat can extract more from a citizen, because the citizen is less likely to pay a larger bribe.This extractive power is modeled as the bureaucrat’s bargaining power ( α ) in our framework.We show that bribery occurs for a larger range of this extractive (bargaining) power in common law countries than in civil law countries. Thus, in this sense, corruption is more likely to occur in common law jurisdictions. We show that strong and unconstrained bureaucracy does not necessarily translate into more corruption. Assuming that a stronger bureaucracy possesses a higher bargaining power ( α ), our model shows that it need not result in more corruption (proposition 3). In fact, a strong bureaucracy is likely to result in less corruption in most cases.

54  Ruben Korsten and Andrew Samuel Third, we show that the size of the bribe need not be lower in a common law country. It has been argued that corruption should be measured not only by the incidence but also by the the magnitude of bribery (Méndez & Sepúlveda, 2010). If so, then our model shows that in many circumstances, the average observed bribe should be higher in common law countries. Fourth, the legal origins literature also argues that common law countries generally implement more efficient outcomes. The logic of this argument is based on the idea that the common law “searches” for the efficient rule through judicial decisions. Within our framework, what this means is that if a case goes to trial, judges learn the true value of θ , so that in subsequent periods the efficient quantity can be chosen. Whereas, in civil law jurisdictions, there is no learning and the outcome is always the ex ante efficient outcome τ. However, our framework shows that this logic is misleading in the presence of bribery. Specifically, bribery can prevent the common law judges from learning the true magnitude of the external cost and indeed cause excessive harm. Thus, our framework questions the validity of such arguments. Our findings also provide insights into further avenues for thinking about the relationship between bribery and a legal system.As other papers on the common law legal system show, the presence of ambiguous laws creates incentives for the regulator to settle (“pre-​trial”) out of court with a defendant (Parameswaran & Samuel, 2019). In some ways, a pre-​trial settlement is not unlike a preemptive bribe (e.g., Samuel, 2009).17 Since pre-​trial settlement is widespread in the common law tradition, it is arguably a substitute type of “bribery,” so that comparing corruption across the two systems is not straightforward. Similarly, our chapter focused primarily on identifying a causal pathway between bribery and a country’s legal system. However, as many others have noted (see a.o. Basu, 2018), corruption can manifest itself in other ways besides as a bribe. For example, judges themselves may be bribed. Or the legal system itself may become corrupt and subject to special interests. As our findings suggest, even in the relatively straightforward instance of corruption as bribery, the relationship between legal systems and corruption is not straightforward. These results show that there is need for formal modeling to better interpret the correlations observed between legal systems and corruption. Specifically, despite the sizable empirical and theoretical literature connecting corruption to legal systems, there is little formal modeling on this topic. In part, this is because, to date, there has been no way to formally model the two legal systems (Guerriero, 2016). However, as our analysis has shown, formal modeling can be useful because it can clarify the empirical findings in the “legal origins” literature,18 suggest new causal pathways and identify questions that remain unanswered.

Notes 1 Or the official may be a judge resolving a dispute. We discuss the angle of judicial corruption in the Conclusion. 2 We call this “the method of deciding cases” (Dainow, 1966) of the respective legal systems.

Legal Systems and Corruption  55 3 Or rather, in most cases, the accused will not find it beneficial to contest the penalty in court. 4 See Dainow (1966), Zweigert and Kotz (1998), Cooper (1950). 5 “If we may generalize, the European is given to making plans, to regulating things in advance, and therefore, in terms of law, to drawing up rules and systematizing them. He approaches life with fixed ideas, and operates deductively.The Englishman improvises, never making a decision until he has to: ‘we’ll cross that bridge when we come to it’ … he is an empiricist. Only experience counts for him; theorizing has little appeal; and so he is not given abstract rules of law. Convinced, perhaps from living by the sea, that life will controvert the best-​laid plans, the Englishman is more at home with case-​law proceeding cautiously step-​by-​step than with legislation which purports to lay down rules for the solution of all future cases” (Zweigert & Kotz, 1998, 70). 6 And it is argued here that this happens frequently, as common law statutes tend to be difficult to apply and rules can sometimes be ambiguous, in the sense that they do not codify every likely scenario in which the law may come into effect (Zweigert & Kotz, 1998). 7 The law of France and countries that –​through conquest and colonization by France –​established civil codes based on the Napoleonic Code, that is, most countries of continental Europe and their former colonies. 8 According to La Porta et al. (2008), the reason for this is that these two legal systems (French civil law versus common law) include the largest samples of countries and because they would have the most distinct approaches to law. 9 See a.o. David and Brierley (1978), Dainow (1966) and Zweigert and Kotz (1998). 10 Referring to the application of deductive reasoning to arrive at a conclusion based on two propositions that are asserted or assumed to be true. 11 Nowadays, the common law is the law of England and its former colonies (e.g., the United States –​except for Louisiana on a state level; Australia; and India –​except for Goa and Pondicherry), as the legal system spread through conquest and colonization by England. 12 Yet, with regard to statutory drafting, common law statutes tend to be much more complex and lengthy than their civil law counterparts and less easy to apply (Zweigert & Kotz, 1998). 13 Obviously with the necessary caution, since the legal systems are moving closer to each other: in common law countries there is an increase of legislative action and, on the other hand, statute is losing the preeminence it once held in civil law jurisdictions (Zweigert & Kotz, 1998; see also David & Brierley, 1978; Dainow, 1966; Zweigert & Kotz, 1998; La Porta et al., 2008). 14 Despite the impression created by La Porta, Lopez-​de Silanes and Shleifer (2008), this only goes for the constitutional/​supreme courts (the highest courts): virtually all jurisdictions appoint judges for life. However, in civil law countries, constitutional court judges usually serve fixed terms (David & Brierley, 1978). 15 But, in a common law country, the judge must give effect to a “narrow” law, whereas judges in civil law can sometimes interpret the law creatively where the law is broadly stated (see above and Cooper, 1950; Dainow, 1966). 16 Note that if q > μ and the functionary’s decision is challenged by the firm, the firm will be fined qE(θ) by the court regardless. 17 A preemptive bribe is paid by a violator to prevent the investigator from carrying out a full investigation, which is not unlike a pre-​trial settlement.

56  Ruben Korsten and Andrew Samuel 18 We have tried to give a nuanced illustration of (some) legal practices in the world. By no means do we wish to join the (rather flawed –​in our eyes) discussion of whether “common law” is better than “civil law” (in terms of combating corruption) or vice versa. Rather, we look at the approaches to judicial decision-​making on their merits in order to make sound policy statements.

References Ahlering Beth and Deakin. Simon Labor regulation, corporate governance, and legal origin: A case of institutional complementarity? Law & Society Review, 41(4): 865–​908, 2007. Aquinas, Thomas Gerald, Phelan, Bernard Kenny, J and Ignatius Theodore Eschmann. De Regno: On Kingship. Divine Providence Press, 2014. Basu, Kaushik. The Republic of Beliefs. Princeton University Press, 2018. Beatrice, Jaluzot and Meiselles, Michaela. Civil law consequences of corruption and bribery in france. In The Civil Law Consequences of Corruption (pp. 225–​238). Nomos Verlagsgesellschaft mbH & Co. KG, 2009. Ben-​Bassat Avi and Dahan, Momi. Social rights in the constitution and in practice. Journal of Comparative Economics, 36(1):103–​119, 2008. Buck, Stuart. The common law and the environment in the courts. Case Western Reserve Law Review, 58:621, 2007. Carter v. Chotiner, Cal. 1930. Code Civil [C. civ.] [Civil Code] (Fr.)., 1804. Code de l’environnement [C. env.] [Environmental Code] (Fr.)., 2001. Coffee Jr., John C Future as history: The prospects for global convergence in corporate governance and its implications. Northwestern University Law Review, 93:641, 1998. Cooper, Thomas Mackay. The common and the civil law: A Scot’s view. Harvard Law Review, 63(3):468–​475, 1950. Dainow, Joseph The civil law and the common law: Some points of comparison. American Journal of Comparative Law, 15:419, 1966. David, Rene and Brierley, John EC. Major Legal Systems in the World Today: An Introduction to the Comparative Study of Law. Simon and Schuster, 1978. Djankov, Simeon, Porta, Rafael La Florencio, Silanes, Lopez-​de, and Shleifer, Andrei. The regulation of entry. Quarterly Journal of Economics, 117(1):1–​37, 2002. Dreher, Axel, Kotsogianni, Christos and McCorriston, Steve. How do institutions affect corruption and the shadow economy? International Tax and Public Finance, 16(6):773–​ 796, 2009. Gelter, Martin and Grechenig, Kristoffel R. History of law and economics. MPI Collective Goods Preprint, 2014(5), 2014. Germani, Anna Rita et al.The environmental enforcement in the civil and the common law systems: A case on the economic effects of legal institutions. Dipartimento di Scienze Economiche, Matematiche e Statistiche, Universita’di …, 2007. Guerriero, Carmine. Endogenous legal traditions. International Review of Law and Economics, 46:49–​69, 2016. Huglo, Christian. The environment and climate change law review: France. In Environment and Climate Change Law Review: 5th Edition (pp. 71–​81), The Law Reviews Law Business Research Limited, 2020. Johnson, Simon, Porta, Rafael La, Florencio, Silanes, Lopez-​de and Shleifer. Andrei Tunneling. American Economic Review, 90(2):22–​27, 2000.

Legal Systems and Corruption  57 Loengard, Janet S. Common law and custom: Windows, light, and privacy in late medieval england. In Laws, Lawyers and Texts (pp. 279–​298). Brill, 2012. LOI n° 2016-​1087 du 8 août 2016 pour la reconquête de la biodiversité, de la nature et des paysages [Law No. 2016-​1087 of 8 August 2016 on the Recovery of Biodiversity, Nature and Landscapes awarded damages for ecological damage] (Fr.). (2016). Mahoney, Paul G. The common law and economic growth: Hayek might be right. Journal of Legal Studies, 30(2):503–​525, 2001. Meiners, Roger E and Yandle, Bruce. Constitutional choice for the control of water pollution. Constitutional Political Economy, 3(3):359–​380, 1992. Méndez, Fabio and Sepúlveda, Facundo. What do we talk about when we talk about corruption? Journal of Law, Economics, & Organization, 26(3):493–​514, 2010. Naidoo,Vinothan. The politics of anti-​corruption enforcement in South Africa. Journal of Contemporary African Studies, 31(4):523–​542, 2013. New Jersey v. New York, S. Ct. 1931. OECD. OECD Environmental Performance Reviews: France. OECD, 1996. Parameswaran, Giri and Samuel, A. The evolution of the common law with strategic plaintiffs. 2019. http://​gpara​mes.sites.haverf​ord.edu/​wp-​cont​ent/​uplo​ads/​2022/​04/​ Parameswar​anSa​muel​_​Evo​luti​onCo​mmon​Law.pdf Parameswaran, Giri. Endogenous cases and the evolution of the common law. RAND Journal of Economics, 49(4):791–​818, 2018. Porta, Rafael La, Silanes, Florencio Lopez-​ de and Shleifer, Andrei. The economic consequences of legal origins. Journal of Economic Literature, 46(2):285–​332, 2008. Porta, Rafael La, Silanes, Florencio Lopez-​de, Shleifer, Andrei, and Vishny, Robert W Law and finance. Journal of Political Economy, 106(6):1113–​1155, 1998. Porta, Rafael La, Silanes, Florencio Lopez-​de, Shleifer, Andrei, and Vishny, Robert W Legal determinants of external finance. Journal of Finance, 52(3):1131–​1150, 1997. Porta, Rafael La, Silanes, Florencio Lopez-​de, Shleifer, Andrei, and Vishny, Robert W.The quality of government. Journal of Law, Economics, Organization, 15(1):222–​279, 1999. Posner, Richard A. Economic Analysis of Law. Wolters Kluwer Law & Business, 2014. Samuel, Andrew. Preemptive collusion among corruptible law enforcers. Journal of Economic Behavior & Organization, 71(2):441–​450, 2009. Selma Pressure Treating Co., Inc. v. Osmose Wood Preserving Co. of America, Cal. App. 1990. Seng, Michael P. Aspects of the common law system in the United States. Common Law World Review, 2:6, 2001. Steinway, Daniel M. and Botts, B. Fundamentals of environmental law. Environmental Law Handbook, 19:1–​66, 2007. Treisman, Daniel. The causes of corruption: A cross-​national study. Journal of Public Economics, 76(3):399–​457, 2000. Tullock, Gordon, Owens, Amanda J., and Rowley, Charles Kershaw. The Case against the Common Law. Number 1. Carolina Academic Press, 1997. Village of Wilsonville v. SCA Services, Inc., Ill. 1981. Zweigert, Konrad and Kötz, Hein Introduction to Comparative Law, Volume 3. Clarendon Press, 1998.

5 Corruption and Optimal Enforcement Ajit Mishra

Introduction Any public policy requiring an agency for implementation is likely to be impacted by corruption. Whether it is tax policy or pollution control or provision of benefits and services of different kinds, the government uses an agency –​ tax officers, pollution inspectors, administrative officials –​to administer and implement the policy.1 In this context, corrupt officials can deviate from their stipulated contractual duties for private gains. Note that this is a generalization of the standard notion of corruption as misuse of public office for private gains. Deviant behavior can take different forms, depending on the context. For example, in the case of tax collection, evasion is sought to be deterred through scrutiny and assessment by tax officers, who can impose fines for underreporting of income. But the officer can let the taxpayer escape these fines, in exchange for a bribe. This obviously encourages tax evasion and lowers tax collections. To consider another example in a different context, take the case of a benefit scheme where only individuals fulfilling certain eligibility criteria are meant to receive the benefit. The corrupt official can accept a bribe from an ineligible individual and award benefits. Consequently, the intended beneficiaries are deprived of their entitlements, and policy objectives are not met.The nature and implications of corrupt practices may change depending on the context, but the common thread is that the policy that is deemed to be optimal in the absence of corruption fails to deliver the intended outcome. One possible response to this would be to take the policy as given and supplement it with various anticorruption mechanisms so that nobody engages in corruption.2 It is possible that some deterrence can be achieved, though it need not be complete. But does the initial optimal policy continue to be optimal in this new scenario? Do we still aim for the same level of progressivity in taxation? Do we insist on strict pollution standards or let low levels of pollution go unsanctioned? Do we still insist on targeting or tend toward universality? The objective of this chapter is to highlight this issue of optimal policy choice in the presence of corruption. The question gets harder because of the multidimensional nature of corruption, which takes place in various forms and in different settings. For DOI: 10.4324/9781003142300-5

Corruption and Optimal Enforcement  59 example, two key forms of corruption are collusion and extortion. Both arise from the misuse of public office for private gains by the government agent, but the clients (individuals, firms) are affected differently. In the case of collusion, both benefit from the corrupt act by the agent, which often involves taking client-​favorable action in exchange for a bribe. In the tax collection example, while colluding with the taxpayer, the agent either underreports or hides evidence of underreporting. This reduces the expected cost of evasion for the individual and leads to more individuals choosing to evade taxes. However, the misuse by the agent can take the form of extortion where the government agent seeks a bribe while threatening the individual or firm with client-​ unfavorable action. The tax officer can threaten to report (falsely) evasion if bribe demands are not met. Here, corruption raises the cost of nonevasion or compliance. In the case of benefit transfers, we have already seen how the agent can collude with the ineligible and award benefits. Similarly, eligible individuals may have to bribe the official to receive the transfers, which they are entitled to. We discuss how these two forms of corruption can have different implications and how anticorruption policies for deterring these are likely to be different.3 Optimal policy choice will depend on whether we are dealing with both these forms of corruption or any one of them and their relative strengths. As the two examples above show, both forms of corruption can be seen in various economic settings, including regulation, enforcement and delivery of public services. We use a simple pollution control setting to study the problem of enforcement. We adopt a commonly used agency framework where the regulator employs an inspector to monitor the firms.4 We discuss how both forms of corruption arise in this setting and how different mechanisms are required to deter. However, these anticorruption mechanisms do not simply restore the original problem (the one without corruption). As suggested in the first paragraph, these have the potential to change the optimal policies (toward pollution) in a fundamental way. It is possible that the regulator will not seek to achieve the same level of pollution control due to the cost of monitoring and fighting corruption. But what is surprising is that it may be optimal to err in the opposite direction and overenforce too! We present an example of such overenforcement. We describe the enforcement problem in the next section. In the next two sections, we introduce the problems of collusion and extortion. We present an example of overenforcement and explain why it might be the optimal policy in response to the problem of corruption. Overenforcement refers to the case of preventing a firm from choosing a level of pollution whose private benefit exceeds the associated social cost.

The Enforcement Problem Consider the following simple problem of enforcement of pollution standards. Firms choose whether to pollute or not, denoted by a ∈ {0,1} where a = 0 refers to zero pollution. Firms benefit from polluting, as they save on abatement

60  Ajit Mishra costs, with benefits (gains) given by g (a ) where g (0 ) = 0, g (1) = g > 0. There is a measure of social harm associated with the act of pollution h (a ) with h (0 ) = 0, h (1) = h . Gains are distributed across some interval [0,G ]. Since our focus is on corruption, to make the enforcement problem interesting, we begin with the case where private benefit never exceeds the social cost, h > G . This means that in the absence of any monitoring cost, the regulator wishes to deter all firms from polluting and implement a = 0. We can refer it as the first best level, a FB . She can do so by a system of inspection, rewards and/​ or sanctions. An inspector audits the firms; each firm is inspected with probability θ ; fines (for pollution) or rewards (for nonpollution) are implemented following the outcome of the inspection.We can assume that inspection reveals the true pollution level with certainty. Rewards and fines are denoted by f (a ) ,0 ≤ f (a ) ≤ F where F is the maximum fine. We can always consider rewards, by letting f < 0.5 In general, social welfare is given by: SW = ∑  g j (a j ) − h (a j ) − C (θ) j

where C (θ) is the cost associated with inspection, C ′ (θ) > 0 , a j is action (pollution) chosen by firm j. The first two terms in the square brackets refer to private benefit and cost associated with pollution choice of firms. The third term is the aggregate inspection cost associated with maintaining an inspection regime where each firm is inspected with probability θ. Firms are risk-​ neutral with utility from a given by U (a ) = g (a ) − e (a ) where e (a ) denotes expected cost associated with an action a. In the absence of corruption, it is simply the expected fines. Assuming f (0 ) = 0, a firm will choose a = 0 only if U (1) = g (1) − θ f (1) ≤ U (0 ) = 0. Let g * be the marginal firm that is indifferent between polluting and not; given U (1) = 0, it is clear that the required monitoring probability θ = g * /F . When all firms are being prevented from polluting, clearly θ = G / F .6 The basic enforcement setting can be extended to consider (i) multiple actions or different levels of pollution (as we do later); (ii) heterogeneity among firms; (iii) endogenous θ , which depends on inspector’s costly effort (Mookherjee and Png, 1995); (iv) self-​reporting by firms and rewards for nonpollution and other incentive schemes and (v) imperfect technology where action is not revealed following inspection (Tirole, 1992). For example, it is possible that monitoring leads to the observation of a signal but not the true action. In that case, fines will be based on the signals. Enforcement policy

( f (a ) , θ) may need to change, depending on the degree of imperfection in the

monitoring technology.7 Given that our focus is on the problem of corruption, we stick to this simple setting and explore the implications of corruption by the inspector.

Corruption and Optimal Enforcement  61

Corruption: Collusion A corruptible inspector can misreport the true a . Suppose a = 1 can be reported as a = 0.8 Collusion refers to the underreporting of the true action (pollution) in exchange for a bribe from the firm. Since bribe payment B, if feasible, will be less than the stipulated fine, both the firm and the inspector stand to gain from colluding this way. Following Tirole (1986), there is a sizable literature looking at various implications of and solutions to this problem of collusion. Dilution Clearly, enforcement is diluted now. Firms with gain g > θB will choose to pollute, expecting to pay a bribe B and avoid the fine, once their act is revealed following the inspection. In general, firm’s compliance incentives now depend on e (a ), not the original fine function f (a ). Of course, the extent of dilution will depend on how bribes are determined and whether they depend on a, f and other factors. However, enforcement is diluted but not eliminated altogether. If B = B ( f ), we can raise inspection probability to θ ′ = G / B (F ) > θ and have the originally stipulated level of deterrence but at a higher cost.9 However, depending on the parameter values, such a θ ′ may not exist and the increased cost may make g * < G optimal. The above is suggestive of the possibility that we can implement the outcome that would be realized in the absence of corruption, even without deterring corruption. Note that firms would have to pay bribes, but these payments act as deterrence. This is different from the approach where the policies are geared toward deterring corruption –​ensuring noncorrupt behavior by the inspector. Preventing Collusion Much of the literature focuses on the prevention of such collusion by looking at various mechanisms involving incentive schemes for the agent and the client. Assuming inspectors are purely guided by monetary considerations, a combination of rewards and sanctions can be employed. In some sense, there is another enforcement problem similar to the original one. Suppose, the inspector gets a reward R for reporting a = 1 and nothing for reporting a = 0. In that case, collusion is costly for inspector, as it means forgone reward income. One can find suitably high levels of rewards, such that the inspector is induced to report truthfully. If the social welfare function, like the one specified earlier, does not consider transfers, then it is easy to see that rewards can be made as large as the fines.10 The collusive deal is detected with some probability q . This could be because of (i) a second level of audit, often termed as super-​audit or external audit (Kofman and Lawarree, 1993; Mookherjee and Png ,1995), or (ii) external leaks facilitated by media, civil society or other whistle-​blowers (Brunetti and

62  Ajit Mishra Weder, 2003; Jha and Sarangi, 2017). The inspector is sanctioned after collusion or underreporting is discovered, S > 0 . This could take the form of dismissal from the job and loss of future wage income (Di Tella and Schardorsky, 2003).11 Following this, initial fine f will be restored for the firm and there can be additional sanction for bribe-​giving as well. It is easy to see that collusive bribery will be affected by these. If we assume that bribe is determined according to Nash bargaining solution, as is commonly done, we can derive bribe as a function of these variables: B = B ( f , R, q, S ) .12 For q > 0 , it can be shown that there always exists R * < F , such that for any R > R * , collusion will not take place.Various authors have focused on different aspects of this mechanism and have explored their implications (Tirole, 1992).

Corruption: Extortion and Harassment Extortion refers to the case of actual –​or the threat of –​overreporting by the inspector. The scope of extortion depends on the particular information structure and the ability of the inspector to manipulate information. This will, of course, depend on the action chosen by the firm. If the firm has chosen maximal pollution, overreporting is not feasible. While it is relatively obvious to see why collusion, a Coasian side contract, can occur, it is not obvious how extortion can be explained.The information structure must allow the inspector to report a = 0 as a =1 . In some cases, such framing may be possible, but information may not be manipulable easily. For example, Khalil et al. (2010) argue that it is easier to manipulate evidence in the case of collusion. Information is soft when both collude, and in such an event report α can take any value, α ∈ {0,1, φ} irrespective of the true value of a . Note that α = φ refers to a report of no information. On the other hand, extortion is limited if information is hard and cannot be manipulated, α ∈ {a, φ} .13 Hence extortion can take place only by suppression of true agent favorable information. However, in many nonenforcement settings, extortion can occur with greater ease. Consider a case of the government issuing licenses (or other services) to applicants, who have to comply with certain requirements to be eligible. Collusion takes place when a corrupt officer can issue a license to an ineligible firm in exchange for a bribe. Extortion can also happen when an eligible firm is denied the license and is faced with a bribe demand. Here information manipulability is not the issue, but the officer’s discretion or ability to delay is important in making extortion demands credible. In fact, early discussions on corruption focus on these kinds of bribery. It is often viewed as harassment bribes or speed money, depending on the contexts. As argued in the introduction, the difference between these two forms of bribery is important. Extortion or harassment bribes can grease the wheels in this case. On the incentive side, collusive bribery allows the excessive entry of inefficient firms; extortion will discourage entry of efficient firms. Several researchers have looked at the possible tradeoffs between collusion control and extortion prevention. It is possible that both can be prevented. In

Corruption and Optimal Enforcement  63 some situations, however, optimal policy involves tolerating one or another. We do not attempt to discuss these issues, but rather point the reader to the recent literature on this.14 In terms of the incentive effects, it is clear that extortion possibilities do reduce incentives for compliance, although it is not always straightforward and it also depends on the context. For example, the dominant form of corruption may depend on the regulatory regime. De Chiara and Manna (2019) consider the regulation of innovation, where the product of successful innovation needs to be authorized before its use. They contrast two regimes –​lenient authorization and strict authorization –​and examine how different forms of corruption are related to these. The distinction is based on the notion of the burden of proof. In the former, authorization is granted if there is no conclusive evidence against the product.15 Because corruption is limited to concealment of information only, as in a hard information setting, these two authorization regimes are subject to different types of corruption. The inspector can collude with the firm and conceal unfavorable evidence, whereas she can extort by threatening to conceal favorable evidence in the case of strict authorization. Likewise, Vafai (2010) considers both types of corruption in an organizational setup to show how extortion can be deterred. Interestingly, ignoring extortion may lead to the design of inefficient contracts. Here also collusion takes place when the supervisor hides agent-​unfavorable information and extortion happens as the supervisor threatens to hide agent-​ favorable information. In a similar vein, Samuel and Mishra (2021) and Hong and Yin (2020) find some positive enforcement effects of extortion but in very specialized information settings. For example, in Hong and Yin (2020), officials grant licenses to applicants from a pool of potentially viable as well as nonviable projects. A corrupt officer can collude with a nonviable project owner and award license and can potentially extort the owner of a viable project. It turns out that there are more nonviable projects when the officer can only collude compared to the case where the officer can collude as well as extort. On the other hand, Khalil et al. (2010) consider both forms in a moral hazard setting, similar to our pollution example, and find that collusion is the lesser of the two evils. The regulator may optimally tolerate collusion in some cases, but they show that it is never optimal to allow extortion. Fighting Extortion Anticorruption strategies are often clubbed together, but fighting extortion requires a different set of instruments. Note that the reward for the inspector to deter collusion is likely to encourage extortion. Similarly, monitoring and further inspection may work, but the nature of monitoring is quite different now. In the collusion case, audited firms can be inspected further to uncover true a and collusion can be detected if reported action has been different, α ≠ a . But in the case of extortion, since true a is reported, such inspections are not useful. Of course, there can be direct auditing of inspector’s income and financial transactions –​to detect bribery.

64  Ajit Mishra Given that the interests of the two parties (the inspector and the firm) are not aligned, the firm can always be encouraged to report the extortion threat. In fact, we can have a system of appeals to deter extortion demands. Suppose the firm, having chosen a = 0 , is facing an inspector demanding a bribe. The firm can refuse payment, and if the inspector reports a = 1, the firm can appeal. Assuming the appeal system detects the true a, the firm does not have to pay any fine and the inspector can be penalized too. However, as Mishra and Samuel (2018) argue, the ability to sanction the inspector depends on the liability regime and it cannot be taken for granted. Secondly, even when there is some reimbursement of legal costs for the victimized firm, it still incurs a positive cost, say L > 0. Potentially, extortion can be prevented if L is not very large.16 For example, in the current case, since f (0 ) = 0 and f (1) = F , if L < F , it is clear that inspector’s threat to report a = 1 is not credible. The nonpolluting firm will rather appeal than pay the fine, and knowing this, the inspector will prefer not to report. This means there cannot be any extortion.

Overenforcement? As discussed in the beginning, we return to the theme of how optimal policies are affected under corruption. Note that enforcement policy can be viewed as an incentive scheme for the firms, where a firm’s payoff depends on its action. In the pollution example, this translates to penalties being dependent on pollution activity. Likewise, consider the problem facing a manager who cannot observe the worker’s action (effort), but would like the worker to choose high effort. As in the classic moral hazard case, output depends on agent’s action and a random variable. Output can be low (high), even when the worker has put in high (low) effort. An agent is hired to monitor effort and the agent observes a signal of effort. Now the manager can base worker’s compensation on the agent’s report, with compensation increasing in reported effort level (as in Khalil et al., 2010). The slope of this compensation schedule or the penalty schedule can be referred to as the power of the scheme. High-​powered scheme means payoff being more sensitive to reported action. But, since misreporting is possible, the scope for corruption is higher, greater is this dependence of firm’s or worker’s payoff on agent’s report. For example, in the extreme opposite case, when compensation/​penalty is flat, there is no scope for corruption! It is common to expect that incentive schemes will tend to be less high-​powered in the presence of corruption possibilities. In our setting, this means that corruption possibilities may lead to underenforcement where the regulator allows certain level of pollution (or activity), even though private gains fall short of social costs. To see this, consider a case where inspectors are corruptible, but they engage in bribery only when expected bribe amounts are sufficiently large to compensate for the financial risk or moral costs of bribery. For example, suppose bribery is detected with a given probability and upon detection the agent loses job (a fixed sum of income) irrespective of the size of bribe taken. Clearly, the agent

Corruption and Optimal Enforcement  65 will refrain from accepting small bribes. Hence, collusion can be discouraged by making expected bribe lower. But bribes are going to be dependent on the gain from collusion, which in this case is proportional to the difference in fines: f (1) − f (0 ) . This means that the regulator may wish to settle for a smaller ( f (1) − f (0 )) so that the corresponding bribe is not high/​feasible.This

leads to either (i) higher monitoring probability θ with f (1) < F or (ii) the regulator allowing some firms with high private gains to pollute. However, as the following example shows, we can have the exact opposite case also. Let us extend the previous pollution example to consider three types of firms, j = 1, 2, 3. Like before, firms choose pollution level a ∈ {0,1, 2}. Assume that g j (0 ) = 0, g j (1) = jg1 , and g j ( 2) = jg2 , with g2 > g1 > 0. Note that the private benefit from pollution increases in the level of pollution and the type of the firm. For example, type 3 firm gains more from higher pollution than type 2. Firm’s type is not available to the regulator, and the regulator cannot condition monitoring intensity on the firm’s type. Hence all firms are subject to same rate of monitoring, θ. The regulator chooses fines f (a) and θ to implement the desired level of pollution. Social welfare, like before, is simply SW = ∑  g j (a j ) − h (a j ) − C (θ) with j

C ′ (θ) > 0, where a j is the action chosen by firm type j. As mentioned earlier, fines and transfers do not enter welfare calculations. Now the regulator has to choose action profile {a j }, specifying an action for each firm type. Like the simple case discussed earlier, it could still be a j = 0 for all j, which is likely if social cost is very high compared to private gains of all types. Suppose private benefits and associated harms are as follows: 3 g2 − h2 > 3 g1 − h1 > 2 g1 − h1 > 0 > 2 g2 − h2 > g1 − h1 > g2 − h2 (1) It can be shown that in the absence of any enforcement costs, given (1), the first best pollution profile will be given by:17 a1 = 0, a 2 = 1, a 3 = 2

(2)

This says that type 3 firm will be allowed to pollute at the highest level as its gain outweighs the corresponding social cost. Likewise, type 2 firm is not allowed to pollute at the highest level, but is allowed to do so at the lower level. Type l firm with the least benefit will not be allowed to pollute at all. This outcome will be denoted as a FB = (0,1, 2).18 Heterogeneity in benefits leads to this situation of marginal deterrence, where to deter the highest type will require a very high level of monitoring. So, with costly monitoring, higher types are more likely to be allowed to pollute. Relative to this first best reference level, we can define overenforcement and underenforcement to refer to situations where an action with gain (harm) in

66  Ajit Mishra excess of harm (gain) is being allowed. For example, if monitoring is costly, the regulator may wish to implement (l, l, 2) implying the effective legalization of the low level of pollution. This is a case of underenforcement. Similarly, (0, l, l) or (0, 0, 2) will be the case of overenforcement. Let us consider corruption possibilities. To focus on extortion and its distortionary nature, we focus on the case where collusion is prevented through the use of rewards to the inspectors. Recall that if rewards exceed the net expected bribe (from collusion), the inspector will always report truthfully. We capture this by assuming that there is a transaction cost associated with bribery, so that bribe worth x is equivalent to kx for the inspector, k < 1. In that case, r ≥ k, where reward equals R = rf will be enough to deter collusion. What about extortion possibilities? There are various possibilities: a = 0 can be reported as a = 1 or a = 2 ; a = 1 can be reported as a = 2. As outlined in the previous section, extortion is deterred to the extent firms appeal against this overreporting. But the presence of fixed costs of appeal implies that some extortion demands may not lead to appeal if the firm’s appeal costs exceed expected payments due to overreporting. In the earlier pollution example, only extortion possibility was a = 0 being reported as a = 1. Such reporting can be discouraged by setting f (1) − f (0 ) = F > L . However, the present example has multiple extortion possibilities; to eliminate extortion we need f (1) − f (0 ) > L , f ( 2) − f (0 ) > L and f ( 2) − f (1) > L . As the following example demonstrates, this may not be feasible. Our objective is to show that optimal enforcement policy is affected in a fundamental way. Note that when there is no corruption or the inspector is honest, it will never be optimal to implement pollution levels where the higher types are forced to choose lower levels compared to the first best. This will be referred to as overenforcement. Given (1) and (2), we will never choose to have type 2, choosing a < 2. This will lead to a loss of welfare from forgone net positive gains (as seen in (1)) and possibly higher monitoring cost. This is true for other types as well. But, with extortion, we can show that it may be optimal to overenforce. One example of overenforcement is complete deterrence where all types are choosing zero pollution. Consider the following example.19 Let g1 = 8, g2 = 15, F = 40, h ( 2) = 35, h (1) = 15 , L = 24 and k = 1 / 2. It can be checked that with f (0 ) = 0 , f (1) = 16, f ( 2) = 40 and θ = 2/3 , we can implement a FB = (0,1, 2) in the absence of corruption. With collusion possibilities, we can choose reward schemes r = k to deter collusion. But with extortion possibilities, the effective penalties change to make marginal deterrence difficult. For example, a firm having chosen a = 1 can be reported as a = 2. Since L = 24 = f ( 2) − f (1) , the firm does not appeal and it is happy to pay an extortion bribe to avoid being reported. Since the officer stands to gain k ( f 2 − f 1 ), bribe has to at least be k ( f 2 − f 1 ).20 Hence it is clear that e (1) = f (1) + k ( f 2 − f 1 ),whereas e (2) = f ( 2). This affects marginal deterrence as the expected difference between penalties for different actions has been reduced. Likewise, a = 0 cannot be reported as a = 2, as it would lead to firm appealing (since 40 > 24 ), but it can be reported

Corruption and Optimal Enforcement  67 as a = 1. What this has done is to restrict further regulator’s ability to choose penalties in order to induce the desired pollution choices by firms.21 In the present case, it is not possible to implement (0,1, 2). What about other pollution levels or action profile? Note that we can achieve other levels of deterrence (0, 0, 2) by choosing f (0 ) = 0 , f (1) = f (2) = 40, θ = 3/4. Compared to the first best (0,1, 2), it is clear that there is overenforcement. Even though type 2 firm’s private benefit from medium level of pollution (a = 1) exceeds associated social cost (since 2 (8) > 15 ), type 2 firm is being forced to choose a = 0. It can be shown that for suitable parameter values, this overenforcement is the optimal policy. The intuition lies in the way a mechanism makes different actions extortion-​ proof. An action is extortion-​proof if the inspector does not overreport it. If the threat of overreporting prompts an appeal, then such threat is not credible and the action can be taken to be extortion-​proof. For example, a = 0 is extortion-​ proof since if it is reported as a = 1 or 2, the firm will appeal and incur cost L rather than pay the fine. But if a lower fine for a = 1 were to be chosen to incentivize type 2 firm to pollute intermediate level, we may end up with a f (1) that will make zero pollution activity extortionable.

Conclusion We have used a simple model of enforcement to explain how corruption arises and how anticorruption strategies are developed. This chapter has sketched the broad contours without providing any details. The main objective has been to draw attention to the nature of optimal policies in a corrupt environment. While we have shown how optimal policies do get affected by the presence of corrupt agents and anticorruption policies, we have not covered various other aspects of anticorruption strategies.The approach to anticorruption strategies, basically follows Becker’s approach to crime and punishment (Becker and Stigler, 1974), where the sole objective is to affect material incentives facing the official. Clearly, this approach overlooks the moral, social and cultural aspects of their behavior. For example, even in the simple case of public sector wages, while a negative relation between corruption and wage level is commonly accepted (Van Rijckeghem and Weder, 2001), the impact of higher wages on curbing corruption is not straightforward (Fisman and Golden, 2017; Demigruc-​Kunt et al., 2021). Banuri (2023), in this volume, looks at how pay reforms affect the organizational culture and corruption in a nonlinear way. Anticorruption efforts should not simply focus on monitoring, rewards and sanctions, but should consider issues like transparency, culture and motivation. Recent advances in behavioral economics point to several social and psychological reasons that also make corruption more costly. Banerjee, Dasgupta and Mitra (2023), in this volume, survey the recent literature. There are studies that show that excessive regulations to fight corruption can be counterproductive.The objective of fighting corruption can sometimes lead

68  Ajit Mishra to excessive monitoring and proliferation of laws with adverse implications. There is agreement that transparency and nondiscretionary power are essential to an anticorruption strategy, but the question is how best to bring these in. The problem lies in the fact that we often rely on legislation, new rules and regulations.When we are beginning with a situation of distrust, we often create more rules and regulations. There are, however, instances where this can be taken to a ridiculous level, and we are probably going to stifle the system.22 Basu (2023), in this volume, goes a step further and draws attention to this phenomenon of complex laws and rules getting broken by ordinary citizens without full knowledge and how this can be misused by the ruling class.

Notes 1 Corruption is likely to impact the formulation of policy also, where politicians will choose policies to suit private interests. But we do not go into details of the political economy of corruption here. 2 Of course the optimality of complete deterrence is not to be assumed (see Tirole, 1992; Kofman and Lawarree, 1993). On the other hand, Mishra and Mookherjee (2013) ask if the optimal outcome can be achieved, without complete deterrence of corruption, by altering the initial policy suitably. 3 While collusion has been studied in great depth, the other form of corruption –​ extortion –​features only in the recent literature. 4 It is often termed as the Principal-​Supervisor-​Agent or the Principal-​Agent-​Client framework where the Principal is employing an agent to monitor or provide services to the client. 5 For example, we can have f (0) < 0 and f (1) > 0. In the case of fines, the upper bound on fine can be due to (i) either some judicially accepted level of fine or (ii) limited liability constraints of firms, whose common wealth equals F . 6 This need not be the optimal policy. If costs are high, optimal policy may involve allowing some firms with G ≥ g > g * to pollute. 7 Suppose the inspector observes a signal X ∈ {X B , XG }, depending on the value of a, where X = X B if a =1 and X = X B with probability π,0 < π < 1 and X = XG with (1 − π ) when a = 0. It can be shown that a = 0 can still be implemented by choosing fG = 0, f B = F and θSB = G /F (1 − π ). 8 This assumes that evidence (regarding pollution) can be easily manipulated. However, in some other cases, we can have a situation where evidence (hard information) can be suppressed but not manipulated. Here, the inspector can claim not to have observed anything, but cannot report another action. 9 Basu, Bhattacharya and Mishra (1992) and Mishra and Mookherjee (2013) look at these examples of implementation of the original compliance level even with corruption, the latter in a more general setting. 10 This is nothing but privatized enforcement since the inspector is effectively collecting the stipulated fine, R = f (1). 11 It is not automatic; the higher-​order officers can be corrupt too, so the sanction will be replaced by another round of bribes by the inspector to its superiors. 12 Other types of bargaining methods with well-​defined solutions can be used also. 13 This is different from the earlier literature on collusion following Tirole (1992), where it is commonly assumed that information is hard and it can be only suppressed but cannot manipulated.

Corruption and Optimal Enforcement  69 14 It includes, among others, contributions by Hindriks et al. (1999), Marjit et al. (2000), Polinsky and Shavell (2001), Saha (2003), Khalil, Lawarree and Yun (2010), Vafai (2010), Mishra and Mookherjee (2013), De Chiara and Manna (2019), Hong and Yin (2020), and Samuel and Mishra (2021). 15 This is related to, but different from, the various immunity regimes studied in Mishra and Samuel (2018). 16 There are other ways to make extortion nonfeasible, especially using e-​governance. See the chapter by Jha and Sarangi (2023) in this volume. 17 Earlier, we assumed it to be a1 = a 2 = a 3 = 0 . 18 It is clear that f (0) = 0, f (2) = F , where f (1) will be determined to satisfy the various incentive constraints. 19 For ease of reading, detailed calculations are not provided but can be obtained from the author. This example is based on Mishra (2022). 20 This will obviously depend on the distribution of bargaining power between the firm and the inspector. 21 It may still be possible to achieve marginal deterrence with extortion in some cases, as e (2) > e (1) > e (0), but it is much harder. 22 For example, the United Nations Convention against Corruption says that we must reduce the ad-​hocness in our system and specify everything that is related to contract enforcement to deal with corruption. Kelman (2003) discusses how a US Defense tender document for oatmeal cookies runs into twenty-​six pages of descriptions!

References Banerjee, R., Dasgupta, U. and Mitra, S. 2023. Stopping the Rot I: A review of models and experimental methods of corruption experiments, in The Political Economy of Corruption, Jha, Mishra and Sarangi (eds.), Routledge. Banuri, S. 2023. Attracting honest workers: Organizational culture, corruption, and worker preferences, in The Political Economy of Corruption, Jha, Mishra and Sarangi (eds.), Routledge. Basu, K. 2023. The political economy of corruption control, in The Political Economy of Corruption, Jha, Mishra and Sarangi (eds.), Routledge. Basu, K., Bhattacharya, S. and Mishra, A. 1992. Notes on bribery and control of corruption. Journal of Public Economics 48: 349–​359. Becker, G. S. and Stigler, G. 1974. Law enforcement, malfeasance, and compensation of enforcers. Journal of Legal Studies 3 (1): 1–​18. Brunetti, A. and Weder, B. 2003. A free press is bad news for corruption. Journal of Public Economics 87: 1801–​1824. De Chiara, A. and Manna, F. 2019. Corruption and the regulation of innovation. University of Barcelona Working Paper 2019/​390. Demirguc-​Kunt, A., Lokshin, M. and Kolchin,V. 2021. Effects of public sector wages on corruption. Policy Research Working Paper 9463, World Bank. Di Tella, R. and 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. Fisman, R. and Golden, M. 2017. Corruption: What Everyone Needs to Know. Oxford University Press, Oxford.

70  Ajit Mishra Hindriks, J., Keen, M. and Muthoo, A. 1999. Corruption, extortion and evasion. Journal of Public Economics 74(3):395–​430. Hong, F. and Yin, Z. 2020. Collusion, extortion and the government’s organizational structure. Journal of Economic Behavior and Organization 180(C):1–​23. Jha, C. and Sarangi, S. 2017. Does social media reduce corruption? Information Economics and Policy 39: 60–​71. Jha, C. and Sarangi, S. 2023. Advances in information and communications technology and corruption, in The Political Economy of Corruption, Jha, Mishra and Sarangi (eds.), Routledge. Kelman, S. 2003. Remaking Federal Procurement. The John F. Kennedy School of Government Series Visions on Governance in the 21st century,Working Paper No. 3. Khalil, F., Lawarree, J. and Yun, S. 2010. Bribery versus extortion: Allowing the lesser of two evils. RAND Journal of Economics 41 (1): 179–​198. Kofman, F. and Lawarree, J. 1993. Collusion in hierarchical agency. Econometrica 61 (3): 629–​656. Marjit, S.,Vivekananda, M. and Arijit, M. 2000. Harassment, corruption, and tax policy. European Journal of Political Economy 16: 75–​94. Mishra, A. 2022. Marginal Deterrence with Collusion and Extortion, University of Bath. Mishra, A. and Mookherjee, D. 2013. Controlling Collusion and Extortion:The Twin Faces of Corruption, Boston University. Mishra, A. and Samuel, A. 2018. Law enforcement and wrongful arrests with endogenously (in)competent officers. Economic Inquiry 56: 1417–​1436. Mookherjee, D. and Png, I. P. L. 1995. Corruptible law enforcers: How should they be compensated? Economic Journal 105 (428): 145–​159. Motta, A. 2009. Ex-​ante and ex-​post corruption. Marco Fanno Working Paper No. 94, University of Padua. Png, I. P. L. 1986. Optimal subsidies and damages in the presence of judicial error. International Review of Law and Economics 6: 101–​105. Polinsky, A. M. and Shavell, S. 2001. Corruption and optimal law enforcement. Journal of Public Economics 81: 1–​24. Polinsky, M. and Shavell, S. 2007. The Theory of Public Enforcement of Law: Handbook of Law and Economics,Volume 1, Elsevier. Saha, B. 2003. Harassment, corruption, and tax policy: A reply. European Journal of Political Economy 19 (4): 893–​897. Samuel, A. and Mishra, A. 2021. Does it matter who extorts? Scottish Journal of Political Economy. DOI:10.1111/​sjpe/​12300 Tirole, J. 1986. Hierarchies and bureaucracies: On the role of collusion in organizations. Journal of Law, Economics and Organizations 2 (2): 181–​214. Tirole, J. 1992. Collusion and the theory of organizations, in Advances in Economic Theory, Vol. 2 (ed. J.-​J. Laffont), Cambridge University Press. Vafai, K. 2010. Opportunism in organizations. Journal of Law, Economics and Organizations 26 (1): 158–​181. Van Rijckeghem, C. and Weder, B. 2001. Corruption and the rate of temptation: Do low wages in the civil service cause corruption? Journal of Development Economics 65: 307–​331.

6 A Theory of Joint Evolution of Corruption and Growth Niloy Bose, Richard Cothren and Nazanin Sedaghatkish

Introduction According to Karl Kraus,1 corruption is worse than prostitution because the latter endangers the moral of an individual, whereas the former endangers the moral of an entire country. For an economist, the issue is not one of morality but the fact that corruption and development are closely intertwined. Naturally, corruption is a clandestine activity and is difficult to measure. Since the early 1980s, however, a number of organizations –​most notably World Bank, Business International Corporations, Political Risk Services Incorporated and Transparency International –​have offered measures of perceived corruption on the basis of questionnaire surveys, whose responses are used to compose a ranking of the extent to which corruption is perceived to exist.While differing in various aspects (e.g., coverage, methodology and availability) and while being susceptible to the usual caveat associated with survey data, the indices were found to be highly correlated with each other and all highly correlated with key economic variables. For these reasons, the indices have become widely accepted as providing reliable estimates of corrupt activities. Their publication has given rise to a flurry of empirical investigations into the relationship between corruption, development and other phenomena. The main findings from this literature offer a series of stylized facts, some of which are highlighted below to set the tone of this chapter. First and foremost, there is overwhelming evidence of a significant negative relationship between the incidence of corruption and economic growth, and this relationship is moderated through a variety of channels. For example, corruption hurts growth by lowering investment (Mauro 1995, Knack and Keefer 1995, Gyimah-​Brempong 2002); corruption creates obstacles to doing business and encourages unofficial sectors (e.g., Johnson et al. 1997); corruption reduces inflow of foreign capital (Wei 2000, Lambsdorff 2003); corruption encourages political instability (e.g., Mo 2001) and decreases the quality of public investment through a misallocation of public expenditure (Tanzi and Davoodi 1998, Mauro 1998); corruption reduces human capital formation as well as firms’ growth (see Reinikka and Svensson 2004, Reinikka and

DOI: 10.4324/9781003142300-6

72  Niloy Bose, Richard Cothren and Nazanin Sedaghatkish Svensson 2005, Fisman and Svensson 2007). Naturally, these effects have adverse consequences for a country’s growth performance.2 Second, cross-​ country differences in the incidence of corruption owe much to cross-​country differences in the level of development.3 For example, according to Treisman (2000), rich countries are generally rated as having less corruption than poor countries, with as much as 50–​73 percent of the variation in corruption indices being explained by variations in per capita income levels. This view is shared by several other studies (e.g., Montinola and Jackman 2002, Paldam 2002, Rauch and Evans 2000). However, the relationship between corruption and growth is not necessarily linear. In fact, the data points to a considerable diversity in the incidence of corruption among middle-​income countries, compared to the uniformly high levels of corruption among low-​ income countries and the uniformly low levels of corruption among high-​ income countries (Blackburn et al. 2010). Finally, even a cursory inspection of the data reveals that many of the low and middle countries where corruption was rampant in the past are among the most corrupt countries today.4 Therefore, countries may be drawn into a vicious circle of high corruption from which there is no easy escape. This conjures up the ideas of the presence of multiple equilibria in the corruption–​development nexus. Haque and Kneller (2009) provide a formal account of such persistence in the data by identifying corruption development clubs where countries appear to become trapped in high corruption–​low development or low corruption–​high development regimes. The above empirical observations pose many important challenges to theoretical research on the macroeconomics of misgovernance. The first challenge is to provide a rigorous, formal theoretical account of why corruption can be damaging to economic development. Ehrlich and Lui (1999) and Sarte (2000) offer some answers. The former develops a model in which opportunities to profit from bureaucratic corruption create incentives for individuals to compete for the privilege of holding public office. These incentives lead to a diversion of resources away from growth-​promoting activities (investment in human capital) toward power-​seeking activities (investment in political capital). The latter constructs a framework in which rent-​seeking bureaucrats restrict the entry of firms into the formal sector of the economy, which has a better system of property rights and law enforcement than the informal sector.When the cost of formality is high, growth is reduced relative to the free-​entry case. These analyses are revealing about the way in which corruption can have adverse effects on the prospective fortunes of an economy. They are less clear about why corruption may arise in the first place, why corruption may persist or decline over time and why corruption may vary across otherwise similar countries. The second challenge is to explain why corruption may itself be affected by development. Although a considerable amount of theoretical research has been directed toward understanding the causes of corruption and its implication for efficiency and welfare (e.g., Banerjee 1997, Carrillo 2000, Mookherjee and Png 1994, Rose-​Ackerman 1975, 1999, 2013, Ackerman 1978, Shleifer and Vishny

A Theory of Joint Evolution of Corruption  73 1993, Andvig and Moene 1990, Cadot 1987, Tirole 1996), these research are generally framed in a static partial equilibrium framework and less suited for analyzing how the level of development could alter the costs and benefits of engaging in corrupt activities within the context of fully specified dynamic general equilibrium models. The above two challenges can be addressed by proposing a unified framework that respects the notion that corruption not only influences but also is influenced by economic development. Some progress has been made in this direction where feedback loop from growth to corruption is combined with standard mechanism through which corruption reduces growth (Blackburn et al. 2006, Mauro 2004, Blackburn et al. 2010, 2011, Aidt et al. 2008, Blackburn and Forgues-​Puccio 2006). These frameworks not only account for the joint evolution of corruption and development but also shed light on the persistence and nonlinearity in the corruption–​development relationship.The main goal of this chapter is to present a simple benchmark analysis to capture the essence of such a framework. To set the stage, consider an economy in which public agents (bureaucrats) are delegated the responsibility for collecting taxes from private individuals (households) on behalf of the political elite (the government). Bureaucrats may exploit their powers of public office to collude with households in bribery and tax evasion. The precise effects of corruption can manifest in a variety of ways. The level of corruption may determine the tax burden and the available resources for public spending. For example, the greater the level of corruption, the higher the tax burden and the lower the public spending toward productive infrastructure if the government is to balance its budget. The higher level of corruption can also reduce resources available for productive private investments and therefore can influence other economy-​wide variables such as wage rates and the returns to private capital. Significantly, some of these state variables are also relevant when deciding whether or not to engage in corruption. For example, a higher tax rate or lower wages for bureaucrats can create an ideal environment for bureaucrats and households to collude to evade taxes. In this chapter, we present a simple model that respects the notion that corruption not only influences but is also influenced by economic development. By admitting this two-​way causality, the analysis can capture the joint evolution of corruption and development along with a rich set of possibilities, including multiple equilibria and threshold effects.

The Environment Corruption is a many-​faceted phenomenon. The discussions in this chapter, however, will pivot around public sector corruption, which is broadly defined as illegal, or unauthorized, profiteering by public officials who exploit their positions in public office for personal gain. To many observers, corruption in public office is an inevitable aspect of state intervention that typically entails some transfer of responsibility from the government to a bureaucracy.

74  Niloy Bose, Richard Cothren and Nazanin Sedaghatkish Now, to fix the idea, consider an economy with a constant population of two-​ period lived agents belonging to overlapping generations of nonaltruistic families. Agents of each generation are divided into two groups of citizens –​private individuals (or households) and public servants. For simplicity, we abstract from issues relating to occupational choice and assume that agents are differentiated at birth according to their skills, and some agents lack the skills necessary to become a bureaucrat, while others possess such skills. Further, to economize notations, we normalize the size of each group to 1. At birth, each household is endowed with a certain amount of labor endowment. To be specific, we assume that a fraction µ ⊂ (0,1) of households are endowed with λ > 1 units of labor, while the remaining fraction, (1 − µ ), are endowed with only one unit of labor. Depending on the endowment, a household is either a low-​income earner and exempt from paying taxes, or a high-​ income earner and liable to pay taxes. In each case, the household saves its entire income net of taxes as capital, which is rented to firms at the market interest rate, rt +1 , to finance old-​age consumption. Taxes, τt , are lump-​ sum and collected by bureaucrats on behalf of the government, which requires funding public expenditures. In the absence of corruption, a household with λ units of labor endowment is expected to earn a net income of ( λwt − τt ). Corruption arises from the incentive of a bureaucrat to conspire with a household in concealing information (the household income) from the government. In doing this, a bureaucrat expects to gain from bribe income, bt , and the household expects to gain from tax evasion. With probability, p , the household and bureaucrat succeed in their conspiracy, and the high-​income households’ net income is λwt − bt. With probability (1 − p ) , their collusion is exposed and the household is forced to pay its tax, implying a net income of λwt − bt − τt . Given these outcomes, we may write the expected lifetime utility of a high-​income household as r ( λwt − τt ) if bt = 0 ut =  t +1 (6.1) rt +1  λwt − bt − (1 − p ) τt  if bt > 0 For this present analysis, we assume that each young bureaucrat of generation t is paid the market salary, wt , from supplying inelastically his unit labor endowment to the government. Like all households, all bureaucrats save their entire income to finance old-​age consumption. Depending on the circumstance, a bureaucrat may or may not engage in corruption. If the latter, a bureaucrat receives an income of wt , with certainty. If the former, then his expected income depends on the bribes that he receives, the chances of being caught, the resources spent on avoiding detection and the penalties incurred if he is exposed. In general, corrupt individuals may try to remain inconspicuous by hiding their illegal income, by investing this income differently from legal income and by altering their patterns of expenditure.We assume here that a corrupt bureaucrat

A Theory of Joint Evolution of Corruption  75 must dispose of all side payments immediately if he is to stand any chance of not being caught. The concealment of bribes however comes with a cost. Specifically, to conceal the bribe money, a corrupt bureaucrat must seek help from a group of agents specializing in investing illegal income, which makes it difficult for the government to trace the source of illegal funds. Such activities may include investing in informal sectors or laundering illegal income outside the country. In return, a corrupt bureaucrat must pay a specialist a commission, Ct = ρbt , where ρ ∈ (0,1) and bt represents the volume of bribe or illegal income that a corrupt bureaucrat wishes to conceal. We assume that the bribe, bt , earns a lower rate of return5 than rt +1 .This assumption captures the idea that hidden investments in the informal sector are unable to access the same support infrastructure that investments in the formal sector can. As a consequence, the productivity of the former is lower than that of the latter. We postulate that the return on bribe income, ς , is a fraction of what is offered in the formal sector, that is, ς = εrt +1 where ε ∈ (0,1) . Accordingly, the net rate of return a corrupt bureaucrat earns on his illegal income is given by χt +1 = ςrt +1 − ρ . It is also the case that income from bribe that corrupt bureaucrats conceal and launder does not contribute to the amount of savings available for capital formation.6 With the above description of events, we may write the initial net income of a corrupt bureaucrat as wt + µbt with probability p and wt + µbt − f t with probability 1 − p . Here f t denotes the fine imposed on the bureaucrat on being caught. To deter corruption, it is optimal for the government to set f t = wt + µbt . Given that wt is invested at the official market rate, rt +1 , while the illegal income, bt, earns a net rate of return χt +1 = ςrt +1 − ρ, we summarize the expected payoff to a bureaucrat as r w if b = 0 νt =  t +1 t t (6.2) p r w + r − b if b 0   > ς ρ µ ( ) + 1 + 1 t t t t t    We view the government as providing public services that contribute to the efficiency of output production (e.g., Barro 1990). Expenditure on these services, gt , is assumed to be a fixed proportion, θ ∈ (0,1), of output. The government also incurs expenditures on bureaucrats’ salaries. The government finances its expenditures each period by running a continuously balanced budget. Its revenues consist of the taxes collected by bureaucrats from high-​ income households, plus any fines imposed on bureaucrats who are caught while engaging in corruption. Finally, the output in this economy is produced using the technology: β γ yt = Altα kt gt (6.3)

( A > 0, α, β, γ ∈ (0,1) , β + γ < 1) where lt denotes labor and kt denotes capital.7 At time t , an output-​producing firm hires labor from the contemporary

76  Niloy Bose, Richard Cothren and Nazanin Sedaghatkish young cohort of households, whereas the supply of capital comes from the savings accumulated by the young households at time t-​1. Both capital and labor are sold to an output producer at competitively determined factor prices and given by β γ wt = Aαltα −1kt gt (6.4)

β −1 γ rt = Aβltα kt gt (6.5)

With the description of the economy in hand, we now proceed to develop a simple framework suitable for the joint determination of corruption and economic growth. To achieve this, we first note a set of state variables such as wage rate, interest rate and taxes that evolve along the growth path and then ask how the propensity toward corruption at a particular time t is influenced by the contemporary values of these state variables. Next, we draw the line in the reverse direction where we seek to learn how the incidences of corruption at a point in time shape the values of these state variables. Finally, we combine these two sets of information to capture the joint evolution of corruption and economic growth.

To Be or Not to Be Corrupt In the analysis that follows, we study the individual incentives of private and public agents to engage in corruption. For the bribery and tax evasion to take place, both high-​income households and bureaucrats must find it advantageous to conspire with each other in concealing information from the government. A high-​income household is willing to pay a bribe if the expected utility from doing so is no less than its expected utility from not engaging in bribery. Equation (6.1) outlines the expected utility of a household under two alternate scenarios. A simple comparison of the two expressions yields that the maximum bribe a household is willing to concede is given by bt = pτt (6.6) Intuitively, the household is prepared to bribe a bureaucrat by no more than what it expects to save in taxes. Similarly, a bureaucrat is willing to accept a bribe if he expects to be no worse off from doing this than from not doing this. From equation (6.2), this requires that bt ≥

(1 − p ) rt +1wt (6.7) µ p (ςrt +1 − ρ)

A Theory of Joint Evolution of Corruption  77 For corruption to take place, both (6.6) and (6.7) must be satisfied simultaneously. This yields the condition pτt ≥

(1 − p ) rt +1wt (6.8) µ p (ςrt +1 − ρ)

Equation (6.8) suggests that the condition that dictates corruption at time t depends on the economy-​wide variables τt , rt +1 and wt . A bureaucrat is more likely to be corrupt the higher is the level of taxes, the higher is the rate of interest and the lower is the level of wages. However, as it will become transparent in the following section, the level of corruption at time t may influence the values of some of these variables inviting the possibility of multiple, frequency-​dependent equilibria. This two-​way causal relationship between the level of corruption and the state of the economy lies at the core of our analysis that is to follow.

Corruption, Tax Rate, Rate of Return and Capital Dynamics The previous analysis sets out the condition based upon which a bureaucrat decides whether or not to engage in corruption. The analysis reveals that such decision depends upon economy-​wide variables –​ wt , τt and rt +1. Now, imagine that period t begins with a given value of kt. As a result, the wage rate, wt , is predetermined.This, however, is not true for both τt and rt +1, which are affected by the time t level of corruption. To see this, recall that gt = θyt and observe that, in equilibrium, lt = l = ( λµ + 1 − µ ).8 From (6.3), (6.4) and (6.5), we then 1/(1− γ )

have gt = Φθktϕ , wt = Φl −1αktϕ and rt = Φβktϕ −1 , where Φ = ( Al α θ γ ) β . Now consider the following two extreme cases: and ϕ = 1− γ Case 1: Absence of Corruption

In this case, we evaluate the effects of corruption on economy-​wide variables when none of the bureaucrats engage in corruption. The government obtains the maximum tax revenue of µτt, which is used to finance its expenditures on public services, gt , and bureaucrats’ salaries, wt . The tax imposed on each high-​income household is then determined from the government’s budget constraint as τˆ t =

gt + wt Φ (θl + α ) ϕ = kt ≡ τˆktϕ (6.9) µ lµ

78  Niloy Bose, Richard Cothren and Nazanin Sedaghatkish Recall that the total savings in the economy at time t contribute toward capital formation at time t + 1 . These savings comprise total savings of low-​ income households, (1 − µ ) wt , of high-​income households, µ ( λwt − τˆ t ) , and of bureaucrats, wt . After collecting these terms together and exploiting (6.9), we may derive the following expression for capital accumulation: ϕ kˆt +1 = lwt − gt = Φ ( α − θ) kt (6.10)

where we assume α > θ, which holds in practice since α and θ are simply the shares of labor and government expenditure in the national income. We also note that since rˆt +1 = Φβkˆtϕ+−11, equation (6.10) may be used to obtain ϕ (ϕ −1) rˆt +1 = Φβ[ Φ( α − θ )]ϕ −1 kt (6.11)

Case 2: Full Prevalence of Corruption Now consider the case where all bureaucrats are corrupt. Clearly, this has different implications for tax rates, rate of return on capital and the path of capital accumulation. To see this, note that a fraction p of the bureaucrat–​ household coalitions will be able to evade detection by the government, while the remaining fraction, 1 − p, will be exposed and penalized. The government’s tax receipts from the former are zero, and from the latter, the government can recuperate µτt from high-​income households and wt + µbt from dishonest bureaucrats as fines. As before, total government expenditure equals expenditures on public services, gt , plus expenditures on bureaucrats’ salaries, wt . It follows from the government’s budget constraint that the tax imposed on each high-​income household is

τ t =

gt + 1 − (1 − p ) wt − (1 − p ) µbt

(1 − p ) µ

where (following equation (6.6)) bt = pτ t .  After substituting the expressions for gt and wt , we rewrite the above expression as τ t = Φ {θl + pα} ktϕ ≡ τktϕ (6.12) {1 − p 2 }l µ It is easy to verify that when there are ample opportunities for the household’s bureaucrat coalition to evade taxes, so that p > 1 − p 2 , the relation τ t > τˆ t holds for any given existing capital stock, kt . Intuitively, the prospect

A Theory of Joint Evolution of Corruption  79 of lost tax revenues under corruption means that the government must raise taxes on high-​income households in order to satisfy its budget constraint. As before, we now evaluate the effects of corruption on the capital dynamics and on the rate of return on capital. The total savings of households comprise the savings of low-​income households, (1 − µ ) wt , and of the savings of high-​ income households, µ  λwt − bt − (1 − p ) τ t  , that engage in tax evasion. Total savings of bureaucrats that contribute toward capital formation come from the wages of only those bureaucrats who can evade detection and are given by pwt. Together with (6.12), the terms may be combined to express the capital accumulation process as ϕ kt +1 = lwt − gt − pµbt = Φ ( α − θ) − p 2 µτ  kt (6.13) ϕ −1 where we assume that [⋅] > 0. Denoting rt +1 = Φβkt +1 , equation (6.13) may be used to obtain

rt +1 = Φβ[ Φ ( α − θ) − p 2 µτ ]ϕ −1 kt

ϕ (ϕ −1)

(6.14)

The expressions derived above lead to the following observations. For any given existing stock of capital, kt , equations (6.10) and (6.13) imply kˆt +1 > kt +1 , and (6.11) and (6.14) imply rˆt +1 < rt +1 . Thus, capital accumulation is higher and interest rates are lower in the absence of any corruption than in the presence of corruption. Intuitively, since τ t > τˆ t , households are therefore willing to pay a larger bribe to evade tax liabilities. Higher taxes, together with the costly concealment of bribe income, reduce savings and capital accumulation, resulting in higher interest rates. In Figure 6.1, we denote the capital accumulation paths for the corruption and no corruption case as path K and path Kˆ and their corresponding steady states by k * and kˆ*, respectively.

The Joint Evolution of Corruption and Growth The analysis presented above brings out two key results. According to equation (6.8), for a given value of kt , the incentive to engage in corruption at time t depends on the tax rate, τt , as well as on the rental rate of capital, rt +1 . At the same time, these two variables are influenced by the level of corruption that is present in the economy at time t. Therefore, the relationship between the incidence of corruption and the state of the economy is fundamentally two-​way causal. Now, we combine these two results to account for the joint evolution of corruption and growth along the path of development. This, however, requires some additional groundwork.

80  Niloy Bose, Richard Cothren and Nazanin Sedaghatkish As a first step, we rewrite equation (6.8) as rt +1 ≥

ρ p 2 µτt ≡ Ω (., τt , wt ) (6.15) p µτt ς − (1 − p ) wt 2

(

)

~ = Ω (., τ = τ~ , w ) ^ = Ω ., τ = τ^ t , w and Ω Next, we define two variables Ω t t t t t where the expressions for τ t and τ t are given by equations (6.9) and ^>Ω ~ . Note that Ω (., τ , w ) as defined by equation (6.12), respectively. Ω t t (6.15) is decreasing in τt . Since τ~ > τ^ , for a given kt (or equivalently for ^>Ω ~ . After substituting w = Φl −1αk ϕ, a given value of wt ), we have Ω t t and the values of τˆ t and τ t from equations (6.9) and (6.12), we obtain the  and Ω  as explicit expressions for Ω  Ω=

ρ p 2l µτˆ (6.16) 2 ˆ p l µτς − Φ (1 − p ) α

and ~ Ω=

ρ p 2l µτ~ (6.17) 2 ~ − Φ (1 − p ) α p l µτς

Finally, recall from equations (6.11) and (6.14) that rˆt +1 < rt +1 , and both r t +1 and rt +1 are decreasing in kt . Given these observations, we may define two  and r~ (k c ) = Ω ~, critical levels of capital, k1c and k2c , that satisfy rˆt +1 (k1c ) = Ω t +1 2 c respectively.We illustrate this in Figure 6.2. Clearly, for all kt k1 , rt +1 (.) >  and ^ . Similarly, for all k < k c , we have r~ (.) > Ω ~ , and for for all kt > k1c , r^t +1 (.) < Ω t 2 t +1 c c c ~ all kt > k2 , r~t +1 (.) < Ω . Evidently, k1 < k2 . It is now easy to pin down the joint evolution of corruption and the level of development.To see this, consider the case where kt < k1c < k2c . Next, consider a behavior profile where corruption is pervasive. As a result, kt +1 = kt +1 , rt +1 = rt +1 , ~ holds for k = k c , and since r is decreasing in k ,  . Since r~t +1 = Ω and Ω = Ω t 2 t +1 t ~ when kt < k2c . Equation (6.15) then implies that no bureauwe have r~t +1 > Ω crat has an incentive to deviate from corrupt behavior. To see that this is a unique equilibrium, consider a behavior profile at the other extreme where at time t no bureaucrat engages in corruption. As a result, kt +1 = kˆt +1 , rt +1 = rˆt +1 ,  . Again, given that rˆ is decreasing in k and that rˆ = Ω  holds and Ω = Ω t +1 t t +1 c c ^. Again, according to equation (6.15), for kt = k1 , kt < k1 implies r^ t +1 > Ω the optimal behavior of a bureaucrat in this situation will be to deviate and engage in corruption. Therefore, noncorrupt behavior cannot exist as equilibrium behavior when the level of development is low, characterized by kt < k1c.

A Theory of Joint Evolution of Corruption  81 The only equilibrium is one in which all bureaucrats are corrupt. A similar argument establishes that for kt > k2c , there exists a unique equilibrium where corrupt behaviors are absent. The predictions that follow from the above analysis accord well with the empirical observations of a high incidence of corruption among low-​income countries and a low incidence of corruption among high-​income countries. However, another notable feature in the data –​one that has received very little publicity –​is the much greater diversity in corruption levels among middle-​ income countries.To account for this, consider the case where k1c < kt < k2c , and  hold simultaneously. As a result, there are ~ and rˆ < Ω the relations r~t +1 > Ω t +1 two candidate equilibria –​one that supports a high incidence of corruption and the other that supports minimum corruption –​that are frequency-​dependent and that are equally likely to arise. This follows from our earlier result that, for any given level of capital, a higher (lower) incidence of corruption is associated with a higher (lower) level of taxes as the government strives to maintain a balanced budget. Higher (lower) taxes means that households are willing to pay larger (smaller) bribes, which, in turn, means that each bureaucrat has a stronger (weaker) incentive to engage in rent-​seeking. In this way, a bureaucrat’s compliance in corruption may depend critically on the compliance of other bureaucrats. In fact, the diversity of equilibrium may be even greater if we allow for mix-​strategy equilibria. To see this, consider the possibility that ε fraction of bureaucrats engage in corruption and the rest do not. The previous analysis suggests that ε will have implications for tax rates τt and the rate of return rt +1 . Also, it is easy to establish that for each kt ∈ (k1c , k2c ), there exists an κ * ∈ (0,1) for which equation (6.8) holds with equality. In this case, the bureaucrats who are otherwise identical may engage in heterogonous behaviors: only κ * fraction of bureaucrats will engage in corruption and the rest will remain honest. Moreover, a middle-​income country may settle in one of three equilibria where the incidence of corruption is high, low or somewhere in between. Surely, the long-​run dynamics of the economy rests on the relationships of the initial capital stock, k0 , with the two critical values, k1c and k2c , and two steady states, kˆ* and k *, as defined by equations (6.10) and (6.13), respectively. Of course, the exact relationship between these variables depends on a variety of exogenous parameters of the model, many of which could potentially represent the legal and/​ or institutional framework of the economy. We conclude this section by highlighting a few possibilities where corruption and poverty can coexist as persistent, rather than transient, phenomena. In Figure 6.1, consider the case where k0 < k * < k1c . In this case, the economy will grow along the lower accumulation path with a high incidence of corruption and will converge to a lower steady state, k * . In other words, if the economy is poor and corrupt, to begin with, then it will be destined to remain poor and corrupt unless fundamental changes take place so as to dictate otherwise. Interestingly, such a prospect does not necessarily disappear when k0 exceeds k *. For example, consider the possibility where k * < k1c < k0 < k2c , and recall that a bad equilibrium with a high incidence of corruption is a distinct possibility

82  Niloy Bose, Richard Cothren and Nazanin Sedaghatkish

Figure 6.1 Capital dynamics.

in the range kt ∈ (k1c , k2c ) . Thus, even with a moderate level of prosperity, an economy can grow along a lower path and be destined toward a poverty trap with a high level of corruption. Such a possibility disappears only when k0 > k2c . Thus, in the joint evolution of corruption and development, a threshold level of prosperity must be attained to ensure further prosperity with a low incidence of corruption.

Conclusion Although the relationship between corruption and growth is complex and moderated by several other factors, including a country’s political and institutional infrastructure,9 the empirical evidence broadly supports the view that good-​quality governance is essential for sustained economic development and that corruption in the public sector is a major impediment to growth and prosperity.10 Corrupt behaviors, however, do not breed in vacuums and are not

A Theory of Joint Evolution of Corruption  83

Figure 6.2 Joint determination of growth and corruption.

impervious to changes in prevailing economic and social conditions. Change in economic and social environments alters the cost–​benefit tradeoffs of engaging in corruption. Therefore, we must view the relationship between corruption and growth within a framework that respects the notion that corruption not only influences but is also influenced by economic development. This chapter offers one such framework that can be exploited to capture the joint evolution of corruption and growth and explains a series of stylized facts present in the data.

Notes 1 Karl Kraus (April 28, 1874–​June 12, 1936) was an Austrian writer and journalist and was nominated for the Nobel Prize in Literature three times. 2 There exists a competing view that corruption can enhance efficiency and raise growth in the presence of cumbersome bureaucratic regulations. Bribes, for instance, are sometimes accepted in exchange for overcoming institutional rigidities that raise inefficiencies (Leff 1964, Huntington 1968, Leys 1970, Lui 1985, Beck and Maher 1986). There is however little evidence to support this efficiency-​enhancing view. For example, Ades and Di Tella (1997) find little evidence of any beneficial effects of corruption in countries mired with red-​tape, while Kaufmann and Wei (1999) conclude that the use of bribes to speed up individual transactions with bureaucrats is largely self-​defeating as the number of transactions tends to increase. 3 Other factors that appear to shape the corruption are the colonial heritage, religious tradition, legal system, federal structure, democratization and openness to trade of a country.

84  Niloy Bose, Richard Cothren and Nazanin Sedaghatkish 4 For example, according to the data from Transparency International, countries such as Bangladesh, Nigeria, Pakistan, Cameroon, India and Kenya have displayed little or no improvement in their corruption since the early 1980s. 5 It is easy to find examples to justify this assumption. Often governments issue bearer bonds in fixed denominations, the proceeds from which on reaching maturity are allowed to be introduced into regular books of accounts without questions being asked about the money that was originally exchanged for the bonds (e.g., the 1981 Special Bearer Bond Act in India).The rate of interest on such bonds is substantially lower than the market rate. Another example is the principle of voluntary disclosure where governments allow individuals to report previously undisclosed income that is taxed at a high rate but grant individuals immunity from prosecution. 6 This strong assumption merely serves to save on notation. One could allow for some but not all illegal income to contribute to productive investments. Alternately, one could assume that investments in informal sector are less productive than investment in the formal sector. Either way, the extent of corruption would impact capital accumulation. 7 The parameter restriction β + γ < 1 ensures the existence of a steady-​state level of capital associated with a strictly concave capital accumulation path. 8 This latter expression defines equilibrium in the labor market, where the total supply of labor is equal to the labor supply of high-​income households, λµ , plus the labor supply of low-​income households, (1 − µ ) . 9 For example, Aidt et al. (2008) and Méon and Weill (2010) indicate that corruption has regime-​specific effects on growth, with weaker effects in countries with weak institutions. 10 For more details, please refer to Ugur (2014) and Campos et al. (2010) for a meta-​ analysis of the literature on corruption and economic growth.

References Ackerman,Rose (1978),Corruption:A Study in Political Economy,NewYork:Academic Press. Ades, Alberto, and Rafael Di Tella (1997), “The new economics of corruption: A survey and some new results.” Political Studies, 45, 496–​515. Aidt, Toke, Jayasri Dutta and Vania Sena (2008), “Governance regimes, corruption and growth: Theory and evidence.” Journal of Comparative Economics, 36, 195–​220. Andvig, Jens Chr, and Karl Ove Moene (1990), “How corruption may corrupt.” Journal of Economic Behavior & Organization, 13, 63–​76. Banerjee, Abhijit V. (1997), “A theory of misgovernance.” Quarterly Journal of Economics, 112, 1289–​1332. Barro, Robert J. (1990), “Government spending in a simple model of endogenous growth.” Journal of Political Economy, 98, S103–​S125. Beck, Paul J., and Michael W. Maher (1986), “A comparison of bribery and bidding in thin markets.” Economics Letters, 20, 1–​5. Blackburn, Keith, Niloy Bose and M. Emranul Haque (2006), “The incidence and persistence of corruption in economic development.” Journal of Economic Dynamics and Control, 30, 2447–​2467. Blackburn, Keith, Niloy Bose and M. Emranul Haque (2010), “Endogenous corruption in economic development.” Journal of Economic Studies, 37(1), 4–​25. Blackburn, Keith, Niloy Bose and M. Emranul Haque (2011),“Public expenditures, bureaucratic corruption and economic development.” Manchester School, 79, 405–​428.

A Theory of Joint Evolution of Corruption  85 Blackburn, Keith, and Gonzalo F. Forgues-​ Puccio (2006), “Financial liberalisation, bureaucratic corruption and economic growth,” Proceedings of the German Development Economics Conference, Berlin 2006 8, Verein für Socialpolitik, Research Committee Development Economics. Cadot, Olivier (1987), “Corruption as a gamble.” Journal of Public Economics, 33, 223–​244. Campos, Nauro F., Ralitza D. Dimova and Ahmad Saleh (2010), “Whither corruption? A quantitative survey of the literature on corruption and growth,” IZA Discussion Papers No 5334. Carrillo, Juan D. (2000), “Graft, bribes, and the practice of corruption.” Journal of Economics & Management Strategy, 9, 257–​288. Ehrlich, Isaac, and Francis T. Lui (1999), “Bureaucratic corruption and endogenous economic growth.” Journal of Political Economy, 107, S270–​S293. Fisman, Raymond, and Jakob Svensson (2007), “Are corruption and taxation really harmful to growth? Firm level evidence.” Journal of Development Economics, 83, 63–​75. Gyimah-​Brempong, Kwabena (2002), “Corruption, economic growth, and income inequality in Africa.” Economics of Governance, 3, 183–​209. Haque, M. Emranul, and Richard Kneller (2009), “Corruption clubs: Endogenous thresholds in corruption and development.” Economics of Governance, 10, 345–​373. Huntington, Samuel P. (1968), Political Order in Changing Societies.Yale University Press. Johnson, Simon, Daniel Kaufmann, Andrei Shleifer, Marshall I. Goldman and Martin L. Weitzman (1997), “The unofficial economy in transition.” Brookings Papers on Economic Activity, 159–​239. Kaufmann, Daniel, and Shang-​Jin Wei (1999), “Does grease money speed up the wheels of commerce?” NBER Working Paper No. 7093. Knack, Stephen, and Philip Keefer (1995), “Institutions and economic performance: Cross-​ country tests using alternative institutional measures.” Economics & Politics, 7, 207–​227. Lambsdorff, Johann Graf (2003), “How corruption affects persistent capital flows.” Economics of Governance, 4, 229–​243. Leff, Nathaniel H. (1964), “Economic development through bureaucratic corruption.” American Behavioral Scientist, 8, 8–​14. Leys, Colin (1970), “New states and the concept of corruption,” in: Heidenheimer, Arnold (ed). Political Corruption (New York: Holt), 341–​345. Lui, Francis T. (1985), “An equilibrium queuing model of bribery.” Journal of Political Economy, 93, 760–​781. Mauro, Paolo (1995), “Corruption and growth.” Quarterly Journal of Economics, 110, 681–​712. Mauro, Paolo (1998), “Corruption and the composition of government expenditure.” Journal of Public Economics, 69, 263–​279. Mauro, Paolo (2004), “The persistence of corruption and slow economic growth.” IMF Staff Papers, 51, 1–​18. Meon, Pierre-​Guillaume, and Laurent Weill (2010), “Is corruption an efficient grease?” World Development, 38, 244–​259. Mo, Pak Hung (2001), “Corruption and economic growth.” Journal of Comparative Economics, 29, 66–​79. Montinola, Gabriella R., and Robert W. Jackman (2002), “Sources of corruption: A cross-​country study.” British Journal of Political Science, 32, 147–​170. Mookherjee, Dilip, and Ivan P. L. Png (1994), “Marginal deterrence in enforcement of law.” Journal of Political Economy, 102, 1039–​1066.

86  Niloy Bose, Richard Cothren and Nazanin Sedaghatkish Paldam, Martin (2002), “The cross-​country pattern of corruption: Economics, culture and the seesaw dynamics.” European Journal of Political Economy, 18, 215–​240. Rauch, James E., and Peter B. Evans (2000), “Bureaucratic structure and bureaucratic performance in less developed countries.” Journal of Public Economics, 75, 49–​71. Reinikka, Ritva, and Jakob Svensson (2004), “Local capture: Evidence from a central government transfer program in Uganda.” Quarterly Journal of Economics, 119, 679–​705. Reinikka, Ritva, and Jakob Svensson (2005), “Fighting corruption to improve schooling: Evidence from a newspaper campaign in Uganda.” Journal of the European Economic Association, 3, 259–​267. Rose-​ Ackerman, Susan (1975), “The economics of corruption.” Journal of Public Economics, 4, 187–​203. Rose-​Ackerman, Susan (1999),“Political corruption and democracy.” Connecticut Journal of International Law, 14, 363. Rose-​Ackerman, Susan (2013), Corruption: A Study in Political Economy. Academic Press. Sarte, Pierre-​Daniel G. (2000), “Informality and rent-​seeking bureaucracies in a model of long-​run growth.” Journal of Monetary Economics, 46, 173–​197. Shleifer, Andrei, and Robert W. Vishny (1993), “Corruption.” Quarterly Journal of Economics, 108, 599–​617. Tanzi, Vito, and Hamid Davoodi (1998), “Corruption, public investment, and growth.” In The Welfare State, Public Investment, and Growth, 41–​60. Springer. Tirole, Jean (1996), “A theory of collective reputations (with applications to the persistence of corruption and to firm quality).” Review of Economic Studies, 63, 1–​22. Treisman, Daniel (2000), “The causes of corruption: A cross-​national study.” Journal of Public Economics, 76, 399–​457. Ugur, Mehmet (2014), “Corruption’s direct effects on per-​capita income growth: A meta-​analysis.” Journal of Economic Surveys, 28, 472–​490. Wei, Shang-​Jin (2000), “How taxing is corruption on international investors?” Review of Economics and Statistics, 82, 1–​11.

7 Corruption and the Financial Sector An Examination of the Literature Arusha Cooray

Introduction Corruption is an area that has generated a significant volume of research over the years.1 Corruption has been found to influence growth (Tanzi and Davoodi 2002, Mauro 1995, Mo 2001), investment (Mauro 1995, Brunetti et al. 1998, Campos et al. 1999), foreign direct investment (Wei 2000, Abed and Davoodi 2002), productivity (Lambsdorff 2003), financial sector development (Cooray and Schneider 2018), emigration (Cooray and Schneider 2016), among other spheres. Studies also show that more corrupt countries face higher inflation (Al-​Marhubi 2000), have larger shadow economies (Johnson et al. 2002, Dreher and Schneider 2010) and channel less expenditure into important areas such as education and health (Mauro 1998). A financial system plays an important role in an economy’s development, by channelling funds from savers to borrowers. The absence of a well-​functioning financial system can reduce the effectiveness of monetary policy, lead to capital flight and increase the risk of financial crises. These conditions may be exacerbated in corrupt environments as credit is channelled away from investors with the most productive investment opportunities to those with political connections, distorting the allocation of resources and generating problems of adverse selection and moral hazard (Cooray and Schneider 2018). The recent financial crisis demonstrated how a dysfunctional financial sector could destabilise the whole globe.Therefore, it is important to have a better understanding of the relation between corruption and the financial sector to minimise the adverse effects of corruption. There has been increased emphasis on the fight against corruption in the recent past, and hence, a study such as this would give policymakers an insight of corruption in the financial sector, placing them in a better position to take the necessary measures to curb corruption.2 There are two competing hypotheses on the effect of corruption on economic activity. One is the ‘sand the wheels’ hypothesis, and the other is the ‘grease the wheels’ hypothesis. According to the former, corruption is detrimental to economic activity. The alternative argument is that corruption can promote economic activity in a second-​best scenario. This view is supported by Nye (1967), Johnson (1975) and Wedeman (1997), among others. The DOI: 10.4324/9781003142300-7

88  Arusha Cooray empirical literature on the effect of corruption on the financial sector has been mixed. While this study does not profess to cover the voluminous literature on corruption and the financial sector in its entirety, it surveys the literature in the area, with the aim of providing a better understanding of the channels through which corruption affects the financial sector. Corruption in the financial sector not only reduces the return on new financial investments and projects but also reduces the return on existing financial projects. A stable economic environment is an important prerequisite for financial investors. A corrupt environment generates uncertainty in returns. How does corruption impact upon the financial sector? Corruption can affect the financial sector through a number of channels. One is through the allocation of loans. In the presence of corruption, the distribution of loans is based on political connections rather than on economic merit (Beck et al. 2006). Under these conditions, credit could be channelled away from productive to unproductive sectors, reducing the allocation of credit to the private sector. Loans granted on the basis of political connections to unproductive, high-​r isk sectors could lead to an increase in the volume of non-​performing loans, forcing banks to increase lending rates, which in turn increases interest rate spreads. Two, corruption can reduce financial sector efficiency by requiring project approvals to pass through many hands (Shleifer and Vishny 1993).This leads to an increase in the number of financial transactions, reducing productivity and efficiency. Three, corruption can affect the financial sector through the lack of transparency, high regulation and lack of proper supervisory measures (Shleifer and Vishny 1993, Williams and Beare 2003, Beck et al. 2006). The secrecy involved in bribe payments reduces the transparency of financial transactions. During periods of financial instability in particular, appropriate disclosure measures of a country’s financial position are necessary for taking corrective measures. The lack of transparency in the presence of corruption could lead to delays in the disclosure of the true state of a country’s financial position (Kaufmann 2010), aggravating conditions of instability and reducing financial sector activity and efficiency (increase non-​ performing loans and interest rate spreads) as the recent financial crisis demonstrated. Four, the high levels of insider trading that takes place in corrupt environments can also affect the efficiency of the financial sector (Du and Wei 2004, Toader et al. 2017). Five, several studies find that the government ownership of banks contributes to financial sector inefficiency (La Porta et al. 2002, Bath et al. 2004, Dicn 2005). All of the aforementioned arguments serve to limit financial sector activity. According to some researchers, corruption could ‘grease’ the wheels of economic activity depending on whether the corrupt monies are consumed or invested, and whether they are retained at home or sent abroad (Johnson 1975, Wedeman 1997). Therefore, according to these arguments, if corrupt monies were retained at home generating investment, then corruption could, in fact, have a positive effect on financial sector development. Leff (1964), Leys (1965) and Huntington (1968) similarly note that corruption could promote investment and insure against bad policies promoting growth in countries with

Corruption and the Financial Sector  89 weak institutions. Heckelman and Powell (2010) argue along the same lines, stating that corruption can promote growth in countries with limited economic freedom; however, the positive impacts of corruption decrease with the increase in economic freedom. Studies also show that corruption could promote economic activity by helping to circumvent bureaucratic delays (Aidt 2003), easing the adverse effects of regulation (Dreher and Gassebner 2013) and increasing lending through the granting of bribes (Fungacova et al. 2015, Cheng et al. 2013). The main purpose of this study is to survey the literature on the relation between corruption and the financial sector, with a view to improving the understanding of the channels through which corruption affects the financial sector. The rest of this study is organised as follows. Section 7.2 surveys the literature on corruption and financial sector activity, and Section 7.3 concludes.

The Literature This section discusses the empirical literature on corruption and financial sector development focusing on (1) studies which hold that corruption is detrimental to financial sector activity –​the sand the wheels hypothesis; (2) studies which maintain that corruption promotes financial sector activity –​the grease the wheels hypothesis and (3) other studies –​those which suggest a non-​linear association in the relation between corruption and economic activity, which explore the relation between corruption, financial sector development and growth and which examine the converse, that is, the effect of financial sector development on corruption. Sand the Wheels Hypothesis Government Ownership of Banks The ‘sand the wheels’ hypothesis suggests that corruption can adversely affect economic activity. This literature highlights the government ownership of banks, inadequate supervisory policies, insider trading, high regulation, political connections, the absence of transparency, among factors which can intensify corruption in the financial sector (Beck et al. 2006, Barth et al. 2004, Du and Wei 2000, Wei and Sievers 1999, Shleifer and Vishny 1993). Several studies show that the government ownership of banks can exacerbate financial sector inefficiency. Employing data on the government ownership of banks from ninety-​two countries around the world, La Porta et al. (2002) test the ‘development’ and ‘political’ theories. Development theories claim that the government ownership of banks enables the government to collect savings and channel these funds towards socially desirable long-​term projects, promoting financial and economic growth. Political theories, on the other hand, state that the government ownership of banks enables the government to finance inefficient, socially undesirable but politically desirable projects. While their results

90  Arusha Cooray indicate some support for the development view, in general the results are found to be more consistent with the political view. That is, the government ownership of banks is associated with lower economic, financial sector and productivity growth, particularly in the developing countries, where the government ownership of banks is high. Similar findings are documented by Bath et al. (2004) who use several regulatory and supervisory controls to investigate two theories of bank regulation and supervision for a sample of 107 countries: the helping hand view, according to which governments that implement strict official oversight of bank activities reduce market distortions and thereby improve bank performance and stability; and the grabbing hand view, which holds that countries which implement official oversight of banks generate higher levels of government corruption without parallel improvements in bank performance or stability. Their results support the grabbing hand theory which predicts that countries with regulatory barriers to bank entry, restrictions on bank activities, greater supervisory power and higher government ownership of banks are more corrupt exhibiting no improvements in bank performance and stability. Evidence along the same lines is documented by Dinc (2005) in a study of political influences on government-​owned banks in emerging economies. He shows that government-​owned banks increase their lending in election years relative to private banks, showing that politicians use government-​owned banks to direct rents to their supporters. Regulation, Intervention and Lack of Supervision Studies also show that there exist greater regulatory restrictions and interventions and a lack of transparency in the presence of corruption, slowing down financial sector development. Shleifer and Vishny (1993) find that corruption can reduce financial sector efficiency due to the secrecy involved in bribe payments and the need for project approvals to pass through many hands.These views are echoed by Williams and Beare (2003), who argue that the lack of transparency diminishes the credibility of a financial system, reducing investor confidence and increasing market volatility.Wei and Sievers (1999) similarly observe that in countries with a higher degree of corruption there exists a higher correlation between inadequate supervisory policies and bank instability. Examining the association between bank supervisory policies and corrupt lending practices on firm’s ability to raise external finance, bank supervision and corruption in lending in 2,500 firms across thirty-​seven countries, Beck et al. (2006) find that political and regulatory intervention is higher in countries with stronger supervisory agencies, hindering the allocation of bank loans. Powerful supervisory agencies are found to reduce the integrity of bank lending. They also observe that private monitoring has positive effects on the integrity of bank lending in countries with strong legal systems and bureaucratic institutions. Kafumann (2010) also observes that the lack of transparency in the presence of corruption leads to delays in the disclosure of the true state of a country’s financial position.

Corruption and the Financial Sector  91 This in turn can lead to delays in taking corrective action. Toader et al. (2017) show that countries which do not observe a corporate governance code and are not members of the European Union are affected more by corruption.They argue that by implementing rigorous corporate governance practices, countries could reduce the adverse effects of corruption. In a study of whether the effect of credit information sharing is based on the degree of corruption in a sample of 120 countries over the 2004–​2017 period, Son et al. (2020) find that information sharing reduces the negative effects of corruption on financial sector development, with public credit registries playing a more important role in this regard compared to private credit bureaus. Insider Trading and Risk-​Taking Behaviour Du and Wei (2004) investigate the role of insider trading as an explanation for cross-​country differences in stock market volatility. They find that countries with a greater degree of insider trading have more volatile stock markets. Similar findings are reported by Chen et al. (2015), who find that corruption increases banks’ risk-​taking behaviour. Using bank-​level data from over 1,200 banks in thirty-​five emerging economies during the 2000–​2012 period, they find that higher levels of corruption increase the risk-​taking behaviour of banks, consistent with the ‘sand the wheel’ hypothesis. These arguments are reiterated by Ben-​Ali et al. (2020), who show that corruption increases the probability of banking crises for a sample of thirty-​eight countries over the period 2000–​2017. This is attributed to excessive risk-​taking in corrupt environments, which in turn increases interest margins of banks, attracting more risky borrowers, increasing non-​performing loans and the probability of crises. Toader et al. (2017) in a study of corruption on banking stability in a sample of emerging markets also confirm that lower levels of corruption are associated with greater bank stability and fewer credit losses. Non-​performing Loans Higher levels of corruption are also associated with higher non-​performing loans. Goel and Hasan (2011), investigating the effects of corruption on bad loans in a sample of countries, find that higher levels of corruption are associated with more bad loans. Loan defaults are found to be lower in faster-​growing economies, in countries with better-​developed financial systems and in countries in the Euro zone, ceteris paribus. In a study of the impact of corruption on the asset quality of banks in twenty-​two emerging market economies over the 2008–​2012 period, Bougatef (2016) also finds that corruption aggravates the problem of non-​performing loans through the misallocation of loanable funds. Division of the sample into two based on the level of corruption shows that collateral and bankruptcy laws reduce the impact of corruption on loan portfolios in low-​corrupt countries.

92  Arusha Cooray Political Connections The studies of Khawaja and Mian (2005), Clasessens et al. (2008) and Chen et al. (2013) examine the relation between financial sector development and politically connected firms. Khwaja and Mian (2005), employing a dataset on loans for over 90,000 firms that represent corporate lending in Pakistan between 1996 and 2002, examine the volume of rents to politically connected firms in the banking sector. They observe that politically connected firms borrow 45 per cent more than other firms and have 50 per cent more default rates. They also find that political rents increase with the power of the firm’s politician and whether the politician or politicians party is in power. Similar findings are documented by Clasessens et al. (2008) for Brazil. Using campaign contribution data for Brazilian firms, they find that firms which provided contributions to (elected) federal deputies faced higher stock returns compared to firms that did not in the 1998 and 2002 elections. The results suggest the importance of political connections for firm finance in Brazil. Chen et al. (2013) observe the same for China. They show that bribery, rather than firm’s performance, largely influences the degree to which private firms can access bank credit. These firms are also found to pay more in terms of bribes. Satisfactory firm performance determines firm’s access to loans, only by the big four banks. For loans granted by smaller banks, bribes rather than performance determine access to loans. Financial Policy Haggard and Lee (1993) observe that in developing countries often the personal interests of policymakers lead to a distortion in the allocation of financial resources. They argue that it is therefore important to consider the political economy features of a nation when investigating its financial sector. According to them, economic and political factors play a major role in influencing changes in financial policy such as financial liberalisation in developing countries. They also note that politically independent states are better able to control corruption. The importance of political economy factors in finance is also emphasised by Blackburn and Forgues-​Puccio (2010), who develop a theoretical model to examine how the liberalisation of financial markets impacts upon the long-​run development of an economy in which the level of corruption could change endogenously. They find that corruption affects economic development adversely, but its effects are more severe in the presence of financial market liberalisation and in financially more open economies compared to financially closed economies, and in poor countries rather than rich countries.They also find that financial liberalisation promotes development in countries in which governance mechanisms are strong, but not in those in which governance structures are weak. Their results are consistent with theoretical predications highlighting the need to consider the political economy aspects of financial liberalisation.

Corruption and the Financial Sector  93 This section has summarised the findings of studies which show why corruption can have detrimental effects on the financial sector.The next section will go on to examine studies which show the converse. Grease the Wheels Hypothesis According to this hypothesis, corruption can have a positive effect on financial sector development in countries with a weak rule of law and poor institutions by hastening financial contract enforcement through the circumvention of restrictive rules (Bardhan 1997). Competition for Resources Some researchers argue that corruption can improve financial sector efficiency by creating competition for limited government resources, leading to a more efficient provision of financial services than would otherwise occur (Aidt 2003, Lambsdorff 2002). For example, a bribe to a bank official who is allocating credit which is limited in supply may be viewed as a payment for ensuring that resources go to the person who is most likely to use them efficiently –​in other words the person who can pay the highest bribe (World Bank 1997). Generate Investment Johnson (1975) argues that if corrupt monies are invested in productive activities with forward and backward linkages, corruption might in fact increase growth rates.This view is supported by Wedeman (1997), who identifies between three types of corruption –​rent-​scraping, looting and dividend collecting.3 Wedeman notes that economic policies designed with the objective of making rents (rent-​ scraping) could have one of two effects. It could lead to a fall in the rate of return on capital investment leading to capital outflows; alternatively, it could lead to new investments and growth, in which case the money would remain at home. Wedeman argues that looting generates insecurity, leading to the consumption of money or the transfer of money out of the country. Dividend-​collecting, on the contrary, depends on the success of firms to generate profits, which gives officials the incentive to implement policies to promote investment. Overcoming Regulation and Weak Institutions Dreher and Gassebner (2013), examining if corruption is an efficient grease leading to a fall in the negative impact of regulation on entrepreneurship in highly controlled economies, find support for the grease in the wheels hypothesis. In a sample of forty-​three countries with data covering the 2003–​2005 period, they find that corruption leads to firm’s entry in highly regulated nations. Meon and Weil (2010) document similar findings for a panel of sixty-​ nine developed and developing countries. They find that corruption is less

94  Arusha Cooray detrimental to promoting economic activity in countries where institutions are less effective. Corruption could even be positively associated with efficiency in countries with very weak institutions.These findings are supported by Mendoza et al. (2015), who similarly find that corruption promotes commerce in SMEs in the Philippines, particularly in cities with weak business environments which are delayed by bureaucratic red tape. Higher Borrowing, Lending and Credit Constraints Fungacova et al. (2015) find that bribery leads to an increase in firms’ bank debt ratios in a sample of firms in fourteen transition countries. Higher levels of bribery, therefore, lead to more bank lending through bribes given to bank officials. Bribery is found to lead to easier access to bank credit in the presence of weak financial sectors. Similar conclusions are reached by Cheng et al. (2013) for a group of firms in China. According to them, bribery rather than firm’s performance determines access to loans. They argue that in developing countries with weak legal frameworks, corruption plays a role in improving efficiency and supporting private firm growth. Employing a detailed dataset on Thai firms before the Asian crisis of 1997, Charumilind et al. (2006) similarly note that the presence of connections was the most important factor determining access to long-​term bank debt prior to the financial crisis. They find that firms with connections to banks and politicians had greater access to long-​ term debt compared to firms with no ties. Connected firms required lower collateral, received more long-​term loans and used fewer short-​term loans compared to those without connections. They also observe that as these politically connected firms were able to obtain more long-​term loans and were less credit-​constrained, they were better able to withstand the financial crisis of 1997, compared to those with no connections. Hence, in a second-​best scenario, corruption can have positive effects on the financial sector. Other Studies Non-​linearity Studies also suggest that the relation between corruption and growth may be non-​linear (Méndez and Sepúlveda 2006, Aidt et al. 2008). Méndez and Sepúlveda (2006) show that the relationship between corruption and growth is quadratic, and that this relationship depends on the degree of political freedom in a country. It is therefore not unreasonable to expect corruption to have non-​linear effects on financial sector development with the financial sector initially progressing with corruption and then falling as corruption continues to increase.

Corruption and the Financial Sector  95 Finance, Growth and Corruption Ahlin and Pang (2008) examine the question of whether financial development and the control of corruption act as substitutes in promoting growth. Developing a model in which both low corruption and financial development expedite the undertaking of productive projects, they find that the improvement of both leads to positive effects as well reflects a substitutability between them. Song et al. (2020) examine the relation between financial sector development, economic growth and corruption for a sample of 142 countries. They find that financial sector development is affected by both growth and corruption. While growth has a positive effect on financial sector development, corruption has a negative effect. Splitting the sample into developed and developing countries, they do not observe a causal relation between the three variables for the developed countries, but for the developing countries, growth promotes financial sector development by reducing corruption. Jallas (2016) develops a theoretical model to see if lower levels of corruption lead to financial sector development and economic growth. Assuming that technologically more advanced firms are located in developed countries, and less advanced firms in poorer countries, he concludes that stronger institutions or low levels of corruption in developed countries are growth-​enhancing due to more efficient financial systems compared to poorer countries. Stronger institutions increase bank entry by reducing banking screening costs and fostering innovation. From Finance to Corruption There are studies which also look at the converse, that is, the effect of financial sector development on corruption. Jha (2020) finds that the removal of entry barriers to the financial sector, through greater autonomy in credit supply decisions, securities market development and increased supervision, could help reduce corruption.The results suggest that developing countries could experience greater benefits compared to developed countries through financial liberalisation. Altunbaş and Thornton (2012) similarly show that financial development as measured by bank credit to the private sector as a share of GDP reduces corruption in a panel of 107 developed and developing countries. This is attributed to the better monitorin g of borrowers by a developed financial sector, which leads to a more efficient allocation of resources. Exploring the role of financial sector development on the control of corruption in a sample of 140 countries over the 1996–​2015 period, Sharma and Paramati (2020) show that financial development plays an important role in reducing the growth of corruption across the panel by enhancing competition among banks and reducing credit costs.

Conclusions This study has reviewed the literature on the role of corruption in affecting financial sector development. The study leads to the following conclusions. First, there is overwhelming support for the argument that corruption damages

96  Arusha Cooray financial sector development. Second, corruption promotes financial sector development in a second-​best scenario. Third, strong institutions are found to reduce the adverse effects of corruption on the financial sector. Given the significance of the financial sector for economic growth, it is important to promote strong institutions through sound macroeconomic management, greater transparency, effective financial regulation and supervisory mechanisms and bureaucratic accountability. As these inefficiencies are created by government officials in the first place, it is important for countries with weak institutions to get the assistance of international bodies to strengthen the rule of law and improve governance mechanisms. A stronger financial system will increase the efficiency of loan allocation, the efficacy of monetary policy and reduce the risk of financial crises.

Acknowledgement I wish to thank Justin Esarey for valuable comments.

Notes 1 Corruption is defined as the abuse of public power for private gain (Buehn and Schneider 2009). 2 See Bahoo (2020) for a bibliometric review of corruption in banks, Bardhan (1997) for a review of literature on corruption and development, and Dimant and Tosato (2017) for a survey of literature on corruption. 3 Wedeman (1997) defines looting as theft of public funds and property, rent-​scraping as the manipulation of macroeconomic policy to produce rents, and dividend-​ collecting as the transfer of a certain proportion of profits earned by privately owned firms to government officials.

References Abed G and Davoodi H (2002) ‘Corruption, Structural Reforms, and Economic Performance in the Transition Economies’, in George T. Abed and Sanjeev Gupta (Eds), Governance, Corruption, & Economic Performance, Washington, DC: International Monetary Fund Publication Services, pp. 489–​537. Ahlin C and Pang J (2008) ‘Are Financial Development and Corruption Control Substitutes in Promoting Growth?’ Journal of Development Economics, 86, 414–​433. Aidt T (2003) ‘Economic Analysis of Corruption: A Survey’, Economic Journal, 113, 632–​652. Aidt T, Dutta J and Sena V (2008) Governance regimes, corruption and growth: Theory and evidence, Al-​Marhubi F (2000) ‘Corruption and Inflation’, Economics Letters, 66, 199–​202. Altunbaş, Y., & Thornton, J. (2012). Fiscal Decentralization and Governance. Public Finance Review, 40(1), 66–​85. Bahoo S (2020) ‘Corruption in Banks: A Bibliometric Review and Agenda’, Finance Research Letters, 35, 1011499. Bardhan P (1997) ‘Corruption and Development: A Review of Issues’, Journal of Economic Literature, 35, 1320–​1346.

Corruption and the Financial Sector  97 Barth J, Caprio Jr G and Levine R (2004) ‘Bank Regulation and Supervision: What Works Best?’, Journal of Financial Intermediation, 13, 205–​248. Beck T, Demirguc-​Kunt A and Levine R (2006) ‘Bank Supervision and Corruption in Lending’, Journal of Monetary Economic, 53, 2131–​2163. Ben-​Ali M, Fhima F and Nouira R (2020) ‘How Does Corruption Undermine Banking Stability? A Threshold Nonlinear Framework’, Journal of Behavioural and Experimental Finance, 27, 100365. Blackburn K and Forgues-​Puccio GF (2010) ‘Financial Liberalization, Bureaucratic Corruption and Economic Development’, Journal of International Money and Finance, 29, 1321–​1339. Bougatef K (2016) ‘How Corruption Affects Loan Portfolio Quality in Emerging Markets?’ Journal of Financial Crime, 23(4), 769–​785. Brunetti A, Kisunko G and Weder B (1998) ‘Credibility of Rules and Economic Growth: Evidence from a Worldwide Survey of the Private Sector’, World Bank Economic Review, 12, 353–​384. Campos E, Lien D and Pradhan S (1999) ‘The Impact of Corruption on Investment: Predictability Matters’, World Development, 27, 1059–​1067. Charumilind C, Kali R and Wiwattanakantang Y (2006) ‘Connected Lending: Thailand before the Financial Crisis’, Journal of Business, 79, 181–​218. Chen M, Jeon BN,Wang R andWu J (2015) ‘Corruption and Bank Risk-​Taking: Evidence from Emerging Economies’, Emerging Markets Review, 24, 122–​148. Chen Y, Liu M and Su J (2013) ‘Greasing the Wheels of Bank Lending: Evidence from Private Firms in China’, Journal of Banking and Finance, 37(7), 2533–​2545 Claessens S, Feijen E and Laeven L (2008) ‘Political Connections and Preferential Access to Finance: The Role of Campaign Contributions’, Journal of Financial Economics, 88, 554–​580. Cooray A and Schneider F (2016) ‘Does Corruption Promote Emigration? An Empirical Examination’, Journal of Population Economics, 29(1), 293–​310. Cooray A and Schneider F (2018) ‘Does Corruption Throw Sand into or Grease the Wheels of Financial Sector Development?’, Public Choice, 177, 111–​133. Dimant E and Tosato G (2018) ‘Causes and Effects of Corruption: What Has Past Decade’s Empirical Research Taught Us? A Survey’, Journal of Economic Surveys, 32, 335–​356 Dinc S (2005) ‘Politicians and Banks: Political Influences on Government-​Owned Banks in Emerging Markets’, Journal of Financial Economics, 77, 453–​479. Dreher A and Gassenbner M (2013) ‘Greasing the Wheels? The Impact of Regulation and Corruption on Firm Entry’, Public Choice, 155, 413–​432. Dreher A and Schneider F (2010) ‘Corruption and the Shadow Economy: An Empirical Analysis’, Public Choice, 144, 215–​238. Du J and Wei S (2004) ‘Does Insider Trading Raise Market Volatility?’, Economic Journal, 114, 916–​942. Fungacova Z, Kochanova A and Weill L (2015) ‘Does Money Buy Credit? Firm-​Level Evidence on Bribery and Bank Debt’, World Development, 68, 308–​322. Goel RK and Hasan I (2011) ‘Economy-​ Wide Corruption and Bad Loans in Banking: International Evidence’, Applied Financial Economics, 21(7), 455–​461. Haggard S and Lee C (1993) ‘The Political Dimension of Finance in Economic Development’, in S. Haggard, C. Lee and S. Maxfield (Eds), The Politics of Finance in Developing Countries, New York: Cornell University Press, pp. 3–​20. Heckelman J and Powell B (2010) ‘Corruption and the Institutional Environment for Growth’, Comparative Economic Studies, 52, 351–​378.

98  Arusha Cooray Huntington S (1968) Political Order in Changing Societies, New Haven:Yale University Press. Jalles J (2016) ‘A New Theory of Innovation and Growth: The Role of Banking Intermediation and Corruption’, Studies in Economics and Finance, 33(4), 488–​500. Jha C (2020) ‘Financial Reforms and Corruption: Which Dimensions Matter?’, International Review of Finance, 20:2, 515–​527. Johnson E (1975) ‘An Economic Analysis of Corrupt Government, with Special Application for Less Developed Countries’, Kyklos, 28, 47–​61. Johnson S, McMillan J and Woodruff C (2002) ‘Property Rights and Finance’, American Economic Review, 92, 1335–​1356. Journal of Comparative Economics, 36, 195–​220. Kaufmann D (2010) ‘Can Corruption Adversely Affect Public Finances in Industrialized Countries?’, Brookings (www.brooki​ngs.edu/​resea​rch/​opini​ons/​2010/​04/​19-​cor​ rupt​ion-​kaufm​ann, retrieved July 2013). Khwaja AI and Mian A (2005) ‘Do Lenders Favor Politically Connected Firms: Rent Provision in an Emerging Financial Market’, Quarterly Journal of Economics, 120(4), 1371–​1411. La Porta R, Lopez-​de-​Silanes F, Shleifer A and Vishny RW (1998) ‘Law and Finance’, Journal of Political Economy, 106, 1133–​1155. La Porta, R., Lopez-​De-​Silanes, F. and Shleifer, A. (2002), Government Ownership of Banks. The Journal of Finance, 57: 265–​301. Lambsdorff JG (2002) ‘Corruption and Rent Seeking’, Public Choice, 113, 97–​125. Lambsdorff J-​G (2003) How Corruption affects Productivity, Kyklos, 4, 457–​474. Leff N (1964) ‘Economic Development through Bureaucratic Corruption’, American Behavioral Scientist, 82, 337–​341. Leys C (1965) ‘What Is the Problem about Corruption?’, Journal of Modern African Studies, 3, 215–​230. Reprinted in Political Corruption: A Handbook, AJ Heidenheimer, M Johnston and VT LeVine (Eds), Oxford: Transaction Books, 1989, pp. 51–​66. Mauro P (1995) ‘Corruption and Growth’, Quarterly Journal of Economics, 110, 681–​712. Mauro P (1998) ‘Corruption and the Composition of Public Expenditure’, Journal of Public Economics, 69, 263–​279. Mendez F and Sepulveda F (2006) ‘Corruption, Growth and Political Regimes: Cross Country Evidence’, European Journal of Political Economy, 22, 82–​98. Mendoza R, Lim R and Lopez A (2015) ‘Grease or Sand in the Wheels of Commerce? Firm Level Evidence on Corruption and SMES’, Journal of International Development, 27, 415–​439. Meon P and Weil L (2010) ‘Is Corruption an Efficient Grease?’, World Development, 38, 244–​259. Mo P (2001) ‘Corruption and Economic Growth’, Journal of Comparative Economics, 29, 66–​79. Nye J (1967) ‘Corruption and Political Development:A Cost–​Benefit Analysis’, American Political Science Review, 61, 417–​427. Rose-​Ackerman S (1999) Corruption and Government, Causes, Consequences and Reform, Cambridge: Cambridge University Press. Sharma, C, Paramati, SR (2021). Does financial development reduce the level of corruption? Evidence from a global sample of 140 countries. Inernational Journal of Financial Economics, 26: 5093–​5109. Shleifer A andVishny R (1993) ‘Corruption’, Quarterly Journal of Economics, 108, 599–​617. Son TH, Khanh HCG and Liem NT (2020) ‘Credit Information Sharing, Corruption and Financial Development: International Evidence’, Cogent Business & Management, 7(1). DOI: 10.1080/​23311975.2020.1851856

Corruption and the Financial Sector  99 Song C, Chang C and Gong Q (2020) ‘Economic Growth, Corruption, and Financial Development: Global Evidence’, Economic Modelling, 94, 822–​830. Tanzi V and Davoodi H (2002) ‘Corruption, Public Investment, and Growth’, in GT Abed and S. Gupta (Eds), Governance, Corruption, & Economic Performance,Washington, DC: International Monetary Fund Publication Services, pp. 280–​299. Toader T, Onofrei M, Popescu A and Andrieș AM (2017) ‘Corruption and Banking Stability: Evidence from Emerging Economies’, Emerging Markets Finance and Trade, 54(4), 591–​617. Wedeman A (1997) ‘Looters, Rent-​Scrapers, and Dividend-​Collectors: Corruption and Growth in Zaire, South Korea, and the Philippines’, Journal of Developing Areas, 31, 457–​478. Wei S (2000) ‘How Taxing Is Corruption on International Investors?’, Review of Economics and Statistics, 82, 1–​11. Wei S and Sievers S (1999) ‘The Cost of Crony Capitalism’, in Asian Competitiveness Report, Geneva: World Economic Forum, pp. 50–​55. Williams J and Beare M (2003) ‘The Business of Bribery: Globalization, Economic Liberalization and the “Problem” of Corruption’, in M. Beare (Ed.), Critical Reflections on Transnational Organized Crime, Money Laundering and Corruption,Toronto: University of Toronto, pp. 88–​119. World Bank (1997) Helping Countries Combat Corruption: The Role of the World Bank, Washington, DC: World Bank.

8 Stopping the Rot I A Review of Models and Experimental Methods of Corruption Experiments Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra

Introduction Corruption impedes economic growth by misallocating talent, technology, capital, investment and entrepreneurial skills (Svensson, 2005). Over the years, corruption has attracted the attention of social scientists. Economists, in particular, have been extensively involved in analyzing the phenomenon and understanding its diverse forms, leading to some seminal work in the area (Becker, 1968; Krueger, 1974; Rose-​Ackerman, 1975). While corruption typically goes hand in hand with a failure of governance and poor state capacity, the academic investigation into corruption has gone beyond the realm of political economy, entrenching itself firmly into the ambits of behavioral science. Past and current academic literature seems to have arrived at the conclusion that corruption needs to be looked at as an individual choice problem, subject to incentives and institutions. Consequently, a richer understanding of this phenomenon necessitates a framework of not just the economic costs and benefits but also the psychology behind the genesis of such unethical choices. This behavioral science–​led lens of looking at corruption is now well developed and offers social and psychological insights on human behavior that are unique and valuable at the same time (Abbink and Hennig-​ Schmidt, 2006; Basu, 2015; Gneezy et al., 2019). Corruption is particularly difficult to study since anyone involved in such activities has obvious reasons to be discreet. Researchers have taken many different approaches to overcome this problem. In particular, the use of laboratory experiments to quantify corruption and evaluate the economic, social and psychological impacts of anticorruption policies has been a prominent path of research in the corruption literature. Typically, in a laboratory experiment, subjects participate as decision-​ makers in constructed/​ artifactual situations with incentive structures mimicking those of a real-​world choice environment. Such experiments can be particularly useful when a researcher is interested in unearthing the behavioral aspects of corrupt decision-​making as well as the saliency of anticorruption policies. A comprehensive review of this literature can be found in Abbink and Serra (2012). However, the number of papers that have explored the psychological underpinnings of corruption has grown DOI: 10.4324/9781003142300-8

Stopping the Rot I  101 exponentially in the last ten years. Consequently, we will focus primarily on lab experiments that have been published post Abbink and Serra (2012). To situate our review better, we start by laying out some of the prominent frameworks for analyzing corrupt behavior and their implications toward policy formulations as well as experimental evaluations. Next, we explain the usefulness of lab experiments along with the extant understanding of the implications for their generalizability, representativeness and external validity. Subsequently, we discuss the typology of corruption games to help understand the canonical structures that have been used to analyze corruption in the experimental literature. We conclude by synthesizing the main strands of the behavioral research that have been discussed in this chapter and propose some avenues of research going forward.

Models of Corruption Becker’s Model, Self-​Interest and Reciprocity The decision to engage in corruption or abstain from it, ceteris paribus, depends on individuals’ tendencies to maximize costs and benefits to get the best payoff out of their choice. In other words, the choice of whether to engage in corruption is determined by comparing the potential gains and losses from their decision. Such an opportunity may arise whenever there is a scope for private gains. In such cases, an individual chooses corruption when the expected benefits outweigh the expected costs. Becker (1968), in his seminal paper, hypothesized that a rational agent’s decision to commit a crime is like making any other economic choice, that is, she compares costs and benefits from a criminal act and maximizes the expected utility, which may be given as: EU (offense ) = pU (Y − f ) + (1 − p )U (Y ) where Y indicates payoff (monetary and psychic) from the offense, U is the utility function and f is the penalty.This model has been widely institutionalized to explain corruption (Abbink et al., 2002; Banerjee and Mitra, 2018; Friesen, 2012; Khadjavi, 2015; Schulze and Frank, 2003). Becker’s model has remained the foundational framework for the widely adopted deterrence hypothesis of corruption. The deterrence hypothesis suggests that any measure that increases the cost of corrupt behavior by increasing the penalty ( f ) should deter the propensity of being corrupt by decreasing the expected utility from committing the offense. The empirical validity of the deterrence hypothesis has been tested frequently. Eide (2000) provides a detailed review of nonexperimental studies that use data from states, police regions, campuses and individuals and finds that most studies corroborate the hypothesis. However, there are important methodological issues in such empirical studies (Friesen, 2012). First, the individual-​level data used in

102  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra these empirical studies mostly come from the criminal justice system. This creates sample selection issues primarily because these data include only those with a history of offense or are self-​reporting individuals. Second, when using aggregate data, it is necessary to consider the endogeneity of crime rates and enforcement parameters. Finally, the constructed measures of probabilities and penalties used in the estimations are typically proxies based on past data and are likely to differ from actual individual perceptions. Besides the standard criticisms of empirical studies, Anderson (2002) questions the effectiveness of the deterrence hypothesis on the grounds of satisfying some basic assumptions. Using results from their prison survey, he argues that criminals are often unaware of the two necessary conditions for the deterrence hypothesis: (1) sufficient information about the probabilities of being detected and (2) consequent punishment meted out.Thus, criminals fail to behave as rational agents of Becker’s model. Many of these concerns can be appropriately addressed in a controlled laboratory setting to complement and improve some of these empirical findings. In particular, laboratory environments allow for direct observations of individual compliance decisions and, importantly, allow the experimenter to vary the probability and severity of punishment in a controlled manner to allow for a more direct test of the underlying theories and a cleaner understanding of causality (Friesen, 2012). Becker’s model assumes that an agent pursues self-​interest, that is, maximizes her own payoff. However, we often tend to reward kind actions and punish unkind ones (Falk and Fischbacher, 2006), which Becker’s standard model does not capture. A simple representation of such a model can be depicted as follows. The utility (U ) from bribe (b) is the sum of pecuniary payoff ( π ) from the expected utility of Becker’s self-​regarding model and utility ( σ ) from the desire to reciprocate. U (b ) = π (b ) + ρσ (b ) where ρ is the reciprocity parameter. Positive and higher the value of ρ, more important is reciprocity utility than utility arising from self-​interest. If ρ = 0, an individual behaves in self-​interest, and there is no element of reciprocity. In the context of corruption, bribes indicate an explicit request for positive reciprocity from the person receiving the bribe. The briber’s success depends on the reciprocity of the bribe-​taker. For example, in Gneezy et al. (2019), when players can bribe the referee to win, bribes may distort the decision-​ maker’s judgment. If the referee keeps the bribe regardless of the winner, bribes no longer influence the referee’s decision. This implies that if bribe-​ receiving has no contingency on the briber’s success, then reciprocity ceases to exist. Manipulating the experimental design, Gneezy et al. (2019) clearly distinguish between self-​interest and reciprocity. The results suggest that bribes influence referees out of self-​interest, implying self-​interest subdues the desire to reciprocate.

Stopping the Rot I  103 Gift Exchange The exchange of gifts is a universal phenomenon that aims to strengthen bonds and create reciprocity in a society. However, on receiving the gift, what is expected in return? When can gifts be labeled as bribes? Both gifts and bribes are the quintessential examples of social and institutional complexity. However, both have three standard functions at the societal level. First, they trigger reciprocity; second, they regulate the exchange process; third, they necessitate a quid pro quo (Graycar and Jancsics, 2017).The effect of gift-​giving is important to understand to predict the possible economic distortions that may arise. It is observed that even when there is no monetary incentive, and there is a complete understanding that a gift is given for a selfish reason, recipients still reciprocate. Also, individuals may show negative reciprocity when a gift is expected but not received (Malmendier and Schmidt, 2017). Following this, Czap and Czap (2019) look at the behavior of bribe recipients in an environment where rejecting a bribe is an option. Allowing an option to refuse a gift liberates people from any quid pro quo on receiving the bribe involuntarily. Also, recipients of such bribes ended up punishing the bribers. On the other hand, when accepting a bribe is voluntary, the public official becomes an accomplice and becomes more likely to reciprocate. This implies that in situations where accepting the bribe is optional, reciprocity may be higher. Republic of Beliefs Recently, Basu (2015, 2020), noting that the anticorruption policies based on the Beckerian model of crime and punishment have met with mixed successes, introduces an interesting take on explaining the modest success of the crime prevention policies. He points out that the Beckerian framework assumes that once a law is enacted, violators will automatically be put to justice. However, if we pause to consider who oversees enforcing the law, the answer invariably would be that an enforcer must be waiting in the wings, ready to serve justice in the case of a transgression. Basu argues that here lies the problem: it is assumed in the classical law and economics framework that the enforcer acts like a mere robot and detects and implements the law without fail whenever required. In fact, the power of this model of crime and punishment lies in the understanding that everyone needs to believe that the police will act judiciously, failing which the local court will act appropriately, failing which the chief justice can act appropriately, and so on. However, the deterrent of fines and punishments end up merely as “ink on paper” if everybody instead decides to ignore them, accruing the same payoffs from actions chosen before the enactment of a new law (Basu, 1993). Consequently, only if the newly enacted law changes the associated beliefs of the players will they abide by the law, and the enacted law would have the potential to curb corrupt behavior. Dasgupta and Radoniqi (2021) test this key idea in a laboratory environment. They manipulate the subjects’ beliefs to find the support for Basu’s core proposal that

104  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra suggests that introducing an anticorruption law on its own does not guarantee a move toward an equilibrium with less corruption. This is especially true for developing countries distraught with leakages of the existing legal enforcement machinery. One needs to find ways to change ingrained beliefs that go hand in hand with the introduction of a new law to experience positive changes in law-​abiding behavior.

Use of Lab Experiments as a Methodological Tool Laboratory experiments can provide a useful platform to study preferences, strategic behavior and the effectiveness of policies. By introducing different treatments, the experimenter is able to control both environmental and population characteristics within the experiment. This allows the researchers to disentangle the effects of possible deterrents of corruption and identify its micro determinants. The process enables them to comment on causal factors related to corruption (Armantier and Boly, 2012). In the real world, all factors work in tandem, and it is nearly impossible to observe the cause and effect of one separated from others. Varying the institutional environment across treatments enables researchers to test mechanisms that are at play and explore the conditions under which policy measures may work. Setting aside the issue of external validity for a moment, laboratory experiments can allow researchers and policymakers to overcome the concern of the lack of observability of corrupt behavior in more natural settings. Importantly, if the treatment effect is of primary interest instead of the absolute level, which is often the case when choosing from a set of policy prescriptions, a laboratory corruption game is a practical and relatively inexpensive tool for researchers (Banerjee, 2016). However, a frequent criticism of laboratory-​driven corruption experiments is about the external validity and generalizability of their results. In response, experimental economists have dedicated studies to establishing the external validity of lab-​generated results over the years. The following subsection discusses selective studies in the last two decades that aim to address external validity concerns by answering the following questions:To what extent laboratory results are replicable in the field? Do results hold across cultures? Are there variations in corrupt behavior when the subject pool changes? Can policy regimes be replicated in the lab convincingly? And importantly, how do corruption experiments account for psychological factors such as framing, moral costs and Hawthorne effects? External Validity and Generalizability A prominent criticism of the lack of generalizability of lab-​ based results comes from Levitt and List (2007). However, this paved the way for the lab experimentalists to establish the importance of lab experiments as a valid tool to collect data for research, particularly from the point of view of generalizability of lab-​based results (Alm et al., 2015; Camerer, 2011). Kessler and Vesterlund (2015) put forward the following set of arguments in favor of the

Stopping the Rot I  105 external validity of lab experiments. First, they argue that lab experiments test the general principles of behavior. Therefore, they are by definition generalizable. Second, since lab experiments differ from field experiments in terms of incentives, rules, norms and so on, it is only relevant to establish external validity in directional or qualitative results and not in terms of the quantitative effects. Contrasting nonexperimental and field-​based experimental research to laboratory experiments, Falk and Heckman (2009) argued that the criticism of lab experiments on “realism” and “generalizability” is often misguided, and the choice of method should ultimately be a matter of the underlying research question. Field and Lab-​Based Experiments: Corruption and Dishonest Behavior While observational data-​based research is thought to be the gold standard for generalizability, field experiment enjoys a reputation that is a cut above lab-​ based ones. Armantier and Boly (2012) are one of the first to explicitly establish the external validity of lab vs. field. They ran their experiments in three environments –​one lab experiment and one field experiment in a developing country (Burkina Faso), as well as a lab experiment in a developed country.The lab experiment was designed to model an environment where a player could bribe a grader to obtain a better grade.The field experiment measured whether a part-​time grader of an actual exam increased grades in exchange for real money tucked inside the answering script. Results show that the probability of accepting bribes is statistically identical in the lab and field environments in Burkina Faso. This provides support for the idea that lab-​and field-​based measures identify the same behavioral primitive. Cohn and Maréchal (2018) reached a similar conclusion when comparing the incentivized coin-​toss task in the laboratory to different indicators of misconduct at school. In particular, they found that students who cheated more in the laboratory were more likely to misbehave at school, establishing that the outcomes from their laboratory experiment on cheating are reliable predictors of rule-​violating behavior in the field. Dai et al. (2018) furnish further evidence by experimenting with bus passengers to study attitudes toward dishonesty. They confirm that the proportion of dishonest participants in laboratory die-​rolling experiments significantly correlates with the proportion of bus passengers without a valid ticket. More recently, Cingl and Korbel (2020) conducted a lab-​in-​the-​field experiment with 303 youths incarcerated in juvenile detention centers. They showed that cheating behavior measured in the laboratory using the coin-​flip task significantly correlates with long-​term field measures such as the official reasons for detention and short-​term outcomes such as misbehavior reported by the detention center staff. These results reflect very positively on the usefulness of the low-​cost laboratory-​based simple tests so far as measuring the moral firmness of the participants is concerned. Further, Armand et al. (2021) use a novel corruption game framework where citizens interact with actual political leaders to show that citizens correctly predict their leaders’ corrupt behavior.

106  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra Citizens providing larger bribes are more likely to be rewarded by leaders who appropriate more public resources. Is it possible that external validity is a function of the institution implementing the intervention (Bold et al., 2013)? The statistically indistinguishable observations across the field and lab (Armantier and Boly, 2012) may result from the implementing agency (here experimenter). In fact, Armantier and Boly (2012) suggest that the external validity of their study should not be considered a definitive answer. The reason partly is the adjustment of several aspects (disclosure of participation, explicit instruction on consequences and knowledge that consequences are purely monetary) of the experimental design to reflect the inherent difference between field and laboratory. Future studies can account for these and other adjustments to confirm the robustness of the results. Besides, the propensity to engage in corruption depends on the image of the agency providing the ground for corruption. Laboratory experiments administered by the context-​free experimenter may not truly represent the image of an institution. Experimental studies allowing for treatment variation of reputation at the institutional level may reflect more on the external validity of laboratory experiments. Other than institutional concerns, lab experiments must account for external validity threats arising from interpersonal connections with long-​term consequences. Studies focusing on the role of social ties before investigating corrupt behavior may address such concerns. One such study by Di Zheng et al. (2020) shows little effect of such social relations on the decision to bribe. However, the results were obtained with no repeated interaction or penalty, and participants did not reject bribes important to real-​life situations. Future research exploring corruption and social ties taking other factors into account may help to account for interpersonal connection in the true sense. Cross-​Cultural Evidence Often economic models are constructed free of cultural contexts. Consequently, the issue of external validity comes to the forefront in experiments testing a certain theory that, in reality, may be influenced by cultural contexts. Therefore, experiments run in one particular country need not be generalizable in the way a context-​free theory of economic preferences might predict. So, do beliefs and behavior related to corruption change with context? In one of the early investigations of this issue, Cameron et al. (2009) conducted a corruption experiment in four countries: Australia (Melbourne), India (Delhi), Indonesia (Jakarta) and Singapore. Based on a three-​person sequential-​move game involving a firm, a government official and a citizen, the study finds more variation in the propensities to punish corrupt behavior than in the propensities for engaging in corrupt behavior across cultures. In a similar vein, in a within-​country analysis, Buonanno et al. (2020) find that subjects from municipalities with historically low amounts of civic capital are significantly more likely to engage in corruption. Barr and Serra (2010) looked at a student

Stopping the Rot I  107 sample from thirty-​ four countries at Oxford University. They found that corrupt behavior in the lab at the individual level was in line with the levels of corruption prevailing in subjects’ home countries. In contrast to the above results, Armantier and Boly (2012) (discussed in section “External Validity and Generalizability”) in their study found that raw differences in behavior exist when it comes to corruption behavior in Burkina Faso and Canada. Still, after controlling for observable differences (e.g., gender, age, ability), the magnitude of several treatment effects becomes statistically indistinguishable across the two countries. Pascual-​Ezama et al. (2015) further validate these observations. In an incentivized cheating experiment across sixteen countries, they found no statistically significant differences across sample countries regarding their honesty levels. A recent exciting study (Cohn et al., 2019) examines the tradeoff between honesty and self-​interest. In 355 cities in forty countries, over 17,000 lost wallets with differing amounts of money were turned in at public and private institutions. The main result suggests that citizens were more likely to return wallets that contained more money in all the countries. These papers suggest that there are broad similarities in dishonest behavior at the individual level, independent of the subjects’ cultural origin. However, in some cases, there seems to be an influence of cultural norms in affecting subjects’ attitudes toward corruption. It is important to underscore the point that context-​specific responses in experimental results are driven by a complex combination of individuals’ perceptions and institutional settings. Hence, looking for a homogeneous context-​free economic behavior, especially in the realms of corruption, might not be appropriate. Variation in Subject Characteristics and Testing of Policy Regimes The other two major criticisms of laboratory experiments are (1) the use of undergraduates as decision-​makers in the experiments and (2) limited replication of real-​world settings. In the context of corruption experiments, this is certainly a valid concern. However, experimental economists have tried to address these issues of generalizability. First, recent studies highlight that treatment effects do not vary across student and nonstudent samples (like prisoners, bus passengers, etc.), making a stronger case for the student sample’s representativeness (Dai et al., 2018; Fosgaard, 2020; Khadjavi, 2015). Second, the lab facilitates convenient access to comparative statistical analysis, allowing for testing of different real-​world policy regimes before an actual expensive roll-​out in the field. It enables researchers to contrast competing hypotheses. For example, in countries like the United States, the United Kingdom, France, Germany and India, the bribe-​giver and recipient are equally culpable and face penalties (symmetric liability). However, prescribed legal punishment for the bribe-​giver is comparatively mild in China, Japan and Russia. Abbink et al. (2014) reproduce these two real-​world scenarios of legal environments in the laboratory to study the efficacy of an asymmetric punishment policy in reducing corrupt practices. They find support for asymmetric liability as a better anticorruption

108  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra mechanism to fight harassment bribery. Similarly, Abbink and Wu (2017) replicate two regimes where both the client and the official may self-​report (symmetric reward mechanism), and only one party may self-​report in combating collusive bribery. However, in the future, studies comparing corrupt behavior across the lab and nationally representative samples can help strengthen the claim of generalizability of the results, specifically the quantitative ones, from corruption experiments. Other Concerns Framing Effects, Moral Cost and Experimenter Demand Effect Economic theory has traditionally focused on context-​free descriptions of choice situations. However, a large body of work due to Tversky and Kahneman (1981) has emerged, establishing that context as well as how the choice problems are framed influence decision-​makers. Although moral costs were salient in the classical economics discussions (Smith, 1759), modeling moral costs in the standard neoclassical economics is very new. Psychological games, in particular, have provided a possible framework for analyzing moral costs (Charness and Dufwenberg, 2006). This can certainly be of concern, especially in the context of corruption experiments with implications for moral standings and other forms of psychological utility payoffs. In this section, we address the following interrelated questions: First, to be representative of corruption in the natural environment, does one need to explain a situation to the participating subjects using language indicative of corruption, or is a frameless description equally useful? Second, do framing effects, particularly in corruption, have implications for moral standings influencing behavior? And third, are laboratory subjects more inclined to make choices to maintain a positive self-​image? Abbink and Serra (2012) concluded the results of framing effects to be mixed in corruption. Banerjee (2016) revisits both the concern of framing and moral cost, comparing the results from a harassment bribery game with a strategically identical but neutrally framed ultimatum game.The results show engagement in corruption, and the average bribe amount is lesser in the bribery frame. However, an additional treatment showed that the treatment effect was attributed to a change in the sense of entitlement rather than framing. These results, driven by this subtle change in entitlement in the bribery game as opposed to the ultimatum game, show the existence of moral costs. Further, Agranov, Dasgupta and Schotter (2018) find that introducing psychological costs typically reduces lying behavior in a communication game between a buyer and a seller. However, as soon as sellers are put in an environment where they need to compete with the buyers, even a high amount of moral costs does not prevent the sellers from frequently lying to cheat the buyers. Gneezy et al. (2019) examined the effect

Stopping the Rot I  109 of moral costs associated with unfairly altering judgments in a bribery experiment involving two participants and a referee. The study shows that when referees can avoid the moral costs involved with altering judgment, self-​interest influences decisions to a greater extent. The detailed review by Alekseev et al. (2017) suggests that in most cases across different domains, using framed language in a given context produces no change in behavior and sometimes proves to be more helpful in understanding the experimental environment. A key difference between laboratory and field experiments is that subjects know that their behavior will be monitored and analyzed by an experimenter upon entering the laboratory. Gneezy et al. (2018) find that more participants tend to lie partially whenever the experimenter cannot observe their outcomes. Abeler et al. (2019) state that the primary motivators for truth-​telling are the preference for being honest and being seen as honest. In the context of corruption, Armantier and Boly (2012) suggest that being scrutinized did not influence the subjects’ corrupt behavior. Subjects in the field did not know that they were participating in their experiment, unlike the laboratory setup. However, the results from the field and laboratory were statistically indistinguishable. Although this may indicate less concern for corrupt behavior, more research on experimenter demand explicitly concerning corruption will help us understand this better.

Experimental Corruption Games: Typology To better identify the scope of our review, we lay down the principal forms of corruption studies using experimental corruption games. The three most observed forms of corruption found in the literature are (1) harassment bribery; (2) collusive bribery and (3) embezzlement. Harassment bribery is the extortion of bribes for services that citizens are already entitled to (Basu, 2011). Although denial of services is not possible legally, an official can prevent the delivery of service beyond a point when the citizens no longer require it. Such coerced transactions hinder social welfare by acting as a regressive tax and disrupting access to public services (Abbink et al., 2014). On the other hand, collusive bribes are paid to access services that citizens are not legally entitled to. Harassment bribery talks about instances where officials extort bribes to do what “they are supposed to do.” In contrast, collusive bribery is characterized by the cases where officials receive a bribe to do “what they are not supposed to do” (Abbink and Serra, 2012). For instance, an act of bribery may be categorized as harassment bribery when a citizen passes a driving test and is eligible for a driving license; however, the same bribery, when a citizen does not pass the test and still obtains a license, may be categorized as collusive bribery. Collusive bribery, unlike harassment bribery, also allows the bribers to benefit from corruption, and neither of the agents has any incentive to report (Abbink and Wu, 2017). Corrupt acts by bureaucrats are not necessarily restricted to demanding bribes in exchange for providing services, which may

110  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra often include appropriation of public funds or embezzlement for private gains (Rose-​Ackerman, 2007). This third category of corruption has been frequently studied through experimental corruption games. In addition to the above, one can also consider forms of corruption that do not necessarily have a pecuniary angle like bribe receipt, but instead manifest through acts of favoritism. Favoritism, nepotism or cronyism are used interchangeably to depict situations where decisions benefit individuals from social relationships with decision-​makers. Several studies have looked at nepotism as a form of discrimination in hiring decisions, and it is included in the most common measure of corruption –​the Corruption Perception Index (CPI). However, these studies have not addressed corruption in public service directly in the same way as experimental research on bribery (Sheheryar Banuri and Eckel, 2012). A recent unique study by Di Zheng et al. (2020) explores the effect of social ties on corruption. It shows that even when there is no scope for bribery, corruption can nevertheless occur as a result of social ties between individuals. However, when bribery is possible, corruption based on social ties is replaced by traditional bribe-​based corruption.

Conclusion Newer theories have been proposed for understanding the persistence or emergence of corruption across countries. For example, it has been established by experimental data that reciprocity between two transacting parties can often serve as the means of the continuance of corruption; consequently, models such as the gift exchange are being looked at frequently in explaining observed behavior, which in turn provides clues to the continuance of corruption. Also, there has been a clear move toward understanding and reinforcing whether results from laboratory experiments are generalizable enough in the context of corruption. This is certainly of critical importance since, at the end of the day, experiments are supposed to serve as a testing bed for the efficacy of different policy propositions, and the external validity of the results is important. This line of investigation has moved in many dimensions, such as trying to extrapolate experimental lab results into field experiment settings, testing whether results from a set of experiments are readily transferrable across countries and societies and verifying whether a set of results are common enough when subject groups are changed. The results are largely reassuring, and although the importance of contexts seems to be important in evaluating and designing experiments on corruption, the literature has broadly confirmed that laboratory experiments can serve as an important complementary tool to other methods of evaluating the efficacy of anticorruption policies. Lab experiments have continued to be used as a low-​ cost method for devising as well as testing theoretical predictions and policy prescriptions in

Stopping the Rot I  111 order to understand how the incentives/​disincentives align themselves in mitigating corruption. This line of research has been particularly popular as rolling out large-​scale field interventions and policies can be costly, and results from laboratory experiments can often be the reassuring first step. Interestingly, policy proposals on fighting asymmetric bribery, which were verified in experimental findings, seem to have been adopted by the Delhi state government as evinced by a hoarding in the city of Delhi, India. There has also been a steady stream of papers trying to understand the importance of framing effects and experimenter demand effects. This is a rich area to pursue research, especially in the context of corruption experiments where there can be clear moral transgressions involved, and hence the agent has to incur not just monetary costs or benefits but also deals with moral costs as well. Consequently, a context-​free environment need not be ideal, especially keeping in mind external validity implications and the well-​established results that decision-​makers are mindful of contexts. The importance of moral costs and making it salient to reduce corrupt behavior seems to have been a theme that is being investigated more, especially under the newer framework of psychological game theory. Finally, while some of the classic experiments focused more on the institutional environments involving citizens or businesses interacting with officials who are in a position to grant or deny favors, recent work has branched out into other drivers of corruption. For example, whether behavior can be classified as cheating or lying, or dishonest behavior more generally, and experimenters have introduced a rich set of games to focus on each of these types of corruption. This direction is particularly helpful in looking into additional areas with the help of the newer experimental games and frameworks. The literature has also delved into social norms, institutions, beliefs, contagion effects, in particular, as well as the importance of self-​selection into corrupt environments. Results from these experiments importantly point out that a narrow monetary cost–​benefit calculation is restricted in predicting corrupt behavior. But certain types of corruption might be more influenced by the existing social norms, and so punishments and/​or rewards might not be in itself sufficient to minimize corrupt practices. Instead, policies need to focus on changing the beliefs of citizens, try and make moral frames more salient and look for ways to nudge social norms in the desirable direction. In part two of the review, we explore the current body of literature pertaining to these issues and offer some recommendations for future research.

Acknowledgments We thank Danila Serra, Lata Gangadharan, Eugen Dimant, Robert Gillanders and an anonymous referee for their comments and suggestions. We also thank the editors of this volume, Chandan Jha and Sudipta Sarangi, for being patient with us, particularly with our liberal interpretation of deadlines.

112  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra

References Abbink, K. & Hennig-​Schmidt, H. (2006). Neutral versus loaded instructions in a bribery experiment. Experimental Economics, 9(2), 103–​121. Abbink, K. & Serra, D. (2012). Anticorruption policies: Lessons from the lab. In New advances in experimental research on corruption (pp. 77–​115). Emerald Publishing. Abbink, K. & Wu, K. (2017). Reward self-​reporting to deter corruption: An experiment on mitigating collusive bribery. Journal of Economic Behavior & Organization, 133, 256–​272. 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. Abbink, K., Irlenbusch, B. & Renner, E. (2002). An experimental bribery game. Journal of Law, Economics, and Organization, 18(2), 428–​454. Agranov, M., Dasgupta, U. & Schotter, A. (2018). Trust Me: Communication and Competition in Psychological Games. Working Paper. https://agranov.caltech.edu/ documents/13169/manuscript_jan20.pdf. Agranov, M., Dasgupta, U. & Schotter,A. (2020). Trust Me: Communication and Competition in Psychological Games. Mimeo, CALTECH. Alekseev, A., Charness, G. & Gneezy, U. (2017). Experimental methods: When and why contextual instructions are important. Journal of Economic Behavior & Organization, 134, 48–​59. Alm, J., Bloomquist, K. M. & McKee, M. (2015). On the external validity of laboratory tax compliance experiments. Economic Inquiry, 53(2), 1170–​1186. Anderson, L. R. & Stafford, S. L. (2003). Punishment in a regulatory setting: Experimental evidence from the VCM. Journal of Regulatory Economics, 24(1), 91–​110. Armand, A., Coutts, A., Vicente, P. C. & Vilela, I. (2021). Measuring corruption in the field using behavioral games. CEPR Discussion Paper No. DP16596. Armantier, O. & Boly, A. (2012). On the external validity of laboratory experiments on corruption. In New advances in experimental research on corruption (pp. 117–​144). Emerald Publishing, UK. Banerjee, R. & Mitra, A. (2018). On monetary and non-​monetary interventions to combat corruption. Journal of Economic Behavior & Organization, 149, 332–​355. Banerjee, R. (2016). On the interpretation of bribery in a laboratory corruption game: Moral frames and social norms. Experimental Economics, 19(1), 240–​267. Banuri, S. & Eckel, C. (2012). Experiments in culture and corruption: A review. Emerald Publishing. Barr, A. & Serra, D. (2010). Corruption and culture: An experimental analysis. Journal of Public Economics, 94(11–​12), 862–​869. Basu, K. (1993). Lectures in industrial organization theory. Blackwell. Basu, K. (2011). Why, for a class of bribes, the act of giving a bribe should be treated as legal. Ministry of Finance Government of India Working Paper No. 1/​2011-​DEA (March 2011) https://mpra.ub.uni-muenchen.de/50335/ Basu, K. (2015).The republic of beliefs: a new approach to? law and economics? Policy Research Working Paper No. 7259, World Bank, Washington, DC. https://papers. ssrn.com/sol3/papers.cfm?abstract_id=2606453 Basu, K. (2020). Republic of Beliefs: A new approach to law and economics. S.l.: Princeton University Press. Becker, G. S. (1968). Crime and punishment: An economic approach. In The economic dimensions of crime (pp. 13–​68). Springer.

Stopping the Rot I  113 Bold, T., Kimenyi, M., Mwabu, G., Ng’ang’a, A. & Sandefur, J. (2013). Scaling up what works: Experimental evidence on external validity in Kenyan education. Center for Global Development Working Paper no. 321. Available at SSRN: https://ssrn.com/ abstract=2241240 Buonanno, P., Puca, M. & Sbriglia, P. (2020). Corruption and within-​country stereotypes. SSRN 3617590. https://​pap​ers.ssrn.com/​sol3/​pap​ers.cfm?abst​ract​_​id=​3617​590 Camerer, C. (2011).The promise and success of lab-​field generalizability in experimental economics: A critical reply to Levitt and List. SSRN 1977749. https://papers.ssrn. com/sol3/papers.cfm?abstract_id=1977749 Cameron, L., Chaudhuri, A., Erkal, N. & Gangadharan, L. (2009). Propensities to engage in and punish corrupt behavior: Experimental evidence from Australia, India, Indonesia and Singapore. Journal of Public Economics, 93(7–​8), 843–​851. Charness, G. & Dufwenberg, M. (2006). Promises and partnership. Econometrica, 74(6), 1579–​1601. Cingl,L.& Korbel,V.(2020).External validity of a laboratory measure of cheating:Evidence from Czech juvenile detention centers. Economics Letters, 191, 109094. Cohn, A. & Maréchal, M. A. (2018). Laboratory measure of cheating predicts school misconduct. Economic Journal, 128(615), 2743–​2754. Cohn, A., Maréchal, M. A.,Tannenbaum, D. & Zünd, C. L. (2019). Civic honesty around the globe. Science, 365(6448), 70–​73. Czap, H. J. & Czap, N.V. (2019). “I gave you more”: Discretionary power in a corruption experiment. Journal of Interdisciplinary Economics, 32(2), 200–​217. Dai, Z., Galeotti, F. & Villeval, M. C. (2018). Cheating in the lab predicts fraud in the field: An experiment in public transportation. Management Science, 64(3), 1081–​1100. Dasgupta, U. & Radoniqi, F. (2021). Republic of beliefs: An experimental investigation. IZA Discussion Paper, No. 14130. https://​pap​ers.ssrn.com/​sol3/​pap​ers.cfm?abst​ract​_​id=​ 3794​076 Di Zheng, J., Schram, A. & Doğan, G. (2020). Friend or foe? Social ties in bribery and corruption. Experimental Economics, 24(3), 854–​882. Eide, E. (2000). Economics of criminal behavior. Encyclopedia of Law and Economics, 5(2), 345–​389. Falk, A. & Fischbacher, U. (2006). A theory of reciprocity. Games and Economic Behavior, 54(2), 293–​315. Falk, A. & Heckman, J. J. (2009). Lab experiments are a major source of knowledge in the social sciences. Science, 326(5952), 535–​538. Fosgaard, T. R. (2020). Students cheat more: Comparing the dishonesty of a student sample and a representative sample in the laboratory. Scandinavian Journal of Economics, 122(1), 257–​279. Friesen, L. (2012). Certainty of punishment versus severity of punishment: An experimental investigation. Southern Economic Journal, 79(2), 399–​421. Gneezy, U., Kajackaite, A. & Sobel, J. (2018). Lying aversion and the size of the lie. American Economic Review, 108(2), 419–​453. Gneezy, U., Saccardo, S. & Van Veldhuizen, R. (2019). Bribery: Behavioral drivers of distorted decisions. Journal of the European Economic Association, 17(3), 917–​946. Graycar, A. & Jancsics, D. (2017). Gift giving and corruption. International Journal of Public Administration, 40(12), 1013–​1023. Kessler, J. B. &Vesterlund, L. (2015). The external validity of laboratory experiments: Qualitative rather than quantitative effects. Handbook of Experimental Economic Methodology http://dx.doi.org/10.1093/acprof:oso/ 9780195328325.003.0020

114  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra Khadjavi, M. (2015). On the interaction of deterrence and emotions. Journal of Law, Economics, & Organization, 31(2), 287–​319. Krueger, A. O. (1974). The political economy of the rent-​seeking society. American Economic Review, 64(3), 291–​303. Levitt, S. D. & List, J. A. (2007). About the real world? Journal of Economic Perspectives, 21(2), 153–​174. Malmendier, U. & Schmidt, K. M. (2017). You owe me. American Economic Review, 107(2), 493–​526. Pascual-​Ezama, D., Fosgaard, T. R., Cardenas, J. C., Kujal, P., Veszteg, R., de Liaño, B. G.-​G., Gunia, B., Weichselbaumer, D., Hilken, K. & Antinyan, A. (2015). Context-​ dependent cheating: Experimental evidence from 16 countries. Journal of Economic Behavior & Organization, 116, 379–​386. Rose-​Ackerman, S. (1975). The economics of corruption. Journal of Public Economics, 4(2), 187–​203. Rose-​Ackerman, S. (2007). International handbook on the economics of corruption. Edward Elgar Publishing. Schulze, G. G. & Frank, B. (2003). Deterrence versus intrinsic motivation: Experimental evidence on the determinants of corruptibility. Economics of Governance, 4(2), 143–​160. Smith, A. (1759). The theory of moral sentiments. Penguin. Svensson, J. (2005). Eight questions about corruption. Journal of Economic Perspectives, 19(3), 19–​42. Tversky, A. & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453–​458.

9 Stopping the Rot II Consequences, Causes and Policy Lessons from the Recent Experiments on Corruption Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra

Introduction No country is immune to the major social and economic phenomenon of corruption. It is characterized by the lack of transparency, accountability and different sorts of institutional weaknesses. In addition to impeding economic growth and imposing immediate economic costs, this incurs a high social cost (Ades and Tella, 1996; Enste and Heldman, 2018). Continued corruption has implications in terms of weakening the fabric of the society by decreasing social trust, reducing the voluntary provision of public goods and encouraging tax avoidance. Abbink and Serra’s review in 2012 was an important piece in this area of research. They dwelled on the lessons learned from the anticorruption experimental literature for policy implications till then. Their review focused on important design issues as well as lessons learned about possible pathways for extrinsic as well as intrinsic motivations that can help in mitigating corruption. Since then, this literature has become bigger and provided different directions of pursuit, both in experimental work and in theoretical proposals that have been tested in the laboratory. In this review, we attempt to inform the reader of some of these strands of explorations, especially those that have appeared after Abbink and Serra (2012), starting with the consequences and causes of corruption.

Consequences of Corruption We start with a discussion of the consequences of corruption in light of the behavioral channels. In particular, we focus on the effects of corruption on trust, contribution to the public good, tax compliance and inequality. Impact on Social Trust and Growth How does corruption affect social trust? Social trust can be viewed as a set of beliefs about the behavior of other individuals in a society (Denzau and North, 2000). It is not surprising that the scope of corruption will be generally DOI: 10.4324/9781003142300-9

116  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra muted in societies with such a higher aggregate level of mutual beliefs. A cross-​ country experimental study supports this hypothesis by showing that trust enables bribery in the low-​corruption countries, but finds no evidence that trust facilitates bribery in the high-​corruption countries (Jiang et al., 2015). Bjørnskov (2021) uses the data collected by Cohn et al. (2019) on wallet return rates across thirty-​eight countries to show that the likelihood of returning the wallet and social trust are strongly correlated. Rothstein and Eek (2009) experimentally explore this relationship the other way around –​that is, does corruption affect social trust? Scenario experiments in low-​ trust, high-​ corruption Romania and in high-​ trust, low-​ corruption Sweden show that when people come across corrupt healthcare workers and local public officials, they lose trust in these institutions, and people in society in general.1 The proposed causal mechanism is that people use some shortcut or heuristic to determine how much to trust by inferring from the behavior of public officials. Similarly, Richey (2010) used US election data to find that residents from states with higher corruption levels demonstrate lower generalized trust. Moreover, in a literature review on corruption, social trust and growth, Serritzlew et al. (2014) show that the absence of corruption and high levels of social trust foster economic growth. It also indicates that corruption has a causal effect on social trust, while the opposite is uncertain.To get a clear causal link, Banerjee (2016) randomized subjects into either a harassment bribery game or a strategically identical but differently framed ultimatum game, followed by a trust game. People in the bribery game treatment trust less than those in the ultimatum game treatment. Demand for bribes violates social appropriateness norms; and norm violation, in turn, negatively affects belief about the matched partner’s trustworthiness reflected in the trust game. Besides, experimental evidence shows that increased social trust boosts e-​commerce shopping (Mutz, 2005) and generates greater support for free trade (Nguyen and Bernauer, 2019), two very relevant areas in the present context of economic growth. In this regard, an important open question remains. An extensive literature identifies the positive effect of social trust on economic performance, such as economic growth and per capita GDP (Knack and Keefer, 1997). While the literature mainly focuses on the high cost of formal contracts and lower investments in a low-​trust environment (e.g., Zak and Knack, 2001), a second-​ order effect of social trust on economic performance may play out through higher levels of corruption. The literature is yet to examine this channel. Impact on Public Good Provision and Tax Compliance Contribution to public goods is indicative of cooperation among members of society. Not surprisingly, then, in an artifactual field experiment from rural Liberia, Beekman et al. (2014) show that corruption undermines incentives for voluntary contributions to local public goods. In a similar vein, Cagala et al. (2017), in a public good laboratory experiment, show that when subjects are

Stopping the Rot II  117 matched to an expropriating administrator, the contributions to public goods are significantly lower. Moreover, contributors are more likely to free-​r ide and less inclined to behave prosocially in the presence of corruption. Similarly, Buffat and Senn (2017) show that, overall, contributions to the public good are reduced by 30 percent when participants can bribe the authority. Tax compliance has similarities with the provision of public goods, even though compliance is a mandatory act and not really voluntary. The similarity comes from the fact that people may readily cheat on their taxes because they believe that the corrupt state will inefficiently use their tax contribution.2 Individual dishonesty is substantially influenced by the incidence of rule violations in society (such as tax evasion, political fraud and corruption) (Gächter and Schulz, 2016). Like Becker (1968), the seminal works of Allingham and Sandmo (1972) and Srinivasan (1973) provide economics of crime approach to tax evasion by weighing the costs and benefits of compliance with the expected utility of tax evasion. The experimental and survey results of Cummings et al. (2009) support the hypothesis that tax compliance increases with individual perceptions of good governance. The linkage between tax evasion and corruption is causally identified by Banerjee et al. (2020) in a public good laboratory experiment.They find that the likelihood of embezzlement by public officials increases tax evasion by citizens, and further, tax evasion causally leads to more embezzlement. Not only that, policies that decrease embezzlement increase tax compliance; however, those that directly impose a penalty on tax evasion do not affect embezzlement. Of course, there are heterogeneities in the extent of such spillover effects. Zhang et al. (2016) show that British participants are more likely to underdeclare their incomes in a public goods experiment than Italians, who have a lower perceived level of corruption. Impact on Economic Inequality So far, evidence suggests that corruption is largely regressive in nature. Low-​ income people suffer disproportionately in corrupt countries, especially when it comes to receiving public services. Corruption can then aggravate inequality in the following ways. First, corruption allows the rich to substitute for public services that are not readily available to the poor, for example, bribes divert public goods to the people who can pay most (Peiffer and Rose, 2018). Second, the poorer segment of the society is typically not in a position to refuse bribe demands and is more likely to pay bribes than the wealthy (Justesen and Bjørnskov, 2014). Third, since the economically powerless segments also lack political power, they are more likely to be at the receiving end of the wrath of corrupt public officials who are interested in exacting any retribution while letting more affluent citizens off with just warnings.This is because the corrupt officers are mindful of the fact that the wealthy might also have substantial political connections and power. For example, in a field experiment, Fried et al. (2010) used four different cars: two newer luxury cars and two older inexpensive cars. Inexpensive cars were pulled over and experienced bribe requests

118  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra much more frequently than luxury ones. In a subsequent qualitative analysis by interviewing, the officers revealed they were afraid of personal connections that luxury car drivers might have with the judiciary and superior officials. Moreover, from experience, officials find that higher-​status people have a better understanding of the law, and it is more time-​consuming to interact with them than poorer individuals. Thus, poorer citizens are easy targets for efficient and less risky petty corruption. However, in their empirical study, Peiffer and Rose (2018) propose an alternate explanation for the vulnerability of poor citizens to bribery in terms of institutional differences.They argue that bribery is the result of a two-​step interaction process with public services. First, the citizens must come in contact with public officials for the provision of the service.The second step, paying a bribe, is contingent on the first step. Further, being poor increases the likelihood of public services demand and thus contacts public services more frequently than wealthier individuals. This suggests it is the exposure to public servants instead of the “easy target” hypothesis that the poor are more likely to pay disproportionately higher bribes. Their data indicate that when exposure differences were taken into account, poverty had no independent effect on bribing. A more direct investigation of the impact of corruption on inequality is exercised by Markussen et al. (2020). A large-​scale lab-​in-​the-​field public goods experiment with over 1,300 participants across rural Vietnam shows that inequality adversely affects aggregate contributions. This result is attributed to high-​endowment individuals who contribute a significantly smaller share than individuals with lower endowments. Further, this effect of inequality is amplified in high-​corruption settings. Overall, though, the literature unequivocally suggests that corruption is not just a nuisance but a systemic feature of the economy and significantly affects economic development by lowering welfare in multiple dimensions. Thus, appropriate anticorruption policies can not only improve governance but also enhance economic growth by reducing these adverse effects of corruption. To design appropriate policies, however, one needs to delve deeper into the causes of corruption, a topic of discussion for our next section.

Causes of Corruption The public choice theory assigns the causality of corruption to a rational individual who maximizes utility, calculating the expected costs and benefits of engaging in corruption. While this theory is useful if viewed in isolation from other determining factors, can it be looked upon as the primary trigger for corrupt behavior? After all, if we are supposed to readily engage in corruption whenever it appears to be a good deal, why do we observe that only some people engage in such practices while others refrain from it even when there are obvious gains to be made? It appears that the public choice theory is insensitive to the larger social context when assigning motivators for corrupt behavior (De Graaf, 2007). The phenomenon of corruption is highly complex. Given the various forms of corruption, there is a clear need to understand its causes

Stopping the Rot II  119 better. We review some of the experimentally validated theories of the causes of corruption next. Importance of Culture Norms, Social Norms and Institutions The social and cultural norms –​or shared beliefs about what should or should not be done –​shape the values of individuals, which in turn may affect their decision. Can these factors be causal mechanisms in explaining why people make corrupt choices? Barr and Serra (2010) predict the individuals who will engage in corruption in a bribery game based on the level of corruption in their home country. While the predictions were successful for Oxford University’s undergraduate students, they did not hold for graduate students in the university, who spent a more considerable amount of time in the United Kingdom. Further, Zhang (2015) employs a within-​country corruption experiment between northern and southern Italian university students. Southern Italians, although having a reputation for being more corrupt, behaves less corruptly than their counterparts.The author associates social stigma as one of the reasons for these surprisingly opposite observations in their experiment. The southern Italians are willing to bear small monetary losses with the expectation that they will be able to upgrade their social status. Further, Zhang (2018) experimentally tests the propensity to report corruption between southern and northern Italians under two different institutional contexts –​“strict enforcement” and “lax enforcement” regimes. In the lax enforcement regime, there is no significant difference between the northern and southern Italians. However, southern Italians are more likely to report wrongdoings, suggesting that being culturally exposed to corruption may instigate greater accountability in the presence of stringent enforcement institutions. This theme is born out in Salmon and Serra (2017), where individuals having ancestors from countries with lower levels of corruption are observed to engage in less corruption when there is a possibility of being socially observed (detailed discussion in “Social Observability” section). Recently, Harri et al. (2020) replicated the experimental methods of Abbink et al. (2002) and Abbink and Hennig-​Schmidt (2006) in Albania, one of the highly corrupt countries in Europe. When the results are compared with Abbink et al. (2002) and Abbink and Hennig-​Schmidt (2006), the overall magnitude of corruption does not differ significantly because of differences in culture between Germany and Albania. However, the likelihood of corruption turns out to be higher among participants in the Albania experiment. Social norms also influence decisions involving nepotism. Banuri and Eckel (2012) discuss the effect of culture on nepotism through a modified version of the trust game conducted in the United States and Pakistan. They argue that in countries with a weak rule of law, engaging in nepotism mitigates the risk of betrayal when the outcome of the transactions is contingent on trust.The results show that willingness to incur efficiency costs to get paired with an in-​group member is significantly higher among subjects from Pakistan. Additionally, the effects of social and cultural

120  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra norms can also be observed in peer reporting. Recently, Romaniuc et al. (2021) compared peer reporting in Moldova and France. They claim that Moldovans’ willingness to collaborate with state authorities may be impacted by the Soviet legacy of using private informants. Their results imply that Moldovans engage in less peer reporting and find it less socially acceptable than French participants. Apart from differences in corrupt behavior arising out of cultural and social norms and the background of the nations, Abbink et al. (2018) take a different approach by explicitly creating social norms in the laboratory. In doing so, they disentangle the effect of descriptive social norms from injunctive social norms. Their results establish the causal link between descriptive social norms and bribe offers. On being paired with a likely corrupt partner, the probability of proposing a bribe increased more than twice compared to when paired with a likely honest partner. Altogether these studies indicate that sociocultural and institutional norms do provide some causal mechanisms for corruption. Whether it be offering more bribes, reporting, favoritism, or altering behavior due to social observability, these norms impact individual behavior of being corrupt. Two caveats may be kept in mind in this regard. First, we should, of course, be cautious about interpreting these results and be mindful about not prejudging or stereotyping individuals based on their cultural backgrounds (Salmon and Serra, 2017). In fact, the experimental evidence does not always conform with behaving dishonestly on account of cultural and social norms. Additional research on a broader range of countries with varying norms would provide a more definitive answer in disentangling the possible mechanisms where the experimental results are not aligned with the predictions. Second, we must recognize that cultures and norms are endogenous to institutions, policies, and other economic factors. So it is possible that the heterogeneity in unethical behavior attributed to culture may have its roots in other economic primitives playing out in the long run. To the extent that an individual considers these shared beliefs as given to her decisions, we may consider culture exogenously affecting ethical choices. Effect of Contagion: The Bad Apple Theory When the peer behavior/​information tends to influence the others in the vicinity to act similarly, the observed phenomenon is termed as contagion effect. To what extent may this induce corrupt acts? Gino et al. (2009) provide the first experimental evidence on the contagion effect concerning unethical behavior. Their experiment employs participants to solve simple mathematical problems where participants can cheat by misreporting. The experimental confederates cheat by finishing a task impossibly quickly and leaving the room with the maximum reward.3 Observing this, the unethical behavior among in-​group members increased. Other experiments have observed the contagion effect with more institutionally formal setups like tax compliance, embezzling public funds, and bribery. Blaufus et al. (2017) investigate the impact of public disclosure on tax evasion behavior.Varying the tax privacy from complete privacy

Stopping the Rot II  121 to full disclosure, they find that tax compliance is lower when an individual observes noncompliance of other individuals, supporting the contagion effect. In a framed laboratory, Boly et al. (2019) provide evidence of embezzlement’s contagion effect. The experiment involves two public officials, A and B, who make decisions regarding embezzlement from separate funds. Before making the decision, Official B observes Official A’s decisions. They find that a corrupt Official A increases the probability of embezzlement by Official B. Schram et al. (2022) observe similar results in bribery, implementing a laboratory experiment of corruption based on a real effort task. The treatments differ in the type of information provided to the subjects about others’ choices of bribe amounts. The results are consistent with the contagion effect. Not only peers but even corrupt leaders and policymakers also pose a threat in proliferating unethical behavior. d’Adda et al. (2017) furnish experimental evidence of how the acts of leaders affect the ethical conduct of their followers. When “leaders” can make public statements and have the discretion to allocate financial rewards, they significantly foment the unethical behavior of the followers. Similarly, Banerjee et al. (2020) also find that greater tax evasion among citizens is associated with greater embezzlement among public officials. The above studies suggest individuals’ behavior of dishonesty, tax compliance, embezzlement, and bribery, to a large extent, is affected by moral spillover effects.

Effectiveness of Anticorruption Policies Abbink and Serra (2012) discussed lessons from the laboratory on the efficacies of some of the proposed anticorruption policies.This review extends the Abbink and Serra (2012) work to include new experimentally validated anticorruption policies.We start with the monetary measures that alter an individual’s financial incentives: wages and salaries, deterrence hypothesis, and severity vs. certainty of punishment. Next, we review the experimentally validated nonmonetary anticorruption measures: awareness about negative externalities, moral costs, competition, and social observability. Third, we discuss some additional policies that can be implemented at the institutional level. Specifically, we discuss the effectiveness of staff rotation, monitoring and role of technology, the role of intermediaries, transparency, whistleblowing and reporting regimes (symmetric vs. asymmetric), and gender roles. Using Monetary Incentives Wages, Salaries and Rewards The scope of increasing benefits to deter decision-​makers from corrupt acts is limited in the cost–​benefit approach (Becker, 1968) and focuses on changing the public officials’ wages and salaries. Abbink and Serra (2012), in their review, suggest that the majority of studies examining changes in public official wages

122  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra find that paying fair salaries to officials decreases bribe-​taking.Van Rijckeghem and Weder (2001) show that a 1-​point change in relative wages and salaries of public officials is associated with a 0.5-​point decrease in the corruption index. This result holds for public educators too. Borcan et al. (2014) investigate the outcome of corruptible high-​stake exams shortly after an unexpected 25 percent wage cut in Romania. It was observed that public schools “performed” better than private schools, the latter not being affected by the policy. This difference is attributed to the surge in corruption induced by the reduction in wages of public educators. However, such a connection between the increase in public servants’ wages and a consequent reduction in corrupt practices does not necessarily hold true. In fact, when the government of Ghana doubled its police officers’ salaries partly to mitigate petty corruption, instead of decreasing petty corruption, it significantly increased the police efforts to collect bribes, the value of bribes, and the overall bribe amounts (Foltz and Opoku-​Agyemang, 2015). Similarly, Navot et al. (2016) provide global evidence using WVS data, suggesting a positive association between public servants’ wages and the tolerance of corruption. They argue that the effect of increasing wages and salaries may be crowded out by pecuniary incentives to advance their self-​interest and motivate justifications for bribe acceptances. Further, Barfort et al. (2019), in a survey experiment in Denmark, the world’s least corrupt country, found that increasing public sector wages would attract more dishonest candidates to public service in Denmark. The public official salaries are necessary but insufficient for lowering corruption (Gans-​Morse et al., 2018). Also, solely relying on increasing government wages to reduce corruption can be very costly (An and Kweon, 2017). Thus, influencing the decision to engage in unethical behavior through wages and salaries does not appear to be a universally optimal policy, the context does seem to matter. However, there is additional support for the reward–​ corruption nexus. Rewards need not be in the form of regular salaries but can be based on good performances. A vignette survey of South Korean bureaucrats who believe their promotion is contingent on their performance shows less tolerance toward corruption (Kwon, 2014). Further, in a field experiment, Khan et al. (2016) show that tax revenues are higher when officials’ wages are associated with collection levels. In addition, Abbink et al. (2020) test the effectiveness of such a reward mechanism in their novel laboratory experiment, which is based on the officials’ collective performance in lowering overall corruption in society. This reward mechanism can be implemented without individual monitoring of the officials’ objective measures (like fines collected and demanded bribes).The results suggest that such a reward system not only significantly reduces the incidences of extortion but also there is a moderate reduction in law-​breaking among citizens. Deterrence Hypothesis and Certainty vs. Severity of Punishment As discussed previously, the standard workhorse theory of Becker predicts the deterrence hypothesis. The lab experiment evidence with student participants

Stopping the Rot II  123 broadly supports the deterrence hypothesis using Becker’s framework.4 All such studies suggest that both the magnitude and the probabilities of punishment lower the susceptibility of corrupt indulgences. However, is one more effective than the other? The experimentalists advance the research in deterrence hypothesis by comparing these two policy levers: severity of punishment vs. the probability of punishment. According to Becker, risk-​averse individuals are deterred more by increases in the severity of punishment than an equivalent rise in the punishment’s probability. First, Abbink et al. (2002) show that even an exogenously determined minor threat to disqualification (“sudden death”) is a strong deterrent to corrupt behavior. In recent literature, Banerjee and Mitra (2018) extend this line of research in an experimental setup of corruption: a harassment bribery game between citizens and public officials. They introduce two audit treatment arms with identical expected payoffs. The first treatment was one with a low audit probability of audit and high fine (LP), and the second was one with a higher audit probability and low fine (HP). Their results show that the average bribe and bribe demands are lower in audit treatment significantly, suggesting deterrence works. More importantly, deterrence from LP treatment is significantly more effective in comparison to HP treatment. Also, when it comes to beliefs, citizens expect more honesty from public officials in LP than from HP treatment. Similarly, Laske et al. (2018) vary the probability and fine associated with being audited in twenty different treatments. Across all the treatments, they found that fines were more effective than the monetarily equivalent probability of punishment. Overall, the evidence suggests that fines and penalties work better than the probability of detection as an effective anticorruption policy. However, Galeotti et al. (2021) examine such mechanisms in their quasi-​ experimental analysis using public transport. They identified fraudsters and nonfraudsters based on fare evasion in public transport subjected to inspections and no inspections conditions. The intrinsic honesty of the same person was measured by misappropriating money in the second stage. When the passenger deboards, a person from the research team asks the passenger whether they lost the banknote seemingly picked up from the ground in front of the passenger.5 The acceptance of the banknote provided the experimenters with the measure of the intrinsic honesty of the passenger. Later, they test whether this intrinsic honesty correlates with compliance in buying tickets while using public transport. Their results suggest that the exposure to deterrence practices increased the unethical behavior of not only fraudsters but nonfraudsters, too, especially when inspection teams are larger. This indicates that deterrence may crowd out the moral cost of not engaging in unethical behavior. Additionally, in close collaboration with the previous section (Wages, Salaries and Rewards), further research comparing the incentive mechanisms with penalty schemes focusing on the benefits and drawbacks of the two systems will be useful for policy formulation.

124  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra Introducing Competition Rose-​Ackerman (1978) proposed that introducing competition at the level of bribe receivers will reduce corruption: when a bureaucrat distributes a scarce benefit, the presence of competing bids decreases the equilibrium quantity of corruption. On the one hand, less competition means firms enjoy higher rents so that bureaucrats with control rights over them are motivated to ask for higher incentives to engage in malfeasant behavior. Also, higher rents imply that the public would be more likely to rewrite the bureaucrat’s contract and to have more resources to control bureaucratic engagements. In this regard, less competition indicates less corruption. The effect of competition on corruption remains ambiguous in theory (Ades and Di Tella, 1999). However, Ryvkin and Serra (2020) experimentally investigate the effect of introducing competition among public officials granting licenses to citizens. Citizens can obtain the license by paying a bribe in addition to the license fee, or they can visit another officer.Visiting other officers incurred a fixed cost, replicating the search cost in finding and going to different offices in real life. The results are contrasted with monopolistic treatment conditions having no option for the citizen to choose another official. The results show when citizens can obtain a license from multiple offices, bribe sizes reduce by more than 50 percent in both one-​shot and repeated transactions. However, the effect of competition diminishes when officials can communicate with each other, though such collusions are not sustainable in the long run. Thus, changing the structure and organization of the service delivery system and ensuring greater competition have high chances of succeeding as an instrument against bureaucratic corruption. Using Nonmonetary Measures Moral Costs While engaging in corrupt behavior, the decision-​maker is typically well aware of their transgressions and faces moral costs leading to cognitive dissonance (Festinger, 1957). Can such moral costs be effectively used to dissuade corrupt practices then? Morality viewed as a personal conviction can also be rationalized and incorporated into an expected utility framework. Rose-​Ackerman (1975) was the first to show agents maximize a utility function with a cost term arising from moral costs. However, the conditions to appeal to morality not only rely on formal structure but also on informal structures such as moral competencies, learning, and education policy. One way to influence moral cost is through the salaries of public officials (Abbink et al., 2002). One may argue, for example, that if government positions are paid worse than comparable other jobs, the moral costs of corruption are reduced. In such a situation, highly paid public officials should be less likely to involve in corrupt activities. However, the results show there is no significant difference in behavior for differently paid public officials and as we discussed, such a policy can rather encourage corruption. Besides

Stopping the Rot II  125 monetary interventions through penalty and probability of audit, Banerjee and Mitra (2018) leverage an intervention through ethics education to influence the moral cost. Specifically, they study the effect of an ethics course that is part of a standard MBA program. They hypothesize that the proportion of public officials who demand a bribe and the bribe demand is lower for the subjects who go through ethics education than those who do not. Although ethics education has a small immediate effect on the likelihood but not on the amount of bribe demand measured just after the intervention, the effect ceases to exist a month after the course ends. Besides pecuniary and nonpecuniary measures that affect the moral cost in demanding a bribe, the moral cost can be there even after taking the bribe, especially when individuals change their decision in response to the bribe. In a bribery experiment involving two players and a referee, Gneezy et al. (2019) show that when referees have scope for avoiding the moral costs associated with bending the judgment, self-​interest is the major driver for such decisions. Further, when moral costs increase through treatment interventions, the impact of bribes is lower. However, the quality and magnitude of moral costs can be context-​dependent and vary greatly depending on the type of corrupt behavior in which the decision-​maker is involved. Though moral costs exist even among corrupt decision-​makers, the evidence so far does not find it as an effective policy lever in curbing corrupt behavior. Future studies should aim to design and experimentally test mechanisms that experiment with the salience of moral costs and, in addition, consider institutions that can sustain the salience for a longer period of time for lasting impacts. Awareness about Negative Externalities While corruption typically involves interactions between two parties, there are serious negative externalities concerning third parties, the society members. Do awareness about these negative externalities intrigue kindness and social preferences of parties directly involved in the corrupt acts? Abbink and Serra (2012) find differing results from their review and call for additional research to test the intrinsic motivations generated by individuals’ awareness of the negative externalities due to corruption. Among recent studies, Senci et al. (2019) found that information about negative externalities was an effective bribery deterrent. Bribe offers and acceptances were further discouraged when combined with information about the rights and duties of the agents. Moreover, when evaluating wrongdoing, Guerra and Zhuravleva (2019) find that in the “high negative externality” treatment, both bribe offers and bribe acceptances are significantly lower than “low negative externality” treatment. However, Abbink and Serra (2012) speculate that subjects have complete knowledge about how much harm they were causing to others in the experiments. In real life, it is often unclear about the consequences of corruption to others. Therefore, individuals may be more likely to shun the adverse effects arising from their corrupt engagements. Thus, awareness about negative externality may not be much effective in deterring corruption.

126  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra Social Observability Shared expectations, unwritten rules, and beliefs form the basis of societal virtues. Individuals usually follow such social norms and strive to appear more prosocial when others observe (Pan and Houser, 2017; Salmon and Serra, 2017). Can social observability then provide further impetus to deter dishonest behavior? Increasing social visibility of one’s actions and identity may influence observed behavior by increasing the stigma associated with actions and choices that are not coherent with social norms. Cason et al. (2016) vary social visibility and observability with two different levels of information disclosure in audits to examine such possibilities. First, “Low information” condition, where only anonymous reporting choices are disclosed. Second, the “High information” present digital photos of all participants and provide feedback on their actual output and compliance.6 They find that the visibility of actions has no significant impact on output or reporting in either of the treatments. However, these results are somewhat in contrast with Salmon and Serra (2017), who employ three treatments with varying degrees of social observability of the officials’ actions. In the first treatment, corrupt actions are completely hidden from others. In the second one, actions are visible to the victim, and in the third, others, including the victim, observe their actions.The third treatment leads to a statistically significant decrease in the propensity to offer bribes. However, the effect is only true among individuals who identify culturally with countries characterized by low levels of corruption. Besides the cultural setting, the differing results from this study to previous studies may be driven by an additional aspect of instant feedback from society members. If this is true, this may suggest that immediate consequences of actions with public display of corrupt behavior, that is, social observability accompanied by immediate feedback, accentuates the stigma associated with unethical behavior. Future studies can experimentally test the effectiveness of such an anticorruption mechanism. Butler et al. (2020) build on the findings of Salmon and Serra (2017) in the context of whistleblowing in corporate organizations. The possibility of “social judgment” may act as an incentive or deterrent against blowing the whistle depending on the negative externality. The results suggest “social judgment” serves as a deterrent when the public does not feel directly affected by the negative externalities caused by corporate fraud, and on the other hand, it may act as an incentive when the opposite holds. By and large, instigating social image concerns appears to be circumstantial in its function of deterring corruption. Some Additional Policy Levers Besides the pecuniary and nonpecuniary policies that can be implemented to alter individual incentives in engaging in corruption, we discuss some additional policy levers, with a focus on changes at the institutional level, that may dissuade corruption.

Stopping the Rot II  127 Staff Rotation An experimental design mimics the staff rotation at the institutional level when there is random rematching of the subjects in the role of officials and clients.7 Abbink (2004) finds such an intervention reduces the average corrupt choices by nearly two-​ thirds. Among recent evidence, Bühren (2020) experimentally analyzes staff rotation in China and Germany. Staff rotation reduced the propensity to reciprocate favors on receiving the bribe. Additionally, experimental results suggest that German bribers anticipated such actions from public officials, and both frequency and amount of bribes were lower in staff rotation. Additionally, Fišar et al. (2021) find that staff rotation does not influence the proportion of firms offering bribes but reduces the share of bribe acceptance. Thus, it appears staff rotation successfully deters corruption among bribe-​ takers, but its effectiveness in bribe-​giving is still not promising. For developed countries, staff rotation can indeed lower the average corruption. However, in developing nations, where bribe-​giving acts as a norm (e.g., harassment bribery), this will only deter the bribe-​takers. Role of Monitoring and Technology Abbink and Serra’s (2012) review relates monitoring to the principal’s capacity, which determines the probability of detection. Additionally, they also talk about various types of monitoring, particularly human monitoring. They suggest the experimental results show a strong effect of monitoring in discouraging corrupt behavior. Further, Lagunes (2017) highlights that civil society oversight results in lower costs of projects, resulting in overall efficiency gains. In a field experiment in Peru, treatment districts were assigned to an organized civil society to monitor public officials’ executing public work. The results show no significant difference in the completion rate of the projects between monitored and nonmonitored districts. However, treated districts completed the same work at 51.39 percent lower cost relative to the untreated districts. Gans-​ Morse et al. (2018) provide an interdisciplinary review of anticorruption policies, focusing on reducing corruption among public officials. Collating research from economists, political scientists, sociologists, and anthropologists, they suggest that hitherto the most promising evidence of successful anticorruption policies pertains to monitoring-​based policies. This is strongly supported by Abbink et al. (2020) in their experimental investigation. They show that the mere presence of police, even if they are corrupt, substantially reduces crime compared to a baseline without police presence. Furthermore, the scope of monitoring has been tremendously improved with the advent of technology. Muralidharan et al. (2016) assess the impact of smartcards in a large-​ scale randomized experiment in Andhra Pradesh. The treated beneficiaries of PDS are provided with smartcards linked to the respective bank account that can be used to receive government transfers directly.8 The results suggest that smartcards not only ensured faster delivery and

128  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra more predictability, but it also entailed less corruption in the payment process. The “leakage” of funds in the process was significantly lower. More importantly, all these positive results didn’t affect the accessibility of the program, and more than 90 percent of beneficiaries preferred the new system. The use of technology in monitoring corruption is in coherence with the idea of the “Republic of Beliefs” (Basu, 2020). Without technology, there is a higher propensity for human interaction, leading to an outside equilibrium formed by a system of beliefs that corruption exists at every level. The use of technology helps mitigate the belief that there is a way out after committing a crime. The Asymmetric Regime and Whistleblowing When it comes to harassment bribery, the bribe-​giver does not have much choice of not offering a bribe. Basu (2011) proposes that bribe offering should be legitimate activity in all such cases, that is, harassment bribe-​givers should have complete immunity from any penalties or punishments imposed by the state. This mechanism where bribe-​takers are culpable but bribe-​givers have legal immunity is commonly known as asymmetric liability. Can asymmetric liability increase bribery incidences too? Dreze (2011) argues that immunity to bribe-​givers can make such acts morally acceptable, thereby shifting the social norm associated with petty bribery and, therefore, encouraging acceptance of bribe demand. Also, in whistleblowing, refunding the bribe is cumbersome in developing countries, given lower prosecution rates and slow justice delivery. Even if these concerns are addressed, there may always be future apprehensions of retaliation if the official still manages to remain in the office. In an experimental setup, Abbink et al. (2014) test both Basu’s policy recommendation and Dreze’s concerns.Their experimental design consists of four treatments in a harassment bribery game. First, in “Symmetric” treatment, both citizens and officials are fined when caught. This is in contrast with the second treatment, “Asymmetric,” where the action is taken only against the officials, and bribers have full impunity for whistleblowing with a full refund of the bribe amount. The apprehensions about the bribe refund and retaliation that can deter whistleblowing are tested in the next two treatments. In the third treatment,“Retaliation,” there is a scope for the official to retaliate after escaping conviction and consequently lower citizens’ payoffs. The fourth treatment, “No Refund,” tests the propensity to report incentives to bribery even if there is no monetary incentive to do so. Results suggest that legal immunity in asymmetric liability increases reporting and reduces bribe demand compared to symmetric liability. Further, contrasting the results from retaliation and no refund means the refund of bribes does not necessarily drive reporting. Instead, nonmonetary factors like intrinsic motivation play a significant role. However, the major obstacle to a successful implementation of a leniency program is when officials can retaliate.

Stopping the Rot II  129 In the retaliation treatment, bribe demand and reporting incidences drive down to symmetric liability treatment levels.Thus, it suggests leniency policies should be accompanied by staff rotation, whistle-​blower anonymity, and speedy prosecution rates. Engel et al. (2016) experimentally analyze symmetric versus asymmetric mechanisms in the context of collusive bribery. Their bribery game replicates a scenario of a corrupt exchange between official and citizen, where the officer can reciprocate in favor of the citizen on receiving the bribe. In such arrangements, both players have incentives in not disclosing their corrupt exchange. In contrast to Abbink et al. (2014), their results suggest that the frequency of corruption is higher in asymmetric punishment. However, collusive bribery and harassment bribery is fundamentally different. The common interest in collusive bribery is the exchange of favors, and asymmetric liability may seem unfair to one or both indulging parties. In the case of harassment bribery, the citizen may be more likely to report from a feeling of being wronged under asymmetric liability. As such, a comparison of Engel et al. (2016) and Abbink et al. (2014) may not be meaningful. In addition to the asymmetric regime, Boly and Gillanders (2018) talk about institutional asymmetry in terms of legal equality between public officials. Legal inequality is when public official A chooses the level of detection and punishment that’s only applicable to public official B. Whereas, when public official A also faces the same detection level and punishment, its legal equality. Their experimental investigation finds that even if the public officials are not corrupt, legal equality significantly distorts institutions by facilitating a lower detection level. Thus, institutions with legal equality among officials may further aggravate corruption. The previous section (Role of monitoring and technology) discussed the effectiveness of monitoring in deterring corruption, with a particular emphasis on the top-​down approach. Citizen reporting or whistleblowing, on the other hand, is commonly referred to as bottom-​up interventions. Ryvkin et al.’s (2017) investigation of online reporting of corrupt offices and officials is an example of such bottom-​up interventions. Serra (2012) provides a detailed comparison of the two approaches, additionally conducting a lab experiment that creates an accountability system by combining bottom-​up monitoring with formal top-​down auditing. Her experimental results suggest even in a weak institutional environment where the likelihood of both formal punishment and fine are low; this “combined” approach could be highly effective in reducing corruption. Further, Butler et al. (2020) evaluate two policies that may encourage whistleblowing: first, by using financial incentives, and second, by providing protection from public scrutiny and social judgment. They provide strong evidence of the effectiveness of monetary incentives for whistleblowing, and this effect further increases when there is scope for social judgment. In a recent study, Buckenmaier et al. (2020) show that the opportunity to blow the whistle decreases the collusion and bribe acceptance rate keeping the bribe offers unaffected. In addition, the results suggest a positive spillover effect of

130  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra leniency. The observations persist even after the mechanism has been removed. Overall, the effects of asymmetric policies and whistleblowing seem impressive, especially in dealing with harassment bribery. Role of Intermediaries Intermediaries may facilitate corruption by reducing bribers’ and bribees’ moral or psychological costs (Drugov et al., 2014). The experimental data confirm that intermediaries lower the moral costs of citizens and officials and, thus, increase corruption. On a different note, Attanasi et al. (2019) investigate the moral costs of intermediaries in embezzling donations before transferring them to recipients. The results show that guilt aversion induced by donors’ and recipients’ expectations of transfer amount does reduce embezzlement. Taking this forward, Di Falco et al. (2020), in a similar setting, additionally vary the number of intermediaries in transfer chains. Donors are not only less generous in the presence of intermediaries, but shorter transfer chains also reduce aggregate embezzlement. Therefore, policy designs should try to eliminate intermediaries, and wherever it’s not feasible, aiming to keep the transfer chain at a minimum will put off corruption level to a greater extent. Transparency What if individuals were able to harness the benefits of transparency and utilize them as a means of gaining access to essential public services that would otherwise be available only through corruption? Peisakhin and Pinto (2010), in a field experiment, demonstrate that India’s information law ensuring transparency is almost as effective as bribery in helping the poor secure access to an essential public service. Their results show that greater transparency lowers corruption even in highly hierarchical and unequal societies. García-​Gallego et al. (2020) conducted an experiment in which two firms compete for a public project by submitting offers that include quality levels and a bribe to a public official who decides the winner. Their investigation consists of transparency in the sense of observability by a third party, who may or may not be affected by the transactions. In the treatment group that tests the impact of transparency, there is a passive subject in the role of a citizen who observes the interaction between firms and public officials. They argue the difference observed in this treatment can be attributed to intrinsic motivation like shame, guilt, or just getting observed by a third party –​the “audience effect.” The result shows that audience effects significantly lower the firm’s bribe placement, and officials accept the highest bribes significantly fewer times than baseline with no audience effect. Similarly, Di Falco et al. (2015) show embezzlement from charity funds decreases with transparency in donations. Thus, ensuring transparency can be extremely useful in deterring corrupt practices.

Stopping the Rot II  131 Gender Roles Alatas et al. (2009), in their cross-​ country investigation, highlight gender differences in behavior when confronted with a common bribery problem. The data from Australia (Melbourne), India (Delhi), Indonesia (Jakarta), and Singapore show that the variations in women’s behavior toward corruption are much more significant than in men. Can this difference be utilized to combat corruption? Frank et al.’s (2011) review on gender differences in corruption suggests that with the involvement of women, a potentially corrupt transaction is more likely to fail. Further, in their review, Chaudhuri (2012) concludes that women behave more prosocially and are less corrupt than men, or gender differences are insignificant. None of the evidence suggests that men are less corrupt.The experimental results of Rivas (2013) show that women are indeed less corrupt than men. Moreover, Fišar et al.’s (2016) findings suggest that women believe in the prevalence of corruption more than men do. They additionally show that males are more likely to offer bribes, while females are less likely to conform to a norm of bribe-​giving. Other studies show that women are more willing to punish corrupt actors (Guerra and Zhuravleva, 2019, 2022). On top of that, higher negative externality leads women to offer and accept fewer bribes than men and punish corrupt actors more severely than men. Detkova et al. (2021) show female public procurement officials in Russia consider corruption an obstacle to development, but no such results surface for male bureaucrats. The heterogeneity between male and female bureaucrats is observed even at the high-​level positions. When it comes to gender roles, recognizing the problem seems to be the first step in supporting an anticorruption policy. The latest UNDP global report (UNDP, 2017) that women’s representation in public administration on an average is 45 percent, with a variation of 3–​77 percent across countries. Out of that, only 20 percent of the country has reached gender parity in public administration.The experimental evidence and facts suggest that ensuring gender composition and even over-​representation of women in administrative services can effectively deter corruption.

Conclusion We end with some thoughts on where this literature may go from here. Our primary position is that the next decade of corruption research using experimental tools needs to involve novel data and cutting-​edge empirical methods, all directed toward increasing the precision and generalizability of the findings. We contemplate such a direction may offer novel and interesting insights and, in the process, have a greater impact on policy. The policy footprint of lab experiments on corruption has been limited till now due to the fact that it has, by and large, focused on qualitative or directional results. However, the field of economics is rapidly moving toward the quantification of effect sizes, and the inquiry for quantitative effects demands a larger and, more importantly, representative sample size. While other topics on behavioral aspects have been

132  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra studied using a representative online sample in the US and Europe, this now seems increasingly possible in developing countries, too, with online samples recruited through survey platforms such as Qualtrics. In this context, one line of research that appears promising relates to structurally estimating moral costs in line with the estimation of other preference parameters. Just like it is extremely useful to know the distribution of risk preference or time preference parameter in a country or across the world, it is valuable to understand how moral cost is distributed spatially and temporally. Notice, unethical choices, be it lying, evading taxes, or indulging in corruption, are eventually functions of some moral cost parameter, and any credible welfare analysis must be founded on its estimation. Related, we find that there is now a general consensus that preferences are amenable to changes when suitably intervened. When looked at from the angle of a preference, it becomes amply clear that moral costs too may be altered with appropriate interventions. Such interventions have been few and far between in the literature. We think there is substantial value to be added by examining the effect of sustained, long-​term interventions aimed at changing the moral cost. Admittedly such interventions are costly and are difficult to run, but they may offer valuable insights into the mechanisms that promote ethical choice. Interestingly, the recent advances in network economics can offer deep insights into the micro functions of the contagion effect, a topic that has generated considerable interest among researchers of corruption. The broad stroke of contagion effect was indeed useful in identifying peer effects. However, the next generation of studies may focus on key questions such as how corruption behavior defuses through networks and the role of central nodes in promoting corruption, as well as restraining it. For example, in an environment characterized by limited state capacity, should punitive institutions focus on deterring individuals occupying central nodes in networks? Is the spillover of deterrence effect larger when such central nodes are targeted? This issue of network assumes even more significance in the fractured developing world where decisive nodes are occupied by people of certain privileged social groups. This may impose a differentially higher cost on people from underprivileged groups as they may not have the networks necessary to navigate through the bureaucracy and red tape. More importantly, they may face a higher cost of accessing public goods and services due to the discriminatory attitudes of the people from the privileged community. In such environments, policies aimed at mitigating discrimination may, in fact, double up as a corruption mitigating tool. As our final thoughts, we note that in spite of some claims that laboratory experiments on corruption may have exhausted their scope, our review suggests that there are exciting possibilities in the future if the literature reimagines itself with new tools, data, and ideas.We believe that the footprints of research related to corruption, and more broadly, public choice theory, will only expand in the days to come.

Stopping the Rot II  133

Notes 1 In the scenarios, the individual seeks immediate assistance from the police/​doctor while also attempting to communicate with another individual who resides in the “unknown” country. 2 In this section we only focus on impact of corruption on tax compliance behavior and not on tax compliance and tax evasion per se. See Alm and Malézieux (2021) and Slemrod (2019) for a detailed discussion in tax compliance and tax evasion respectively. 3 The experimenters hired a professional actor as their confederate to indicate that the person was lying or cheating. 4 For discussion see Khadjavi (2015, 2018). 5 They hired professional actor who was part of the research team. 6 In this experimental setup to investigate audit mechanisms, participants make two distinct decisions –​the actual level of output and the reporting level of output. 7 For detailed discussion, see Abbink and Serra (2012). 8 Smartcards are biometrically authenticated payments infrastructure.

References Abbink, K. (2004). Staff rotation as an anticorruption policy: An experimental study. European Journal of Political Economy, 20(4), 887–​906. 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. Abbink, K., Freidin, E., Gangadharan, L. & Moro, R. (2018). The effect of social norms on bribe offers. Journal of Law, Economics, and Organization, 34(3), 457–​474. Abbink, K. & Hennig-​Schmidt, H. (2006). Neutral versus loaded instructions in a bribery experiment. Experimental Economics, 9(2), 103–​121. Abbink, K., Irlenbusch, B. & Renner, E. (2002). An experimental bribery game. Journal of Law, Economics, and Organization, 18(2), 428–​454. Abbink, K., Ryvkin, D. & Serra, D. (2020). Corrupt police. Games and Economic Behavior, 123, 101–​119. Abbink, K. & Serra, D. (2012). Anticorruption policies: Lessons from the lab. In New advances in experimental research on corruption (pp. 77–​115). Emerald Publishing. Ades, A. & Di Tella, R. (1996). The causes and consequences of corruption: A review of recent empirical contributions. IDs Bulletin, 27(2), 6–​11. 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. Allingham, M. G. & Sandmo, A. (1972). Income tax evasion: A theoretical analysis. Taxation: Critical Perspectives on the World Economy, 3, 323–​338. Alm, J. & Malézieux, A. (2021). 40 years of tax evasion games: A meta-​ analysis. Experimental Economics, 24(3), 699–​750. An, W. & Kweon, Y. (2017). Do higher government wages induce less corruption? Cross-​country panel evidence. Journal of Policy Modeling, 39(5), 809–​826. Attanasi, G., Rimbaud, C. & Villeval, M. C. (2019). Embezzlement and guilt aversion. Journal of Economic Behavior & Organization, 167, 409–​429.

134  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra Banerjee, R. (2016). On the interpretation of bribery in a laboratory corruption game: moral frames and social norms. Experimental Economics, 19(1), 240–​267. Banerjee, R., Boly, A. & Gillanders, R. (2020). Anti-​tax evasion, anti-​corruption and public good provision: An experimental analysis of policy spillovers. SSRN. https://​ pap​ers.ssrn.com/​sol3/​pap​ers.cfm?abst​ract​_​id=​3652​411 Banerjee, R. & Mitra, A. (2018). On monetary and non-​monetary interventions to combat corruption. Journal of Economic Behavior & Organization, 149, 332–​355. Banuri, S. & Eckel, C. (2012). Experiments in culture and corruption: A review. Emerald Publishing. Barfort, S., Harmon, N. A., Hjorth, F. & Olsen, A. L. (2019). Sustaining honesty in public service:The role of selection. American Economic Journal: Economic Policy, 11(4), 96–​123. Barr, A. & Serra, D. (2010). Corruption and culture: An experimental analysis. Journal of Public Economics, 94(11–​12), 862–​869. Basu, K. (2011). Why, for a class of bribes, the act of giving a bribe should be treated as legal. Ministry of Finance Government of India Working Paper No. 1/​2011-​DEA (March 2011) https://​mpra.ub.uni-​muenc​hen.de/​50335/​ Basu, K. (2020). Republic of Beliefs: A new approach to law and economics. S.l.: Princeton University Press. Becker, G. S. (1968). Crime and punishment: An economic approach. In The economic dimensions of crime (pp. 13–​68). Springer. Beekman, G., Bulte, E. & Nillesen, E. (2014). Corruption, investments and contributions to public goods: Experimental evidence from rural Liberia. Journal of Public Economics, 115, 37–​47. Bjørnskov, C. (2021). Civic honesty and cultures of trust. Journal of Behavioral and Experimental Economics, 92, 101693. Blaufus, K., Bob, J., Otto, P. E. & Wolf, N. (2017). The effect of tax privacy on tax compliance: An experimental investigation. European Accounting Review, 26(3), 561–​580. Boly, A. & Gillanders, R. (2018). Anti-​corruption policy making, discretionary power and institutional quality: An experimental analysis. Journal of Economic Behavior & Organization, 152, 314–​327. Boly, A., Gillanders, R. & Miettinen, T. (2019). Deterrence, contagion, and legitimacy in anticorruption policy making: An experimental analysis. Journal of Legal Studies, 48(2), 275–​305. 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. Buckenmaier, J., Dimant, E. & Mittone, L. (2020). Effects of institutional history and leniency on collusive corruption and tax evasion. Journal of Economic Behavior & Organization, 175, 296–​313. Buffat, J. & Senn, J. (2017). Corruption and cooperation.Working paper No. 260, University of Zurich, Department of Economics, Zurich. Bühren, C. (2020). Staff rotation as an anti-​ corruption policy in China and in Germany: An experimental comparison. Jahrbücher Für Nationalökonomie Und Statistik, 240(1), 1–​18. Butler, J. V, Serra, D. & Spagnolo, G. (2020). Motivating whistleblowers. Management Science, 66(2), 605–​621. Cagala, T., Glogowsky, U., Grimm, V., Rincke, J. & Tuset Cueva, A. (2017). Does corruption affect the private provision of public goods? SSRN 3112049. https://​ pap​ers.ssrn.com/​sol3/​pap​ers.cfm?abst​ract​_​id=​3112​049

Stopping the Rot II  135 Cason, T. N., Friesen, L. & Gangadharan, L. (2016). Regulatory performance of audit tournaments and compliance observability. European Economic Review, 85, 288–​306. Chaudhuri, A. (2012). Gender and corruption: A survey of the experimental evidence. In New advances in experimental research on corruption (pp. 13–​49). Emerald Group Publishing Limited. Cohn, A., Maréchal, M. A.,Tannenbaum, D. & Zünd, C. L. (2019). Civic honesty around the globe. Science, 365(6448), 70–​73. Cummings, R. G., Martinez-​Vazquez, J., McKee, M. & Torgler, B. (2009). Tax morale affects tax compliance: Evidence from surveys and an artefactual field experiment. Journal of Economic Behavior & Organization, 70(3), 447–​457. d’Adda, G., Darai, D., Pavanini, N. & Weber, R. A. (2017). Do leaders affect ethical conduct? Journal of the European Economic Association, 15(6), 1177–​1213. De Graaf, G. (2007). Causes of corruption: Towards a contextual theory of corruption. Public Administration Quarterly 31(1–​2):, 39–​86. Denzau, A. T. & North, D. C. (2000). Shared mental models: Ideologies and institutions. In Elements of reason: Cognition, choice, and the bounds of rationality, ed. A Lupia, MC McCubbins, SL Popkin (pp. 23–​46). Cambridge Univ. Press Detkova, P., Tkachenko, A. & Yakovlev, A. (2021). Gender heterogeneity of bureaucrats in attitude to corruption: Evidence from list experiment. Journal of Economic Behavior & Organization, 189, 217–​233. Di Falco, S., Magdalou, B., Masclet, D., Villeval, M. C. & Willinger, M. (2015). Embezzlement: Does transparency of information matter? An experiment in Tanzania. Post-​Print halshs-​01184420, HAL. url: https://​ideas.repec.org/​p/​hal/​jou​r nl/​hal​shs-​ 01184​420.html Di Falco, S., Magdalou, B., Masclet, D., Villeval, M. C. & Willinger, M. (2020). Can shorter transfer chains and transparency reduce embezzlement? Review of Behavioral Economics, 7(2). Dreze, J. (2011). The bribing game. Indian Express, 23. (April 23, 2011) Drugov, M., Hamman, J. & Serra, D. (2014). Intermediaries in corruption: An experiment. Experimental Economics, 17(1), 78–​99. Engel, C., Goerg, S. J. & Yu, G. (2016). Symmetric vs. asymmetric punishment regimes for collusive bribery. American Law and Economics Review, 18(2), 506–​556. Enste, D. H. & Heldman, C. (2018).The consequences of corruption. In Handbook on the geographies of corruption (pp. 106–​119). Edward Elgar Publishing. Festinger, L. (1957). A theory of cognitive dissonance (vol. 2). Stanford University Press. Fišar, M., Krčál, O., Staněk, R. & Špalek, J. (2021). Committed to reciprocate on a bribe or blow the whistle: The effects of periodical staff-​rotation in public administration. Public Performance & Management Review, 44(2), 404–​424. Fišar, M., Kubák, M., Špalek, J. & Tremewan, J. (2016). Gender differences in beliefs and actions in a framed corruption experiment. Journal of Behavioral and Experimental Economics, 63, 69–​82. Foltz, J. D. & Opoku-​Agyemang, K. A. (2015). Do higher salaries lower petty corruption? A policy experiment on West Africa’s highways. Working paper, University of Wisconsin–​Madison and University of California, Berkeley. Frank, B., Lambsdorff, J. G. & Boehm, F. (2011). Gender and corruption: Lessons from laboratory corruption experiments. European Journal of Development Research, 23(1), 59–​71.

136  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra Fried, B. J., Lagunes, P. & Venkataramani, A. (2010). Corruption and inequality at the crossroad: A multimethod study of bribery and discrimination in Latin America. Latin American Research Review 45 (1):, 76–​97. Gächter, S., & Schulz, J. F. (2016). Intrinsic honesty and the prevalence of rule violations across societies. Nature, 531(7595), 496–​499. Galeotti, F., Maggian, V. & Villeval, M. C. (2021). Fraud deterrence institutions reduce intrinsic honesty. Economic Journal, 131(638), 2508–​2528. Gans-​Morse, J., Borges, M., Makarin, A., Mannah-​Blankson, T., Nickow, A. & Zhang, D. (2018). Reducing bureaucratic corruption: Interdisciplinary perspectives on what works. World Development, 105, 171–​188. García-​Gallego, A., Georgantzis, N., Jaber-​López,T. & Michailidou, G. (2020). Audience effects and other-​regarding preferences against corruption: Experimental evidence. Journal of Economic Behavior & Organization, 180, 159–​173. Gino, F., Ayal, S. & Ariely, D. (2009). Contagion and differentiation in unethical behavior: The effect of one bad apple on the barrel. Psychological Science, 20(3), 393–​398. Gneezy, U., Saccardo, S., & Van Veldhuizen, R. (2019). Bribery: Behavioral drivers of distorted decisions. Journal of the European Economic Association, 17(3), 917–​946. Guerra, A. & Zhuravleva,T. (2019). Social norms in corruption:A bribery experiment,Working Paper, EUI MWP, 2019/​07, European University Institute. url: https://​cad​mus.eui. eu/​han​dle/​1814/​63589 Guerra, A., & Zhuravleva, T. (2022). Do women always behave as corruption cleaners?. Public Choice, 191(1), 173–​192. Harri, A., Zhllima, E., Imami, D. & Coatney, K. T. (2020). Effects of subject pool culture and institutional environment on corruption: Experimental evidence from Albania. Economic Systems, 44(2), 100783. Jiang, T., Lindemans, J. W. & Bicchieri, C. (2015). Can trust facilitate bribery? Experimental evidence from China, Italy, Japan, and the Netherlands. Social Cognition, 33(5), 483–​504. Justesen, M. K. & Bjørnskov, C. (2014). Exploiting the poor: Bureaucratic corruption and poverty in Africa. World Development, 58, 106–​115. Khadjavi, M. (2015). On the interaction of deterrence and emotions. Journal of Law, Economics, & Organization, 31(2), 287–​319. Khadjavi, M. (2018). Deterrence works for criminals. European Journal of Law and Economics, 46(1), 165–​178. Khan, A. Q., Khwaja, A. I. & Olken, B. A. (2016). Tax farming redux: Experimental evidence on performance pay for tax collectors. Quarterly Journal of Economics, 131(1), 219–​271. Knack, S., & Keefer, P. (1997). Does social capital have an economic payoff? A cross-​ country investigation. The Quarterly Journal of Economics, 112(4), 1251–​1288. Kwon, I. (2014). Motivation, discretion, and corruption. Journal of Public Administration Research and Theory, 24(3), 765–​794. Lagunes, P. (2017). Guardians of accountability: A field experiment on corruption and inefficiency in Peru’s local public works. International Growth Center. Laske, K., Saccardo, S. & Gneezy, U. (2018). Do fines deter unethical behavior? The effect of systematically varying the size and probability of punishment. SSRN, April 5. https://​pap​ers.ssrn.com/​sol3/​pap​ers.cfm?abst​ract​_​id=​3157​387 Markussen, T., Sharma, S., Singhal, S., & Tarp, F. (2021). Inequality, institutions and cooperation. European Economic Review, 138, 103842.

Stopping the Rot II  137 Muralidharan, K., Niehaus, P. & Sukhtankar, S. (2016). Building state capacity: Evidence from biometric smartcards in India. American Economic Review, 106(10), 2895–​2929. Mutz, D. C. (2005). Social trust and e-​commerce: Experimental evidence for the effects of social trust on individuals’ economic behavior. Public Opinion Quarterly, 69(3), 393–​416. Navot, D., Reingewertz, Y. & Cohen, N. (2016). Speed or greed? High wages and corruption among public servants. Administration & Society, 48(5), 580–​601. Nguyen, Q. & Bernauer, T. (2019). Does social trust affect public support for international trade? Insights from an experiment in Vietnam. Political Studies, 67(2), 440–​458. North, D. C. (1997). Where have we been and where are we going? In Economics, Values and Organization (pp. 491–​508), ed.A. Ben-​Ner and L. Putterman. Cambridge: Cambridge University Press. Pan, X. & Houser, D. (2017). Social approval, competition and cooperation. Experimental Economics, 20(2), 309–​332. Peiffer, C. & Rose, R. (2018). Why are the poor more vulnerable to bribery in Africa? The institutional effects of services. Journal of Development Studies, 54(1), 18–​29. 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. Richey, S. (2010). The impact of corruption on social trust. American Politics Research, 38(4), 676–​690. Rivas, M. F. (2013). An experiment on corruption and gender. Bulletin of Economic Research, 65(1), 10–​42. Romaniuc, R., Dubois, D., Dimant, E., Lupusor,A. & Prohnitchi,V. (2021). Understanding cross-​cultural differences in peer reporting practices: Evidence from tax evasion games in Moldova and France. Public Choice, 190(1–​2), 127–​147. Rose-​Ackerman, S. (1978). Corruption A Study in Political Economy. Academic Press. Rose-​Ackerman, S. (1975). The economics of corruption. Journal of Public Economics, 4(2), 187–​203. Rothstein, B. & Eek, D. (2009). Political corruption and social trust: An experimental approach. Rationality and Society, 21(1), 81–​112. Ryvkin, D. & Serra, D. (2020). Corruption and competition among bureaucrats: An experimental study. Journal of Economic Behavior & Organization, 175, 439–​451. Ryvkin, D., Serra, D. & Tremewan, J. (2017). I paid a bribe: An experiment on information sharing and extortionary corruption. European Economic Review, 94, 1–​22. Salmon, T. C. & Serra, D. (2017). Corruption, social judgment and culture: An experiment. Journal of Economic Behavior & Organization, 142, 64–​78. Schram, A., Di Zheng, J. & Zhuravleva, T. (2022). Corruption: A cross-​country comparison of contagion and conformism. Journal of Economic Behavior & Organization, 193, 497–​518. Senci, C. M., Hasrun, H., Moro, R. & Freidin, E. (2019). The influence of prescriptive norms and negative externalities on bribery decisions in the lab. Rationality and Society, 31(3), 287–​312. Serra, D. (2012). Combining top-​down and bottom-​up accountability: evidence from a bribery experiment. Journal of Law, Economics, & Organization, 28(3), 569–​587. Serritzlew, S., Sønderskov, K. M. & Svendsen, G. T. (2014). Do corruption and social trust affect economic growth? A review. Journal of Comparative Policy Analysis: Research and Practice, 16(2), 121–​139.

138  Ritwik Banerjee, Utteeyo Dasgupta and Satarupa Mitra Slemrod, J. (2019).Tax compliance and enforcement. Journal of Economic Literature, 57(4), 904–​954. Srinivasan, T. N. (1973). Tax evasion: A model. Journal of Public Economics, 2, 339–​346. UNDP. (2017). Global report: Gender Equality in Public Administration (GEPA). www. undp.org/​publi​cati​ons/​glo​bal-​rep​ort-​gen​der-​equal​ity-​pub​lic-​adm​inis​trat​ion-​gepa. 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. Zak, P. J., & Knack, S. (2001).Trust and growth. The Economic Journal, 111(470), 295–​321. Zhang, N. (2015). Changing a “culture” of corruption: Evidence from an economic experiment in Italy. Rationality and Society, 27(4), 387–​413. Zhang, N. (2018). Institutions, norms, and accountability: A corruption experiment with northern and southern Italians. Journal of Experimental Political Science, 5(1), 11–​25. Zhang, N., Andrighetto, G., Ottone, S., Ponzano, F. & Steinmo, S. (2016). “Willing to pay?”Tax compliance in Britain and Italy: An experimental analysis. PloS One, 11(2), e0150277.

10 The Past, Present and Future of Research on Gender and Corruption Justin Esarey and N.Valdes

Dollar et al. (2001) were the first to report an interesting finding from cross-​ national time-​series data: countries with a greater proportion of women in their parliament also experience less corruption in their government according to measures like Transparency International’s (2019) Corruption Perception Index. This finding was quickly confirmed by Swamy et al. (2001), who also discovered an individual-​level relationship between tolerance for corruption and gender in the World Values Survey data (Inglehart et al., 2020). The relationship between gender and corruption they discovered was substantively strong and theoretically surprising. This discovery touched off a flurry of research attempting to understand how the gender–​corruption relationship works and how to apply it as a policy solution to the pressing problem of corruption around the world. In this chapter, we describe how research into gender and corruption has evolved over time and what we believe are the most promising avenues for future study. Although there is now little doubt among scholars that gender is causally entwined with tolerance for and participation in corruption, a common understanding of the processes that relate the two has not been reached despite years of sustained empirical research with causal inference techniques (Angrist and Krueger, 1999). We speculate that there is a confusing pattern of results in prior work despite sound research design because the relationship between gender and corruption is the product of multiple mechanisms that push and pull against one another. In some contexts, mechanisms linking gender and corruption push in the same direction and create a strong, statistically significant relationship between the two. In other contexts, these mechanisms can cancel each other out to create a weak relationship or none at all. Such findings do not translate easily into actionable policy precisely because (a) there are multiple causal mechanisms underlying them and (b) the strength and direction of these mechanisms is context sensitive. Our key argument is that future research should focus on separately identifying multiple causal mechanisms that make women less susceptible to corruption than men, with particular emphasis on how institutional, social or individual characteristics can condition or modify the relationship.This may require generating new theories and testing hypotheses from them, but (as we will explain in this chapter) many such theories already exist. Therefore, we think the most pressing need DOI: 10.4324/9781003142300-10

140  Justin Esarey and N.Valdes is to empirically separate and measure the strength of various causal pathways leading women to be more resistant to corruption. Special attention must also be paid to simultaneity when groups (countries, agencies, etc.) are the unit of analysis: more women in government can reduce corruption, but corruption can also reduce the participation of women in government. In a nutshell, we think that a simplistically falsificationist research program—​ that is, trying to collect evidence that critically tests the predictions of one particular theory—​is not quite appropriate for this field. Of course, hypotheses should still be derived from theories and tested with evidence. But the pattern of evidence to date implies that we cannot expect one theory to rise above the others in explaining why gender and corruption are connected. Some mechanisms will be operative in some contexts but not others; multiple mechanisms may drive behavior in certain places and times, while one mechanism dominates in another. Our goal must be to empirically separate and identify these mechanisms in contexts and, from this heterogeneity, develop an overarching theoretical framework describing why and how these mechanisms are activated or deactivated.

Empirical Origins The initial work into the relationship between gender and corruption was, for the most part, empirically driven; there was no strong theoretical framework from which the cross-​national association between gender and corruption was derived. Dollar et al. (2001) and Swamy et al. (2001) are holotypes for this stage of research in the field. Both are based on observational data. They offered possible interpretations for their findings but mostly left theory-​testing to future research. When early studies offered an explanation for their findings, they drew on psychological and sociological evidence of a more prosocial orientation of women compared to men. For example, Dollar et al. (2001) cites several studies purporting to show that women are on average more generous and helpful than men.1 One of these studies, Eckel and Grossman (2008), finds that men behave more selfishly than women in dictator games played in a laboratory (Hoffman et al., 1994), where each subject has a chance to unilaterally and anonymously decide how much of a pot of money to share with an anonymous partner. Swamy et al. (2001, p. 51) do not offer a causal explanation for the gender–​ corruption link, but its authors note that they are reassured to learn that our evidence is entirely consistent with the findings of leading criminologists. For instance, Gottfredson and Hirschi (1990, p. 194) show … that arrests for embezzlement per 100,000 white-​collar workers are higher for men for every age group. They also cite a variety of sources to make the case that across age groups, countries, and types of crime, the evidence regarding higher participation of men is remarkably uniform. The following summary statement from a study conducted by

The Past, Present and Future of Research  141 the National Academy of Sciences of the United States reflects the confidence with which the gender differential has been identified in the criminology literature:2 “The most consistent pattern with respect to gender is the extent to which male criminal participation in serious crimes at any age greatly exceeds that of females, regardless of the source of data, crime type, level of involvement, or measure of participation.” Evidence also suggests that, although both men and women in the United States vote based on what they perceive as the nation’s economic interests, men are (at least in some situations) more strongly influenced by their personal self-​interest than women (Kam, 2009; Clarke et al., 2005; Welch and Hibbing, 1992).3 Against this background of prior knowledge, which “impl[ies] that women will be less likely to sacrifice the common good for personal (material) gain,” it is logical to deduce that “increased female participation leads to more honest government” (Dollar et al., 2001, p. 424). There was some initial skepticism of whether the correlations reported by this early work represented a causal relationship between women and corruption. For example, Sung (2003) examined the possibility that initial findings were spurious, as the relationship between women in parliament and corruption becomes statistically and substantively weaker once a measure of democracy is controlled for. But this skepticism was quickly dampened by subsequent research that seemed to confirm the initial findings. Laboratory experiments like Schulze and Frank (2003), Alatas et al. (2009) and Beaman et al. (2009) demonstrated that spuriousness cannot fully explain the link between gender and corruption: at least in some environments (as described in the next section), women are less likely to engage in bribery even when facing identical incentives to men. Rivas (2013) found that women were no less likely than men to accept bribes but were less likely to reciprocate the bribe with preferential treatment. Indeed, some of this early work is optimistic that its results can translate directly into policy. Swamy et al. (2001, p. 53) concludes by saying that “given this evidence, we suspect the gender differential in corruption will be stable in the medium term, and policy initiatives [to increase women’s representation] … will indeed reduce corruption.” Even more surprisingly, some policymakers implemented feminization initiatives contemporaneously with the discovery of the gender–​corruption link (Moore, 1999; McDermott, 1999; UN Women, 2011). The increasing role for women in government worldwide had an unclear effect on corruption. For example, the establishment of a female-​only traffic police force seems to have decreased incidences of bribery in Peru, although that perception was not universally shared nor did it result in larger-​scale changes in the organization’s culture (Karim, 2011). Moreover, when studying women’s police stations that specialize in pursuing domestic and sexual violence cases, a case study in four Latin American countries found that “an all-​female staff does not necessarily guarantee better service quality” (UN Women, 2011, p. 5). It is not hard to come up with examples of countries with greatly increasing

142  Justin Esarey and N.Valdes levels of women’s representation but persistent corruption. Figure 10.1 shows two such examples: Mexico (Figure 10.1a) and South Africa (Figure 10.1b). In both cases, women’s representation in the lower house of the legislature has markedly increased between 1995 and 2020, reaching approximate parity by the end of the time period. Both countries still have substantial corruption according to the Varieties of Democracy Project’s Political Corruption Index (Coppedge et al., 2021a) and the Transparency International’s (2019) Corruption Perception Index. The uncertain effectiveness of feminization to reduce corruption (and the natural progress of scientific inquiry) led to the next stage of the research program studying corruption and gender: questioning the nature of the causal linkage between the two. If women were intrinsically and universally more resistant to corruption than men, as Dollar et al. (2001) suggested on the basis of sociological differences between men and women, we would expect to see greater women’s representation in government to be associated with lower corruption in that government virtually everywhere. But, as we see in Figure 10.1, that is not so. This is the basis for the puzzle that the second stage of research sought to solve.

Context Sensitivity and Reverse Causality As highlighted by the previous section, the early experiments designed to study the relationship between gender and corruption confirmed a causal link between the two. These papers bolstered the credibility of Dollar et al. (2001) and Swamy et al. (2001): their discoveries could not be explained away as mere confounding or an artifact of measurement. These experiments also suggested that feminization initiatives could actually work to reduce corruption. But these experiments also indicated that this relationship is complex: gender differences in behavior exist in some contexts but not in others. For example, Alatas et al. (2009) found that women were less likely to engage in corrupt activity but in only one of the four countries they studied. Field and lab experiments by Armantier and Boly (2013) on Canadian and Burkinabe subjects not only found no gender difference in willingness to accept a bribe but also that bribed women provide more corrupt benefits to their patron than bribed men. Puzzling findings like these motivated efforts to refine the causal explanation for the gender–​corruption link. Some of the research conducted at this time specifically studied the heterogeneity of the relationship between gender and corruption in different contexts. For example, Esarey and Chirillo (2013) showed that greater women’s representation is only associated with lower corruption in democracies, not dictatorships. Additionally, they found that the individual-​level gender gap in tolerance for corruption observed in surveys only exists in democracies as well. Esarey and Schwindt-​Bayer (2018) and Schwindt-​Bayer and Tavits (2016) further indicate that even within democracies, only those with relatively strong

newgenrtpdf

% Women in Lower House

TI Corruption Perception Index

V−Dem Overall Corruption

% Women in Lower House

V−Dem Overall Corruption

60

80

% women / corruption rating 0−100

60

40

40

20

20

0 1995

2000

2005

2010

2015

2020

1995

2000

2005

2010

year

year

(a) Mexico

(b) South Africa

2015

2020

Figure 10.1 Change in women’s representation and corruption over time for two countries. Note: The graphs show the change in corruption and women’s representation over time for two countries, Mexico (Figure 10.1a) and South Africa (Figure 10.1b). Corruption is measured using the Varieties of Democracy Political Corruption Index (Coppedge et al., 2021a) and the Transparency International’s (2019) Corruption Perception Index. Women’s representation is measured by the proportion of women in the lower house of the legislature from the V-​Dem v. 11.1 dataset (Coppedge et al., 2021b).

The Past, Present and Future of Research  143

% women / corruption rating 0−100

TI Corruption Perception Index

144  Justin Esarey and N.Valdes mechanisms of voter accountability have a strong negative association between women in government and corruption. This wave of research attempted to use patterns in the gender–​corruption relationship to rule out some causal explanations. For example, if a gender–​ corruption link only exists in democracies with strong accountability to voters, it is much more difficult to explain that link with women’s greater intrinsic moral aversion to corruption. If women were intrinsically averse to corruption, they would presumably be more likely than men to avoid it when serving in dictatorial governments or democracies without clarity of responsibility for policy outcomes. Esarey and Chirillo (2013) propose the alternative that voters may provide an extrinsic incentive for female politicians to avoid corruption by punishing them more harshly than men for corruption scandals; this incentive would only be relevant in democracies with clarity of responsibility (that stigmatize corruption) and not dictatorships (which operate via patronage, personalistic rule and/​or nomenklatura systems). Still, a large number of potential explanations were consistent with the evidence in the literature at this point. Esarey and Chirillo (2013) note that women’s well-​documented greater aversion to risk (Jianakoplos and Bernasek, 1998; Watson and McNaughton, 2007; Eckel and Grossman, 2008) might make them more likely than men to avoid corruption when it is risky (in democracies that punish it) but equally likely to participate in corruption when it is not risky (in dictatorships where corrupt behaviors are integral to governance). Or perhaps women are relative newcomers to the political scene and therefore do not have the opportunities to engage in corruption than more established politicians do (Goetz, 2007; Alhassan-​Alolo, 2007), but would be willing to do so once given the opportunity. Maybe cultural factors like masculinity (competitiveness) or power distance (acceptance of power inequality) mediate the link between women’s representation and corruption (Debski et al., 2018). Widespread corruption might also reduce women’s involvement in politics instead of (or in addition to) the other way around. Many of the arguments for this direction of causality focus on the effect of corruption on the recruitment of new candidates for office and/​or appointed bureaucratic positions.The key insight is that corruption creates the need for secrecy and loyalty among politicians to avoid exposing their corrupt activities. This need leads politicians to recruit and appoint people they believe they can trust, people who are like them in many ways—​including gender. Bjarnegård (2013) offers both cross-​ national statistical evidence and a detailed case study of Thailand to support her claim that “the clientelist context specifically carries with it incentives for individuals to accumulate homosocial capital” (p. 11), meaning that male-​ dominated governments are more likely to recruit candidates like themselves in corrupt, clientelist systems. Thus, when “(male) elites involved in cabinet formation will tend to appoint ministers whom they can trust with secretive tasks” (Stockemer and Sundström, 2019, p. 83), the result is that women are excluded from high-​ level executive branch positions. Along similar lines, Stockemer (2011, p. 697) argues that women are excluded from office because in corrupt

The Past, Present and Future of Research  145 countries “political seats can be bought and public officials are elected based on often male-​dominated clientelistic networks.” Sundström and Wängnerud (2016, p. 355) make a similar argument that “shadowy arrangements affect the recruitment of women in two ways: (i) they pose a direct obstacle to women when male-​dominated networks influence political parties’ candidate selection, and (ii) they pose an indirect obstacle when they influence citizens’ everyday life experiences and make them reluctant to engage in political matters” with this second effect being “larger for women than for men since women, generally speaking, have less power in society.” Summarily, this stage of research about gender and corruption succeeded at demonstrating that the relationship was far more complex than originally believed. It emphasized the need to rule out simultaneity in future observational studies. Experiments verified that women are (at least sometimes) less likely to approve of and participate in corruption but cannot determine how much this causal relationship explains the connection we see in cross-​national panel data where bidirectional causal linkages are possible. This tranche of research also failed to narrow the set of potential explanations for why more women’s representation is associated with lower corruption; indeed, the number of potential explanations that were broadly consistent with the available evidence grew substantially. Thus, this research underscored a pressing need for more observational research designed to produce a clear empirical picture of why, how, and when women in government are more resistant to corruption.

Theoretical Proliferation To rule out simultaneity and establish defensible causal links between gender and corruption in real-​world governance, the cutting edge of research has taken full advantage of observational causal inference research designs (Angrist and Krueger, 1999). This approach has clarified some important questions. For example, studies using a variety of different instrumental variables for women’s participation in government have confirmed that it does cause reduced corruption in government, at least in some contexts (Jha and Sarangi, 2018; Paweenawat, 2018; Esarey and Schwindt-​Bayer, 2019).4 It has also firmly established the bidirectionality of causal relationships between gender and corruption in government. For example, the instrumental variables approach taken by Esarey and Schwindt-​Bayer (2019) showed that corruption caused reduced women’s representation in addition to women’s representation reducing corruption. Stockemer and Sundström (2019) also found a negative relationship flowing from corruption to a reduced proportion of women in a country’s cabinet. Thus, we now know that it is imprudent to draw policy conclusions from the links we see in studies like Dollar et al. (2001). Observational data is produced by a simultaneous relationship between gender and corruption that is analogous to data on price and quantity that are produced by the intersection of supply and demand curves. If we want to know how much increasing women’s representation will reduce corruption, we must work considerably harder.

146  Justin Esarey and N.Valdes In most cases, careful empirical investigation suggests multiple theoretical mechanisms of influence. A pattern of confusing, contradictory findings characterizes the empirical record for most of the possible explanations for a link between gender and corruption. There are a large number of instances of this phenomenon in recently published work. For example, new findings have resurrected explanations for the gender–​ corruption link that had been apparently falsified in the past. Esarey and Chirillo (2013) concluded that intrinsic gender differences were ruled out by their finding that women are only more resistant to corruption in democracies (see also Boehm, 2015). But Bauhr et al. (2018) argues that it is women representatives’ stronger interest in public service delivery (and their greater propensity to break up clientelist networks) that results in reduced corruption when they are elected to office (p. 1060). In particular, because women officials pursue policies “that benefit traditionally female-​oriented sectors, such as education and health care,” the need for petty corruption to obtain these services is reduced among the primary consumers (women citizens). Their conclusion is supported by data from the European Union in 2013, where they found that (a) greater women’s representation in regional legislatures is associated with lower levels of grand and petty corruption in the region but (b) that female citizens’ experience with corruption is substantially more reduced than men’s. This evidence suggests that an intrinsic difference in the policy preferences of female politicians drives part of the reduction in corruption that they create. Such evidence supports the argument of Wängnerud (2020, p. 2) that “an influx of women into political institutions is accompanied with an influx of empathetic and other-​regarding values and that the important change, leading to lower levels of corruption, is that self-​regarding values, rather than individual men, are replaced.” Thus, we now have reason to believe that there are intrinsic—​albeit not necessarily natural, permanent, or essential—​differences in how men and women think about corruption. On the other hand, new evidence has also emerged to support the idea that increased risk aversion among women does indeed reduce their willingness to engage in corruption. Barnes and Beaulieu (2019) conducted a survey experiment in the United States designed to test how respondents evaluated a politician’s likelihood of accepting illegal payments from lobbyists. The experiment placed multiple explanations into competition, examining whether voters thought women would be less susceptible to corruption because (a) they were more honest; (b) they were more averse to the risk of being caught; or (c) they were excluded from opportunities for corruption.They “find strong evidence to suggest that perceptions of risk aversion help to explain why female politicians reduce suspicions of corruption” but “that honesty and marginalization do not have similar effects” (p. 158).5 Thus, the risk aversion hypothesis offered by (among others) Esarey and Chirillo (2013) is vindicated by new evidence. The perceived intrinsic honesty of women only influenced male subjects in the experiment, partially but not fully consistent with Wängnerud’s (2020) theory

The Past, Present and Future of Research  147 about the key role of other-​regarding values in mediating the link between gender and corruption. What about the hypothesis that women politicians face harsher punishment from voters for engaging in corruption compared to otherwise equivalent men? Survey experiments conducted by Batista Pereira (2020) in Brazil suggest that corrupt women politicians face harsher punishment from voters than corrupt men, but this same study finds no such relationship for essentially identical survey experiments in Mexico. A survey vignette experiment administered in the United Kingdom by Eggers et al. (2018) finds that male voters treat men and women members of parliament (MPs) involved in an expenses scandal equally, but that female voters are less likely to vote for implicated women MPs compared to implicated men MPs. Another survey experiment by Schwindt-​ Bayer et al. (2018) finds no evidence for differential punishment of women at all in either Brazil or the United States. As yet another example of this confusing pattern of results, consider the following set of studies about the relationship between women’s longevity in office and their susceptibility to corruption. Bauhr and Charron (2020) use a regression discontinuity design on data from French municipalities to show that newly elected women mayors who narrowly won election are associated with less corruption than comparable male incumbents, but that female incumbents are not; they argue that political outsiders who are not yet integrated into networks of corruption provide only a temporary respite from corruption until these newcomers are integrated into the system. Afridi et al. (2017) find the opposite result in India, where corruption and efficiency are at first worse in localities randomly assigned to have electoral quotas for women but eventually move toward parity with those localities without quotas. Meanwhile, Brollo and Troiano (2016) find (using a regression discontinuity design) that municipalities in Brazil with female mayors are consistently less corrupt than those with male mayors. It is difficult to draw any firm conclusion about whether electing women to public office reduces corruption from this record of research. In sum, the overall impact of the most recent phase of research about corruption and women’s representation was not to narrow the range of theoretical explanations for the relationship between the two, but to provide support for many different explanations and expand the number of viable theories. More concerningly, as a whole these studies provided contradictory evidence in support of and opposition to essentially all of these theories. In that sense research has failed to produce a clear empirical picture of why, how, and when women in government are more resistant to corruption.

Conclusion Researchers studying the link between representation of women in government and the corruption of that government have failed to arrive upon a unified explanation for the relationship between gender and corruption. The

148  Justin Esarey and N.Valdes current literature renders a sort of dodo bird verdict in the competition among theories: “everyone has won and all must have prizes” (Luborsky et al., 1975).6 Based on this pattern of findings, we recommend that the research community redirect their efforts away from studies designed to falsify (or critically test) a particular explanation for the gender–​corruption relationship. Such studies produce evidence for, against, or indifferent to various explanations depending on details of the research design (Incerti, 2020), the institutions and culture of the country studied, characteristics of the local and global environment peculiar to the time of the study, the nature of the treatment and perhaps other factors we have not considered. But we do not recommend giving up. There are important conclusions that can be drawn from the current literature and these conclusions suggest a way forward. The evidence consistently suggests that causality flows in both directions and is strong enough to be politically important. We think the best interpretation of an otherwise confusing pattern of confirmation and falsification is that many causal mechanisms for the gender–​corruption relationship are in simultaneous operation and the magnitude of any particular mechanism’s effect depends on the context in which it is evaluated. For this reason, we believe that the future of this research program lies in determining where and when these explanations will tend to work together to produce a large causal relationship between women’s representation and corruption. We speculate that the extremely strong empirical relationship between corruption in a country and women’s share of legislative seats in that country that was discovered by Dollar et al. (2001) and Swamy et al. (2001) is not the result of a single mechanism, but multiple mechanisms that are simultaneously influencing behavior. Where these mechanisms push in the same direction, they produce a large, politically important causal link between gender and corruption. Where they push in opposing directions—​such as in dictatorships (Esarey and Chirillo, 2013)—​they may cancel each other out. And, of course, these separate mechanisms may be individually more or less activated by any policy designed to decrease corruption by boosting women’s representation in government. Even in contexts where we observe (for example) that greater women’s participation reduces corruption, artificially or exogenously increasing women’s participation via a policy change may not have the same effect if the nature of the policy neutralizes or reverses the mechanism by which the causal link operates. Consequently, we think that the next stage of gender and corruption research should not be focused on conducting critical tests of the predictions of various theories. These tests will continue to be a part of this research, but the focus should be on being able to separately identify and measure the strength of various mechanisms linking gender to corruption in a particular context. Sometimes, a study will fail to reject the null hypothesis that a particular mechanism has no influence on behavior in that context. But we should expect that more than one mechanism is operating at once. When studying groups (e.g., countries), we should also expect that these mechanisms are bidirectional.

The Past, Present and Future of Research  149 As one example of the kind of research we have in mind, Stensöta et al. (2015) studies the effect of increasing women’s representation in the administrative (bureaucratic) state on corruption. They study the bureaucracy because its institutional context may change how gender influences corruption when compared to similar relationships in the legislature. Indeed, they find that “the curbing effect of women representatives on corruption is greater in the electoral than in the bureaucracy arena” and that the magnitude of the relationship in the bureaucracy is inversely proportional to the administration’s organizational strength and identity (pp. 492–​493). We see this work not as an attempt to support one theory or falsify another, but instead as an attempt to determine how various theories apply to behavior in a context where some may be activated and others deactivated to produce an overall causal relationship between women and corruption. Another model for future work comes from Pereira and André Melo (2015) and Pavão (2018). They provide evidence from Brazil that voters may not hold elected officials accountable for corruption if they do not believe that there is a viable alternative candidate who is genuinely anticorruption7 or if they believe that they benefit from their representative’s ability to bring public spending to their area. Thus, even if disproportionate voter punishment of corrupt female politicians can explain why women are less involved in corruption than men, we would not expect this pressure to operate in environments with pervasive corruption or when women are able to direct disproportionate spending to their constituents.These papers approach their research questions from the perspective of determining which explanations work in a particular context and why (or why not). They are trying to isolate individual mechanisms and figure out what makes them stronger or weaker in a given context. Eventually, that kind of evidence should enable us to develop and test an overarching theory linking these mechanisms together. That is why we believe this sort of work will be the most impactful during the next phase of the gender and corruption research program.

Notes 1 It is worth noting that the first meta-​analytic study they cite, Eagly and Crowley (1986), actually comes to the opposite conclusion. Dollar et al. summarize this study by saying that “women are more likely to exhibit ‘helping’ behavior” (p. 423), but the study itself concludes that “in general men helped more than women and women received more help than men” (p. 283). 2 Blumstein et al. (1986), cited in Gottfredson and Hirschi (1990). 3 For example, “Men rely significantly on how their households are doing financially in their decision calculus.When times are tough for them and their own, they punish the incumbent. Women, on the other hand, rely significantly less on household economic evaluations in their vote choice” (Kam, 2009, p. 620). 4 For example, Esarey and Schwindt-​Bayer (2019) examine only democracies on the basis of earlier findings from Esarey and Schwindt-​Bayer (2018) and Esarey and Chirillo (2013) and thus its findings are limited to that context.

150  Justin Esarey and N.Valdes 5 See also Barnes et al. (2018), which comes to similar conclusions. 6 From p. 995: “The subtitle you will recognize since it is from Alice in Wonderland—​it was the dodo bird who handed down this happy verdict after judging the race.” This phrase characterized their contemporaneous assessment of the state of research comparing efficacy of different modes of psychotherapy. 7 Klašnja et al. (2016) come to a similar conclusion based on their study of Slovakians.

References Afridi, F., Iversen,V. and Sharan, M. R. (2017). Women political leaders, corruption, and learning: Evidence from a large public program in India. Economic Development and Cultural Change, 66(1):1–​30. Alatas, V., Cameron, L., Chaudhuri, A., Erkal, N. and Gangadharan, L. (2009). Gender, culture, and corruption: Insights from an experimental analysis. Southern Economic Journal, 75(3):663–​680. Alhassan-​Alolo, N. (2007). Gender and corruption: Testing the new consensus. Public Administration and Development, 237:227–​237. Angrist, J. D. and Krueger, A. B. (1999). Empirical strategies in labor economics. In Handbook of labor economics, vol. 3 (pp. 1277–​1366). Elsevier. Armantier, O. and Boly, A. (2013). Comparing corruption in the laboratory and in the field in Burkina Faso and in Canada. Economic Journal, 123(573):1168–​1187. Barnes,T. D. and Beaulieu, E. (2019).Women politicians, institutions, and perceptions of corruption. Comparative Political Studies, 52(1):134–​167. Barnes, T. D., Beaulieu, E. and Saxton, G. W. (2018). Restoring trust in the police: Why female officers reduce suspicions of corruption. Governance, 31(1):143–​161. Batista Pereira, F. (2020). Do female politicians face stronger backlash for corruption allegations? Evidence from survey-​ experiments in Brazil and Mexico. Political Behavior. https://​doi.org/​10.1007/​s11​109-​020-​09602-​9. Bauhr, M. and Charron, N. (2020). Will women executives reduce corruption? Marginalization and network inclusion. Comparative Political Studies. https://​doi.org/​ 10.1177/​00104​1402​0970​218. Bauhr, M., Charron, N. and Wängnerud, L. (2018). Exclusion or interests? Why females in elected office reduce petty and grand corruption. European Journal of Political Research. https://​doi.org/​10.1111/​1475-​6765.12300. Beaman, L., Chattopadhyay, R., Duflo, E., Pande, R. and Topalova, P. (2009). Powerful women: Does exposure reduce bias? Quarterly Journal of Economics, 124(4):1497–​1540. Bjarnegård, E. (2013). Gender, informal institutions and political recruitment: Explaining male dominance in parliamentary representation. Palgrave Macmillan. Blumstein, A., Cohen, J., Roth, J. A. and Visher, C. A., editors (1986). Criminal careers and “career criminals,” vol. 1. National Academies Press. Boehm, F. (2015). Are men and women equally corrupt? U4 Brief. www.cmi.no/​publi​ cati​ons/​5851-​are-​men-​and-​women-​equa​lly-​corr​upt Brollo, F. and Troiano, U. (2016). What happens when a woman wins an election? Evidence from close races in Brazil. Journal of Development Economics, 122:28–​45. Clarke, H. D., Stewart, M. C., Ault, M. and Elliott, E. (2005). Men, women and the dynamics of presidential approval. British Journal of Political Science, 35(1):31–​51. Coppedge, M., Gerring, J., Knutsen, C., Lindberg, S. I., Teorell, J., Altman, D., Bernhard, M., Cornell, A., Fish, M., Gastaldi, L., et al. (2021a). V-​Dem codebook v11.1. www.v-​ dem.net/​en/​data/​data/​v-​dem-​data​set-​v111.

The Past, Present and Future of Research  151 Coppedge, M., Gerring, J., Knutsen, C. H., Lindberg, S. I.,Teorell, J., Alizada, N., Altman, D., Bernhard, M., Cornell, A., Fish, M. S., et al. (2021b). V-​Dem dataset v11.1. www.v-​ dem.net/​en/​data/​data/​v-​dem-​data​set-​v111. Debski, J., Jetter, M., Mösle, S. and Stadelmann, D. (2018). Gender and corruption: The neglected role of culture. European Journal of Political Economy, 55:526–​537. Dollar, D., Fisman, R. and Gatti, R. (2001). Are women really the “fairer” sex? Corruption and women in government. Journal of Economic Behavior & Organization, 46(4):423–​429. Eagly, A. H. and Crowley, M. (1986). Gender and helping behavior: A meta-​analytic review of the social psychological literature. Psychological Bulletin, 100(3):283. Eckel, C. C. and Grossman, P. J. (2008). Men, women, and risk aversion: Experimental evidence. In C. Plott and V. Smith, editors, Handbook of experimental economic results, vol. 1 (pp. 1061–​1073). Elsevier. Eggers, A. C., Vivyan, N. and Wagner, M. (2018). Corruption, accountability, and gender: Do female politicians face higher standards in public life? Journal of Politics, 80(1):321–​326. Esarey, J. and Chirillo, G. (2013). “Fairer sex” or purity myth? Corruption, gender, and institutional context. Politics and Gender, 9(4):390–​413. Esarey, J. and Schwindt-​Bayer, L. (2018). Women’s representation, accountability, and corruption in democracies. British Journal of Political Science, 48(3):659–​690. Esarey, J. and Schwindt-​ Bayer, L. (2019). Estimating causal relationships between women’s representation in government and corruption. Comparative Political Studies, 52(11):1713–​1741. Goetz, A. M. (2007). Political cleaners: Women as the new anti-​corruption. Development and Change, 38(1):87–​105. Gottfredson, M. R. and Hirschi, T. (1990). A general theory of crime. Stanford University Press. Hoffman, E., McCabe, K., Shachat, K. and Smith,V. (1994). Preferences, property rights, and anonymity in bargaining games. Games and Economic Behavior, 7(3):346–​380. Incerti, T. (2020). Corruption information and vote share: A meta-​analysis and lessons for experimental design. American Political Science Review, 114(3):761–​774. Inglehart, R., Haerpfer, C., Moreno, A., Welzel, C., Kizilova, K., Diez-​Medrano, J., Lagos, M., Norris, P., Ponarin, E., Puranen, B., editors (2020). World values survey: All rounds: Country-​pooled datafile. JD Systems Institute and WVSA Secretariat. www. worldv​alue​ssur​vey.org/​WVSD​ocum​enta​tion​WVL.jsp. Jha, C. K. and Sarangi, S. (2018).Women and corruption:What positions must they hold to make a difference? Journal of Economic Behavior & Organization, 151:219–​233. Jianakoplos, N. A. and Bernasek, A. (1998). Are women more risk averse? Economic Inquiry, 36(4):620–​630. Kam, C. D. (2009). Gender and economic voting, revisited. Electoral Studies, 28(4):615–​624. Karim, S. (2011). Madame officer. Americas Quarterly, 5(3). www. americasquarterly.org/​ node/​2802. Klašnja, M., Tucker, J. A. and Deegan-​Krause, K. (2016). Pocketbook vs. sociotropic corruption voting. British Journal of Political Science, 46(1):67–​94. Luborsky, L., Singer, B. and Luborsky, L. (1975). Comparative studies of psychotherapies: Is it true that “everyone has won and all must have prizes”? Archives of General Psychiatry, 32(8):995–​1008. https://​doi.org/​10.1001/​archp​syc.1975. 01760260059004. McDermott, J. (1999). International: Women police ride in on a ticket of honesty. Daily Telegraph. July 31.

152  Justin Esarey and N.Valdes Moore, M. (1999). Mexico City’s stop sign to bribery; to halt corruption, women traffic cops replace wen. Washington Post. July 31. www.highb​eam. com/​doc/​1P2-​605613. html. Pavão, N. (2018). Corruption as the only option: The limits to electoral accountability. Journal of Politics, 80(3):996–​1010. Paweenawat, S. W. (2018). The gender-​corruption nexus in Asia. Asian-​Pacific Economic Literature, 32(1):18–​28. https://​doi.org/​10.1111/​apel.12214. Pereira, C. and André Melo, M. (2015). Reelecting corrupt incumbents in exchange for public goods: Rouba mas faz in Brazil. Latin American Research Review, 50(4):88–​115. Rivas, M. F. (2013). An experiment on corruption and gender. Bulletin of Economic Research, 65(1):10–​42. http://​online​libr​ary.wiley.com/​doi/​10.1111/​j.1467-​8586. 2012.00450.x/​abstract. Schulze, G. G. and Frank, B. (2003). Deterrence versus intrinsic motivation: Experimental evidence on the determinants of corruptibility. Economics of Governance, 4:143–​160. Schwindt-​ Bayer, L. A., Esarey, J. and Schumacher, E. (2018). Gender and citizen responses to corruption among politicians: The US and Brazil. In H. Stensöta and L. Wängnerud, editors, Gender and corruption (pp. 59–​82). Palgrave Macmillan. Schwindt-​Bayer, L. A. and Tavits, M. (2016). Clarity of responsibility, accountability and corruption. Cambridge University Press. Stensöta, H., Wängnerud, L. and Svensson, R. (2015). Gender and corruption: The mediating power of institutional logics. Governance, 28(4):475–​496. Stockemer, D. (2011). Women’s parliamentary representation in Africa: The impact of democracy and corruption on the number of female deputies in national parliaments. Political Studies, 59(3):693–​712. Stockemer, D. and Sundström, A. (2019). Corruption and women in cabinets: Informal barriers to recruitment in the executive. Governance, 32(1):83–​102. Sundström, A. and Wängnerud, L. (2016). Corruption as an obstacle to women’s political representation: Evidence from local councils in 18 European countries. Party Politics, 22(3):364–​369. Sung, H.-​E. (2003). Fairer sex or fairer system? Gender and corruption revisited. Social Forces, 82:703–​723. Swamy, A., Knack, S., Lee, Y. and Azfar, O. (2001). Gender and corruption. Journal of Development Economics, 64(1):25–​55. Transparency International. (2019). Corruption perceptions index overview. https://​ bit.ly/​2Zyd​9DC. UN Women Virtual Knowledge Centre to End Violence against Women and Girls. (2011). Women’s police stations/​ units. www.endvaw​now.org/​uplo​ads/​ browser/​ files/​security_​wps_​case_​study.pdf. Wängnerud, L. (2020).Why women in elected assemblies reduce levels of corruption:The indirect approach. ECPR General Conference Online, August 24–​28. https://​ecpr.eu/​ Filest​ore/​paperp​ropo​sal/​ 20c5788c-​29a0-​4464-​9834-​76c888414f7a.pdf. Watson, J. and McNaughton, M. (2007). Gender differences in risk aversion and expected retirement benefits. Financial Analysts Journal, 63(4):52–​62. Welch, S. and Hibbing, J. (1992). Financial conditions, gender, and voting in American national elections. Journal of Politics, 54(1):197–​213.

11 The Culture-​Corruption Hypothesis Revisited Organizational Culture, Corruption and Worker Preferences Sheheryar Banuri

Introduction Since Mauro’s seminal paper in 1995, the empirical literature on corruption has grown exponentially. This increase has been largely attributed to improvements in the measurement of corruption. This line of research has typically focused on survey-​based measures of corruption perceptions (Treisman, 2000; Knack and Keefer, 1995), expert informed measures (Mauro, 1995; Rauch and Evans, 2000), government official behaviour (Fisman and Miguel, 2007; Hanna and Wang, 2017) and even behaviour of the general public (Cameron et al., 2009; Barr and Serra, 2010; Banuri and Eckel, 2015; Banuri and Keefer, 2016a; Olsen et al., 2019). Improvements in the measurement of such clandestine activity have led to a deeper understanding of the effects of corruption, with the literature demonstrating negative impacts on economic growth, educational attainment and institutional quality (Lambsdorff, 2006). Corruption is widespread, and differs substantially, both within and across countries (defined as the use of public office for private gain1 –​ Treisman, 2007). One of the outstanding issues in this literature is the relationship between corruption and culture (defined as a set of shared norms, values and attitudes2 –​ Desmet et al., 2017). The “Culture of Corruption” hypothesis (Barr and Serra, 2010) states that a prescriptive social norm exists which promotes “abstinence from corruption” in some contexts but not in others. That is, engaging in corruption is a type of social norm that exists in some settings but not in others. In this chapter I explore the culture–​corruption relationship using empirical papers focusing specifically on said relationship. I first review papers that unpack the relationship between culture and corruption by examining the empirical literature that establishes the linkage between corruption perceptions and corrupt behaviour and asking whether this behaviour is particular to public officials or prevalent in the general population. Next, I focus on whether perceptions of corruption impact selection: Are workers who perceive greater levels of corruption in the public sector more likely to join the public sector? Following this, I then focus on a small literature on socialization: whether corruption and increase or decrease over time (and through interventions). DOI: 10.4324/9781003142300-11

154  Sheheryar Banuri Finally, I ask what the role of pecuniary incentives is in attracting workers with the “right” preferences into public sector roles. Within each section I highlight the need for further research and critically examine the methods used to answer the questions.This review focuses primarily on lab and field experiments, but other studies, including theoretical and observational, are also included to the extent possible.

A Culture of Corruption? First, I turn to the empirical literature reporting the correlates of corruption behaviour/​perceptions (Treisman, 2000; Rauch and Evans, 2000) to identify empirical regularities distinguishing countries with strong norms, values and attitudes of corruption (a culture of corruption) from countries with weak norms, values and attitudes of corruption. This literature predicts that individuals emerging from countries with weak norms, values and attitudes of corruption are more likely to display corrupt behaviour in their everyday lives. Fisman and Miguel (2007) document precisely this, showing that (prior to a law change in 2003) diplomats from corrupt countries, stationed in New York, were more likely to have a higher number of unpaid parking tickets. The authors note that this constitutes a revealed preference measure of corrupt behaviour because mission personnel (and families) benefitted from protections from minor legal infractions (such as parking tickets) due to their diplomatic status. The authors argue that parking illegally constitutes a behavioural measure of corruption because the individual in question benefits directly, but is protected by their diplomatic status, and hence is able to use their public office for private gain. Countries with stronger corruption norms would be expected to amass a greater number of infractions (controlling for mission size and diplomat income levels). Fisman and Miguel (2007) show that the number of unpaid parking tickets (prior to the rule change) correlated with the control of corruption measure from the World Bank’s World Governance Indicators dataset (Kaufmann et al., 2005). This measure is a perceptions-​based measure of corruption, taken from “enterprise, citizen, and expert survey respondents.” The relationship suggests that countries with a high degree of perceptions of corruption have diplomatic missions (public officials) that engage in a higher degree of corrupt practices. This paper provides an important validation of perceptions-​based measures as well as a methodological innovation in the measurement of corruption using revealed preferences. The paper documents individuals from a specific area of the public sector exhibiting behaviour consistent with perceptions of the public sector at large. At the same time, the paper is limited in scope, as diplomats are a specific type of public agent and not representative of the public sector. The paper is also notable to be among the first economics paper to document actual behaviour of public officials and to demonstrate the relationship between perceptions and behaviour, something that the empirical literature had argued for, but had not demonstrated.

The Culture-Corruption Hypothesis  155 Barr and Serra (2010) report the results of an experiment carried out at Oxford University that replicates the result from Fisman and Miguel (2007) using an incentivized laboratory experiment with a student sample. Barr and Serra (2010) have subjects play a one-​shot-​adapted version of a bribery game (introduced by Abbink et al., 2002).The game consists of three players –​a “private citizen,” a “public official” and “other members of society.” Private citizens start the game by choosing whether, and how much, to offer to the public official as a “bribe.” Upon observing the bribe offer, public official chooses whether to accept or reject the offer. If the bribe is accepted, other members of society incur a cost, simulating the negative impact on welfare due to corrupt practices. Many of the laboratory versions of bribery games covered in this chapter share similar features, so it is worth discussing these in detail here. Essentially bribery games simulate real-​world bribery settings by putting individuals in a social dilemma where accepting bribes is individually beneficial (both for the bribe initiator and the official) but imposes a social cost (usually on an individual or set of individuals) that outweighs the benefits incurred to bribers and officials.3 Barr and Serra’s (2010) experiment is notable in that the subject pool comprises university students from 34 countries, with a full two-​thirds of the sample coming from outside the United Kingdom and from countries with differing levels of corruption. Barr and Serra (2010) find that undergraduate (but not graduate) students originating from countries with higher perceptions of corruption were more likely to offer and to accept a bribe. They interpret this result as similar to the one reported by Fisman and Miguel (2007): countries with higher perceptions of corruption also generate citizens that are more tolerant of bribery. Hence private citizens emerging from corrupt cultures are more likely to engage in corrupt practices, providing suggestive evidence that norms, values and attitudes of corruption exist within broader society, rather than just limited to some public sector institutions (as shown by Fisman and Miguel, 2007). Taken together, the papers show that corruption permeates culture and even affects behaviour of those entering a new culture. These experiments are useful in that they demonstrate how individuals behave when faced with the exact same decision (and the exact same consequences). Average citizens (undergraduate students) and a set of public officials (diplomats) are more likely to engage in corrupt practices if they come from a country with higher perceptions of corruption. Importantly, however, Barr and Serra (2010) find that this relationship between corruption perceptions and corruption behaviour exists within undergraduate students, but not graduate students.When Barr and Serra (2010) restrict their analysis to graduate students, the relationship between baseline corruption levels in the country of origin and bribing behaviour does not hold, casting some doubt on the culture-​corruption hypothesis.4 This conditionality indicates that while corrupt practices may indeed be a cultural phenomenon (i.e., shared norms of corruption permeate all aspects of society), preferences can drive some of this behaviour and that corruption norms can be impacted

156  Sheheryar Banuri by (i) selecting honest workers and (ii) providing secondary socialization to those joining public office. Differences in behaviour across cultures are found in experiments using different tasks. For example, Gachter and Schulz (2016) use a die-​roll task with student samples in 23 countries.5 The die-​roll task asks students to roll a die that is unobservable to anyone other than the student and then to report the number rolled. Students are paid according to the number reported, with higher numbers paying out higher amounts. Even though the probability of rolling any one number is equal, the use of incentives increases the likelihood of higher numbers being reported. This allows the authors to construct a measure of cheating by comparing the reported distribution with the (known) theoretical distribution. Gachter and Schulz (2016) construct an index of the prevalence of rule violations for each country (using data on corruption, tax evasion and fraudulent politics). They find that students from societies with high rule violations report systematically higher numbers (i.e., are more dishonest in the die-​roll task), consistent with the culture-​corruption hypothesis. However, this paper is additionally notable for constructing the index using data from 2003. That is, the data for rule violations are obtained from a time when none of the sample would have been aware of rule violations or engaged in any rule-​ breaking themselves.6 Cameron et al. (2009) conduct a one-​shot bribery game, with student subjects in three distinct roles –​a “Firm” (i.e., briber), an “Official” (i.e., bribee) and a “Citizen” (i.e., individual incurring a negative externality). The authors implement two conditions –​one where bribes are welfare-​enhancing and another where bribes are welfare-​reducing. They conduct the experiment in four countries with very different levels of corruption: from lowest to highest perceptions of corruption, Singapore (ranked 3rd in 2009 according to the Corruption Perceptions Index), Australia (ranked 8th), India (84th) and Indonesia (111th). In addition, the authors use loaded language using terms like “bribe” and “punishment” in their instructions, which departs from neutral language used in many other lab studies. The authors asked whether subjects in countries with high perceived corruption engage in greater corrupt practices and punish such practices less than similar samples in countries with less perceived corruption. As above, the idea here was that in individuals from countries with shared norms (i.e., culture) of corruption are more tolerant of corrupt practices. Their results are puzzling: while Indian subjects were the most tolerant of corruption (i.e., engaging in more bribery and less punishment of corruption), the culture that was least tolerant of corruption was Indonesia. Australian and Singaporean subjects displayed behaviour that was statistically indistinguishable, but significantly less tolerant than Indian subjects and more tolerant than Indonesian subjects.The authors note that the relationship between culture and corruption is likely to be more complex, with both Singapore and Indonesian subjects displaying tolerance of corruption that is not in line with corruption perceptions.

The Culture-Corruption Hypothesis  157 Banuri and Eckel (2015) conduct a lab experiment in two countries with very different levels of perceived corruption, the United States (ranked 16th in 2015) and Pakistan (ranked 117th). The experiment conducts a repeated version of the three-​ person corruption game, conducted over 30 rounds. However, this paper departs from previous literature by implementing punishment for a short time (simulating a “crackdown” –​a short-​lived policy focusing on punishing corruption). The first ten rounds are conducted with no punishment, the next ten with punishment and then a final ten with no punishment. The results show that in a repeated framework, bribe offers and acceptances are statistically indistinguishable across cultures in the absence of punishment. Differences arise within the punishment rounds, where corrupt practices in the United States (low corruption setting) are lower than in Pakistan (the high corruption setting). Once punishment is removed, corruption rebounds back to pre-​crackdown levels. Again, corrupt behaviour (in the absence of punishment) does not differ in the direction predicted by corruptions perceptions indices. Finally, another paper that reports similar (lack of) effects across cultures is by Armantier and Boly (2013), who conduct experiments with student subjects in two countries that differ in their levels of perceived corruption: Canada (ranked 9th in 2013) and Burkina Faso (ranked 83rd). Their game is a bespoke version of a field experiment conducted in Burkina Faso (first reported in Armantier and Boly, 2011), where workers were hired to grade examinations, but some exams have a small bribe and a note requesting beneficial treatment. The lab experiments implemented the same conditions as in the field, except that the subjects knew they were participating in an economics experiment. Hence, the experimental “game” constitutes subjects hired to mark exams, with a small bribe and a request for a favour attached to some of the exam papers. This paper finds remarkably similar results between the two countries once controlling for observable characteristics (gender, age, ability). Indeed, the only difference reported is the effect of a high bribe which increases bribe acceptance in Burkina Faso, but not in Canada, though this may be the result of different magnitudes of the bribe increase across cultures. Nevertheless, these results from lab samples all show that corruption behaviour varies across cultures but does not always align with corruption perceptions. Overall the studies presented in this section point to a culture of corruption affecting behaviour. Public officials and university students from countries with higher perceived levels of corruption were more likely to engage in corrupt practices, consistent with the culture-​corruption hypothesis. If the culture-​corruption hypothesis is true, the average individual from a corrupt country would be likely to exhibit corrupt behaviour. However, data from lab experiments, using non-​public officials, conducted in different countries, find behaviour out of step with that predicted by corruption indices. While differences in design might be plausible reasons for these inconsistencies, it may also be the case that individuals do not differ in corrupt practices, but public officials do. That is, public officials from countries with higher perceived corruption are indeed more corrupt due to selection, rather than culture. In

158  Sheheryar Banuri other words, the public sector in high corruption settings attract workers with stronger preferences for corruption.

Do Corrupt Workers Select into Public Sector Jobs? A nascent literature in economics and public administration focuses on hiring workers with the “right” preferences. Individuals vary in their preferences for corruption.This affects the extent to which these workers are resistant to norms of corruption in the public sector. If the public sector attracts workers with strong preferences for corruption, they are more likely to engage in corrupt acts, while if the public sector attracts workers with strong preferences for prosociality, they are less likely to engage in corrupt acts, even when there are strong financial incentives to do so. This points to the importance of non-​ monetary incentives in the decision to engage in corrupt practices. The public administration literature has highlighted the role of public service motivation (Perry, 1996) as an important driver of effort in the public sector (and relatedly of lower corrupt practices). Wilson (1989) highlights this issue in his seminal work on the American bureaucracy, stating that given the state of contract enforcement and weak monitoring in the public sector, “what is surprising is that bureaucrats work at all … rather than shirk at every opportunity” (p. 17). This literature focuses on differing worker motives and demonstrates that many employees (including those that join the public sector) are motivated by the outcomes of their actions rather than by monetary incentives (Prendergast, 2007). Another strand of the literature focuses on the role of intrinsic motivation and effort (Perry, 1996; Besley and Ghatak, 2005; Deci and Ryan, 2008, among many others). This literature finds that monetary incentives work well for workers with little intrinsic motivation but are inefficient for motivated workers (Frey and Oberholzer-​Gee, 1997; Deci and Ryan, 2008; Delfgaauw and Dur, 2008; Carpenter and Gong, 2016; Banuri and Keefer, 2016a). The intuition here is that while workers are influenced by extrinsic incentives (e.g., monetary incentives associated with corrupt acts), motivated workers are likely to exert effort even in the absence of extrinsic incentives, and hence offering these types of workers monetary incentives leads to inefficiencies. In other words, these worker types have intrinsic incentives that yield behaviour different from that predicted by extrinsic incentives. For the public sector, one area that has explicitly been highlighted is that of a mission match (Besley and Ghatak, 2005): the idea that workers derive utility from working for organizations with a shared mission. Papers on worker selection into the public sector concern themselves with two main (interrelated) worker preferences: prosocial/​public service motivation and preferences for dishonesty. These papers measure either prosociality (using dictator games with a charity recipient, such as Banuri and Keefer, 2016a) or dishonesty (using the die-​roll task, such as Barfort et al., 2019). While prosociality and dishonesty are theoretically distinct concepts, Banuri and Keefer (2016a)

The Culture-Corruption Hypothesis  159 show that those exhibiting greater prosociality (in Indonesia) work harder for a prosocial mission. Hanna and Wang (2017) show that public sector nurses (in India) with a greater preference for dishonesty are also more likely to engage in corrupt practices (absenteeism). Furthermore, Hanna and Wang (2017) also show that in a student sample in India, dishonest individuals are more likely to report wanting a job in the public sector, while prosocial individuals are less likely to report wanting a job in the public sector. Barfolt et al. (2019) find similar results using a sample in Denmark.Taken together, these studies point to the relationship between prosociality, dishonesty and selection into the public sector. Missions matter: Carpenter and Gong (2016) conduct an experiment with university students and manipulate (1) whether the work aligns with the students’ personal mission and (2) the incentive structure (unconditional or conditional pay). The authors measure political preferences (a proxy for the students’ personal mission) by asking them to declare their political affiliation. The experiment then asks students to generate mailers requesting funds on behalf of their preferred political party or the rival political party. Since the authors knew subjects’ political preferences ahead of time, they were able to distinguish between matched (aligned) and unmatched (not aligned) workers. The authors randomly assigned students to either a piece rate or a flat rate contract. They find that (consistent with theory) students whose political preferences matched the task are much more productive, and that the effect of extrinsic incentives (piece rate) is much stronger for mismatched workers than for matched workers. In other words, matched workers work harder, especially when the financial incentive to work is weak, consistent with other literature.7 Mechanisms exist to attract different types of workers, and worker pay is an important aspect. Hanna and Wang (2016) conduct an experiment on dishonesty and selection into the public sector in India using a sample of students and a sample of public sector nurses.The student sample was recruited from final year students at seven large universities in Bangalore. These students were recruited from majors that could join either the public or the private sector upon graduating. These students participated in a series of incentivized games designed to measure preferences for dishonesty using (1) a die-​roll task where subjects are asked to roll a die and report the outcome but are incentivized to report higher numbers and (2) a message task using a cheap-​talk sender-​receiver game, where subjects are incentivized to lie to maximize their own payoff. In addition, Hanna and Wang (2017) also measured prosocial preferences using donations to a set of well-​known charities.The public sector nurses’ sample was administered the die-​roll task. For the student sample, the authors were interested in the preferences of those that signalled intent to join the public sector. For the public sector nursing sample, the authors were interested in actual corrupt behaviour (measured by absenteeism: collecting paychecks for days not actually worked). The results show that those students reporting higher die rolls (more dishonest) were significantly more likely to want a government job. Those that donated more to the charity (more prosocial) were significantly less likely to

160  Sheheryar Banuri want a government job. Hence, in India (a country with high levels of perceived corruption), those that exhibited stronger preferences for dishonesty and lower prosocial preferences were more likely to prefer government jobs. Using the public sector nurses sample, they show that greater preferences for dishonesty predict absenteeism. In further work, Dhaliwal and Hanna (2017) show that low nurse attendance at work is related to lower birth weight of babies, highlighting the consequences of engaging in corrupt acts. In the context of India, a country where perceptions of corruption are high, those with strong preferences for dishonesty or weak preferences for prosociality are choosing to join public sector jobs, indicating some evidence in favour of a selection hypothesis: corrupt individuals are more attracted to public sector jobs in corrupt societies. Banerjee et al. (2015) report similar findings in India: they conduct an incentivized corruption game where subjects can overreport effort provided by matched workers to benefit themselves and can underpay matched workers to benefit themselves. The authors conduct this game with two subject pools –​private sector aspirants (students at the International Management Institute in Delhi, who are highly likely to join the private sector) and public sector aspirants (students who were preparing for the Union Public Service Commission Exam, a credible signal of joining public service).The authors find a significantly higher proportion of the public sector aspirants are more likely to overreport and to underpay, meaning that public sector aspirants were more likely to exhibit corrupt behaviour. Barfort et al. (2019) conduct a survey with students in Denmark (a low corruption context) undertaking majors that are equally likely to go on to work for the public or the private sector. The authors implement an online survey experiment and asked participants to play the incentivized die-​roll task. They also measure aspirations to join the public sector.8 Furthermore, they also measure prosocial preferences (dictator game with charity recipient) and the importance of wages in their choice of career. They report three main results: First, public service aspirants are less likely to engage in cheating behaviour. Second, this relationship is largely driven by prosocial preferences and (more tentatively) pecuniary compensation.This means that those that are more prosocial are both less likely to cheat and more likely to want a job in the public sector, and those that rank wages as an important consideration are more likely to cheat and less likely to want a public sector job. Third, they also report the results from a game asking participants to choose between two hypothetical jobs –​one in the private sector and another in the public sector. In the game they vary the pay differential between the two jobs and find that as the pay gap between the public and the private sector reduces, more dishonest individuals are attracted into the public sector. Hence, this study demonstrates the role of selection in sustaining low levels of corruption in Denmark: In a country with low perceptions of corruption (among the lowest in the world), those that are less dishonest and more prosocial are attracted into the public sector. The study also suggests that raising public sector wages will have the unintended effects of attracting less prosocial workers into the public sector.

The Culture-Corruption Hypothesis  161 Finally, Olsen et al. (2019) report the results of an online die-​roll task conducted with male9 public sector aspirants from ten countries that vary in the overall level of corruption perceptions: Denmark (ranked 1st in the Corruption Perceptions Index in 2019), Sweden (ranked 4th), Singapore (ranked 4th), Germany (ranked 9th), the United Kingdom (ranked 12th), Morocco (ranked 80th), Indonesia (ranked 85th),Thailand (ranked 101st), Egypt (ranked 106th) and Algeria (ranked 106th). Importantly, this survey focused on male-​only student participants, and only if they chose the public sector in response to a question about the sector they would want to work in the future. The authors find that participants in high corruption contexts are more dishonest than participants in low corruption contexts, correlating the results obtained in the game with measures of corruption such as the Corruption Perceptions Index and the Worldwide Governance Indicators dataset. These results are consistent with those reported by Gachter and Schulz (2016), discussed in the preceding section, who find that students from societies with high rule violations are also more dishonest. Overall, the papers conclude that corruption is (in part) a matter of selection: those cultures with high perceptions of corruption attract more dishonest/​ less prosocial workers, while those cultures with low perceptions of corruption attract more honest/​more prosocial workers. However, there are a few papers that do not follow the pattern above, indicating that (again) there is more to the corruption-​culture story. Kolstad and Lindkvist (2013) conduct an incentivized dictator game with student recipients with a sample of medical and nursing students in Dar es Salaam, Tanzania (a country and is high on corruption perceptions –​ranked 111th in 2013). Contrary to the studies above, students that displayed greater prosocial motivation were more likely to prefer working in public facilities. Similarly, Banuri and Keefer (2016a) report the results of a dictator game (with a charity recipient) conducted with accounting students from two institutions of higher education in Jakarta, Indonesia (a country with high corruption perceptions). One sample comes from the University of Indonesia, which constitutes a sample of general workers, while the other sample comes from the State College of Accountancy, where students study tuition free in exchange for serving the government of Indonesia upon graduation. Students in the latter have accepted entry into a school that constitutes an ironclad decision to work in the public sector but have not actually been socialized into the public sector.The authors find that those that chose the public sector university are more prosocial (donate more to charity). Alatas et al. (2009) conduct an incentivized bribery game with two subject pools in Jakarta, Indonesia –​students from the University of Indonesia and students from the State College of Administration (STIA), which contains mid-​career public servants. The authors find that public officials are less likely to engage in corrupt practices relative to student subjects. They also report that students aspiring to join the public sector show behaviour statistically indistinguishable from students wanting to join other sectors (Banuri and Keefer, 2016a, also report the same result).

162  Sheheryar Banuri Overall, the results from these papers are mixed vis-​à-​vis selection and corrupt behaviour. While Hanna and Wang (2017) and Banerjee et al. (2015) show that dishonest workers aspire to public sector jobs in a high corruption environment (India), Barfort et al. (2019) show that honest workers aspire to public sector jobs in a low corruption environment (Denmark), and Olsen et al. (2019) find that higher dishonest behaviour of male public sector career aspirants in ten countries correlates higher levels of corruption, other papers report the opposite (Kolstad and Lindkvist, 2013; Banuri and Keefer, 2016a; Cameron et al., 2009). This indicates the need for more research into the type of workers selecting into public sector jobs and whether the overall level of corruption perceptions impact the type of worker that aspires to join the public sector.

Do Public Sector Workers Become More/​Less Corrupt over Time? A related question is whether working in a corrupt public sector increases the likelihood of public workers engaging in corrupt acts over time.That is, what is the effect of socialization in public sector organizations on corrupt practices. In this section I highlight the role of socialization, which many papers consider to be the primary mode of transmission of corruption norms. A series of theoretical papers support the idea of the intergenerational transmission of preferences generally (Bisin and Verdier, 2001;Tabellini, 2008; Bulte and Horan, 2011; see Bisin and Verdier, 2011 for a review), and of corruption specifically (Andvig and Moene, 1990; Hauk and Saez-​Marti, 2002). However, empirical evidence on the transmission of preferences in the public sector is rare. Banuri and Keefer (2016b) present data using an incentivized dictator game conducted with mid-​career public officials in Indonesia (previously used by Alatas et al., 2009). The authors measure prosocial behaviour using a dictator game with a charity recipient. Students at the State College of Administration (STIA –​ Sekolah Tinggi Ilmu Administrasi) are mid-​career public officials who come from all over Indonesia to take evening classes in public administration, typically for the purposes of enhancing the probability of promotion within the public sector. Students from this institution were previously used by Alatas et al. (2009) in their study comparing behaviour with standard university samples. This sample is particularly unique in that it contains public officials at various stages of their career and hence contains subjects at various ranks, tenures and ages. Banuri and Keefer (2016b) report that those with greater time spent in the public sector are more prosocial than those with less time spent in the public sector. This result holds when controlling for rank, income, age, perceived effectiveness of the charity and a host of other factors including gender. This paper provides suggestive evidence in favour of socialization in the public sector: those that have spent more time in a prosocial organization become more prosocial. This is particularly remarkable given that the Indonesian public

The Culture-Corruption Hypothesis  163 sector is generally in the middle of the corruption distribution and hence would not be particularly prone to less prosociality over time.10 The results are consistent with the secondary socialization results reported by Barr and Serra (2010), with individuals from corrupt countries exhibiting lower levels of corrupt behaviour the more time they spent in a low corruption context (i.e., the United Kingdom). Nevertheless, this represents an area of fruitful research as there is little empirical evidence, and much of the data from Indonesia is inconsistent with data from other (similarly corrupt) contexts.

Changing the Culture of Corruption In a tangential literature on tax compliance (which shares many behavioural features with corruption, such as preferences for dishonesty), studies show that the treatment effects of experiments aimed at increasing compliance behaviour are mediated by previous compliance behaviour (Coleman, 1996; Slemrod, Blumenthal and Christian, 2001; Dwenger et al., 2016).11 This literature provides evidence in favour of the existence of agents with preference types, which mediates the effectiveness of interventions aimed at increasing compliance. What is interesting to note here is that this literature focuses on the general public but finds that those with strong preferences for honesty/​compliance are less likely to engage in evasion/​corrupt practices at baseline. Such individuals are subsequently less effected by interventions. Importantly, this literature coupled with the literature on motivation provides evidence in favour of worker selection as a way to impact the link between corruption and culture. For example, both Coleman (1996) and Slemrod, Blumenthal and Christian (2001) report positive effects of the increased threat of audits but find that these effects are concentrated among low and middle income taxpayers. For high income tax payers, the effects are either negative (Slemrod, Blumenthal and Christian, 2001) or mixed (Coleman, 1996). Dwenger et al. (2016) can unpack the heterogeneity further by focusing on the effects of (i) correcting perceptions of audit probabilities; (ii) increasing audit probabilities; and (iii) manipulating intrinsic incentives to comply (by focusing on norms, moral appeals or providing social recognition).The context was a binding local church tax in Germany, but the important feature in this study is that the authors can link tax payments to income data. In this way, they can precisely observe those that evade (pay less than their income levels warrant), those that comply (pay exactly what is owed) and those that donate (pay more than what is owed). Hence the authors can observe treatment effects on different types of (intrinsically motivated) agents. The authors report that compliers are unaffected by changes in deterrence, while the positive impacts are concentrated among evaders. The authors also document a strong bunching of exact compliance (people paying exactly what they owed), providing evidence for preferences for honesty driving behaviour. Finally, and perhaps most importantly, the impact of rewards (either extrinsic or intrinsic) occurs mainly through donors (those

164  Sheheryar Banuri contributing more than warranted), who increase their overpayments. Rewards also reduced the payments of evaders. The authors interpret this finding as rewards signalling the voluntary aspect of donations and downplay the mandatory aspect of donations.The key point here is that an explicit focus on rewards carries heterogeneous effects based on underlying preferences. Manipulating pecuniary incentives to attract certain types of workers to the public sector have been tested in a variety of contexts, though this is still an emerging literature. Dal Bo, Finan and Rossi (2013) present the results of a field experiment for a specific type of public sector worker in Mexico. The authors ask a simple question: does increasing (advertised) wages affect the quality of applicant (and thereby worker selection) into the public sector? Quality is measured in two ways –​ability (measured by previous earnings and IQ) and motivation (using Perry’s 1996 scale of public service motivation). The experiment randomized two sets of advertised wages (either 3,750 pesos or 5,000 pesos per month) across multiple locations.The authors find that the recruitment sites that advertised higher wages attracted more applicants, and that these applicants were of higher ability (as measured by either previous wages or IQ scores). In addition, the increase in wage did not reduce the overall levels of public service motivation in the subject pool (inconsistent with the results reported by Barfort et al., 2019). In fact, the results show an increase in the public service motivation index in the treatment sites relative to the control sites. The authors find that this increase in both ability and public service motivation is due to a positive correlation between ability and public service motivation. In their sample, those with higher ability also had a higher public service motivation index, hence the higher wage increasing both ability and motivation. Ashraf et al. (2020) conduct a field experiment in the recruitment drive for a new health worker cadre in Zambia. Their experiment manipulates either pecuniary or social incentives. The pecuniary incentive poster emphasized the following sentence: “Become a community health worker to gain skills and boost your career!” while the social incentives poster emphasized the following sentence: “Want to serve your community? Become a community health worker!” The authors measure ability (test scores), career ambition and prosociality. The authors find that emphasizing career incentives increases the average ability level in the applicant pool. Prosociality in the applicant pool drops overall but is only significantly lower in the lower ranked candidates, and no different in the higher ranked candidates. Overall, this means that the selected candidates are of higher ability and no different on prosociality, similar to the results reported by Dal Bo et al. (2013). By contrast, Deserranno (2019) conducts a field experiment recruiting Community Health Promoters in Uganda. These individuals are all female, recruited from microfinance groups managed by the implementing NGO (BRAC). The position is part time and has flexible hours. Unlike the other two experiments, however, while these positions are indeed prosocial, they are not public officials. The experiment manipulates the level of compensation by emphasizing the maximum (200,000 UGX per month), the average (30,000

The Culture-Corruption Hypothesis  165 UGX per month) or the minimum (7,000 UGX per month). The author measures prosocial motivation through previous experience of volunteering in the health sector, a survey measure of the importance of serving the community and an incentivized dictator game with an NGO recipient in line with previous literature. The results show that higher pay crowds out prosocial workers, such that the applicant pool is significantly more prosocial in the treatment emphasizing lower pay levels. The three field experiments discussed above present conflicting findings. Dal Bo et al. (2013) and Ashraf et al. (2020) report higher ability with no decline in prosociality in the overall applicant pool with higher wages, whereas Deserrano (2019) finds the opposite –​no reduction in ability, but a drop in prosociality. Banuri and Keefer (2016a) conduct a lab experiment using incentivized measures of both ability and prosociality. The subjects are students at the University of Indonesia who are asked to participate in a lab experiment for pay and select between two types of contracts –​a low wage public sector contract or a private sector contract. The treatment asks workers to choose between a high wage public sector contract or the same private sector contract discussed earlier. In this fashion the experiment simulates the sector choice in a lab setting. Importantly, and unlike previous studies, the measures of ability and prosocial motivation are completely independent and incentivized. The results show that the small increase in unconditional wages (small relative to large increases implemented in the field) have no impact on the ability of the worker selecting into the high wage public sector contract. However, the workers are significantly less prosocial, in line with the results from Deserrano (2019). One final point is whether interventions can affect behaviour of those that have already selected into the public sector. Harris et al. (2022) reports the results of an honesty intervention in Ghana, where police workers are randomly assigned to an ethics training programme designed to reduce corrupt practices. They find that police officers undergoing the programme are less likely to engage in dishonest behaviour and report a higher degree of honest values.12 These results suggest that interventions focusing on socialization can be effective in reducing corrupt behaviour, but more research is needed in this area.

Conclusions In this chapter, I reviewed the empirical (largely experimental) literature on the corruption-​culture hypothesis. The review points to interesting directions for future research. First, the issue of whether a culture of corruption exists in the wider population (beyond the public sector) is still an open one. Is it the case that random individuals from societies with high levels of corruption are more likely to have a preference for corruption? The papers reviewed have mixed answers, mainly due to the differences in sampling and methods applied. While it may well be true that public officials from more corrupt countries are more likely to engage in corrupt behaviour in a controlled setting, is this reflective of

166  Sheheryar Banuri the population, or are public officials a self-​selected group based on their own perceptions of corruption and their preferences for dishonesty/​prosociality? Furthermore, can workers be socialized away from engaging in corrupt practices? This is an open area of research, with very few papers. Data from Indonesia shows that (at least in one context) public officials are becoming more prosocial, and a specific targeted intervention in Accra was able to change behaviour among a specific type of public worker. Whether these types of interventions can be used to change behaviour in other types of public workers and in workers that are not in the public eye is another open question. Finally, there is some literature demonstrating that workers choosing to join the public sector have distinct preferences, and pecuniary incentives can affect the type of workers that selects into the public sector. The results from studies in Denmark, Zambia and Indonesia suggest that increasing wages is not the answer, but other papers report differential impacts. More research is needed to reconcile these findings. The literature on prosocial motivation and effort further highlights the importance of baseline motives and incentives. The focus here is primarily on workers in the mission-​based sector (i.e., public sector or non-​governmental organizations). Prosocial/​public sector motivation is critical (Besley and Ghatak, 2005; Carpenter and Gong, 2016; Banuri and Keefer, 2016a), where individuals are motivated primarily by the mission of the organization, rather than by extrinsic incentives. Overall, the literature demonstrates the importance of worker selection. Individual preferences differ in terms of truth-​telling (Gneezy, Kajackaite and Sobel, 2018; Abeler, Nosenzo and Raymond, 2019).The underlying mechanism has been argued to be the presence of moral costs (Abbink, 2000; Cox et al., 2016), in that those with higher levels of moral costs are likely to act ethically. Similarly, the literature on prosocial motivation (Besley and Ghatak, 2005; Carpenter and Gong, 2016; Banuri and Keefer, 2016a) find that people that are motivated to work for the public good are more likely to exert high effort even in the presence of weak incentives. Thus, emphasizing the prosocial nature of avoiding corrupt practices is likely to have an effect, but on a particular type of worker. Attracting motivated workers into the public sector is an important policy tool. Worker pay and emphasis on missions both play key roles, and the core lesson here is that public sector managers should aim to attract and retain workers with strong prosocial/​honest preferences.

Notes 1 See Jain (2001) for an extensive discussion on the definition of corruption. 2 See Fernández (2011) for an extensive discussion on the definitions of culture. 3 See Abbink et al. (2002) and Cameron et al. (2009) for a discussion of the impact of social costs on bribing behaviour. 4 The authors further test for a socialization effect by controlling for the number of years spent in the UK, reasoning that graduate students may have spent more years

The Culture-Corruption Hypothesis  167 in the UK and potentially been socialized into engaging in less corrupt practices. They find little evidence of this, however. While they do find that those who have spent more time in the UK are less likely to engage in corrupt practices (providing some evidence for secondary socialization), the difference in behaviour between graduate and undergraduate students remains unexplained. 5 Yet another experiment on dishonest behaviour across cultures is conducted by Cohn et al. (2019), who hand “lost” wallets to front desk employees (in public and private institutions) in 40 countries and measure the likelihood of being contacted by the employees. The authors do not analyse the relationship with culture directly, though they do report a higher likelihood of return in cultures with a higher share of protestants (p. 62 in supplement). Another set of literature that I do not cover (Treisman, 2000; Brunetti and Weder, 2003; Chowdhury and Shyamal, 2004; Gokcekus and Knorich, 2006; Gokcekus, 2008), show that corruption perceptions are lower in countries with a higher share of protestants, attributed to the religious traditions of protestants in denouncing abuses by public officials. 6 Note, however, that this behaviour (dishonesty) is similar to, but not the same as corrupt behaviour, though the two are correlated. Hanna and Wang (2017) demonstrate the relationship between dishonesty and corrupt behaviour in India (covered later in the chapter). 7 Other papers report similar results (Frey and Oberholzer-​Gee, 1997; Deci, Koestner and Ryan, 1999; Besley and Ghatak, 2005; Delfgaauw and Dur, 2008; Banuri and Keefer, 2016a; Jones, Tonin and Vlassopoulos, 2018; Tonin and Vlassopoulos, 2010; Fehrler and Kosfeld, 2014; Gerhards, 2015; DellaVigna and Pope, 2018; among others), that providing individuals with a (prosocial) mission increases effort. 8 They ask subjects to rank eight viable jobs in order of attractiveness, and elicit the probability with which they are to work in each job. Their measure of aspiring to work in the public sector is if the public sector position ranked amongst the top two positions. 9 The sample of male public sector aspirants is notable. The literature on gender and corruption reports (unambiguously) that women are less likely to engage in corrupt acts than men (Swamy et al. 2001; Alatas at al. 2009; Frank et al. 2011; Rivas, 2013). While one interpretation of these results is due to women being more prosocial than men (see Andreoni and Vesterlund, 2001; Dollar et al. 2001), however Frank et al. (2011) report that other factors associated with gender (such as avoiding the risk of detection, or less trusting of the partner in a corrupt act) might be more likely to drive such behaviour. 10 These results may be consistent with a moral cleansing effect reported by Sachdeva et al. (2009) and Branas-​Garza et al. (2013), though the authors do not test this hypothesis. The effect suggests that those that engage in less prosocial acts (such as corruption) may exhibit higher prosocial behaviour in a bid to “cleanse” themselves of previous transgressions. 11 See Barbuta-​Miu (2011) and Marandu et al. (2015) for useful reviews on the determinants of tax compliance. 12 It is not immediately clear that ethics training programs are effective, however. For example, Banerjee and Mitra (2018) report the results of a lab experiment using an ethics training intervention for MBA students in India. They find that while the proportion of subjects demanding bribes reduces in the immediate aftermath of the intervention, these effects do not persist after 4 weeks, indicating that these interventions might be short-​lived.

168  Sheheryar Banuri

References Abbink, Klaus. Fair salaries and the moral costs of corruption. Bonn Econ Discussion Paper No. 1/​2000, 2000. Abbink, Klaus, Bernd Irlenbusch and Elke Renner. “An experimental bribery game.” Journal of Law, Economics, and Organization 18, no. 2 (2002): 428–​454. Abeler, Johannes, Daniele Nosenzo and Collin Raymond. “Preferences for truth-​ telling.” Econometrica 87, no. 4 (2019): 1115–​1153. Alatas, Vivi, Lisa Cameron, Ananish Chaudhuri, Nisvan Erkal and Lata Gangadharan. “Subject pool effects in a corruption experiment: A comparison of Indonesian public servants and Indonesian students.” Experimental Economics 12, no. 1 (2009): 113–​132. Andreoni, James, and Lise Vesterlund. “Which is the fair sex? Gender differences in altruism.” Quarterly Journal of Economics 116, no. 1 (2001): 293–​312. Andvig, Jens Chr, and Karl Ove Moene. “How corruption may corrupt.” Journal of Economic Behavior & Organization 13, no. 1 (1990): 63–​76. Armantier, Olivier, and Amadou Boly. “A controlled field experiment on corruption.” European Economic Review 55, no. 8 (2011): 1072–​1082. Armantier, Olivier, and Amadou Boly. “Comparing corruption in the laboratory and in the field in Burkina Faso and in Canada.” Economic Journal 123, no. 573 (2013): 1168–​1187. Ashraf, Nava, Oriana Bandiera, Edward Davenport and Scott S. Lee. “Losing prosociality in the quest for talent? Sorting, selection, and productivity in the delivery of public services.” American Economic Review 110, no. 5 (2020): 1355–​1394. Banerjee, Ritwik, and Arnab Mitra. "On monetary and non-​monetary interventions to combat corruption." Journal of economic behavior & organization 149 (2018): 332–​355. Banerjee, Ritwik, Tushi Baul and Tanya Rosenblat. “On self-​selection of the corrupt into the public sector.” Economics Letters 127 (2015): 43–​46. Banuri, Sheheryar, and Catherine Eckel. “Cracking down on bribery.” Social Choice and Welfare 45, no. 3 (2015): 579–​600. Banuri, Sheheryar, and Philip Keefer.“Pro-​social motivation, effort and the call to public service.” European Economic Review 83 (2016a): 139–​164. Banuri, Sheheryar, and Philip Keefer. “Mellowing with tenure? Socialization increases prosocial behavior in public organizations.” In Experiments in organizational economics, pp. 127–​140. Emerald Publishing, 2016b. Barbuta-​Misu, N. A review of factors for tax compliance. Economics and Applied Informatics 1 (2011): 69–​76 Barfort, Sebastian, Nikolaj A. Harmon, Frederik Hjorth and Asmus Leth Olsen. “Sustaining honesty in public service: The role of selection.” American Economic Journal: Economic Policy 11, no. 4 (2019): 96–​123. Barr, Abigail, and Danila Serra. “Corruption and culture: An experimental analysis.” Journal of Public Economics 94, no. 11–​12 (2010): 862–​869. Besley, Timothy, and Maitreesh Ghatak. “Competition and incentives with motivated agents.” American Economic Review 95, no. 3 (2005): 616–​636. Bisin, Alberto, and Thierry Verdier. “The economics of cultural transmission and the dynamics of preferences.” Journal of Economic Theory 97, no. 2 (2001): 298–​319. Bisin, Alberto, and Thierry Verdier. “The economics of cultural transmission and socialization.” In Handbook of social economics, vol. 1, pp. 339–​416. North-​Holland, 2011.

The Culture-Corruption Hypothesis  169 Brañas-​Garza, Pablo, Marisa Bucheli, María Paz Espinosa and Teresa García-​Muñoz. “Moral cleansing and moral licenses: Experimental evidence.” Economics & Philosophy 29, no. 2 (2013): 199–​212. Brunetti, Aymo, and Beatrice Weder. “A free press is bad news for corruption.” Journal of Public Economics 87, no. 7–​8 (2003): 1801–​1824. Bulte, Erwin H., and Richard D. Horan. “Intergenerational transmission of preferences.” Economics Letters 112, no. 1 (2011): 85–​87. Cameron, Lisa,Ananish Chaudhuri, Nisvan Erkal and Lata Gangadharan.“Propensities to engage in and punish corrupt behavior: Experimental evidence from Australia, India, Indonesia and Singapore.” Journal of Public Economics 93, no. 7–​8 (2009): 843–​851. Carpenter, Jeffrey, and Erick Gong. “Motivating agents: How much does the mission matter?” Journal of Labor Economics 34, no. 1 (2016): 211–​236. Chowdhury, Shyamal K. “The effect of democracy and press freedom on corruption: An empirical test.” Economics Letters 85, no. 1 (2004): 93–​101. Coleman, Stephen. “The Minnesota income tax compliance experiment: State tax results.” (1996). www.reve​nue.state.mn.us/​sites/​defa​ult/​files/​2011-​11/​resear​ch_​r​ epor​ts_​c​onte​nt_​c​ompl​nce.pdf Cohn, Alain, Michel André Maréchal, David Tannenbaum, and Christian Lukas Zünd. "Civic honesty around the globe." Science 365, no. 6448 (2019): 70–​73. Cox, James C., John A. List, Michael Price, Vjollca Sadiraj and Anya Samek. Moral costs and rational choice: Theory and experimental evidence. National Bureau of Economic Research, 2016. Dal Bó, Ernesto, Frederico Finan and Martín A. Rossi. “Strengthening state capabilities: The role of financial incentives in the call to public service.” Quarterly Journal of Economics 128, no. 3 (2013): 1169–​1218. Deci, Edward L., and Richard M. Ryan. “Self-​determination theory: A macrotheory of human motivation, development, and health.” Canadian Psychology/​Psychologie canadienne 49, no. 3 (2008): 182. Deci, Edward L., Richard Koestner and Richard M. Ryan. “A meta-​analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation.” Psychological Bulletin 125, no. 6 (1999): 627. Delfgaauw, Josse, and Robert Dur. “Incentives and workers’ motivation in the public sector.” Economic Journal 118, no. 525 (2008): 171–​191. DellaVigna, Stefano, and Devin Pope. “What motivates effort? Evidence and expert forecasts.” Review of Economic Studies 85, no. 2 (2018): 1029–​1069. Deserranno, Erika. “Financial incentives as signals: Experimental evidence from the recruitment of village promoters in Uganda.” American Economic Journal: Applied Economics 11, no. 1 (2019): 277–​317. Desmet, Klaus, Ignacio Ortuño-​Ortín and Romain Wacziarg. “Culture, ethnicity, and diversity.” American Economic Review 107, no. 9 (2017): 2479–​2513. Dhaliwal, Iqbal, and Rema Hanna. “The devil is in the details: The successes and limitations of bureaucratic reform in India.” Journal of Development Economics 124 (2017): 1–​21. Dollar, David, Raymond Fisman and Roberta Gatti. “Are women really the ‘fairer’ sex? Corruption and women in government.” Journal of Economic Behavior & Organization 46, no. 4 (2001): 423–​429. Dwenger, Nadja, Henrik Kleven, Imran Rasul and Johannes Rincke. “Extrinsic and intrinsic motivations for tax compliance: Evidence from a field experiment in Germany.” American Economic Journal: Economic Policy 8, no. 3 (2016): 203–​232.

170  Sheheryar Banuri Fehrler, Sebastian, and Michael Kosfeld. “Pro-​ social missions and worker motivation: An experimental study.” Journal of Economic Behavior & Organization 100 (2014): 99–​110. Fernández, Raquel. "Does culture matter?." Handbook of social economics 1 (2011): 481–​510. Fisman, Raymond, and Edward Miguel. “Corruption, norms, and legal enforcement: Evidence from diplomatic parking tickets.” Journal of Political Economy 115, no. 6 (2007): 1020–​1048. Frank, Björn, Johann Graf Lambsdorff and Frédéric Boehm. “Gender and corruption: Lessons from laboratory corruption experiments.” European Journal of Development Research 23, no. 1 (2011): 59–​71. Frey, Bruno S., and Felix Oberholzer-​Gee. “The cost of price incentives: An empirical analysis of motivation crowding-​out.” American Economic Review 87, no. 4 (1997): 746–​755. Gächter, Simon, and Jonathan F. Schulz. “Intrinsic honesty and the prevalence of rule violations across societies.” Nature 531, no. 7595 (2016): 496–​499. Gerhards, Leonie. “The incentive effects of missions: Evidence from experiments with NGO employees and students.” European Economic Review 79 (2015): 252–​262. Gokcekus, Omer. “Is it protestant tradition or current protestant population that affects corruption?” Economics Letters 99, no. 1 (2008): 59–​62. Gokcekus, Omer, and Jan Knörich. “Does quality of openness affect corruption?” Economics Letters 91, no. 2 (2006): 190–​196. Gneezy, Uri, Agne Kajackaite and Joel Sobel. “Lying aversion and the size of the lie.” American Economic Review 108, no. 2 (2018): 419–​453. Hanna, Rema, and Shing-​ Yi Wang. “Dishonesty and selection into public service: Evidence from India.” American Economic Journal: Economic Policy 9, no. 3 (2017): 262–​290. Harris Donna, Borcan Oana, Serra Danila, Telli Henry, Schettini Bruno and Dercon Stefan. “Proud to belong: The impact of ethics training on police officers in Ghana.” CSAE Working Paper, June 2022. Hauk, Esther, and Maria Saez-​Marti. “On the cultural transmission of corruption.” Journal of Economic Theory 107, no. 2 (2002): 311–​335. Jain, Arvind K., ed. The political economy of corruption, vol. 2. Routledge, 2001. Jones, Daniel, Mirco Tonin and Michael Vlassopoulos. “Paying for what kind of performance? Performance pay and multitasking in mission-​oriented jobs.” July 18, 2018. IZA DP No. 11674. Kaufmann, Daniel, Aart Kraay and Massimo Mastruzzi. “Governance matters IV: Governance indicators for 1996–​2004.” World Bank Policy Research Working Paper 3630, 2005. Kolstad, Julie Riise, and Ida Lindkvist. “Pro-​social preferences and self-​selection into the public health sector: Evidence from an economic experiment.” Health Policy and Planning 28, no. 3 (2013): 320–​327. Knack, Stephen, and Philip Keefer. “Institutions and economic performance: Cross-​ country tests using alternative institutional measures.” Economics & Politics 7, no. 3 (1995): 207–​227. Lambsdorff, Johann Graf. “Causes and consequences of corruption: What do we know from a cross-​section of countries.” International Handbook on the Economics of Corruption 1 (2006): 3–​51. Marandu, Edward E., Christian J. Mbekomize and Alexander N. Ifezue. “Determinants of tax compliance: A review of factors and conceptualizations.” International Journal of Economics and Finance 7, no. 9 (2015): 207–​218.

The Culture-Corruption Hypothesis  171 Mauro, Paolo. “Corruption and growth.” Quarterly Journal of Economics 110, no. 3 (1995): 681–​712. Olsen,Asmus Leth, Frederik Hjorth, Nikolaj Harmon and Sebastian Barfort.“Behavioral dishonesty in the public sector.” Journal of Public Administration Research and Theory 29, no. 4 (2019): 572–​590. Perry, James L. “Measuring public service motivation: An assessment of construct reliability and validity.” Journal of Public Administration Research and Theory 6, no. 1 (1996): 5–​22. Prendergast, Canice. “The motivation and bias of bureaucrats.” American Economic Review 97, no. 1 (2007): 180–​196. Rauch, James E., and Peter B. Evans. “Bureaucratic structure and bureaucratic performance in less developed countries.” Journal of Public Economics 75, no. 1 (2000): 49–​71. Rivas, M. Fernanda. “An experiment on corruption and gender.” Bulletin of Economic Research 65, no. 1 (2013): 10–​42. Sachdeva, S., Iliev, R. and Medin, D. L. 2009. Sinning saints and saintly sinners. The paradox of moral self-​regulation. Psychological Science 20: 523–​528. Slemrod, Joel, Marsha Blumenthal and Charles Christian. “Taxpayer response to an increased probability of audit: Evidence from a controlled experiment in Minnesota.” Journal of Public Economics 79, no. 3 (2001): 455–​483. Swamy, Anand, Stephen Knack, Young Lee and Omar Azfar. “Gender and corruption.” Journal of Development Economics 64, no. 1 (2001): 25–​55. Tabellini, Guido. “Institutions and culture.” Journal of the European Economic Association 6, no. 2–​3 (2008): 255–​294. Tonin, Mirco, and Michael Vlassopoulos. “Disentangling the sources of pro-​socially motivated effort: A field experiment.” Journal of Public Economics 94, no. 11–​12 (2010): 1086–​1092. Treisman, Daniel. “The causes of corruption: A cross-​national study.” Journal of Public Economics 76, no. 3 (2000): 399–​457. Treisman, Daniel. “What have we learned about the causes of corruption from ten years of cross-​national empirical research?” Annual Review of Political Science 10, no. 1 (2007): 211–​244. Wilson, J. Q., ed. Bureaucracy: What government agencies do and why they do it, 1989. Basic Books.

12 How Advances in Information and Communication Technologies Impact Corruption? Chandan Kumar Jha and Sudipta Sarangi

Introduction Over the last two decades, information and communication technologies (ICTs) have made rapid progress and become embedded in every sphere of our lives. Therefore, it is not surprising that various studies have investigated whether ICTs can play a role in fighting corruption. Although corruption can be defined in many ways, the most common definition describes it as the abuse of power entrusted by the state for private gain. By facilitating the spread of information in different ways, various types of ICTs can impact corruption through different mechanisms. First, opportunities for corrupt practices in many instances arise due to the lack of information, and ICTs can play an important role in providing access to information and promoting transparency regarding government services and projects. Second, corrupt activities often involve an interaction between a government official and a citizen seeking a government service that provides the opportunity for demanding a bribe. ICTs allow for the use of electronic government (e-​government) procedures that can reduce or even eliminate the need for a direct interaction between the official and the citizen, reducing rent-​seeking opportunities. Third, many forms of ICTs, such as the internet and social media, can be used to create awareness about the prevalence of corruption in the country and mobilize support to fight corruption.1 In this chapter, we discuss how advances in ICTs can help reduce corruption and review the empirical evidence on the relationship between various forms of ICTs and corruption. As one can anticipate, studies investigating the effects of ICTs on corruption need to overcome many challenges to make robust inferences.The first and foremost issue is the availability of corruption data that is comparable over time and space. Since corruption by definition is an illegal activity, obtaining data on actual corruption is very difficult.2 Cross-​country studies typically use one or more of the following three perception-​based corruption indices: the Control of Corruption Index (CCI) by the World Bank’s World Governance Indicators (WGI), the International Country Risk Guide (ICRG) Corruption Index by the PRS Group and Corruption Perception DOI: 10.4324/9781003142300-12

How Advances in ICTs Impact Corruption?  173 Index (CPI) by the leading nongovernment anticorruption organization Transparency International.3 It is important to note that these perception-​based corruption measures are quite informative and useful for at least two important reasons. First, perception-​based corruption measures have been reported to be highly correlated with experience-​based measures of corruption (Mocan, 2008). Second, and more importantly, many have argued that perception-​based measures of corruption are important because individual and business decisions are influenced by the perception of corruption (Donchev and Ujhelyi, 2014; Kaufmann, Kraay and Mastruzzi, 2011), and therefore it has a direct impact on economic activity. The second issue, which is particularly problematic for cross-​ country comparisons, is the lack of variability in the corruption index over time. Corruption tends to be persistent even when anticorruption measures and incentives are in place (Mishra, 2006). It is therefore not surprising that corruption indices in general have little variation over time, making it difficult to carry out a panel analysis because one cannot introduce country dummies in the model. Therefore, researchers often have to rely on cross-​sectional analyses. Cross-​country (cross-​sectional) analyses are, however, subject to bias arising out of both unobserved country-​specific fixed and time-​variant factors, which are difficult, if not impossible, to rule out. A few studies, such as Andersen (2009), have utilized the first-​difference approach to minimize the bias arising out of omitted country-​specific fixed factors. By utilizing the difference between the (explained and explanatory) variables over two relatively distant points in time, the first-​difference approach eliminates the concerns that the estimates could be biased due to the omission of country-​specific time-​invariant factors. This approach, however, does not rule out the possibility that the estimates could be biased due to the omission of country-​specific time-​variant factors. Despite small variations in corruption indices, many studies (e.g., Elbahnasawy, 2014; Lio, Liu and Ou, 2011) have been able to use panel data methods to investigate the link between various forms of ICTs and corruption. A limited number of studies have thus been able to control for time-​variant observed factors such as the quality of institutions and economic development that might confound the relationship between ICTs and corruption. Another issue is potential endogeneity due to reverse causality as corruption might impact the prevalence of ICTs through various channels including infrastructure investments and content censorship. Corruption in a country might impact investment in infrastructure (Collier and Hoeffler, 2005), thereby affecting the advancement of and access to the ICT itself. Moreover, corrupt regimes have a strong incentive to control the use of ICTs or curb freedom on the net that might expose their wrongdoings. Evidence suggests that governments in many countries have used a variety of ways to control access to the internet and curb freedom on the internet using vague and flexible security laws (Knutsen, 2015; Wagner, 2012; Freedom House, 2009b). To address endogeneity concerns arising out of omitted variable bias and simultaneity, most studies have employed one of two econometric methods in

174  Chandan Kumar Jha and Sudipta Sarangi the ICT corruption literature. The first is instrumental variable (IV) analysis where one needs to find an exogenous instrument that is highly correlated with the explanatory variable of interest but does not affect the outcome variable except through its effect on the variable being instrumented. The other method is the use of dynamic panel data models such as the System Generalized Method of Moments (SGMM) that uses lags of the explanatory variables as their instruments. Studies such as Jha and Sarangi (2017), Andersen et al. (2011) and Martins,Veiga and Fernandes (2021) have used an IV analysis, while others such as Kanyam, Kostandini and Ferreira (2017), Elbahnasawy (2014) and Lio, Liu and Ou (2011) have employed dynamic panel data models to address endogeneity concerns. Finally, the effects of ICTs on corruption might be contingent on other factors that could amplify or weaken the effects of ICTs on corruption. These include factors like the quality of institutions, the strength of the civil society and cultural norms. Advances in ICTs might also be complementary to other factors such as the effects of press freedom that significantly impact corruption (Jha and Sarangi, 2017). This chapter reviews the empirical ICT corruption literature and discusses ways in which various forms of ICTs can help fight corruption.4 We describe measures of various forms of ICTs, their data sources and outline the key methodologies utilized by different studies to address endogeneity issues and to take into account the conditionality and complementarity between ICTs and other factors in investigating the association between ICTs and corruption. It is worth keeping in mind that many forms of ICTs are interrelated. For instance, one must have access to the internet to use social media or avail e-​government services. For the sake of clarity, we have broadly categorized our discussion in following three sections. In the next section, we discuss the role e-​government can play in curbing corruption followed by a section on the role of the internet and mobile phones in reducing corruption. The last two sections review the recent literature that documents a significant association between the most recent advances in ICTs, that is, social media and corruption, and offer discussion and concluding remarks, respectively.

E-​Government and Corruption E-​government refers to the use of ICTs by governments for providing different types of government services to beneficiaries that could be individuals, citizens or businesses. One of the most commonly used measures of e-​governance, known as the E-​Government Development Index (EGDI), comes from the Global E-​Government Readiness Reports compiled by the UN.5 The EGDI is designed to measure the willingness and capacity of governments to use online and mobile technology for executing government functions and providing government services to the citizens (United Nations, 2010). It is a weighted average of three important dimensions of e-​government.

How Advances in ICTs Impact Corruption?  175 • • •

The Online Service Index reflects the “scope and quality of online services.” The Telecommunication and Infrastructure Index is based on the ownership and use of personal computers, internet and telephones and the subscriptions of mobile phones and fixed broadband. Finally, the Human Capital Index consists of adult literacy rates and school gross enrollment ratios and reflects the capacity of the citizens to utilize online government services.

Many studies have also used individual components of the EGDI as a measure of e-​government. Another widely used measure of e-​government is the index compiled by Darrell M.West of Brown University (West, 2007).6 The computation of this index takes into account the availability of information, the delivery of services and public access to government websites (see West, 2007, for the complete methodology). Finally, a few studies have constructed a variable known as “e-​government maturity” using the Online Service Index that measures the quality and maturity of online services provided by the government (Arayankalam, Khan and Krishnan, 2021). The extent of the provision of government services through e-​government covers the entire spectrum: While certain services may be provided in their entirety, for others only information regarding their provision might be available through ICTs. In either case though, the use of e-​government can play an important role in promoting transparency and accountability. The easy availability of information outlining the process for availing the government services and required documents and fees not only increases transparency but also ensures accountability of the government officials responsible for providing these services. In other words, empowered by this knowledge, the beneficiaries can hold government officials accountable and refuse to accept delayed or deferred services and/​or pay bribes. For instance, if the required documents for availing a particular service are clearly listed on a government website, then it is very difficult for an official to ask for a bribe citing the inadequacy of the documents provided. Essentially, the availability of information regarding the process of the provision of the government services and the time within which these services should be provided reduces a government official’s discretionary powers that are often used to demand bribes. Next, for certain government services, e-​government can be used to provide services directly to the beneficiaries either in part or in their entirety. This reduces or eliminates the need for interaction between the beneficiary and intermediary rent-​seeking government officials, thereby reducing the opportunities for bribes. The existing literature provides extensive evidence on the association between e-​government and corruption using within-​country case studies as well as cross-​country analysis. First, at a conceptual level, Bhatnagar (2003) argues that the delivery of government services electronically requires rules and procedures to be standardized across regions so that they can be coded into a computer system.This in itself reduces the public officials’ discretion and hence

176  Chandan Kumar Jha and Sudipta Sarangi opportunities for arbitrary actions. Additionally, there is a greater likelihood that a wrongdoing will be discovered. As a result, e-​government acts as a deterrent to corruption. Researchers have analyzed examples from different parts of the world to assess whether specific e-​government projects have led to a decline in corruption. Below we review some of these interesting case studies. One of the most impressive examples of the use of e-​government that has reportedly saved 6.7 million farmers approximately 1.32 million days in waiting times and over Rs. 806 million (roughly 18 million US dollars at the then exchange rate of Rs. 45 per USD) in bribes is the Bhoomi (land in English) project from the Indian state of Karnataka (World Bank, 2004; Bhatnagar, 2003).7 The Bhoomi project led to a decline in corruption by allowing ordinary citizens to obtain a copy of the “Record of Rights, Tenancy and Crops (RTC)” using computer kiosks and limiting the bribery opportunities available to government officials responsible for maintaining the land records. Another example of such citizen empowerment is provided in the study by Kim, Kim and Lee (2009) who examine the effects of OPEN (Online Procedures ENhancement for civil application), an anticorruption system in the Seoul metropolitan government in South Korea, on corruption. OPEN provides information on administrative procedures and allows citizens to apply for public services in various areas such as urban planning, sanitation and housing and construction. After the introduction of OPEN, citizens can not only apply for qualified services but also check the status of their applications online rather than having to physically meet the officer-​in-​charge. Thus, OPEN eliminates the need for citizens to interact with the administrative officials. Moreover, the OPEN standardizes many of the processes, increasing the objectivity of the officials while reducing their discretion. Consequently, transparency increased and corruption declined in Seoul municipality after the introduction of OPEN (Kim, Kim and Lee, 2009; Bhatnagar, 2003).8 Several cross-​country studies have investigated whether the negative association between e-​government and corruption holds in a cross-​country analysis. The findings of cross-​country studies echo the findings of the micro-​level case studies documenting a significant, negative association between e-​government and corruption. For instance, Andersen (2009) examines the relationship between e-​government and corruption using data from 1996 to 2006 for 149 countries across the world. Using the CCI as a measure of corruption and the 2006 e-​government index from West (2007), he finds that an improvement in the e-​government index from the tenth to the ninetieth percentile is associated with an improvement in the corruption index from the tenth to the twenty-​ third percentile. Elbahnasawy (2014) uses CPI as the measure of corruption and EGDI as the measure of e-​government to find that e-​government and corruption are negatively correlated in an analysis covering 160 countries from 1995 to 2009. Starke, Naab and Scherer (2016) find that greater online service delivery (the measure of e-​government) is negatively associated with corruption in 157 countries in 2013. Both Elbahnasawy (2014) and Starke, Naab and Scherer (2016) find that the effect of e-​government on corruption

How Advances in ICTs Impact Corruption?  177 gets stronger with internet expansion. Park and Kim (2020) use panel data for over 200 countries and find that the Online Service Index and E-​Participation Index from the UN e-​government survey (United Nations, 2010) are negatively associated with corruption measured by the CCI. Srivastava, Teo and Devaraj (2016) explore the relationship between e-​government and corruption in three different national institutions (political, legal and media institutions) and national stakeholder service systems (business and citizen systems) data for 63 countries across the world from 2004 to 2007. They find that the level of e-​government development in a country is negatively and significantly associated with the level of corruption in all three institutions. Martins, Veiga and Fernandes (2021) use data for over 170 countries for the period 2002–​2017 and an IV analysis to show that greater e-​government usage, digital public services for business and online service completion are negatively and significantly associated with the level of corruption. Some recent studies have examined whether the effectiveness of e-​ government on corruption varies across countries based on a third factor. For instance, Arayankalam, Khan and Krishnan (2021) explore the role of government administrative effectiveness in influencing the relationship between e-​government maturity and corruption. The authors find that e-​government maturity is positively associated with government administrative effectiveness measured by the rule of law from the World Justice Project (WJP),9 which in turn is negatively associated with legislative, executive and judicial corruption. Further, the size of the Virtual Social Network (VSN) is found to moderate the relationships between e-​government maturity and government administrative effectiveness, and government administrative effectiveness and corruption.10 Nam (2018) uses the Online Service Index as a proxy for e-​government (United Nations, 2010) and finds the index to be negatively associated with corruption. The authors further find that the effects of e-​government on corruption is weakened in countries that are characterized by cultures of high power distance and high uncertainty avoidance.11 The evidence regarding the benefits of e-​government on corruption reduction is overwhelming. Case studies from several countries in the world present significant evidence that e-​government enhances transparency, encourages accountability and reduces discretionary powers, leading to a significant decline in corruption. The external validity of these finding has been confirmed by cross-​country studies that exploit the variations in the intensity of e-​government across countries and time. Finally, the evidence suggests that the effectiveness of e-​government on corruption can be enhanced by factors such as government effectiveness and weakened by cultural traits characterized by greater power distance and uncertainty avoidance. These findings suggest that the policymakers should pay close attention to cultural and institutional factors while designing e-​government services to maximize its benefits for corruption reduction. Studies focusing on how cultural and institutional factors, however, are both in their scope and numbers. Hence, there is a need for further research to identify such factors that affect the

178  Chandan Kumar Jha and Sudipta Sarangi efficacy of e-​government on corruption. Interestingly, a recent study by Martins, Veiga and Fernandes (2021) finds that e-​government has greater effects on corruption in low-​income countries, suggesting that even with lower resources at their disposal, providing services through e-​government might be worth it.

Mobile Phones, Internet and Corruption While e-​governance has had an impact on corruption, it is not surprising that by having revolutionized the provision of and access to information, mobile phones and the internet also play an important role in the reduction of corruption by removing information asymmetries. Researchers have utilized the proportion of the population in a country using cellphones and access to the internet to investigate whether the ability of the citizens to communicate and exchange information matters for corruption. Mobile phones and the internet provide cheap and easy access to information. Moreover, they can deliver information to previously inaccessible areas. In fact today it is used by traditional media such as the print and television media as well to communicate with a larger audience at the latter’s convenience. Most traditional media platforms, therefore, today have a presence on the internet where people can view the content they desire whenever they want (though sometimes this may require a paid subscription). Besides empowering citizens, a greater availability of, and easy access to, information contributes to establishing more open and free societies, which, in turn, lead to lower levels of corruption. Finally, a greater access to information promotes transparency in the provision of public services and, hence, promotes accountability of government officials. In addition to facilitating the exchange of information, access to the internet and mobile technology also allows citizens to report instances of corruption to authorities.12 Thus mobile phones and the internet add a new dimension that is not available with e-​ governance. They provide a means to monitor the activities of government officials and report on their activities. For instance, mobile phones can be used to record conversations with officials demanding bribes and other instances of corruption. Such recordings can then be shared using messaging services, the internet and social media. Consequently, it is argued that the expansion of mobile and internet services should result in a decline in corruption. In one of the earliest studies, DiRienzo et al. (2007) use the Digital Access Index (DAI) provided by the International Telecommunication Union (ITU) to study the effects of access to information on corruption. The DAI is computed on the basis of the availability of ICT infrastructure, quality of ICT services, affordability of access, the users’ ability to access and process the information and ICT usage.13 Motivated by earlier studies that document the effects of transparency and access to information on corruption, the authors hypothesize a negative association between the DAI and corruption. In a cross-​country analysis of 85 countries, the authors find that DAI is negatively associated with corruption measured by CPI. Sassi and Ali (2017) focus on African countries for two reasons: the region is characterized by very high levels of corruption,

How Advances in ICTs Impact Corruption?  179 and traditional anticorruption policies have been ineffective in the region (Transparency International, 2015). In a study covering 47 African countries, using the SGMM analysis, Sassi and Ali (2017) find that the expansion of both the internet and mobile phones is negatively and significantly associated with corruption in Africa. They also find that the effects of ICT diffusion on corruption is greater in the presence of better law enforcement measured by the rule-​of-​law index. Kanyam, Kostandini and Ferreira (2017) have similar findings: mobile phone penetration is negatively associated with corruption in sample of 44 sub-​Saharan African countries for the period 2002–​2014. Many studies have hypothesized a negative association between internet penetration and corruption based on the above discussion in a cross-​country setting. As discussed earlier, potential endogeneity could be an issue and corruption could also be highly persistent over time (Mishra, 2006). To address endogeneity concerns, some of the recent studies have resorted to an IV analysis. For instance, Andersen et al. (2011) examine the relationship between internet penetration and corruption both across countries and across the US states. They instrument internet penetration with lightning intensity that causes power disruptions impacting the cost of ICT use. It is a strong instrument since by increasing the cost of ICT use, it is negatively associated with internet penetration. It is arguably a valid instrument as the lightning intensity does not affect corruption except through by having an impact on the use of the internet. They find that internet adoption is negatively associated with corruption both across countries and across the US states. A key advantage of within-​country studies like this examining the association between internet penetration and corruption across the US states eliminates the concerns that this association could have been driven by country-​specific unobserved fixed factors. The negative association between internet penetration and corruption has also been confirmed by a recent study by Jha and Sarangi (2017), who use an IV analysis to explore the association between social media and corruption. The authors instrument internet penetration with technological adoption in 1500 AD arguing that nations’ past technological advantages have persisted over the long run and exist even today (Comin, Easterly and Gong, 2010). Thus, countries that had better technology would have greater internet penetration today making it a strong instrument. At the same time, there is little reason to expect that a country’s technological state would have a direct effect on today’s level of corruption. Other studies such as Lio, Liu and Ou (2011) and Elbahnasawy (2014) use dynamic panel estimation to address the persistence of corruption. Lio, Liu and Ou (2011) use dynamic panel data models to investigate the association between internet adoption and corruption for 70 countries over 1998–​2005. They find internet adoption to be significantly, negatively associated with corruption. Elbahnasawy (2014) uses SGMM estimation to investigate the association between internet penetration and corruption and finds a significant, negative association between the two. Goel, Nelson and Naretta (2012) take a different approach to the issue. They examine whether greater internet

180  Chandan Kumar Jha and Sudipta Sarangi corruption awareness can reduce corruption by creating awareness about the nature and type of corruption in a country. To measure internet corruption awareness in a country, the authors conduct Google and Yahoo search for keywords “corruption” and “bribery” with that country’s name. For instance, the search term for measuring internet corruption awareness in Albania would be a variation of “Albania corruption bribery.” The authors argue that in countries with high awareness of the presence of corruption as indicated by the number of related search terms, the policymakers will have greater incentive to constrain corruption. They find that their measure of internet awareness of corruption per capita is significantly, negatively associated with corruption in a sample of 150 countries. To summarize, there is robust evidence that mobile phones and internet adoption can have significant, negative effects on corruption. By enabling better dissemination of information, mobile phones and the internet promote greater accountability and transparency while also reducing the scope for bribe demands. Further, they also provide a previously unavailable monitoring technology that allows citizens to record and anonymously share instances of corruption. Hence, it is not surprising that the empirical evidence stacks up strongly in favor of the negative effects of these communication technologies and corruption.

Social Media and Corruption In recent years a number of studies have started to focus their attention on the most recent form of ICTs –​social media –​that has become an essential part of our daily lives. Social media platforms use the internet to facilitate communication and conversations and allow users to generate and share information with other users on the platform. Studies investigating the effects of social media on corruption measure social media in various ways. For instance, Jha and Sarangi (2017) use the proportion of the population on Facebook as a proxy for social media. Tang et al. (2019) utilize the World Economic Forum’s (WEF) Global IT Report (Dutta, Geiger and Lanvin, 2015) for data on social media usage across countries. Here social media usage captures how extensively social media platforms like Facebook, Twitter, LinkedIn, and so on are used by people in a country. The variable is constructed using the following survey question: “In your country, how widely used are virtual social networks (e.g., Facebook, Twitter, LinkedIn) for professional and personal communication?” Within-​ country studies like Enikolopov, Petrova and Sonin (2018) and Qin, Strömberg and Wu (2017) have looked at the effects of influential bloggers in Russia and Sina Weibo (a Chinese microblogging website) posts in China, respectively. Let us first understand why social media works differently from information dissemination on the internet. An attractive feature of social media is its ability to provide multi-​way and quick means of communication that is hard to monitor and curb unlike traditional media like radio and television that provide only one-​way communication. It is well known that governments

How Advances in ICTs Impact Corruption?  181 in many countries censor content exposing political corruption, police brutality and human rights violations by controlling print and broadcast media (Freedom House, 2009b, 2009a). While governments around the world also clamped down on the content shared on the internet,14 it is much harder to do so because social media not only allows for multi-​way means of communication but also affords the users anonymity that makes authorities to monitor and identify them (Freedom House, 2009b). These features of social media allow citizens to share their experiences with corrupt officials with a broader audience without depending on the mainstream media to report the incident (Jha and Sarangi, 2017). The authors further argue that people feel compelled to act on corruption experiences shared by their friends and family members on social media platforms to show solidarity. This can create additional pressure on policymakers to act on the information. Moreover, social media can be used to facilitate mass mobilization against corruption by anticorruption activists by exchanging information with the like-​minded people. In fact, social media has been used to share live videos of the protests to prevent the authorities from using excessive force against protestors. Some major events provide support to the arguments that social media can be used to mobilize support against corruption and organize anticorruption protests. For instance, social media has been credited for playing an important role in facilitating the large scale anticorruption protests in India that became one of Time’s top ten world news for the year 2011 (Jha, 2014; Bong et al., 2012).15 Anticorruption activists extensively used social media to organize protests by communicating information about the time and place of such protests. Several recent studies have empirically examined the relationship between social media and corruption both in cross-​country and within-​country settings (Jha and Sarangi, 2017; Enikolopov, Petrova and Sonin, 2018; Tang et al., 2019). Enikolopov, Petrova and Sonin (2018) study the link between social media and corruption by focusing on the blog of an influential Russian blogger and “shareholder” activist, Alexey Navalny. They find that that blog posts exposing corruption in state-​controlled companies has a causal negative effect on their market returns.The blog posts are positively associated with management turnover reflecting increased accountability. The authors conclude that social media has the potential to reduce corruption even in countries where traditional media is strongly regulated and censored. An additional mechanism linking social media with corruption is the surveillance function provided by social media as discussed by Qin, Strömberg and Wu (2017) in the Chinese context but may be applicable to other countries as well. Governments can track social media activities that can provide them with a better picture of the public’s opinion regarding the prevalence of corruption among local officials and politicians. Social media thus can be used to identify the extent of corruption in a region which may not be feasible for local newspapers since they are subject local government control.The policymakers can use this information to take appropriate actions before the threat of anticorruption protests arise. Consistently, Qin, Strömberg and Wu (2017) find that Sina Weibo posts in China are flooded with

182  Chandan Kumar Jha and Sudipta Sarangi corruption allegations and they can predict future corruption charges against specific individuals up to a year in advance. Using an IV analysis, Jha and Sarangi (2017) study the relationship between social media and corruption. In a cross-​country analysis of over 150 countries, they find that social media, proxied by Facebook penetration, is negatively associated with corruption. Note that this association is shown to be significant after controlling for the internet diffusion suggesting that the effects of social media on corruption is in addition to other ways such as the use of e-​government in which the internet can influence corruption. Their findings further suggest that social media plays a complementary role to free press in reducing corruption. The role of press freedom in constraining corruption by acting as an external control has been documented by Brunetti and Weder (2003). Tang et al. (2019) find a negative association between social media and corruption in a panel analysis of 62 countries over the years 2011–​2015. They further find that the effect of social media on corruption is stronger in countries characterized by “loose cultures” that are characterized by weak social norms and greater tolerance for deviance from these norms than tight cultures that have strict social norms and little tolerance for noncompliance. Although the literature on social media and corruption is still relatively new, the findings seem to be robust: Social media usage is negatively associated with corruption. The current evidence also suggests that social media can be more effective in reducing corruption in countries where press enjoys greater freedom and in countries that with strict social norms. Further research on the effects of social media on corruption should incorporate better measures and richer data on social media to assess its effects on corruption and whether the association between the two is mediated by other cultural or institutional factors.

Concluding Remarks In this section, we would like to draw attention to the fact that the effectiveness of ICTs in constraining corruption might need to be complemented with other aspects of an economy. These features require a collective effort from various stakeholders including the citizens, civil society and the government. For instance, a greater awareness about the existence of corruption does not automatically mean that corruption will fall. Instead it requires that citizens demand action against corruption and refuse to pay bribes. And, while the provision of e-​government requires necessary infrastructure such as computers and the internet connection to be provided by the government, the end-​user must be willing and able to utilize these services. Thus, the citizens need to be both informed and educated to avail government services that are offered digitally. Similarly, many studies have argued that one of the ways e-​government and the internet reduce corruption is by promoting transparency. However, the empirical evidence suggests that while transparency does reduce corruption, it may not be sufficient in itself and may need to be complemented with other

How Advances in ICTs Impact Corruption?  183 policies (Peisakhin, 2012; Kolstad and Wiig, 2009). Kolstad and Wiig (2009), for example, argue that in order for transparency to have a significant impact on corruption, agents with increased access to information must also be able to process that information and possess the ability and incentive to act on it. Thus, the effects of transparency on corruption will also heavily depend on citizen’s level of education and their ability to hold corrupt government and bureaucrats accountable. Informed and educated citizens will be able to use e-​government more effectively and can assess the validity and implications of information available on the net. The latter is an important concern in the age of social media which is rife with false information and political propaganda (see Kumar and Shah 2018 for a review of literature on social media’s role in spreading false information). Kossow and Kukutschka (2017) argue that ICTs can be used by enlightened citizens to inform themselves about corruption and mobilize support to fight it. The authors use share of the population that used the internet at least once during the last year, the percentage of the population with broadband access, newspaper circulation and the share of total population that are Facebook users to proxy for enlightened citizens. They find these measures to be negatively correlated with corruption. For many forms of ICTs such as the internet and social media to be effective against corruption, it is imperative that content is not censored and unbiased information is available to the public. Moreover, it is crucial that the freedom on the net is ensured and anticorruption activists are protected from retaliation by corrupt and authoritarian states. In this regard, the role of civil society becomes very important as they must fight to ensure freedom on the net.16 Governments in many countries around the world have been found to clamp down on internet freedom or spy on their own citizens using vague laws and technological means, resulting in self-​censorship by citizens on the web (Freedom House, 2015, 2009b).This problem is quite widespread –​even countries considered to be champions of democracy like the United States and United Kingdom have been reported to spy on their own citizens. Such government practices will undermine the effectiveness of ICTs on corruption by promotion self-​censorship by citizens. Existing studies have highlighted the role of civil society in this context. For instance, Kossow and Kukutschka (2017) find that the ICTs can be more effective in fighting corruption when the civil society is more organized. They argue that ICTs and strong civil society are complementary in their effectiveness against corruption. To measure organized civil society, the authors use the Core Civil Society Index (CCSI) based on the Varieties of Democracy (V-​Dem) dataset.17 The CCSI “captures the robustness of civil society by measuring the level of citizen activism and the organizational environment for CSOs –​namely state repression and entry/​exit control” (Bernhard et al., 2015). In the same vein, Jha and Sarangi (2017) argue that one of the ways social media can reduce corruption is by being complementary to press freedom, that is, making it possible that the press has a wider reach. They find that the effect of social media is indeed stronger in countries where the press enjoys greater freedom.

184  Chandan Kumar Jha and Sudipta Sarangi Finally, the empirical evidence suggests that strong institutions are also important for ICTs to be effective in reducing corruption. In an analysis of 90 countries for 2012, Bhattacherjee and Shrivastava (2018) find that ICT use is positively and significantly associated with both certainty and celerity of punishment in corruption-​related crimes. Moreover, they find that ICT laws positively and significantly moderate these relationships. In other words, if two countries have similar ICT use, the country with the stronger ICT laws would have higher certainty and celerity of punishment for corruption and this, in turn, will lead to lower corruption. This finding is also supported by Sassi and Ali (2017) who find that ICT diffusion is more effective in reducing corruption in the presence of stronger rule of law. In sum, ICTs have been powerful tools against corruption, but a collective action from the government, civil society and citizens is required in order to realize their full potential. Technology by its very nature will change and it will create new ways of affecting corruption. It will also create possibly new channels of complementarity. Hence, research in this area will need to keep up with evolving technology to keep providing us with insights about the relationship between ICTs and corruption.

Notes 1 Interestingly, ICT has been referred to as “liberation technology” because it allows for a decentralized and open-​access multi-​way communication at very low cost with the potential to foster political activism and promote mass political mobilization (Shirky, 2011; Diamond, 2010). Various forms of ICTs, including mobile phones, internet and social media, have been found to be instrumental in mobilizing citizens for anticorruption and prodemocracy movements in many countries and credited with the success of these protests in many countries (Enikolopov, Makarin and Petrova, 2020; Manacorda and Tesei, 2020; Jha and Kodila-​Tedika, 2020; Breuer, Landman and Farquhar, 2015; Jha, 2014; Bong et al., 2012; Howard et al., 2011). 2 The closest options are survey data such as the World Bank Enterprise Survey (WBES) (www.enterp​r ise​surv​eys.org/​en/​enterp​r ise​surv​eys) and the International Crime Victimization Survey (ICVS) (https://​wp.unil.ch/​icvs/​) that record individuals’ and businesses’ corruption experiences with the government officials. These survey data do have some limitations. First, they are not available for every country of the world for any given year. Second, while the WBES survey focuses on firms, ICVS focuses on individuals. Hence, it is reasonable to argue that neither of these two captures the true extent of corruption that affects both business and individuals in the country. Moreover, these surveys might not reflect the existence of grand corruption involving politicians and high-​level government officials as small businesses and individuals are unlikely to be involved in a direct exchange with them. 3 For methodology of the CCI, see Kaufmann, Kraay and Mastruzzi (2011); data available at https://​info.worldb​ank.org/​gov​erna​nce/​wgi/​. For methodology of the ICRG corruption index, see Howell (2011); data available at https://​epub.prsgr​oup. com/​produ​cts/​icrg. For CPI data and methodology, visit www.trans​pare​ncy.org/​en/​ cpi/​2020/​index.

How Advances in ICTs Impact Corruption?  185 4 Note that the majority of the literature reports that ICTs help reduce corruption. Only recently some studies have highlighted that certain types of ICTs can promote corruption by studying the role of virtual and cryptocurrencies in facilitating corrupt transaction and laundering corruption money and other financial crimes (Teichmann and Falker, 2021; Choo, 2015). 5 https://​publi​cadm​inis​trat​ion.un.org/​ego​vkb/​en-​us/​about/​meth​odol​ogy. 6 www.ins​idep​olit​ics.org/​egovtd​ata.html. 7 The Bhoomi project, started in 1991, computerized 20 million land records owned by 6.7 million farmers that used to be maintained by approximately 9,000 village accountants. Prior to this, the transfer of ownership of land or obtaining a copy of the RTC –​a document required for various purposes including to get a loan –​typically involved a bribe. 8 See Bhatnagar (2003, 2004) and Bhuiyan (2011) for many other case studies from different parts of the world documenting the negative association between e-​ government and corruption. 9 The WJP is an independent organization with an objective to promote the rule of law worldwide. It provides the rule of law index for 139 countries that is based on household and expert surveys. More information available at https://​worl​djus​tice​ proj​ect.org. 10 The diffusion of VSN is constructed using responses to the following survey question from the WEF’s Global IT Report (Dutta, Geiger and Lanvin, 2015): “In your country, how widely are virtual social networks (Facebook, Twitter, LinkedIn, etc.) used?” 11 Interestingly, other measures of culture such as individualism and masculinity are not found to moderate the relationship between e-​government and corruption although studies have documented a significant relationship between individualism and corruption (Jha and Panda, 2017). 12 For instance, the web platform “ipaidabribe.com” allows the victims of corruption in India to report the instances of bribery and its details (the name of the corrupt official, bribe amount and the services availed) anonymously. This information is used to create awareness and “to argue for improving governance systems and procedures, tightening law enforcement and regulation and thereby reduce the scope for corruption in obtaining services from the government” (ipaidabribe. com). 13 For further details on DAI, visit www.itu.int/​ITU-​D/​ict/​dai. 14 In many instances, by the time a government is able to remove the content it does not want to be shared, it is too late. For instance, Washington Post reports that the documentary Under the Dome had millions of online views before it was taken down by the Chinese government. See www.was​hing​tonp​ost.com/​news/​ene​rgy-​envi​ ronm​ent/​wp/​2015/​03/​16/​this-​docu​ment​ary-​went-​viral-​in-​china-​then-​it-​was-​ censo​red-​it-​wont-​be-​forgot​ten/​ (accessed June 27, 2021). 15 http://​cont​ent.time.com/​time/​speci​als/​packa​ges/​arti​cle/​0,28804,2101​344_​2101​ 368_​2101​650,00.html (accessed July 11, 2021). 16 See Jha (2017) and Diamond (2010) for detailed discussions on the role of civil society in determining the effectiveness of civil society on corruption and democracy, respectively. 17 For more information, visit https://​v-​dem.net.

186  Chandan Kumar Jha and Sudipta Sarangi

References Andersen, Thomas Barnebeck (2009). “E-​Government as an anti-​corruption strategy.” Information Economics and Policy, 21: 201–​210. Andersen, Thomas Barnebeck, Jeanet Bentzen, Carl-​Johan Dalgaard and Pablo Selaya (2011). “Does the internet reduce corruption? Evidence from US states and across countries.” World Bank Economic Review, 25: 387–​417. Arayankalam, Jithesh, Anupriya Khan and Satish Krishnan (2021). “How to deal with corruption? Examining the roles of e-​government maturity, government administrative effectiveness, and virtual social networks diffusion.” International Journal of Information Management, 58: 102203. Bernhard, Michael, Eitan Tzelgov, Dong-​Joon Jung, Michael Coppedge and Staffan I. Lindberg (2015). “The varieties of democracy core civil society index.” V-​Dem Working Paper, 13. Bhatnagar, Subhash (2003).“E-​government and access to information.” Global Corruption Report, 2003: 24–​32. Bhatnagar, Subhash (2004). E-​government: From vision to implementation –​A practical guide with case studies. Sage. Bhattacherjee, Anol, and Utkarsh Shrivastava (2018). “The effects of ICT use and ICT laws on corruption: A general deterrence theory perspective.” Government Information Quarterly, 35: 703–​712. Bhuiyan, Shahjahan H. (2011).“Modernizing Bangladesh public administration through e-​governance: Benefits and challenges.” Government Information Quarterly, 28: 54–​65. Bong, Sasha, Kenneth Chung, Karen Parkinson, Andrew Peppard, Justin Rabbach and Nicole Thiher (2012).“Analyzing social media momentum.” Workshop in International Public Affairs. Breuer, Anita, Todd Landman and Dorothea Farquhar (2015). “Social media and protest mobilization: Evidence from the Tunisian revolution.” Democratization, 22: 764–​792. Brunetti, Aymo, and Beatrice Weder (2003). “A free press is bad news for corruption.” Journal of Public Economics, 87: 1801–​1824. Choo, Kim-​Kwang Raymond (2015).“Cryptocurrency and virtual currency: Corruption and money laundering/​terrorism financing risks?” In Handbook of digital currency. Elsevier, pp. 283–​307. Collier, Paul, and Anke Hoeffler (2005). “The economic cost of corruption in infrastructure.” In Global corruption report. Transparency International, pp. 12–​19. Comin, Diego, William Easterly and Erick Gong (2010). “Was the wealth of nations determined in 1000 BC?” American Economic Journal: Macroeconomics: 65–​97. Diamond, Larry (2010). “Liberation technology.” Journal of Democracy, 21: 69–​83. DiRienzo, Cassandra E., Jayoti Das, Kathryn T. Cort and John Burbridge (2007). “Corruption and the role of information.” Journal of International Business Studies, 38: 320–​332. Donchev, Dilyan, and Gergely Ujhelyi (2014). “What do corruption indices measure?” Economics & Politics, 26: 309–​331. Dutta, Soumitra, Thierry Geiger and Bruno Lanvin (2015). The global information technology report 2015. World Economic Forum. Elbahnasawy, Nasr G. (2014). “E-​government, internet adoption, and corruption: An empirical investigation.” World Development, 57: 114–​126. Enikolopov, Ruben, Alexey Makarin and Maria Petrova (2020). “Social media and protest participation: Evidence from Russia.” Econometrica, 88: 1479–​1514.

How Advances in ICTs Impact Corruption?  187 Enikolopov, Ruben, Maria Petrova and Konstantin Sonin (2018). “Social media and corruption.” American Economic Journal: Applied Economics, 10: 150–​174. Freedom House (2009a). Freedom of the press 2008: A global survey of media independence. Rowman & Littlefield. Freedom House (2009b). Freedom on the net: A global assessment of internet and digital media. https://​freed​omho​use.org/​sites/​defa​ult/​files/​2020-​02/​FOT​N_​Fr​eedo​m_​On​ The%20N​et_​F​ull%20Repo​rt_​2​009.pdf Freedom House (2015). Freedom on the net 2015. https://​freed​omho​use.org/​sites/​defa​ ult/​files/​2020-​02/​FH_​FOTN_​20​15Re​port​_​Ove​rvie​wEss​ayFi​nal.pdf Goel, Rajeev K., Michael A. Nelson and Michael A. Naretta (2012). “The internet as an indicator of corruption awareness.” European Journal of Political Economy, 28: 64–​75. Howard, Philip N., Aiden Duffy, Deen Freelon, Muzammil M. Hussain, Will Mari and Marwa Mazaid (2011). “Opening closed regimes: What was the role of social media during the Arab Spring?” SSRN 2595096. Howell, Llewellyn D. (2011). International country risk guide methodology. PRS Group. Jha, Chandan, and Bibhudutta Panda (2017). “Individualism and corruption: A cross-​ country analysis.” Economic Papers: A Journal of Applied Economics and Policy, 36: 60–​74. Jha, Chandan Kumar (2014). “Can social media and the internet reduce corruption?” International Growth Centre. www.the​igc.org/​blog/​can-​soc​ial-​media-​and-​the-​ inter​net-​help-​red​uce-​cor​rupt​ion/​ Jha, Chandan Kumar (2017). “Information control, transparency, and social media: Implications for corruption.” In Political scandal, corruption, and legitimacy in the age of social media. IGI Global, pp. 51–​75. Jha, Chandan Kumar, and Oasis Kodila-​Tedika (2020). “Does social media promote democracy? Some empirical evidence.” Journal of Policy Modeling, 42: 271–​290. Jha, Chandan Kumar, and Sudipta Sarangi (2017). “Does social media reduce corruption?” Information Economics and Policy, 39: 60–​71. Kanyam, Daniel A., Genti Kostandini and Susana Ferreira (2017). “The mobile phone revolution: Have mobile phones and the internet reduced corruption in Sub-​Saharan Africa?” World Development, 99: 271–​284. Kaufmann, Daniel, Aart Kraay and Massimo Mastruzzi (2011). “The worldwide governance indicators: Methodology and analytical issues.” Hague Journal on the Rule of Law, 3: 220–​246. Kim, Seongcheol, Hyun Jeong Kim and Heejin Lee (2009). “An institutional analysis of an e-​government system for anti-​corruption: The case of OPEN.” Government Information Quarterly, 26: 42–​50. Knutsen, Carl Henrik (2015). “Why democracies outgrow autocracies in the long run: Civil liberties, information flows and technological change.” Kyklos, 68: 357–​384. Kolstad, Ivar, and Arne Wiig (2009). “Is transparency the key to reducing corruption in resource-​r ich countries?” World Development, 37: 521–​532. Kossow, Niklas, and Roberto Martínez Barranco Kukutschka (2017). “Civil society and online connectivity: Controlling corruption on the net?” Crime, Law and Social Change, 68: 459–​476. Kumar, Srijan, and Neil Shah (2018). “False information on web and social media: A Survey.” arXiv:1804.08559. Lio, Mon-​Chi, Meng-​Chun Liu and Yi-​Pey Ou (2011). “Can the internet reduce corruption? A cross-​country study based on dynamic panel data models.” Government Information Quarterly, 28: 47–​53.

188  Chandan Kumar Jha and Sudipta Sarangi Manacorda, Marco, and Andrea Tesei (2020). “Liberation technology: Mobile phones and political mobilization in Africa.” Econometrica, 88: 533–​567. Martins, João, Linda Veiga and Bruno Fernandes (2021). Does electronic government deter corruption? Evidence from across the world. Technical report. NIPE-​ Universidade do Minho. Mishra, Ajit (2006). “Persistence of corruption: Some theoretical perspectives.” World Development, 34: 349–​358. Mocan, Naci (2008). “What determines corruption? International evidence from microdata.” Economic Inquiry, 46: 493–​510. Nam, Taewoo (2018). “Examining the anti-​corruption effect of e-​government and the moderating effect of national culture: A cross-​country study.” Government Information Quarterly, 35: 273–​282. Park, Chul Hyun, and Koomin Kim (2020). “E-​government as an anti-​corruption tool: Panel data analysis across countries.” International Review of Administrative Sciences, 86: 691–​707. Peisakhin, Leonid (2012). “Transparency and corruption: Evidence from India.” Journal of Law and Economics, 55: 129–​149. Qin, Bei, David Strömberg and Yanhui Wu (2017). “Why does China allow freer social media? Protests versus surveillance and propaganda.” Journal of Economic Perspectives, 31: 117–​140. Sassi, Seifallah, and Mohamed Sami Ben Ali (2017). “Corruption in Africa: What role does ICT diffusion play.” Telecommunications Policy, 41: 662–​669. Shirky, Clay (2011). “The political power of social media: Technology, the public sphere, and political change.” Foreign Affairs: 28–​41. Srivastava, Shirish C., Thompson S. H. Teo and Sarv Devaraj (2016). “You can’t bribe a computer: Dealing with the societal challenge of corruption through ICT.” MIS Quarterly, 40: 511–​526. Starke, Christopher, Teresa K. Naab and Helmut Scherer (2016). “Free to expose corruption: The impact of media freedom, internet access and governmental online service delivery on corruption.” International Journal of Communication, 10: 21. Tang, Zhenya, Leida Chen, Zhongyun Zhou, Merrill Warkentin and Mark L. Gillenson (2019). “The effects of social media use on control of corruption and moderating role of cultural tightness-​looseness.” Government Information Quarterly, 36: 101384. Teichmann, Fabian Maximilian Johannes, and Marie-​ Christin Falker (2021). “Cryptocurrencies and financial crime: Solutions from Liechtenstein.” Journal of Money Laundering Control, 24: 775–​788. Transparency International. (2015). “Corruption perceptions index 2015.” www.trans​ pare​ncy.org/​en/​cpi/​2015 (accessed May 19, 2022). United Nations. (2010). “United Nations e-​ government survey 2010: Leveraging e-​government at a time of financial and economic crisis.” Wagner, Ben (2012). “Push-​button-​autocracy in Tunisia: Analysing the role of internet infrastructure, institutions and international markets in creating a Tunisian censorship regime.” Telecommunications Policy, 36: 484–​492. West, Darrell (2007). “Global e-​government, 2007.” www.ins​idep​olit​ics.org/​ego​vt07​int. pdf (accessed June 26, 2021). World Bank. (2004). World development report 2004: Making services work for poor people. World Bank.

13 Corruption in Europe The Underestimated Devil and the Role of the European Union Ina Kubbe and Stoyan Panov

Introduction From a comparative perspective, Europe is often considered a continent that suffers less from corruption.Yet, European countries exhibit an array of corrupt activities and are characterized by large cross-​national and progressive differences in the extent of corruption. An example of this is provided by Transparency International’s Corruption Perception Index 2021 –​this index, among others, indicates that EU member states, in particular, continue to represent the best performing region in the world (Corruption Perception Index 2021). Denmark, Finland and New Zealand are the highest scoring countries (88 out of 100, where 0 is highly corrupt and 100 is very clean), followed by Sweden, Norway and Singapore (85). However, we also find countries such as Romania (45), Hungary (43) and Bulgaria (42) at the bottom of the index. These findings are also in line with other measurements of corruption such as those provided by the World Bank (2021) or the Eurobarometer (2021). Constant scandals in nearly every European state illustrate that corrupt activities still persist over time. Since the beginning of the economic crisis in 2007 –​further exacerbated by the beginning of the COVID-​19 pandemic in 2020 –​country corruption values have continuously increased, in particular in Southern European and Central and Eastern European countries. At the same time, the often perceived “clean North” also exhibits signs of not being as clean as it seemed. Corruption, broadly defined as the abuse of entrusted power for private gain (e.g., Transparency International, 2021a), can take different forms, including bribery, favoritism or sextortion. It is also often combined with organized crime as corruption cases in the Netherlands or Italy often demonstrate (e.g., Henley, 2019). Research still remains to be conducted to identify not only all the different types of corruption but also how they relate to each other. We do know, however, that corruption is detrimental to economic, social, political and sustainable development.1 The fight against corruption has therefore been high on the agenda of international and multilateral organizations, such as the EU, for a long time. However, Europe in general and the EU in particular are still looking for effective anticorruption strategies (European Commission, 2021a). DOI: 10.4324/9781003142300-13

190  Ina Kubbe and Stoyan Panov The chapter provides an overview of the causes, consequences and trends in the area of corruption in EU member states and discusses the anticorruption strategies and mechanisms of the Union. The study analyzes and assesses data from the most recent Special Eurobarometer 502 (SEB 502) that covers the perception of corruption in the business sector and citizen’s perceptions. The findings below indicate that it is worth examining whether the same methodology is appropriate to use as the regions are fairly heterogeneous: even within the regions, there are some noticeable differences. The added value of our project is rooted in mixing and matching perceptions of corruption with existing legal frameworks on EU level on the subject matter. The analysis is limited to the actors and legislative efforts on supranational level only. In that sense, the transposition in each EU member state and subsequent implementation on domestic level are out of scope of the current chapter. We look at secondary EU legislation such as Directive 2017/​1371 on the fight against fraud to the Union’s financial interests by means of criminal law (“PIF” directive). The analysis of the EU anticorruption legal framework begins with corruption as a “euro-​crime” under Article 83(1) of the Treaty on the Functioning of the European Union (TFEU), which states that acts of corruption constitute particularly “serious crimes with a cross-​ border dimension resulting from the nature or impact of such offenses or from a special need to combat them on a common basis.” With the adoption of the Stockholm Program in 2009,2 the European Commission further received a political mandate to measure efforts in the fight against corruption as well as to develop a comprehensive EU anticorruption policy in cooperation with the Council of Europe Group of States against Corruption (GRECO) (Horizon, 2020). This article reflects on the EU’s increasing anticorruption efforts such as the establishment of the European Public Prosecutor’s Office (EPPO) in addition to the already existing OLAF and the EU’s mechanism on the protection of the EU’s budget in case of breaches of the rule of law, which includes corruption or other breaches to the implementation of the Union budget.

Corruption (Perception) in European Member States Corrupt practices vary widely in Europe across North-​Western, Central and Eastern Europe and Southern Europe. In the following sections, each region is described in more detail to portray underlying attitudes to corruption and uncover legislative loopholes in the treatment of corruption as a crime. Suggestions for improvement are also provided.The results are also summarized in the table in the appendix which provides an overview of the main forms of corruption in the public and private sector, the failures in institutions regarding its anticorruption efforts, prevalent norms and informal practices that might enable corruption and recommendations.

Corruption in Europe  191 North-​Western Europe According to the latest SEB 502,3 the perception of and experience with corruption in North-​Western Europe are still very low compared with the other EU member states and from a global perspective. In particular, the Scandinavian countries (e.g., Denmark, Sweden, Finland) show very low corruption scores, but so do Luxembourg, Germany and the Netherlands. Yet the results give reasons for concern and question the modalities for measuring corrupt activities in general. This needs to be taken into account especially in the context of COVID-​19, which might provide new insights regarding corrupt practices, in the health care sector and beyond. Many countries in the North of Europe are still characterized by a very fragmented anticorruption landscape, unfinished reforms (e.g., Ireland; see Devitt, 2021) and underresourced and underpowered key institutions. The most concerning issue regarding the capacity of institutions remains threats to their independence and risk of political-​driven control over them. This also explains the deterioration of anticorruption norms and why trust in political institutions remains comparatively low across the region such as in Austria, Ireland, Netherlands or Germany. The general perception is that the level of corruption has become more widespread since 2017 –​especially in Denmark, Ireland and Austria, followed by the Netherlands and Sweden.4 This increase can be attributed to a higher awareness of corruption among the general public due to the higher coverage of scandals by the mass media (e.g., the Ibiza affair in Austria) (Oltermann, 2019). Still, many citizens, for example in Austria, consider corrupt practices to be acceptable. In the Netherlands, known as one of the least corrupt countries (e.g., Corruption Perception Index, 2021), 21% of respondents consider it to be totally acceptable to give money or to do someone a favor to receive public administration or a public service.5 In Germany, bribe-​taking and power abuse are seen most prevalent in political parties (46%), private companies (43%) and among politicians (40%).The negative perception of parties can be explained by a persisting lack of transparent party financing and several scandals surrounding party sponsoring (e.g., European Parliamentary Research Service, 2021). In most of the countries in the North, the public believes that corruption is mainly related to business –​mostly via too close links between business and politics. Negative perceptions of private companies are also commonplace in Germany, with a number of scandals impacting public perception, such as the Siemens foreign bribery scandal (Schubert and Miller, 2008), the Volkswagen emission scandal (Hotten, 2015) or the Cum Ex Tax scandal (European Parliament, 2018). Corruption scandals, such as the “German Mask Scandals” (DW Germany, 2021), also indicate how pervasive illegal practices are in society, in particular within parties that collaborate with private companies. Furthermore, money laundering has been described as a key problem, notably

192  Ina Kubbe and Stoyan Panov in the German real estate sector (e.g., Trautvetter and Henn, 2020). Also, the extensive lobbying of the German industry might contribute to the image that companies exert undue influence. This is also the case for the Netherlands and Denmark, which still lack lobbying regulations (e.g., Kergueno et al., 2021). Swedish respondents find corruption to be present foremost among private companies, banks and financial institutions. This perception might be a result of the money-​laundering scandals emerging in two of the top four leading banks in Sweden in 2019, Swedbank and SEB, which gained significant international attention, share price losses and ensuing investigations in several states (Hoikkala, 2020). Yet, it is a result that clearly diverges from the EU average as well. Furthermore, almost 70% of Austrians agree that too close links between business and politics lead to corruption. In particular, in small countries like Austria and Luxembourg, it is common to use informal ways to receive privileged information or to render a service to maintain good social relations. It is a practice not necessarily experienced or seen as real corruption and is considered a part of the business culture in many countries (SEB 502). Although the healthcare sector is generally very vulnerable to corruption,6 it is not considered a sensitive sector in North-​Western Europe (e.g., Finland, Sweden, Ireland, Netherlands). An exception is Austria, where almost 20% reported that they had to give an extra payment or a valuable gift to a nurse or a doctor or make a hospital donation.7 The practice of “envelope medicine” has existed in Austrian hospitals for many years (Sommersguter-​Reichmann and Stepan, 2017). In Germany, a similar trend exists. Since the introduction of the German Law Combatting Corruption in the Health Sector in 2016, the official corruption statistics first noted a considerable increase in cases of severe corruption for the year 2019 with a total 138 cases (in 2018: two cases).Thus, the health care sector took a share of 32.4% among all cases of severe corruption in Germany. Cases of acceptance of bribes also rose from to 135 (in 2018, 40); cases of granting bribes rose to 146 (in 2018, 29) (Muschel, 2020). The tendency might have increased due to COVID-​19. Moreover, general regulations regarding “revolving doors” are either weak or limited (e.g., Netherlands) or do not exist at all (e.g., Denmark and the Netherlands), as is the case for conflict of interest in political offices (e.g., Finland, Germany). For instance, in December 2020 it emerged that a former head of the Danish Defense Ministry was also employed as a lobbyist by a defense consulting firm (Altinget, 2020). This is not unlawful because of the lack of revolving door regulations in Denmark. There are no general rules regarding duties to declare financial interests for civil servants and ministers in the country (SEB 502). Generally, the SEB 502 data indicates that all countries in North-​Western Europe need better guidelines on how and where to report corruption. Overall, the public greatly hesitates to report corruption for several reasons. First and foremost, citizens do not know where to report it; secondly, corruption seems to be difficult to prove; thirdly, a great fear of not receiving adequate protection

Corruption in Europe  193 exists (Denmark, Sweden, Finland, Netherlands, Germany, Austria, Ireland). Almost 70% in Austria do not know where to report a case of corruption. Although citizens have comparatively high levels of trust in political institutions such as the judiciary and the police, in particular in Sweden, Finland and Denmark, it is worth noting that reporting corruption is perceived as very accusatory and “taboo.” A strong cultural norm “not to betray others” often leads to the decision not to report corruption, in particular in Austria, where a substantially higher percentage of Austrians (28%) compared to the EU average (19%) decide not to report a corruption case due to this norm. Similarly, a substantially higher share of Austrians (24%), compared to the EU average (16%), says that it is not worth the effort of reporting it. Moreover, Sweden stands out slightly in the results concerning the reason not to report being that people do not know where to report (37% compared to the EU 28 average of 22%), that there is a lack of protection for those who report (31% compared to the EU 28 average of 29%) and that it is not worth the effort (19% compared to the EU 28 average of 16%) (SEB 502). In Germany, many citizens believe that everyone knows about cases, but no one reports them (15%) or that reporting is not worth the effort (20%). While none of the reported numbers shows a significant change compared to 2017, these perceptions reflect the persisting neglect of the topic of whistleblowing –​ not only in Germany –​for instance in terms of providing anonymous reporting channels or adequate legal protection. Yet, in Sweden, whistleblower protection has improved through a strengthened regulation concerning retaliation; however, this particular aspect in the Swedish context might be explained by media reporting on maltreatment of whistleblowers in conjunction with notable corruption scandals (SEB 502). The protection of whistleblowers has been inconsistent across the Northern and Western states and whistleblowers have not always been adequately protected against retaliation. There is currently no general legislation on the protection of whistleblowers in Finland.8 In general, EU member states have been slow in transposing the whistleblower directive in their respective legal orders (SEB 502). Furthermore, there is a clear need for better education and transparent information for the public regarding corrupt practices (e.g., Ireland, Luxembourg). In Luxembourg, for instance, high scores of the “no opinion” answer evidence a lack of knowledge of the general population which can only find its source in the lack of information and education on the subject. This increase of the “without opinion” responses is an important trend that is visible across the majority of Luxembourg answers. It can be explained by the lack of global campaign and political will to communicate on corruption and to increase awareness of the population, which does not feel concerned (SEB 502). Judicial Cooperation in Criminal Matters (Article 83 TFEU), Protection of the Financial Interests of the EU (Article 325 TFEU) and Financial Fraud and Corruption: Directive 2017/​1371

194  Ina Kubbe and Stoyan Panov The word “corruption” is mentioned once in the Treaty on the Functioning of the EU (TFEU). Under the heading “Judicial Cooperation in Criminal Matters” chapter, Article 83 TFEU empowers the European Parliament and the Council in accordance with the ordinary legislative procedure to establish minimum rules on the definition of various criminal offenses and corresponding sanctions with a particular focus on cross-​border criminal activity. Corruption is considered a serious crime, along with crimes such as terrorism, illicit drug trafficking, human trafficking, organized crime among others.9 The legal base for legislative activities against financial fraud including corruption on the EU level is also found in Article 325 TFEU. It enables the EU and member states to “counter fraud and any other illegal activities affecting the financial interests of the Union.”10 In order to achieve this, various measures can be implemented, “which shall act as a deterrent and be such as to afford effective protection in the Member States, and in all the Union’s institutions.”11 Member states are obliged to treat threats and correspondingly take countermeasures against fraud affecting the financial interest of the EU on the principle of equivalence as if the fraud is affecting their own domestic financial interests.12 It is noticeable that the EU’s anticorruption framework is based on a financial-​oriented or budget protection approach, while there is a current process of linking grand corruption with rule of law and democracy backsliding (Panov, 2019). What particular anticorruption frameworks exist at the EU level to respond to the identified problems in Northern Europe which are also prevalent in other parts of Europe? The provided examples of legislation on EU level are thematically grouped but they are applicable to all EU member states unless otherwise stipulated in the respective legal acts. One of the recurring themes in Europe in general is financial fraud and corruption practices. The most consequential legislative initiative in the field of anticorruption efforts on EU level is Directive 2017/​1371 on the fight against fraud to the EU’s financial interests by means of criminal law. This legislative framework aims to establish minimum rules concerning the definition of criminal offenses against fraud and other illegal activities affecting EU’s financial interests.13 It obliges the EU member states to “take the necessary measures to ensure that fraud affecting the Union’s financial interests constitutes a criminal offence when committed intentionally.”14 The directive defines criminal offenses of fraud affecting the EU’s financial interests as well as revenue arising from VAT own resources in cross-​border schemes, involving a total damage of at least 10 million euro.15 It requires member states to criminalize money laundering and property derived from criminal offenses under Article 1(3) of Directive 2015/​849. The most relevant provision for this chapter is contained in Article 4(2) of Directive 2017/​1371 which states that “Member States shall take the necessary measures to ensure that passive and active corruption, when committed intentionally, constitute criminal offences” in their respective domestic legal orders. Article 4(2) of the Directive defines “passive corruption” as “the action of a public official who,

Corruption in Europe  195 directly or through an intermediary, requests or receives advantages of any kind, for himself or for a third party, or accepts a promise of such an advantage, to act or to refrain from acting in accordance with his duty or in the exercise of his functions in a way which damages or is likely to damage the Union’s financial interests.” Active corruption per Article 4(2) includes the act of an individual “who promises, offers or gives, directly or through an intermediary, an advantage of any kind to a public official for himself or for a third party for him to act or to refrain from acting in accordance with his duty or in the exercise of his functions in a way which damages or is likely to damage the Union’s financial interests.” The Directive also includes an aggravating circumstance when the criminal offense is committed within a criminal organization.16 The Directive lists the relevant institutions which are tasked with cooperation in the fight against the criminal offenses, including corruption: OLAF and the Commission, Eurojust, the European Public Prosecutor’s along with the member states.17 Council Framework Decision 2003/​568/​JHA on Combating Corruption in the Private Sector Another noticeable recurring theme in Northern Europe is corruption in the private sector. In 2003, Council Framework Decision 2003/​568 was drafted in response to the assessment that corruption has a cross-​border effect, thus requiring action on European level. Building on the Tampere agenda in the area of Justice and Home Affairs (CEPS, 2020),18 there was a push to establish a set of minimum rules of standardization of the criminal offense of corruption in the private sector across the EU member states. Public and private corruption was deemed to pose “a threat to a law-​abiding society” as well as exerting a negative effect on commercial activities and economic development.19 The Framework Decision requires EU member states to take the necessary measures to criminalize corruption when carried out in the course of business activities.20 Additionally, it includes penalties for private corruption by obliging member states to introduce criminal penalties of a maximum of at least one to three years of imprisonment in their respective jurisdictions.21 The Decision includes liability of legal persons which should be concurrent with the liability of natural persons for the commission of private corruption. The criminal or noncriminal fines and other sanctions for legal persons include inter alia the exclusion from entitlement to public benefit or aid, or the temporary or permanent disqualification from the practice of commercial activities.22 Directive (EU) 2018/​ 1673 on Combating Money Laundering by Criminal Law Money laundering is a noticeable issue, reflected in the perception of corruption-​related problems in Northern Europe. Here perceptions meet the law to a certain degree as the EU has been significantly active in this area.

196  Ina Kubbe and Stoyan Panov The directive on combating money laundering aims through criminalization to create more efficient cross-​border cooperation between the relevant competent authorities by providing a sufficiently uniform definition of the criminal activities constituting the predicate offenses.23 The aim is to respond to the impact of money laundering by public officials through more severe penalties for the criminal activities in the respective national frameworks such as maximum terms of imprisonment of at least four years.24 The criminalized activities include corruption.25 The money-​laundering offenses linked to the criminal activities include inter alia the intentional commission of conversion or transfer of property, knowingly derived from the criminal activity, and the acquisition, possession or use of property, derived from criminal activity at the time of the receipt.26 The EU has also been recently active in responding to financial crimes of the EU budgets by legislating the Market Abuse Directive and the Regulation to counter white-​collar crimes as part of the provisions under Article 114 TFEU.27 The legal framework of the Directive establishes liability of legal persons for the money-​laundering offenses for any person, who has acted “either individually or as part of an organ of the legal person and having a leading position within the legal person” as manifested in the power of representation of the legal person, an authority to take decisions for the legal person, or an authority to exercise control within the legal person.28 The sanctions include exclusion from entitlement to public benefits or aid, temporary or permanent exclusion from access to public funding, tender procedures, grants and concessions, temporary or permanent disqualification from commercial activity, among others.29 The Directive is to be transposed in the legal orders of the EU member states by December 3, 2020.30

Central and Eastern Europe The situation in Central and Eastern Europe (CEE) is manifold compared to the rest of Europe (e.g., Kubbe and Cvetanoska, 2021). Actual experience with corruption across the EU (5%, 2019 EU average) is far less common than the perception of corruption in general (71%, 2019 EU average) and the perception of the influence of corruption in the daily life of citizens (26%, 2019 EU average) in CEE. All countries are either at the EU average (Estonia and Latvia) or above the EU (5%). In the 2013–​2019 period, most of the countries registered a decreasing trend of actual experience with corruption.31 Hungary is the only country where the respondents perceive corruption as increased both at the national level and the regional/​local public institutions while Hungary and Bulgaria experienced an increase in the share of respondents personally affected by corruption in the 2013–​2019 period. Corruption continues to be perceived as widespread by CEE citizens. With the exception of Estonia, the SEB 502 report indicates that all countries in the region are well above the EU average (71% in 2019). Based on the respondents’ perceptions, it appears that the Baltic states perform better

Corruption in Europe  197 in controlling corruption compared with the middle such as Slovakia and the Czech Republic and other countries in the region including Hungary, Romania and Bulgaria. Thus, the Baltic countries and the Czech Republic are below the EU average in 2019 (42%), which implies that less than 42% of their citizens perceive an increase in the level of corruption in the last three years. The remaining countries are above the EU average, as more than 42% of their citizens perceive an increase in corruption in the last three years. Still, Estonia is the champion in the region due to the fact that it has managed to significantly drop perceived levels of corruption in just seven years. Bulgaria shows an opposite trend as the only country in the region that indicates an increasing trend from 41% in 2013 to 51% in 2019 as it is possible that corruption has become concentrated in specific domains such as the political sector, management of EU funds and public procurement in general, the healthcare system, the integrity of the judiciary, to mention a few, while there is also a clear lack of citizens’ trust in public institutions (SEB 502; Transparency International, 2021b). In all CEE countries a gap exists between the perception of corruption in general and the perception of the impact of corruption in daily life.The share of respondents that are personally affected by corruption is much smaller than that of those concerned with corruption in the national public institutions. Bulgaria (7%) and Hungary (13%) showed an increase of people personally affected by corruption during 2013–​2019. The rest of the countries in CEE indicate only small dynamics without perceivable trends (SEB 502). Most of the respondents consider the political sector as most affected by corruption. However, 53% in the analyzed Eastern Europe region point out corruption in political parties which is lower than the EU average (56%). In all countries at least 40% of respondents perceive widespread corruption within political parties. Also, in the CEE countries public procurement corruption or building permits corruption is perceived as a corrupt sector by at least 30% of respondents. This situation is similar in the case of private companies and banks and financial institutions. The second and third most corrupt sectors reported being the healthcare system (51%) and public procurement system (48%). In general, the situation seems to improve in all areas as fewer citizens perceive widespread corruption compared to the previous report (European Commission, 2021a). In other areas, such as healthcare, police and customs, substantial differences exist in the perceptions across countries. For example, corruption in healthcare is perceived more acutely in Lithuania (71%) or Bulgaria (53%) than in Estonia (20%). Also, corruption within police and customs is more problematic in Bulgaria (61%) than in Estonia (14%). The Bulgarian police and customs are perceived as the most corrupt public institutions in the country, at the highest level in the EU (SEB 502). In Lithuania, the public perception about the integrity of courts changed after the judicial system was affected by a massive corruption scandal with senior judges accused of acting as an organized group and accepting bribes in early 2019 (Reuters, 2019).

198  Ina Kubbe and Stoyan Panov Furthermore, the majority of the CEE countries show a decline in the perception that it is acceptable to give money for public services. This change is particularly observed in Lithuania (from 42% in 2013 to 23% in 2019), Latvia (from 38% to 25%) and Slovakia (from 29% to 20%). Romania and Hungary are the countries that have a reversed trend (Romania from 20% to 37%; Hungary from 39% to 43%). Estonia and Bulgaria are below the 2019 EU average (16%), meaning that fewer citizens of those countries consider giving money for public services acceptable (SEB 502). Moreover, although more respondents consider gifts as acceptable in relation to the public administration compared with the average in the EU, the levels of acceptance decreased in the region. Regarding the acceptance of doing favors in relation to the public sector, all CEE countries, except Estonia, are above the EU average. This means that more citizens in the region consider doing a favor an acceptable behavior in relation to the public administration, compared with the average number in the Union (SEB 502). In all countries, except Romania, the levels of acceptance of doing a favor dropped.32 The perception of this corruption risk tends to be constant over the 2013–​2019 period, yet more fluctuation is observed in Romania and the Czech Republic. Estonia is the only country where people perceive a lower risk of collusion between business and politics. Estonia continues to perform better than the rest of the countries in the region also when it comes to the need for political connections for succeeding in business. Only Bulgaria and Hungary remain above the 2019 European average (71%). Similar to North-​Western Europe, the majority of citizens agree that too close links between business and politics lead to corruption (except Estonia which is above the 2019 EU average). Bulgaria is the only country in the region where there is an increase in the number of people agreeing that the only way to succeed in business is to have political connections. This might be explained with endemic problems with enforcement against grand corruption and issues with the independence of the judiciary and prosecutorial services. Most of the citizens in the region agree that favoritism and corruption hamper business competition (SEB 502). Yet, some companies in the region do not consider gifts and little favors as corrupt or unethical behavior. Except in Estonia and Latvia, the share of citizens concerned by the entrenchment of corruption into the business sector is higher than the EU average (61%). Regarding actual experience with corruption, all countries in the Eastern Europe region are either at the EU average (Estonia and Latvia) or above the 5% EU average. Considering the entrenchment of corruption into the business sector, the findings are mixed. On the one hand, there are countries including Estonia, Lithuania, Latvia, Czech Republic and Slovakia where the number of citizens concerned with this risk has dropped. On the other hand, there are countries where the number of citizens concerned with this risk has increased such as in Hungary and Bulgaria. Only in Romania no perceivable trend exists according to the data.

Corruption in Europe  199 Experience with corruption in healthcare across the EU (5%, 2019 EU average) is far less common than the perception of corruption in healthcare (27%, 2019 EU average). In the CEE region, the picture is very diverse. Only the Czech Republic, Slovakia and Estonia are in line with the 2019 EU average or below it.The other countries experience constant problems with corruption in the healthcare sector. Romania is in first place (19% reported unofficial payments), followed closely by Hungary, Bulgaria, Lithuania and Latvia (SEB 502). In terms of law enforcement, the police are trusted to deal with corruption complaints by the majority of citizens in all countries except Lithuania, where the mass media is more trusted than the police. Mass media is also highly trusted in Slovakia, followed by the justice system including anticorruption agencies. Only half of the CEE countries established specialized anticorruption agencies and many citizens in these countries trust these agencies to a higher degree (Lithuania 29%, Latvia 32%, Bulgaria 18% and Romania 30%) than the justice sector (below 10%) (SEB 502). Poland, in particular, seems to be a special case with its weak institutions. Despite the ongoing European Commission’s proceedings regarding Poland’s infringements of the rule of law (European Commission, 2021c) and introduction of the regulation, designed to protect EU funds from being misused by EU governments that bend the rule of law (European Parliament, 2020b), the government in Warsaw ignores international recommendations. The European Commission’s demands are not met, while the government plans further reforms, aimed at gaining more control over the judiciary (Gałczyńska, 2021). The lack of independence of prosecution services and the process of judiciary capture by the ruling parties are problematic. A significant issue of the Polish anticorruption framework is the lack of a legally binding definition of the conflict of interest. Such a lack of conflict-​of-​interest provision may be a direct breach of the provision of conflict of interest in the implementation of the EU budget by national authorities.33 Considering the obstacles in reporting corruption cases, most of the citizens in Eastern Europe (45%) believe that people decide not to report a case of corruption because it is difficult to prove anything. For 35% reporting is pointless because those responsible will not be punished. The average perceptions in Eastern Europe are in line with the EU average. Regarding the social acceptance of corruption practices, the perception in Eastern Europe has improved over the last decades. In the last seven years, most of the countries in the region including Lithuania, Latvia, Slovakia, Estonia, Bulgaria and the Czech Republic have increased in the number of respondents considering it unacceptable to give money for public services.Yet, Romania and Hungary are the countries that have a reversed trend. The levels of acceptance of giving gifts to the public sector decreased in most of the countries such as in Lithuania, Slovakia, Latvia, Estonia, Hungary and Bulgaria, while Romania and the Czech Republic have a reversed trend from 2013 onwards. In all CEE

200  Ina Kubbe and Stoyan Panov countries, except Romania, the level of acceptance of doing a favor dropped in the last seven years (SEB 502). In general, significant differences within the CEE region exist regarding the perceived level of corruption –​with worse trends in Romania and improvements in Estonia, despite current corruption scandals in the country (Henley, 2021). In particular, the judiciary seems more vulnerable to corruption in the CEE (such as in Poland, Lithuania) due to a lack of legal regulations and concrete definitions, a deterioration of the rule of law and declining institutional trust, compared to North-​Western Europe. The EPPO As outlined in the previous section, the situation in CEE is vastly heterogeneous. The region can be classified as experiencing structural transformations and institutional changes in anticorruption policy and efforts. In that line, it is appropriate to reflect on two EU mechanisms with institutional anticorruption relevance. One of the recurring themes in CEE is the low trust in national institutions and the perception that reporting corruption is futile as national institutions are not effective or independent. A possible answer on EU level can be found in the recently established EPPO.34 The significance of the establishment of the EPPO is in the structure, function and competence in terms of its antifraud and anticorruption direction and indirect influence on the rule of law in the participating member states.35 It is the first EU body with powers to adopt decisions against individuals within the field of criminal law (Giuffrida, 2017, p.1). It is fully functioning for the participating 22 EU member states as of 1 June 2021 (European Commission, 2021b). The EPPO structure is characterized as “an indivisible Union body operating as one single Office” with a decentralized structure and it shall be organized at central and decentralized levels.36 The European Chief Prosecutor is the EPPO head, with responsibilities to organize the work of the EPPO, to guide its activities and to take decisions in accordance with the Regulation and the relevant rules of procedure.37 The College, consisting of the European Chief Prosecutor (ECP) and one European Prosecutor (EP) from each participating member state, is tasked with the general oversight of the EPPO activities and with taking decisions of strategic matters with focus on ensuring coherence, efficiency and consistency in the EPPO prosecution policy.38 The EP, responsible for the supervision of the investigation or prosecution, also participates in the deliberation of the Permanent Chambers.39 The EP supervises the investigations and prosecutions of the European Delegated Prosecutor (EDP) handling the case in the relevant member state(s) of origin.40 The complex structure of the EPPO is noticeable at the decentralized level around at least two European Delegated Prosecutors in each member state.41 The EDPs act on behalf of the EPPO in their respective member states and have the same powers as national prosecutors in investigations, prosecutions and bringing cases to judgment along with their

Corruption in Europe  201 specific powers.42 The EDPs have a “dual hat” function as they act as national prosecutors as long as the two functions do not conflict with their obligations under the EPPO Regulation (Ligeti and Simonato, 2013, p.15).43 The European moment is found in the ability of the national judicial and law enforcement to notify without undue delay the EPPO about investigations of offenses within the competence of the EPPO in order for the EPPO to determine whether to exercise the evocation privilege.44 The EPPO’s task is to “investigate, prosecute and bring to judgment the perpetrators of offenses against the Union’s financial interests” in pursuance of Directive 2017/​1371 (Giuffrida, 2017, p.5). The material competence scope is extended per Article 22(3) of the EPPO Regulation because the EPPO has competence for “any other criminal offence” that is “inextricably linked” to the PIF criminal conduct under Article 22(1). The efficient investigation of the PIF offenses may require “an extension of the investigation to other offences under national law,” inextricably linked to the PIF offense(s). The qualification “inextricably linked” is assessed in line of the relevant case law with a focus on the relevant criterion of the material facts or substantially the same facts. The PIF offense must carry a higher maximum sanction than the ancillary offense as well as the ancillary nature of the offense which is instrumental to the PIF offense where the main purpose of the ancillary crime is to create the conditions to commit the PIF offense. In this manner, the ancillary offense can encompass various acts or omissions related to corruption. The EPPO started functioning on June 1, 2021. There are hopes that the competence and structure of the EPPO would allow for more proactive investigations in some EU member states where the trust in the investigative and prosecutorial authorities is low or compromised as observed in the perceptions above. In November 2021, the first conviction in an EPPO case was issued in Slovakia against a former mayor for committing an attempted offense against the financial interests of the EU (EPPO, 2021). Protection of the General Regime of Conditionality for the Protection of the EU Budget and Breaches of the Rule-​of-​Law Mechanism The European Commission has been active in another aspect related to the protection of the financial interests of the EU, that is, with regard to graft and improper tender procedures with EU budget funds. In May 2018, the Commission issued a proposal for a Regulation for the Protection of the EU’s Budget in Case of Deficiencies of the Rule of Law in the EU Member States. The link between the rule of law and the financial interests of the Union is that the former is a prerequisite for the confidence of EU spending.45 At the July 2020 EU Summit, the EU member states committed to passing a historically high Multiannual Financial Framework as well as a recovery pandemic fund Next Generation EU in an effort to respond to the devastating effect on the EU economy as a result of the COVID-​19 pandemic. One of the main points of contention is the conditionality mechanism which would be triggered

202  Ina Kubbe and Stoyan Panov when breaches of the rule of law take place in EU member states (European Parliament, 2021c; EUObserver, 2020). The main reason for the regulation is the recent backsliding in domestic checks and balances in cases such as Hungary and Poland, as well as the disbursement and conditionality mechanism of the EU budget and the COVID-​ 19 recovery fund in loans and grants to EU member states (Kirst, 2021, pp.101–​110). Such regress has affected the rule of law in some member states, which poses serious concerns for the whole Union. Additionally, the mechanism builds on the efforts of the Commission to integrate the rule-​of-​law evaluations as well as the Cooperation and Verification Mechanisms in the cases of Romania and Bulgaria. Hence, the mechanism can be seen as a logical continuation of the line of institutional reforms and further integration on the European level. The mechanism allows the EU institutions to complement the existing mechanisms, such as the criminal law–​oriented EPPO, in order to ensure effective control and respect of the rule of law and protect the EU’s financial interests.46 The breaches of the principles of the rule of law are focused on endangering the independence of the judiciary, failing to prevent, correct and sanction arbitrary or unlawful decisions by public authorities, withholding financial resources affecting their proper functioning, and limiting the availability and effectiveness of legal remedies, including limitations on effective investigation and prosecution.47 The measures are to be implemented when the breaches of the principles of the rule of law “affect or seriously risk affecting the sound financial management of the EU budget or the protection of the financial interests of the Union in a sufficiently direct way.”48 If the Commission determines that the breaches of the rule of law have occurred and the remedial measures by the breaching member state do not suffice after communicating with the affected member state per the procedure of Article 6 of Regulation 2020/​2092, the Commission renders a sanctioning proposal, sent to the Council, which on its side adopts a decision within a month from receiving the proposal. The Council may amend the Commission’s proposal, acting by qualified majority voting.49 The measures include suspension of payments, a suspension of disbursement of installments in full or in part or an early repayment of loans guaranteed by the Union budget, a suspension or reduction of the economic advantage under an instrument guaranteed by the Union budget, or a prohibition to enter in new agreements on loans.50 The consolidated version of Regulation 2020/​2092 was formalized in mid-​ December 2020 by using the ordinary legislative procedures and corresponding votes at the Council and European Parliament. Hungary and Poland voted against the proposal after lifting their veto which was imposed on the EU budget and the recovery fund at the EU Summit on December 11–​12, 2020. The reached compromise by the EU member states in the form of Regulation 2020/​2092, described above, was reviewed by the Court of Justice of the EU (von der Burchard, 2020). The reached compromise by the EU member states

Corruption in Europe  203 in the form of Regulation 2020/​2092, described above, successfully passed the review by the Court of Justice of the EU.

Southern Europe According to the SEB 502 data, the majority of the Southern European (SE) countries perform better with respect to the levels of acceptability of corruption than the EU average. This result relates to all three scores (23% always /​sometimes acceptable for doing a favor and giving a gift and 16% for giving money). Only a small number of countries including Portugal, Croatia and Spain, stand out for positive developments since the Special Eurobarometer 470 (2017). With 8%, Portugal shows the lowest level of acceptability of corruption within the EU. All countries, except France, consider corruption to be widespread at levels above the EU average. Similar to countries in the North and Central and Eastern Europe, political parties are among the three most frequently mentioned institutions in all SE countries, “politicians” in all, except Greece and Cyprus, while “officials awarding public tenders” only in Croatia, Italy, Portugal and Slovenia (SEB 502). All countries, with the exception of France, show that the perception of being personally affected by corruption at levels is above the EU average. Cyprus, Portugal, Spain and Greece are among the five countries that are most affected. Looking at developments since 2017, the situation has significantly deteriorated in Malta, Portugal, Greece and Cyprus, yet to a lesser extent in France and Slovenia. No significant change has taken place in Spain and Italy. The deterioration in Malta and Portugal can be explained by the political scandal surrounding the murder of journalist Daphne Caruana (BBC, 2021). All SE countries –​except France and Greece –​consider the situation to have deteriorated at levels above the EU average over the last three years. Malta (74%), Cyprus and Croatia (69%) and Slovenia (58%) are the countries that report the highest deterioration –​compared with the EU average of 42%. On corruption in local or regional public institutions, all SE countries agree on the presence of corruption at levels above the EU average (68%).The ranking is actually led by six of the countries including Croatia (93%), Portugal (90%), Spain and Greece (89%), Cyprus (88%) and Italy (82%). Similarly, regarding national public institutions, all SE countries agree on the presence of corruption at levels above the EU average (70%). With regard to the financing of political parties, all SE countries disagree on sufficient transparency and supervision being in place at levels that are either equal (for Italy, 55%) or higher than the EU average. On bribery and the use of connections as the easiest way to obtain public services, respondents in all SE countries agree at levels above the EU average (64%). Four SE countries are among the five most affected societies (Cyprus, 91%; Greece, 90%; Croatia, 88% and Slovenia, 85%). Concerning the statement on corruption being part of the business culture, all SE countries, bar France (57%), agree on this at levels above the EU average (61%).

204  Ina Kubbe and Stoyan Panov Similarly, regarding the statement on the links between business and politics leading to corruption, all countries, except France, agree on this at levels above the EU average. Moreover, the SE countries agree that the only way to succeed in business is to have political connections at levels above the EU average (51%), with the exception of France. Four SE countries rank among the top five (Cyprus, 87%; Croatia, 80%; Greece, 76%; Slovenia, 74%). On favoritism and corruption hampering business competition, all SE countries agree on this at levels above the EU average (63%): Portugal (87%), Greece (85%), Spain (84%) and Slovenia (81%) are among the top five. Turning to perceptions about how countries have dealt with corruption, and starting with successful prosecutions, all SE countries rather disagree than agree. Disagreement is particularly pronounced in Slovenia (72%), Cyprus (70%), Croatia (70%), with no other EU countries pronouncing disagreement above 63% (EU average of 50%). Concerning the statement that government efforts to combat corruption are effective, all SE countries rather disagree at levels above the EU average (53% disagree). On the question of personally knowing anyone who takes or has taken bribes, all SE countries, bar Italy (7%) and Portugal (7%), answer in the affirmative at levels above the EU average (11%). Two SE countries count among the top five, including Croatia (27%) and Greece (25%). Looking more specifically at bribery and health care and the question on “extra payment or a valuable gift … or make a donation,” only three SE countries answer in the affirmative at levels above the EU average (5%), namely Greece (14%), Croatia (7%) and France (5%). It is worth noting that Cyprus, Spain and Portugal (all at 2%) count among the countries with the lowest levels of affirmative responses, only surpassed by the Netherlands, Finland (both 1%) and Sweden (0%). On the question of having experienced or witnessed a case of corruption, all SE countries, bar Portugal (3%), answer in the affirmative at levels above the EU average (5%). The ranking is led by Croatia (15%); the top ten also includes Slovenia (10%) and Cyprus and Greece (both at 9%). On knowledge as to where to report corruption, Greece (65%), Portugal (55%), Spain (53%) and Italy (49%) respond rather “yes” than “no” (Greece, Portugal and Spain rank among the top five together with Finland and Latvia), while it is the opposite for Slovenia (48%), Cyprus (46%), France (42%), Croatia (40%) and Malta (33%) in comparison to the EU average of 44% “yes” and 53% “no.” Looking at the reasons for not reporting corruption, most SE countries largely coincide with perceptions at the EU level on the most frequently mentioned topics. “Difficult to prove anything” is among the three most frequently mentioned items in all SE countries, “reporting it would be pointless …” in all except France, and “there is no protection …” in all SE states, except Greece. Turning to trust in authorities to report corruption to, most SE countries largely coincide with perceptions at the EU level on the most frequently mentioned items. Interestingly, the police are the most frequently mentioned institution that can be trusted in the SE countries; the judicial sector is among

Corruption in Europe  205 the three most frequently mentioned items in all SE counties, except Croatia, Cyprus, Malta and Slovenia; and the National Ombudsman is among the three most frequently mentioned institutions in Greece, Spain, Cyprus, Malta and Slovenia. The SE countries seem rather homogenous in their responses, except France. Whistleblowing Protection: Directive 2019/​1937 of 23 October 2019, and Facilitation of Financial Information and Cooperation: Directive (EU) 2019/​1153 A noticeable divergent practice in some European countries from Southern Europe is the issue of whistleblower protection which is essential for the investigation of corruption. The EU has legislated in the area by passing Directive 2019/​1937, which deals with protection of reports of breaches of EU law that are harmful to the public interest. The purpose of the Directive is “to enhance the enforcement of Union law and policies…by laying down common minimum standards providing for a high level of protection of persons reporting breaches of Union law.”51 The scope of the breaches includes breaches of the financial interests of the EU under Article 325 TFEU.52 The Directive lays down various mechanisms of protection of whistleblowers such as internal reporting, external reporting, public disclosures and various protection measures, among others. Finally, one of the latest attempts on the EU level focuses on facilitating the use of financial information in prevention, detection, investigation and prosecution of serious crimes such as corruption. The aim is to enhance the cooperation and access to information by Financial Intelligence Units (FIUs) in order to improve combating money laundering, financial crimes and prevention of tax crimes in the EU.53 Especially relevant is the access and use of financial information and bank account information by the competent authorities.54 Corruption is listed as a serious criminal offense, following the categorization in Annex to Regulation 2016/​794. The crux of the directive is the enhanced access to and search of bank account information by the competent authorities of member states for the purposes of prevention, investigation or prosecution of serious crimes, including the identification, tracing and freezing of assets.55 This facilitation is done primarily through the obligation of national FIUs to cooperate with other competent authorities to reasoned requests for financial information or analysis when there is an ongoing detection, investigation or prosecution of serious criminal offenses including corruption.56 The Directive is to be transposed by August 1, 2021.57

Discussion and Conclusion: The Devil Is in the Detail Corruption varies widely across European countries, but on the continental level in Europe corruption is still approached as a monolith, which results in a lack of context-​specific approaches.What is noticeable across board in the three

206  Ina Kubbe and Stoyan Panov European regions by analyzing the data collected for the purposes of this study is that there is an overall deterioration in the perception of corruption as well as of anticorruption norms. While most legislative efforts on the EU level are either institutional-​oriented or specifically linked to financial crimes of the EU funds, the broader anticorruption framework may need to be strengthened on the European level. The European Commission has been active in linking the corruption perceptions across the EU, anticorruption strategies in the member states as well as outlining the necessary reforms to strengthen the capacity to fight corruption with the rule-​of-​law situation in all member states in its rule-​ of-​law report (European Commission 2021d, 10–​15). The business sector is also the focus of corruption with respect to its influence on public administration and political parties and government. In that respect, the introduction of the EPPO is a welcome development along with private sector corruption initiatives. Lobbying regulations also vary widely across EU member states and need a common strategy. The political sector is perceived to be most affected by corruption which begs the question of whether the EU should be more involved through political means in the backsliding of democracy and rule of law which are inherently linked to grand corruption and state capture (Kubbe and Cvetanoska, 2021). One positive development is the mechanism to protect the EU budget when rule-​ of-​ law backsliding is observed in member states. Public procurement problems and instances of bribery are classic examples of corruption where the EU has provided the legal framework but the issue remains in the implementation of the legal norms in the respective member states. In particular, in the northern part of Europe, collusive bidding is on the rise in procurement procedures (SEB 502). Furthermore, the findings indicate that the EU comes short when addressing sector-​specific corruption such as graft in the healthcare sector. This may be due to the micro-​level acts and the presumption that such instances of sector-​ specific corruption are better handled on a national level. Nonetheless, the increase of corruption in the healthcare sector may necessitate a stronger supranational focus, especially in times of the COVID pandemic.Therefore, this study recommends for each region to include more efficient anticorruption programs into the healthcare sector), regulations of combining private and public practices in the healthcare sector and an improvement of working conditions and remuneration for doctors and nurses and all supporting personnel to prevent and lower corruption and informal practices in short and long term. There are also noticeable issues with the reporting of corruption.Throughout all regions of the EU, whistleblowing is often considered as a taboo and has a negative reputation, either due to a lack of trust in the punishment system or as perceiving it as a form of blackening. Therefore, reporting is undermined and perpetrators are not sanctioned. Thus, the whistleblower protection legislation, as well as the EPPO, are welcomed anticorruption developments. However, most EU member states are seriously lagging in transposing the EU Directive on Whistleblower Protection, but these delays must not be used as an excuse to

Corruption in Europe  207 take shortcuts that undermine transparency and inclusiveness. EU institutions can also alleviate the decreasing trust in national anticorruption strategies. There are recurring issues throughout the EU such as a noticeable need for capacity building of the criminal justice systems in order to fight corruption and implement the EU anticorruption framework along with national law. Challenges in cross-​border investigations of corruption remain and the EPPO will need to play an essential role to respond to this issue. Moreover, the current anticorruption framework needs to be strengthened and updated in order to improve prevention and integrity checks throughout the EU member states. It is crucial to follow how the EU will respond to backsliding in democracy and rule of law and weakening of various institutions in some EU member states. Some sectors remain prone to corruption, such as the healthcare sector. Difficulties are also seen in connecting the dots between corruption and affected sectors in society such as the role of corruption in media freedom deterioration, independence of the judiciary or political party funding. One ponders whether corruption is the symptom or the cause in such complex situations. It is also worth examining whether the same methodology in measuring corruption is appropriate to use as the regions are fairly heterogeneous (even within the regions, there are some noticeable differences as illustrated above). Different member states or subregions experience some particular corruption-​related problems. The chapter also notes that most EU responses expectedly target the macro-​level instances of corruption, while surveys, based on citizen’s perceptions and experiences, also capture micro-​level corruption. Contextualization is important in such instances and the EU institutions should pay closer attention to the trends and patterns on micro level. Ultimately, the big question is how EU member states should be methodologically assessed. Although the same assessment criteria should be applied, there is also a need for context-​specific approaches that may include historical, cultural, political and economic circumstances in the member states and regions. Connecting the dots is an important task in order to appropriately determine what further legislative initiatives might be needed. Hence, it is difficult to label Europe as a monolith when it comes down to complex phenomena such as corruption. More specific analytics should be used by the EU and the member states as well as the corresponding fine-​tuned responses are necessary. If the EU is the doctor who needs to provide individualized treatment on a general assessment of all patients, it becomes difficult to understand how such treatment would indeed respond to the specific symptom. In the end, it is very often about every single detail in definitions of corrupt practices, measurements, strategies or law regulations to fight corruption in depth and in the long term. Nonetheless, perceptions in society do matter as even though they may not be currently matched by reality and existing mechanisms, such notions and impressions may have powerful negative effects on future development and efforts in anticorruption, in particular related to trust in institutions such as anticorruption agencies or the European Union (Transparency International, 2021b; Wanat and Bayer, 2021).

208  Ina Kubbe and Stoyan Panov

Notes 1 Corruption disturbs macroeconomic and fiscal stability, stunts economic growth by creating business uncertainty, exacerbates inflation and promotes social inequality and poverty. It violates the fundamental principles of democracy such as equality, fairness, transparency and accountability. 2 The Stockholm Programme establishes a framework for EU action on issues of citizenship, justice, security, asylum, immigration and visa policy for the period 2010–​14. 3 The Special Eurobarometer 502 looks, for example, at the acceptability of corruption on the basis of the criteria of doing a favor, giving a gift and giving money. 4 See country-​specific data in SEB 502. 5 About half of the respondents in the Netherlands still perceive tendering procedures and permit delivery to be the weakest procedures with widespread corruption. 6 The European Commission Directorate General Migration and Home Affairs commissioned in 2017 a study on corruption in the healthcare sector and identified six types of corrupt activities: (1) bribery in the medical service delivery; (2) procurement corruption; (3) improper marketing relations; (4) misuse of (high)-​ level positions; (5) undue reimbursement claims; (6) fraud and embezzlement of medicines and medical devices. 7 An increase of 8% compared to the previous surveys in 2017 (5%) and 2013 (5%). 8 Directive (EU) 2019/​ 1937 (the Whistleblower Directive) is still transposed in Finland. 9 Art 83(1) TFEU. 10 Art 325(1) TFEU. 11 Art 325(1) TFEU. 12 See Art 325(2) TFEU. 13 Art 1 Dir. 2017/​1371. 14 Art 3(1) Dir. 2017/​1371. 15 Art 3(2) Dir. 2017/​1371. Art 4 Dir. 2017/​1371 includes other criminal offences affecting the EU’s financial interests. 16 Art 8 Dir. 2017/​1371. 17 Art 15(1) Dir. 2017/​1371. 18 The Tampere Programme provided the framework for European integration in Justice and Home Affairs. 19 See pt. 9 of Preamble of Decision 2003/​568. 20 Art 1 Decision 2003/​568. 21 Art 4 Decision 2003/​568. 22 Art 6 Decision 2003/​568. 23 See pt. 5 of the preamble of Dir. 2018/​1673. 24 See pt. 14 of the preamble of Dir. 2018/​1673. See also Art 5 Dir. 2018/​1673. 25 Art 2(1)(h) Dir. 2018/​1673. 26 Art 3(1)(a)–​(c) Dir. 2018/​1673. 27 See Directive on criminal sanctions for insider dealing and market manipulation (Directive 2014/​57/​EU, L173/​179) and Regulation on market abuse (Regulation (EU) 596/​2014). 28 Art 7 Dir. 2018/​1673. 29 Art 8 Dir. 2018/​1673. 30 Art 13 Dir. 2018/​1673.

Corruption in Europe  209 31 Estonia –​2%, Lithuania –​17%, Latvia –​3%, Czech Republic –​6%, Bulgaria –​2%. In 2019, Romania (+​3%), Slovakia (+​1%) and Hungary (23%) showed an increase in 2019 of people who have actual experience with corruption compared to 2017 (SEB 502). 32 Romania registered once again a reverse trend, where the acceptance increased from 20% to 37% (SEB 502). 33 Art 61 of Regulation 2018/​1046 on the financial rules applicable to the general budget of the Union (July 18, 2018). 34 Council Regulation (EU) 2017/​1939, OJ L283 (October 12, 2017). 35 As of the current moment, all EU member states with the exception of Denmark, Ireland, Sweden, Poland and Hungary participate in the EPPO framework. 36 Art 8(2–​3) Regulation 2017/​1939. 37 Art 11(1) Regulation 2017/​1939. 38 Art 9(1–​2) Regulation 2017/​1939. 39 Art 10(9) Regulation 2017/​1939. 40 Art 12(1) Regulation 2017/​1939. 41 Art 13(2) Regulation 2017/​1939. 42 Art 13(1) Regulation 2017/​1939. 43 Art 13(3) Regulation 2017/​1939. 44 Art 24(2) Regulation 2017/​1939. 45 Proposal for a Regulation of the European Parliament and of the Council on the Protection of the Union’s Budget in Case of Generalized Deficiencies as regards the Rule of Law in the Member States (May 2, 2018) COM (2018) 324 final, 1. 46 Art 4(2) of Regulation (EU, Euratom) 2020/​2092 of the European Parliament and of the Council of December 16, 2020, on a general regime of conditionality for the protection of the Union budget. 47 Art 3 Regulation 2020/​2092. 48 Art 4 Regulation 2020/​2092. 49 Art 6(11) Regulation 2020/​2092. 50 Art 5 Regulation 2020/​2092. 51 Art 1 Dir. 2019/​1937. 52 See Art 2(1)(b) Dir. 2019/​1937. 53 Pt. 2 of Preamble Dir. 2019/​1153. 54 See Art 1 Dir. 2019/​1153. 55 Art 4 Dir. 2019/​1153. 56 Art 7(1) Dir. 2019/​1153. 57 Art 23 Dir. 2019/​1153.

References Altinget (2020, December 27) Fhv. Chief of Defense new lobbyist at Rud Pedersen. www.altin​get.dk/​navne​nyt/​fhv-​forsv​arsc​hef-​ny-​lobby​ist-​hos-​rud-​peder​sen. BBC News (2021, July 30) Daphne Caruana Galizia: Malta responsible for journalist death –​Inquiry. www.bbc.com/​news/​world-​eur​ope-​58012​903. CEPS (2020, May 12) 20 year anniversary of the Tampere Programme. www.ceps.eu/​ ceps-​publi​cati​ons/​20-​year-​anni​vers​ary-​of-​the-​tamp​ere-​progra​mme. Corruption Perception Index (2021, January 08 30) Index 2021. www.trans​pare​ncy. org/​en/​cpi/​2021.​

210  Ina Kubbe and Stoyan Panov Coun​cil Framework Decision 2003/​568/​JHA. Combating corruption in the private sector. Council of the European Union. http://​data.eur​opa.eu/​eli/​dec_​fr​amw/​ 2003/​568/​oj. Devitt, J. (2021, 18 March) Is Ireland a “safe haven” for the world’s dirty money? www. trans​pare​ncy.org/​en/​blog/​is-​irel​and-​a-​safe-​haven-​for-​the-​wor​lds-​dirty-​money?utm​ _​sou​rce=​new​slet​ter&utm​_​med​ium=​email&utm_​c​ampa​ign=​wee​kly-​19-​03-​2021. Directive 2017/​1371. Fight against fraud to the Union’s financial interests by means of criminal law. European Parliament, Council of the European Union. http://​data.eur​ opa.eu/​eli/​dir/​2017/​1371/​oj. Directive 2018/​ 1673. Combating money laundering by criminal law. European Parliament, Council of the European Union. http://​data.eur​opa.eu/​eli/​dir/​2018/​ 1673/​oj. Directive 2019/​1153. Rules facilitating the use of financial and other information for the prevention, detection, investigation or prosecution of certain criminal offences. European Parliament, Council of the European Union. http://​data.eur​opa.eu/​eli/​ dir/​2019/​1153/​oj. Directive 2019/​ 1937. Protection of persons who report breaches of Union law. European Parliament, Council of the European Union. http://​data.eur​opa.eu/​eli/​ dir/​2019/​1937/​oj. DW Germany (2021, March 11) German mask scandal: “Unforgivable violations of ethical standards.” www.dw.com/​en/​ger​man-​mask-​scan​dal-​unfor​g iva​ble-​vio​lati​ons-​ of-​ethi​cal-​standa​rds/​a-​56833​018. EUObserver (2020, November 6) Deal reached on linking EU funds to the rule of law. https://​euo​bser​ver.com/​instit​utio​nal/​149​967. European Commission (2021a, November 30) Corruption. https://​ec.eur​opa.eu/​ home-​affa​irs/​what-​we-​do/​polic​ies/​corrup​tion​_​en. European Commission (2021b, May 26) Protecting the EU budget: European Public Prosecutor’s Office will start operating on 1 June. https://​ec.eur​opa.eu/​com​miss​ ion/​pres​scor​ner/​det​ail/​en/​ip_​21_​2​591. European Commission (2021c, January 27) Rule of law: Commission adopts next step in the infringement procedure to protect judicial independence of Polish judges. https://​ec.eur​opa.eu/​com​miss​ion/​pres​scor​ner/​det​ail/​en/​ip_​21_​224. European Commission (2021d, July 20) 2021 Rule of Law Report: The Rule of Law Situation in the European Union. COM(2021) 700 final. European Parliament (2018) The Cum-​ex files –​Information document. www.europ​ arl.eur​opa.eu/​cmsd​ata/​158​435/​2018-​11-​26%20-​%20Info​r mat​ion%20pa​per%20 on%20Cum-​ex%20-​%20Cum-​cum.pdf. European Parliament (2020, December 16). Parliament approves the “rule of law conditionality” for access to EU funds. www.europ​arl.eur​opa.eu/​news/​en/​press-​room/​ 20201​211I​PR93​622/​par​liam​ent-​appro​ves-​the-​r ule-​of-​law-​con​diti​onal​ity-​for-​acc​ ess-​to-​eu-​funds. European Parliament (2021, March 29) Rule of law: New mechanism aims to protect EU budget and values. www.europ​arl.eur​opa.eu/​news/​en/​headli​nes/​eu-​affa​irs/​ 20201​001S​TO88​311/​r ule-​of-​law-​new-​mechan​ism-​aims-​to-​prot​ect-​eu-​bud​get-​ and-​val​ues. European Parliamentary Research Service (2021) Statute and funding of European political parties under Regulation 1141/​2014. European Parliamentary Research Service Ex-​ Post Evaluation Service. www.europ​arl.eur​opa.eu/​RegD​ata/​etu​des/​ STUD/​2021/​662​646/​EPRS_​STU(2021)662646​_​EN.pdf.

Corruption in Europe  211 European Public Prosecutor’s Office (2021, 23 November) First conviction in EPPO case: Former Slovak mayor sentenced to 3 years of conditional imprisonment. www. eppo.eur​opa.eu/​en/​news/​first-​con​vict​ion-​eppo-​case-​for​mer-​slo​vak-​mayor-​senten​ ced-​3-​years-​cond​itio​nal-​impri​sonm​ent. Gałczyńska M. (2021, January 3) Pięć lat rewolucji PiS w sądownictwie. Co jeszcze przed nami? https://​wia​domo​sci.onet.pl/​tylko-​w-​one​cie/​refo​r ma-​sado​wnic​twa-​ pis-​anal​iza-​magdy-​galcz​ynsk​iej/​0fpz​l16. Giuffrida, F. (2017) The European Public Prosecutor’s Office: King without kingdom? CEPS Research Report No 2017/​03. Henley, J. (2019, July 16) Dutch police are being infiltrated by criminal gangs, report says. www.theg​uard​ian.com/​world/​2019/​jul/​16/​dutch-​pol​ice-​crimi​nal-​gangs. Henley, J. (2021, January, 13) Estonian government collapses over corruption investigation. www.theg​uard​ian.com/​world/​2021/​jan/​13/​eston​ian-​gov​ernm​ent-​collap​ses over-​corruption-​investigation. Hoikkala, H. (2020, May 27) Sweden’s money laundering affair brings bankers closer to police. www.bnnbl​oomb​erg.ca/​swe​den-​s-​money-​lau​nder​ing-​aff​air-​bri​ngs-​bank​ers-​ clo​ser-​to-​pol​ice-​1.1441​978. Horizon 2020. DIGIWHIST –​The digital whistle-​blower: Fiscal transparency, risk assessment and impact of good governance policies assessed. https://​cor​dis.eur​opa. eu/​proj​ect/​id/​645​852. Hotten, R. (2015, December 10) Volkswagen: The scandal explained. www.bbc.com/​ news/​busin​ess-​34324​772. Kergueno, R., Aiossa Lucinda, N. A., Nuri, P., Corser, S., Teixeira, V. and van Hulten, M. (2021) Deep pockets, open doors: Big tech lobbying in Brussels. Transparency International EU. Pp. 1–​21. Kirst, N. (2021) Rule of law conditionality: The long-​awaited step towards a solution of the rule of law crisis in the European Union? European Papers, 6(1), 101–​110. Kubbe, I. & Cvetanoska, L. (2021) Corruption, democracy and (non-​development):The role of the European Union. In G. Crawford and A.-​G. Abdulai (Eds.), Handbook on Democracy and Development. Edward Elgar Publishing, pp. 293–​312. Ligeti, K. & Simonato M. (2013) The European Public Prosecutor’s Office: Towards a truly prosecution service? New Journal of European Criminal Law, 4(1–​2). pp. 7–​21. Muschel, R. (2020, 2 December) Mehr Betrugsfälle im Gesundheitswesen. www.rnz. de/​poli​tik/​suedw​est_​arti​kel,-​baden-​wuert​temb​erg-​mehr-​betrug​sfae​lle-​im-​gesun​ dhei​tswe​sen-​_​arid,589​032.html. Oltermann, P. (2019, 20 May) Austria’s “Ibiza scandal”: What happened and why does it matter? www.theg​uard​ian.com/​world/​2019/​may/​20/​aust​r ia-​ibiza-​scan​dal-​sting-​ operat​ion-​what-​happe​ned-​why-​does-​it-​mat​ter. Panov, S. (2019) The EU’s trifecta mechanisms: Analysis of EU’s response to the challenges to the rule of law and corruption. Kyiv-​Mohyla Law and Politics Journal, 5, 83–​117. Regulation 2017/​1939. Enhanced cooperation on the establishment of the European Public Prosecutor’s Office (“the EPPO”). Council of the European Union. http://​ data.eur​opa.eu/​eli/​reg/​2017/​1939/​oj. Regulation 2018/​1046. Financial rules applicable to the general budget of the Union. Council of the European Union. http://​data.eur​opa.eu/​eli/​reg/​2018/​1046/​oj. Regulation 2020/​2092. A general regime of conditionality for the protection of the Union budget. Council of the European Union. http://​data.eur​opa.eu/​eli/​reg/​ 2020/​2092/​oj.

212  Ina Kubbe and Stoyan Panov Reuters (2019, February 20) Lithuania arrests eight top judges in anti-​corruption crackdown. www.reut​ers.com/​arti​cle/​us-​lithua​nia-​cor​rupt​ion-​idUSKC​N1Q9​22O. Schubert, S. & Miller, T. C. (2008, December 20) At Siemens, bribery was just a line item. www.nyti​mes.com/​2008/​12/​21/​busin​ess/​worldb​usin​ess/​21siem​ens.html. Sommersguter-​ Reichmann, M. and Stepan, A. (2017) Hospital physician payment mechanisms in Austria: Do they provide gateways to institutional corruption? Health Economics Review, 7(11), 1–​13. Special Eurobarometer 470 (2017) Corruption. https://​data.eur​opa.eu/​data/​datas​ets/​ s21​76_​8​8_​2_​470_​eng?loc​ale=​en. Special Eurobarometer 502 (2021, November 29) Corruption. https://​data.eur​opa.eu/​ data/​datas​ets/​s22​47_​9​2_​4_​502_​eng?loc​ale=​en. Transparency International (2021a, November 29) What is corruption? www.trans​pare​ ncy.org/​en/​what-​is-​cor​rupt​ion. Transparency International (2021b, November 29) People see low political integrity throughout the European Union. www.trans​pare​ncy.org/​en/​news/​low-​politi​cal-​ integr​ity-​thr​ough​out-​the-​europ​ean-​union-​gcb-​eu-​2021. Trautvetter, C. & Henn, M. (2020) Transparency Register. No Transparency. A Research Report on Anonymity in Berlin’s Real Estate Market. Rosa Luxemburg Stiftung. von der Burchard, H. (2020, December 17) EU top court gears up for rule-​of-​law battle (of its life). Politico. www.polit​ico.eu/​arti​cle/​pol​and-​hung​ary-​rule-​of-​law-​court-​of-​ just​ice-​of-​the-​europ​ean-​union-​gears-​up-​bat​tle-​of-​its-​life. Wanat, Z. & Bayer, L. (2021, November 19) Brussels takes step toward rule-​of-​law penalty process with Poland, Hungary. www.polit​ico.eu/​arti​cle/​eu-​rule-​of-​law-​pena​ lty-​proc​ess-​pol​and-​hung​ary/​amp. World Bank (2021, November 29) CPIA transparency, accountability, and corruption in the public sector rating. https://​data.worldb​ank.org/​indica​tor/​IQ.CPA.TRAN. XQ?end=​2020&start=​2020&view=​map.

newgenrtpdf

Annex Comparison Table Eurobarometer Region

Types of corruption in the public sector

Types of corruption in the private sector

Norms/​informal practices

-​ Fast-​changing -​ Informal ways to legislation and obtain privileged policies information -​ Complex (Favoritism) administrative -​ Gift giving procedures to health care -​ Lack of party employees financing as gratitude -​ Weak regulations (“envelope regarding revolving medicine”) doors (Austria) -​ Taboo to report corruption-​ strong cultural norm “not to betray others” -​ > often leads to decision not to report corruption

Author’s recommendations -​ Increase of transparency in terms of party financing and party sponsoring -​ Strict regulations regarding revolving doors policies -​ Anti-​corruption programs in healthcare sector (especially urgent in light of COVID-​19) -​ Increase of global campaign and political will to communicate on corruption and to increase awareness of seemingly unconcerned population -​ Better guidelines how and where to report corruption -​ Providing anonymous reporting channels or adequate legal protection -​ Better whistleblower protection -​ Increase awareness toward corruption -​ Stronger role of investigative journalism (continued)

Corruption in Europe  213

North and -​ Favoritism -​ Close Western -​ Bribe-​taking and relationships in Europe power abuse in construction political parties, industry, banking, private companies real estate -​ Lobbying business, political -​ Institutional parties, between corruption police or politicians and municipalities business -​ Healthcare sector of -​ Quid pro quo concern practices in public tenders, building permits -​ Collusive bidding in procurement procedures -​ Money Laundering

Failures in institutions

newgenrtpdf

Types of corruption in the public sector

Types of corruption in the private sector

Central and Eastern Europe

-​ Misuse of EU funds - Conflict of -​ Bribery interest in private -​ Political sector sector, not as most affected sanctioned in by widespread some CEE states corruption -​ Healthcare system -​ Public procurement system -​ Building permits -​ Favoritism -​ Campaign donations for political parties by businesses (state capture) -​ Tax fraud,VAT non-​ payment, kickbacks -​ Systematic rent-​ seeking by public officials

Failures in institutions

Norms/​informal practices

Author’s recommendations

-​ Lack of integrity of judicial system -​ Varying lack of trust in public institutions -​ Trust in anticorruption agencies slightly higher than in judiciary -​ Wide variation of effectiveness of rule-​of-​law institutions -​ Lack of whistleblower protection (negative stereotypes) -​ Enforcement of grand corruption is problematic; hardest to catch offenders as political leaders exert influence over judiciary and prosecutors -​ Low institutional trust

-​ Money, gifts, favors to public sector officials -​ High tolerances of deviant behavior -​ Informal payments in healthcare sector -​ Lack of reporting corruption -​ Social acceptance of corruption is decreasing -​ some companies do not consider gifts and little favors as corrupt or unethical behavior.

-​ Anti-​corruption programs in healthcare sector (especially urgent in light of COVID-​19) -​ Regulations of combining private and public practices in healthcare sector -​ Improving working conditions and remuneration for doctors, nurses -​ Reforms in judiciary sector and improving and guaranteeing independence of judiciary -​ Increasing efficiency in anti-​ corruption investigations -​ Better implementation and application of anti-​corruption frameworks -​ Easing access for the general public to report corruption -​ Improving whistleblower protection -​ More transparent regulation of party financing -​ Improving conflict of interest enforcement -​ Enhanced role of EU through EPPO and Rule-​of-​Law Mechanism.

214  Ina Kubbe and Stoyan Panov

Region

Southern Europe

-​ Bribery -​ Political parties corruption -​ Political party financing -​ Favoritism -​ Golden Visas -​ Widespread local and regional public procurement problems.

-​ Corruption as part of business culture -​ Links between business and politics -​ Favoring friends and /​or family members in business

-​ Corruption in public institutions is higher than EU average -​ Weak law enforcement -​ Low trust that prosecution of corruption cases is successful -​ Government efforts to combat corruption unfavorable -​Reporting of corruption is heterogeneous in region -​ Lack of independent judiciary

-​ Reliance on connections to succeed in business -​ Politicization of corruption cases

-​ Better public corruption prevention -​ Enhanced responses to bribery -​ Better regulation of lobbyism and business influence on politics -​ Sector-​specific regulations such as in the healthcare sector -​ Increasing the trust in law enforcement authorities in anti-​ corruption investigations -​ Enhanced whistleblower protection -​ Targeted investigation and prosecution of unregulated links between business and politics

Corruption in Europe  215

14 Tackling Corruption Practical Perspectives Chandan Kumar Jha and Neelesh Kumar Sah

Introduction The General Assembly of the United Nations, concerned about the threats posed by corruption to the stability and security of societies, institutional and democratic strength, ethical values, rule of law and justice and sustainable development, adopted the United Nations Convention against Corruption (UNCAC) in 2003.1 It urged all States and competent regional economic integration organizations to sign and ratify the UNCAC. The UNCAC was a result of longstanding efforts on part of different stakeholders for many years, documented through different United Nations Resolutions. Its stated objectives are to promote international cooperation against corruption, strengthen measures to effectively fight corruption and enhance the integrity and accountability of public affairs and proper management of public property (Article 1, Chapter 1) (United Nations, 2004). While the fight against corruption has been in vogue since time immemorial, the adoption of the UNCAC reinvigorated efforts to tackle corruption at the global scale. Notably, the UNCAC does not define corruption but prescribes measures on how to deal with corruption. It also leaves scope to the participating nations to adapt the convention as per their specific conditions. Almost two decades have passed since the convention was adopted, yet corruption remains a global problem with citizens in developing countries being affected worst and facing corruption in almost all areas of government, including the police, courts and other public services such as school, utilities and hospitals. As per the Global Corruption Barometer (GCB),2 nearly 20 percent of those who used a public service in the last year paid a bribe in Asia (GCB Asia, 2020). The corresponding numbers are more than 25% for Africa (GCB, 2019) and approximately 33% in the Pacific region (GCB, 2021). It is therefore not surprising that researchers in various academic fields, including economics, sociology and political science, have studied the issue of corruption in many countries from different perspectives and have suggested various anticorruption policies. However, most of these policies have largely been unsuccessful as is evident by the extent of bribery and corruption captured by surveys done by DOI: 10.4324/9781003142300-14

Tackling Corruption  217 organizations such as the World Bank and Transparency International that publish corruption indices used by researchers. In this chapter, we take a different approach and look at the issue of corruption from a practical perspective. We argue that a successful fight against corruption requires more than a thorough understanding of corruption as a problem and the extent of its prevalence.To successfully fight corruption, we must be able to detect where it occurs and identify what conditions lead to opportunities for corrupt practices and sustain them over the long run. Therefore, it is imperative that we learn the processes involved in the provision of a public service. In this chapter, we describe the processes involved in the provision of a typical service provided by government agencies. We discuss how such processes are vulnerable to corrupt practices and why anticorruption measures do not produce desired results. We discuss potential ways to fix these processes to minimize the system’s vulnerability to corruption. We use a case study from the Tax Department of India designed to simplify the processes involved in tax filings. Using the case study, we make a critical assessment of information technology (IT)-​based systems in reducing corruption and the challenges that arise in building and instituting such systems.We conclude with lessons learned from past experiences and offer a few suggestions for the future.

Corruption: The Problem While corruption is universally recognized as a problem, finding an apt solution always presents a challenge for anyone, whether the government or an administrator. It is so because corruption manifests itself in different forms and magnitudes. Our focus in this chapter is not to define corruption but to identify how corrupt practices take place in the real world in day-​to-​day activities. Although crucial for tackling it, defining corruption is not enough from a practical perspective because corrupt practices –​as typically defined –​are not easily observable and even harder to prove in most cases. Consider, for instance, the most commonly used definitions of corruption in academic research: “the abuse of entrusted power for private gain” by the leading anticorruption organization Transparency International3 or the “use of public office for private gain” by the World Bank.4 In most cases, detecting that a personal gain was made by abusing the entrusted authority is very difficult, if not impossible. It is even more challenging to prove the occurrence of bribery and establish its blame on a specific individual. There are no consequences for the corrupt officials if the incidence of bribery cannot be proved in a court of law. Consequently, there is no incentive for a government official to abstain from corrupt practices.5 The first step in fighting corruption successfully, therefore, is the ability to detect the abuse of authority at each stage of the public service provision. To the extent that corruption involves deriving personal gains through the abuse of the entrusted authority, the focus has to be on identifying and tackling the abuse of authority. The abuse of authority manifests itself in different ways.6 We

218  Chandan Kumar Jha and Neelesh Kumar Sah identify and discuss various parameters that influence the abuse of the entrusted authority.

Public Service Provision: The Process Authority is the responsibility bestowed upon an individual to carry out a task. Tasks are nothing but part of a process where various tasks lead to the accomplishment of an objective. To understand this, let us look at any service that the public authorities dispense, say, the issue of licenses. A license is basically a permit to carry out a specified activity. Various activities require a license. For instance, driving a vehicle, carrying out certain professions like those of a lawyer, a physician, an architect or running a business. Each class of licenses essentially goes through a system of application, evaluation of the application, an examination of the applicant and payment of requisite fees (in some cases). So, there is a process flow involved in issuing licenses. In many cases, there are too many processes requiring multiple points of contacts, making it very difficult for a single person (the beneficiary) to understand and go through all these processes. Moreover, the official handling the case often has some discretion in the process. Using this discretionary power, the public official has the ability to delay the delivery of the service. The delay in the provision of the service is costly to the beneficiary. The official, being empowered with this knowledge, may choose to extract rents from the applicant. The rent extraction process often involves a throbbing system of outside (nongovernment) agents, also known as middlemen. These agents emerge in the form of solution providers and, despite being unauthorized, can get the work done for additional fees.The existence of these middlemen works well for service-​seekers because they get a single window resolution to their requirements. Authority shares a part of the additional fees charged by the agents and, in return, provides the services in a timely manner. In other cases, service-​seekers may try to influence the officials’ decision by bribing, especially if the decision is not expected to be in their favor.7 For instance, an individual seeking a driver license might offer a bribe if he/​she does not know how to drive. A beneficiary might offer a bribe to speed up the process if collecting required documents is costly and cumbersome. A business might use bribery to avoid taxes and regulations or to secure a government contract.8 In such cases, the middlemen (agents) exist who get the work done even when the requirements for the service is not fulfilled by the beneficiary.9 Of course, there is an additional fee for that, which is shared by the government official and the middlemen.

Factors Leading to the Emergence and Persistence of Corruption The previous section outlines how a simple process of providing a service like issuing a license gets mired in corruption. Let’s rewind a bit and try to see why we land in a corrupt system.10

Tackling Corruption  219 The system involves a process that could be long and cumbersome for the user. The application process might be difficult to understand and follow through with. Finally, the application system might be difficult to access and nonresponsive. To avoid turning into a corrupt system, at the minimum, these factors need to be addressed. This can be done by addressing the nature of the process for getting the work done, keeping in mind the needs of the end user. A few other aspects that make a system vulnerable to abuse are discretion available in decision-​ making at each stage –​at the time of filing of an application, processing of the application, or the disposal of the application. Those handling the stages of process flow could abuse their authority at every stage of the process. The absence of suitable grievance mechanisms makes the actors in the system unaccountable to the requirements of due performance. Redressal of grievances plays a major role in building citizen-​friendly, responsive administration.11 Weak internal controls further make the system vulnerable to abuse. Internal controls are a combination of Control Environment, Risk Assessment, Control Activities, Information and Communication and Monitoring, as laid out by the Committee of Sponsoring Organizations of the Treadway Commission (COSO).12 COSO also outlines seventeen principles against these five components of internal control. Weak internal controls exemplified through ineffective implementation of one or more of these principles cause the system to be vulnerable to abuse.Weak internal controls can be compounded by a weak oversight mechanism, like, internal and external audits, ombudsman, grievance mechanisms, vigilance, anticorruption bodies, regulators and parliamentary oversight. These oversight mechanisms, when functioning from outside of the organization, are known as external controls. Absence of such oversight mechanisms or such oversight mechanisms being weak indicate problems in overall governance of the systems and organizations.13

Addressing the Problem In the previous section, we outlined and discussed some of the features of a system that make it vulnerable to corruption. In this section, we discuss how these features can be modified to reduce the endemic of corruption.

Process Simplification Process simplification is an approach through which various processes are to be simplified. The process simplification entails reductions in the number of steps and the number of processes to complete a transaction or a job. Processes emanate from some legislation. A maze of legislations with their associated rules and regulations lead to multiple processes that can prove to be cumbersome. Complicated and cumbersome processes involve excess time and money, creating a sense of anguish for the citizen complying with the rules. Another aspect that influences processes is the clarity in roles and responsibilities of every entity involved in the process and the type of information processed at every step.

220  Chandan Kumar Jha and Neelesh Kumar Sah Unclear roles and responsibilities lead to duplication and redundant processes being in place. This makes the process inefficient and, at the same time, cumbersome for citizens. Since the advent of computers for the last two to three decades, this area has received a lot of attention and focus. Many processes have been simplified and made accessible to the end-​users. Some of the examples, from India, are the tax administration, land record computerization, railway reservation system, passport issuance and many other services.14 The process simplification does not only reduce the system’s vulnerability to corrupt practices but also promotes the compliance. For example, the simplification of tedious tax return filing process in India led to the number of returns filed in the country to increase from a range of around 34 million in 2009–​10 to about 64 million in 2019–​20.15

Reducing/​Eliminating Discretion Discretion is one of the most potent reasons for inducement to corrupt practices. The discretion to take a decision one way or the other, can be abused to extract bribes from the end-​users.16 Discretion in a process arises when multiple options to a decision are available to the decision maker, the administrative authority. It cannot be wished away as it is impossible to lay down rules, regulations and processes for all possibilities and eventualities for a decision. Since it is a necessary “evil,” it is important to figure out how to address it. It is for the procedures to ensure that the abuse of discretion does not go unnoticed.17 The gaze on all decisions involving discretion has the potential of reducing its abuse. While eliminating discretion altogether might not be feasible in most cases, serious effort should be placed in minimizing the discretion in the decision-​making processes. Some of key considerations with respect to discretion, as outlined by the Ombudsman Western Australia in their guidelines to decision-​making, can be listed as below:18 • • • •

•

Administrative procedures should be followed: While exercising discretion, there may be conditions that some consultations are required, or some basis is to be applied. Relevant information to be relied upon: The discretionary decision should rely on necessary information by collecting the same, if required. Act without bias: The decision maker must act in fairness and without a bias. Consistency in applying discretion: One of the elements in ensuring that the discretionary decision is unbiased is the consistency of the decision taken. Change of stand and inconsistency leads to the doubt in the fairness of the discretion exercised. Transparency in decision-​making: While the discretionary decision may be essential part of a process, transparency makes the abuse of the process less likely. It should be endeavored to make the discretionary decisions available readily.19

Tackling Corruption  221 •

Maintenance of records: Discretionary decisions should be recorded and kept for future reference. It may also include information on the facts considered for making the decision.

Interface of Officials with the Citizen An interface is an essential requirement for transactions to occur between any government department and the citizen. A physical interface between the officials and the citizen provides an opportunity for a direct interaction between the two, giving rise to the possibility of an exchange of bribe. Eliminating the need for a direct interaction between the government official and the citizen will reduce opportunities to engage into corrupt practices. One of the ways to reduce the direct interaction between the official and the service-​seeker is to ensure that there is a single window for the provision of services to the citizen. A single window or point of contact ensures that the beneficiaries do not have to follow their case from one desk to the other, exposing them to the possibility of corrupt transactions multiple times. If such a single window is virtual then the possibility of corrupt transactions is greatly minimized. Various government agencies around the world have recently been using technology for the purpose of reducing and/​or eliminating the need for a physical interaction between the citizens and the government officials. For instance, the Income Tax Department in India has adopted a faceless assessment system, eliminating the need for interface between the taxpayers and the tax officials for the scrutiny of their assessment. The system has been made faceless and the allocation of cases for scrutiny assessment has been randomized with layers of work allocation mechanisms and review and control mechanisms. As such, it is now very difficult for the tax officials to carry out compromised assessments. As a result, opportunities for taxpayer’s harassment has significantly declined. Similarly, a railway ticket can be bought online without any physical interface. Land record transactions like getting the land record details, payment of land revenue etc. are increasingly done online.The use of technology has allowed the government agencies to provide various services virtually –​a process known as electronic government or e-​government in short –​leading to significant declines in corrupt practices. The use of the e-​government has simplified the process of the provision of public services and reduced the discretion of the officials and promoted accountability, leading to a decline in corruption (see Chapter 12; Jha and Sarangi (2022)).20

Lax Oversight Mechanisms Oversight mechanisms ensure adherence to rules and regulations. They also ensure that the systems and processes function properly and exceptions, if any, are identified. The oversight mechanisms could be internal or external. The internal oversight starts from the monitoring and control activities whereas the external oversight encompasses the legislative or parliamentary, the independent

222  Chandan Kumar Jha and Neelesh Kumar Sah bodies (like the audit offices, human right institution, ombudsmen, etc.), or the civil society itself. The oversight mechanisms, when they function well and in tandem while ensuring adherence to rules and regulations, assure the citizenry of accountability, absence of corrupt/​divergent practices, illegality and unfair practices.21 The absence of such mechanisms or a lax mechanism, therefore, brings about the nonaccountability of the system and its operatives toward the organization, government and citizens. It leads to a system where due process cannot be assured. Consequently, there is an inherent tendency to engage in divergent/​ unfair and corrupt practices. Lax oversight mechanisms could lead to a situation where the operatives of the system may create a situation where the routine, rule-​based system fails or is crippled and they can maneuver it in the manner they wish, facilitating their rent-​seeking behavior. It will therefore be essential that the elements of oversight are inherently intertwined with any system and oversight mechanisms are implemented and monitored closely. The internal oversight mechanisms of monitoring and control are typically adequately addressed in IT-​ based systems. Systems of internal audit could be good means of ensuring accountability and adherence to prescribed rules and regulations. The system of internal audits makes the organizations self-​sufficient and self-​driven toward ensuring internal accountability. The systems of external oversight mechanism play an important role in ensuring the third party, unbiased, assurance on the organizations’ functioning. Independent institutions like the Regulators, Supreme Audit Institutions, Vigilance Institutions and Human Rights Institutions add to this system of assurance. They also act as measures of a deterrent for the operatives to engage in divergent/​ unfair/​ corrupt behaviors and contribute toward making the system robust and responsive.

Difficult Recourse to Grievances Grievances are obvious fallouts of working in any organization. A responsive organization puts mechanisms in place to pick up grievances and redress them. To address grievances effectively, it is important to have a robust mechanism. A system of identification/​receipt of grievances, enumerating them and a system of follow-​up on the grievances that ensures that each and every grievance is duly addressed are signs of a good grievance mechanism. A corrupt or unfair system will thrive in the absence of a difficult-​to-​access grievance mechanism. This, like the absence of oversight mechanisms, emboldens the operatives acting in a corrupt manner because the likelihood of any repercussions for their corrupt practices is extremely small. Clearly, operatives engaging in unfair and corrupt practices will have an interest in ensuring that the grievance redressal and oversight mechanisms, as well as internal controls of the organization, become dysfunctional. In organizations, today, simple grievance interfaces supported by elaborate redressal mechanisms could be put in place through the use of IT-​based

Tackling Corruption  223 grievance systems. An example of such a system is Centralized Public Grievance Redress and Monitoring System (CPGRAMS) in India. CPGRAMS provides an online platform for citizens to register their grievances regarding service delivery that are resolved through a close monitoring mechanism at the highest levels.22 The existence of effective grievance mechanisms can promote the accountability of the government systems, enhance governance quality and foster confidence in the citizenry about the government systems.23

Lack of Information on the Organization and Its Working Adequate information is essential to making decisions and crucial for seeking any service. Information on organizational structure, processes, its operatives, that is, the officials and their responsibilities, would enable any interaction of an outsider with the organization simple and smooth.The absence of this information has potential for anxiety and behavior on part of the service-​seeker, which may not be in order. Further, the lack of information on the organization and its workings makes its functioning opaque. As such, deficient information can have twin impacts. First, the outsider (the service-​seeker) is not assured of an adequate response and thus may be tempted to solicit favors or services through unfair means. Second, the operatives of the system, the government officials, aware of the lack of information to the outsider, can engage in rent-​seeking.24 The simplest solution, in this case, is that all information is made public and easily accessible. While building systems for such information sharing has the potential of burdening the organization, it is a great way for providing assurance to beneficiaries and promoting conformance to rules and regulations.The result would be lower corruption and a just and fair system. Many government bodies around the world have made significant progress in this direction. The Income Tax Department in India has brought to the fore a lot of information, be it the information on tax return filing, tax return processing, tax payments, refunds, etcetera. A taxpayer has a lot more information available to be able to duly file their tax return. Now, the Income Tax Department provides a prefilled tax return based on information available with the department and the taxpayer needs to only confirm the details while filing their return. This simple mechanism ensures that no income is hidden by the taxpayer. At the same time, knowledge of the mechanism ensures better tax compliance by the taxpayer, and there is little scope available to the tax officials to make changes to the reported income.

Case Study: Computerization of Income Tax Department, India There has always been a general perception that the tax department in India was very corrupt. To the credit of the tax departments, they have tried to continuously improve their systems to ensure that the menace of

224  Chandan Kumar Jha and Neelesh Kumar Sah corrupt practices is nipped in bud by creating a system that is fast, accessible, simple, and responsive. Income Tax Department The Income Tax Department in India is the administrative department for tax administration. The department derives its mandate through the Income Tax Act, 1961 (the Act) –​a Central Act passed by the Indian Parliament. As per the Act, tax is leviable on any income derived by a person in India. A “person” could be an individual, a firm, a company, a cooperative society, or any artificial juridical person. All entities are subject to tax at particular rates with some exemptions built in for entities involved in specific activities or specific income.The Act has 298 sections and fourteen schedules. The Act boasts of 131 Income Tax Rules. Add to these various judgments of the Income Tax Appellate Tribunals, High Courts of all Provinces in India, and the Supreme Court of India, and the tax matters become complex and daunting. The tax administration involves the filing of income tax returns by the assessees (the persons). The person has to get a Permanent Account Number (PAN) issued by the department.The returns filed by the assessee are assessed by Income Tax officials and assessment order is issued. The order may indicate a tax liability on the assessee by issuing a tax demand notice, a refund due to the assessee by issuing a refund order, or neither refund nor liability. The department has a system of collecting information on the income of every person. While carrying out their assessment, the income tax official is supposed to verify whether all incomes have been correctly reported by the taxpayer. The assessing officer may issue various notices to the assessee to further scrutinize his/​her records to ascertain/​confirm the correctness and completeness of income. So, we have a system where the taxpayer files his tax return to the department. The department carries out an internal assessment, which is inaccessible to the taxpayer. The assessments are to be completed over a period of around one and half years of the tax return filing. The taxpayer only receives an Income Tax Order at the end of the process of assessment and gets to know whether there has been any addition to his income (thus requiring more tax and associated penalties to be paid) or not. If there is a tax liability or a clarification is sought, the process could be agonizingly long. In case there was a refund, the refund may not be deposited in the taxpayer’s bank account and getting this refund released could be an arduous task. In many instances, the process could take as long as few years after the filing of the return.The process could be even more agonizing if it involved a notice for scrutiny. A cumbersome tax filing system, a nontransparent tax return processing system, and uncertainty of the outcome lead to a situation where the taxpayer is inclined to try to find ways to ascertain the status of their assessment

Tackling Corruption  225 and, if possible, influence the outcome. The system was vulnerable to manipulations. No wonder, the officials of the department were perceived to be involved in corrupt practices. The department acknowledged this issue and steadily brought about many changes to streamline the processes, made possible by computerization and automation. The steps in brief are: Simplification of the tax return forms: Making it simple to fill tax return forms for taxpayers, especially salaried and small business firms. Computerization of the complete workflow: Issue of the assessee identification: The PAN. Simplification of PAN issuance process in a timely manner. E-​filing of income tax returns: Faceless returns filing mechanism, making it possible for the taxpayers to remotely file tax returns and receive an acknowledgment. Centralized Processing Center for the returns: The processing of returns, earlier called summary assessment, conducted from a centralized location for the whole country, making the system faceless for the processing of returns before selection for scrutiny. Refund Banker: Leading to a reduction in time taken to transfer tax refund in the taxpayers’ accounts. E-​payment system: Makes payment of various tax liabilities like the tax, interest, penalty, etc. online, eliminating the physical interaction with tax department officials. E-​TDS (Tax Deduction at Source): Enables linking with other computerized systems like CPC, AST, OLTAS, etc., ensuring seamless linking of tax deducted at source to be linked with the tax claims filed by the assessee through his returns. Online Tax Accounting System (OLTAS): The system integrating all tax payments by the taxpayer in a single ledger account. The system is useful while processing the tax return and determining any pending tax liability. Annual Information Return: Information collected by the department on high-​value transactions attributed to the taxpayer. Used during the processing of the tax returns to determine the completeness and correctness of the returned income by taxpayers. Ayakar Sampark Kendra (Income Tax Contact Center) –​The Customer service center of the department, providing tax and return-​related information to taxpayers.

226  Chandan Kumar Jha and Neelesh Kumar Sah AST (Assessment Information System): The module used to complete scrutiny assessments. The returns are selected for scrutiny assessment after a risk assessment is carried out on the returns processed summarily or based on some intelligence. This stage of assessment was continuously under flak for having potential for corruption and harassment, as taxpayers were summoned for scrutiny assessments. Faceless Assessment System: This latest development is a response to the allegations of potential corruption and harassment of taxpayers during the scrutiny assessment. Additionally, a system of randomness in allocating assessments to assessing officers has been introduced. This has eliminated the assessee–​assessing officer direct interface. While the Income Tax Department has undertaken the huge task of simplifying tax laws and forms and computerized the processes, its job may not yet be over. It has had its share of problems in coming this far. Some of the issues faced by the department in using computerized solutions could be the following:25 1. Concurrent use of computerized and manual systems: When there is a change in a system from manual or legacy to automated or computerized, the two systems are often allowed to operate in parallel. If allowed for too long, the system has a tendency to fall back to legacy systems.This could happen due to various reasons including the vested interest of those who have been used to the legacy system.These vested interests could also be driven by the element of corrupt behavior facilitated by the legacy systems. 2. Inadequate involvement of officials in system analysis and testing: Another reason for failure is that officials whose systems are being computerized are not consulted in the system analysis and testing. The officials working on a system have the deep knowledge of the process and, more importantly, of various exceptional situations that arise in carrying out the process and their solutions. A vanilla system developed on a common understanding of the system is bound to fail in handling exceptions. As a result, disruptions arise, which is fit for being abused by those who offer a quick resolution to the public/​citizen to their cases. 3. Resistance from employees for the change:There is an inherent inertia to change. Any system improvement may also face similar issues. Those with vested interest try to make full use of such inertia to continue with the legacy system, often manual or archaic, and a possible breeding ground for corrupt behavior. 4. Data entry errors leading to data quality issues: Conditions favorable to corrupt behavior occur when there is a glut in receiving due services. Any errors in data entry generally impact the beneficiary rather than the official who made the input error. The absence of accountability puts entire burden on

Tackling Corruption  227 the beneficiary while the person making the error is completely off the hook.The existence of such situations increase the system’s vulnerability to corruption. While any human work is prone to errors, there is a need for a process in place that can identify those that make deliberate errors with an objective to extort bribes. Moreover, not only there should be least-​ effort (in part of the beneficiary) process to correct such errors, but there should also be consequences for those deliberately making such errors. Again, technology can be useful in detecting such cases. 5. Inadequate change management plan: As pointed out above, an inadequate change management plan from legacy to new system has the potential of failure or there is a possibility that an inadequate or error-​prone system is being implemented. A system not yielding desired results leads to a situation where it becomes vulnerable to corrupt behavior. 6. Inadequate connectivity: For the success of a country wide or large system, connectivity is a must. Inadequate connectivity creates an unreliable system and leads to a delay in the service provision, making the system vulnerable to corrupt practices where the agents may surface to get the work done for a fee. In the absence of adequate connectivity, merely the existence of e-​government services would have no benefits in corruption reduction as beneficiaries would be forced to use in-​person services, making them susceptible to bribe demands by the officials. The discussion above reads like an appraisal report of IT system development and implementation issues involved in the process. So, while deciding on tackling corruption by invoking IT-​based systems, it will be important to be aware of the issues that such changes bring with them. Some, if not all, of the problems related to IT system development and implementation lead to situations where the system becomes vulnerable to abuse and corruption. The lesson from this case study is that if we wish to use technology to fight corruption then we must take into account not only the best practices in technology and governance but also their interaction. IT solutions have to be very well thought of, designed, and implemented to obviate its manipulation and to bring in a transparent, efficient, convenient, and incorruptible system.26

Concluding Remarks Corruption does not have a one size fits all solution. However, there are accepted enablers that help reduce corruption, technology being one of the most potent, provided they are planned in a manner to identify and plug the vulnerabilities of corruption in an as widespread manner as possible. In this chapter, we have outlined various processes involved in providing government services to end-​users. We have discussed how these processes give rise to opportunities for corruption and what conditions make corruption persist over time. A case study from the tax department in India has been discussed that shows how technology can be used to simplify the process to reduce opportunities for

228  Chandan Kumar Jha and Neelesh Kumar Sah corruption. Using this case study, we discuss what kinds of challenges arise in designing and implementing new systems designed to reduce corruption. For instance, while IT-​based systems (or the use of e-​government) can automatically reduce certain opportunities for corruption by impacting factors like eliminating the need for a physical interaction between the government official and the beneficiary, other issues that facilitate corrupt practices remain or are only inadequately addressed. For instance, while IT-​based systems might simplify and improve the process leading to an increase in transparency and accountability to certain extent, it does not make the system foolproof against corruption. For instance, it is imperative to establish accountability for human errors like inputting data in the system, and making the resolution in such cases an official accountability. In such cases, although the errors are made by the officials, the victims are the beneficiaries that face psychological distress and monetary costs to get these errors fixed. Such factors need to be addressed beyond just moving from one system to another. To build an IT-​based system that yields desired results and is as robust against corrupt practices as possible, it is crucial to understand the complex nature of the process that is often well known only to the insiders (officials of the department). Therefore, insiders’ input must be sought at every stage of designing the system to predict any loopholes that might remain in the system and might be abused for rent-​seeking purposes. Understanding corruption brings out one thing: the involved parties game the system willingly or unwillingly. Willing parties take advantage of the weak systems of management, control, and oversight. Unwilling parties are forced by the willing, dominant party into the corrupt act. This is accentuated by the fact that the unwilling party has a need and desire to get the services and it may be not readily available unless s/​he becomes a party to the corrupt behavior.While the willing party’s behavior needs a strong oversight mechanism, the unwilling party’s behavior calls for a larger solution where the services are as abundant as to assure the service-​seeker of the service availability and delivery in due course. This is thus related to the larger question of the allocation of resources and their utilization toward the service of people … which is the domain of politics. Further, legal regimes and the strength of legal institutions matter as it is not sufficient to be only able to detect corrupt practices but there must be consequences to those abusing their authority and engaging in corruption practices. In a symmetric liability policy, both bribe-​givers and bribe-​takers are legally equally punishable, which discourages the unwilling party to report having been engaged in bribery to the authorities. In an asymmetric liability policy, only bribe-​takers are subject to legal sanctions, which might encourage the victims of harassment bribery to report the incident.27

Notes 1 United Nations General Assembly, vide its resolution 58/​4 of 31 October 2003; Resolution adopted by the General Assembly, [on the report of the Third Committee (A/​55/​593)], 55/​61 –​An effective international legal instrument against corruption.

Tackling Corruption  229 2 www.trans​pare​ncy.org/​en/​gcb. 3 www.trans​pare​ncy.org/​en/​what-​is-​cor​rupt​ion (accessed August 5, 2022). 4 The World Bank’s “Control of Corruption captures perceptions of the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as ‘capture’ of the state by elites and private interests.” www.trans​pare​ncy.org/​en/​what-​is-​cor​rupt​ion (accessed August 5, 2022). 5 Huther and Shah (2000) present a simple framework in which a public official’s decision to engage in a corrupt practice depends on the expected gains and costs. In this simple framework, public officials can be discouraged to get involved in corrupt practices by implementing policies that either lower the gains from or increase the costs of a corrupt transaction. One of the ways to enhance the costs of undertaking a corrupt transaction is to increase the likelihood that bribery is detected and penalties are imposed. 6 See Graycar (2015) for a discussion on various types, activities, sectors and places (TASP) of corruption. 7 Distinguishing between “extortionary” and “collusive” bribery is useful in this context. In extortionary or harassment bribery, the government officials extract bribes for the services that the citizens are entitled to. For instance, a citizen seeking a passport after submitting the required documents. On the other hand, in collusive bribery, the service-​seekers use bribery to get undue benefits. For example, an individual might use bribery to avoid a traffic ticket, or a corporation might use bribery to avoid business regulations or taxes. 8 Alm et al. (2016) on the use of corruption by firms for tax evasion purposes. 9 In an experimental study based in India, Bertrand et al. (2007) find that the middlemen can get around the required procedures, and by making extralegal payments through these intermediaries, one can obtain a driver’s license without knowing how to drive. 10 For various factors that facilitate corruption (in infrastructure), visit “Why Corruption Occurs” page by the Global Infrastructure Anti-​Corruption Centre (https://​gia​ccen​tre.org/​why-​cor​rupt​ion-​occ​urs, accessed August 4, 2022). Many of these factors are relevant to corruption in other sectors. 11 A study sponsored by the Department of Administrative Reforms and Public Grievances, Government of India, concluded that “the state of public grievances serves as a barometer to gauge the efficiency and effectiveness of the administrative processes and polices” (Arora, 2008). 12 www.coso.org/ ​ S ha ​ red%20Do ​ c ume ​ n ts/​ F ramew​ o rk- ​ E xecut ​ ive- ​ S umm ​ a ry.pdf (accessed 31 July 2022). 13 The role of oversight mechanism and internal and external audits and controls in reducing corruption has been examined by various studies (see, e.g., Jiménez, 2009; Dorotinsky and Pradhan, 2007). Sometimes external audits and oversight might even lead to higher corruption. Even parliamentary oversight might facilitate corruption through their influence on bureaucratic system (Brierley, 2020). In this context, the role of an active civil society becomes extremely important. Interested readers should refer to Dorotinsky and Pradhan (2007) for a detailed discussion on the role of oversight mechanisms on corruption. On the role of civil society in ensuring the efficacy of internal and external controls in promoting transparent and accountability and reducing corruption, see Dorotinsky and Pradhan (2007) and Jha (2017). 14 The evidence suggests that computerization, at least in certain cases, has been successful in reducing corruption. For instance, Bhatnagar (2003) documents that

230  Chandan Kumar Jha and Neelesh Kumar Sah the Bhoomi (translated land in English) project that computerized 20 million land records owned by 6.7 million farmers in the Indian state of Karnataka caused a significant decline in corruption. The project, started in 1991, reportedly led to a saving of Rs. 806 million (roughly 18 million US Dollars at the then exchange rate of Rs. 45 per USD) for 6.7 million farmers in bribes. 15 Reports of the Comptroller and Auditor General of India, Union Government, Department of Revenue (Direct Taxes), for years 2010–​11 and 2021 (Report No. 8). 16 Bureaucratic discretion might be used for private gains rather than in public interest and is generally bad for businesses and investments as it increase uncertainty (Kurniawan, 2021; Beazer, 2012). However, some bureaucratic discretion might be good (Decarolis et al., 2021) and even necessary. The quality of institutions might determine whether and what extent of discretionary power might be optimal. 17 Shrivastava (2016) suggests some ways to identify possible corrupt practices in the provision of public services. 18 Ombudsman Western Australia, 2019. Guidelines on Decision-​ making (revised April 2019). www.ombuds​man.wa.gov.au/​Publi​cati​ons/​Docume​nts/​gui​deli​nes/​ Bin​der-​Decis​ion-​Mak​ing.pdf (accessed July 20, 2022). 19 While some studies show that transparency can lead to lower corruption even in highly hierarchical societies (Peisakhin and Pinto, 2010), others have argued that transparency may not be enough in reducing corruption and might need to be complemented with other policies to be effective against corruption (Lindstedt and Naurin, 2010). Further, it also matters whether transparency is controlled by the agents or nonagents (such as a free press) with the latter being more effective in reducing corruption (Lindstedt and Naurin, 2010). 20 Several studies have explored the role of e-​government in reducing corruption. See Jha and Sarangi (2022) for a discussion of various ways in which e-​government can reduce corruption and a review of empirical literature on this association. These include promoting transparency and accountability leading to a reduction in the scope for officials’ discretion and minimizing the need for a physical interaction between government officials and beneficiaries. 21 It is important to emphasize that additional conditions like an active civil society might be required for oversight mechanisms, internal or external, to be effective against corruption, else such oversights might even lead to an increase in corruption (Dorotinsky and Pradhan, 2007). 22 Grievance Analysis Report, Department of Administrative Reforms and Public Grievances (DARPG), Government of India. www.darpg.gov.in/​node/​1863; CPGRAMS portal: https://​pgpor​tal.gov.in. 23 Although there is not much work on this, Mohapatra (2016) and Paul (2011) provide some discussion on the importance of grievance redressal mechanism for governance and corruption. 24 Studies have shown that access to information is negatively correlated with corruption across countries (see, e.g., DiRienzo et al., 2007). 25 See Sah (2009) for a detailed discussion on some of these issues. 26 See Jha and Sarangi (2022) for a discussion and existing empirical evidence on many ways in which information and communication technologies (ICTs) can help in the fight against corruption (see Chapter 11). 27 An asymmetric liability policy for “harassment bribes” was proposed by Basu (2011). Subsequently, studies have shown that asymmetric liability (with additional incentives) can reduce bribery in theoretical and experimental settings (Basu, Basu,

Tackling Corruption  231 and Cordella, 2016; Dufwenberg and Spagnolo, 2015; Jha, 2015; Abbink et al., 2014). Dreze (2011) has criticized this suggestion.

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. Alm, J., Martinez-​Vazquez, J. & McClellan, C. (2016). Corruption and firm tax evasion. Journal of Economic Behavior & Organization, 124, 146–​163. Arora, D. (2008). Public grievance redress and monitoring system in Government of India ministries and departments. https://​darpg.gov.in/​sites/​defa​ult/​files/​IIPA​_​Rep​ ort_​GRM.pdf (accessed August 4, 2022). Basu, Kaushik (2011): Why, for a Class of Bribes, the Act of Giving a Bribe should be Treated as Legal. Published in: Ministry of Finance Government of India Working Paper No. 1/​2011-​DEA (March 2011). Basu, K., Basu, K. & Cordella, T. (2016). Asymmetric punishment as an instrument of corruption control. Journal of Public Economic Theory, 18(6), 831–​856. Beazer, Q. H. (2012). Bureaucratic discretion, business investment, and uncertainty. The Journal of Politics, 74(3), 637–​652. 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. Bhatnagar, S. (2003). E-​government and access to information. Global Corruption Report, 2003, 24–​32. Brierley, S. (2020). Unprincipled principals: Co-​opted bureaucrats and corruption in Ghana. American Journal of Political Science, 64(2), 209–​222. Decarolis, F., Fisman, R., Pinotti, P. & Vannutelli, S. (2021). Rules, discretion, and corruption in procurement: Evidence from Italian government contracting. National Bureau of Economic Research, w28209. DiRienzo, C. E., Das, J., Cort, K. T. & Burbridge, J. (2007). Corruption and the role of information. Journal of International Business Studies, 38(2), 320–​332. Dorotinsky, W. & Pradhan, S. (2007). Exploring corruption in public financial management. In The many faces of corruption: Tracking vulnerabilities at the sector level (pp. 267–​ 294). The World Bank, Washington, DC. Dreze, J. (2011). The bribing game. Indian Express, April 23, 2011. http://​econ​dse.org/​ wp-​cont​ent/​uplo​ads/​2012/​09/​JD-​The-​Brib​ing-​Game-​2011.pdf (accessed June 28, 2022). Dufwenberg, M. & Spagnolo, G. (2015). Legalizing bribe giving. Economic Inquiry, 53(2), 836–​853. GCB. (2019). Global corruption barometer: Africa 2019. Transparency International. www.trans​pare​ncy.org/​en/​gcb/​afr​ica/​afr​ica-​2019 GCB. (2020). Global corruption barometer: Asia 2020. Transparency International. www. trans​pare​ncy.org/​en/​gcb/​asia/​asia-​2020 GCB. (2021). Global corruption barometer: Pacific 2021. Transparency International. www.trans​pare​ncy.org/​en/​gcb/​paci​fic/​paci​fic-​2021 Graycar,A. (2015). Corruption: Classification and analysis. Policy and Society, 34(2), 87–​96. Huther, J. & Shah, A. (2000). Anti-​corruption policies and programs: A framework for evaluation, vol. 2501. World Bank Publications.

232  Chandan Kumar Jha and Neelesh Kumar Sah Jha, C. K. (2015). Essays on corruption. LSU Doctoral Dissertations, 2669. https://​dig​ital​ comm​ons.lsu.edu/​grads​choo​l_​di​sser​tati​ons/​2669. Jha, C. K. (2017). Information control, transparency, and social media: Implications for corruption. In Political scandal, corruption, and legitimacy in the age of social media (pp. 51–​ 75). IGI Global. Jha, C. K. & Sarangi, S. (2022). How advances in information and communication technologies impact corruption? SSRN. https://​ssrn.com/​abstr​act=​4158​888. Jiménez, F. (2009). Building boom and political corruption in Spain. South European Society and Politics, 14(3), 255–​272. Kurniawan, T. (2021). Discretion as a factor in corruption: A case from Indonesia. Public Integrity. https://​doi.org/​10.1080/​10999​922.2021.1975​939. Lindstedt, C. & Naurin, D. (2010). Transparency is not enough: Making transparency effective in reducing corruption. International Political Science Review, 31(3), 301–​322. Mohapatra, B. P. (2016). Good governance, public institutions and grievance redressal mechanism in India: Can citizen grievance mechanism be able to enhance the performance of public institutions? Journal of Governance & Public Policy, 6(1), 59. Paul, S. (2011). Fighting corruption. Economic and Political Weekly, 17–​19. 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. Sah, N. (2009). Audit perspectives for e-​ governance projects. In Fostering e-​ governance: Selected compendium of Indian initiatives (pp. 32–​44).The Icfai University Press. Srivastava, B. (2016).Tackling corruption with agents & ICT:A vision. arXiv:1701.06426. United Nations. (2004). United Nations convention against corruption. United Nations. www.unodc.org/​docume​nts/​bruss​els/​UN_​Co​nven​tion​_​Aga​inst​_​Cor ​r upt​ion.pdf (accessed July 23, 2022).

Index

Note: Page numbers in italics refer to figures and those in bold type refer to tables. Endnotes are indicated by the page number followed by “n” and the endnote number e.g., 208n1 refers to endnote number 1 on page 208. absence of corruption 2, 58, 60, 61, 66, 74, 77–​78, 116 absenteeism 159, 160 abuse of authority 217–​218, 219, 228 accountability 4, 5, 16, 96, 115, 119, 129, 144, 149, 208n1; and ICTs 175, 177, 178, 180, 181, 183; tackling corruption 216, 221, 222, 223, 226–​227, 228, 229n13, 230n20 active corruption 194, 195 activism, anticorruption 181, 183, 184n1 activist government 35, 36 adversarial attitude 36 Afrobarometer surveys 17, 19–​20 agency framework 59, 68n4 allocation of resources 87, 95, 228 ambiguous laws 54 anonymity 129, 140, 180, 181, 185n12, 193, 213 anticorruption 1–​2, 3, 4, 5, 15, 21, 58, 59, 104, 107–​108, 115, 149, 173, 176, 217, 219; in Europe 191, 194, 199, 200, 206, 214, 215 anticorruption activism 181, 183, 184n1 anticorruption agencies 173, 199, 207, 214, 217 anticorruption frameworks 194, 199, 206, 207 anticorruption policies 3, 21, 22–​23, 23, 24, 25, 59, 67, 100, 103, 110, 118, 179, 181–​182, 190, 200; effectiveness of 121–​131, 179, 216–​217 anticorruption reform 24 anticorruption strategies 1, 15, 63, 67, 68, 189, 190, 206, 207

Aquinas, Thomas 34 arrest 7, 10, 11, 12, 140–​141 asymmetric bribery 111 asymmetric punishment 107, 129 asymmetric regimes 128–​130 attitudes of corruption 154, 155 audits 61–​62, 123, 125, 133n6, 163, 222 Austria 191, 192, 193, 213 authority 12, 16, 34, 43, 117, 196, 218, 220; abuse of 217–​218, 219, 228 authorization 63 bad apple theory 120–​121 bank accounts 25, 127–​128, 205, 224 banks 90, 91, 92, 93, 94, 95 bargaining power 43, 44, 53, 69n20 Becker’s crime enforcement model 1, 8, 13–​14n1, 18, 67, 101–​102, 103, 122–​123 behavior: corrupt/​corruption see corrupt behavior; deviant 58, 214; dishonest 105–​106, 107, 111, 126; equilibrium 42; human 8, 100; opportunistic 3; risk-​taking 91 beliefs 21, 106, 111, 115–​116, 123, 126; republic of 103–​104, 128; shared 119, 120; see also norms Bhoomi project 176, 185n7, 229–​230n14 bias 173–​174, 183, 220, 222 blogging 180, 181 borrowing constraints 94 bribery: asymmetric 111; collusive 62, 108, 109–​110, 129, 229n7; extortionary 229n7; harassment 108, 109–​110, 116, 123, 127, 128, 129, 130, 228,

234 Index 229n7; model with 43–​53, 46, 47, 50; preemptive 54, 55n17 bribery games 108, 116, 119, 123, 128, 129, 155, 156, 161 Bulgaria 189, 196, 197, 198, 199, 202, 209n31 bureaucracy 2, 11, 32, 53, 73, 132, 149, 158 bureaucratic corruption 1, 72, 124 bureaucratic discretion 5, 36, 43, 44, 62, 121, 175–​176, 218, 219, 220–​221, 230n16, 230n20 bureaucratic power 32 bureaucrats 32, 53, 109–​110, 122, 124, 131, 158, 183; and growth 72, 73, 74, 75, 76, 77, 78, 79, 81, 83n2 Burkina Faso 24, 105, 107, 157 business 71, 94, 173, 177, 184n2, 218, 225, 229n7; competition in 198, 204; in Europe 190, 191, 192, 195, 203, 206, 208n1, 213, 215 Canada 107, 157 capital dynamics 77–​79, 82 Catholic Church 34 causal pathways: between bribery and a country’s legal system 54; between a country’s legal origins and the quality of its government 53; linking legal origin to corruption 32 causes of corruption 3, 72–​73, 115, 118–​121, 190 CCI (Control of Corruption Index) 172–​173, 176, 177, 184n3 CCSI (Core Civil Society Index) 183 CEE (Central and Eastern Europe) 189, 190, 196–​200, 203, 214 cell phones 4, 174, 175, 178–​180, 184n1 censorship 173, 181, 183 Central and Eastern Europe (CEE) 189, 190, 196–​200, 203, 214 certainty vs. severity of punishment 121, 122–​123 church 34, 35, 163 citizen reporting 121, 126, 128–​130, 193, 205, 206–​207, 208n8, 213, 214, 215 civil law 2, 31, 32–​33, 34–​35, 36–​37, 38, 39, 41–​43, 44, 45, 47, 50, 51, 52, 53, 54, 55n8, 55n12, 55n13, 55n14, 55n15, 56n18 civil society 4, 61–​62, 127, 174, 182, 183, 184, 185n16, 222, 229n13, 230n21 codification, of the law 34, 36

collateral damage, to the economy 12 collective action 16, 20, 184 collusion 2, 59, 61–​62, 62–​63, 65, 66, 68n3, 68n13, 74, 124, 129, 198 collusive bribery 62, 108, 109–​110, 129, 229n7 common law 2, 31–​32, 33, 34, 35, 36, 37, 38, 39, 40–​41, 43, 44, 45, 47, 51, 52–​53, 54, 55n6, 55n8, 55n11, 55n12, 55n13, 55n15, 56n18 competition 95, 121, 124, 146, 148; business 198, 204; for resources 93, 124 compliance 16, 18–​20, 20–​21, 25, 102, 123, 126, 182; and collusion 61; and corruption 22, 22, 24, 26, 81; and extortion 63; regulatory 2, 20; standard model of 18, 20, 22; tax 2, 19, 59, 115, 116–​117, 120, 121, 133n2, 163, 167n11, 220, 223 conditional mean bribe 53 conflicts of interest 192, 199, 214 consequences of corruption 3, 16, 72, 115–​118, 125, 126, 160, 190, 228 contagion 16, 18, 111, 120–​121, 132 context sensitivity, and reverse causality 142–​145, 143 contracts 16, 35, 36, 62, 63, 69n22, 93, 116, 124, 158, 159, 165, 218 Control of Corruption Index (CCI) 172–​173, 176, 177, 184n3 Convention against Corruption (UNCAC), of United Nations 69n22, 216 cooperation 15, 16, 116, 202, 205; cross-​border 196; international 24–​25, 216; judicial 193, 194, 195 Core Civil Society Index (CCSI) 183 corporate governance 91 corrupt workers, and selection into public sector jobs 158–​162 corruption: absence of 2, 58, 60, 61, 66, 74, 77–​78, 116; active 194, 195; bureaucratic 1, 72, 124; causes of 3, 72–​73, 115, 118–​121, 190; culture of 163–​165; definition of 33, 166n1, 172, 207, 217; deterrence hypothesis of 101–​102, 121, 122–​123; and development 72–​73; emergence of 218–​219; in Europe 189–​209; exacerbation by corruption control of 8–​11, 9, 11; financial sector 87–​96; full prevalence of 78–​79; grand 43, 184n2, 194, 198, 206, 214; harassment 17,

Index  235 62–​63; high 2, 24, 72, 116, 118, 157, 158, 161, 162; imitative effect of 18; joint evolution of growth and 71–​84, 82, 83; over time 162–​163; perception of see corruption perceptions; persistence of 3, 5, 15, 179, 218–​219; and policy 12–​13; tackling of 216–​231; and trust 17–​20; see also perceived corruption corruption compliance cycle 26 corruption control 1, 2, 6–​11, 9, 11, 12–​13, 92 Corruption Eradication Commission (KPK) 13 corruption experiments 100–​111 corruption norms 154, 155–​156, 162 corruption perceptions 20, 153, 155, 156, 157, 161, 162, 167n5, 206; see also perceived corruption Corruption Perception(s) Index (CPI) 17, 110, 139, 142, 143, 156, 161, 173, 176, 178, 184n3, 189, 191 corruption research 1, 5, 131, 148, 149 corruption trap 2, 15 corruption-​culture hypothesis 153–​167 corruption–​gender relationship 139, 144, 148 Council Framework Decision 2003/​568 195 courts 24, 33–​34, 35, 36, 37, 38, 39, 40, 43, 55n14, 197, 216, 224 COVID-​19 pandemic 18–​19, 21, 189, 191, 192, 201–​202, 206, 213, 214 CPI see Corruption Perception(s) Index (CPI) credit 87, 88, 91, 92, 93, 94, 95 crime 10–​11, 67, 101, 102, 103, 127, 128, 140–​141, 190, 194, 196, 201; enforcement of 1, 102; euro-​ 190; organized 189, 194; prevention of 103 criminal justice systems 102, 207 Croatia 203, 204, 205 crony capitalism 10, 11 cronyism 1, 6–​14, 9, 11, 110 cronyism-​corruption trap 11 cross-​border cooperation 196 cross-​country settings 12, 72, 91, 131, 172, 173, 175, 176–​177, 178–​179, 181, 182 cross-​cultural evidence 106–​107 cultural norms 13, 107, 119–​120, 174 culture-​corruption hypothesis 4, 153–​167

cultures 3, 104, 106, 120, 155, 156, 157, 161, 167n5, 177, 182 Cyprus 203, 204, 205 Czech Republic 197, 198, 199–​200, 209n31 DAI (Digital Access Index) 178, 185n13 decision-​making 16, 21, 37, 56n18, 100, 220, 230n18 definition of corruption 33, 166n1, 172, 207, 217 delivery of public services 59, 109, 124, 146, 175–​176, 176–​177, 218, 223, 228 democracies 4, 10, 17, 24, 183, 185n16, 194, 206, 207, 208n1; and gender 141, 142–​144, 143, 146, 149n4 Denmark 122, 159, 160, 161, 162, 166, 189, 191, 192, 193, 209n35 detection of corruption 2, 23, 23, 64–​65, 74, 78, 79, 123, 127, 129, 167n9, 205 deterrence 22–​23, 25, 58, 61, 65, 66, 67, 68n2, 69n21, 132, 163 deterrence hypothesis, of corruption 101–​102, 121, 122–​123 development: and corruption 72–​73; economic 72, 73, 82, 83, 92, 118, 173, 195; financial sector 3, 87, 88, 89, 90, 91, 92, 93, 94, 95–​96; sustainable 15, 189, 216 deviant behavior 58, 214 dictator game 160, 161, 162, 165 die-​roll task 156, 158, 159, 160, 161 Digital Access Index (DAI) 178, 185n13 digital technology 12, 174, 178, 179 dilution 2, 61 diplomatic status 154 Directive 2017/​1371 190, 193–​194, 194–​195, 201 Directive 2018/​1673 195–​196 Directive 2019/​1937 205, 206–​207, 208n8 disclosure 84n5, 88, 90–​91, 106, 120–​121, 126 discretion, bureaucratic 5, 36, 43, 44, 62, 121, 175–​176, 218, 219, 220–​221, 230n16, 230n20 dishonesty 105–​106, 107, 111, 117, 121, 126, 158–​159, 160, 163, 166, 167n6 disputes 31, 33, 35, 36, 37, 54n1 dynamic panel data models 174, 179 Ebola 19 economic activity 87–​88, 88–​89, 94, 173

236 Index economic development 72, 73, 82, 83, 92, 118, 173, 195 economic growth 31, 71, 76, 84n10, 89, 95, 96, 100, 115, 116, 118, 153, 208n1 economic inequality 117–​118 efficiency 3, 15, 40, 53, 119, 127, 147, 200, 214, 229n11; financial sector 88, 90, 93, 94, 96; and growth 72–​73, 75, 83n2 effort 64, 121, 158, 160, 166, 167n7, 182, 193 EGDI (E-​Government Development Index) 174, 175, 176 e-​governance 69n16, 174, 178 e-​government 4, 172, 174–​178, 182–​183, 185n11, 221, 227, 228, 230n20 E-​Government Development Index (EGDI) 174, 175, 176 e-​government maturity 175, 177 elites 15, 16, 73, 144, 229n4 embezzlement 20, 23, 25, 109, 110, 117, 121, 130, 140, 208n6 emergence of corruption 218–​219 empirical origins of initial work into gender and corruption 140–​142 endogeneity 102, 173–​174, 179 enforcement: law 15, 20, 72, 179, 185n12, 199, 201, 215; non-​ 62; optimal 58–​69; over 2, 59, 64–​67; policy for 60, 64, 66; problem of 59–​60, 61; under 64–​65, 65–​66 Enlightenment 34, 36 environmental law 33, 37 E-​Participation Index 177 EPPO (European Public Prosecutor’s Office) 190, 200–​201, 202, 206, 207, 209n35, 214 equilibrium analysis 41–​43 Estonia 196, 197, 198, 199, 200, 209n31 ethics 120, 121, 125, 132, 165, 166, 167n12, 216 EU (European Union) 4, 91, 146, 189–​209; Budget 196, 199, 201–​203, 206 Eurobarometer 189, 190, 203, 208n3, 213–​215 euro-​crime 190 European Public Prosecutor’s Office (EPPO) 190, 200–​201, 202, 206, 207, 209n35, 214 European Union (EU) 4, 91, 146, 189–​209 executive branch 39, 144

expected utility 18, 74, 76, 101, 102, 117, 124 experimental corruption games, typology of 109–​110 experimental methods 100–​111 experimenter demand effect 108–​109, 111 experiments: field 18, 105, 109, 110, 116, 117, 122, 127, 130, 154, 157, 164–​165; laboratory see laboratory experiments exploitation 17, 31, 73, 177 external validity 3, 101, 104–​105, 106, 110, 111, 177 extortion 2, 59, 62–​64, 66, 67, 68n3, 69n16, 69n21, 109, 122 extortionary bribery 229n7 extortion-​proof action 67 extractive power 2, 53 extrinsic incentives 144, 158, 159, 166 fairness 13, 15, 17, 21, 23, 208n1, 220 favoritism 110, 120, 189, 198, 204, 213, 214, 215 favors 1, 31, 111, 127, 129, 198, 214, 223 females see women feminization 141, 142 field experiments 18, 105, 109, 110, 116, 117, 122, 127, 130, 154, 157, 164–​165 fighting extortion 63–​64 financial policy 92–​93 financial regulation 88, 89, 90–​91, 93–​94, 96, 214 financial sector: corruption in 87–​96; development of 3, 87, 88, 89, 90, 91, 92, 93, 94, 95–​96 financial systems 87, 90, 91, 95, 96 Finland 189, 191, 192, 193, 204, 208n8 framing 62, 104, 108–​109, 111 France 25, 37, 55n7, 107, 120, 203, 204, 205 fraud 117, 126, 190, 193–​195, 208n6, 214 freedoms 4, 36, 89, 94, 173, 174, 182, 183, 207 friends 7, 10, 11, 12, 181, 215 full prevalence of corruption 78–​79 functionaries 40, 43, 55n16 game theory 12, 111 GCB (Global Corruption Barometer) 216, 229n2 gender 1, 3–​4, 107, 121, 131, 139–​150, 143, 157, 162, 167n9

Index  237 gender–​corruption relationship 139, 144, 148 General Assembly, of United Nations 216, 228n1 generalizability 101, 104–​105, 107, 108, 131 Germany 32, 35, 107, 119, 127, 161, 163, 191, 192, 193 gift exchange 103, 110 Global Corruption Barometer (GCB) 216, 229n2 Global IT Report 180, 185n10 good governance 82, 117 governance 1, 3, 16, 19, 92, 96, 100, 118, 144, 145, 185n12, 219, 223, 227, 230n23; corporate 91; e-​ 69n16, 174, 178; good 82, 117; see also World Governance Indicators (WGI) government effectiveness 19, 177 government ownership, of banks 88, 89–​90 government services 4, 5, 172, 174, 175–​176, 177, 182, 227 grand corruption 43, 184n2, 194, 198, 206, 214 ‘grease the wheels’ hypothesis 87–​88, 88–​89, 93–​94 Greece 203, 204, 205 grievances 219, 222–​223, 229n11, 230n22, 230n23 growth: impact of corruption on 115–​116; economic 31, 71, 76, 84n10, 89, 95, 96, 100, 115, 116, 118, 153, 208n1; joint evolution of corruption and 71–​84, 82, 83 harassment 2, 221, 226 harassment bribery 108, 109–​110, 127, 128, 129, 130, 228, 229n7, 230n27 harassment bribery game 108, 116, 123, 128 harassment corruption 17, 62–​63 healthcare 116, 146, 191, 192, 197, 199, 204, 206, 207, 208n6, 213, 214, 215 high corruption, vicious circle of 2, 24, 72, 116, 118, 157, 158, 161, 162 honesty 18, 23, 25, 66, 81, 107, 109, 120, 123, 141, 146–​147; culture-​corruption hypothesis 156, 161, 162, 163, 165, 166; dis-​ 105, 117, 121, 158–​159, 160, 163, 166, 167n6; legal systems 43, 44, 45, 46, 47, 48, 49, 51, 52

households 73, 74, 75, 76, 77, 78, 79, 81, 84n8, 149n3, 185n9 human behavior 8, 100 Human Capital Index 175 Hungary 189, 196, 197, 198, 199, 202, 209n31, 209n35 ICRG (International Country Risk Guide) Corruption Index 172, 184n3 ICTs (information and communication technologies) 4, 172–​185, 227–​228 IMF (International Monetary Fund) 25 imitative effect of corruption 18 incentives 76–​77; extrinsic 144, 158, 159, 166; intrinsic 158, 163; monetary/​ pecuniary 4, 103, 121–​124, 128, 129, 154, 158, 164, 166 incentivized bribery game 161 incentivized dictator game 161, 162, 165 incentivized laboratory experiment 155 Income Tax Department, India 221, 223–​227 India 5, 55n11, 84n4, 84n5, 106, 107, 111, 131, 147, 181, 185n12; culture-​ corruption hypothesis 156, 159, 160, 162, 167n6, 167n12; tackling corruption 217, 220, 221, 223–​227, 229n9, 229n11, 230n15, 230n22 Indonesia 6, 7–​8, 13, 20, 106, 131, 156, 159, 161, 162–​163, 165, 166 information and communication technologies (ICTs) 4, 172–​185, 227–​228 information availability 223 information disclosure 126 information structure 62 inquisitorial attitude 36 insider trading 88, 89, 91 inspection 59, 60, 61, 62, 63, 64, 66, 67, 68n7, 68n8, 68n10, 68n11, 69n20, 72, 123 institutional asymmetry 129 institutional environments 25, 104, 111, 129 institutional norms 120 institutional trust 15–​26, 22, 23, 200, 214 institutions: importance of 119–​120; strong 3, 25, 96, 184; weak 25, 84n9, 89, 93–​94, 96, 129, 199 instrumental variable (IV) analysis 145, 174 interface of officials with citizen 221, 226 intermediaries 121, 130, 229n9

238 Index internal controls 219, 222 international cooperation 24–​25, 216 International Country Risk Guide (ICRG) Corruption Index 172, 184n3 International Monetary Fund (IMF) 25 International Telecommunication Union (ITU) 178 internet 4, 172, 173, 174, 175, 177, 178–​180, 181, 182, 183, 184n1 interpretation of the law 33, 34, 35 interventions 18, 25, 90, 111, 125, 129, 132, 153–​154, 163, 165, 166, 167n12 intrinsic honesty 123, 146–​147 intrinsic incentives 158, 163 intrinsic motivation 19, 115, 125, 128, 130, 158 investment 16, 71, 72, 73, 75, 84n6, 87, 88–​89, 93, 100, 116, 173, 230n16 Ireland 21, 191, 192, 193, 209n35 Italy 117, 119, 189, 203, 204 ITU (International Telecommunication Union) 178 IV (instrumental variable) analysis 145, 174 joint evolution of corruption and growth 71–​84, 82, 83 judge-​made law 38, 44 judges 2, 32, 33, 35, 36, 37, 38, 40, 54, 54n1, 55n14, 55n15, 197 judicial cooperation 193, 194, 195 judiciary 2, 12, 35–​36, 39, 118, 193, 197, 198, 199, 200, 202, 207, 214, 215 jurisdictions: civil law 31, 35, 36, 38, 39, 40, 41, 42–​43, 44, 45, 46, 47, 51, 53, 54, 55n13; common law 34, 36, 38, 39–​40, 41, 42–​43, 44–​45, 46, 47, 47, 51, 53 justice 15, 17, 34, 103, 128, 199, 202–​203, 208n2, 216; criminal 102, 207; procedural 19, 21 kickbacks 31, 214 KPK (Corruption Eradication Commission) 13 Kraus, Karl 71, 83n1 laboratory experiments 3, 20, 22, 100, 104, 105, 106, 107, 110, 111, 116–​117, 121, 122–​123, 129, 132, 141, 155, 157, 165, 167n12 Latin America 17, 141 Latvia 196, 198, 199, 204, 209n31

law: ambiguous 54; civil see civil law; codification of 34, 36; common see common law; enforcement of 15, 20, 72, 179, 185n12, 199, 201, 215; environmental 33, 37; interpretation of 33, 34, 35; Roman 34; rule of see rule of law; statute 34 lax oversight 221–​222 legal interpretation 33, 34, 35 legal origins 32, 33–​38, 53, 54 legal systems 2, 31–​56, 46, 47, 50, 90 legislation 33, 34, 35, 38, 39, 55n5, 68, 190, 193, 194, 206, 213, 219 legislative branch, of government 39 legitimacy 2, 15–​26, 22, 23 legitimacy effect 21, 22, 23, 23 legitimacy trap 20–​25, 22, 23, 26 lending 88, 89, 90, 92, 94 lenient authorization 63 Lithuania 197, 198, 199, 200, 209n31 loans 88, 90, 91, 92, 94, 202 Luxembourg 191, 192, 193 males see men Malta 203, 204, 205 marginal deterrence 65–​66, 69n21 markets 13, 16, 74, 75, 84n5, 84n8, 90, 91, 92, 95, 181, 196, 208n27 maximization problem 44 men, and corruption 131, 139, 140–​141, 142, 144, 145, 146, 147, 149, 149n1, 149n3, 167n9 methodology 71, 175, 184n3, 190, 207 Mexico 18, 142, 143, 147, 164 missing expenditures 20 missions 158, 159, 166, 167n7 mistrust 15, 18, 19, 20, 21, 22, 25 mobile phones 4, 174, 175, 178–​180, 184n1 models 1, 3, 73, 100–​111; with bribery 43–​53, 46, 47, 50; dynamic panel data 174, 179 Moldova 120 monetary incentives 4, 103, 121–​124, 128, 129, 154, 158, 164, 166 money laundering 191–​192, 194, 195–​196, 205, 213 monitoring 25, 59, 60, 63, 65–​66, 67, 68, 90, 121, 122, 129, 158, 219, 221–​222, 223; technologies for 60, 127–​128, 129, 180 moral costs 4, 64, 104, 108–​109, 111, 121, 124, 125, 130, 132, 166

Index  239 moral hazard 63, 64, 87 morality 34, 71, 124 motivation: intrinsic 19, 115, 125, 128, 130, 158; prosocial 161, 165, 166; public service 158, 164 negative externalities 121, 125, 126 negligence 37 Netherlands 189, 191, 192, 193, 204, 208n5 networks 3, 132, 145, 146, 147; social 177, 180, 185n10 Nigeria 25, 84n4 nonenforcement 62 non-​linearity 94 non-​performing loans 91 norms: corruption 154, 155–​156, 162; cultural 13, 107, 119–​120, 174; institutional 120; social 111, 119–​120, 126, 182; sociocultural 120 North-​Western Europe 191–​196, 198, 200 nuisance 37–​38, 118 observational causal inference research designs 145 Ombudsman Western Australia 220–​221, 230n18 one-​shot bribery game 156 Online Procedures ENhancement for civil application (OPEN) 176 online reporting, of corrupt offices and officials 129 Online Service Index 175, 177 OPEN (Online Procedures ENhancement for civil application) 176 opponents of political leaders 10, 11, 11, 18 opportunistic behavior 3 opposition, political 2, 7, 13, 24 optimal enforcement 58–​69 organizational culture 4, 67, 153–​167 organized crime 189, 194 other-​regarding values 146, 147 overenforcement 2, 59, 64–​67 overreporting 62, 66, 67 oversight 90, 127, 200, 219, 222, 228, 229n13, 230n21 Pakistan 84n4, 92, 119, 157 pandemic, COVID-​19 18–​19, 21, 189, 191, 192, 201–​202, 206, 213, 214

pecuniary incentives 4, 103, 121–​124, 128, 129, 154, 158, 164, 166 penalties 40–​41, 55n3, 64, 66–​67, 74–​75, 195, 196; corruption experiments 101, 102, 106, 107, 117, 123, 125, 128; tackling corruption 224, 225, 229n5 perceived corruption 16, 17, 19–​20, 71, 156, 157–​158, 160, 173; in Europe 190, 191, 195, 196, 197, 198, 199, 203, 206 persistence of corruption 3, 5, 15, 179, 218–​219 “PIF” directive 190, 193–​194, 194–​195, 201 Poland 199, 200, 202, 209n35 policing 17, 33, 38, 101, 103, 122, 127, 133n1, 141, 165, 181, 216; in Europe 193, 197, 199, 204–​205, 213 policy: anti-​corruption see anti-​ corruption policies; and corruption 12–​13; enforcement of 60, 64, 66; financial 92–​93; public 58 policy lessons 115–​133 policy regimes 104, 107–​108 political allegiance 10–​11 political connections 87, 88, 89, 92, 117–​118, 198, 204 political leadership 1–​2, 7, 8, 10, 11, 12, 105–​106, 214 political opponents 10, 11, 11, 18 political opposition 2, 7, 13, 24 political power 7, 117 political science 5, 16, 216 political sector 197, 206, 214 politics 6, 144, 156, 191, 192, 198, 204, 215, 228 pollution 33, 37, 38, 58, 59–​60, 61, 62, 63, 64, 65, 66, 67, 68n8 Portugal 203, 204 power: bargaining 43, 44, 53, 69n20; bureaucratic 32; extractive 2, 53; political preemptive bribery 54, 55n17 pre-​trial settlement 54, 55n17 prevention of collusion 61–​62 Principal-​Agent-​Client framework 59, 68n4 Principal-​Supervisor-​Agent framework 59, 68n4 principles of reason 34 private gain 1, 31, 33, 58, 59, 64, 65, 96n1, 101, 110, 153, 154, 172, 189, 217, 229n4, 230n16

240 Index private individuals 73, 74; see also households private nuisance 38 private sector 88, 95, 159, 160, 165, 190, 195, 206, 213, 214 procedural justice 19, 21 process simplification 219–​220 productivity 73, 75, 84n6, 87, 88, 90, 93, 95, 159 profit-​maximizing quantity 45, 47, 48, 51 profits 38, 40, 41, 44, 45, 46, 48, 49, 52, 93, 96n3 property 35, 37, 38, 72, 96n3, 194, 196, 216 prosecution 84n5, 128, 129, 199, 200–​201, 202, 204, 205, 215 prosociality 4, 25, 126, 140; and culture-​ corruption hypothesis 158, 159–​160, 161, 162–​163, 164, 165, 166, 167n7, 166n9, 166n10 psychological utility 108 public administration 131, 158, 162, 191, 198, 206 public agents see bureaucrats public choice theory 118–​119, 132 public good provision 19, 20, 116–​117 public nuisance 38 public office 1, 31, 33, 58, 59, 72, 73, 147, 153, 154, 156, 217 public policy 58 public procurement 131, 197, 206, 214, 215 public sector wages/​worker pay 4, 67, 73, 77, 79, 121–​122, 159, 160, 164, 165, 166 public service delivery 59, 109, 124, 146, 175–​176, 176–​177, 218, 223, 228 public service motivation 158, 164 public service provision 217, 218 public spending 73, 149 public transport 123 public trust 17, 22, 25 punishment 16, 18, 23–​24, 67, 102, 103, 121, 147, 149, 156, 157, 184, 206; asymmetric 107, 129; certainty vs. severity of 121, 122–​123 Putin,Vladimir rate of return 75, 77–​79, 81, 93 reciprocity 19, 101–​102, 103, 110 reform 15, 16, 20–​25, 22, 23, 36, 67, 191, 199, 202, 206, 214 regime legitimacy 17

regulation 16, 32, 33, 36, 37, 38, 39, 44, 59, 63, 185n12, 193, 196, 215; financial 88, 89, 90–​91, 93–​94, 96, 214 Regulation 2020/​2092 202–​203 regulatory compliance 2, 20 rent extraction 5, 218 reporting: citizen 121, 126, 128–​130, 193, 205, 206–​207, 208n8, 213, 214, 215; online 129; over 62, 66, 67; self-​ 60, 102 republic of beliefs 103–​104, 128 resistance 5, 24, 226 resources 15, 17, 21, 72, 73, 74, 92, 106, 178, 194, 202; allocation of 87, 95, 228; competition for 93, 124 reverse causality, and context sensitivity 142–​145 ‘revolving doors’ 192, 213 reward–​corruption nexus 122 rewards 61, 63, 66, 102, 108, 120, 122 risk aversion 3, 146–​147 risk-​taking behavior 91 Roman law 34 Romania 116, 122, 189, 197, 198, 199, 200, 202, 209n31, 209n32 rule of law 3, 38, 93, 96, 119, 177, 179, 184, 185n9, 216; in Europe 190, 194, 199, 200, 201–​203, 206, 207, 209n45, 214 rule violations 117, 156, 161 salaries 75, 77, 78, 121–​122, 124 ‘sand the wheels’ hypothesis 87, 89–​93 scandals 17, 144, 147, 189, 191, 192, 193, 197, 200, 203 scrutiny 58, 129, 221, 224, 225, 226 SEB 502 see Special Eurobarometer 502 (SEB 502) secondary socialization 156, 163, 166–​167n4 security 19, 173, 208n2, 216 selection: into public sector jobs 158–​162; worker 158–​159, 163, 164, 166 self-​interest 13, 13–​14n1, 101–​102, 107, 109, 122, 125, 141 self-​reporting 60, 102 severity vs. certainty of punishment 121, 122–​123 SGMM (System Generalized Method of Moments) 174, 179–​180 shame costs 15 shared beliefs 119, 120; see also norms

Index  241 Singapore 24, 106, 131, 156, 161, 189 Slovakia 150n7, 197, 198, 199, 201, 209n31 Slovenia 203, 204, 205 smartcards 127–​128, 133n8 social media 4, 12, 172, 174, 178, 179, 180–​182, 183, 184n1 social networks 177, 180, 185n10 social norms 111, 119–​120, 126, 182 social observability 120, 121, 126 social trust 115–​116 social visibility 126 socialization 153–​154, 156, 162–​163, 165, 166–​167n4 societies 2, 3, 13, 16, 23–​24, 26, 110, 116, 130, 156, 160, 161, 165, 178, 203, 216, 230n19 sociocultural norms 120 South Africa 19, 20, 142, 143 Southern Europe 190, 203–​205, 215 Spain 17, 203, 204, 205 Special Eurobarometer 470, 203 Special Eurobarometer 502 (SEB 502), 190, 191, 192–​193, 196, 197, 198, 199, 200, 203, 206, 208n4, 209n31, 209n32 stability 18, 71, 88, 90, 91, 208n1, 216 staff rotation 121, 127, 129 standard model of compliance 18, 20, 22 state capacity 100, 132 state capture 206, 214 statute law 34 stigma 15, 119, 126, 144 strict authorization 63 strict liability 37, 39 strong institutions 3, 25, 96, 184 subject characteristics 107–​108 sub-​Saharan Africa 15, 17, 19, 179 supervision, of financial sector 90–​91, 95 supporters, of political leaders 2, 7, 10, 11, 11 sustainable development 15, 189, 216 Sustainable Development Goals, of United Nations 15 Sweden 19, 116, 161, 189, 191, 192, 193, 204, 209n35 System Generalized Method of Moments (SGMM) 174, 179–​180 tax burden 73 tax collection 58, 59 tax compliance 2, 19, 115, 116–​117, 120, 121, 133n2, 163, 167n11, 220, 223

tax evasion 2, 16, 20, 23, 58, 59, 73, 74, 76, 79, 117, 120, 121, 133n2, 156, 163, 229n8 tax filings 5, 217, 224–​225 tax morale 19–​20 tax privacy 120–​121 tax rate 73, 77–​79, 81 taxation 19, 20, 58 technologies 1, 5, 75–​76, 100, 121, 221; digital 12, 174, 178, 179; information and communication (ICTs) 4, 172–​185, 227–​228; monitoring 60, 127–​128, 129, 180 Telecommunication and Infrastructure Index 175 TFEU (Treaty on the Functioning of the EU) 190, 193–​194, 196, 205 torts 37–​38 tradeoffs, between collusion control and extortion prevention 62–​63 transparency 4, 5, 67, 68, 115, 121, 130–​131; in Europe 203, 207, 208n1, 213; in financial sector 88, 89, 90–​91, 96; and ICTs 172, 175, 176, 177, 178, 180, 182–​183; and tackling corruption 220, 228, 230n19, 230n20 Transparency International 6, 13, 71, 84n4, 139, 142, 143, 173, 179, 189, 197, 207, 217 Treaty on the Functioning of the EU (TFEU) 190, 193–​194, 196, 205 trespass 37 trust: and corruption 17–​20; institutional 15–​26, 22, 23, 200, 214; mis-​ 15, 18, 19, 20, 21, 22, 25; public 17, 22, 25; social 115–​116 Uganda 19, 164 UNCAC (United Nations Convention against Corruption) 69n22, 216 unconstrained bureaucracy 53 unconstrained dishonest profits 49 underenforcement 64–​65, 65–​66 United Kingdom 4, 21, 25, 31, 107, 119, 147, 155, 161, 163, 166–​167n4, 183 United Nations 174, 177; Convention against Corruption (UNCAC) 69n22, 216; General Assembly 216, 228n1; Sustainable Development Goals 15 United States 4, 19, 37–​38, 55n11, 107, 119, 141, 146, 147, 157, 183 utility 60, 101, 102, 158; expected 18, 74, 76, 101, 102, 117, 124; maximization of

242 Index 118, 124; psychological 108; reciprocity 102 values 119, 146, 147, 153, 154, 155, 165, 216 vicious circle of corruption 15–​26, 22, 23, 72 Virtual Social Network (VSN) 177, 180, 185n10 wages, public sector 4, 67, 73, 77, 79, 121–​122, 159, 160, 164, 165, 166 weak institutions 25, 84n9, 89, 93–​94, 96, 129, 199 Weber, Max 21 WEF (World Economic Forum) 180, 185n10 WGI (World Governance Indicators) 154, 172–​173 whistleblowing 121, 126, 128–​130, 193, 205, 206–​207, 208n8, 213, 214, 215 within-​country settings 106–​107, 119, 175, 179, 181

WJP (World Justice Project) 177, 185n9 women: corruption 131, 139–​140, 141–​142, 143, 144–​145, 146–​147, 148, 149; politicians 144, 146, 147, 149; representation of 131, 141, 142, 143, 144, 145, 146, 147, 148, 149 worker pay, public sector 4, 67, 73, 77, 79, 121–​122, 159, 160, 164, 165, 166 worker preferences 153–​167 worker selection 158–​159, 163, 164, 166 World Bank 18, 25, 71, 93, 154, 172–​173, 176, 184n2, 189, 217, 229n4 World Economic Forum (WEF) 180, 185n10 World Governance Indicators (WGI) 154, 161, 172–​173 World Justice Project (WJP) 177, 185n9 World Values Surveys 16 Zambia 164, 166 zero tolerance 24, 26