66815 Draining Development? Controlling Flows of Illicit Funds from Developing Countries edited by Peter Reuter DRAINING DEVELOPMENT? DRAINING DEVELOPMENT? Controlling Flows of Illicit Funds from Developing Countries Edited by PETER REUTER THE WORLD BANK © 2012 International Bank for Reconstruction and Development / International Development Association or The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org 1 2 3 4 15 14 13 12 This volume is a product of the staff of The World Bank with external contributions. The �ndings, interpretations, and conclusions expressed in this volume do not necessarily re- flect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to the work is given. For permission to reproduce any part of this work for commercial purposes, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Of�ce of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN (paper): 978-0-8213-8869-3 ISBN (electronic): 978-0-8213-8932-4 DOI: 10.1596/978-0-8213-8869-3 Library of Congress Cataloging-in-Publication Data Draining development? : controlling flows of illicit funds from developing countries / edited by Peter Reuter. p. cm. Includes bibliographical references and index. ISBN 978-0-8213-8869-3 — ISBN 978-0-8213-8932-4 (electronic) 1. Tax administration and procedure—Developing countries. 2. Money laundering— Developing countries. 3. Tax evasion—Developing countries. 4. Transfer pricing— Taxation—Law and legislation—Developing countries. I. Reuter, Peter, 1944– HJ2351.7.D73 2011 364.1’33—dc23 2011039746 Cover design by Drew Fasick of the Fasick Design Group. Contents Foreword ix Acknowledgments xi Contributors xiii Abbreviations xv 1. Introduction and Overview: The Dynamics of Illicit Flows 1 Peter Reuter Part I. The Political Economy of Illicit Flows 19 2. Governance and Illicit Flows 21 Stephanie Blankenburg and Mushtaq Khan 3. The Political Economy of Controlling Tax Evasion and Illicit Flows 69 Max Everest-Phillips 4. Tax Evasion and Tax Avoidance: The Role of International Pro�t Shifting 109 Clemens Fuest and Nadine Riedel Part II. Illegal Markets 143 5. Illicit Capital Flows and Money Laundering in Colombia 145 Francisco E. Thoumi and Marcela Anzola v vi Contents 6. Human Traf�cking and International Financial Flows 171 Pierre Kopp Part III. To What Extent Do Corporations Facilitate Illicit Flows? 203 7. Transfer Price Manipulation 205 Lorraine Eden 8. The Role of Transfer Pricing in Illicit Financial Flows 235 Carlos A. Leite 9. Accounting for the Missing Billions 265 Richard Murphy 10. Trade Mispricing and Illicit Flows 309 Volker Nitsch Part IV. Policy Interventions 335 1 1. Tax Havens and Illicit Flows 337 Alex Cobham 12. How Well Do Anti–Money Laundering Controls Work in Developing Countries? 373 Michael Levi 13. The Kleptocrat’s Portfolio Decisions 415 James Maton and Tim Daniel Part V. Conclusions and the Path Forward 455 14. The Practical Political Economy of Illicit Flows 457 Mick Moore Contents vii 15. Policy and Research Implications of Illicit Flows 483 Peter Reuter Boxes 8.1. A Trading Intermediary: A Case of Management Foresight? 238 8.2. A Trading Intermediary: Redux 240 8.3. A Measure of Tax Evasion in the United States 250 8.4. Zambia: Tax Collection in the Copper Sector 256 12.1. A Suspicious Activity Report Leads to a Transnational Corruption Investigation 402 Figures 2.1. Definitions of Capital Flight 36 2.2. Choosing Attainable Benchmarks for Policy Design 60 3.1. The Political Economy of an Effective State and Tax System 85 3.2. The Political Economy Model of Tax Evasion Driving Illicit Capital Flows 87 5.1. Colombia: Outward Foreign Direct Investment, 1994–2010 160 6.1. Human Trafficking: Supply and Demand 194 7.1. Estimating Trade Mispricing from International Trade Data 219 9.1. Corporate Tax Rates, Initial Results, 1997–2008 273 9.2. Corporate Tax Rates, Weighted by GDP, 1997–2008 275 9.3. Corporate Income Tax, Africa, 1980–2005 279 9.4. Top 20 Subsidiaries by Secrecy Jurisdiction, Population, and GDP 300 11.1. Intensity of Economic Exposure: Implied Ratios of Secrecy Jurisdiction Shares to GDP, by Income Group 361 11.2. Intensity of Economic Exposure: Implied Ratios of Secrecy Jurisdiction Shares to GDP, by Region 362 11.3. Zambian Trade: Swiss Role, 1995–2008 363 12.1. The Case of Montesinos: Laundering Maneuvers If Monies Were Collected in Peru 391 viii Contents Tables 1.1. Crossborder Flows of Global Dirty Money 3 2.1. Logical Framework for Identifying Illicit Flows from Developing Countries 53 2.2. Illicit Financial Flows in Different Contexts 58 3.1. Proposed Typology: Illegal and Illicit Capital Flows and Capital Flight 71 5.1. The Results of Anti–money Laundering Efforts: Number of Seized and Forfeited Assets 156 6.1. The Financial Flows Generated by Human Trafficking 182 9.1. Summary Data: Corporate Tax Rate, Including Secrecy Jurisdiction Locations, 2008 276 9.2. Summary Data: Corporate Tax Rate, Excluding Secrecy Jurisdiction Locations, 2008 277 9.3. Summary Data: Tax Rate on Foreign Source Income, Selected Secrecy Jurisdiction Locations, 2008 278 9.4. Multinational Corporations Surveyed by the Tax Justice Network 297 9.5. Secrecy Jurisdiction Locations of Multinational Corporation Subsidiaries 298 10.1. Asymmetries in Australia–European Union Trade: A Statistical Practitioner’s Assessment 316 10.2. Motives for Mispricing in International Trade 320 10.3. Examples of Product Categories in the U.S. Tariff Schedule, 2009 324 10.4. Largest Trade Gaps, 2004 328 10.5. Underreporting of Exports by Country, 2002–06 330 10.6. Correlation of Exporter-Specific Average Trade Gaps 331 12.1. Vulnerabilities in Financial Intermediation 386 14.1. Summary Statistics on Sources of Government Revenue, by Country Category 469 14.2. Major Political, Institutional, and Policy Characteristics Affecting Decisions on the Location of Large-Scale Private Capital 472 Foreword The subject of Draining Development? is one that we feel very strongly about. Estimates on the magnitude of illicit �nancial flows from developing countries vary enormously, but even the most conservative suggest that the total outflow exceeds signi�cantly the amount of of�cial development assis- tance from the Organisation of Economic Co-operation and Development countries. When we look at the �nancing needs of the developing world— what it will take to achieve the Millennium Development Goals—it is clear that development aid alone is insuf�cient. To have any hope of reaching the Millennium Development Goals, poor countries must attack the illicit out- flow of monies and recover what is now illegally held abroad. Nigeria’s experience illustrates the challenges. In the early 2000s, the Nigerian government made a concerted push to recover the funds that ille- gally fled their country. They encountered problems both in Nigeria and abroad: outdated rules and poorly trained staff at home, and overseas there was bank secrecy, opaque corporate and trust vehicles, and time consuming procedures precluding cooperation. The playing �eld was tilted against Nigeria’s national interests. No developing country should have to repeat what Nigeria had to go through to recover funds that rightly belong to its citizens. The effort required thousands of hours of investigative work to locate where the money was hidden, and then the government had to pay millions of dollars in legal fees to recover just a portion of it. Although recovery is becoming easier, too many obstacles remain. Laws governing trusts and corporations in many countries still provide anonym- ity to asset holders. Even where the laws are up to date, enforcement often ix x Foreword continues to lag. A recent Global Witness report on banks and money laun- dering provided a damning indictment of the enforcement efforts of many nations. No matter how successful we are in reducing barriers to asset recovery, it will continue to be a time-consuming and expensive proposition. Differ- ences in legal systems, the demands of due process, the need to respect prop- erty rights—all of these factors ensure that recovery will never be a straight- forward affair. That is why Draining Development? and the information and analysis it provides on illicit �nancial flows are so important. Far better than reducing barriers to asset recovery is to not have to recover the assets in the �rst place. If we can �nd ways to stop the illegal outflow of money, developing coun- tries can refocus their energies on development. But as the chapters in this volume show, money flows out illicitly for many reasons, including tax evasion, the smuggling of illegal goods, the traf- �cking of human beings and other organized crimes, the manipulation of transfer prices and trade mispricing, customs fraud, the failure of money laundering controls, terrorist �nancing, and bribery. In addition to depleting already meager public coffers and hiding the pro�ts of crime, illicit �nancial flows pose a risk to the stability of global �nancial markets; contribute to suboptimal investment decisions; undermine tax morale and accountability between citizen and state; and add to growing income inequality both within and between countries. The consequences are incalculable. It is our contention that the illicit outflow of money from developing countries, in particular, has not received the attention it deserves—either from the development community or from policy makers in developed or developing countries. This book represents an important step toward gar- nering the attention the issue deserves. Norway is proud to have assisted the World Bank in this endeavour, and together we would like to thank the authors and the editor for the �ne work they have produced. Erik Solheim, Minister of the Environment and International Development, Norway Otaviano Canuto, Vice President, World Bank Acknowledgments This project was initiated by Norway’s Ministry of Foreign Affairs, which provided funding to the World Bank. Harald Tollan was the Ministry staff member who oversaw the grant; he gave helpful guidance on the project at several critical points. At the World Bank, Rick Messick served as more than a project monitor. He was an active participant from beginning to end, helping identify authors, commenting on my ideas for topics and providing comments on my own chapters. He also navigated the complexities of the World Bank procure- ment system to enable (all but one) conference participants to actually reach Washington on time, no mean achievement. Amir Farmanes, then a doctoral student at the School of Public Policy at the University of Maryland served as project assistant and played a major role in organizing the September 2009 conference at which the �rst versions of the papers were delivered. The discussants at the conference included: Odd-Helge Fjeldstad (Chris- tian Michelsen Institute, Bergen, Norway); Phil Keefer (World Bank); Grace Pérez-Navarro (OECD, Paris); Mick Keen (IMF); Richard Danziger (Interna- tional Organization for Migration, Geneva, Switzerland); Marijn Verhoeven (World Bank); Ted Moran (Georgetown University); Rob Weiner (George- town University); Dilip Ratha (World Bank); Robert Kudrle (U. Minnesota, Minneapolis); Sony Kapoor (Re-De�ne); Nicos Passas (Northeastern Uni- versity); Victoria Green�eld (US Naval Academy); Gabriel DeMombynes (World Bank); Simon Pak (Pennsylvania State University, Great Valley); Alan Carter (HM Revenue and Custom, London); Jack Blum (Baker, Hostettler, Washington); Maxwell Nkole (former chief of Zambian Anti-Corruption agency). Their comments materially helped strengthen the book. I am also xi xii Acknowledgments grateful to the other speakers at the conference: Nuhu Ribadu (then at the Center for Global Development), Raymond Baker (Global Financial Integ- rity) and Stephen Shay, (U.S. Treasury Department). Stephanie Blankenburg and Mushtaq Khan served as gracious hosts for a preliminary working conference to discuss early drafts of the papers at the School of Oriental and African Studies in London. Max Everest-Phillips in addition to writing one of the book chapters, was invaluable in identifying a number of the other authors for important topics. Finally, I would like to thank Robert Zimmerman for editing the manu- script, and Susan Graham, Stephen McGroarty, and Andres Meneses of the World Bank’s Of�ce of the Publisher for managing the production of the book. Contributors Marcela Anzola is a legal scholar at the University of Texas, Austin. Stephanie Blankenburg is at the Department of Economics, School of Orien- tal and African Studies, University of London. Alex Cobham is Chief Policy Adviser, Christian Aid, London. Tim Daniel is a London-based partner in the U.S. law �rm Edwards Wildman Palmer LLP. Lorraine Eden is Professor of Management, Mays Business School, Texas A&M University, College Station, Texas. Max Everest-Phillips is Director of the Governance and Institutional Devel- opment Division, Commonwealth Secretariat, London. Clemens Fuest is at the Oxford University Centre for Business Taxation, Oxford. Mushtaq Khan is at the Department of Economics, School of Oriental and African Studies, University of London. Pierre Kopp is Professor of Economics, Department of Economics, Panthéon- Sorbonne University, University of Paris 1. Carlos A. Leite is a former economist at the International Monetary Fund and was a tax specialist with Deloitte Touche Tohmatsu Limited, Toronto, at the time of writing. Michael Levi is Professor of Criminology, School of Social Sciences, Cardiff University, Cardiff, United Kingdom. xiii xiv Contributors James Maton is a London-based partner in the U.S. law �rm Edwards Wild- man Palmer LLP. Mick Moore is at the Institute for Development Studies, University of Sus- sex, Brighton, United Kingdom. Richard Murphy is Research Director of the Mapping the Faultlines Project at the Tax Justice Network, London. Volker Nitsch is Professor of Economics, Department of Law and Economics, Technische Universität Darmstadt, Darmstadt, Germany. Peter Reuter is Professor at the School of Public Policy and Department of Criminology, University of Maryland, College Park. Nadine Riedel is at the Oxford University Centre for Business Taxation, Oxford. Francisco E. Thoumi is Tinker Visiting Professor of Latin American Studies, University of Texas, Austin. Abbreviations AFDI Annual Inquiry into Foreign Direct Investment (U.K. Of�ce for National Statistics) ALP arm’s-length principle AML anti–money laundering ATM automated teller machine BLS U.S. Bureau of Labor Statistics BNP BNP Paribas (Suisse) SA CBN Central Bank of Nigeria c.i.f. cost, insurance, and freight CIT corporate income tax DNE National Narcotics Directorate (Colombia) DTA double tax agreement EU European Union FATF Financial Action Task Force FDI foreign direct investment FDN Nicaraguan Democratic Foundation Fiscalía Attorney General’s Of�ce (Colombia) FIU �nancial intelligence unit f.o.b. free on board FSI �nancial secrecy index GDP gross domestic product HS harmonized system (product classi�cation) IAS International Accounting Standard (International Accounting Standards Board) IFF illicit �nancial flow IFRS International Financial Reporting Standard xv xvi Abbreviations ILO International Labour Organization IMF International Monetary Fund IOM International Organization for Migration IRS Internal Revenue Service (United States) KACC Kenya Anti-Corruption Commission MNC multinational corporation MNE multinational enterprise MOF Ministry of Finance (Zambia) NGO nongovernmental organization OECD Organisation for Economic Co-operation and Development OFC offshore �nancial center OFDI outward foreign direct investment PEP politically exposed person SAR special administrative region (China) SAR suspicious activity report SITC Standard International Trade Classi�cation TJN Tax Justice Network TPM transfer price manipulation UNCAC United Nations Convention against Corruption UNCTAD United Nations Conference on Trade and Development VAT value added tax Note: All dollar amounts are U.S. dollars (US$) unless otherwise indicated. 1 Introduction and Overview: The Dynamics of Illicit Flows Peter Reuter Abstract A remarkable consensus emerged during the �rst decade of the 21st cen- tury on the need for the wealthy countries of the world to increase their aid sharply to developed states. At Monterrey, Mexico, in 2002, the world’s leaders committed to a “substantial increase in of�cial develop- ment assistance [to help] developing countries achieve internationally agreed development goals and objectives� (United Nations 2003, 14). Three years later, at Gleneagles, United Kingdom, the heads of the G-8 nations reaf�rmed this commitment, and in Doha, Qatar, in 2008, the nations of the world again recognized the need to increase development aid and pledged to do so. However, although development assistance steadily increased over the decade, rising from US$58 billion in 2000 to a projected US$125 billion in 2010, it still falls far short of what the developed world promised, and what many believe poor countries need if their citizens are to escape poverty. As it became clear over the decade that the wealthy world’s commit- ments would not match the rhetoric, the development community began searching for other sources of funding to �ll the development �nance gap. At the same time, reports of massive, illegal outflows from developing 1 2 Draining Development countries began appearing. In 2004, Transparency International estimated that 10 of the most notoriously corrupt heads of state in developing coun- tries may have, together, spirited as much as US$60 billion out of their countries during their respective tenures in of�ce (Transparency Interna- tional 2004). In 2005, Raymond Baker estimated that more than US$540 billion flowed out of developing countries each year thanks to a combina- tion of tax evasion, fraud in international trade, drug traf�cking, and cor- ruption (Baker 2005). In 2007, Christian Aid and the Tax Justice Network produced studies reporting similar �gures (Kapoor 2007; TJN 2007). These reports and their implications were not lost on those looking for ways to �ll the development �nance gap. If, in fact, the amount of money illicitly flowing out of developing countries was anywhere near the amounts estimated in these reports, even partially staunching the flow held signi�cant promise for �lling the gap. Two questions thus immediately arose. Are the outflows large enough to justify efforts to staunch them? And, if so, what can be done? In 2008, the Norwegian government asked the World Bank to under- take a research project that would address these questions. The Bank, in turn, commissioned the editor of this book to organize a conference with authors of original papers and to edit the proceedings. The purpose of this book is to assess what is known about the composition of illicit flows, the processes that generate these flows, the role of tax havens in facilitating them, and the effectiveness of programs aimed at either pre- venting the flows or locating and recouping them once they have left. The book provides the �rst collection of analytic contributions, as opposed to advocacy essays and black box estimates, on illicit �nancial flows (IFFs). Some of the chapters present new empirical �ndings; oth- ers, new conceptual insights. All of them enrich the understanding of the dynamics of the illicit flows phenomenon. The book does not offer a new estimate of the global total of these flows because the phenomenon is too poorly understood. The chapters are based on papers �rst presented at a September 2009 conference at the World Bank. Each paper had one or two assigned dis- cussants, and the revisions reflect the often searching critiques of the discussants, as well as additional comments from the editor and from two external peer reviewers. The chapters have been written to be acces- sible to nonexperts. Introduction and Overview: The Dynamics of Illicit Flows 3 Following this introduction, the book has �ve parts: I. The Political Economy of Illicit Flows; II. Illegal Markets; III. To What Extent Do Cor- porations Facilitate Illicit Flows?; IV. Policy Interventions; and V. Con- clusions and the Path Forward. The Short History of Illicit Financial Flow Estimates Raymond Baker’s Capitalism’s Achilles Heel: Dirty Money and How to Renew the Free-Market System, published in 2005, gave shape to the topic. Baker reported on the results of 550 interviews he had conducted with senior business executives around the world in the early 1990s to estimate the extent of illicit flows involving corporate mechanisms, mostly in transactions between unrelated parties. He examined govern- ment and academic reports on other sources of the flows, such as illegal markets, and, accompanied by explicit statements on the reasons for his judgments, offered assessments of the amounts that might be flowing internationally out of developing countries as a consequence. A sum- mary of his quantitative �ndings is provided in table 1.1. Baker estimates that over 60 percent of total illicit flows arise from legal commercial activities, and most of the remainder from criminal activity. The former leave the developing world through three channels: the mispricing of goods traded between independent parties, the distor- tion of transfer prices charged on goods traded within a multinational �rm, and fraudulent transactions. Baker provides minimal detail about the derivation of these �g- ures. Thus, there is no information on the methods used to convert the Table 1.1. Crossborder Flows of Global Dirty Money US$, billions Global From developing or transitional countries Type High Low High Low Criminal 549 331 238 169 Corrupt 50 30 40 20 Commercial 1,000 700 500 350 Total 1,599 1,061 778 539 Source: Baker 2005. 4 Draining Development interviews with 550 senior corporate executives in 11 nations into esti- mates of trade mispricing in the global economy. This is not a simple task. For instance, the nature of the sample is critical to such an exercise (in terms of dimensions such as the commodities and services involved or the countries with which these are traded). Similarly, between the early 1990s, when Baker did his interviews, and 2005, when the estimate was published, there were many changes that could have either raised or lowered the share of mispriced transactions. For the criminal revenues, Baker uses �gures that are often cited in the semiprofessional literature. Consider drug markets, for example: this represents the most well studied illegal industry. Baker uses �gures from the United Nations Of�ce on Drugs and Crime that cite a total of US$400 billion in retail sales and US$120 billion in wholesale revenues. There is, however, a critical literature suggesting that these are substantial over- estimates, perhaps twice the true value.1 As Baker and his colleagues have acknowledged, given the scarcity of available data, the numbers Baker presents are not precise estimates, but, rather, are indications of orders of magnitude that are meant to prompt academic researchers, the International Money Fund (IMF), and the World Bank to collect more accurate and complete data and devise more rigorous estimating techniques. Nonetheless, as is the norm with such work, popular accounts of Baker’s conclusions have focused on the money amounts, which are in the hundreds of billions of dollars, and ignored the caveats that accompany them. Moisés Naím’s Illicit: How Smugglers, Traf�ckers, and Copycats Are Hijacking the Global Economy, also published in 2005, has added to the prominence of the topic. Naím’s emphasis is on the expansion of the variety of illicit international trade involving banned goods, such as drugs, or the counterfeit and theft of intellectual property. He offers no original estimates, but includes alarmingly high numbers, such as an international trade volume in illegal drugs amounting to US$900 billion. Though Naím’s book does not use the term “illicit �nancial flow,� it rein- forces the sense that there are large flows of dirty money from the devel- oping world to the developed nations. Global Financial Integrity, an organization founded by Raymond Baker in 2006, has turned out a number of reports on illicit flows.2 These have Introduction and Overview: The Dynamics of Illicit Flows 5 attracted a great deal of media attention. Beyond one that is much dis- cussed in this volume (Kar and Cartwright-Smith 2008), it has also pub- lished studies showing the importance of these flows to speci�c countries and regions (on Africa, for example, see Kar and Cartwright-Smith 2010). However, as with the estimates in Baker’s book, popular accounts generally ignore the caveats attached to the estimates in favor of the raw numbers. Though several contributors to this volume have raised questions about the validity of the current estimates, there is no doubt that these numbers have helped galvanize attention on illicit flows and ways to deal with them. In addition to the work mentioned above, there has been a steady flow of reports from organizations such as the Tax Justice Network (for example, TJN 2007) and Christian Aid that focus, in particular, on the role of contracts involving multinational corporations. These contracts have apparently allowed the exploitation of natural resources by the cor- porations through failure to specify properly the price at which the gold, timber, and so on is to be exported from developing countries. The transfer may be legal, but, it is often alleged, the underlying contract is the result of corrupt dealings between of�cials and the multinational corporations; the flows are thus appropriately classi�ed as illicit. One disappointing note is that, since the appearance of Baker’s vol- ume in 2005 to broad acclaim, our volume is the �rst substantial attempt to address the issue from a scholarly perspective. Of the over 100 Google Scholar citations to Capitalism’s Achilles Heel, none is from a major aca- demic journal.3 When this volume was in �nal preparation, an odd af�rmation of the reality of IFFs appeared fleetingly on the Internet site of the Bank of China.4 The headline-catching sentence was as follows: According to a research report published by the Chinese Academy of Social Sciences, since the middle of the 1990s, the overall number of the escaped Party and Government cadres, of�cials in the judicial and public security branches, senior-level administrators in the state-owned enterprises, as well as staff in the Chinese institutions stationed abroad, added up to 16,000–18,000, and these corruptors have taken with them around RMB 800 billion (circa US$100 billion). The report then detailed many major cases, identifying not only the offenders, but also the methods that they had used to move the money 6 Draining Development out of the country. The RMB 800 billion was not intended to be an esti- mate of the gross flows. It was an estimate of how much had been detected in roughly a 10-year period, surely a modest fraction of all IFFs, particularly because it included only of�cial corruption and not other sources such as tax evasion and criminal earnings. The Policy Response to Illicit Financial Flows Norway has led government efforts to focus attention on the issue. In 2008, the Norwegian government created a commission to prepare a report on illicit flows.5 It has also funded, with other nations, the Task Force on Financial Integrity and Economic Development, for which Global Financial Integrity serves as the secretariat, and funded the confer- ence that led to this volume.6 The governments of Germany and the Neth- erlands have also paid attention to the topic. This is reflected, for example, by Germany’s sponsorship of a side event on tax flight at the fall 2008 meetings of the World Bank and IMF and the funding by the Netherlands of an ongoing World Bank study on illicit flows in East Africa in 2010. Despite (or because of ?) the absence of serious scholarly attention, illicit flows have become a topic of high-level policy discussion. For example, the G-20 Pittsburgh summit in September 2009, near the height of the global �scal crisis, addressed the issue in its communiqué.7 It was also discussed at the October 2009 annual meetings of the World Bank and IMF Boards (Development Committee 2010). IFFs have often been identi�ed as a contributing factor in the current global �nancial crisis and a source of instability in the world �nancial system because of the illicitness.8 Tax havens, more appropriately called secrecy jurisdictions, have been inculpated in many major scandals over the last decade; they have been under attack because they can undermine effective �nancial regulation in other nations, both developed and devel- oping.9 Once the spotlight is turned on them, it is hard not to notice that they also serve as the destination for the bribes received by many dic- tators; for example, James Maton and Tim Daniel, in chapter 13 in this volume, describe how Sani Abacha, the dictator of Nigeria in the late 1990s, kept substantial liquid funds on the Isle of Jersey. More importantly, IFFs have been identi�ed as a major impediment to growth and to the development of sound �nancial systems and gover- Introduction and Overview: The Dynamics of Illicit Flows 7 nance in developing countries, an issue dealt with by Mick Moore in chap- ter 14 in this volume. Attention to the issue has also grown as the demand for greater resources to �nance development has collided with the shrink- ing volume of funds arriving from developed countries following the �nancial crisis and, more generally, with the lack of support by taxpayers. The greater attention has highlighted that illicit flows may be engendered by of�cial development assistance given that aid represents, in some coun- tries, a substantial share of the money available to central governments. It also highlights that misuse of the funds helps undermine support for for- eign aid. Still, if there are large flows of illicit funds out and if they can be curbed, the development �nance gap can be met painlessly. De�nition is an important source of controversy in the study of this phenomenon.10 “Illicit �nancial flows� is an ill-de�ned term, and the boundaries are disputed. “Illicit� is not the same as “illegal.� No one will deny that the deposit of Mobutu Sese Seko’s corrupt earnings in Swiss banks is an illicit flow. However, there will be less agreement, for exam- ple, about flows that represent efforts to evade the arbitrarily adminis- tered taxes and regulations of corrupt governments; efforts to evade the oppressive economic legislation of President Ferdinand Marcos might have been illegal, but they were not necessarily illicit. Perhaps the de�n- ing characteristic of illicit is that (1) the acts involved are themselves illegal (corruption or tax evasion) in a regime that has some democratic legitimacy, or (2) the funds are the indirect fruits of illegal acts (for example, bene�ts given to those who have provided illegal funding for a presidential election). Thus, illicit funds are not merely the consequence of bad public policy and do not include all international illegal �nancial flows from illegitimate regimes. The questions about de�nition are taken up in a number of chapters, notably, those of Stephanie Blankenburg and Mushtaq Khan (chapter 2) and Max Everest-Phillips (chapter 3). The remainder of this chapter gives brief summaries of the individual chapters. The Political Economy of Illicit Flows In chapter 2, Stephanie Blankenburg and Mushtaq Khan, economists at the School of Oriental and African Studies, in London, develop an ana- lytical framework that identi�es core determinants and drivers of illicit 8 Draining Development capital flight from developing countries. They begin with an extensive history of the use of the term capital flight and the variety of concepts and measures that have been attached to this label. This is an essential starting point because the central estimates of the scale of illicit flows are based on methods developed to estimate the volume of capital flight. Moreover, the concept of capital flight has morphed a number of times since it was �rst formulated by Kindleberger (1937); even now, there is considerable ambiguity as to whether capital flight refers to short- or long-term capital movements. Blankenburg and Khan then relate systematically the concepts of dirty money, illegal capital flows, and capital flight, which are often used almost synonymously, but which represent different behaviors and chal- lenges. Some de�nitions are legalistic and rule oriented; others are broad and oriented toward de�ning the phenomenon in terms of motivations or effects. A signi�cant component of the illicit funds leaving developing coun- tries results from the structural features of the societies of these countries, such as laws that advance the interests of a ruling elite; paradoxically, these features cause domestic capital to seek pro�table investment oppor- tunities in more advanced economies. Policy measures to control and reduce illicit capital flight from developing economies will be effective only to the extent that they take account of these structural investment constraints. Market-improving good governance reforms are insuf�cient and can sometimes even be counterproductive in this respect. The heart of Blankenburg and Khan’s argument is that, in many developing countries, governments lack legitimacy; government policies do not represent the result of the working out of a bargain among vari- ous interest groups, but, rather, the imposition of the power of a small set of economic actors. Illegal �nancial flows may not be illicit flows because of the state’s lack of legitimacy. Max Everest-Phillips, a governance adviser at the U.K. Department for International Development at the time of writing, also examines, in chap- ter 3, the determinants of IFFs through a political economy analysis. His focus, though, is tax evasion. Reviewing the experiences of a large array of countries and tax systems, he argues that the root cause of all illicit capital outflows is ultimately not poor policy or capacity constraints in adminis- tration, but the failure of political will. Controlling tax evasion (and the Introduction and Overview: The Dynamics of Illicit Flows 9 other dimensions of illicit flows: criminality and corruption) requires political incentives to build an effective state. An effective state includes effective tax systems, which derive from formal and informal institutional arrangements (political settlements) that establish the ambition to pro- mote prosperity and raise public revenue. Such a commitment arises if political leaders and taxpayers perceive the need for effective tax systems so as to provide the state with the resources required to enforce property rights, deliver political stability, and promote economic growth. The extent and the form of tax evasion derive from the political consensus to tax effectively and to develop the administrative capacity to do so. In turn, this shapes and reflects the intrinsic willingness to pay taxes (tax morale) of taxpayers. The political economy of controlling illicit capital flows, including tackling tax evasion, demonstrates the necessity of addressing ineffective governance and the weak state legitimacy of many developing countries. If regime leaders and elites are not prepared to tax themselves and prevent free-riding, genuine political ownership of efforts to curb illicit capital flows will be problematic. In chapter 4, Clemens Fuest and Nadine Riedel, economists at Oxford University, assess tax evasion and avoidance in developing countries and the role of international pro�t shifting, which may serve as a channel for illicit �nancial flows. They begin, as does Volker Nitsch in chapter 10, by expressing skepticism about the existing estimates of the extent of tax losses to developing countries that occur as a result of the movement of capital from developing countries to tax havens. The available data are limited, and estimates reflect broad assumptions, often selected in a way that generates high estimates. For example, it is often assumed that the capital would, if kept at home, pay the full tax rate, even though there are numerous loopholes that allow corporations to incur much lower tax obligations. Empirical evidence on the magnitude of the problem and on the factors driving income shifting is scarce and con�ned to studies of developed nations. Fuest and Riedel discuss methods and available data sets that can be used to gain new insights into the problem of corporate income shifting. The authors argue that the results of many existing studies on tax evasion and avoidance in developing countries are dif�- cult to interpret, mainly because the measurement concepts used have a number of drawbacks. They discuss alternative methods and data sets 10 Draining Development and present descriptive evidence that supports the view that pro�t shift- ing takes place from many developing countries into tax havens. Illegal Markets The international flows of funds in illegal markets, apart from the mar- kets for drugs, have not been the subject of systematic empirical study. Nonetheless, it is routinely asserted that they also generate large flows from developing countries, where cocaine and heroin are produced and from which many humans are traf�cked. For example, as noted above, Raymond Baker (2005) estimates that illegal markets account for US$168 billion–US$231 billion of the illicit flows from developing nations. In “Illicit Capital Flows and Money Laundering in Colombia� (chap- ter 5), Francisco Thoumi, a Colombian economist, and Marcela Anzola, a Colombian legal scholar, assess illicit flows in Colombia primarily gen- erated by the trade in illegal drugs. They show that Colombia’s principal problem is the inflow rather than outflow of drug moneys and that this has had important economic and political consequences for the nation. For example, these flows have exacerbated the concentration of land ownership in rural areas and, by providing access to foreign funds, also exacerbated the long-running insurgency. Though Colombia has enacted model asset-recovery and anti–money laundering (AML) control laws, a detailed examination of the record of the last decade indicates that it has failed to provide a serious threat to the �nancial well-being of traf�ckers. In turn, this reflects the failure of the Colombian state to establish norms of compliance with �nancial laws. Thoumi and Anzola argue that the key to these failures of well- written laws, as well as an important factor in Colombia’s prominence in the international cocaine trade, is the broad lack of compliance with legal norms and the general weakness of the state. In “Human Smuggling and International Financial Flows� (chapter 6), Pierre Kopp, an economist at the University of Paris (Sorbonne), examines the markets for human smuggling in the �rst such essay by an economist. These markets are varied, ranging from successfully assisting workers to cross illegally into the United States to coercively employing sex workers from Thailand in Belgium. Drawing on basic economic con- cepts, Kopp shows that the �nancial flows vary by type of traf�cking. In Introduction and Overview: The Dynamics of Illicit Flows 11 some of these markets, there will be flows back to the country of human origin, but, in others, the value added will occur and remain in the des- tination country. Kopp concludes that there are no authoritative esti- mates of the scale of these markets and little prospect that such estimates will be generated either globally or at the national level; prices for ser- vices vary substantially, and there is no systematic basis for estimating the scale of the human traf�cking itself. To What Extent Do Corporations Facilitate Illicit Flows? The chapters in part III mostly deal with inherently technical questions that cannot be avoided in seeking to understand the flow between the developing and developed worlds. In particular, transfer pricing, the methods by which multinational �rms price transactions between af�li- ates in different countries with varying tax rates, may permit large trans- fers of taxable revenue that are properly viewed as illicit, even if not for- mally illegal. Trade mispricing, in which export or import documents carry false prices, also may be important both as a source of tax evasion and as a channel for movement of illicit funds. In “Transfer Price Manipulation� (chapter 7), Lorraine Eden of the Mays Business School at Texas A&M University provides an overview of the methods by which �rms set prices in these internal, but international, transactions. There is a well-established set of theoretical principles and, in some nations, generally in the Organisation for Economic Co-operation and Development (OECD), a detailed set of rules to implement these principles. However, in the real world, there are many potential sources of deviation that allow for considerable discretion and potential abuse. Moreover, some countries, mostly in the developing world, have not cre- ated rules speci�c to the setting of these prices. Eden reviews the thin empirical literature on the extent of income transfer by transfer price manipulation (TPM). Whether through analy- sis of individual transactions or more aggregate methods, most studies �nd evidence that TPM occurs in response to changing corporate tax rate differentials. However, the research task is complicated by the large variety of factors that influence a corporation’s incentives for TPM; ad valorem taxes, restrictions on the repatriation of pro�ts, and political instability all play a role. The literature is also dominated by studies 12 Draining Development focused on where the data are available, particularly the United States. Eden concludes that TPM occurs, but she is not able to estimate how important it is anywhere, let alone in individual developing countries. More detail that is speci�c to the case of developing countries is pro- vided in “The Role of Transfer Pricing in Illicit Financial Flows� (chapter 8) by Carlos Leite, a former IMF economist and Deloitte Touche Toh- matsu analyst in Toronto. Leite illustrates the inherent uncertainties in many dimensions of transactions. For example, what is a reasonable method of allocating risk among different af�liates in a multinational corporation? He uses the example of an oil company in a developing country. The company entered into a complex contract that generated unexpectedly high pro�ts in the country, which lacked any speci�c regu- lations for transfer pricing. The company was thus able to shift the prof- its to the country in which its oil trading subsidiary resided and in which it may have paid only a 1 percent tax. Should this be treated as tax avoid- ance, particularly given the �rm’s legitimate fear of expropriation by the new regime in the developing country, or tax evasion, given that any reasonable set of transfer pricing rules would have required declaration of the income in the developing country? Leite’s chapter emphasizes the extent to which many transfer pricing decisions that affect developing countries lie outside the realm of the OECD principles. Richard Murphy, an accountant and a principal analyst in the Tax Justice Network effort to bring attention to the corporate and tax haven roles in illicit flows, tackles the scale of the corporate flows in “Account- ing for the Missing Billions� (chapter 9). Basing his analysis on the approach of an auditor, Murphy �rst tests the hypothesis that substantial transfer mispricing by major corporations might contribute to a loss of at least US$160 billion a year to developing countries in the context of the total likely corporate pro�ts tax paid or not paid worldwide in a year. Second, he explores the hypothesis that activities in developing coun- tries and in the extractive industries in particular might be especially prone to this abuse. This reflects both the lack of any effective monitor- ing capability in many developing countries and the fact that transfer prices can greatly affect not only taxes on corporate pro�ts, the usual focus of transfer mispricing, but royalties and other price-sensitive taxes. Third, Murphy considers whether this sum could be hidden from view within the accounts or �nancial statements of the multinational corpo- Introduction and Overview: The Dynamics of Illicit Flows 13 rations that might be perpetrating the mispricing. Finally, he explores the possibility that the corporations might use secrecy jurisdictions (usually called tax havens) to assist in hiding these transactions from view. In each case, Murphy �nds that the hypothesis is plausible and that, as a consequence, losses of the estimated amount are also plausible, although not proven to exist as a result of this work. In addition, the chapter identi�es the potentially important role of the Big Four �rms of accountants with regard to this issue based on their signi�cant presence in secrecy jurisdictions, their role as auditors of most of the world’s mul- tinational corporations, and their important role in setting International Financial Reporting Standards. Volker Nitsch, an economist at Technische Universität Darmstadt, Ger- many, addresses a related issue in “Trade Mispricing and Illicit Flows� (chapter 10). A potential vehicle for the unrecorded movement of capital out of a country is the falsi�cation of invoices in international trade trans- actions. In contrast to transfer pricing, this involves transactions between formally unrelated parties. Exporters may understate the export revenue on their invoices (whether by giving low �gures for the price or quantity), and importers may overstate import costs, while their trading partners are instructed to deposit the balance for their bene�t in a foreign account. The chapter critically reviews empirical approaches to quantify the extent of trade mispricing. Various reasons for discrepancies in bilateral trade statis- tics are discussed, and incentives for faking trade invoices other than capi- tal flight are highlighted. Overall, the accuracy and reliability of estimates of IFFs based on trade mispricing are questioned. Policy Interventions Alex Cobham, an economist with Christian Aid, examines the role of tax havens in “Tax Havens and Illicit Flows� (chapter 11). This chapter makes three contributions. First, it provides a brief, critical survey of the state of knowledge on the impact of havens on development. Second, it uses existing data to extend that knowledge by examining the detail on bilat- eral trade and �nancial flows between havens and developing countries and identifying the exposure of developing countries of different types. Third, the chapter sets out a research agenda that would allow greater certainty about the scale of the impact of tax havens on development. 14 Draining Development Two results emerge. First, the exposure of developing countries to tax havens is on a par with, if not more severe than, that of high-income OECD countries. Thus, efforts to ensure that developing countries ben- e�t from moves to require greater transparency in terms of international tax cooperation in particular may be of great value. Second, the differ- ences in developing-country exposure across different income groups and regions are substantial, and recognition of this must lead to more detailed and careful study and, over time, policy responses. OECD member countries and others, through the intergovernmental Financial Action Task Force, have invested considerable effort in the implementation of a comprehensive set of laws and regulations against money laundering in all nations; this is seen as a major tool to prevent illicit outflows.11 Michael Levi, a criminologist at Cardiff University, in “How Well Do Anti–Money Laundering Controls Work in Developing Countries?� (chapter 12), examines how well the AML system in devel- oping countries has worked for those countries. Levi �nds evidence that, for many countries, perhaps particularly those most in need of effective AML controls, the regime has minimal capacity to detect or punish violations. Indeed, there are few instances in which domestic AML controls have generated cases against kleptocrats or their families. Levi examines �ve cases of grand corruption to show the extent to which the flows of funds could plausibly have been detected by a domestic system of rules and �nds the results quite mixed. He concludes by not- ing the many dif�culties of creating an effective AML regime if the gov- ernment is thoroughly corrupted, reinforcing a message in the Daniel and Maton chapter. Though illicit flows are a relatively recent issue, components of the response to the flows have been prominent for some time. In particular, there has been growing emphasis since the 1990s on the recovery of sto- len assets and the spread of AML controls around the globe. Tim Daniel and James Maton, London-based lawyers at Edwards Wildman UK who have represented governments in successful suits against corrupt of�cials in a number of developing countries, describe and analyze experiences in attempting to recover stolen assets in “The Kleptocrat’s Portfolio Decision� (chapter 13). Though success is rarely complete in that not all the assets are recovered, Daniel and Maton �nd growing willingness on the part of British and Swiss courts to render Introduction and Overview: The Dynamics of Illicit Flows 15 judgments that facilitate the collection of such assets. They provide a particularly detailed and compelling account of the recovery of assets from the estate and family of Sani Abache, the military ruler of Nigeria in the 1990s. The experience in recent cases suggests that the use of secrecy jurisdictions to hide bene�cial ownership is often only a modest obstacle to recovery. More often, the problem lies in the home country, where the successor government may be unwilling to provide the legal and evidentiary support for effective recovery or where the local courts thwart overseas suits; the case against Tommy Suharto in Indonesia, whereby a Guernsey court froze assets while waiting for action by the Indonesian government, which is still not forthcoming years later, pro- vides one illustration of the problem. Conclusions Mick Moore, a tax economist at the Institute of Development Studies at the University of Sussex, assesses some of the consequences of IFFs in “The Practical Political Economy of Illicit Flows� (chapter 14). The increasing scope to expatriate capital illicitly exacerbates problems of corruption, low investment, the unequal sharing of tax burdens across different parts of the private sector, the low legitimacy of private enter- prise, and relatively authoritarian and exclusionary governance. The international community is already developing a range of interlock- ing tools to deal with the nexus of problems around illicit capital flows, capital flight, corruption, money laundering, tax avoidance, tax havens, and transfer mispricing. Improvements in the design of these tools and greater vigor in implementation should have especially bene�cial effects within many of the poorest countries, notably, in increasing private investment and economic growth, reducing the popular mistrust of private enterprise, and providing more space for more democratic gov- ernance. More effective international action against illicit capital flows would be complementary rather than competitive with attempts to improve from within the quality of public institutions in the poorest countries. The �nal chapter (15) presents the editor’s overview of the topic and how the issues should be dealt with in terms of future research. The chapter argues that illegal markets, though signi�cant in absolute terms, 16 Draining Development are unlikely to be the source of substantial illicit flows out of developing countries because most of the value added is in rich countries, and much of it that is earned by developing county nationals is repatriated home. It also argues that, whatever the problems with existing IFF estimates, the phenomenon is large enough to command serious attention. Illicit flows is a coherent concept in policy terms; what links the international outflows from tax evasion, corruption, and drug markets is the fact that they can all, in principle, be stopped by the same set of policies and laws. A central question is whether there is a set of measures that has the pros- pect of making a substantial difference to these flows. The alternative view is that these outflows are much less of a problem than the underly- ing phenomena that generate them and that the debate on outflows may actually divert attention from these phenomena. The chapter suggests a research path for clarifying this issue. Illicit �nancial flows have complex origins and varied consequences and present dif�cult policy options. The research program appropriate to study them will also be complex, varied, and dif�cult. Notes 1. For an early critique, see Reuter (1996). For a much more detailed analysis pub- lished after 2005, see Kilmer and Pacula (2009). 2. See http://www.g�p.org/. 3. A 2010 paper written for the United Nations Department of Economic and Social Affairs by a well-known development economist does make extensive use of Baker (see FitzGerald 2010). 4. The odd feature was that the item was withdrawn within a few hours of appear- ing on the website. Even in that brief period, it captured public attention. See, for example, Chen and Pansey (2011). Apparently, the report, “Investigation on the Asset Transfer Routes Used by Corrupt Of�cials and the Corresponding Sur- veillance Methods,� was given a Bank of China award for the quality of the research but was never intended to be made public. The translated report can be found on the website of this volume, http://go.worldbank.org/N2HMRB4G20. 5. See Norway, Minister of the Environment and International Development, 2009, Tax Havens and Development: Status, Analyses and Measures, Oslo: Gov- ernment Commission on Capital Flight from Poor Countries, http://www .regjeringen.no/upload/UD/Vedlegg/Utvikling/tax_report.pdf. 6. See Task Force on Financial Integrity and Economic Development, Washington, DC, http://www.�nancialtaskforce.org/. Introduction and Overview: The Dynamics of Illicit Flows 17 7. Leaders’ Statement, G-20 Pittsburgh Summit, September 24–25, 2009, paragraphs 22 and 42, http://www.g20.org/Documents/pittsburgh_summit _leaders_statement_250909.pdf. 8. See “Final Report from the Task Force on the Development Impact of Illicit Financial Flows� at http://www.leadinggroup.org/IMG/pdf_Final_report_Task _Force_EN.pdf. 9. Wall Street Journal, 2009, “Tax Havens Pledge to Ease Secrecy Laws,� March 13, http://online.wsj.com/article/SB123685028900906181.html. 10. A typical de�nition of illicit, presented at dictionary.com, shows two meanings. “1. not legally permitted or authorized; unlicensed; unlawful,� and “2. disap- proved of or not permitted for moral or ethical reasons.� See http://dictionary .reference.com/browse/illicit. 11. See Financial Action Task Force, Paris, http://www.fatf-ga�.org/. References Baker, R. W. 2005. Capitalism’s Achilles Heel: Dirty Money and How to Renew the Free- Market System. Hoboken, NJ: John Wiley & Sons. Chen, L., and G. Pansey. 2011. “Capital Flight Accompanies Corruption in China, Bank Warns.� EpochTimes, June 25. http://www.theepochtimes.com/n2/china /capital-flight-accompanies-corruption-in-china-bank-warns-57712.html. Development Committee (Joint Ministerial Committee of the Boards of Governors of the Bank and the Fund on the Transfer of Real Resources to Developing Countries). 2010. “Strengthening Governance and Accountability Review, Results, and Roadmap.� Report DC2010–0007, April 21. http://siteresources .worldbank.org/DEVCOMMINT/Documentation/22553919/DC2010-0007% 28E%29Governance.pdf. FitzGerald, V. 2010. “International Tax Cooperation and International Development Finance.� Paper prepared for the 2010 World Economic and Social Survey, United Nations Department of Economic and Social Affairs, New York. http:// www.un.org/esa/analysis/wess/wess2010workshop/wess2010_�tzgerald.pdf. Kapoor, S. 2007. “Haemorrhaging Money: A Christian Aid Brie�ng on the Problem of Illicit Capital Flight.� Christian Aid, London. Kar, D., and D. Cartwright-Smith. 2008. “Illicit Financial Flows from Developing Countries, 2002–2006.� Global Financial Integrity, Washington, DC. http://www . g�p.org/storage/g�p/economist%20-%20�nal%20version%201-2-09.pdf. ———. 2010. “Illicit Financial Flows from Africa: Hidden Resource for Devel- opment.� Global Financial Integrity, Washington, DC. http://www.g�p.org /storage/g�p/documents/reports/g�_africareport_web.pdf. Kilmer, B., and R. Pacula. 2009. “Estimating the Size of the Global Drug Market: A Demand-Side Approach.� In A Report on Global Illicit Drug Markets, 1998–2007, ed. P. Reuter and F. Trautmann, 99–156. Brussels: European Commission. 18 Draining Development Kindleberger, C. P. 1937. International Short-Term Capital Movements. New York: Augustus Kelley. Naím, M. 2005. Illicit: How Smugglers, Traf�ckers, and Copycats Are Hijacking the Global Economy. New York: Anchor Books. Reuter, P. 1996. “The Mismeasurement of Illegal Drug Markets: The Implications of Its Irrelevance.� In Exploring the Underground Economy: Studies of Illegal and Unreported Activity, ed. S. Pozo, 63–80. Kalamazoo, MI: W. E. Upjohn Institute for Employment Research. TJN (Tax Justice Network). 2007. Closing the Floodgates: Collecting Tax to Pay for Development. London: TJN. http://www.innovative�nance-oslo.no/pop.cfm? FuseAction=Doc&pAction=View&pDocumentId=11607. Transparency International. 2004. Global Corruption Report 2004. London: Pluto Press. http://www.transparency.org/publications/gcr/gcr_2004. United Nations. 2003. “Monterrey Consensus of the International Conference on Financing for Development: The Final Text of Agreements and Commitments Adopted at the International Conference on Financing for Development, Monterrey, Mexico, 18–22 March 2002.� Department of Public Information, United Nations, New York. http://www.un.org/esa/ffd/monterrey/Monterrey Consensus.pdf. Part I The Political Economy of Illicit Flows 2 Governance and Illicit Flows Stephanie Blankenburg and Mushtaq Khan Abstract The concern about illicit capital flows from developing countries reflects a variety of relevant policy issues, but is often motivated by weakly for- mulated underlying analytical frameworks. We review the literature on illicit capital flows and suggest that the common underlying concern that motivates the different approaches is the identi�cation of flows that potentially damage economic development. Implicitly, if these flows could be blocked, the result would be an improvement in social out- comes. Illicit flows can be illegal, but they need not be if the legal frame- work does not reflect social interests or does not cover the relevant flows. A minimal de�nition of an illicit capital flow has to consider both the direct and the indirect effects of the flow and has to assess these effects in the context of the speci�c political settlement of the country in question. To demonstrate the implications in simpli�ed form, we distinguish among advanced countries, intermediate developers, and fragile devel- oping countries. The types of flows that would be considered illicit are shown to be signi�cantly different in each of these cases. The analysis provides a rigorous way of identifying policy-relevant illicit flows in developing countries. Given the potential importance of these flows, it is 21 22 Draining Development vital to have a rigorous framework that at least ensures that we minimize the chances of causing inadvertent damage through well-intentioned policies. Indeed, the analysis shows that many loose de�nitions of illicit capital flows are problematic in this sense. Introduction The concept of illicit capital flows has come to prominence relatively recently, reflecting growing concerns about the rami�cations of an insuf- �ciently regulated and apparently increasingly predatory international �nancial system. In advanced economies, the 2008 global �nancial crisis brought into sharp relief the growing gap between the effectiveness of national regulatory tools and the global operations of private �nancial agencies. As illustrated, for example, by the standoff between the U.S. Securities and Exchange Commission and Goldman Sachs employees over the collateralized debt obligation deal, Abacus 2007–AC 1, the debate in advanced countries has highlighted important ambiguities about what constitutes legitimate �nancial market behavior. In the case of developing countries, the international concern about illicit capital flows is motivated primarily by concerns that vital develop- mental resources are being lost to these economies because of the ease with which capital flight can flourish in the context of a burgeoning, yet opaque international �nancial system (for example, see Eurodad 2008a; Global Witness 2009; Kar and Cartwright-Smith 2008; Baker 2005). Closely related to this is the idea that illicit capital flows from developing economies are indicative of deeper structural problems of political gov- ernance in these countries. Finally, there also are worries about how illicit capital flows from developing countries may affect advanced coun- tries through diverse mechanisms, such as directly or indirectly �nanc- ing crime or terror. In this chapter, we are concerned mainly with the impact of illicit capital flows on developing countries and the effective- ness of policy to control such outflows. The next section situates the literature on illicit capital flows within the wider context of economic analyses of capital flight from developing economies and provides our de�nitions. The following section elabo- rates our de�nition of illicit capital flows. We simplify the range of varia- Governance and Illicit Flows 23 tion across countries using a three-tier typology of different economic and political contexts: advanced economies; an intermediate group of normal developing countries in which a stable political settlement exists even though institutions still have a large element of informality; and a �nal tier of fragile and vulnerable developing countries in which the political settlement is collapsing, and economic processes, including illicit capital flight, are driven by the collapsing polity. A de�nition of illicit capital flight has to be consistently applicable across these substan- tially different institutional and political contexts. Keeping this in mind, we de�ne illicit capital flows as flows that imply economic damage for a society given its existing economic and political structure. The penulti- mate section develops some basic policy tools to operationalize this ana- lytical framework. The �nal section concludes. Illicit Capital Flows and Capital Flight: Concepts and De�nitions The term illicit has strong moral undertones, but a closer look at the actual use of the term almost always reveals an underlying concern with the developmental damage that particular capital flows can inflict. We believe that, from a policy perspective, this has to be made explicit, as well as the precise methodology that is being applied to determine the damage. We start by de�ning an illicit capital flow as a flow that has a negative impact on an economy if all direct and indirect effects in the context of the speci�c political economy of the society are taken into account. Direct effects refer to the immediate impact of a particular illicit capital flow on a country’s economic growth performance, for example, through reduced private domestic investment or adverse effects on tax revenue and public investment. Indirect effects are feedback effects on economic growth that arise from the role played by illicit capi- tal flows in the sustainability of the social and political structure and dynamics of a country. For clarity of exposition, we specify the political economy of a society in terms of the political settlement. A political set- tlement is a reproducible structure of formal and informal institutions with an associated distribution of bene�ts that reflects a sustainable dis- tribution of power. Sustainability requires that the formal and informal 24 Draining Development institutional arrangements that govern societal interaction in a country and the distribution of bene�ts to which they give rise achieve a suf�- cient degree of compatibility between economic productivity and politi- cal stability to allow the society to reproduce itself without an escalation of conflict (Khan 2010). There are two important elements in this de�nition. First, the judg- ment of impact has to consider both direct and indirect effects because we are interested in the overall net effect of particular �nancial flows on the developmental prospects of a country. Second, the speci�c political settle- ment of the country is important because the indirect effects, in particu- lar, depend on the interplay between the economic and political structure of the country, and this can vary greatly across contexts. To simplify, we focus on three broad variants of political settlements, but �ner distinc- tions can be made. The assessment of both direct and indirect effects is necessarily based on a counterfactual assessment of what would happen if the illicit capital flow in question could hypothetically be blocked. The assessment is counterfactual because, in many cases, the flows cannot actually be blocked or only partially so. The assessment is therefore subject to the analytical perspective of the observer, but we believe that this needs to be done in an explicit way to facilitate public debate. Our de�nition allows us to make sense of the perception that not all illegal flows are nec- essarily illicit, while some legal flows may be illicit. The task of policy is to identify both the particular capital flows that can be classi�ed as illicit, but also the subset of illicit flows that can be feasibly targeted. What constitutes damage in the sense of a negative developmental impact depends on how we de�ne development. When illicit capital flows are equated with illegal outflows (as in Kar and Cartwright-Smith 2008; Baker and Nordin 2007), the implicit suggestion is that adherence to prevailing legal rules is suf�cient for promoting the social good.1 Yet, the use of the notion of illicitness also suggests that damaging develop- mental outcomes may not always correspond to violations of the law and that, therefore, social, economic, and political damage needs to be more precisely de�ned. Moreover, if we are not to suffer the criticism of pater- nalism, our criteria have to be widely accepted as legitimate in that soci- ety. In practice, it is dif�cult to establish the criteria that measure devel- opment and therefore can be used to identify damage in any society, but particularly in developing ones. A social consensus may not exist if there Governance and Illicit Flows 25 are deep divisions about social goals. However, as a �rst step, we can insist that analysts and observers making judgments about illicit flows at least make their own criteria explicit. This will allow us to see if these criteria are so far away from what are likely to be the minimal shared assumptions in a society as to make the analysis problematic. In addi- tion, because the effect of capital flows (and economic policies in gen- eral) is often heavily disputed, we also require an explicit reference to analytical models that identify how particular capital flows affect par- ticular developmental goals. Making all this explicit is important because observers may disagree in their choices on these issues. Our de�nition suggests that capital flows that are strictly within the law may be illicit if they damage society and that, conversely, flows that evade or avoid the law may sometimes be benign, and blocking some of these flows may have adverse consequences. However, all illegal capital flows may be judged illicit from a broader perspective if the violation of laws is judged to be damaging for development regardless of the speci�c outcomes associated with lawbreaking. We argue that this position is easier to sustain in advanced countries where the formal structure of rules is more or less effectively enforced by rule-based states and where political processes ensure that legal frameworks are relatively closely integrated with the evolution of socially acceptable compromises to sus- tain political stability and economic growth. The ongoing global �nan- cial and economic crisis has demonstrated that this picture does not always hold true even in advanced countries. Meanwhile, developing countries are typically characterized by signi�cant informality in social organization; laws, if they exist, are, in general, weakly enforced and do not typically reflect worked-out social compromises and economic pro- grams, and signi�cant aspects of the economy and polity therefore oper- ate through informal arrangements and informality in the enforcement of formal rules (Khan 2005, 2010). Violations of formal rules are wide- spread in these contexts and do not necessarily provide even a �rst approximation of social damage. To keep the analysis broad enough to include different types of societ- ies, what then should we be looking for to judge damage to society? Is damage to be judged by the effects of particular flows on economic growth, or income distribution, or different measures of poverty, or some combination of the above? We suggest that we should use the least 26 Draining Development demanding way of judging damage because the more minimal our requirement, the more likely we are to �nd broad support across observ- ers who may disagree at the level of more detailed speci�cations. An illicit �nancial flow (IFF), according to our minimal de�nition, is one that has an overall negative effect on economic growth, taking into account both direct and indirect effects in the context of the speci�c politi- cal settlement of a country. A flow that has not directly affected economic growth, but has undermined the viability of a given political settlement without preparing the ground for an alternative sustainable political settlement may be seen to have, indirectly, a negative effect after we account for adjustments that are likely as a result of a decline in political stability. It is now increasingly recognized that the different ways in which the political and social order is constructed in developing coun- tries have important implications for how institutions and the economy function (North et al. 2007). Because the construction of political settle- ments (and, in North’s terminology, of social orders) differs signi�cantly across societies, the economic and political effects of particular �nancial flows are also likely to be different. Our de�nition has some overlaps with, but also important points of departure from, the way in which damaging capital flight has been ana- lyzed in the economics literature. The loss of developmental resources through capital flows from poor to rich economies has been an important topic of development economics since its inception. In particular, the lit- erature on capital flight had made the problem of abnormal capital out- flows from developing countries its main interest long before the policy focus on illicit capital outflows from such economies emerged. A brief review of this literature allows us to identify systematically different core drivers of capital flight highlighted by different strands of the literature. Our review also establishes the problems associated with ignoring the structural differences in the types of political settlements across countries, problems that our de�nition of illicit capital flight speci�cally addresses. From capital flight as abnormal capital outflows to single-driver models of capital flight In 1937, Charles Kindleberger famously de�ned capital flight as “abnor- mal� capital outflows “propelled from a country . . . by . . . any one or more of a complex list of fears and suspicions� (Kindleberger 1937, 158). Governance and Illicit Flows 27 A specialist of European �nancial history, Kindleberger had in mind well-known episodes of European capital flight going back to the 16th century, as well as the troubles of crisis-ridden Europe in the 1920s and 1930s. In most of these historical cases, this “complex list of fears and suspicions� can be attributed to speci�c events or to exceptionally dis- ruptive periods of political, social, and economic change or confronta- tion. To differentiate ex post between abnormal capital outflows driven by profound uncertainty (fear and suspicion) about the future on the one hand and, on the other hand, normal capital outflows driven by usual business considerations was a rather straightforward exercise in those contexts. For policy purposes, however, such ex post analyses are of limited rel- evance. What matters is suf�cient ex ante knowledge of the factors that differentiate abnormal from normal situations. Such ex ante knowl- edge—to facilitate effective preventive policy options—will have to include consideration of the determinants of precisely those economic, political, and social factors that may trigger the disruption that we regard as obvious from an ex post perspective. There may not be any general or abstract solution in the sense that what quali�es as capital flight requir- ing a policy response may differ across countries. A half-century on from his original contribution, Kindleberger himself struck a considerably more cautious note in this respect, as follows: It is dif�cult—perhaps impossible—to make a rigorous de�nition of capital flight for the purpose of devising policies to cope with it. Do we restrict cases to domestic capital sent abroad, or should foreign capital precipi- tously pulled out of a country be included? What about the capital that emigrants take with them, especially when the people involved are being persecuted . . . ? Does it make a difference whether the emigration is likely to be permanent or temporary, to the extent that anyone can tell ex ante? And what about the cases where there is no net export of capital, but capital is being returned to the country as foreign investment . . . ? Is there a valid distinction to be made between capital that is expatriated on a long-term basis for fear of con�scatory taxation, and domestic speculation against the national currency through buying foreign exchange that is ostensibly inter- ested in short-term pro�ts? (Kindleberger 1990, 326–27) As attention shifted to capital flight from contemporary developing economies, in particular in the wake of the Latin American debt crisis of 28 Draining Development the early 1980s, a growing consensus emerged about the dif�culty of isolating speci�c determinants of abnormal capital outflows ex ante (Dooley 1986; Khan and Ul Haque 1985; Vos 1992). While there was some agreement on basic characteristics of abnormal outflows (that they tend to be permanent, not primarily aimed at asset diversi�cation, and not generating recorded foreign exchange income), the original effort to identify the characteristics of abnormal capital outflows was replaced by de�nitions of capital flight in terms of a single-core driver, perceived dif- ferently by different approaches (Dooley 1986, 1988; Cuddington 1987). Capital flight as portfolio choice First, the portfolio approach adopts standard models of expected utility maximization by rational economic agents to explain capital flight as a portfolio diversi�cation response to higher foreign returns relative to domestic returns on assets (Khan and Ul Haque 1985; Lessard and Wil- liamson 1987; Dooley 1988; Collier, Hoeffler, and Pattillo 2001). More speci�cally, this involves a counterfactual comparison of after-tax domes- tic and foreign returns, adjusted for a range of variables, such as expected depreciation, volatility of returns, liquidity premiums, and various indi- cators of investment risk, including indexes of corruption. In this view, capital flight is caused by the existence of market distortions and asym- metric risks in developing countries (relative to advanced economies). The underlying market-theoretical model of economic development builds on four core premises: (1) economic behavior relevant to capital flight is correctly described by expected utility maximization, (2) markets exist universally (and thus can be distorted), (3) individual agents possess the ability to compute the probabilities of investment risk globally and on the basis of counterfactual investment models, and (4) computable prob- abilities can be attached to all events. These are obviously fairly restrictive assumptions in any case. In the present context, however, the most important conceptual drawback is illustrated by the following remark of a Brazilian economist: Why is it that when an American puts money abroad it is called “foreign investment� and when an Argentinean does the same, it is called “capital flight�? Why is it that when an American company puts 30 percent of its equity abroad, it is called “strategic diversi�cation� and when a Bolivian Governance and Illicit Flows 29 businessman puts only 4 percent abroad, it is called “lack of con�dence�? (Cumby and Levich 1987, quoted in Franko 2003, 89) The obvious answer is that, relative to overall domestic resources, more domestic capital tends to flee for longer from Argentina and Bolivia than from the United States and that this is so because structural uncer- tainty about future investment opportunities is higher in the former economies. To prevent relatively scarce capital from voting with its feet in situations of great structural uncertainty, developing countries are more likely to impose regulatory barriers on free capital movement, thus turning what might simply appear to be good business sense into illegal capital flight. By focusing on a single broad motive for capital flight, namely, utility maximization in the presence of differential policy regimes and invest- ment risks, the portfolio approach conflates short-term utility (and pro�t) maximization with structural political and economic uncertain- ties. No systematic distinction is made between the drivers of asymmet- ric investment risks in developing economies, which may range from inflation and exchange rate depreciation to expectations of con�scatory taxation and outright politically motivated expropriation. Similarly, policy-induced market distortions can range from short-term �scal and monetary policies, common to all economies, to policies promoting long-term structural and institutional changes with much more wide- ranging (and often more uncertain) implications for future investment opportunities. This failure to distinguish between different drivers of capital flight considerably weakens the effectiveness of the policy impli- cations arising from this approach. This consists essentially in the rec- ommendation of market-friendly reforms to eliminate such market dis- tortions on the assumption that these will also minimize asymmetric investment risks. If, however, markets do not as yet exist or suffer from fundamental structural weaknesses, market-friendly reforms may be insuf�cient to minimize differential investment risk relative to, for example, advanced economies. This problem is, in fact, at least partially recognized by advocates of this approach. Some authors use a narrow statistical measure of capital flight, the hot money measure, which limits capital flight to short-term speculative capital outflows of the private nonbank sector (taken to be 30 Draining Development the primary source of net errors and omissions in a country’s balance of payments) (for example, see Cuddington 1986, 1987). In contrast, the most widely used statistical de�nition of capital flight, often referred to as the residual measure, includes recorded and nonrecorded acquisitions of medium- and short-term assets and uses broader estimates of capital inflows (World Bank 1985; Erbe 1985; Myrvin and Hughes Hallett 1992).2 Cuddington thus reemphasizes the idea that abnormal capital outflows of money running away are a reaction to exceptional circum- stances and events rather than ordinary business, although, in times of extensive deregulation of international �nancial markets, this conjecture may be less convincing. Other authors have taken a different route to distinguish portfolio capital flight from other forms of capital flight, in particular that induced by extreme political instability, such as civil war (Tornell and Velasco 1992; Collier 1999). In the latter case, capital flight can occur despite lower actual foreign returns relative to potential domestic returns (if peace could be achieved) and is part of a wider out- flow of productive resources, including labor. The social controls approach to capital flight A second strand of the capital flight literature de�nes capital flight as “the movement of private capital from one jurisdiction to another in order to reduce the actual or potential level of social control over capital� (Boyce and Zarsky 1988, 192). To the extent that such capital flight is motivated primarily by the pursuit of private economic gain, this social control de�nition is not radically different from the portfolio approach. In both cases, capital flight occurs in response to policy intervention. However, the “social controls� approach rests upon an explicit premise absent in much conventional economic theory, namely that individual control over capital is rarely absolute or uncon- tested, but rather subject to social constraints, the character and extent of which vary through time. Unlike many authors, Boyce and Zarsky, there- fore do not consider capital flight to be necessarily “abnormal.� (Rishi and Boyce 1990, 1645) As with the portfolio approach, there is no systematic distinction between fundamentally different drivers of capital flight. All capital flight occurs in response to government controls, whether these concern Governance and Illicit Flows 31 short-term macroeconomic stabilization through �scal and monetary policy measures or more long-term structural interventions. However, in this case, the underlying model of the economy is not that of universal and competitive markets inhabited by maximizers of expected utility and pro�ts, but rather that of a mixed economy in which markets are one set of several institutions ultimately governed by a (welfare or devel- opmental) state or, more broadly, by social interventions from outside the market sphere. Different from the portfolio approach, the main pol- icy implication is therefore not simply a focus to promote markets, but to minimize capital flight through the strengthening of existing social controls or the introduction of alternative, more effective administrative measures to control private capital movements. The social controls approach remains ambiguous, however, about the origin and legitimacy of social (capital) controls. A weak version of the approach identi�es governments as the core players in social control, and no explicit judgment is made about the legitimacy or effectiveness of speci�c government controls. This position comes closest to the port- folio approach. In this vein, Walter, for example, argues that capital flight appears to consist of a subset of international asset redeployments or portfolio adjustments—undertaken in response to signi�cant perceived deterioration in risk/return pro�les associated with assets located in a particular country—that occur in the presence of conflict between objec- tives of asset holders and the government. It may or may not violate the law. It is always considered by authorities to violate an implied social con- tract. (Walter 1987, 105) By contrast, the strong version adopts the more heroic assumption that social (capital) controls reflect some kind of a social consensus about the ways in which economic development is best achieved. The implied social contract here is not one merely perceived by government authorities to exist, but one based on genuine social approval. A recent example is the observation of Epstein, as follows: When people hear the term “capital flight� they think of money running away from one country to a money “haven� abroad, in the process doing harm to the home economy and society. People probably have the idea that money runs away for any of a number of reasons: to avoid taxation; to avoid con�scation; in search of better treatment, or of higher returns 32 Draining Development somewhere else. In any event, people have a sense that capital flight is in some way illicit, in some way bad for the home country, unless, of course, capital is fleeing unfair discrimination, as in the case of Nazi persecution. (Epstein 2006, 3; italics added) For Epstein (2006, 3–4), capital flight is therefore an “inherently polit- ical phenomenon� and also mainly “the prerogative of those—usually the wealthy—with access to foreign exchange.� This view entails a more explicitly normative position than either the portfolio approach or the weak version of the social controls approach in that capital flight is char- acterized not only as economically damaging for development, but also as illegitimate from the perspective of an existing consensus about the social (developmental) good. Capital flight as dirty money Both the portfolio approach and the social controls approach broadly interpret capital flight as a response by private capital to expectations of lower domestic returns relative to foreign returns on assets. Both are out- come oriented in that their primary concern is with the macroeconomic analysis of the perceived damage inflicted by capital flight on developing economies rather than with the origins of flight capital or the methods by which it is being transferred abroad. They differ with regard to their ana- lytical benchmark models of a normal or ideal economy. Whereas the portfolio approach subscribes to variants of the standard model of a com- petitive free-market economy, the social controls approach adopts a mixed economy model in which the social control of private capital, for example by a welfare state, is normal. This translates into different single-driver models of capital flight. In the �rst case, the driver is simply pro�t (or util- ity) maximization (or the minimization of investment risk) at a global level, and the damage done arises not from capital flight per se, but from market distortions that lower relative returns on domestic private assets. In the second case, the driver is the avoidance of policy controls, and the damage arises from private capital breaking implicit or explicit social con- tracts. These models overlap only in so far as (1) market distortions arise from domestic policy interventions rather than other exogenous shocks and (2) social controls are judged, in any particular context, to be poorly designed so as to undermine rather than promote economic development and thus violate the implicit or explicit social contract. Governance and Illicit Flows 33 The more recent literature on illicit capital flows adopts an explicitly normative perspective on capital flight that focuses primarily on adher- ence to the law or perceived good practice. Illicit capital flows are flows that break the implicit rules asserted as desirable by the observer. A cross- border movement of capital that, at any stage from its generation to its use, involves the deployment of illegal or abusive activities or practices is illicit. Probably the most well known example of this procedural or rule-based perspective on capital flight is Baker’s notion of dirty money that distin- guishes criminal, corrupt, and commercial forms of illicit capital flows (a term used interchangeably with dirty money) (Baker 2005; Baker and Nordin 2007). Criminal flows encompass “a boundless range of villainous activities including racketeering, traf�cking in counterfeit and contraband goods, alien smuggling, slave trading, embezzlement, forgery, securities fraud, credit fraud, burglary, sexual exploitation, prostitution, and more� (Baker 2005, 23). Corrupt flows stem “from bribery and theft by (foreign) government of�cials� trying to hide the proceeds from such activities abroad, and the commercial component of dirty money stems from tax evasion and mispriced or falsi�ed asset swaps, including trade misinvoic- ing and abusive transfer pricing (Baker 2005, 23). Of these components of dirty money, illicit commercial flows have been estimated to account for around two-thirds of all illicit outflows, and proceeds from corruption for the smallest part, around 5 percent of the total (Eurodad 2008b). This rule-based dirty money approach leads to a conceptual de�ni- tion of capital flight that is narrower than the de�nitions adopted by more conventional approaches. First, de�nitions of capital flight here are mostly limited to unrecorded capital flows. Thus, Kapoor (2007, 6–7), for example, de�nes capital flight as the “unrecorded and (mostly) untaxed illicit leakage of capital and resources out of a country,� a de�ni- tion taken up by Heggstad and Fjeldstad (2010, 7), who argue that its characteristics include that the resources are domestic wealth that is permanently put out of reach for domestic authorities. Much of the value is unrecorded, and attempts to hide the origin, destination and true ownership of the capital are parts of the concept. This does not necessarily translate into the adoption of narrow statistical measures of capital flight, such as hot money estimates that take account 34 Draining Development only of unrecorded capital flows in the balance of payments. Rather, measures of illicit capital flows attempt to capture as many likely con- duits for such flows as possible, in practice integrating different conven- tional measures of capital flight with estimates of trade misinvoicing (Kar and Cartwright-Smith 2008). Second, consistent with the conceptual focus on dirty money, the core driver of capital flight is also more narrowly de�ned, as follows: The term flight capital is most commonly applied in reference to money that shifts out of developing countries, usually into western economies. Motivations for such shifts are usually regarded as portfolio diversi�ca- tion or fears of political or economic instability or fears of taxation or inflation or con�scation. All of these are valid explanations for the phe- nomenon, yet the most common motivation appears to be, instead, a desire for the hidden accumulation of wealth. (Kar and Cartwright-Smith 2008, 2; emphasis added) However, capital flight in response to return differentials, asymmet- ric investment risks, or perceived macroeconomic mismanagement through social controls can also occur in the open, in particular in a highly deregulated �nancial environment. Thus, the main concern of conventional approaches to capital flight is with wealth accumulation abroad in general, rather than exclusively or even primarily with hidden wealth accumulation. Finally, illicit capital flows are also often equated with illegal outflows, on the grounds that funds originating in (or intended for) illicit activi- ties have to be hidden and will eventually disappear from any records in the transferring country (Kar and Cartwright-Smith 2008; Baker and Nordin 2007). This narrows conventional de�nitions of capital flight because the latter’s focus on capital outflows that are, in some sense, damaging to the economy is not limited to illegal outflows. How exten- sive the overlap between illegal and damaging capital outflows is will depend on how closely and effectively legal frameworks encapsulate the social or economic good, de�ned in terms of the respective underlying benchmark models of an ideal or licit state of affairs. In fact, the equation of illegal and illicit capital outflows is not sys- tematically sustained, even in the dirty money approach. At least some of the activities included in the analysis of dirty money are not illegal, but Governance and Illicit Flows 35 refer to practices characterized as abusive, such as aspects of transfer pricing, speci�c uses made of tax havens, and corrupt activities that have not necessarily been outlawed. Rather, the procedural or rule-based focus on adherence to the law in this approach to capital flight is based on an implicit claim that adherence to the law will promote economic development or, more generally, the social good. If the principal motiva- tion for capital flight is, in fact, “the external, often hidden, accumulation of wealth, and this far outweighs concerns about taxes� (Baker and Nor- din 2007, 2), the policy response has to be directed primarily at eliminat- ing the dirty money structure through the enforcement of national and global standards of �nancial transparency and democratic accountabil- ity. The core obstacle to economic development then becomes the lack of good governance, corruption, and the absence of or weaknesses in the rule of law (rather than speci�c legislation). Thus, the dirty money approach to illicit capital flows from developing countries differs from more conventional de�nitions of capital flight mainly in that it appears to be rooted less in outcome-oriented models of economic growth and development, but in a speci�c rule-based liberal model of good governance and a good polity, perceived to be a necessary condition for achieving any more substantially de�ned social good. From the above, we see that the literature identi�es three core drivers of capital flight from developing economies, each implying a different underlying view of what constitutes damage to these economies and each adopting a monocausal perspective. Figure 2.1 summarizes features of these approaches, as follows: 1. In the portfolio approach, social damage is a result of interference with competitive markets that are presumed to otherwise exist. Capi- tal flight is driven by economic incentives to escape such interference given pro�t-maximizing investment strategies. The best way to elim- inate capital flight in this perspective is to remove the damaging gov- ernment interventions in competitive markets. This largely ignores the fact that competitive markets require government regulation even in advanced economies and that extensive market failures imply that signi�cant government intervention may be required to achieve developmental outcomes in developing economies. 36 Draining Development 2. The social controls approach is almost the mirror image of the port- folio approach. All social controls over private capital movements are presumed to reflect a legitimate social contract that is welfare enhanc- ing. As a result, capital outflows that violate social controls imposed by a welfare-developmental state are damaging and illicit, presum- ably because the indirect effects of violating the social contract will be socially damaging. The best way to eliminate capital flight is to rein- force social controls. This largely ignores the fact that not all formal social controls and regulations in developing countries are legitimate, growth enhancing, or politically viable. 3. In the dirty money approach, social damage is the result of violating the rule of law. Capital flight is driven by the desire to accumulate wealth by hiding from the rule of law. No substantial view on what constitutes social, economic, or political damage is typically offered. This largely ignores the fact that, in developing countries in particu- lar, existing legal frameworks may not adequately encapsulate the economic and political conditions necessary to achieve whichever substantial idea of the social good is adopted (economic development, social justice, the preservation of speci�c human rights, and so on). Figure 2.1. De�nitions of Capital Flight Portfolio approach Social good: Competitive markets Core driver: “Illicit� policy interference with competitive markets Policy response: Market-friendly reforms Social controls approach outcome broad Social good: Benevolent developmental state/mixed economy oriented Core driver: Evasion of legitimate social controls of private capital movement Policy response: Reinforce controls Illicit capital flows Social good: The rule of law/good governance Core driver: Hidden accumulation of wealth Policy response: Global bans and regulation Illegal capital flows narrow rule oriented Source: Author compilation. Governance and Illicit Flows 37 Illicit Capital Outflows: An Alternative Policy Framework The discussion above suggests that the search for a conceptual de�nition of capital flight in general and of illicit capital outflows from developing countries in particular has been motivated by differing perceptions of what constitutes the social or developmental good. Implicit differences in the values and theoretical models of observers explain the signi�cant differences in how illicit capital flows have been de�ned. These hidden differences are not conducive for developing effective policy responses on the basis of which some minimum agreement can be reached. We believe we can go to the core of the problem by de�ning illicit capital flows as outflows that cause damage to the economic development of the country, taking into account all direct and indirect effects that are likely, given the speci�c political settlement. An implicit assumption in many conventional approaches is that all flight capital, however de�ned, will yield a higher rate of social return in developing economies if it can be retained domestically (Schneider 2003a; Cumby and Levich 1987; Walter 1987). From a policy perspective, locking in potential flight capital is supposed to reduce the loss of devel- opmental resources directly and indirectly stabilize domestic �nancial markets and improve the domestic tax base (Cuddington 1986). In con- trast, we argue that the problems faced by countries with capital flight can be different in nature. The widely shared premise of a general nega- tive relationship between capital outflows and domestic capital accumu- lation simply is not valid (Gordon and Levine 1989). Dynamic links among capital flows, economic growth, technological change, and politi- cal constraints mean that, even in advanced economies, the regulation of capital outflows is an uphill and evolving task. In developing countries, too, some capital outflows may be desirable to sustain development. Moreover, the idea underlying some of the illicit capital flow literature that adherence to the law is suf�cient for identifying the social or eco- nomic good clearly does not always apply. On the basis of our minimalist de�nition of illicit capital flows, we proceed to develop a simple three-tier typology of economic and politi- cal constellations and governance structures and the ways in which these pose policy challenges in controlling illicit capital outflows. Spe- ci�cally, we distinguish between advanced economies (in opposition to 38 Draining Development developing countries in general), and, within the latter group, we distin- guish between normal or intermediate developing countries and fragile developing countries. Advanced economies: The differences with respect to developing economies Advanced economies are likely to be supported by rule-based states and political institutions such that the social compromises for political sta- bility and the economic policies required for sustaining growth—the political settlement—are reflected and codi�ed in an evolving set of laws. In these contexts, it is not unreasonable to expect that as a �rst approximation, damaging capital flows are likely to be capital flows that violate existing laws. Illegal capital flows are likely to be damaging and can justi�ably be described as illicit in this context. However, some dam- aging flows may not be illegal if laws do not fully reflect these conditions or have not caught up with the changing economic and political condi- tions involved in sustaining growth. For instance, the proliferation of inadequately regulated �nancial instruments that resulted in the �nan- cial crisis of 2008 was driven by �nancial flows that were, in many cases, not illegal, but turned out, ex post, to be seriously damaging. A focus on identifying damaging �nancial flows independently of the existing legal framework may therefore be helpful even in advanced countries, though, most of the time, a focus on illegal �nancial flows may be adequate. Advanced economies, by de�nition, have extensive productive sec- tors, which is why they have high levels of average income. As a result, formal taxation can play an important role not only in providing public goods, but also in sustaining signi�cant levels of formal, tax-�nanced redistribution. Both help to achieve political stability and sustain exist- ing political settlements. Taxation may be strongly contested by the rich, but an implicit social contract is likely to exist whereby social interde- pendence is recognized and feasible levels of redistributive taxation are negotiated. Signi�cant �scal resources also provide the resources to pro- tect property rights and enforce a rule of law effectively. Tax evasion and capital flight in this context are likely to represent individual greed and free-riding behavior rather than escape routes from unsustainable levels of taxation. Moreover, if economic policies and redistributive taxation represent the outcome of political negotiations Governance and Illicit Flows 39 between different groups, illegal capital flows would also be illicit in the sense of damaging the sustainability of the political settlement and thereby, possibly, having damaging indirect effects on economic growth. Transparency, accountability, and the enforcement of the rule of law can therefore be regarded as mechanisms for limiting tax-avoiding capital flight and other forms of illicit capital flows in advanced countries. Nonetheless, even within advanced countries, if legal arrangements cease to reflect economic and redistributive arrangements acceptable to major social constituencies, the correspondence between illegal and illicit can break down. The more recent discussion of illicit capital flows in advanced countries has to be considered in a context in which the increase in the bargaining power of the rich to de�ne economic laws in their own interest has exceeded the pace at which other social groups have accepted these changes. This allows us to make sense of the more recent preoccupation with IFFs in many countries of the Organisation for Economic Co-operation and Development (OECD) where the con- cern with dirty money has been closely associated with changes in the structure of many of these economies as a result of the decline in manu- facturing and a growth in the service sectors, in particular, �nance. The result has been not only a gradual change in the formal structure of taxa- tion and redistribution in ways that reflect the regressive changes in power, but also a growing tendency of the super-rich to evade even exist- ing legal redistributive arrangements through �nancial innovations or outright illegal capital flows. The avoidance and evasion of taxation have not been uncontested in advanced countries. Many social groups have criticized these developments and questioned the legitimacy of capital flows that seek to avoid and evade taxes. This constitutes an important part of the concern with illicit capital flows in advanced countries. Palma (2005) points to the signi�cant decline in manufacturing prof- its as a share of total pro�ts in advanced countries over the last three decades and the concomitant search for global �nancial, technological, and resource rents (see also Smithin 1996). The decline in manufac- turing can at least partly explain the signi�cant change in income dis- tribution in OECD countries, generally in favor of the highest income groups. For example, in the United States, the average real income of the bottom 120 million families remained roughly stagnant between 1973 and 2006, while that of the top 0.01 percent of income earners increased 40 Draining Development 8.5 times, meaning that the multiple between the two income groups rose from 115 to 970. During the four years of economic expansion under the George W. Bush Administration, 73 percent of total (pretax) income growth accrued to the top 1 percent of income earners com- pared with 45 percent during the seven years of economic expansion under the Clinton Administration (Palma 2009). The structural change in advanced countries is obviously complex and differs across countries, but there has been a signi�cant structural change over the last few decades, and the redistributive arrangements that emerged out of the postwar political consensus have faced renegotiation as a result. The contemporary concern with illicit capital flows in advanced coun- tries is motivated by a growing dissatisfaction both with the enforcement of law and with the extent to which the law in critical areas still effectively supports the social good. Narrower debates about illegal capital flows have focused on capital movements that have flouted existing legal redis- tributive arrangements because of the growing power of the rich to evade taxes through complex �nancial instruments and the threat of relocation to other jurisdictions. But a more profound concern with illicit capital flows is based on a feeling of unease that the increase in the political power of these new sectors has enabled them to change laws without the acquiescence of groups in society that would once have had a say. This includes changes in laws that have weakened regulatory control over capi- tal seeking to move much more freely in search of risky pro�ts or tax havens. Even legal capital flows could be deemed to be illicit from this perspective if economic welfare or political stability were threatened by a unilateral rede�nition of the implicit social contract. The steady deregulation of �nancial flows in advanced countries in the 1980s and 1990s is a case in point. Changes in regulatory structures increased returns to the �nancial sector, arguably at the expense of greater systemic risk for the rest of society, and diminished the ability of states to tax these sectors (Eatwell and Taylor 2000). Some of the associated �nan- cial flows could therefore easily be judged to have been illicit according to our de�nition. Underlying this judgment is the implicit claim that the restructuring of law is based on an emergent distribution of power that is not legitimate because it eventually produces economic and political costs in which many social constituencies no longer acquiesce and that thereby threatens to become unsustainable (Crotty 2009; Wray 2009; Pollin 2003). Governance and Illicit Flows 41 The concern here is that some of these �nancial flows were illicit either because they were directly damaging to the economy or because they dam- aged the political settlement in unsustainable ways and may eventually bring about economic costs in the form of social protests and declining political stability. These are obviously matters of judgment, but our de�ni- tion provides a consistent way of structuring the policy debate without getting locked into a monocausal de�nition. As long as the new sectors grew rapidly, the critique that some capital flows were illicit (even if legal) remained a fringe argument. However, the �nancial crisis of 2008, largely the result of excessive risk taking by an insuf�ciently regulated �nancial sector, showed, ex post, that some of these �nancial flows were not only questionable in terms of their legiti- macy and their impact on social agreements, they were damaging in a straightforward sense of economic viability. If the political process in advanced countries responds to these pressures by enacting laws that com- bine socially acceptable redistributive arrangements with regulatory struc- tures that make sense for sustained economic growth, the law could once again reflect a broadbased social compromise grounded on maintaining politically sustainable economic growth. Under these circumstances, illicit capital flows will once again become coterminous with illegal capital flows. Even so, the basic features of advanced countries that we discuss above have often created an expectation that a rule of law can and should be enforced at a global level, making illicit capital flows relatively easy to identify and target. Unfortunately, while this analysis makes sense at the level of individual advanced countries, it falls apart as an analytical framework at a global level. It ignores the obvious fact that global laws guiding economic policy or capital flows would only be legitimate if there were a global social consensus based on the same principles of redistributive taxation and social provision on which cohesive societies are constructed at the level of individual countries. It also presumes that a global agency would have the �scal resources to enforce laws at the global level in the same way that individual advanced countries can enforce national laws. None of these presumptions are reasonable given the current global gaps among countries in terms of economics and pol- itics. By our de�nition, the absence of a sustainable global political set- tlement based on global redistributive and enforcement capabilities makes problematic any attempt to de�ne IFFs at the global level. 42 Draining Development To understand this more clearly, the differences between the political settlements in advanced and developing countries need to be spelled out. First, developing countries are structurally different because their inter- nal political stability and economic development are not (and cannot normally be) organized through formal rules and laws to the extent observed in advanced countries. In developing countries, the legal frame- work is not an adequate guide for identifying illicit capital flows even as a �rst approximation. There are two essential limitations on a purely legal analysis of what is illicit in the typical developing country. First, in the realm of politics, the internal political stability of developing countries is not solely or even primarily based on social agreements consolidated through legal �scal redistributions. For a variety of reasons, including limited �scal space and more intense conflicts given the context of social transformations, redistributive �scal arrangements are typically less transparent and less formal compared with those in advanced countries. The political problem involves delivering resources to powerful constit- uencies that would otherwise be the source of political instability in a context of �scal scarcity. If politically powerful constituencies have to be accommodated legiti- mately and transparently, acceptable redistributions to more deserving groups such as the severely poor have to be agreed upon simultaneously to achieve political legitimacy. The �scal sums typically do not add up in developing countries for a redistributive package that would pass the test of public legitimacy, as well as provide the redistribution required by powerful groups. As a result, it is not surprising that the critical redistri- butions to powerful constituencies typically occur through patron-client politics and other mechanisms characterized by limited transparency. If successful, these arrangements achieve political stabilization by incorpo- rating suf�cient numbers of politically powerful factions within the rul- ing coalition. Even in developing countries where signi�cant �scal redis- tribution takes place, critical parts of the overall system of political redistribution are not based on transparently negotiated arrangements codi�ed in �scal laws and economic policies (Khan 2005). Second, the economies of developing countries are also signi�cantly different from the economies of advanced countries. Their formal or regulated modern sectors are normally a small part of the economy, and a much larger informal sector is unregulated or only partially regulated. Governance and Illicit Flows 43 Much of the economy, including the formal sector, suffers from low pro- ductivity and does not generate a big enough taxable surplus to pay for across-the-board protection of property rights and adequate economic regulation. The achievement of competitiveness also often requires peri- ods of government assistance and strong links between business and politics. This is because, even though wages may be low, the productivity of the modern sector is often even lower. Developing countries have poor infrastructure and poor skills in labor and management, and, in particular, they lack much of the tacit knowledge required to use mod- ern technologies ef�ciently, even the most labor-intensive ones. The strategies through which developing countries progress up the technology ladder while maintaining their internal political arrange- ments can differ substantially across countries (Khan and Blankenburg 2009). These strategies usually involve creating opportunities and condi- tions for pro�table investment in at least a few sectors at a time. Because the creation of pro�table conditions across the board is beyond the �scal capacities of developing countries, these strategies inevitably create priv- ileges for the modern sector and, often, for particular subsectors through government interventions in prices, exchange rates, interest rates, regu- lations, taxes, subsidies, and other policy instruments. The strategies for assisting learning in countries with industrial policies, such as the East Asian tigers, are well known. But the growth of modern sectors in all developing countries has required accidents or smaller-scale policy interventions that overcame the built-in disadvantages of operating pro�tably given the adverse initial conditions (Khan 2009). Some of this assistance may be formal and legal, but other aspects of assistance may be informal and even illegal. For instance, some �rms may informally have privileged access to land, licenses, and other public resources. Some of these privileges may be important in offsetting initial low produc- tivity or the higher costs created by poor infrastructure and the poor enforcement of property rights and the rule of law. Deliberately or oth- erwise, these arrangements can assist some �rms in starting production in adverse conditions and engage in learning by doing, but they can also simply provide privileges to unproductive groups (Khan and Jomo 2000; Khan 2006). Thus, in many cases, business-government links in developing coun- tries are predatory from the perspective of the broader society. Resources 44 Draining Development captured by privileged �rms in the modern sector are wasted, and, in these cases, the modern sector remains inef�cient at signi�cant social cost. However, in other cases, periods of hand-holding and bailouts do lead to the emergence of global competitiveness through formal and informal links between emerging enterprises and the state. The ef�cacy of developmental strategies depends on the nature of the relationships between business and politics, the compulsions on both sides to generate productivity growth over time, the time horizons, and so on, but not in any simple way on the degree to which formal laws are upheld (Khan and Blankenburg 2009; Khan 2009). It is not surprising that all developing countries fail the test of adher- ence to a rule of law and political accountability. Yet, some developing countries perform much better than others in terms of politically sus- tainable growth that eventually results in poverty reduction, economic development, and movement toward the economic and political condi- tions of advanced countries. These observations suggest that what con- stitutes a damaging �nancial flow may be more dif�cult to identify in developing countries. Because of the signi�cant differences in the eco- nomic and political conditions across developing countries, we �nd it useful to distinguish between normal developing countries, which we call intermediate developers, and fragile developers that suffer from more serious crises in their political settlements. Normal developing countries: The intermediate developers Our term intermediate developers refers to the typical or normal devel- oping country in which internal political and economic arrangements sustain political stability. Their internal political settlements can be quite varied, but also differ from those in advanced countries in that their reproduction typically requires signi�cant informal arrangements in both redistributive arrangements and the organization of production. Nonetheless, in most developing countries, there is a political settlement that has characteristics of reproducibility, and these societies can sustain economic and political viability. This does not mean that the govern- ments are universally recognized as legitimate, nor are violence and con- flict entirely absent, but the political arrangements are able to achieve development (at different rates) without descending into unsustainable levels of violence.3 Countries in this category include, for example, China Governance and Illicit Flows 45 and most countries of South Asia, Southeast Asia, and Latin America. In contrast, fragile developers are countries, such as the Democratic Repub- lic of Congo or Somalia, in which a minimally sustainable political set- tlement among the contending forces in society does not exist and in which the fundamental problem is to construct this in the �rst place. Our analysis of illicit capital flows from intermediate developers can be simpli�ed by distinguishing between political and economic actors according to their motivations for making decisions about �nancial flows. The former are likely to be concerned about threats to the pro- cesses through which they accumulate resources and about the political threats to these resources; the latter are likely to be primarily concerned with pro�t opportunities and expropriation risk. In reality, political and economic actors may sometimes be the same persons; in this case, we have to look at the motivations jointly. The simpli�cation may nonethe- less help one to think through the different analytical issues involved so that appropriate policies can be identi�ed in particular cases. Financial outflows driven by political actors. An obvious reason why political actors in developing countries may engage in capital flight is that their opponents may expropriate their assets if the opponents come to power. A signi�cant amount of the resource accumulation is likely to have violated some aspect of the structure of formal laws. The legality can be questioned by the next ruling coalition for a number of reasons, including expeditious political reasons, for instance, to undermine the ability of previous ruling factions to return to power. Let us assume that this capital flight is immediately damaging because it represents a loss of resources. The issue from the perspective of an analysis of (illicit) capital flight involves assessing the consequences of hypothetically blocking speci�c �nancial outflows in these circumstances. The important point is that patron-client politics cannot be immediately replaced by �scal politics because these countries are developing economies with a limited tax base. Therefore, the fundamental mechanisms through which politi- cal entrepreneurs gain access to economic resources are unlikely to dis- appear in the short run in most developing countries, with the exception of those that are close to constructing Weberian states. Given the nature of political settlements in developing countries, attempting to block �nancial outflows driven by politicians is unlikely to 46 Draining Development lead to a liberal rule of law because any new ruling coalition will also require off-budget resources to maintain political stability and will keep resources available for elections (and for their own accumulation). If these resources can be expropriated by a new coalition after an election, this can signi�cantly increase the stakes during elections. Expropriating the �nancial resources of former politicians may therefore have the par- adoxical effect of increasing instability (the case of Thailand after 2006, for example). It can increase the intensity with which assets of the cur- rent ruling coalition are attacked by the opposition, and it can increase the intensity of opposition by excluded coalitions as they attempt to pro- tect their assets from expropriation. Indeed, the evidence from stable intermediate developers such as Brazil or India suggests that the stability of the political settlement in the presence of competition between patron-client parties requires a degree of maturity whereby new coali- tions understand that it is not in their interest to expropriate the previ- ous coalition fully. An informal live-and-let-live rule of law guiding the behavior of political coalitions can reduce the costs of losing and allow elections to mature beyond winner-takes-all contests. However, this type of informal understanding is vulnerable. If the competition between political factions has not achieved a level of maturity that informally sets limits on what can be clawed back from a previous ruling coalition, a premature attempt at restricting �nancial flows may have the unintended effect of signi�cantly raising the stakes in political conflicts. If the ruling coalition cannot protect some of its assets in other jurisdictions, it may feel obliged to use violence or intimidation to stay in power, and this can increase the likelihood of eventual expro- priation. Paradoxically, some amount of flexibility in politically driven �nancial flows at early stages of state building may help lower the stakes at moments of regime change. Thus, restrictions on political �nancial flows are only likely to improve social outcomes if live-and-let-live com- promises between political coalitions have already been established. For instance, in more stable political settlements such as in Argentina, Brazil, or India, a gradual increase on restrictions on �nancial outflows could lead to better social outcomes as long as current politicians feel that the risk of domestic expropriation is low and the restrictions simply restrict excessive illegal expropriation by political players. If formal rules restrict political accumulation that is beyond what is normally required to sus- Governance and Illicit Flows 47 tain political operations in this political settlement, then �nancial flows that violate these rules can justi�ably be considered illicit. In contrast, if the informal understanding between competing parties is still vulnera- ble and live-and-let-live arrangements have not become entrenched (as in Bangladesh, Bolivia, Thailand, and, to an extent, República Bolivari- ana de Venezuela), attempts to limit �nancial outflows driven by politi- cal players are likely to be evaded or, if forcefully enforced, can occasion- ally have damaging consequences in raising the stakes during elections. According to our de�nition, we should not consider all capital outflows by political actors to be illicit in these contexts, with critically important policy implications. Clearly, these judgments reflect an attempt to take into account direct and indirect effects and are open to a degree of dis- agreement. However, we believe that these judgments have to be made and that they are best made explicitly. Financial outflows driven by economic actors. Financial outflows driven by economic actors can be motivated by concerns about expropriation or low pro�tability (in addition to tax evasion and tax avoidance). These factors help explain the frequent paradox that economic actors in devel- oping countries often shift assets to advanced-country jurisdictions where tax rates are higher and the returns achieved, say, on bank depos- its, are nominally lower (Tornell and Velasco 1992). Some of these flows are damaging, and blocking them (if that were possible) is likely to leave society more well off. Others are not damaging, or, even if they are dam- aging, attempts to block them would not be positive for society after all the direct and indirect effects are weighed. Financial outflows with net negative effects. The simplest cases involve capital flight whereby both the direct and indirect effects are negative, such as those driven by tax evasion or tax avoidance. The direct effect of these outflows is likely to be negative if tax revenue and, with it, public invest- ment is reduced, and, in addition, the indirect effect is also likely to be negative if the taxes are socially legitimate and their loss undermines polit- ical stability. The capital flight in these cases is clearly illicit. Another clear- cut case is that of theft of public resources with the collusion of political actors. A particularly serious example is the two-way capital flow involved in the odious debt buildup resulting if external borrowing by governments 48 Draining Development is turned, more or less directly, into private asset accumulation abroad by domestic residents. Examples include the Philippines and several Sub- Saharan African economies (Boyce and Ndikumana 2001; Cerra, Rishi, and Saxena 2008; Hermes and Lensink 1992; Vos 1992). Odious debt has no redeeming features, and, if it can be blocked, the developing country is likely to be more well off in terms of direct investment effects. In normal cases where a political settlement involving powerful domestic constituen- cies exists, theft on this scale by a subset of the ruling coalition is likely to undermine the political settlement and have additional negative effects on growth. Blocking these �nancial outflows is therefore likely to have a posi- tive effect on growth through both direct and indirect effects. The flows are thus rightly classi�ed as illicit. Many cases are more complex. Consider a plausible case wherein cap- ital flight is driven by attempts to evade environmental restrictions, labor laws, or other socially desirable regulations that reduce pro�ts. In prin- ciple, these could be welfare-enhancing regulations, and capital flight to evade them could be judged illicit. However, a more careful evaluation suggests that the issue may vary from case to case. Because developing countries are often competing on narrow margins with other developing countries, investors may threaten to leave and begin to transfer resources away from a country. What should the policy response be? If social poli- cies in the developing country are signi�cantly out of line with competi- tors, these policies may have made the country uncompetitive. Yet, removing all social protections is also not desirable. The real issue in this case is coordination in social policy that takes into account differences in initial conditions across countries, not an easy task. If such a coordinated policy structure is not possible across countries, the enforcement of restrictions on capital flight is unlikely to improve growth because domestic investors may become globally uncompetitive. These capital flows would therefore not be illicit according to our de�nition because blocking capital flight without deeper policy coordination may fail to improve economic outcomes in a particular country. This has important policy implications: all our effort should not be put into trying to block capital flight regardless of the underlying causes. Rather, the policy focus should be either to achieve the coordination of social policies across developing countries or, more realistic, to change policies such that reg- ulations are aligned across similar countries. Governance and Illicit Flows 49 An even more important example concerns policies to overcome major market failures such as those constraining technological capabili- ties. Growth in developing countries is typically constrained by low prof- itability because of the absence of formal and informal policies to address market failures and, in particular, the problem of low productivity because of missing tacit knowledge about modern production processes (Khan 2009). In the absence of policies that assist technological capabil- ity development, capital is likely to flow out of the developing country despite low wages. Countervailing policies, variously described as tech- nology policies or industrial policies, involve the provision of incentives for investors to invest in particular sectors and to put in the effort to acquire the missing tacit knowledge. In the presence of such policies, some restrictions on capital movements may be potentially bene�cial. The provision of incentives to invest in dif�cult processes of technology acquisition and learning may be wasted if domestic investors can claim the assistance without delivering domestic capability development. Unrestricted �nancial outflows may allow domestic investors to escape sanctions attached to poor performance. If the policy is well designed and the state has the capability to enforce it, restrictions on �nancial outflows may be socially bene�cial. Some amount of capital flight may still take place, and liberal economists may want to argue that this is a justi�cation for removing the policy and returning to a competitive market. However, if the market failures constraining investments are sig- ni�cant, this may be the wrong response, if a credible technology policy exists. The restrictions on capital flight in East Asian countries in the 1960s and 1970s worked dramatically because they combined signi�cant incentives for domestic investment with credible restrictions on capital flight. In such cases, capital flight would be directly damaging in the sense of lost investment and not have any indirect positive effects either. It would therefore be illicit according to our de�nition. The mirror image of this is capital flight driven by the absence of prof- itable domestic opportunities in a country without effective technology policies. Capital is likely to seek offshore investment opportunities. Even if this capital flight appears to be damaging in an immediate sense, it may not be. This is because attempting to block capital fleeing low prof- itability is unlikely on its own to solve the deeper problems of growth. Indeed, enforcing restrictions on capital outflows in a context of low 50 Draining Development pro�tability and absent policies to correct market failures could para- doxically make economic performance worse by reducing the incomes of nationals. Capital flight in these contexts is not necessarily damaging in terms of direct effects on growth, and the indirect effects may also not be negative unless we optimistically believe that blocking these flows will force the government into adopting the appropriate policies for tackling market failures. Because this is an unlikely scenario, it would be mislead- ing to classify these �nancial outflows as illicit. There is an important policy implication: developing countries need to design policies to address low productivity and to absorb new technologies. Attempting to block �nancial outflows without solving these problems will not neces- sarily improve social outcomes. A different and even more obvious case is one in which capital flight is induced by the presence of bad policies such as the protection of domestic monopolies that disadvantage investors who are not politically connected or privileged with monopoly rights. In this situation, if capi- tal leaves the country, the direct effects may appear to be damaging in the sense of lost investment, but may not be if domestic investment oppor- tunities are poor. The problem is not the capital flight, but the growth- reducing arrangements that induce it. Blocking �nancial outflows could, in an extreme case, lead to the consumption of capital by some investors because pro�table investment opportunities may be unavailable. The indirect effects of blocking �nancial outflows in this context may also not be positive, and the �nancial outflow is not usefully described as illicit. The appropriate response would be to remove some of the under- lying restrictions on investment. Finally, a particularly interesting set of cases concerns �nancial out- flows associated with activities that are directly growth-sustaining, but have signi�cant negative effects on a society’s political settlement and, therefore, on long-run growth through indirect effects. A classic example is the business associated with narcotics and drugs. For many countries, including relatively developed countries like Mexico, the income from the production and marketing of drugs is a signi�cant contributor to overall economic activity. By some measures, there could be a signi�cant positive effect on growth as a direct effect. However, given the legal restrictions on the business in many countries, the activity inevitably involves criminality and massive hidden rents that disrupt the under- Governance and Illicit Flows 51 lying political settlement of the producing country. The indirect nega- tive effects are likely to far outweigh any positive direct effect. It would be quite consistent with our de�nition to describe the �nancial flows asso- ciated with these sectors as illicit. Financial outflows with positive or neutral effects. We discuss in passing above a number of examples in which �nancial outflows do not have a net negative effect if both direct and indirect effects are accounted for. In some cases, this can involve making dif�cult judgments that are speci�c to the context. For instance, in the presence of signi�cant market failures facing investors, it may sometimes be useful not to enforce restrictions on capital flight too excessively. In textbook models, developing coun- tries lack capital and, therefore, capital should flow in if policies are undistorted. In reality, the productivity in developing countries is so low that most investments are not pro�table, and temporary incentives are needed to attract investments (Khan 2009). It is possible that some of these incentives have to be made available in other jurisdictions to be credibly secure from expropriation. In these contexts, the strict enforce- ment of restrictions on �nancial flows may reduce the degree of freedom states have in constructing credible incentives for investors taking risks in technology absorption and learning. For example, a signi�cant part of the foreign direct investments in India in the 1980s and 1990s came from jurisdictions such as Mauritius. A plausible interpretation is that much of this was Indian domestic capi- tal going through Mauritius to come back for reinvestment in India. One side of this flow was clearly a hidden �nancial outflow, the other a trans- parent inflow in the form of foreign direct investment. As a result, the developing country may not be a net capital loser, and, indeed, the incen- tives provided through this arrangement may make new productive investment possible in areas where investment may not otherwise have taken place. As in the case of odious debt, recycling is also only one side of a two-way flow and is generally assumed to be driven by tax and regu- latory arbitrage and to be harmful for society (for example, see Kant 1998, 2002; Schneider 2003b, 2003c). However, in the case of countries with low productivity and missing tacit knowledge, the rents captured in this way may, in some situations, serve to make investments more attrac- tive and thereby increase net investments. The direct effect may be to 52 Draining Development promote growth. Recycling through a foreign jurisdiction may also be a mechanism for hiding the source of funds in cases in which much of the initial capital base of emerging capitalists has involved questionable pro- cesses of accumulation that could be challenged by their competitors in terms of a formal interpretation of laws. Declaring these flows illicit for the purpose of blocking them may be a mistake. The direct effect is likely to be a reduction in investment because these types of accumulation may continue to remain hidden or be diverted into criminal activities. The expectation of positive indirect effects, for instance by creating dis- incentives for accumulating resources through questionable informal processes, may also be misguided given the structural informality in developing countries discussed elsewhere above. These are matters of judgment in particular cases. Sometimes, capital outflows may appear to be illicit simply because they are disallowed by ill-considered laws that cannot be enforced. If the laws were enforced, society might become even less well off. For instance, in some developing countries, vital imports may be illegal for no obvi- ous reason, forcing importers to engage in illegal �nancial transfers to get around the restrictions. In many developing countries, remitting for- eign exchange out of the country may also be disallowed for many pur- poses, including vitally important ones. An example is the widespread practice of illegally remitting foreign exchange from Bangladesh to for- eign employment agencies that want a commission for arranging over- seas employment for Bangladeshi workers. If, as a result, domestic work- ers are able to �nd employment on better terms than in the domestic market, their higher incomes are likely to have a positive effect on wel- fare, and their remittances are likely to support domestic growth. If ille- gal �nancial outflows involve payments to people smugglers, and most domestic workers end up less well off, the direct effects alone would make us classify the �nancial flows as illicit. A careful analysis is required in each case, but, in many cases, the problem may be an inappropriate legal or regulatory structure. Table 2.1 summarizes some of the examples we discuss. Only flows falling in the middle row of the table are properly illicit according to our de�nition. De�ning as illicit the other capital flows listed may be a policy error even in the case of �nancial flows in which the direct effect appears to be damaging. Nor is the legal-illegal divide of much use in the typical Governance and Illicit Flows 53 Table 2.1. Logical Framework for Identifying Illicit Flows from Developing Countries Direct effect on growth Net effect on growth Direct effect negative Direct effect positive or neutral (including indirect effects) (flow can be legal) (flow can be illegal) Net effect negative Examples: Capital outflows to Examples: Financial flows associated (flow is illicit even if legal) evade/avoid legitimate taxes with the production of drugs and Evasion of restrictions that support narcotics an effective industrial policy Financial outflows of political actors in the presence of live-and-let-live agreements Net effect positive or Examples: Financial outflows of Examples: Unauthorized payments neutral political actors in the absence of to overseas agencies to provide jobs (flow is not illicit even if informal agreements restricting that are better than domestic illegal) expropriation opportunities (Bangladesh) Recycling that bypasses critical market failures in developing econ- omies and increases investment by domestic investors in their own country Outflows from countries in which pro�tability is low because of absent industrial policy or presence of damaging policies such as protec- tion for domestic monopolies Source: Author compilation. developing-country case if the aim is to identify �nancial flows that need to be blocked to make the developing country more well off. In any par- ticular country, many �nancial flows may be simultaneously driven by different underlying causes. The judgment that has to be made involves identifying the drivers of the most signi�cant �nancial flows and assess- ing whether these are damaging in the context of the speci�c political settlement. Developing countries in crisis: The fragile developers Fragile developing countries are ones in which the internal political set- tlement is close to collapsing or has collapsed, and political factions are engaged in violent conflict. This group includes, for example, Afghani- stan, the Democratic Republic of Congo, and Somalia. It also includes a number of other developing countries in North and Sub-Saharan Africa, 54 Draining Development Latin America and the Caribbean, and Central Asia in which political settlements are highly vulnerable to collapse in the near future. IFFs may appear to be at the heart of their fragility and, indeed, may be fueling conflicts. However, we argue that it is dif�cult to de�ne what is illicit in such a context of conflict. Fundamental disagreements about the distri- bution of bene�ts are unlikely to be resolved without recourse to sys- temic violence. Violence is, nonetheless, not the distinctive feature of fragile coun- tries; there may also be pockets of intense violence in intermediate devel- opers such as Bolivia, Brazil, India, or Thailand. Rather, fragile countries are characterized by a signi�cant breakdown of the political settlement and, in extreme cases, also of social order. While pockets of rudimentary social order may spontaneously emerge in such societies, this is largely limited to the organization of violence and subnational economies sup- porting the economy of violence. There is a grey area between interme- diate developers facing growing internal conflicts and a developing country classi�ed as fragile. Nonetheless, in intermediate developers, while a few political groups and factions may be engaging in signi�cant violence, most signi�cant political factions are engaged in the normal patron-client politics of rent seeking and redistribution using the formal and informal mechanisms through which political settlements are con- structed in these countries. A political settlement is possible because there is a viable distribution of resources across the most powerful groups that reflects their relative power and that can be reproduced over time. This is a necessary condition describing a sustainable end to sig- ni�cant violence and the emergence of a political settlement. The de�ning characteristic of fragility is that a sustainable balance of power and a corresponding distribution of bene�ts across powerful political actors have not emerged. Violence is the process through which contending groups are attempting to establish and test the distribution of power on which a future political settlement could emerge. But this may take a long time because the assessment by different groups of what they can achieve may be unrealistic, and some groups may believe that, by �ghting long enough, they can militarily or even physically wipe out the opposition. In some cases, this belief may be realistic (Sri Lanka in 2010); in other cases, the attempt to wipe out the opposition can result in a bloody stalemate until negotiations about a different distribution of Governance and Illicit Flows 55 bene�ts can begin (perhaps Afghanistan in 2010). The analysis of what is socially damaging needs to be fundamentally reevaluated and rede- �ned in these contexts, and this has implications for our assessment of illicit flows. The historical examples tell us that apparently predatory resource extraction has been the precursor of the emergence of viable political settlements that have generated longer-term social order and viable states (Tilly 1985, 1990). Yet, in contemporary fragile societies, new circum- stances make it less likely that conflicts will result in the evolution of rela- tionships between organizers of violence and their constituents that resemble state building. First, natural resources can give organizers of violence in some societies access to previously unimaginable amounts of purchasing power, destroying incentives for internal coalition building with economic constituents in the sense described by Tilly. Second, the presence of advanced countries that manufacture sophisticated weapons and can pump in vast quantities of resources in the form of military assis- tance or aid also changes the incentives of domestic organizers of vio- lence. Instead of having to recognize internal distributions of power and promote productive capabilities, these people recognize that the chances of winning now depend at least partly on international alliances and the ability to play along with donor discourses. Domestic organizers who try to �ght wars by taxing their constituents in sustainable ways are likely to be annihilated by opponents who focus on acquiring foreign friends. These considerations should give us serious cause for concern in talk- ing about IFFs in these contexts in which the indirect effects of �nancial flows through the promotion or destruction of political stability are likely to far outweigh the direct effects on growth. By its nature the con- struction of politically stable settlement involves winners and losers: strategies of state or polity formation are not neutral in any sense. The terminology of illicit flows in these contexts should preferably be avoided, or its use should be restricted to �nancial flows that are illicit explicitly from the perspective of the observer. For instance, the flow of narcotics incomes, grey or black market transactions in the global arms market, or sales of natural resources by warlords may go against the legiti- mate interests of outsiders, and they are entitled to declare the associated �nancial flows as illicit from the perspective of their interests. This is justi�- able if these flows are causing damage to the interests of other countries. 56 Draining Development We should not, however, pretend that blocking particular flows is in any way neutral or necessarily bene�cial for the construction of viable political settlements, because there are likely to be many possible settle- ments that different groups are trying to impose. Moreover, competing groups of outsiders are also likely to be providing aid, guns, and, some- times, their own troops to their clients within the country. There is no easy way to claim that some of these resource flows are legitimate and constructive and others are not without exposing signi�cant political partiality toward particular groups. The real problem is that we do not know the outcomes of these con- flicts, and the internal distribution of power is both unstable and chang- ing. Because the sustainability of the eventual sociopolitical order that may appear depends on the emergence of a sustainable distribution of power among the key groups engaged in conflict, we cannot properly identify ex ante the resource flows that are consistent with a sustainable political settlement. The resource flows are likely to help determine the political settlement as much as sustain it. It follows that the analysis and identi�cation of illicit capital flows must be different across intermediate developers and across fragile developers. Fragile societies typically do not have internationally com- petitive sectors in their economies, nor do their competing leaders have the capacity to encourage productive sectors effectively. The adverse conditions created by conflict mean that signi�cant countervailing poli- cies would have to be adopted to encourage productive investment in these contexts. Though this is not impossible, as the example of the Pal- estinian Authority in the �ve years immediately following the Oslo Accords shows (Khan 2004), the likely direct growth effect of particular �nancial flows is a moot question in most cases of fragility. Globally competitive economic activities do take place in many con- flicts, but these are of a different nature from the development of tech- nological and entrepreneurial skills with which intermediate developers have to grapple. One example is natural resource extraction that is a spe- cial type of economic activity because it does not require much domestic technical and entrepreneurial capability. The returns may be large enough for some foreign investments or for extraction based on arti- sanal technologies even in war zones. A similar argument applies to the Governance and Illicit Flows 57 cultivation of plants associated with the manufacture of narcotics. The �nancial flows from these activities are likely to be directly controlled by political actors and may provide the funds for sustaining the conflict. The only type of capital flight that is likely to emanate directly from the decisions of economic actors is the obvious one of attempting to escape destruction or appropriation. But if individuals try to move their assets out of a war zone to avoid expropriation, it would be unhelpful to char- acterize these as illicit. Of most concern in fragile societies are the flows organized by politi- cal actors, as these are bound to be connected with ongoing conflicts in some way. In the case of intermediate developers, our concern is that, under some conditions, restrictions on �nancial flows may inadvertently increase the stakes for holding on to power. A similar, but obviously more serious set of uncertainties affects the analysis of conflicts. In the- ory, if all parties to a conflict were blocked from accessing the outside world, this may have a positive effect in forcing them to recognize the existing distribution of power and reaching a compromise more quickly than otherwise. In the real world, a total sealing off is unlikely. Some par- ties to the conflict are likely to be recognized by outside powers as legiti- mate well before an internal political settlement has been arrived at and receive �nancial and military assistance. Two entirely different outcomes may follow. The less likely is that the groups excluded by the international community recognize the hopelessness of their situation and either capitulate or agree to the settlement that is offered. A more likely outcome, as in Afghanistan, is that external assistance to help one side while attempting to block resource flows to the other by declaring these to be unauthorized or illicit can increase the local legitimacy of the opposition and help to intensify the conflict. In the end, every group is likely to �nd some external allies and ways to funnel resources to �ght a conflict where the stakes are high. Paradoxically, �nding new foreign allies becomes easier if the enhanced legitimacy of the excluded side makes it more likely that they will win. These considerations suggest that a neutral way of de�ning IFFs is particularly dif�cult, perhaps impossible, in the case of fragile developers. Table 2.2 summarizes our analysis in this section. 58 Draining Development Table 2.2. Illicit Financial Flows in Different Contexts Policy focus in Country Main types of addressing illicit type De�ning features Main policy concern illicit capital flows capital flows Advanced High average incomes Laws and �scal programs Mainly flows that violate Strengthen enforcement economies and long-term political should maintain social existing laws (for and regulation. At stability. Political process cohesion and economic example, tax evasion). moments of crisis, responds effectively to growth. Occasionally refer to attempt to bring law economic underperfor- legal flows where laws back into line with broad mance and distributive no longer reflect social social consensus on concerns. consensus or economic economic and political sustainability. goals. Intermediate Lower average incomes, Develop, maintain, and Flows that undermine Economic policies to or normal and politics based on a expand viable develop- developmental strate- enhance pro�tability developers combination of formal ment strategies, in gies. Capital flight in a primarily by addressing and informal (patron- particular the develop- context of a failure to critical market failures. client) redistributive ment of broadbased raise domestic pro�t- Build governance arrangements. Achieves productive sectors. ability is problematic, capabilities to enforce long-term stability but is not necessarily restrictions on �nancial without sustained illicit. Flows associated flows that make these violence. with internationally policies less vulnerable to criminalized activities political contestation. like drugs. Fragile Breakdown of existing State building and Not possible to de�ne A viable political settle- developing political settlements reconstruction of a illicit flows in a neutral ment requires compet- countries resulting in the outbreak viable political way when elites are in ing groups to accept a of sustained violence settlement to allow conflict. The effects of distribution of bene�ts that undermines longer- society to embark on a �nancial flows have to consistent with their term conditions for the sustainable path of be judged primarily in understanding of their maintenance of basic economic development. terms of their effects relative power. Dif�cult sociopolitical order. on the establishment for outsiders to con- of a particular political tribute positively and settlement. easy to prolong conflicts inadvertently. Source: Author compilation. Policy Responses to Illicit Flows Our discussion above suggests that capital outflows from developing countries qualify as illicit according to our de�nition if they have a nega- tive economic impact on a particular country after we take both direct and indirect effects into account in the context of a speci�c political settlement. Both judgments about effects are counterfactual in the sense that we are asking what would directly happen to growth if a particular flow could be blocked, and then we are asking what would happen after the indirect effects arising from adjustments by critical stakeholders to the new situation have taken place. The latter assessment depends on our knowledge about the economy and polity of the society as summarized Governance and Illicit Flows 59 in the political settlement. This methodology allows us to derive some useful policy conclusions in the case of intermediate developers. By con- trast, the use of the illicit flow terminology involves signi�cant dangers in the case of fragile countries. Policy makers should at least be explicitly aware of this. In this section, we discuss a sequential approach for the identi�cation of feasible policy interventions to address illicit capital outflows that sat- isfy our de�nition and that occur in intermediate developers. Assume that a particular capital outflow from an intermediate devel- oper is illicit according to our de�nition. The core policy concern here is a microlevel assessment of how any particular set of measures to restrict the illicit capital outflow affects the macrodynamics of the economy and society in question. While our assessment about the nature of an illicit flow has taken into account the indirect effects of blocking the flow, the implementation of any policy almost always gives rise to unintended consequences, and all the indirect effects may not be understood. More often than not, policy failure is the failure to take such unintended con- sequences into account ex ante. An illustration of such unintended consequences, in this case pertain- ing to the international regulation of �nancial flows, is provided by Gap- per’s analysis (2009) of the Basel Accords I (1998) and II (2004), which were designed to regulate global banks essentially by setting higher capi- tal adequacy standards and improving the measurement of leverage. Ironically, these standards inadvertently accelerated rather than muted �nancial engineering by banks, in particular the securitization of risky mortgage debts. The accords raised the threshold of responsible bank- ing, but simultaneously created incentives to �nd a way around the thresholds that would prove more disastrous to the global economy. As Gapper remarks (2009, 1), “it would be wrong to throw away the entire Basel framework . . . because global banks found ways to game the sys- tem.� The obvious implication is that one must strengthen rather than throw away existing regulations. More generally, in our view, an effective policy response to illicit flows requires a step-by-step, sequential assessment of the macrolevel effects of blocking particular capital flows. To illustrate what this entails in the case of an intermediate developer, consider the example provided by Gulati (1987) and Gordon Nembhard’s analysis (1996) of industrialization 60 Draining Development policies in Brazil and the Republic of Korea from the 1960s to the 1980s. Both authors argue that trade misinvoicing in these countries during this period followed an unusual pattern in that imports tended to be under- invoiced. Both authors explain this in terms of domestic producers and traders, “rather than being preoccupied with evading controls to earn foreign exchange, may be more concerned about meeting export and production targets and maximizing government plans� (Gordon Nemb- hard 1996, 187). In these two countries (but not necessarily in others), there were potentially avoidable losses from trade misinvoicing, and these flows were therefore illicit. Nonetheless, the losses were signi�cantly smaller than the gains from the incentives created for higher investments in sectors promoted by the industrial policy. To decide whether or not to block illicit trade misinvoicing in these circumstances (and, if so, how), the most important aspect of policy design is to prepare a microlevel analysis of the impact of different ways of blocking these flows on the overall economy. This analysis has to take account of economic and political circumstances and the likely impact of particular regulatory strategies. Figure 2.2 illustrates the iterative process required to arrive at policy decisions that avoid unintended negative consequences. Figure 2.2. Choosing Attainable Benchmarks for Policy Design INTERVENTION 1. Policy regime 2. New policy regime ILLICIT PRACTICES INTERVENTION 4. New policy regime 3. New policy regime ILLICIT PRACTICES INTERVENTION 5. New policy regime 6. New policy regime Source: Author compilation. Governance and Illicit Flows 61 Assume that the policy intervention from stage 1 to stage 2 in �gure 2.2 introduces an industrial policy regime to promote manufacturing exports. This, in turn, provides incentives for some illicit practices, giving rise to stage 3. Despite the presence of the illicit practices arising from the implementation of the industrial policy introduced in stage 2, the overall development outcome at stage 3 is reasonably good. There are several possible responses to the observation of illicit practices at stage 3. One option is to abandon the export-promoting industrial policies on the grounds that these create the incentive to underinvoice imports. This would be in line with the portfolio approach to capital flight that adopts competitive markets as the benchmark model. However, if critical mar- ket failures were signi�cant to start with, developmental outcomes at position 1 may be worse than at position 3. Hence, an immediate response to the illicit flow problem at stage 3 that does not take account of devel- opmental problems at stage 1 may be self-defeating. Indeed, in terms of our de�nitions, the �nancial flows at position 3 are not illicit with respect to position 1, though they are illicit with respect to position 2. An obvious option would be to run a more ef�cient customs adminis- tration to raise the transaction costs of import underinvoicing, ideally to the point at which this becomes unpro�table. This is the preferable policy option: it could shift the policy framework from position 3 to position 4. The result would be a reduction in underinvoicing that then takes us to position 5, a combination of the new policy framework and a new (reduced) level of illicit flows. The developmental outcomes at stage 5 would be better than those at stage 3, and, so, this response to the illicit flow problem would be entirely justi�ed. Yet, to get from position 3 to position 5 requires a careful microlevel analysis of the costs and bene�ts of different policies. If import underinvoicing and the illicit pro�ts asso- ciated with it could be eliminated without jeopardizing the participation of private sector �rms in the export promotion program, the policy of better enforcement would be effective and economically justi�ed. In contrast, consider the case (such as Brazil in the 1970s rather than Korea) in which tolerating some import underinvoicing is an informal incentive to ensure private sector participation in a growth-enhancing export promotion program. If the extra incomes from this source are important to ensuring private sector participation in the policy, an 62 Draining Development attempt to remove these rents could inadvertently result in more damag- ing rent seeking by the private sector or the sector’s political refusal to participate in industrial policy programs. If either happened, the econ- omy would be less well off and could revert to the less preferable position at stage 1. This assessment would depend on our understanding of the political settlement and the power of the private sector to resist the impo- sition of policies that the sector perceives to be against its interests. The attempt to control the illicit flows would then have failed in a develop- mental sense even if the illicit flows disappeared. The core policy task in this case would not be to block the illicit capital flows associated with stage 3 (import underinvoicing), but to support the creation of state capacities and an adjustment of the political settlement that would allow the effective implementation of ef�cient customs administration in the future without reliance on informal perks such as import underinvoicing. More generally, this suggests that a careful sequential microanalysis of the macro-outcomes of blocking particular illicit capital flows is required, locating particular strategies in the context of speci�c initial conditions. Conclusion Our core concern has been to show that what constitutes an illicit capital outflow can vary across countries, depending on the core political and economic features. The number of capital outflows that are unequivocally illicit in our minimal de�nition is likely to be more limited than the num- ber that might be de�ned as such in single-driver approaches based on questionable underlying economic models. Quantitative estimates of illicit flows using our de�nition are therefore likely to be considerably less spectacular than, for example, recent estimates of dirty money (see Kar and Cartwright-Smith 2008). This is precisely our point. We argue that, if economic development is to be the main concern in this debate, then the promotion of economic growth in the context of a distribution of bene�ts that is politically viable has to be the condition that any policy intervention must meet. Wider criteria than this to de�ne illicit capital outflows as a target for policy intervention by necessity rely on some broader notion of the social or economic good based on an abstract underlying model that may have little relevance in seeking development in the real world. Governance and Illicit Flows 63 Our analysis also questions the validity of a core premise of the capi- tal flight literature, namely, that all flight capital, if retained domestically, will yield a higher social rate of return (Schneider 2003a; Cumby and Levich 1987; Walter 1987). The purpose of our analytical typology is to show that, from the point of view of economic and political develop- ment, blocking all capital outflows from developing countries is not desirable. Moreover, as we outline in table 2.1, blocking some apparently damaging flows can also do more harm than good if these flows are driven by deeper problems (such as the absence of industrial policy, live- and-let-live rules between competing political factions, or the presence of domestic monopolies). In these cases, we argue, it is a mistake to describe the resultant flows as illicit. The policy implication is not that nothing should be done in such cases. It is rather that the solution is to look for policies that address the underlying structural problems. In the case of fragile developers, our analysis points out the dangers of too eas- ily de�ning what is illicit. Finally, our sequential impact assessment sug- gests that there are dangers even in simplistically attacking capital flows that we do deem illicit. Even in these cases, it is possible that some policy responses are more effective than others, and policies that tackle illicit capital flows in isolation from the wider political settlements of which they are a part may leave society less well off. Consider the following example on anti–money laundering (AML) policies. The cost-effectiveness of the AML regime promoted by the intergovernmental Financial Action Task Force is doubtful even for advanced economies (Reuter and Truman 2004; Schneider 2005).4 Shar- man (2008) extends this analysis to developing countries, with a special focus on Barbados, Mauritius, and Vanuatu, and �nds that AML stan- dards have had a signi�cant net negative impact on these three develop- ing countries. Essentially, the costs imposed on legitimate businesses and states having limited administrative and �nancial capacity far out- weighed the extremely limited bene�ts in terms of convictions or the recovery of illicit assets. Yet, according to Sharman, rather than rethink- ing the policy model, its enforcement was strengthened mainly by black- listing noncompliant countries and through competition between states for international recognition and imitation. Sharman (2008, 651) highlights the case of Malawi, as follows: 64 Draining Development Malawi is not and does not aspire to be an international �nancial center, nor has it been associated with money laundering or the �nancing of ter- rorism. Speaking at an international �nancial summit in September 2006, the Minister of Economics and Planning recounted how his country had come to adopt the standard package of AML regulations. The Minister was told that Malawi needed an AML policy. The Minister replied that Malawi did not have a problem with money laundering, but was informed that this did not matter. When the Minister asked if the package could be adapted for local conditions he was told no, because then Malawi would not meet international standards in this area. The Minister was further informed that a failure to meet international AML standards would make it harder for individuals and �rms in Malawi to transact with the outside world relative to its neighbors, and thus less likely to attract foreign invest- ment. The Minister concluded: “We did as we were told.� Notes 1. “The term, illicit �nancial flows, pertains to the crossborder movement of money that is illegally earned, transferred, or utilized.� (Emphasis in original.) Global Financial Integrity summary fact sheet on illicit �nancial flows, http://www.g�p. org/storage/gfip/documents/illicit%20flows%20from%20developing%20 countries%20overview%20w%20table.pdf. 2. The most commonly applied variant of this residual measure subtracts capital outflows or uses of foreign exchange (the current account de�cit and increases in central bank reserves) from capital inflows (net external borrowing, plus net foreign direct investment). An excess of inflows over outflows is interpreted as indicative of capital flight. Nonbank variants of this measure include net acqui- sitions of foreign assets by the private banking system in capital outflows. 3. Our intermediate developers should not be confused with middle-income coun- tries. Relatively poor and middle-income developing countries can both have sustainable political settlements, and both can suffer from internal political crises that result in fragility as a result of a collapse in the internal political settlements. 4. See Financial Action Task Force, Paris, http://www.fatf-ga�.org/. References Baker, R. W. 2005. Capitalism’s Achilles Heel: Dirty Money and How to Renew the Free- Market System. Hoboken, NJ: John Wiley & Sons. Baker, R. W., and J. Nordin. 2007. “Dirty Money: What the Underworld Understands That Economists Do Not.� Economist’s Voice 4 (1): 1–3. Governance and Illicit Flows 65 Boyce, J. K., and L. Ndikumana. 2001. “Is Africa a Net Creditor? New Estimates of Capital Flight from Severely Indebted Sub-Saharan African Countries, 1970– 96.� Journal of Development Studies 38 (2): 27–56. Boyce, J. K., and L. Zarsky. 1988. “Capital Flight from the Philippines, 1962–1986.� Philippine Journal of Development 15 (2): 191–222. Cerra, V., M. Rishi, and S. C. Saxena. 2008. “Robbing the Riches: Capital Flight, Insti- tutions and Debt.� Journal of Development Studies 40 (8): 1190–213. Collier, P. 1999. “On the Economic Consequences of Civil War.� Oxford Economic Paper 51 (1): 168–83. Collier, P., A. Hoeffler, and C. A. Pattillo. 2001. “Flight Capital as a Portfolio Choice.� World Bank Economic Review 15 (1): 55–80. Crotty, J. 2009. “Structural Causes of the Global Financial Crisis: A Critical Assess- ment of the ‘New Financial Architecture.’ � Cambridge Journal of Economics 33 (4): 539–62. Cuddington, J. T. 1986. Capital Flight: Estimates, Issues, and Explanations. Princeton Studies in International Finance 58 (December). Princeton, NJ: Princeton University. ———. 1987. “Capital Flight.� European Economic Review 31 (1–2): 382–88. Cumby, R. E., and R. M. Levich. 1987. “On the De�nition and Magnitude of Recent Capital Flight.� In Capital Flight and Third World Debt, ed. D. R. Lessard and J. Williamson, 26–67. Washington, DC: Peterson Institute for International Economics. Dooley, M. P. 1986. “Country-Speci�c Risk Premiums, Capital Flight and Net Invest- ment Income Payments in Selected Developing Countries.� Departmental memorandum, March, Research Department, International Monetary Fund, Washington, DC. ———. 1988. “Capital Flight: A Response to Differences in Financial Risks.� IMF Staff Papers 35 (3): 422–36. Eatwell, J., and L. Taylor. 2000. Global Finance at Risk. New York: New York Press. Epstein, G. A., ed. 2006. Capital Flight and Capital Controls in Developing Countries. Cheltenham, U.K.: Edward Elgar. Erbe, S. 1985. “The Flight of Capital from Developing Countries.� Intereconomics 20 (6): 268–75. Eurodad (European Network on Debt and Development). 2008a. “Eurodad Fact Sheet: Capital Flight Diverts Development Finance.� Eurodad Report, May 8, Eurodad, Brussels. ———. 2008b. “Addressing Development’s Black Hole: Regulating Capital Flight.� Eurodad Report, May 8, Eurodad, Brussels. Franko. P. 2003. The Puzzle of Latin American Economic Development. Lanham, MD: Rowman & Little�eld. Gapper, J. 2009. “How Banks Learnt to Play the System.� Financial Times, May 6. 66 Draining Development Global Witness. 2009. Undue Diligence: How Banks Do Business with Corrupt Regimes. Report, March. London: Global Witness. http://www.undue-diligence.org/Pdf/ GW_DueDilligence_FULL_lowres.pdf. Gordon, D., and R. Levine. 1989. “The Problem of Capital Flight: A Cautionary Note.� World Development 12 (2): 237–52. Gordon Nembhard, J. 1996. Capital Control, Financial Regulation, and Industrial Policy in South Korea and Brazil. London: Praeger. Gulati, S. K. 1987. “A Note on Trade Misinvoicing.� In Capital Flight and Third World Debt, ed. D. R. Lessard and J. Williamson, 68–78. Washington, DC: Peterson Institute for International Economics. Heggstad, K., and O.-H. Fjeldstad. 2010. “How Banks Assist Capital Flight from Africa: A Literature Review.� CMI Report R 2010: 6 (January), Chr. Michelsen Institute, Bergen, Norway. Hermes, N., and R. Lensink. 1992. “The Magnitude and Determinants of Capital Flight: The Case for Six Sub-Saharan African Countries.� De Economist 140 (4): 515–30. Hollingshead, A. 2010. “The Implied Tax Revenue Loss from Trade Mispricing.� Global Financial Integrity, Washington, DC. http://www.g�p.org/storage/g�p/ documents/reports/implied%20tax%20revenue%20loss%20report_�nal.pdf. Kant, C. 1998. “Capital Inflows and Capital Flight: Individual Country Experiences.� Journal of Economic Integration 13 (4): 644–61. ———. 2002. “What Is Capital Flight?� World Economy 25 (3): 341–58. Kapoor, S. 2007. “Haemorrhaging Money: A Christian Aid Brie�ng on the Problem of Illicit Capital Flight.� Christian Aid, London. Kar, D., and D. Cartwright-Smith. 2008. “Illicit Financial Flows from Developing Countries, 2002–2006.� Global Financial Integrity, Washington, DC. http://www .g�p.org/storage/g�p/economist%20-%20�nal%20version%201-2-09.pdf. Khan, M. H., ed. 2004. State Formation in Palestine: Viability and Governance during a Social Transformation. With G. Giacaman and I. Amundsen. Routledge Politi- cal Economy of the Middle East and North Africa Series. London: Routledge Curzon. ———. 2005. “Markets, States and Democracy: Patron-Client Networks and the Case for Democracy in Developing Countries.� Democratization 12 (5): 705–24. ———. 2006. “Determinants of Corruption in Developing Countries: The Limits of Conventional Economic Analysis.� In International Handbook on the Economics of Corruption, ed. S. Rose-Ackerman, 216–44. Cheltenham, U.K.: Edward Elgar. ———. 2009. “Learning, Technology Acquisition and Governance Challenges in Developing Countries.� Report, Research Paper Series on Governance for Growth. School of Oriental and African Studies, University of London, London. http://www.d�d.gov.uk/r4/SearchResearchDatabase.asp?OutPutId=181367. ———. 2010. “Political Settlements and the Governance of Growth-Enhancing Institutions.� Report, Research Paper Series on Governance for Growth, School Governance and Illicit Flows 67 of Oriental and African Studies, University of London, London. http://eprints. soas.ac.uk/9968/. Khan, M. H., and S. Blankenburg. 2009. “The Political Economy of Industrial Policy in Asia and Latin America.� In Industrial Policy and Development: The Political Economy of Capabilities Accumulation, ed. M. Cimoli, G. Dosi, and J. E. Stiglitz, 336–77. Initiative for Policy Dialogue Series. Oxford: Oxford University Press. Khan, M. H., and K. S. Jomo, eds. 2000. Rents, Rent-Seeking and Economic Develop- ment: Theory and Evidence in Asia. Cambridge, U.K.: Cambridge University Press. Khan, M. S. and N. Ul Haque. 1985. “Foreign Borrowing and Capital Flight: A For- mal Analysis.� IMF Staff Papers 32 (4): 606–28. Kindleberger, C. P. 1937. International Short-Term Capital Movements. New York: Augustus Kelley. ———. 1990. Historical Economics. New York: Harvester Wheatsheaf. Lessard, D. R., and J. Williamson. 1987. “The Problem and the Policy Responses.� In Capital Flight and Third World Debt, ed. D. R. Lessard and J. Williamson, 201– 54. Washington, DC: Peterson Institute for International Economics. Myrvin, L. A., and A. J. Hughes Hallett. 1992. “How Successfully Do We Measure Capital Flight? The Empirical Evidence from Five Developing Countries.� Jour- nal of Development Studies 28 (3): 538–56. North, D. C., J. J. Wallis, S. B. Webb, and B. R. Weingast. 2007. “Limited Access Orders in the Developing World: A New Approach to the Problem of Development.� Policy Research Working Paper 4359, World Bank, Washington, DC. Palma, G. 2005. “Four Sources of De-Industrialization and a New Concept of the ‘Dutch Disease.’ � In Beyond Reform: Structural Dynamics and Macroeconomic Vulnerability, ed. J. A. Ocampo, 71–116. Washington, DC: World Bank; Berke- ley, CA: Stanford University Press. ———. 2009. “The Revenge of the Market on the Rentiers: Why Neo-liberal Reports of the End of History Turned Out to be Premature.� Special issue, Cambridge Journal of Economics 33 (4): 829–69. Pollin, R. 2003. Contours of Descent: U.S. Economic Fractures and the Landscape of Global Austerity. London: Verso. Reuter, P., and E. M. Truman. 2004. Chasing Dirty Money: The Fight against Money Laundering. Washington, DC: Institute for International Economics. Rishi, M., and J. K. Boyce. 1990. “The Hidden Balance of Payments: Capital Flight and Trade Misinvoicing in India, 1971–1986.� Economic and Political Weekly 25 (30): 1645–48. Schneider, B., ed. 2003a. The Road to International Financial Stability: Are Key Financial Standards the Answer? International Political Economy Series. Lon- don: Overseas Development Institute; Basingstoke, U.K.: Palgrave Macmillan. ———. 2003b. “Measuring Capital Flight: Estimates and Interpretations.� Working Paper 194 (March), Overseas Development Institute, London. 68 Draining Development ———. 2003c. “Resident Capital Outflows: Capital Flight or Normal Outflows? Sta- tistical Interpretations.� Working Paper 195 (March), Overseas Development Institute, London. ———. 2005. “Do Global Standards and Codes Prevent Financial Crises? Some Pro- posals on Modifying the Standards-Based Approach.� UNCTAD Discussion Paper 177 (April), United Nations Conference on Trade and Development, Geneva. Sharman, J. C. 2008. “Power and Discourse in Policy Diffusion: Anti–money Laun- dering in Developing States.� International Studies Quarterly 52 (3): 635–56. Smithin, J. 1996. Macroeconomic Policy and the Future of Capitalism: The Revenge of the Rentiers and the Threat to Prosperity. Cheltenham, U.K.: Edward Elgar. Tilly, C. 1985. “War Making and State Making as Organized Crime.� In Bringing the State Back In, ed. P. B. Evans, D. Rueschemeyer, and T. Skocpol, 161–90. Cam- bridge, U.K.: Cambridge University Press. ———. 1990. Coercion, Capital and European States. Oxford: Blackwell. Tornell, A., and A. Velasco. 1992. “The Tragedy of the Commons and Economic Growth: Why Does Capital Flow from Poor to Rich Countries?� Journal of Polit- ical Economy 100 (6): 1208–31. Vos, R. 1992. “Private Foreign Asset Accumulation, Not Just Capital Flight: Evidence from the Philippines.� Journal of Development Studies 28 (3): 500–37. Walter, I. 1987. “The Mechanisms of Capital Flight.� In Capital Flight and Third World Debt, ed. D. R. Lessard and J. Williamson, 103–28. Washington, DC: Peterson Institute for International Economics. World Bank. 1985. World Development Report 1985: International Capital and Eco- nomic Development. New York: World Bank; New York: Oxford University Press. Wray, R. 2009. “The Rise and Fall of Money Manager Capitalism: A Minskian Approach.� Cambridge Journal of Economics 33 (4): 807–28. 3 The Political Economy of Controlling Tax Evasion and Illicit Flows Max Everest-Phillips Abstract This chapter examines illicit capital flows from developing countries through analysis of one major component of the problem, tax evasion. The political economy de�nition put forward in the chapter clari�es that the root cause of all illicit capital outflows is ultimately not poor policy or capacity constraints in administration, but the failure of political will. Tax evasion and the other major sources of illegal flows (criminality and cor- ruption) flourish in the absence of the political ambition to build a legiti- mate, effective state. Such effectiveness, including the effectiveness of tax systems, derives from formal and informal institutional arrangements (political settlements) that establish state legitimacy, promote prosperity, and raise public revenue. The commitment to these arrangements distin- guishes illicit from illegal. It requires that political leaders and taxpayers perceive the need for effective tax systems to provide the state with the resources necessary to enforce their own property rights, deliver political stability, and promote economic growth. The extent and form of tax eva- sion derive from the political consensus to tax effectively and develop the administrative capacity to do so. This, in turn, shapes and reflects the intrinsic willingness to pay taxes (tax morale) of taxpayers. To contain 69 70 Draining Development illicit capital flows, including tax evasion, the weak legitimacy of the state in many developing countries must be addressed. If regime leaders and elites are not prepared to tax themselves and prevent free-riding, genuine political ownership of efforts to curb illicit capital flows is problematic. Introduction: Effective States, Tax Evasion, and Illicit Capital Flows The G-8 world leaders, meeting in July 2009, recognized “the particu- larly damaging effects of tax evasion for developing countries.�1 Effective states require effective, ef�cient, and equitable tax systems. Creating the commitment of citizens not to evade taxation is a political process cen- tral to state building; cajoling elites to pay taxes has always been an essential step to any state becoming effective. Bad governance manifests itself through an unjust tax system and rampant tax evasion (the illegal avoidance of paying taxation) (Everest-Phillips 2009a).2 Throughout history, tax has been a symbol of state authority; tax evasion an indicator of political resistance.3 Tax evasion is signi�cantly correlated with dis- satisfaction with government, political interference in the economy, and weak governance (see, respectively, Lewis 1982; Kim 2008; Hayoz and Hug 2007). In modern times, widespread tax evasion remains a trigger for and indicator of political instability.4 Political and economic malaise intertwine. For instance, after the Tequila Crisis in Mexico, tax evasion increased from 43 percent of the total take in 1994, to 54 percent in 1995, and to 59 percent in the �rst half of 1996.5 Political economy de�nition: Flight or flow? Illegal or illicit? Considerable confusion has arisen in the nascent literature on illicit cap- ital flows from the failure to develop conceptual clarity between capital flight and capital flows and between illicit and illegal. To establish clear de�nitions requires recognition that effective, legitimate states are cru- cial for development (DFID 2006). Illicit indicates that the activity is generally perceived as illegitimate, which, in turn, requires the state to be regarded as legitimate. International capital transfers become illicit if they originate from an illegal source (evasion, corruption, or criminal- ity) or are illegal by bypassing capital controls, but also immoral in undermining the state’s willingness and capacity to deliver better lives The Political Economy of Controlling Tax Evasion and Illicit Flows 71 for its citizens. (For simplicity, we treat illicit here to mean immoral and illegal, but there also exist potentially immoral, but not illegal capital flows, such as those arising from aggressive tax avoidance.) Flight indi- cates exit from risk; flow is simply a movement of liquid �nancial assets. By this de�nition, illegal capital flight is the rational, but unlawful, shift of assets away from high to low political risk, whereas illicit capital flows take place by choice despite a state’s political legitimacy evinced by a credible commitment to development in the national interest (table 3.1). The frequent conflation of illicit and illegal loses the important moral dimension for developing countries and their partners in the interna- tional community of the need to tackle poverty and deliver the United Nations’ Millennium Development Goals. The three main drivers of illicit capital flows—tax evasion, corrup- tion, and criminality—are, at once, both causes and effects of the fragil- ity of state institutions and, so, challenge perceived legitimacy. Tax eva- sion undermines the funding of the state and, thus, the legitimacy associated with the state through the delivery of public services; corrup- tion weakens the moral legitimacy of the state; and criminality chal- lenges the legitimacy of state authority. However, while the rationale for the illicit capital flows driven by these factors is clear, the relative impor- tance of these factors in different contexts and the connections between them remain uncertain and unquanti�ed.6 This chapter outlines the political economy of the apparently most common and undoubtedly most politically complex and signi�cant of illicit flows, the flows linked to tax evasion. It argues that, because political Table 3.1. Proposed Typology: Illegal and Illicit Capital Flows and Capital Flight Capital movement Illegal Illicit Capital flow (from greed) Unlawful asset transfer to hide Asset transfer that immorally wealth from the state undermines a legitimate state’s capacity and commitment to development Capital flight (from risk) Unlawful asset transfer from a Immoral asset transfer to poor to a good governance reduce exposure to political risk context to reduce political risk (for example, from nascent (for example, from arbitrary democracy) autocratic rule) Source: Author compilation. 72 Draining Development governance is central to the problem of illicit capital flows, tackling the flows requires more than technocratic solutions. Domestic and interna- tional efforts to contain illicit capital flows must include building state legitimacy through a locally credible political vision for sustainable devel- opment. It concludes that the key policy concern in attempts to control illicit capital flows must be how to create the political interests and institu- tional incentive structures so that political leaders and elites will tax them- selves, thereby curtailing the potential bene�ts from those forms of illicit capital flows that may currently be funding their grip on power. Effective states tackle illicit capital flows Where poor political governance reflects and exacerbates political inse- curity, signi�cant illicit capital flight may occur. Any modern state that is not a tax haven or an oil exporter requires resourcing through an effec- tive tax system capable of constraining tax evasion. The state also requires perceived legitimacy so that taxpayers and citizens in general do not regard illegal capital outflows as licit, that is, as morally legitimate. Yet, many developing states are highly aid dependent and lack legitimacy in the eyes of their citizens; so, they need to strengthen their tax base and the willingness of their citizens to pay taxes (Chabal and Daloz 1999; Everest-Phillips 2011). Overcoming this development conundrum requires recognition that politics shape both the root cause and the potential solution. Tax systems, which comprise legislation, policy, and administration, are part of this, reflecting the national consensus on col- lective action, through the political process, for resourcing essential pub- lic goods such as security, rule of law, and the provision of basic services (Timmons 2005). Taxation reflects the intrinsic legitimacy of the state (based on consent, manifested among taxpayers as tax morale, their inherent willingness to pay taxes) and funds the effectiveness of state institutions (manifested in actual compliance).7 The extent of tax evasion thus provides a good indicator of the com- parative quality of governance.8 The state cannot function without rev- enue; how that revenue is raised (the balance between coercion and con- sent) mirrors the relationship between society and the state, but also shapes it: tax is state building in developing countries (Bräutigam, Fjeld- stad, and Moore 2008; Everest-Phillips 2008a, 2010).9 While the size and structure of a country’s economy signi�cantly influence the types and The Political Economy of Controlling Tax Evasion and Illicit Flows 73 yield of the taxes that can be collected, the root cause of weaknesses in tax systems in many countries is ultimately not a limited economy, poor tax policy, or capacity constraints in tax administrations (Stotsky and WoldeMariam 1997). Rather, the problem is that taxation reflects politi- cal attitudes toward the state and the type of state that taxation should fund.10 Tax evasion and the potential illicit capital flows arising from it there- fore flourish if there is a lack of political determination to build the legit- imacy of the state needed to make taxation effective and deliver develop- ment (Therkildsen 2008; Everest-Phillips 2009a).11 For example, across Africa, the postindependence collapse in governance was reflected in tax evasion: thus, in Malawi, tax evasion increased sevenfold between 1972 and 1990.12 Corruption also flourished as the state became the vehicle for personal enrichment by the elite, and the state’s challenged legiti- macy encouraged institutionalized criminality (Bayart, Ellis, and Hibou 1999). As a result of such dynamics, containing all forms of illicit capital flows requires building state legitimacy and improving political and eco- nomic governance. Empirical data consistently con�rm the importance of the perceived quality of governance for tax compliance. Public per- ception of high levels of corruption, for example, is associated with high levels of tax evasion; this is both cause and effect given that higher levels of corruption lower the ratio of taxes to gross domestic product (GDP) (Rose-Ackerman 1999; Ghura 2002). Tax compliance does not depend solely or probably not even largely on a taxpayer’s analysis of cost, ben- e�t, and risk of evasion (Alm and Martinez-Vazquez 2003). Instead, tax- payers evade and illicitly move capital abroad not simply because the potential bene�t outweighs the perceived risk, but because they believe the state lacks the legitimacy of capable, accountable, and responsive governance (Steinmo 1993). Indeed, in the face of rampant corruption, criminality, and waste in the public sector, tax evasion is seen as legitimate (Tanzi and Shome 1993). Widespread tax evasion indicates a lack of tax effort: the state lacks the will to engage with its citizens, and citizens do not believe their interests coincide with those of the state. The �scal social contract— namely, a credible commitment among leaders, elites, and citizens to pay for and deliver sustainable national development—does not exist. The 74 Draining Development popular Russian attitude toward tax evasion exempli�es the relevant political context found in many developing countries, as follows:13 First, [Russian citizens] rightly did not believe that all “the other� taxpay- ers were paying their taxes properly, so it was really no point in being “the only one� who acted honestly. The goods (public, semi-public or private) that the government was going to use the money to produce would simply not be produced because there were too little taxes paid in the �rst place. Secondly, they believed that the tax authorities were corrupted, so that even if they paid their taxes, a signi�cant part of the money would never reach the hospitals or schools, etc. Instead, the money would �ll the pock- ets of the tax bureaucrats. (Rothstein 2001, 477) Thus, tax evasion creates a vicious circle: the perception that others are not paying tax drastically reduces compliance and delegitimizes the state, fueling both capital flight and capital flows. So, for example, the private assets held abroad by Venezuelan citizens more than doubled, from US$23 billion to US$50 billion, between 1998 and 2005 as a result of widespread tax evasion and capital flight arising from the collapse of the political legitimacy of the state as perceived by taxpaying citizens under the divisive populist politics of Chavez’s Bolivarian Revolution (Di John 2009). Only if the state formulates an inclusive vision for eco- nomic growth and development that is politically credible with a broad base of taxpayers and holders of capital, as well as the wider electorate, does rational capital flight from governance risk become illicit capital flow, that is, an immoral undermining of the state’s commitment to pov- erty reduction, sustainable economic growth, and long-term political stability. Evasion into illicit outflows: Looting, rent-scraping, and dividend-collecting Illicit international capital transfers require three drivers: (1) opportu- nity, particularly facilitated by globalization; (2) pull, that is, attraction to more effective governance contexts with less political risk (flight); and (3) push, that is, the lack of a credible domestic commitment to develop- ment (flows). Wedeman (1997) distinguishes three political economies that convert tax evasion into capital flight and outflows: looting, or unin- hibited plundering or systematic theft of public funds and extraction of The Political Economy of Controlling Tax Evasion and Illicit Flows 75 bribes by public of�cials, whereby political insecurity is so endemic that outflows are institutionalized; rent-scraping, or political manipulation to produce rents, thereby allowing the scraping off of these rents by public of�cials so that short-term political gain drives outflows; and dividend- collecting, or transfers of a predictable percentage of the pro�ts earned by private enterprises to government of�cials. Under this last political con- text, the longer-term consensus on development generates the political security that is essential to constrain outflows. These different dynamics result in the wide variations in the effective- ness of tax systems. In a looting context such as Zaire under Mobutu, minimal taxation and endemic tax evasion reflected the failure to build the state capacities required to enforce property rights, and the political settlement collapsed. In a rent-scraping environment such as the Philip- pines under Marcos, the tax system was weak, and tax evasion was wide- spread and increasing; the ratio of taxes to GDP fell from 12.0 percent in 1975 to 9.6 percent in 1984 as political support for the dictatorship col- lapsed (Manasan and Querubin 1987). The �rst �nance minister of the post-Marcos Aquino government later noted that “every successful busi- nessman, lawyer, accountant, doctor, and dentist I know has some form of cash or assets which he began to squirrel abroad after Marcos declared martial law in 1972 and, in the process, frightened every Filipino who had anything to lose� (Boyce and Zarsky 1988, 191). A return to formal democracy did little to alleviate the problem, however, and, a decade later, the politics of vested interests in the Philippines continued as rent- scraping, in which “tax avoidance and evasion are evidently largely the province of the rich� (Devarajan and Hossain 1995). In these rent-seeking contexts, economic opportunities and political order are created and maintained by limiting access to valuable resources (North et al. 2007; North, Wallis, and Weingast 2009). Limited access to the looting of state assets, rent seeking, corruption, criminality, and ease of tax evasion creates the political incentives in any regime to cooperate with the group in power.14 Access to illicit capital flows and tolerance of tax evasion form part of this rent and are distributed to solve the problem of endemic violence and political disorder. The regime, lacking a broad elite coalition for long-term development, may tacitly or expressly allow the “winning coalition� of its power base to evade tax, while trying to constrain open access to the illicit capital flows and to tax evasion that would undermine 76 Draining Development the rent seeking on which the stability of the state in these fragile gover- nance contexts depends (North et al. 2007; North, Wallis, and Weingast 2009). Repressive governments may be more inclined to restrain tax eva- sion by simply avoiding the sources of tax that require a substantial degree of voluntary cooperation (Hettich and Winer 1999). By contrast, in dividend-collecting contexts, the political importance assigned to building an effective state has meant that evasion has trans- lated less into illicit outflows, and developmental states often emerge (Khan 2000). In a dividend-collecting system, economic growth and political stability are encouraged because political leaders and corrupt of�cials perceive the payoffs in the long term. This dynamic explains how developing countries often begin to build a national development project despite the initially low tax revenue that undermines their early capacity to deliver services to citizens, potentially reinforcing the lack of legitimacy and the low tax take. Corruption and tax evasion may be rampant not least so as to fund political stability, but illicit flows are recognized as threatening to sus- tainable development and are tackled (Khan 2005; North et al. 2007; North, Wallis, and Weingast 2009). This is in contrast to the short-term outlook of the looting model, which destroys economic value. So, while, in all these contexts, grand tax evasion arises from collusion among political and business elites and the top levels of government, the evasion– capital flows process is shaped by different political dynamics. Policies and enforcement may be formulated to permit tax evasion in return for political favors, and they may operate through the overt or tacit complic- ity of senior tax administrators and �scal policy of�cials. Without politi- cal support for the high-pro�le prosecution of tax evaders, grand eva- sion results in rampant illicit capital flows in looting or rent-scraping contexts, but the politics of dividend-collecting facilitates a pragmatic focus on controlling capital flows for domestic investment and political stability.15 Petty low-level tax evasion, by contrast, is probably insigni�cant for illicit capital flows, but affects the underlying political economy context. The average size of African shadow economies was estimated at 41.3 per- cent of GDP in 1999/2000, increasing to 43.2 percent in 2002/03), though these �gures hide signi�cant grand evasion because the lack of commit- ment of elites to long-term development morphs into capital flight and The Political Economy of Controlling Tax Evasion and Illicit Flows 77 flows (Schneider, A. 2007; UNECA 2009). The dynamics of petty evasion also have political implications, however, because of what Tendler (2002) refers to as the “devil’s deal,� the tacit understanding in patronage poli- tics that petty tax evasion (evasion that does not lead to signi�cant illicit flows) is tolerated as the quid pro quo for political support that allows grand evasion and illicit flows: “if you vote for me . . . , I won’t collect taxes from you; I won’t make you comply with other tax, environmental or labor regulations; and I will keep the police and inspectors from harassing you� (Tendler 2002, 28). This leaves both vested interests and the informal economy untaxed or undertaxed and deprives the infor- mal sector of the incentive to engage in constructive political bargaining over taxation, the �scal social contract.16 At the same time, rulers are left unrestrained to embezzle and misuse public resources to sustain their grip on power, while the public goods and services necessary for poverty reduction and development are underprovided. This devil’s deal underpins the neo-patrimonial politics behind illicit capital flows. Political clients do not act as taxpaying citizens using their votes to improve governance, and political patrons exploit corruption and tax evasion to fund their grip on power, while protecting their long- term prospects outside the country. So, political dynamics shape the capacity of the state to manage evasion and constrain the incentives for political and economic elites to shift the tax burden to middle-size �rms and the middle class.17 The Political Economy of Tax Evasion The dynamics outlined above indicate that the political economic deter- minants of the extent of tax evasion in any speci�c context reflect a com- bination of six factors, as follows: (1) the structure of the economy that shapes the potential sources of government revenues; (2) the structure of the political system, that is, the political rules by which politicians gain and hold onto the power that shapes political competition and incentives (constitutions, party structures, electoral processes, legislative and executive rules shaping policy choice, and so on) (Stewart 2007; Schneider, F. 2007); (3) the state’s monopoly of violence as the basis for enforcing state tax and other authority (North et al. 2007; North, Wallis, and Weingast 2009); (4) the credibility of the regime’s political time 78 Draining Development horizons that shape the perceptions of tax as an investment in the future prosperity and stability of the state; (5) the potential influence of collec- tive action by taxpayers or other citizen groups to negotiate with politi- cal leadership; and (6) the rationale for leaders and elites to build the broad, long-term economic development of their societies, which is influenced by noneconomic interests such as traditional ethnic, kinship, and communal ties and patterns of patronage, as well as other factors such as international pressures.18 These six political economy factors, in turn, shape citizen under- standing of the legitimacy of taxation.19 Taxpayer attitudes toward this legitimacy are formed by three more immediate forces, as follows: • Tax morale: taxpayer civic responsibility to support or undermine national goals, that is, the extent to which enforcement is accepted as a legitimate and effective exercise of state power • Capacity: the capacity of tax policy and particularly tax administra- tion, including the effectiveness of political and administrative checks that limit rent seeking and patronage and the political support for the ef�ciency of the tax administration, measured by tax effort and the tax compliance costs imposed on taxpayers compared across coun- tries at similar levels of development20 • Perceived fairness: the perceived fairness and effectiveness of taxes and the tax authority (based both on experience and the perception of the extent of the compliance of other taxpayers)21 Tax evasion flourishes in poor governance contexts In light of the above, it is not surprising to �nd that tax evasion is endemic in contexts of poor governance, for example: • In the Philippines in the mid-1980s, with the collapsing Marcos regime, income tax evasion grew to account for nearly 50 percent of the potential yield (Manasan 1988). The restoration of democracy under President Aquino was only partly successful in delivering polit- ical stability, a weak legitimacy reflected in corporate tax evasion on domestic sales tax at between 53 and 63 percent in the early 1990s • (Abinales and Amoroso 2005). uncertainty was reflected in the federal In Nigeria, persistent political government estimation in 2004 that, because of tax evasion, corrup- tion, and weak administration, it had collected only around 10 per- The Political Economy of Controlling Tax Evasion and Illicit Flows 79 cent of the taxes due and that half the revenue collected was then lost or embezzled (OECD and AfDB 2008). • In the Central African Republic, ongoing political tensions have facil- itated evasion and kept the ratio of taxes to GDP at only 8 percent (World Bank 2009a).22 Tax evasion is, however, not static, but fluctuates. A primary variation is the perceived strength of state legitimacy in developing countries, which affects evasion directly through tax morale and indirectly through a lower tax effort (Fjeldstad, Katera, and Ngalewa 2009). This has also been demonstrated in the developed world, in Italy, for example, in the increases in the evasion of the value added tax (VAT) during periods of political uncertainty such as the peak, at 37 percent in the late 1980s, or, in Spain, when shifting political pressures have weakened the adminis- trative effort devoted to tackling tax evasion (Chiarini, Marzano, and Schneider 2009; Esteller-Moré 2005). Nonetheless, the political roots of the tax evasion driving interna- tional capital transfers affect particularly the developing world. In the international development discourse, improving tax systems is too often portrayed as only a technical, apolitical challenge, rather than as a pro- found litmus test of state legitimacy. Public choice theory’s conceptual- ization of politics as merely self-interest is unfortunately reflected too uncritically in the tax literature.23 So, it is worth citing the rich evidence showing that tax systems everywhere reflect shifting political legitimacy, as follows: • In Benin by the late 1990s, during the country’s turbulent post–Cold War transition from a Marxist dictatorship, only 10 percent of tax- payers were regularly paying their taxes (van de Walle 2001). • In The Gambia, in the years immediately before the military coup of 1994, lost revenues rose to 9 percent of GDP, and income tax evasion expanded to 70 percent of the total revenue due (Dia 1993, 1996). • In Kenya by 2001, the chaotic last year of the Moi presidency, the tax gap had reached at least 35 percent, suggesting signi�cant and increas- ing evasion (KIPPRA 2004).24 • In Madagascar, the political turmoil of the late 1990s saw the tax eva- sion rate rise to nearly 60 percent, equivalent to 8.8 percent of GDP 80 Draining Development (compared with only 8.3 percent of GDP collected in taxes) (Gray 2001). • In Niger, the political chaos of the 1990s saw rampant tax evasion trigger a major decline in �scal revenues (Ndulu et al. 2008). • In Pakistan, as the legitimacy of the state collapsed, the size of the informal economy increased from 20 percent of GDP in 1973 to 51 percent in 1995, while tax evasion more than trebled over the same period and continued thereafter within the climate of continuing political turmoil and military rule; the tax-GDP ratio fell from 13.2 percent in 1998 to 10.6 percent in 2006 (Saeed 1996). • In Paraguay, the political optimism and improved tax compliance after the end of the Stroessner dictatorship had faded by 2003, when increasing tax evasion and corruption were depriving the state of about two-thirds of potential revenues (Freedom House 2008).25 • In West Bank and Gaza during the period 1994 to 2000, the declining perceived legitimacy of the Palestinian Authority reflected an appar- ently increasing tolerance of tax administrative procedures that facili- tated evasion and corruption (Fjeldstad and al-Zagha 2004). The political foundations of tax systems are apparent in the contrast between states in the Caribbean and states in Latin America given that Caribbean countries inherited parliamentary institutions rather than the presidential regimes of Latin America. This constitutional difference apparently created a broader acceptance of the need for an effective tax system across the Caribbean because parliamentary systems offer the possibility of negotiating a compromise among elites in forming a gov- ernment. By contrast, Latin American political structures created a vicious circle of political instability and more regressive tax structures (Schneider, Lledo, and Moore 2004). The collapse of the Soviet Union and the resulting political uncertain- ties in the transition countries over the progress toward representation through taxation underline the political foundations of tax evasion (Gehlbach 2008).26 For instance, in Armenia, evasion became endemic as the tax-GDP ratio halved from 29 percent in 1991 to 14.8 percent in 1994. Although it recovered to 17 percent in 1997, the tax-GDP ratio remained well below the level in other transition economies that had a more secure political consensus on national purpose and that had not The Political Economy of Controlling Tax Evasion and Illicit Flows 81 faced such a dramatic loss of legitimacy (for example, tax-GDP ratios in Estonia of 39 percent or in Poland of 44 percent in 1997) (Mkrtchyan 2001). The World Bank, in 2009, concluded that “fundamental political economy dif�culties such as . . . checks to the power of powerful business interests in evading customs and tax payments . . . remain to be seriously addressed� (World Bank 2009c, 4). Armenia, however, highlights the need for further research on the complex international dynamics of illicit capital flight and flows because, despite political and economic fra- gility, it appears, in recent years, to have had sizable unrecorded capital inflows that may reflect other political economy dynamics in the region (Brada, Kutan, and Vukšic 2009). Natural experiments in tax evasion and illicit flows Cross-country comparisons over time illustrate the political dynamics behind state legitimacy and the ef�cacy of tax systems. In the Central America of the 1950s, Costa Rica and Guatemala shared similar charac- teristics in geographical location, size, colonial history, position in the world economy, levels of economic development, and reliance on coffee exports, but showed different patterns of illicit capital flows (Fatehi 1994). The Costa Rican political settlement following a brief civil war in 1948 successfully consolidated democracy and delivered economic growth. Taxpayers recognized that making the state effective required an adequate tax base to fund public services. As a result, the political settle- ment established the political will to tackle tax evasion and minimize illicit capital outflows. The tax compliance and collection rates in Costa Rica are the highest in Central America, and indicate the least evidence of an inclination to capital flight (Torgler 2003). By contrast, Guatemala has the lowest tax take in Central America and has a long history of widespread tax evasion and capital flight (Erbe 1985). The failure of the political class in Guatemala to develop a wider vision for society resulted in civil war in the 1970s and 1980s. Political bargaining during the drafting of the Constitution of 1985 resulted in three articles being inserted deliberately to weaken the tax authority; as a result, tax evasion in the early 1990s was 58 percent.27 The peace accord of 1996 has yet to generate any sense of national purpose so that, in the later half of the 1990s, for every quetzal taxed, 65 centavos were evaded, leaving the Guatemalan state without the resources to improve its legitimacy by 82 Draining Development providing better services for the population (Yashar 1997). The legisla- ture has often appeared to facilitate evasion and avoidance deliberately by making tax laws overly complex. Without political support to tackle evasion by powerful interests, the revenue authorities have instead directed their anti–tax evasion effort toward small taxpayers. The probability of evaders being caught has been low; in 2003, the Guatemalan revenue administration estimated total tax evasion at more than two-thirds of actual collection, with sig- ni�cant variation across revenue bases, at 29 percent for VAT, but 63 per- cent for income taxes.28 If tax evasion is detected, the administrative and legal procedures are complicated and costly for the state. Penalties for tax evasion remain insigni�cant because of the absence of political determi- nation to impose administrative and judicial sanctions in any effective or timely manner, and tax amnesties are too frequent and generous (Sánchez, O. 2009). The total tax-GDP ratio rose from a low of 6.8 per- cent when the civil war ended to 10.3 percent of GDP in 2006; citizens still widely regard tax evasion in Guatemala as ethical because the per- ception remains that “a signi�cant portion of the money collected winds up in the pockets of corrupt politicians or their families and friends� (McGee and Lingle 2005, 489).29 Another natural experiment in the links between illicit outflows and the legitimacy of taxation can be seen in the contrast between Argentina and Chile, despite similar cultures, levels of economic development, his- tory, and federal constitutions. Tax evasion in Argentina has consistently been high, the result of political instability and weak institutions (de Melo 2007). Sporadic efforts to crack down on evasion, including by the Menem presidency in the early 1990s, lacked political sustainability.30 By contrast, in the early 20th century, Chile achieved a political con- sensus against tolerating tax evasion and in favor of alleviating poverty and advancing national development through an effective tax system (Bergman 2009). With Chile’s return to democracy in 1990, taxpayers were persuaded through the Concertación (the Concert of Parties for Democracy) that higher taxation was a small price to pay if democracy was to work by delivering an effective social contract (Boylan 1996; Vihanto 2000; Bergman 2009). Income tax evasion rates in Chile fell from 61.4 percent in 1990 to 42.6 percent in 1995 (Barra and Jarratt 1998). This �scal social contract covered indirect and direct taxation, so The Political Economy of Controlling Tax Evasion and Illicit Flows 83 that Chile, with a rate of 18 percent, collected almost 9 percent of GDP in tax revenue. By another contrast, Mexico, with a rate of 15 percent, collected less than 3 percent of GDP (Tanzi 2000). The mix of political will and administrative capacity was then reflected in overall evasion rates that nearly halved in Chile during the early 1990s, from 30 to 18 percent during the euphoria of the return to civilian rule; as the eupho- ria wore off, evasion had increased again to 25 percent by 1998, but, sub- sequently, because of rising political concern, matched by administrative effort, declined steadily to under 15 percent in 2004 (World Bank 2006). As a result, capital flight and illicit capital flows have been consistently lower in Chile than in neighboring Argentina (Schneider, B. 2001; Di John 2006). Politics and institutions, not geography or economic struc- ture, drive rates of capital flight. Political legitimacy and illicit capital flows Illicit capital flows arising from tax evasion indicate not only the oppor- tunity for free-riding, but also taxpayer perceptions. If the state does not appear to be aligned with taxpayer interests, this lack of faith in a regime’s current legitimacy and its future legitimacy after improving governance shapes the decision to hold assets abroad. However, if taxpaying elites believe in a locally owned political vision for the future, and regimes need their support, evasion may be tolerated by the state, yet not lead to illicit flows. As economic theory suggests, information asymmetries and higher rates of return in capital-scarce developing countries would retain domestic investment rather than encourage capital flow from the developing to the developed world.31 But, if citizens see no credible long- term future political stability, savers and investors will seek to move their capital abroad to contexts of more effective, more stable governance, even at lower returns on investment and higher tax rates.32 Where the state lacks political legitimacy, this capital flight, whether legal or not, would, as suggested above, be widely perceived as licit. Effective tax systems alone are therefore not enough. Bergman (2009, 109) concludes that “moderate ef�ciency in law-abiding societies gener- ates better results than good administration in a world of cheaters.� For the state, the political imperative in curtailing free-riding on taxes is that, while core state functions such as preserving peace and upholding jus- tice are universal expectations, transparent and fair property rights are 84 Draining Development essential to the dividend-collecting transition to effective, politically open access states and open markets (North et al. 2007; North, Wallis, and Weingast 2009). The creation and enforcement of property rights are, however, not costless (Khan 2005). Historically, the need to raise revenue motivated states to de�ne property rights. In turn, this stimu- lated economic growth (Tilly 1975). Taxpayers are politically prepared to pay taxes that fund the state capabilities needed to secure their property rights (Khan 2000). In developing polities, the allocation and de�nition of property rights are continuously evolving, as they did historically in developed countries (Everest-Phillips 2008b; Hoppit 2011). Tackling tax evasion and illicit flows is therefore part of the political bargaining over how and what services, including protection of property rights, the state will provide. If taxpayers do not accept the legitimacy of the state and are not intrinsically prepared to fund the state’s services and the development of their own collective future prosperity, the political incentives for and credibility of the commitment to long-term sustainable development are lacking. Widespread illicit capital flows occur not only because the state is ineffectual, but also because citizens believe that the state has neither the political leadership nor the consensus needed to reform weak or cor- rupt institutions that threaten peace, justice, and the long-term security of private property rights (Friedman et al. 2001). As van de Walle (2001, 53) notes, the root of the problem is “the political logic of a system in which the authority of the state is diverted to enhance private power rather than the public domain.� Despite the democratization that began in the 1990s in many devel- oping countries, the entrenchment of vested interests has ensured that improvements have been modest.33 Improvements in the ratio of tax rev- enue to GDP have also often been short-lived because technical changes dependent on cooperative elites and high tax morale have been over- come by underlying political forces (Orviska and Hudson 2003; Engel- schalk and Nerré 2002). So, rampant tax evasion remains widespread in many developing countries despite the seemingly endless tax policy and administration reforms undertaken over the intervening two decades. Tackling illicit capital flows therefore requires not merely more tight- ening of tax policy and tax administration, but a shift in citizen attitudes The Political Economy of Controlling Tax Evasion and Illicit Flows 85 Figure 3.1. The Political Economy of an Effective State and Tax System Tax system Governance quality Political settlement Source: Author compilation. toward the political institutions shaping the governance context. These operate on three layers, as follows (see �gure 3.1): 1. The political settlement: the informal and formal rules of the game over political power.34 This includes both the formal rules for manag- ing politics (peace treaties, constitutional provisions, parliamentary rules, the organization of political parties, and so on) and the infor- mal networks, family ties, and social norms that shape political behaviors.35 More an iterative process than a critical juncture, the political settlement shapes tax morale as an expression of social order, a collective interest, and a sense of political commitment to improve public institutions, constrain corruption, and deliver public goods (Andreoni, Erard, and Feinstein 1998). The settlement develops from the national political community that delivers the perceived fairness of and trust in a credible state; within this context, the tax system represents a legitimate and effective exercise of state power, while the ef�cient provision of public goods bene�ts taxpayers and ordinary citizens (Lieberman 2003). Politics shapes attitudes and behaviors relative to tax evasion, and the differences across groups within a country are potentially as signi�cant as the differences across coun- tries (Gërxhani and Schram 2002). 2. General quality of governance: state institutional effectiveness builds on the political settlement. Capital flight (rather than capital flows seeking only higher rates of return) moves from weak to effective governance, 86 Draining Development from the generally lower tax contexts of developing countries to the higher tax environment of the developed world. Given overall lower rates of return on such investment, the main rationale would seem to be flight because of a lack of faith in domestic rates of return that derives from a lack of genuine elite commitment to national develop- ment, the political risk arising from political instability, and rampant corruption (Fuest and Riedel 2009). This appears to be con�rmed by estimates of the capital flight from the African continent of about US$400 billion between 1970 and 2005; around US$13 billion per year left the continent between 1991 and 2004, or a huge 7.6 percent of annual GDP (Eurodad 2008). So, in contexts of poor governance, the wide overlap between capital flight and capital flows is reflected in the frequent conflation of the two terms. 3. Tax administrative competence: capacity to contain free-riding and maintain a credible commitment to paying taxes. This is critical to compliance; the state’s legitimacy is reflected in the tax authority’s powers of enforcement (that is, accountable coercion).36 In many developing countries, political influence is often pronounced in the selection and promotion of staff at all levels in the revenue adminis- tration, thereby facilitating evasion and corruption such as the case of Benin in the late 1990s, when 25 percent of revenues were lost (van de Walle 2001).37 Tax evasion is also signi�cantly influenced by the inter- actions between taxpayers and the tax authorities: more respectful behavior on the part of tax authorities is associated with higher rates of tax compliance (Feld and Frey 2002). At the same time, differences in treatment by tax authorities arise from differences in the quality of political participation: “regime type matters in the explanation of the structure of taxation� (Kenny and Winer 2006, 183). Variations in political governance away from political settlements and the quality of the institutions shaping administrative competence there- fore lie at the heart of varying patterns in the tax evasion dimension of illicit capital flows.38 Citizen perceptions of the legitimacy of taxation and of the effectiveness of the state combine in the calculation citizens make of the tax cost/bene�t–burden/contribution equation. Whether taxation is viewed as a value-for-money contribution to society or sim- ply an unproductive burden depends also on how it is raised and used.39 The Political Economy of Controlling Tax Evasion and Illicit Flows 87 Figure 3.2. The Political Economy Model of Tax Evasion Driving Illicit Capital Flows Illicit capital flow International political economy Opportunity Citizen Tax State Domestic perceptions evasion effectiveness political (see figure 3.1) economy Source: Author compilation. Alm and Martinez-Vazquez (2003, 158) �nd that government can decrease evasion “by providing goods that their citizens prefer more, by providing these goods in a more ef�cient manner, or by more effectively emphasizing that taxes are necessary for receipt of government services.� Alm and Martinez-Vazquez (2003, 163) conclude that “compliance is strongly affected by the strength and commitment to the social norm of compliance.� This is unlikely, for instance, in Malawi, where, in 2000, the government admitted that a third of public revenues were stolen annu- ally by civil servants.40 At the other end of the governance spectrum, countries with a strong political settlement and an effective administra- tion, as in Scandinavia, witness the coexistence of large nominal and effective tax rates with low levels of evasion.41 In effective governance contexts, therefore, taxpayers do perceive taxation as less a burden and more of an investment in an effective state.42 Variation in state effectiveness and, therefore, in the opportunity to turn evasion into illicit capital flows is the meeting point of the domestic and the international political economy (�gure 3.2). Efforts to stem capital flight date back to the League of Nations before World War II. The political will to tackle illicit capital flows still requires the combination of effective national commitment to development and international coordination. This has been lacking, because, internationally and domestically, “[tax] is predominantly a matter of political power� (Kaldor 1963, 418). 88 Draining Development Controlling illicit capital flows: Governance beyond technical solutions “Revenue is the chief preoccupation of the state. Nay, more, it is the state.� —Edmund Burke, quoted in O’Brien (2001, 21) The last decade or so has demonstrated that improving governance in developing countries is not simply a technical task of putting capacities in place, but is a problem of political will shaped by institutions and incentives. As a result, the development challenge has evolved from good governance through good-enough governance to engaging with the spe- ci�c political drivers of change (Fritz, Kaiser, and Levy 2009). Adminis- trative and policy reforms alone cannot alter the rampant corruption and tax evasion typical of weak states into the strong compliance of citizen-taxpayers in effective liberal democracies. Tax evasion and cor- ruption translate into capital flight and illicit flows not only if the oppor- tunity arises to place assets abroad. Capital flight arises from political risk; illicit flows arise from a perceived lack of political consensus about building an effective state. “A tax regime is conditioned by the power bal- ances and struggles among the major social groups in a country, includ- ing the relative strengths of political parties that represent their diver- gent interests� (Di John 2008, 1). So, as Ndikumana (2004, 290) points out: “Tax evasion is made easier when the taxpayer is also the political client for the decision-maker.� Such patrimonial politics, so common in many developing countries, thrives on evasion, corruption, and crimi- nality (Chabal and Daloz 1999). Illicit capital flows should, as a consequence, be addressed in the con- text of the political economy of inequality. Nontaxation and wide-scale tax evasion by the rich in high-inequality contexts undermine the capa- bilities and perceived legitimacy of the state (World Bank 2008a). In many developing countries, extreme inequalities create potential politi- cal instability, which, if combined with the lack of a social contract to pay taxes for the common good, fosters illicit capital flows and explains why achieving genuine national ownership of development often proves so problematic. In such environments, tax system complexity can replace brutal repression as the obfuscation of political choice by elites in a situation in which tax morale is weak. An extreme example is offered by Honduras in The Political Economy of Controlling Tax Evasion and Illicit Flows 89 1991. Through legal exemptions alone, a taxpayer there would have required an income 687 times the average per capita income before being liable at the highest marginal tax rate. Given the rampant evasion, the rich in Honduras were, de facto, untaxed. The state consequently lacked broad legitimacy; development stalled; and political instability remained endemic.43 These dynamics were also evident in Madagascar, for example, where the tax-GDP ratio was only 11.8 percent in the early 1990s. This was the outcome of political interest in permitting the tax system to be too complex for the country’s administrative capabilities so as to facilitate widespread evasion for the privileged elites of the regime. This had a disastrous impact on economic growth and job creation; trade taxes discriminated across markets, while pro�ts and wage taxes discouraged investment and employment (IMF 2007). Elites, in other words, often maintain their power base in highly unequal societies by deliberately making tax systems overly complex, at huge cost to long-term development, not least because this generates illicit capital flows. Successful development, however, such as has occurred in East Asia, is often accompanied by the political imperative to deliver development in the face of internal and external threats, com- bined with tight capital controls and, thereby, the absence of global opportunities to hold illicit assets abroad. As globalization over the last few decades has created unparalleled opportunity to move capital easily abroad, the political imperative for leaders and elites to deliver national development appears to have weakened. The democratization of the 1990s does not seem to have signi�cantly altered the incentives for cor- ruption and tax evasion. Indeed, the contemporaneous opportunities that opened up with globalization often allowed the political class to accept political reforms, knowing that they could protect their assets by legally, illegally, or illicitly transferring them abroad. This has been man- ifested in the use of capital flight in response to populist politics such as in Bolivia and República Bolivariana de Venezuela and of illicit flows from many developing countries in the absence of a political commit- ment to sustainable development. If controlling illicit flows is inherently a political problem, tax evasion is particularly important because tax structures are among the most measurable manifestations of the political settlement central to state effectiveness. Yet, in many developing countries, particularly the low- 90 Draining Development income ones, the current political dynamics create little incentive to make the necessary improvements in governance (Bratton and van de Walle 1997; Chabal and Daloz 1999). Tax evasion flourishes if an incom- plete political settlement excludes many from the political system. The politically excluded, in turn, exclude themselves from paying taxes, while the lack of political accountability empowers unaccountable elites to exploit their position both to evade and, in the absence of faith in long- term political stability built on legitimacy, to move evaded assets abroad. A weak, corrupt, and illegitimate state drives political flight as citizens disengage from the state, encouraging more illicit flows to escape from the state’s grasp. Such dynamics explain why countries often achieve an uneasy equilibrium in their �scal systems that reflects the balance of political forces and institutions and remain at this position until shocked into a new equilibrium (Bird, Martinez-Vazquez, and Torgler 2006). Without understanding the underlying political governance determinants of this equilibrium, reforms are unlikely to solve technical problems in the tax system, such as those identi�ed by Le, Moreno-Dodson, and Rojchai- chaninthorn (2008, 16): “tax policies riddled with overly complex struc- tures and multiple—largely ad hoc—incentives that narrow the already limited tax base, create more loopholes for tax avoidance and evasion, intensify the public perception of unfairness of taxes, and generate opportunity for corruption.� As a result, tackling tax evasion is a problem of collective action: funding adequate public goods depends on the discount rates or time horizons of rulers and taxpayers, on the transaction costs involved, and on relative bargaining power (Paul, Miller, and Paul 2006). Intra- and cross-country variation in tax effort and tax-GDP ratios indicates dif- ferences not only in the economy or in levels of development or in regime type, but also in the political will to tackle the collective action problem (Bird and Zolt 2007; Cheibub 1998).44 This applies in devel- oped- and developing-country contexts.45 So, although low-tax coun- tries tend also to have lower income per capita, there are signi�cant variations in tax effort and tax-GDP ratios. Zambia, for example, had a per capita income of US$785 in 2003 and collected 18.1 percent of GDP in tax, whereas Uganda had a per capita income of US$1,167 but col- lected only 11.4 percent of GDP. Indonesia, with a gross national prod- The Political Economy of Controlling Tax Evasion and Illicit Flows 91 uct per capita of US$668, collected 16.5 percent, while Lesotho, with a slightly lower per capita gross national product (US$624), collected more than twice as much, at 36.8 percent of GDP (Teera and Hudson 2004).46 Mozambique’s tax revenue, at 13.4 percent of GDP in 2006, rep- resents a remarkable recovery from civil war, but remains well below that of comparable low-income countries in the region, such as Malawi (18 percent) and Zambia (18 percent).47 Policy Implication: Strengthen the Political Commitment to Effective States Effective tax regimes develop when political leaders need effective tax systems. The necessary political commitment to tackle evasion arises when the political costs of the tax effort are outweighed by the political bene�ts of delivering sustainable development, including by providing the state with the capacity to enforce the property rights of political elites, thereby helping facilitate the emergence of long-term political sta- bility and economic growth. The political incentives around taxes, growth, and property rights are central to this dynamic. An effective tax system offers a general investment in the public goods that the economy and society need. Concepts such as citizenship, fairness, trust, and equal- ity develop practical manifestation through taxation: “the tax system is therefore an effective way of articulating assumptions about the market, consumption and social structures� (Daunton 2001, 21). Because political ideas, interests, incentives, and identities reflect and are reflected through the tax system, the evolution of tax structures demonstrates a state’s political and economic levels of development (Thies 2004; Fjeldstad and Moore 2009). For example, both Finland and Sierra Leone collected 31 percent of total revenue in generally progres- sive taxation on income and property in 1990; yet, Finland was able to collect more than seven times the revenue take, as a share of GDP, than Sierra Leone (Lieberman 2002). Evasion may express politically passive rather than actively subversive attitudes toward government, particu- larly where, in the Chinese phrase, “the Mountain is high and the emperor far away.� Focused mainly on local taxes and fees, Chinese vil- lagers, in the 1980s, developed a moral politics of evasion (taoshui) and tax resistance (kangshui) around the perceived extent of reciprocity and 92 Draining Development mutual obligation with different levels of government. Central to the moral legitimization of evasion was the extent to which the authorities were perceived to be delivering on their responsibilities (zeren), from which people then classi�ed taxes as either reasonable (heli de shui) or unreasonable (buheli de shui) (Ku 2003). The quality of political institutions therefore has a strong observable effect on tax morale (Torgler et al. 2008; Everest-Phillips 2008a, 2009b). Tax morale, the practical expression of support for national develop- ment, as noted above, is particularly signi�cant in developing countries with weak governance quality: citizens across the developing world invariably justify tax evasion through the belief that they do not get an adequate return for the taxes paid.48 For instance, the U.K. Department for International Development–World Bank survey in the Republic of Yemen of private sector tax morale found that 91 percent of respondents felt that the state had to be seen to be using revenues fairly and ef�ciently if tax evasion was to be tackled effectively; 97 percent felt they paid a fair amount of tax; and 100 percent felt that tackling corruption was essen- tial (World Bank 2008b; see also Aljaaidi, Manaf, and Karlinsky 2011). Modern democracies aspire to achieve voluntary tax compliance, that is, all citizens willingly comply without the state needing to resort to any compulsion. In reality, all tax systems settle for quasi-voluntary tax com- pliance, recognizing that the free-riding potential always requires some state authority to compel. Levi (1989, 69) categorizes quasi-voluntary compliance as the outcome of “the sanctions, incentives and reciprocity practices that produce social order and conditional cooperation with- out central state- or ruler-imposed coercion.� Yet, the overwhelming focus on tackling tax evasion in developing countries is on �scal policy and administrative capacity constraints, not on the political governance that makes tax policy and administration effective (Sánchez, O. 2006; Sánchez, J.A. 2008). Policy Implication: Tackle Illicit Capital Flows by Strengthening the Legitimacy of the State So, tackling the tax evasion dimension of illicit capital flows requires that one address the underlying political economy constraints to the long- The Political Economy of Controlling Tax Evasion and Illicit Flows 93 term building of an effective state (Fjeldstad and Moore 2009). Tax is conditional on political performance: The quality of political institutions has a strong observable effect on tax morale . . . not only the overall index, but also the sub-factors of voice and accountability, rule of law, political stability and absence of violence, reg- ulatory quality and control of corruption exercise a strong influence on tax morale. Moreover, trust in the justice system and the parliament also has a highly signi�cant positive effect on tax morale. (Torgler, Schaffner, and Macintyre 2007, 23) In other words, every political system constantly struggles to main- tain its �scal legitimacy: The achievement of signi�cant quasi-voluntary compliance within a pop- ulation is always tenuous . . . free riding, once begun, is likely to increase. Once quasi-voluntary compliance has declined, it is extremely dif�cult to reconstitute. Its reestablishment often requires an extraordinary event— such as war, revolution, or depression—that makes people willing to negotiate a new bargain. (Levi 1989, 69–70) So, tax evasion and the resulting illicit capital flows must be tackled by strengthening the legitimacy of the state in the eyes of current and potential taxpayers. There is evidence in some developing countries, such as Ghana, Indonesia, and Zambia, that, through the messy process of democratic politics, this is happening (Pritchard 2009; Heij 2007; von Soest 2006). Conclusion Further research is undoubtedly needed to understand the relative importance and interconnection of tax evasion, corruption, and crimi- nality in driving illicit capital flows in different contexts, but it is clear that tackling the problem will require building effective states responsive to the needs of their societies. Currently, there does not appear to be reli- able data on the level of illicit capital that flows South-South or South- North. What we do know is that the governance reforms needed to build effective states and to reduce illicit capital outflows require strong politi- cal commitment. Such commitment develops if political institutional 94 Draining Development incentives and interests are aligned with the goal of building state capaci- ties. This will only occur in developing countries if a broad political con- sensus in its favor is created. Administrative efforts to tackle illicit capital flows will only be as effective as the constantly evolving political process of state building allows, because “the history of state revenue production is the history of the evolution of the state� (Levi 1989, 1). Notes 1. “Responsible Leadership for a Sustainable Future,� G-8 Leaders Declaration, paragraph 17d, L’Aquila, Italy, July 8, 2009, http://www.g8italia2009.it/static/ G8_Allegato/G8_Declaration_08_07_09_�nal%2c0.pdf. 2. Recognition of the relationship between tax evasion and effective governance is not new. In the mid-�fth century, the theologian Salvian the Presbyter of Mar- seilles, seeking to make sense of the decline and fall of the Roman Empire that he witnessed around him, concluded that the empire’s collapse lay in the inter- connection between poor governance and a weak tax system. Modern research validates Salvian’s speci�c analysis and general thesis: tax evasion was both cause and symptom of Rome’s decline: for example, see Wickham (2010). 3. On the �scal anthropology of power and reciprocal obligations, see Braithwaite (2003). By contrast, tax avoidance, the legal minimization of tax payments, indi- cates weak policy and ineffectual administration. 4. For example, on the case of Bolivia in the late 1990s, see Mosley (2007); on the long and continuing tradition of Chinese peasant tax evasion as revolt, see Tilly (2003); on Ghana’s value added tax (VAT) riots in 1995, see Pritchard (2009). 5. This involved not merely the evasion of direct taxation; for VAT evasion: see Moreno-Dodson and Wodon (2008). For 1998, the estimated level of evasion was 37.5 percent; in May 2000, the Procuradoría Fiscal de la Federación announced that the overall level of evasion and avoidance represented 35 per- cent of potential collection. 6. Tax evasion, corruption, and criminality are all illegal activities that flourish in domestic and international climates in which personal greed is unrestrained by the political ambition to build legitimate, effective states. Nonetheless, there does not appear to be a single study that has tried to distinguish even conceptu- ally between them, let alone provide clear evidence of their relative importance in different contexts. This chapter is a beginning in addressing the �rst issue. That much remains to be done is suggested by the �ndings of Reuter and Tru- man (2004), according to which tax evasion accounted for over 55 percent of criminal proceeds in the United States in 1990. 7. Tax morale is conceptually distinct from the instrumental willingness based on government performance. The Political Economy of Controlling Tax Evasion and Illicit Flows 95 8. Silvani and Baer (1997) offer a typology for the tax gap, ranging from countries with a tax gap at under 10 percent (for example, Denmark or Singapore), which reflects high political will and strong administration, to countries with a tax gap at over 40 percent (for example, Kenya), which reflects limited political will and weak tax administration. 9. The Organisation for Economic Co-operation and Development offers a de�ni- tion of state building, as follows: “purposeful action to develop the capacity, institutions and legitimacy of the state in relation to an effective political process for negotiating the mutual demands between state and societal groups� (OECD 2008, 14). 10. Musgrave (1996) offers four basic patterns: the service delivery state, the welfare state, the communitarian state, and the government failure state (rather than the market failure state). 11. Legitimacy is a signi�cant component of economic growth in developing- country contexts; see Englebert (2000). 12. As a percentage of actual total tax revenue and of potential tax revenue, tax eva- sion declined between 1972 and 1974, thereafter rose rapidly, and was about 60 percent of actual tax revenue and 37 percent of potential tax revenue in 1990, representing sevenfold and fourfold increases, respectively. See Chipeta (2002); Therkildsen (2001). 13. In 1999, only 3.4 percent of the Russian population thought they could trust the state (Oleinik 1999). 14. It has been calculated that, in India in the 1980s, up to 45 percent of gross national product was produced by distortionary government policies (for example, arti�- cial price setting, the protection of state-owned enterprises, and government monopolies) (Mohammad and Whalley 1984). 15. In Bangladesh, for example, criminal prosecution for grand evasion is almost unknown; in India, tax evasion, estimated at from Rs 400 billion to Rs 1 trillion, is at the heart of the informal economy, in which politically influential indus- tries operate (Vittal 1999). 16. Schneider and Enste (2000) outline the political determinants of tax evasion and the informal sector. 17. On the missing middle, or the absence of medium-size �rms as a result of dis- proportionate tax burdens, see Kauffmann (2005). 18. Wa Wamwere (2003, 176) describes this as follows: “The cream of government service goes to the ruling ethnic elites, the crumbs to the lesser ethnic elites, and dust to members of the so-called ruling ethnic community.� See also Kimenyi (2006). 19. For taxation, four political dimensions of the state then shape the effectiveness of tax systems, as follows: (a) the degree of concentration of state power (for example, a unitary or federal system), (b) the state’s reach into society (for example, the capacity to collect information and attain adequate territorial 96 Draining Development coverage), (c) the autonomy of the state (the degree of capture, isolation, or embedded autonomy of of�cials); and (d) the responsiveness of the state to the population (that is, democratic dialogue). See Hobson (1997). 20. Virmani (1987) models corruption that creates higher rates of evasion. The political economy of tackling corruption in the tax administration also involves corrupt tax policy formulation, especially in tax expenditures, through the dis- cretionary powers of politicians and tax administrators. There is an interesting provision in the Lesotho Revenue Authority Act that requires the tax authority to report on the impact of any concessions or waivers granted during the course of the reporting year. Thus, if ministers wish to grant a concession to a speci�c investor, they may do so, but the nature and the cost of the decision will be reported in the public domain (Charlie Jenkins, former commissioner of the Lesotho Revenue Authority, personal communication, 2009). 21. Thus, the Zambia Revenue Authority has declared that it is committed to the objective of fairness, de�ned as performing of�cial duties in an impartial man- ner, free of political, personal, or other bias; this combines with its mission to maximize and sustain revenue collections through integrated, ef�cient, cost- effective, and transparent systems that are professionally managed to meet the expectations of all stakeholders. See, for example, the Zambia Revenue Author- ity’s “Tax Payer Charter,� at http://www.zra.org.zm/charter.php. 22. Decreases in the tax-GDP ratio can also arise from other factors, such as an increase in GDP or in tax avoidance, but the latter reflects weak administration and taxpayer doubt over policy. 23. For instance, “when tax reform enters the political arena, the subtleties of the key issues are usually lost in the midst of self-serving arguments and misleading simpli�cations� (Slemrod and Bakija 1996, ix). 24. The Kenyan Revenue Authority seems to believe that it is higher, over 40 percent (Waweru 2004). 25. Tax collections increased from 10.3 to 11.9 percent of GDP between 2003 and 2004 and have since stabilized at around 12 percent; see tax system diagnostics in World Bank (2009b). 26. Gehlbach (2008) notes that diversity not only at the national level, but also at the sectoral level was shaped by different political dynamics; the politically powerful were able to secure public goods, while also learning that it was often simpler to evade by running up tax arrears than to hide revenue from the tax authorities. 27. This is based on purchases reported (Guatemala, Ministry of Public Finance, 1990, “Análisis Evaluativo del Sistema Tributario de Guatemala y Apendice Tecnico: Consultoría para Administración Fiscal,� July, Policy Economics Group; KPMG; and Policy Research Program, Georgia State University, Washington, DC). 28. In 2006, VAT evasion was 31.6 percent; this compared favorably with the cases of Mexico (45.7 percent), Argentina (40.4 percent), and Bolivia (39.4 percent), but less favorably with the cases of Colombia (28.0 percent) and Chile (19.7 per- cent), not least because of persisting legal loopholes (World Bank data of 2007). The Political Economy of Controlling Tax Evasion and Illicit Flows 97 29. The Guatemalan government has attempted in recent years to address these problems, not least through the Anti-Evasion Law of 2006 (World Bank data of 2007). 30. Thus, under Menem, sanctions against evasion were strengthened, and the tem- porary closures by the tax authorities of the business premises of enterprises that failed to register for the VAT or to issue invoices rose from 751 in 1990 to 5,021 in the �rst nine months of 1991. However, this momentum subsequently lapsed; see Morisset and Izquierdo (1993). 31. This, for example, seems to apply to recent Chinese illicit capital flows in round- tripping to bene�t from tax incentives that have existed for foreign investment. 32. For this reason, tax havens require good governance, which is often provided or guaranteed by the countries of the Organisation for Economic Co-operation and Development by proxy (for example, Monaco through France) or as colo- nies or crown dependencies (for example, Isle of Man by the United Kingdom) (Dharmapala and Hines 2006). This centrality of the quality of governance may also help to explain why high-tax countries remain high tax, and low-tax coun- tries remain low tax (Markle and Shackelford 2009). 33. Across Africa, for instance, Keen and Mansour (2008) note that improvements in tax-GDP ratios in recent years have been, overall, almost entirely the result of rising commodity prices. 34. A working paper on state building published by the U.K. Department for Inter- national Development (Whaites 2008, 4) de�nes political settlements as “the forging of a common understanding, usually among elites, that their interests or beliefs are served by a particular way of organizing political power.� 35. Bratton and van de Walle (1997) assert that neo-patrimonial regimes in Africa have a cultural system of the quasi-taxation of �scal-type obligations to the extended circle of family, friends, and tribal chiefs, thereby complicating the idea that evasion arises from governments acting as self-serving Leviathans rather than as maximizers of social welfare. 36. Virmani (1987) points out that higher penalties for evasion could reduce tax revenues by making collusion between taxpayer and tax collector more pro�t- able unless penalties for corruption are strengthened and their implementation is given full political backing. 37. Fjeldstad and Tungodden (2003) argue that discretion and �scal corruption contribute to undermining the legitimacy of the tax administration and thereby increase tax evasion. 38. By contrast, Mahon (2004) �nds that tax reform is predicted by the political drive of new administrations, more authoritarian-leaning elected governments, the dominance of the president’s party in the legislature, established electoral systems, closed-list proportional representation, less polarized party systems, and more numerous parties. 39. For example, the political settlement requires regional taxpayers to trust the state to redistribute fairly (Ndulu et al. 2007). 98 Draining Development 40. The Nation, Lilongwe, Malawi, April 27, 2000, page 5. 41. Otherwise, high nominal taxes tend to produce considerable tax evasion, includ- ing the unrecorded activity of the shadow economy (Schneider and Enste 2000). 42. While some authors have claimed there is causation, Friedman et al. (2001) are not able to �nd evidence of a negative effect of taxes on the size of the shadow economy. No study has adequately addressed taxpayer perceptions of taxation not as a burden, but as an investment in the political settlement to tackle weak institutions and fund an effective state. 43. By 1997, under the pressure of the International Monetary Fund, the income subject to the highest marginal rate had been substantially reduced, but it was still at the remarkable level of 104 times the average per capita income (Shome 1999). 44. Measurements of tax effort and tax-GDP ratios are notoriously problematic, but, in the absence of other clear indicators, remain the most widely used basis for cross-country comparisons. 45. Thus, Belgium has far greater evasion rates than the Netherlands (Nam, Parsche, and Schaden 2001). 46. By contrast, one of the lowest rates in the world is in Afghanistan, where the government, in 2008, raised only 6.4 percent of national income, or US$750 million, in national revenue across a population of 25 million. 47. International Monetary Fund data for 2007. See Le, Moreno-Dodson, and Roj- chaichaninthorn (2008) on tax effort variation. 48. For example, for Pakistan, see World Bank (2004). References Abinales, P. N., and D. J. Amoroso. 2005. State and Society in the Philippines. State & Society, East Asia Series. Oxford: Rowman & Little�eld. Aljaaidi, K. S. Y., N. A. A. Manaf, and S. S. Karlinsky. 2011. “Tax Evasion as a Crime: A Survey of Perception in Yemen.� International Journal of Business and Man- agement 6 (9): 190–201. Alm, J., and J. Martinez-Vazquez. 2003. “Institutions, Paradigms, and Tax Evasion in Developing and Transition Countries.� In Public Finance in Developing and Transitional Countries: Essays in Honor of Richard Bird, ed. J. Martinez-Vazquez and J. Alm, 146–78. Studies in Fiscal Federalism and State-Local Finance Series. Cheltenham, U.K.: Edward Elgar. Andreoni, J., B. Erard, and J. Feinstein. 1998. “Tax Compliance.� Journal of Economic Literature 36 (2): 818–60. Barra, P., and M. Jarratt. 1998. “An Analysis of the Chilean Tax System.� Report, Divi- sion of Studies, Servicio de Impuestos Internos, Santiago, Chile. Bayart, J.-F., S. Ellis, and B. Hibou. 1999. The Criminalization of the State in Africa. African Issues. Oxford: James Currey; Bloomington, IN: Indiana University Press. The Political Economy of Controlling Tax Evasion and Illicit Flows 99 Bergman, M. 2009. Tax Evasion and the Rule of Law in Latin America: The Political Culture of Cheating and Compliance in Argentina and Chile. University Park, PA: Pennsylvania State University Press. Bird, R., J. Martinez-Vasquez, and B. Torgler. 2006. “Societal Institutions and Tax Effort in Developing Countries.� In The Challenges of Tax Reform in a Global Economy, ed. J. Alm, J. Martinez-Vazquez, and M. Rider, 283–338. New York: Springer Science+Business Media. Bird, R., and E. Zolt. 2007. “Tax Policy in Emerging Countries.� International Tax Pro- gram Paper 0707, International Tax Program, University of Toronto, Toronto. Boyce, J. K., and L. Zarsky. 1988. “Capital Flight from the Philippines, 1962–1986.� Philippine Journal of Development 15 (2): 191–222. Boylan, D. 1996. “Taxation and Transition: The Politics of the 1990 Chilean Tax Reform.� Latin American Research Review 31 (1): 7–31. Brada, J. C., A. M. Kutan, and G. Vukšic. 2009. “The Costs of Moving Money across Borders and the Volume of Capital Flight: The Case of Russia and Other CIS Countries.� EMG Working Paper WP-EMG-28–2009, Emerging Markets Group, Cass Business School, City University, London. Braithwaite, V. A., ed. 2003. Taxing Democracy: Understanding Tax Avoidance and Evasion. Aldershot, U.K.: Ashgate. Bratton, M., and N. van de Walle. 1997. Democratic Experiments in Africa: Regime Transitions in Comparative Perspective. Cambridge Studies in Comparative Pol- itics. Cambridge, U.K.: Cambridge University Press. Bräutigam, D., O.-H. Fjeldstad, and M. Moore, eds. 2008. Taxation and State-Building in Developing Countries: Capacity and Consent. Cambridge, U.K.: Cambridge University Press. Chabal, P., and J.-P. Daloz. 1999. Africa Works: Disorder as Political Instrument. Afri- can Issues. Oxford: James Currey; Bloomington, IN: Indiana University Press. Cheibub, J. A. 1998. “Political Regimes and the Extractive Capacity of Governments: Taxation in Democracies and Dictatorships.� World Politics 50 (3): 349–76. Chiarini, B., E. Marzano, and F. Schneider. 2009. “Tax Rates and Tax Evasion: An Empirical Analysis and Long-Run Characteristics in Italy.� Working Paper 1–2009 (January), Department of the Treasury, Italian Ministry of Economy and Finance, Rome. Chipeta, C. 2002. “The Second Economy and Tax Yield in Malawi.� AERC Research Paper 113 (March), African Economic Research Consortium, Nairobi. Daunton, M. 2001. Trusting Leviathan: The Politics of Taxation in Britain, 1799–1914. Cambridge, U.K.: Cambridge University Press. de Melo, M. 2007. “Institutional Weakness and the Puzzle of Argentina’s Low Taxa- tion.� Latin American Politics and Society 49 (4): 115–48. Devarajan, S., and S. I. Hossain. 1995. “The Combined Incidence of Taxes and Public Expenditures in the Philippines.� Policy Research Working Paper 1543, World Bank, Washington, DC. 100 Draining Development DFID (U.K. Department for International Development). 2006. Eliminating World Poverty: Making Governance Work for the Poor; A White Paper on International Development. Secretary of State for International Development. London: Her Majesty’s Stationery Of�ce. Dharmapala, D., and J. R. Hines Jr. 2006. “Which Countries Become Tax Havens?� NBER Working Paper 12802, National Bureau of Economic Research, Cam- bridge, MA. Dia, M. 1993. “A Governance Approach to Civil Service Reform in Sub-Saharan Africa.� World Bank Technical Paper 225, World Bank, Washington, DC. ———. 1996. Africa’s Management in the 1990s and Beyond: Reconciling Indigenous and Transplanted Institutions. Directions in Development Series. Washington, DC: World Bank. Di John, J. 2006. “The Political Economy of Taxation and Tax Reform in Developing Countries.� WIDER Research Paper 74, World Institute for Development Eco- nomics Research, Helsinki. ———. 2008. “Why Is the Tax System So Ineffective and Regressive in Latin America?� Development Viewpoint 5 (June), Centre for Development Policy and Research, School of Oriental and African Studies, University of London, London. ———. 2009. From Windfall to Curse? Oil and Industrialization in Venezuela, 1920 to the Present. University Park, PA: Pennsylvania State University Press. Engelschalk, M., and B. Nerré. 2002. “Gorillas, Gender and Tax Morale: Conserva- tionism, French Catholicism and Female Tax Compliance Rates in Rwanda, 1998–2003.� In Forschungsspektrum aktueller Finanzwissenschaft, ed. B. Nerré, 142–76. Heidenau, Germany: PD Verlag. Englebert, P. 2000. State Legitimacy and Development in Africa. Boulder, CO: Lynne Rienner. Erbe, S. 1985. “The Flight of Capital from Developing Countries.� Intereconomics 20 (6): 268–75. Esteller-Moré, A. 2005. “Is There a Connection between the Tax Administration and the Political Power?� International Tax and Public Finance 12 (5): 639–63. Eurodad (European Network on Debt and Development). 2008. “Addressing Devel- opment’s Black Hole: Regulating Capital Flight.� Eurodad Report, May 8, Euro- dad, Brussels. Everest-Phillips, M. 2008a. “Business Tax as State-Building in Developing Countries: Applying Governance Principles in Private Sector Development.� International Journal of Regulation and Governance 8 (2): 123–54. ———. 2008b. “The Myth of ‘Secure Property Rights’: Good Economics as Bad His- tory and Its Impact on International Development.� SPIRU Working Paper 23 (May), Strategic Policy Impact and Research Unit, Overseas Development Insti- tute, London. ———. 2009a. “When Do Elites Pay Taxes? Tax Morale and State-Building in Devel- oping Countries.� Paper presented at the World Institute for Development Eco- nomics Research Conference, “WIDER Conference on the Role of Elites in Eco- nomic Development,� Helsinki, June 12–13. The Political Economy of Controlling Tax Evasion and Illicit Flows 101 ———. 2009b. “Taxation as State-Building: Principles for Reform in Developing Countries.� Paper presented at the 65th Annual Congress of the International Institute of Public Finance, Cape Town, August 13–16. ———. 2010. “State-Building Taxation for Developing Countries: Principles for Reform.� Development Policy Review 28 (1): 75–96. ———. 2011. “Tax, Governance and Development.� Discussion Paper 11, Common- wealth Secretariat, London. Fatehi, K. 1994. “Capital Flight from Latin America as a Barometer of Political Insta- bility.� Journal of Business Research 30 (2): 187–95. Feld, L. P., and B. S. Frey. 2002. “Trust Breeds Trust: How Taxpayers Are Treated.� Economics of Governance 3 (2): 87–99. Fjeldstad, O.-H., and A. al-Zagha. 2004. “Taxation during State Formation: Lessons from Palestine, 1994–2000.� Forum for Development Studies 31 (1): 89–113. Fjeldstad, O.-H., L. Katera, and E. Ngalewa. 2009. “Maybe We Should Pay Tax after All? Citizens’ Views on Taxation in Tanzania.� REPOA Special Paper 29–2009 (April), Research on Poverty Alleviation, Dar es Salaam, Tanzania. Fjeldstad, O.-H., and M. Moore. 2009. “Revenue Authorities and Public Authority in Sub-Saharan Africa.� Journal of Modern African Studies 47 (1): 1–18. Fjeldstad, O.-H., and B. Tungodden. 2003. “Fiscal Corruption: A Vice or a Virtue?� World Development 31 (8): 1459–67. Freedom House. 2008. “Freedom in the World 2008: Paraguay.� Country report (July 2), Freedom House, Washington, DC. http://www.freedomhouse.org/template .cfm?page=22&year=2008&country=7468. Friedman, E., S. Johnson, D. Kaufmann, and P. Zoido. 2001. “Dodging the Grabbing Hand: The Determinants of Unof�cial Activity in 69 Countries.� Journal of Pub- lic Economics 76 (3): 459–93. Fritz, V., K. Kaiser, and B. Levy. 2009. “Problem-Driven Governance and Political Economy Analysis: Good Practice Framework.� World Bank, Washington, DC. Fuest, C., and N. Riedel. 2009. “Tax Evasion, Tax Avoidance and Tax Expenditures in Developing Countries: A Review of the Literature.� Report, Oxford University Centre for Business Taxation, Oxford. Gehlbach, S. 2008. Representation through Taxation: Revenue, Politics, and Develop- ment in Postcommunist States. Cambridge Studies in Comparative Politics. Cambridge, U.K.: Cambridge University Press. Gërxhani, K., and A. Schram. 2002. “Tax Evasion and the Source of Income: An Experimental Study in Albania and the Netherlands.� Discussion paper, Tinber- gen Institute, Amsterdam. Ghura, D. 2002. “Tax Revenue in Sub-Saharan Africa: Effects of Economic Policies and Corruption.� In Governance, Corruption, and Economic Performance, ed. G. T. Abed and S. Gupta, 369–94. Washington, DC: International Monetary Fund. Gray, C. S. 2001. “Enhancing Transparency in Tax Administration in Madagascar and Tanzania.� African Economic Policy Discussion Paper 77 (August), United States Agency for International Development, Washington, DC. 102 Draining Development Hayoz, N., and S. Hug, eds. 2007. Tax Evasion, Trust, and State Capacities. Interdisci- plinary Studies on Central and Eastern Europe 3. Bern: Peter Lang. Heij, G. 2007. “Who Pulled the Strings? A Comparative Study of Indonesian and Vietnamese Tax Reform. PhD thesis, University of Groningen, Groningen, the Netherlands. Hettich, W., and S. L. Winer. 1999. Democratic Choice and Taxation: A Theoretical and Empirical Analysis. Cambridge, U.K.: Cambridge University Press. Hobson, J. M. 1997. The Wealth of States: A Comparative Sociology of International Economic and Political Change. Cambridge Studies in International Relations. Cambridge, U.K.: Cambridge University Press. Hoppit, J. 2011. “Compulsion, Compensation and Property Rights in Britain, 1688– 1833.� Past & Present 210 (1): 93–128. IMF (International Monetary Fund). 2007. “Republic of Madagascar: Selected Issues.� IMF Country Report 07/239 (July), IMF, Washington, DC. Kaldor, N. 1963. “Will Underdeveloped Countries Learn to Tax?� Foreign Affairs 41 (2): 410–18. Kauffmann, C. 2005. “Financing SMEs in Africa.� Policy Insights 7 (May), OECD Development Centre, Organisation for Economic Co-operation and Develop- ment, Paris. Keen, M., and M. Mansour. 2008. “Revenue Mobilization in Sub-Saharan Africa: Challenges from Globalization.� Draft working paper, August 11, Fiscal Affairs Department, International Monetary Fund, Washington, DC. Kenny, L., and S. Winer. 2006. “Tax Systems in the World: An Empirical Investigation into the Importance of Tax Bases, Administration Costs, Scale and Political Regime.� International Tax and Public Finance 13 (2): 181–215. Khan, M. H. 2000. “Rent-Seeking as a Process: Inputs, Rent-Outcomes and Net Effects.� In Rents, Rent-Seeking and Economic Development: Theory and Evidence in Asia, ed. M. H. Khan and K. S. Jomo, 70–104. Cambridge, U.K.: Cambridge University Press. ———. 2005. “Markets, States and Democracy: Patron-Client Networks and the Case for Democracy in Developing Countries.� Democratization 12 (5): 705–24. Kim, S. 2008. “Does Political Intention Affect Tax Evasion?� Journal of Policy Model- ing 30 (3): 401–15. Kimenyi, M. 2006. “Ethnicity, Governance and the Provision of Public Goods.� Jour- nal of African Economies 15 (1): 62–99. KIPPRA (Kenya Institute for Public Policy Research and Analysis). 2004. “Tax Com- pliance Study.� Tax Policy Unit, Macroeconomics Division, KIPPRA, Nairobi. Ku, H. B. 2003. Moral Politics in a South Chinese Village: Responsibility, Reciprocity, and Resistance. Asian Voices. Oxford: Rowman & Little�eld. Le, T. M., B. Moreno-Dodson, and J. Rojchaichaninthorn. 2008. “Expanding Taxable Capacity and Reaching Revenue Potential: Cross-Country Analysis.� Policy Research Working Paper 4559, World Bank, Washington, DC. The Political Economy of Controlling Tax Evasion and Illicit Flows 103 Levi, M. 1989. Of Rule and Revenue. California Series on Social Choice and Political Economy. Berkeley, CA: University of California Press. Lewis, A. 1982. The Psychology of Taxation. Basingstoke: U.K.: Palgrave Macmillan. Lieberman, E. 2002. “Taxation Data as Indicators of State-Society Relations: Possi- bilities and Pitfalls in Cross-National Research.� Studies in Comparative Interna- tional Development 36 (4): 89–115. ———. 2003. Race and Regionalism in the Politics of Taxation in Brazil and South Africa. Cambridge, U.K.: Cambridge University Press. Mahon, J. E. 2004. “Causes of Tax Reform in Latin America, 1977–95.� Latin Ameri- can Research Review 39 (1): 3–30. Manasan, R. G. 1988. “Tax Evasion in the Philippines, 1981–1985.� Philippine Journal of Development 15 (2): 167–88. Manasan, R. G., and R. G. Querubin. 1987. “Revenue Performance of National Gov- ernment Taxes, 1975–1985.� Staff Paper 87–01, Philippine Institute for Devel- opment Studies, Manila. Markle, K., and D. A. Shackelford. 2009. “Do Multinationals or Domestic Firms Face Higher Effective Tax Rates?� NBER Working Paper 15091, National Bureau of Economic Research, Cambridge, MA. McGee, R. W., and C. Lingle. 2005. “The Ethics of Tax Evasion: A Survey of Guatema- lan Opinion.� Paper presented at the 60th International Atlantic Economic Conference, New York, October 6–9. Mkrtchyan, A. V. 2001. “Tax Reform in Armenia.� NISPAcee Occasional Paper 2 (3): 44–56, Network of Institutes and Schools of Public Administration in Central and Eastern Europe, Bratislava, Slovak Republic. Mohammad, S., and J. Whalley. 1984. “Rent Seeking in India: Its Costs and Policy Signi�cance.� Kyklos 37 (3): 387–413. Moreno-Dodson, B., and Q. Wodon, eds. 2008. Public Finance for Poverty Reduction: Concepts and Case Studies from Africa and Latin America. Directions in Devel- opment. Washington, DC: World Bank. Morisset, J., and A. Izquierdo. 1993. “Effects of Tax Reform on Argentina’s Revenues.� Policy Research Working Paper 1192, World Bank, Washington, DC. Mosley, P. 2007. “The ‘Political Poverty Trap’: Bolivia 1999–2007.� WEF Working Paper 0020, World Economy and Finance Research Programme, Birkbeck, Uni- versity of London, London. Musgrave, R. 1996. “The Role of the State in Fiscal Theory.� International Tax and Public Finance 3 (3): 247–58. Nam, C. W., R. Parsche, and B. Schaden. 2001. “Measurement of Value Added Tax Evasion in Selected EU Countries on the Basis of National Accounts Data.� CESifo Working Paper 431, CESifo Group, Munich. Ndikumana, L. 2004. “Fiscal Policy, Conflict and Reconstruction in Burundi and Rwanda.� In Fiscal Policy for Development: Poverty, Reconstruction and Growth, ed. T. Addison and A. Roe, chap. 13. Studies in Development Economics and 104 Draining Development Policy. Helsinki: World Institute for Development Economics Research; Basing- stoke, U.K.: Palgrave Macmillan. Ndulu, B. J., S. A. O’Connell, J.-P. Azam, R. H. Bates, A. K. Fosu, J. W. Gunning, and D. Njinkeu, eds. 2008. Country Case Studies. Vol. 2 of The Political Economy of Economic Growth in Africa, 1960–2000. Oxford: Oxford University Press. Ndulu, B. J., S. A. O’Connell, R. H. Bates, P. Collier, and C. C. Soludo, eds. 2007. An Analytic Survey. Vol. 1 of The Political Economy of Economic Growth in Africa, 1960–2000. Oxford: Oxford University Press. North, D. C., J. J. Wallis, S. B. Webb, and B. R. Weingast. 2007. “Limited Access Orders in the Developing World: A New Approach to the Problem of Development.� Policy Research Working Paper 4359, World Bank, Washington, DC. North, D. C., J. J. Wallis, and B. R. Weingast. 2009. Violence and Social Orders: A Con- ceptual Framework for Interpreting Recorded Human History. Cambridge, U.K.: Cambridge University Press. O’Brien, P. 2001. “Fiscal Exceptionalism: Great Britain and Its European Rivals; From Civil War to Triumph at Trafalgar and Waterloo.� LSE Paper 65/01 (October), Department of Economic History, London School of Economics, London. OECD (Organisation for Economic Co-operation and Development). 2008. “Con- cepts and Dilemmas of State-Building in Fragile Situations: From Fragility to Resilience.� OECD/DAC Discussion Paper, Journal on Development 9 (3), OECD, Paris. http://www.oecd.org/dataoecd/59/51/41100930.pdf. OECD (Organisation for Economic Co-operation and Development) and AfDB (Afri- can Development Bank). 2008. African Economic Outlook 2008. Paris: OECD. Oleinik, A. N. 1999. “A Trustless Society: The Influence of the August 1998 Crisis on the Institutional Organization of the Russians’ Everyday Life.� Unpublished working paper, Stanford University, Stanford, CA. Orviska, M., and J. Hudson. 2003. “Tax Evasion, Civic Duty and the Law Abiding Citizen.� European Journal of Political Economy 19 (1): 83–102. Paul, E. F., F. D. Miller Jr., and J. Paul, eds. 2006. Taxation, Economic Prosperity, and Distributive Justice. Vol. 23, Part 2 of Social Philosophy and Policy. Cambridge, U.K.: Cambridge University Press. Pritchard, W. 2009. “The Politics of Taxation and Implications for Accountability in Ghana 1981–2008.� IDS Working Paper 330, Institute of Development Studies, Brighton, U.K. Reuter, P., and E. M. Truman. 2004. Chasing Dirty Money: The Fight against Money Laundering. Washington, DC: Institute for International Economics. Rose-Ackerman, S. 1999. Corruption and Government: Causes, Consequences, and Reform. Cambridge, U.K.: Cambridge University Press. Rothstein, B. 2001. “Trust, Social Dilemmas, and Collective Memories: On the Rise and Decline of the Swedish Model.� Journal of Theoretical Politics 12 (4): 477–503. Saeed, K. A. 1996. “Principle of Taxation: The Taxation System of Pakistan.� Pakistan Journal of Public Administration, January–June. Sánchez, J. A. 2008. “Shaping Taxation: Economic Elites and Fiscal Decision-Making in Argentina, 1920–1945.� Journal of Latin American Studies 40 (1): 83–108. The Political Economy of Controlling Tax Evasion and Illicit Flows 105 Sánchez, O. 2006. “Tax System Reform in Latin America: Domestic and International Causes.� Review of International Political Economy 13 (5): 772–801. ———. 2009. “Tax Reform Paralysis in Post-Conflict Guatemala.� New Political Economy 14 (1): 101–31. Schneider, A. 2007. “Socioeconomic Change, Political Parties, and Tax Capacity: Structure, Agency, and Process in Central America.� Paper presented at the American Political Science Association’s annual meeting, Chicago, August 30. Schneider, A., V. Lledo, and M. Moore. 2004. “Social Contracts, Fiscal Pacts and Tax Reform in Latin America.� Unpublished working paper, Inter-American Devel- opment Bank, Washington, DC. Schneider, B. 2001. “Measuring Capital Flight: Estimates and Interpretations.� Draft working paper, Overseas Development Institute, London. Schneider, F. 2007. “Shadow Economies and Corruption All Over the World: New Estimates for 145 Countries.� Economics E-Journal 1 (9): 1–66. Schneider, F., and D. H. Enste. 2000. “Shadow Economies: Size, Causes and Conse- quences.� Journal of Economic Literature 38 (1): 77–114. Shome, P. 1999. “Taxation in Latin America: Structural Trends and Impact of Admin- istration.� IMF Working Paper 99/19, International Monetary Fund, Washing- ton, DC. Silvani, C., and K. Baer. 1997. “Designing a Tax Administration Reform Strategy: Experiences and Guidelines.� IMF Working Paper 97/30, International Mone- tary Fund, Washington, DC. Slemrod, J., and J. M. Bakija. 1996. Taxing Ourselves: A Citizen’s Guide to the Great Debate over Tax Reform. Cambridge, MA: MIT Press. Steinmo, S. 1993. Taxation and Democracy: Swedish, British, and American Approaches to Financing the Modern State. New Haven, CT: Yale University Press. Stewart, M., ed. 2007. “Tax Law and Political Institutions.� Special issue, Law in Con- text 24 (2), Federation Press, Annandale NSW, Australia. Stotsky, J. G., and A. WoldeMariam. 1997. “Tax Effort in Sub-Saharan Africa.� IMF Working Paper 97/107, International Monetary Fund, Washington, DC. Tanzi, V. 2000. “Taxation in Latin America in the Last Decade.� Working Paper 76 (December), Center for Research on Economic Development and Policy Reform, Stanford University, Stanford, CA. Tanzi, V., and P. Shome. 1993. “A Primer on Tax Evasion.� IMF Working Paper 93/21, International Monetary Fund, Washington, DC. Teera, J. M., and J. Hudson. 2004. “Tax Performance: A Comparative Study.� Journal of International Development 16 (6): 785–802. Tendler, J. 2002. “Small Firms, the Informal Sector, and the Devil’s Deal.� Bulletin of the Institute of Development Studies 33 (3): 3–33. Therkildsen, O. 2001. “Understanding Taxation in Poor African Countries: A Critical Review of Selected Perspectives.� Forum for Development Studies 2001 28 (1): 99–123. ———. 2008. “Taxation and State Building with a (More) Human Face.� DIIS Policy Brief (October), Danish Institute of International Studies, Copenhagen. 106 Draining Development Thies, C. 2004. “State Building, Interstate and Intrastate Rivalry: A Study of Post- Colonial Developing Country Extractive Efforts, 1975–2000.� International Studies Quarterly 48 (1): 53–72. Tilly, C. 1975. The Formation of National States in Western Europe. Princeton, NJ: Princeton University Press. ———. 2003. The Politics of Collective Violence. Cambridge Studies in Contentious Politics. Cambridge, U.K.: Cambridge University Press. Timmons, J. F. 2005. “The Fiscal Contract: States, Taxes, and Public Services.� World Politics 57 (July): 530–67. Torgler, B. 2003. “Beyond Punishment: A Tax Compliance Experiment with Taxpay- ers in Costa Rica.� Revista de Analisis Economico 18 (1): 27–56. Torgler, B., I. C. Demir, A. Macintyre, and M. Schaffner. 2008. “Causes and Conse- quences of Tax Morale: An Empirical Investigation.� Economic Analysis and Policy 38 (2): 313–39. Torgler, B., M. Schaffner, and A. Macintyre. 2007. “Tax Compliance, Tax Morale and Governance Quality.� CREMA Working Paper 2007–17, Center for Research in Economics, Management and the Arts, Basel, Switzerland. UNECA (United Nations Economic Commission for Africa). 2009. “Enhancing Domestic Resource Mobilization: Challenges and Opportunities in Southern Africa.� Document E/ECA/SA/ICE.XV/2009/3 (May 21), United Nations Eco- nomic and Social Council, Geneva. van de Walle, N. 2001. African Economies and the Politics of Permanent Crisis, 1979– 1999. Political Economy of Institutions and Decisions. Cambridge, U.K.: Cam- bridge University Press. Vihanto, M. 2000. “Tax Evasion in a Transition from Socialism to Capitalism: The Psychology of the Social Contract.� BOFIT Discussion Paper 6/2000, Bank of Finland, Helsinki. Virmani, A. 1987. “Tax Evasion, Corruption, and Administration: Monitoring the People’s Agents under Symmetric Dishonesty.� DRD Discussion Paper 271, Development Research Department, World Bank, Washington, DC. Vittal, N. 1999. “Applying Zero Tolerance to Corruption.� Unpublished paper, Cen- tral Vigilance Commission, New Delhi. von Soest, C. 2006. “How Does Neopatrimonialism Affect the African State? The Case of Tax Collection in Zambia.� GIGA Working Paper 33, German Institute of Global and Area Studies, Hamburg. Wa Wamwere, K. 2003. Negative Ethnicity: From Bias to Genocide. New York: Seven Stories Press. Waweru, M. G. 2004. “Tax Administration in Kenya: Problems and Prospects.� Address presented to the Federation of Kenya Employers, Rift Valley Branch’s 16th Annual General Meeting, Nakuru, Kenya, March 10. http://www.revenue .go.ke/speeches/cgspeechtax100304.html. The Political Economy of Controlling Tax Evasion and Illicit Flows 107 Wedeman, A. 1997. “Looters, Rent-Scrapers, and Dividend-Collectors: Corruption and Growth in Zaire, South Korea, and the Philippines.� Journal of Developing Areas 31 (4): 57–78. Whaites, A. 2008. “States in Development: Understanding State-Building.� Working paper, U.K. Department for International Development, London. Wickham, C. 2010. The Inheritance of Rome: A History of Europe from 400 to 1000. London: Penguin. World Bank. 2004. “Pakistan: Tax Administration Reforms Project.� Project Informa- tion Document, Report 30640, World Bank, Washington, DC. ———. 2006. “Chile: Development Policy Review.� Report 37429 (June), Poverty Reduction and Economic Management Unit, Latin America and the Caribbean Region, World Bank, Washington, DC. ———. 2008a. “The Political Economy of Taxation in Developing Countries: Chal- lenges to Practitioners.� In collaboration with U.K. Department for Interna- tional Development, International Finance Corporation, Washington DC. ———. 2008b. “Tax Perception and Compliance Cost Surveys: A Tool for Tax Reform.� In collaboration with U.K. Department for International Develop- ment. Investment Climate Advisory Services, World Bank, Washington, DC. https://www.wbginvestmentclimate.org/uploads/TPCCS_Consolidated_Web .pdf. ———. 2009a. “Program Document for a Proposed Economic Management and Governance Reform Grant to the Central African Republic.� Report 47559-CF (March 9), World Bank, Washington, DC. ———. 2009b. “Program Document for a Proposed First Public Sector Program- matic Development Policy Loan to the Republic of Paraguay.� Report 47712-PY (March 31), World Bank, Washington, DC. ———. 2009c. “Program Document for a Proposed First Development Policy Oper- ation to the Republic of Armenia.� Report 48605-AM (June 2), World Bank, Washington, DC. Yashar, D. J. 1997. Demanding Democracy: Reform and Reaction in Costa Rica and Guatemala, 1870s–1950s. Stanford, CA: Stanford University Press. 4 Tax Evasion and Tax Avoidance: The Role of International Pro�t Shifting Clemens Fuest and Nadine Riedel Abstract In the debate on the impact of illicit capital flows on developing coun- tries, the view is widespread that pro�t shifting to low-tax jurisdictions undermines the ability of developing countries to raise tax revenue. While the shifting of income out of developed countries is a widely debated issue, empirical evidence on the magnitude of the problem and on the factors driving income shifting is scarce. This chapter reviews the literature on tax avoidance and evasion through border crossing income shifting out of developing countries. We discuss methods and available data sets that can be used to gain new insights into the problem of corporate income shifting. We argue that the results of many existing studies on tax avoidance and evasion in devel- oping countries are dif�cult to interpret, mainly because the measure- ment concepts used have a number of drawbacks. We discuss some alternative methods and data sets and present empirical evidence that supports the view that pro�t shifting out of many developing countries and into tax havens takes place. Additional tables and �gures from this chapter can be found at http://go.worldbank.org /N2HMRB4G20. 109 110 Draining Development? Introduction In the debate on the impact of illicit capital flows on developing coun- tries, the view is widespread that illicit flows undermine the ability of developing countries to raise tax revenue. The reason stated is that illicit capital flows may channel resources to the informal economy or to other jurisdictions, in particular to tax havens, so that the resources escape taxation. A large part of this activity takes place in the shadow economy and largely escapes public attention. Yet, parts of the of�cial economy, particularly multinational �rms, are accused of engaging in tax avoid- ance and tax evasion as well. They are criticized for shifting income out of developing countries and into tax havens to avoid paying corporate income taxes. Since developing countries frequently lack appropriate legislative and administrative resources, they are generally seen to be more vulnerable to income shifting relative to developed countries. While the shifting of income out of developed countries is a widely debated issue, empirical evidence on the magnitude of the problem and on the factors driving income shifting is scarce. This chapter contributes to the debate as follows. First, we review the literature on tax avoidance and evasion through border crossing income shifting out of developing countries. Second, we discuss methods and available data sets that can be used to gain new insights into the problem of corporate income shifting. There is a growing number of empirical studies on corporate pro�t shifting in the countries of the Organisation for Economic Co-operation and Development (OECD). Many of these studies use appropriate data and sophisticated econometric methods, and the results offer valuable insights into corporate pro�t shifting. Unfortunately, almost none of these studies include developing countries. The main reason is that, for developing countries, signi�cantly fewer data are available. A number of studies, mostly published by nongovernmental organizations (NGOs), try to estimate income shifting and the tax revenue losses suffered by developing countries. These studies have the merit of attracting the attention of a wider public to the issue of income shifting out of devel- oping countries. However, the results of these studies are somewhat dif- �cult to interpret, mainly because the measurement concepts used have a number of drawbacks (Fuest and Riedel 2009). Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 111 The setup of the rest of this chapter is as follows. In the next section, we briefly introduce the concept of pro�t shifting by multinational �rms and discuss empirical approaches that have been used to detect pro�t shifting. In the following section, we review existing studies on pro�t shifting out of developing countries. The section builds on and extends our earlier work, in Fuest and Riedel (2009). The subsequent section discusses the particular role of tax havens. In the penultimate section, we suggest and discuss the pros and cons of different econometric identi�- cation strategies and data sets that can be used to gain new insights into the phenomenon of pro�t shifting out of developing countries. We also provide some evidence from one of the data sets that supports the view that a signi�cant amount of pro�t shifting out of developing countries and into tax havens does take place. The last section concludes. Multinational Firms and the Concept of Pro�t Shifting Intra�rm pro�t shifting For purposes of taxation, the pro�ts of a multinational �rm have to be allocated to the individual jurisdictions where the �rm �les for income taxation. This is usually accomplished through the method of separate accounting. Each entity (subsidiary or permanent establishment) of the multinational �rm individually calculates the income it has generated. Transactions between different entities of a multinational �rm (con- trolled transactions) should, in principle, be treated as transactions with third parties (uncontrolled transactions). However, multinational �rms may use controlled transactions to shift income across countries. For instance, they may shift income from high-tax jurisdictions to low-tax jurisdictions using transfer pricing or intra�rm debt. The concept of income shifting raises the question of whether a true or objective distribution of pro�ts earned by the individual entities of a multinational �rm can be identi�ed. Achieving this is complicated for a number of reasons. In particular, the entities of multinational �rms typ- ically jointly use resources speci�c to the �rm such as a common brand name or �rm-speci�c expertise. Pricing these resource flows appropri- ately is dif�cult because goods traded between unrelated parties are usu- ally different. It is an important characteristic of many multinational �rms that the individual entities jointly use resources that could not be 112 Draining Development? used in the same way if they were separate �rms. If they could be used in the same way, there would be no reason to create the multinational �rm in the �rst place. For this reason, it is dif�cult to establish what a pro�t distribution in the absence of pro�t shifting would look like. Most empirical studies on corporate income shifting, however, do not explicitly refer to a hypothetical distribution of pro�ts that would occur in the absence of income shifting. Instead, they focus on particular fac- tors that are likely to drive income shifting, and they try to explore whether these factors affect the distribution of reported income across countries and, if so, how large these effects are. In this chapter, we focus on tax-induced income shifting. Empirical work in this area essentially uses two types of approaches to investigate whether and to what extent �rms shift income to exploit tax differences across countries. The �rst approach looks directly at the use of instruments for pro�t shifting. For instance, some studies focus on income shifting through debt and ask whether, all else being equal, multinational �rms use more debt in high- tax countries relative to low-tax countries (Buettner and Wamser 2007; Huizinga, Laeven, and Nicodeme 2008). Other instruments that have been studied in the context of international pro�t shifting are transfer pricing and the location of intangible assets (Clausing 2003; Dischinger and Riedel 2008). The second approach focuses on the result of tax-induced pro�t shift- ing: the overall pro�tability of individual entities of multinational �rms in different countries. In the presence of tax-induced income shifting, one would expect to observe a negative correlation between reported pro�tability and tax levels (Grubert and Mutti 1991; Huizinga and Lae- ven 2008; Weichenrieder 2009). One drawback of this approach is that a negative correlation between pretax pro�tability and tax levels may even emerge in the absence of income shifting. The reason is that the location of economic activity itself is influenced by taxes. Firms have incentives to locate highly pro�table projects in low-tax jurisdictions. Both approaches deliver estimates of the (marginal) impact of tax dif- ferences on income shifting behavior. Under certain assumptions, these estimates can be used to calculate a hypothetical pro�t distribution across countries that would occur in the absence of tax differences. For instance, Huizinga and Laeven (2008) analyze a sample of European multinational �rms and �nd that, in 1999, the corporate tax base of Ger- Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 113 many, which was the country with the highest corporate tax rate in Europe, would have increased by 14 percent if there had been no tax incentives to shift income to other countries. Pro�t shifting and transactions among unrelated parties Pro�t shifting as discussed in the previous section takes place through transactions between entities of multinational �rms in different coun- tries. If two corporations located in two different countries belong to the same multinational �rm or are controlled by the same interest, it is uncontroversial that transactions between these two �rms may be used to shift pro�ts across borders. However, some authors have argued that transactions between unrelated �rms may also be used to shift pro�ts across borders. This is emphasized, in particular, by Raymond Baker (2005) in his book, Capitalism’s Achilles Heel: Dirty Money and How to Renew the Free-Market System. There, he quanti�es the yearly illicit �nancial flows out of developing countries through the business sector at US$500 billion to US$800 billion. Baker’s book provides a breakdown of this number according to dif- ferent activities. The analysis claims that slightly above 60 percent of these �nancial flows are related to legal commercial activities, whereas the rest is assigned to criminal activity. Baker argues that money earned on legal commercial activities leaves developing countries through three potential channels: the mispricing of goods traded between independent parties, the distortion of transfer prices charged on goods traded within a multinational �rm, and fake transactions. With respect to mispricing between unrelated parties, Baker bases his estimate on 550 interviews he conducted with of�cials from trading companies in 11 economies in the early 1990s: Brazil; France; Germany; Hong Kong SAR, China; India; Italy; the Republic of Korea; the Nether- lands; Taiwan, China; the United Kingdom; and the United States. Because Baker assured anonymity, he does not make the data publicly available, but argues that the data contain appropriate information on trading practices. He reports that the interviewees con�rmed that collusion was com- mon between importers and exporters to draw money out of developing countries. Speci�cally, he states that “mispricing in order to generate kickbacks into foreign bank accounts was treated as a well-understood 114 Draining Development? and normal part of transactions� by the interviewed managers (Baker 2005, 169). As a result of his study, Baker estimates that 50 percent of foreign trade transactions with Latin American countries are mispriced by, on average, around 10 percent, adding to a worldwide average mis- pricing of goods traded between third parties of 5 percent. Similar, slightly larger �gures are reported for countries in Africa and Asia, sug- gesting a level of mispricing at 5 to 7 percent. Studies on International Pro�t Shifting in Developing Countries Most existing empirical studies on tax-induced pro�t shifting focus on OECD countries. Studies on pro�t shifting in developing countries are scarce. Most studies on tax-induced pro�t shifting in developing coun- tries (as well as income shifting undertaken for other reasons) have been published by NGOs. Below, we discuss and criticize some of these stud- ies. It should be emphasized, though, that these studies have the merit of attracting the attention of a wider public to this important issue. The trade mispricing approach Studies based on the trade mispricing approach start with the idea that �rms may manipulate prices of internationally traded goods to shift income across countries. This idea is known from empirical work on income shifting in developed countries (see Clausing 2003 and the lit- erature cited there). The key question is how the manipulation of prices is identi�ed. There are different identi�cation strategies with different implications. As mentioned in the preceding section, Baker (2005) uses interviews to estimate the extent of mispricing in trade transactions with develop- ing countries. He quanti�es the income shifted out of developing coun- tries through mispricing activities by multiplying the low end of his interview-based mispricing estimate (that is, mispricing of 5 percent of import and export value, respectively) with the sum of imports and exports of developing countries, which is equal to approximately US$4 trillion. Given this, he arrives at what he refers to as a lower-bound esti- mate of US$200 billion for capital outflows due to trade. Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 115 The main disadvantage of Baker’s approach to the estimation of capi- tal outflows is that it is based on a relatively small number of interviews, and these interviews are con�dential. Therefore, the results cannot be replicated. Another approach to identifying mispricing is used by Pak (2007), who identi�es abnormally priced import and export transactions through the price �lter matrix method. For example, the method might rely on trade statistics that offer information on the prices of transac- tions for individual product groups such as U.S. trade statistics that offer information about the prices of refrigerators imported into the country in a given year. One might then classify all transactions as overpriced if they involve prices that exceed the average price for imported refrigera- tors by a certain amount (for instance, prices in the upper quartile of the price range), while classifying as underpriced all transactions involving prices suf�ciently below the average price in that product group. On this basis, one calculates the income shifted into and out of the country. Pak (2007, 120) does so as follows: “the dollar amounts are computed by aggregating the amount deviated from [the] lower quartile price for every abnormally low priced U.S. import and the amount deviated from [the] upper quartile price for every abnormally high priced U.S. export.� The analysis in Pak (2007) leads to the estimate that U.S. imports from all other countries were underpriced by approximately US$202 billion in 2005, or 12.1 percent of total imports. The value of U.S. exports in the same year was overpriced by US$50 billion, or 5.5 percent of over- all exports. Zdanowicz, Pak, and Sullivan (1999) investigate the interna- tional merchandise statistics between Brazil and the United States and �nd that the amount of income shifted because of abnormal pricing is between 11.1 percent for underinvoiced exports from Brazil and 15.2 percent for overinvoiced imports to Brazil. Pak, Zanakis, and Zdanowicz (2003) use the same framework to investigate capital outflows from Greece because of the mispricing of internationally traded goods and services. The share of income shifted from Greece to the world varies between 2.0 percent for underinvoiced exports from Greece and 5.9 per- cent for overinvoiced imports to Greece. Another study using this approach has been published recently by Christian Aid (2009), which argues that pro�t shifting out of developing countries through trade mispricing in 2005–07 represented above US$1 116 Draining Development? trillion, giving rise to a yearly tax revenue loss of US$121.8 billion per year. Using the same approach, Christian Aid (2008) calculates a tax rev- enue loss of US$160 billion suffered by developing countries in 2008 because of trade mispricing. The mispricing approach as employed in these studies has the advan- tage of simplicity and transparency. It uses publicly available data, and it is straightforward to replicate the results of existing studies. Unfortu- nately, the results of this type of analysis are dif�cult to interpret, and they effectively reveal little reliable information about income shifting in the context of tax avoidance and evasion. This is so for the following reasons. First, it is likely that, to some extent, price differences within product groups simply reflect quality differences. If there are price differences within product groups, it would be natural to assume that developing countries tend to export low-end, low-price products, whereas devel- oped countries export high-end products at higher prices. Chinese exports are an example of this pattern, as recently demonstrated by Schott (2008). How this affects the results of income shifting calcula- tions depends on whether or not the trade volumes of different coun- tries in a given product group are considered jointly to identify mispric- ing. If they are considered jointly and if the quality pattern is as described above, the mispricing approach systematically overestimates income shifting from developing to developed countries. If they are considered separately, this cannot happen, but, in this case, goods that are classi�ed as overpriced in one country may be counted as underpriced in another country. This is inconsistent. As long as it is not possible to disentangle quality differences and income shifting, the interpretation of numbers generated by the mispricing approach is dif�cult. Second, identifying the highest and the lowest quartile of observed prices as abnormal prices implies that any price distribution with some variance would be diagnosed to include overpricing and underpricing, even if the observed price differences are small or are driven by factors other than mispricing. Empirical analysis should normally allow for the possibility that a hypothesis—in this case, the hypothesis that income is shifted from developing to developed countries—is not supported by the data. This is excluded by assumption, unless all prices within a com- modity group are identical. Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 117 Third, the results are dif�cult to interpret because the counterfactual is not clear. Assume that, in one period, there is only one transaction in the upper quartile price range and only one transaction in the lower quartile price range. All other transactions are priced below the upper quartile price and above the lower quartile price. In this case, the coun- terfactual, which is a hypothetical situation without mispricing, should be that the two mispriced transactions disappear or their prices are adjusted to within the inner quartile price range. But now assume that, in the next period, the two transactions identi�ed as mispriced in the �rst period take place at corrected prices, which are between the upper and lower quartile prices identi�ed for the preceding period; everything else remains the same. In this case, the quartile price ranges for the sec- ond period would change, and transactions that were not identi�ed as mispriced in the previous period are now identi�ed as mispriced. This inconsistency occurs because there is no well-de�ned counterfactual. Fourth, the price �lter method, by construction, identi�es overpriced and underpriced transactions, so that it always identi�es income shifting in two directions: into and out of the country under consideration. Yet, many studies using the approach only report income shifting in one direction and ignore income shifting in the other direction. For instance, Pak (2007) reports underpriced imports into the United States and over- priced U.S. exports, but overpriced imports and underpriced exports (both of which would shift income out of the United States) are neglected. A similar approach is used by Christian Aid (2009) and other studies. The restriction to one direction in income shifting leads to highly misleading results if the �ndings are used to estimate the impact of income shifting on corporate income tax revenue collected by a particu- lar country or group of countries such as in Christian Aid (2008, 2009). A meaningful estimate of the tax revenue effects would have to take into account pro�t shifting in both directions. Consider the following simple example. Assume that there are three exporters of a good in country A. Firm 1 exports the good at a price of 4; �rm 2 exports the good at a price of 8; and �rm 3 exports the good at a price of 12. The mispricing approach would identify the transaction at a price of 4 as underpriced and the transaction at a price of 12 as overpriced. Assume, further, that all �rms have costs of 4 in country A that are deductible from the pro�t tax base 118 Draining Development? in country A. The goods are exported to country B, where all three are sold at a price of 14 to consumers in country B. In this example, the aggregate corporate income tax base in country A is equal to 12. Firm 1 shifts income out of country A, and �rm 3 shifts income into country A. In the absence of trade mispricing, the tax base in country A would be the same. The tax revenue loss of country A because of mispricing is equal to 0. A method that only takes into account �rm 1 and neglects the implications of mispricing by �rm 3 is clearly misleading. The same applies to the impact of income shifting on coun- try B. For illustrative purposes, one might, for instance, consider a devel- oping country with a weak political system and a low corporate tax rate. While some �rms may be willing to shift pro�ts into that country to exploit the low corporate tax rate, others may consider it bene�cial to transfer pro�ts out of the country to hedge against expropriation risks. This might give rise to the heterogeneous transfer price distortions laid out above. While we do not necessarily want to suggest that income is, in reality, shifted into developing countries, we nonetheless consider that an empirical identi�cation approach should allow for the possibility that this might take place. This hypothesis could then be rejected as a result of the analysis. What happens if other taxes are considered? For instance, one might be interested in measuring import duty revenue losses that arise because of avoidance or evasion. Assume that there is a proportional import duty in country B. In this case, it is easy to check that tax revenue is the same in the two cases under consideration, and it is of key importance to take into account both under- and overpriced imports to country B. Note also that, in this case, �rms would have an incentive systematically to understate the import price. If this happens, income is shifted into the country. This is another reason why neglecting income shifting into developing countries is not appropriate in the context of studies on tax avoidance and evasion. Depending on the question asked, it may be meaningful to consider either net flows or gross flows, but reporting flows in one direction only and ignoring the flows in the other direction are not appropriate and may easily lead to misunderstandings. Pak (2007) defends his approach by claiming that the price �lter method he uses is also applied by the U.S. Internal Revenue Service (IRS regulation 482) to deal with transfer pricing issues. This is not correct. Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 119 IRS regulation 482 stipulates that this method can only be applied to uncontrolled transactions (see section 1.482-1e, iii [C]), which means that only transactions between unrelated parties can be taken into account in assessing whether or not a transfer price is acceptable. Trans- actions within multinational �rms must be excluded. The reason is that transactions between unrelated parties are more likely to reflect undistorted prices. Effectively, the IRS approach compares transactions between unrelated �rms to transactions between related �rms. In con- trast, Pak (2007) applies the price �lter method to all transactions, includ- ing transactions between related parties. This is fundamentally different. An approach that uses a method consistent with IRS regulation 482 is applied in a study of trade mispricing by Clausing (2003). This study focuses on U.S. external trade, however, not on developing countries in particular. Clausing (2003) compares the prices in trade transactions between related parties with the prices in transactions between unre- lated parties and shows that the differences are signi�cantly influenced by tax rate differences. For instance, in transactions between related par- ties, the prices of exports from low-tax countries to high-tax countries are higher than the prices in transactions between unrelated parties. This suggests that multinational �rms try to reduce the taxes they have to pay by manipulating transfer prices. Of course, one could argue that mispricing is also likely to occur in transactions between unrelated parties. For instance, an exporter located in a low-tax country and an importer in a high-tax country could agree to increase the price of the transaction. This agreement could include a side payment that the exporter makes to the importer. Such a payment would have to be concealed from the tax authorities. If this happens, the approach used in Clausing (2003) systematically understates the impact of tax differences on pro�t shifting through transfer pricing. However, a mere price manipulation in a transaction between related parties is much easier than a price manipulation, combined with a concealed side payment by which the importer would participate in the tax savings. Pro�tability and pro�t shifting As mentioned above, a second approach to measuring income shifting directly considers the pro�tability of �rms and asks whether the pro�t- ability pattern observed may be explained as a result of income shifting. 120 Draining Development? As in the case of mispricing, the insights provided by this type of study depend on how pro�t shifting is identi�ed. Oxfam (2000) estimates that tax revenue losses arising because of corporate pro�ts shifted out of developing countries are equal to US$50 billion per year. This number is calculated as follows. Oxfam multiplies the stock of foreign direct invest- ment (FDI) in developing countries (US$1.2 trillion in 1998, according to UNCTAD 1999) by a World Bank estimate for the return on FDI of 16 to 18 percent in developing countries. Oxfam argues that the true esti- mate for the return on FDI is even higher since the World Bank �gure does not account for pro�t shifting activities. Thus, they set the rate of return to 20 percent. Next, the paper assumes an average tax rate of 35 percent and thus derives a hypothetical corporate tax payment of around US$85 billion. Since the actual tax payments received are around US$50 billion, this leaves a tax gap of US$35 billion according to Oxfam. Oxfam augments this �gure with revenue forgone because of the evasion of income from �nancial assets held abroad, which is estimated at US$15 billion. This leads to the estimated tax revenue losses of US$50 billion. This approach raises a number of questions. First, one may question the accuracy of the tax gap estimates. An important weakness of this calculation is the assumption that, with perfect compliance, all income from FDI would effectively be taxed at a rate of 35 percent. The issue is not that the average headline corporate tax rate may have been closer to 30 percent, as Oxfam (2000) recognizes. The key issue is that this approach neglects the existence of tax incentives for corporate invest- ment. In the developing world, these investment incentives play a much larger role than they do in developed countries (Klemm 2009; Keen and Mansour 2008). Many developing countries use tax incentives, for example, tax holidays or free economic zones that offer low or zero cor- porate taxes, to attract foreign investment. It is controversial whether these incentives are ef�cient from a national or global welfare point of view (see OECD 2001), but the related revenue impact should be distin- guished from the impact of tax avoidance and evasion. Neglecting this implies that the revenue losses caused by evasion and avoidance are overestimated. Moreover, the return on FDI is not identical with the cor- porate tax base. For instance, if FDI is �nanced by debt, it cannot be expected that the contribution of the investment to the corporate tax Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 121 base will be 20 percent of this investment because interest is deductible from the tax base. Second, one should note that this type of analysis is purely descriptive and does not investigate the factors driving pro�t shifting. Thus, existing tax gap estimates do not help to explain why multinational �rms may have an incentive to transfer pro�ts out of developing economies. Taxa- tion may not be the main factor that causes income shifting out of the developing world. Other factors such as the threat of expropriation or con�scation of private property, economic and political uncertainty, �s- cal de�cits, �nancial repression, or devaluation may be the real driving forces, as pointed out by, for example, de Boyrie, Pak, and Zdanowicz (2005). The implementation of effective policy measures against pro�t shifting from developing countries requires knowledge about the main motivations and incentives behind this type of activity and is thus an important topic for future research. As mentioned above, there is a growing literature that investigates the role played by taxation as a factor driving income shifting. Unfortu- nately, most of this work focuses on OECD countries, rather than devel- oping countries. Grubert and Mutti (1991) analyze pro�t shifting among U.S. multi- national �rms and use a data set that includes developed and developing countries. They show that �rms systematically report higher taxable pro�ts in countries with lower tax rates. In their analysis, �rms in coun- tries with a tax rate of 40 percent would report an average ratio of pretax pro�t and sales of 9.3 percent, whereas �rms in countries with a tax rate of 20 percent would report a pro�t and sales ratio of almost 15.8 percent. This suggests that some pro�t shifting that is motivated by taxation does occur. Azémar (2008) investigates how the effectiveness of law enforcement affects pro�t shifting. Interacting a summary measure of law enforce- ment quality and tax pro�t ratios of �rms in her regressions, she �nds that a low quality in law enforcement accompanies a high sensitivity in tax payments to corporate tax rates. Her interpretation of this observa- tion is that countries with ineffective law enforcement face greater dif�- culties in ef�ciently implementing anti–tax avoidance measures such as thin capitalization rules or transfer pricing corrections. This suggests 122 Draining Development? that developing countries are more vulnerable than developed countries to income shifting. The Role of Tax Havens Tax havens are widely viewed as playing a major role in the tax avoid- ance and tax evasion by multinational �rms, as well as by individual taxpayers. Empirical research on income shifting to tax havens encoun- ters the dif�culty that data on economic activity are scarce in these places. Nonetheless, there are studies on tax avoidance and evasion in tax havens and on the impact of tax havens on tax revenue collection by other countries. Unfortunately, these studies usually do not focus on developing countries. First, NGOs have made estimates of the tax revenues forgone because of the existence of tax havens. But these estimates are partly related to the potential impact of tax havens on tax rates set by other countries. For instance, Oxfam (2000) estimates that developing countries may be los- ing annual tax revenues of at least US$50 billion as a result of tax com- petition and the use of tax havens. It argues as follows: Tax competition, and the implied threat of relocation, has forced devel- oping countries to progressively lower corporate tax rates on foreign investors. Ten years ago these rates were typically in the range of 30–35 percent, broadly equivalent to the prevailing rate in most OECD coun- tries. Today, few developing countries apply corporate tax rates in excess of 20 percent. Ef�ciency considerations account for only a small part of this shift, suggesting that tax competition has been a central consider- ation. If developing countries were applying OECD corporate tax rates their revenues would be at least US$50 billion higher. (Oxfam 2000, 6) The issue here is that the extent is not clear to which the decline in corporate income tax rates that has occurred in both developing and developed countries is caused by tax havens. Tax rate competition would exist even in the absence of tax havens. In addition, some authors argue that, under certain circumstances, tax havens may reduce the intensity of tax competition (Hong and Smart 2010). Of course, the empirical rele- vance of this analysis remains to be investigated. Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 123 However, the main critique of tax havens is not that they force other countries to cut taxes. It is the perception that these tax havens offer opportunities for tax avoidance or tax evasion to multinational �rms and individual taxpayers residing in other countries, so that other coun- tries suffer tax revenue losses. One estimate of tax revenue losses because of the existence of tax havens has been published by the Tax Justice Net- work (TJN 2005). TJN starts with estimates of global wealth in �nancial assets published by banks and consultancy �rms (for example, Capgem- ini and Merrill Lynch 1998; BCG 2003). This is combined with estimates (for 2004) by the Bank for International Settlements of the share of �nancial assets held offshore, though this refers to U.S. asset holdings.1 Based on these numbers, TJN claims that the offshore holdings of �nan- cial assets are valued at approximately US$9.5 trillion. This is augmented by US$2 trillion in non�nancial wealth held offshore in the form of, for example, real estate. TJN thus estimates that, globally, approximately US$11.5 trillion in assets are held offshore. Assuming an average return of 7.5 percent implies that these offshore assets yield US$860 billion. TJN assumes that these assets are taxable at 30 percent and thus calcu- lates a revenue loss of US$255 billion per year (as of 2005). TJN does not attempt to estimate which part of these revenue losses occur in developing countries. Cobham (2005) uses the TJN (2005) results and estimates the share of developing countries. Assuming that the 20 percent of worldwide gross domestic product that is accounted for by middle- and low-income countries represents a credible share of the offshore wealth holdings of the developing world, he �nds that 20 percent of the revenue loss can be assigned to these countries, that is, US$51 billion. Other estimates of these revenue losses use similar methods. Oxfam (2000) calculates the revenue losses caused by the evasion of taxes on income from �nancial assets held abroad at around US$15 billion per year. This result is mainly driven by an estimate of the foreign asset hold- ings of residents in developing countries in 1990 (US$700 billion), which is now outdated. In a more recent study, Oxfam (2009) estimates that US$6.2 trillion of developing-country wealth is held offshore by indi- viduals. This leads to an estimated annual tax loss to developing coun- tries of between US$64 billion and US$124 billion. 124 Draining Development? It is dif�cult to interpret these estimates for tax revenue losses because of offshore wealth holdings. They do not unambiguously over- or under- estimate the revenue losses, but they rely on a large number of strong assumptions. These include the assumptions on the distribution of asset holdings across the developed and the developing world, as well as the taxable rates of return and the average tax rates. In addition, there are several open questions that have to be addressed. First, it is unclear whether all income from offshore wealth holdings is taxable on a resi- dence basis. Some developing countries do not tax the foreign source income of residents because it may not be administratively ef�cient to do so (Howard 2001). Second, even if savings income is taxable on a resi- dence basis, the taxes paid in the source country may be deductible from the taxes owed in the residence country. Third, it is unlikely that all income from �nancial assets held offshore evades taxation in the coun- try of residence of the owners. There may be other than tax reasons for offshore holdings of �nancial assets, and it is possible that the owners of these assets declare their incomes in their countries of residence. To the extent that these assets generate passive investment income, they will also be subject to controlled foreign corporations legislation, which means that this income is excluded from the deferral of home country taxation or the exemption granted to active business income. Additional research is needed to determine, for example, the causal effect that the presence in tax haven countries of multinational af�liates has on the tax revenues paid by these af�liates in the developing world. Such an investigation should follow work by Desai, Foley, and Hines (2006a, 2006b) and Maf�ni (2009). These papers study the role of tax havens for the European Union and the United States. In the next sec- tion, we discuss methods and data sets that may be used to do similar work on developing countries. Data Sets and Identi�cation Strategies to Assess Tax Evasion and Avoidance in Developing Countries Attempts to assess whether and to what extent (multinational) �rms in developing countries engage in international tax evasion and tax avoid- ance activities have long been hampered by a lack of appropriate data. Thus, existing evidence on the issue is largely anecdotal with the excep- Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 125 tion of a small number of studies reviewed in the previous section. As many of these papers face methodological dif�culties and their identi�- cation strategies partly rely on strong assumptions, more research is needed to identify and quantify corporate tax avoidance and evasion in the developing world. In recent years, a number of data sets have become available that may suit this purpose. In the following, we review a selec- tion of these databases and discuss potential identi�cation strategies. The analysis focuses on microdata sources because we consider these more well suited to the identi�cation of corporate pro�t shifting activi- ties than macrolevel data given that the former are less prone to endoge- neity problems in the estimation strategy. Identi�cation strategies The basis for a valid empirical identi�cation strategy is the development and testing of hypotheses that derive effects that are unlikely to capture activities other than corporate pro�t shifting. In the following, we dis- cuss two identi�cation strategies that, from our point of view, largely ful�ll this requirement and that may be applied to microlevel data. The �rst identi�cation strategy builds on the notion that companies in developing countries differ with respect to their flexibility and oppor- tunities to shift income out of the host countries. For example, compa- nies that are part of a multinational group can plausibly undertake pro�t shifting activities more easily than �rms without af�liates in foreign countries. This is because they can transfer pro�ts to an af�liated com- pany abroad. Moreover, �rms that belong to multinational groups with tax haven af�liates have particularly good opportunities to transfer income out of developing countries. Thus, they can be expected to engage in even larger pro�t shifting activities. Consequently, the identi- �cation strategy is to compare pro�t shifting measures for the treatment group of multinational �rms (with tax haven connections) to a control group of national �rms that are expected not to engage in signi�cant pro�t shifting activities. To identify pro�t shifting activities in this context, researchers need detailed information on corporate ownership structures and on com- pany variables that are expected to capture pro�t shifting activities. According to previous empirical studies on corporate shifting activities in the industrialized world, multinational �rms use different channels to 126 Draining Development? transfer taxable resources out of countries, the most important ones being the distortion of intra�rm trade prices and the debt-equity struc- ture, as well as the relocation of pro�table assets such as corporate pat- ents (Clausing 2003; Huizinga and Laeven 2008; Buettner and Wamser 2007; Dischinger and Riedel 2008). Testing for this type of pro�t shifting therefore requires associated information on trade prices, debt levels, and patents. As this is often not available, researchers may also exploit information on corporate pretax pro�ts and corporate tax payments because pro�t shifting outflows are expected to lower both variables. Thus, following the above identi�cation strategy, we expect that multi- national �rms in developing countries (especially those with a tax haven connection) should report lower pretax pro�ts per unit of assets, pay lower taxes per unit of assets and per unit of pro�t, hold higher fractions of (intra�rm) debt, and exhibit stronger distortions of intra�rm trade prices (that is, enlarged import prices and diminished export prices) than the control group of national �rms. The obvious challenge of this identi�cation strategy is to account empirically for a potential selection of �rms with differing characteristics into the control group (national �rms) and the treatment group (multinational �rms [with a tax haven connection]). Strategies to solve this problem have been presented in earlier papers for the developed world (for example, see Desai, Foley, and Hines 2006a, 2006b; Maf�ni 2009; Egger, Eggert, and Winner 2007). If, after accounting for all these issues, no differences between the consid- ered pro�t shifting variables for national and multinational �rms are found, the pro�t shifting hypothesis is rejected. A second identi�cation strategy starts with a question: why do com- panies shift pro�ts out of the developing world? One motive might be that they want to save on tax payments. A second might be that they draw their money out of corrupt and politically unstable countries where they are prone to threats of expropriation. To test for these hypotheses implies to determine whether companies in countries with a high tax rate or a high corruption rate report lower pretax pro�ts per assets, pay lower taxes on their assets and pro�ts, and show higher debt- to-equity ratios and more strongly distorted intra�rm transfer prices. In this case, the challenge is to ensure that the identi�ed effect between taxation-corruption and the pro�t shifting measures is not driven by an unobserved heterogeneity of �rms that are located in high-tax (high- Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 127 corruption) countries and low-tax (low-corruption) countries. This requires the inclusion of a set of control variables that capture differ- ences between af�liates and host countries. The most convincing approach involves including a set of af�liate �xed effects into a panel data regression, which implies that the researcher accounts for all time- constant af�liate differences, and the identi�cation is achieved via cor- porate adjustments to changes in taxes or the corruption level. For this identi�cation strategy to be applied, stringent data requirements have to be met. Ideally, accounting data on pretax pro�ts, tax payments, debt levels, or intra�rm transfer prices should be available in panel format for several years; the same should be true of information on corporate taxes and the level of corruption and political stability. Data sets A number of data sets may ful�ll the requirements associated with the two identi�cation strategies laid out above. In the following, we present a selection of these databases. In a �rst step, we discuss one of the data sources, Orbis, in some depth. In a second step, we briefly describe alter- nate databases available for the purpose of testing tax avoidance in the developing world. The tables referenced in this discussion are contained on the website of this volume.2 Orbis. The Orbis data provided by Bureau van Dijk contain information on companies worldwide (see table 4A.1 on the website for a description of the data). While the majority of �rms in Orbis are located in industri- alized economies, the data also include information on countries in the developing world. Orbis is available in different versions, but the largest version of the data set comprises 85 million �rms (as of 2011), and new data are added constantly. The data are collected from various (partly private and partly of�cial) sources that may differ across countries. Con- sequently, it is a well-known problem of the Orbis data that the �rm coverage differs across countries, and some economies are poorly repre- sented in relative terms. This problem is especially pronounced in devel- oping countries. While the �rm coverage tends to be particularly limited in Africa, information on a suf�ciently large number of �rms is reported for several economies in Asia and Latin America. Hence, we think that focusing on those developing economies for which good information is 128 Draining Development? available allows the Orbis data to be used for the purpose of identifying corporate pro�t shifting behavior. Orbis provides balance sheet information and data on pro�t and loss account items. Thus, it contains detailed information on pretax and after-tax pro�t (both operating pro�ts and �nancial, plus operating pro�ts), as well as tax payments and debt variables. Moreover, informa- tion on ownership links is included, especially links to all direct and indirect shareholders of the �rm, as well as to subsidiaries within the multinational group. Finally, Orbis provides address information (postal code and city name), which may allow researchers to determine the loca- tion of �rms (or at least the headquarters) within developing countries and thus to identify, for instance, companies located in free economic zones or to determine the role of taxes levied at the regional level. It is important to stress that, as with other databases provided by private sources, there may be issues regarding the quality of the Orbis data. We believe these problems can be handled by rigorous plausibility checks and data cleaning. In the following, we present and discuss descriptive statistics on some countries in the Orbis data. Note that the Orbis version available to us contains large �rms only. Analyzing the full Orbis version, which also accounts for smaller �rms, is likely to enhance the �rm coverage com- pared with our exercise. Moreover, we restrict the analysis to Asia and only employ data for countries with a certain suf�cient threshold of �rm coverage. The economies included in the sample are China; India; Indo- nesia; Malaysia; Pakistan; the Philippines; Taiwan, China; and Thailand. Our �nal data set is a cross section of 87,561 �rms for the year 2006 (see table 4A.2 on the website for the country distribution). Because the Orbis data provide information on ownership connec- tions among �rms, they allow us to pursue the �rst identi�cation strat- egy described in the previous section. We establish the de�nition that a �rm in a developing country maintains a direct link to a foreign econ- omy if it directly owns a subsidiary in a foreign country through posses- sion of at least 50 percent of the ownership shares or if it is directly owned by a parent �rm in a foreign country through a holding of at least 50 percent of the ownership shares. Moreover, we adopt a second, less- restrictive de�nition of a multinational �rm that applies if any af�liate within the multinational group is located in a foreign country (including Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 129 subsidiaries of the immediate or ultimate global owner that do not have a direct ownership link to the �rm under consideration). According to these de�nitions, 2,202 �rms in our data set exhibit a direct ownership link to a foreign af�liate, and 2,807 �rms belong to a multinational group in the broader sense that at least one af�liate in the group is located in a foreign country (irrespective of the existence of either a direct or an indirect ownership link). We face the potential challenge that information on some af�liates within the multinational group is missing in the data. If missing af�liates are located in foreign countries (from the point of view of the �rm under consideration), then we might declare that a corporation is a national �rm, although it is actually part of a multinational group. Such a mis- classi�cation introduces noise into our analysis and may be expected to bias the results against us. This implies that the results should be inter- preted as a lower bound to the true effect. We undertake a similar exercise to identify �rms with ownership links to af�liates in a tax haven country. A tax haven is identi�ed according to the OECD’s list of tax havens. We establish the de�nition that a group of �rms has a direct ownership link to a tax haven country if they directly own a tax haven subsidiary through at least 50 percent of the ownership shares or are directly owned through at least 50 percent of the ownership shares by a foreign parent �rm in a tax haven. In a second step, we adopt a broader de�nition of tax haven links that identi�es �rms in our data that belong to multinational groups with a tax haven af�liate (irrespec- tive of the existence of a direct or indirect ownership link). According to these de�nitions, 207 �rms in our data show a direct ownership link to a tax haven country, while 691 corporations belong to groups that have a tax haven af�liate (directly or indirectly connected to the �rm under consideration). Thus, according to the broader de�nition, 25 percent of the multinational �rms in our sample have an ownership link to a tax haven country, while only 9 percent of the �rms have a direct ownership link to a foreign tax haven. We might face the problem that information may not be available for all af�liates belonging to a multinational group. Because a large fraction of multinational �rms based in industrialized countries are known to operate subsidiaries in tax haven countries, we run a cross-check on the data and restrict the analysis to �rms that are owned by immediate and 130 Draining Development? ultimate global owners (in foreign countries) and on which information on the subsidiary list of the owners is available. Among this subgroup of �rms, we �nd that 63 percent belong to multinational groups with a tax haven af�liate. One should note that missing information on tax haven connections introduces noise into the estimation and biases our results against our working hypothesis, meaning that it would lead us to under- estimate the impact of tax haven presence. Thus, we compare different corporate variables that are expected to capture pro�t shifting activities among the subgroup of �rms as de�ned above. The Orbis data contain information on a wide range of accounting variables, including unconsolidated corporate pretax pro�t, tax payments, and debt levels. In table 4A.3 on the website, we provide descriptive statis- tics that discriminate among, �rst, all �rms in the data; second, �rms belonging to groups that own af�liates in a foreign country; third, �rms with a direct ownership link via a parent �rm or a subsidiary to a foreign country (this is thus a subset of the second group of �rms); fourth, �rms belonging to a multinational group with an af�liate in a tax haven coun- try; and, �fth, �rms having a direct ownership link (via a direct parent �rm or subsidiary) in a tax haven country. The rationale behind investi- gating groups with direct ownership links to foreign countries and tax havens separately is that group af�liates connected through direct owner- ship are presumed to be more closely tied in an economic sense. This is expected to facilitate pro�t shifting between the entities. As table 4A.3 shows, the �rms included in the analysis show average total asset investments of US$23.3 million. Multinational �rms possess larger asset stocks, with US$98.1 million and US$94.9 million for �rms with any link to a foreign af�liate and �rms with a direct link, respec- tively. Moreover, among the multinational �rms, corporations with a tax haven link are reported to have higher total asset investments than other multinational �rms: at US$140.5 million for �rms with any link to a tax haven country and US$172.3 million for �rms with a direct link to a tax haven. All the differences are statistically signi�cant as indicated by the 95 percent con�dence intervals around the mean. Table 4A.3 also presents the pretax pro�t per total assets reported by the companies in our sample. This may be considered a proxy for the corporate tax base of the �rms. The average pretax pro�tability of the �rms in our sample is estimated at 0.092. Multinational �rms (irrespec- Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 131 tive of direct or indirect ownership links to foreign countries) show a lower pretax pro�tability, at 0.071, on average, which is thus signi�cantly lower than the average pretax pro�tability of the full sample as indicated by the 95 percent con�dence interval around the mean. This result may seem counterintuitive at �rst sight because multinational �rms com- monly show larger productivity rates than national �rms, which suggests that they also report larger pretax pro�tability values (Helpman, Melitz, and Yeaple 2004). However, as indicated above, they may equally encoun- ter more opportunities to engage in tax avoidance and tax evasion through international channels, and this may lower the reported pretax pro�ts in our sample countries (that do not comprise any tax haven). Or, alternatively, the higher international mobility of their investments may endow them with greater bargaining power with respect to host country governments and allow them to obtain a lower tax base relative to less mobile national �rms. However, �rms belonging to multinational groups that include tax haven af�liates do not report signi�cantly lower pro�tability rates than national �rms. Because multinational �rms with a link to tax havens are presumed to face more opportunities for tax avoidance and evasion and probably also exhibit more willingness to take up these opportunities, this suggests that the operations of multinationals with a tax haven con- nection are more pro�table than the operations of other multinational �rms. This picture prevails if we restrict the pro�tability variable to rates above 0. Table 4A.3 also shows the tax payments per total assets reported by the �rms in our sample. The broad picture resembles the picture for the pro�tability rates. While national �rms pay the highest taxes per total assets reported, at an average of 0.018, the tax payments per total assets of multinational �rms are signi�cantly smaller, at 0.015 and 0.014, respectively, for �rms belonging to multinational groups in general and �rms with a direct ownership link to a foreign country. The subgroup of multinational �rms with a link to tax havens does not make signi�cantly lower tax payments per total assets than national �rms. (This outcome is likely driven by the pro�tability pattern discussed in the paragraphs above.) Furthermore, we report descriptive statistics on tax payments per pretax pro�t, which are a proxy for the average tax rate of the observed 132 Draining Development? �rms. The results suggest that national �rms face the highest average tax rate, at 20.0 percent, which is signi�cantly higher than the average tax rate of multinational �rms, at 16.9 and 16.4 percent, respectively, for �rms belonging to multinational groups in general and �rms with a direct ownership link to a foreign country. Among multinationals, the lowest average tax rate is experienced by �rms belonging to multina- tional groups with a link to tax havens. These �rms pay 13.2 and 11.2 percent taxes on their pro�ts, respectively, for �rms with any tax haven link and �rms with a direct tax haven link, which is signi�cantly lower than the rates faced by multinational �rms in general and, thus, also signi�cantly lower than the rates faced by national �rms. This suggests that �rms with a tax haven connection manage to reduce their corporate tax burden signi�cantly. To account for the fact that the characteristics of �rms belonging to multinational groups with a link to tax havens might differ in ways that may determine the described pro�t shifting measures, we also run a set of regressions to attempt to control for some of the potential sources of heterogeneity. Table 4A.4 on the website presents the results of a simple ordinary least squares model that regresses the pretax pro�tability (de�ned as pretax pro�ts over total assets) of the �rms in our sample on two dummy variables, which indicate entities belonging to multina- tional groups and entities belonging to groups with tax haven af�liates. In speci�cations (1) to (4), these de�nitions require the �rm to have a direct ownership link to a foreign �rm and a tax haven af�liate, respec- tively, while the multinational and tax haven de�nitions in speci�cations (5) to (8) also allow for indirect connections. Speci�cation (1) presents the regression results without any control variables, which derives �nd- ings that are analogous to the descriptive statistics in table 4A.3. The results suggest that �rms belonging to multinational groups have signi�- cantly lower reported pretax pro�ts per total assets than national �rms, whereas the pretax pro�tability of �rms belonging to multinational groups with a tax haven af�liation does not statistically differ from the corresponding pro�tability of national �rms. In speci�cations (2) to (4), we include additional control variables within the estimation framework to account for heterogeneity in other �rm characteristics. Thus, in speci�cation (2), we add a full set of coun- try �xed effects that absorb time-constant heterogeneity in the pretax Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 133 pro�tability of �rms in various sample countries. This, for example, accounts for time-constant differences in accounting methods across our sample countries. It renders insigni�cant the coef�cient estimate for the dummy variable indicating multinational �rms and suggests that the pretax pro�tability of multinational �rms generally and of multina- tional �rms with tax haven connections in particular does not differ from the pretax pro�tability of national entities. In speci�cation (3), we add a full set of two-digit industry dummies to account for heterogene- ity in the pro�tability ratios of different industries; this does not change the results. Finally, speci�cation (4) additionally controls for the fact that pro�tability rates may vary according to �rm size; it includes the loga- rithm of the total assets of �rms as an additional control variable. The coef�cient estimate for the size effect turns out to be negative and statis- tically signi�cant at the 1 percent level. This indicates decreasing returns to scale, that is, in our sample, large �rms tend to report pretax pro�t- ability rates that are lower than the corresponding rates among small �rms. Accounting for this, in turn, derives positive and statistically sig- ni�cant coef�cient estimates for the dummy variables indicating multi- national �rms (with tax haven connections). Hence, conditional on their size, multinational �rms (with tax haven connections) report larger pre- tax pro�tability than their national counterparts, which is in line with previous evidence in the literature that suggests the higher produc- tivity—and, in consequence, the higher pro�tability—of multinational �rms. Speci�cations (5) to (8) reestimate the regression model to account for multinational �rm and tax haven de�nitions that capture direct and indirect ownership links. This derives comparable results. In table 4A.5 on the website, we repeat the exercise and determine the connection between multinational ownership links (to tax haven af�li- ates) and the corporate tax payments per assets. Analogously to the pre- vious table, speci�cation (1) regresses the corporate tax payment ratio on dummy variables indicating direct ownership links to foreign �rms (in tax haven countries). The results indicate that multinational �rms, in general, tend to pay signi�cantly lower taxes on their total asset stock than national �rms, while the tax payments of multinational �rms with a tax haven connection do not signi�cantly differ. This result prevails if we account for a full set of country �xed effects and industry �xed effects in speci�cations (2) and (3). In speci�cation (4), we additionally include 134 Draining Development? a size control (the logarithm of total assets). Similar to the results in the previous table, the coef�cient estimate for the size variable is negative and signi�cant, suggesting that the tax payments per total assets decrease with �rm size. Moreover, conditional on �rm size, the speci�cations indicate that multinational �rms in general, but especially those with a tax haven connection report larger tax payments per total assets than national �rms. This result may arise because multinational �rms tend to show larger underlying productivity and, hence, earn greater pro�ts per total assets, which lead to larger tax payments. Last, we assess the difference in the average effective tax rate of national and multinational �rms (with a tax haven connection) as mea- sured by tax payments over the pretax pro�ts of the �rms (see table 4A.6 on the website). To do this, we restrict our sample to �rms that show both a positive pretax pro�t and nonnegative tax payments. Speci�ca- tion (1) regresses the average tax burden on dummy variables indicating multinational �rms (with tax haven connections) as determined by direct ownership links. The regression results suggest that multinational �rms, in general, are not subject to a lower average tax rate than their national counterparts, while multinational corporations with tax haven links report an average tax rate that is lower by 5 percentage points. Additionally, controlling for a full set of country �xed effects and indus- try �xed effects equally renders the coef�cient estimate for the multi- national dummy negative and statistically signi�cant, suggesting that multinational �rms in general and multinational �rms with a tax haven connection pay lower taxes on their reported pretax pro�ts compared with national �rms. This result is, moreover, robust against the inclusion of a size control in the model, as presented in speci�cation (4). Quanti- tatively, the results suggest that multinational �rms pay 1 percentage point lower taxes on pretax pro�t than national corporations, while multinational �rms with a tax haven connection pay 4.4 percentage points lower taxes on their pro�t. This means that the average effective tax rate of multinational �rms with a tax haven connection is around 3.4 percentage points lower than the average effective tax rate of multina- tional �rms without a tax haven connection. In speci�cations (5) to (8), we rerun the regression to account for indirect ownership links in the de�nition of the multinational dummy (with tax haven connections) and �nd comparable results. Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 135 Thus, summing up, the results suggest that multinational �rms tend to report higher pretax pro�ts and tax payments per assets than compa- rable national �rms, which may mean that they have a higher underlying productivity level. However, we also �nd that multinational �rms, espe- cially those with a tax haven connection, face a signi�cantly lower aver- age tax burden, that is, they make lower average tax payments per pretax pro�ts. Note that some caution is warranted in interpreting these results. As discussed above, an in-depth analysis requires that we account for the selection of different �rms in the groups of national corporations, multi- nationals, and multinationals with an ownership link to avoid producing results that are driven by unobserved heterogeneity between groups. In interpreting the results, one should also keep in mind that the majority of information exploited is derived from �rms in China. Future work should aim to run analyses on a broader basis, including information on other countries in Africa, Asia, Eastern Europe, and South America. One important issue is also to assess to what extent the difference in the aver- age tax rates of national �rms and multinationals (with a tax haven con- nection) is driven by the fact that multinationals tend to bene�t from the location in special economic zones or receive special tax breaks from governments through other avenues. This is methodologically feasible as demonstrated by previous studies on the industrialized countries (Desai, Foley, and Hines 2006a, 2006b; Maf�ni 2009; Egger, Eggert, and Winner 2007). Moreover, it would be necessary to account for any time-varying differences in host country characteristics such as heterogeneity in accounting and tax base legislation. A rigorous analysis that accounts for these issues is beyond the scope of this chapter and is relegated to future research. The structure of the Orbis data would also allow one to pursue the second identi�cation strategy, that is, determine how changes in the cor- porate tax rate and corruption parameters affect pro�t shifting variables. This is possible because Orbis includes rich information on �rm and group characteristics that may be used as control variables, and Orbis is available in a panel structure that allows one to control for time- constant differences between af�liates. In the course of this chapter, we determine only the correlation between the corporate statutory tax rate (obtained from various sources) and a corruption index (obtained from 136 Draining Development? Transparency International) with the pro�t shifting variables named above, and we �nd small correlations only. This does not necessarily mean that �rms in the sample do not engage in pro�t shifting behavior, but it does indicate that pro�t shifting measures are determined by sev- eral factors that correlate with the tax rate and corruption indexes. These factors have to be accounted for in a regression framework to make meaningful statements about the effect of taxes and corruption on pro�t shifting behavior. This is left to future research. Other data sources. Apart from Orbis, several other data sets may be used to analyze tax evasion and tax avoidance in developing economies. A database that is comparable with Orbis is Compustat Global, which is provided by Standard and Poor’s. The data encompass �rm-level infor- mation on balance sheet items and pro�t and loss accounts of compa- nies around the world. Thus, information on pretax pro�ts, corporate tax payments, debt levels, interest payments, and research and develop- ment expenditure is included, which allows one to identify corporate pro�t shifting and assess the importance of different pro�t shifting channels out of the developing world. In total, Compustat covers more than 30,000 companies in 100 countries, including several developing countries. The coverage is especially good in the Asia and Paci�c region, where information on almost 16,000 �rms is available (see table 4B.1 on the website for a description). The data have some drawbacks. First, they do not cover ownership information, that is, it is not possible to link subsidiaries and parent �rms in the data. Consequently, Compustat does not allow one to apply the �rst identi�cation strategy because foreign �rms and tax haven af�li- ates cannot be systematically identi�ed. However, because the data are available in panel format for several years, they allow one to pursue the second identi�cation strategy to examine how tax rate changes and changes in the corruption index affect pro�t shifting measures. A second drawback of the data is that information is only provided for companies that are listed on a stock exchange. This imposes a sample restriction. Nonetheless, large (listed) �rms are likely to be the main pro�t shifters, and, thus, pro�t shifting effects should still be identi�able. Third, as stressed above, the quality of the data sets provided by pri- vate institutions has been criticized in the past. Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 137 To address this last concern, researchers may consider using data sources provided by of�cial institutions that have become available recently and that allow tax avoidance and evasion to be investigated in the developing world. These data sets encompass information on out- ward investments by multinational �rms. The most widely known data sets of this sort are the Direct Investment and Multinational Compa- nies Database of the U.S. Bureau of Economic Analysis, the Deutsche Bundesbank’s Microdatabase on Direct Investment, and the U.K. Annual Inquiry into Direct Investment Abroad (respectively, see tables 4B.2, 4B.3, and 4B.4 on the website). All three data sets contain information on multinational parent �rms in the respective countries and their for- eign subsidiaries, including subsidiaries in the developing world. The main advantage of these data sources is that the reporting is mandatory by national law, which suggests that the quality of the reported informa- tion is high. Moreover, information on directly and indirectly held af�li- ates is available. Only the U.K. data set is restricted to information on directly held subsidiaries. Despite some limitations, the data sets thus allow for studies based on the �rst identi�cation strategy. Moreover, the data include several variables that capture pro�t shift- ing behavior such as information on after-tax pro�ts, tax payments, and debt ratios. Nonetheless, the data are somewhat less comprehensive than the data in Orbis or Compustat. Thus, both the German and the U.K. data sets include only information on company pro�ts after taxation, not on pretax pro�ts. This is a disadvantage because information on pretax pro�ts is important for the analysis of pro�t shifting. Unlike the U.K. and the U.S. data sets, the German data set also does not report any information on the tax bills of the foreign subsidiaries. Still, the data sets possess a major advantage: they include information on intra�rm lend- ing and intra�rm interest flows, which allows one to test for debt shifting activities between af�liates. Additionally, in the United States, researchers at the U.S. Treasury have access to con�dential U.S. �rm-level data that are not available to the general public. The data cover information on U.S. tax returns and include variables on the tax payments, pro�ts, and investments of U.S. multinationals in the United States and information on the income and tax payments of foreign-controlled companies, including subsidiaries in developing economies (see table 4B.5 on the website). Finally, table 4B.6 138 Draining Development? on the website lists other country-speci�c �rm-level data sets that con- tain information on the foreign activities of domestic companies. Data on international trade prices may be exploited to investigate pro�t shifting out of developing countries that takes place through trade mispricing by multinational �rms. Previous studies reviewed above use the United States Merchandise Trade Databases, which contain price data according to import-harmonized commodity codes and export- harmonized commodity codes.3 However, these data hardly allow for a clear identi�cation strategy to ferret out pro�t shifting behavior. Because the data do not discriminate trade between related and unrelated parties, our �rst identi�cation strategy is not applicable. This shortcoming can, however, be addressed by using a different data source. Thus, Clausing (2003) exploits data from the International Price Program of the U.S. Bureau of Labor Statistics, which publishes information on 700 aggre- gate export and import price indexes. The data differentiate trade between related and unrelated parties, which makes the control group approach feasible. Assuming that pro�t shifting via trade price distor- tions is negligible between third parties, the trade prices of intra�rm trade between multinational af�liates can be compared with trade pric- ing between third parties (see Clausing 2003). A different strategy for the development of a valid identi�cation strategy would involve exploiting the panel dimension of the trade price data; this has not been done in existing studies. In this way, one might investigate how trade prices are affected by policy reforms such as tax rate changes or changes in the level of law enforcement or political stability in partner countries, which implies that the countries without tax reforms would serve as a control group. Conclusion Tax avoidance and evasion through pro�t shifting out of developing countries represent an important and widely debated issue. Yet, rela- tively little is known about the magnitude of and the factors driving pro�t shifting in the developing world. The main reason is that the avail- able data are limited in terms of both quality and quantity. The results of most existing empirical studies on developing countries are dif�cult to interpret because the methods used to measure income shifting raise a Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 139 number of problems. We have suggested and discussed several data sets, some of which have become available recently, and methods to under- take more research on pro�t shifting in the developing world. While these approaches also have limitations, they do have the potential to improve our understanding of this important issue and to inform poli- cies directed at crowding back tax avoidance and evasion in the develop- ing world. Notes 1. The website of the Bank for International Settlements is at http://www.bis.org/. 2. See http://go.worldbank.org/N2HMRB4G20. 3. See USA Trade Online (database), U.S. Department of Commerce, Washington, DC, https://www.usatradeonline.gov/. References Azémar, C. 2008. “International Corporate Taxation and U.S. Multinationals’ Behav- iour: An Integrated Approach.� Discussion Paper (August), Department of Eco- nomics, University of Glasgow, Glasgow. Baker, R. W. 2005. Capitalism’s Achilles Heel: Dirty Money and How to Renew the Free- Market System. Hoboken, NJ: John Wiley & Sons. BCG (Boston Consulting Group). 2003. “Winning in a Challenging Market: Global Wealth 2003.� BCG, Boston. Buettner, T., and G. Wamser. 2007. “Intercompany Loans and Pro�t Shifting: Evi- dence from Company-Level Data.� CESifo Working Paper 1959, CESifo Group, Munich. Capgemini and Merrill Lynch. 1998. “World Wealth Report 1998.� Capgemini and Merrill Lynch Global Wealth Management, New York. Christian Aid. 2008. “Death and Taxes: The True Toll of Tax Dodging.� Christian Aid Report, May. Christian Aid, London. ———. 2009. “False Pro�ts: Robbing the Poor to Keep the Rich Tax-Free.� Christian Aid Report, March, Christian Aid, London. Clausing, K. A. 2003. “Tax-Motivated Transfer Pricing and US Intra�rm Trade Prices.� Journal of Public Economics 87 (9–10): 2207–23. Cobham, A. 2005. “Tax Evasion, Tax Avoidance and Development Finance.� QEH Working Paper 129, Queen Elizabeth House, University of Oxford, Oxford. http://www3.qeh.ox.ac.uk/pdf/qehwp/qehwps129.pdf. de Boyrie, M. E., S. J. Pak, and J. S. Zdanowicz. 2005. “Estimating the Magnitude of Capital Flight Due to Abnormal Pricing in International Trade: The Russia- USA Case.� Accounting Forum 29 (3): 249–70. 140 Draining Development? Desai, M. A., C. F. Foley, and J. R. Hines Jr. 2004. “Foreign Direct Investment in a World of Multiple Taxes.� Journal of Public Economics 88 (12): 2727–44. ———. 2006a. “Do Tax Havens Divert Economic Activity?� Economics Letters 90 (2): 219–24. ———. 2006b. “The Demand for Tax Haven Operations.� Journal of Public Econom- ics 90 (3): 513–31. Dischinger, M., and N. Riedel. 2008. “Corporate Taxes and the Location of Intangi- ble Assets within Multinational Firms.� Discussion Paper in Economics 5294, Department of Economics, University of Munich, Munich. Egger, P., W. Eggert and H. Winner. 2007. “Saving Taxes through Foreign Plant Own- ership.� CESifo Working Paper 1887, CESifo Group, Munich. Fuest, C., and N. Riedel. 2009. “Tax Evasion, Tax Avoidance and Tax Expenditures in Developing Countries: A Review of the Literature.� Report, Oxford University Centre for Business Taxation, Oxford. Grubert, H., and J. Mutti. 1991. “Taxes, Tariffs and Transfer Pricing in Multi- national Corporation Decision Making.� Review of Economics and Statistics 73 (2): 285–93. ———. 2000. “Do Taxes Influence Where U.S. Corporations Invest?� National Tax Journal 53 (4): 825–39. Helpman, E., M. J. Melitz, and S. R. Yeaple. 2004. “Export Versus FDI with Hetero- geneous Firms.� American Economic Review 94 (1): 300–16. Hong, Q., and M. Smart. 2010. “In Praise of Tax Havens: International Tax Planning and Foreign Direct Investment.� European Economic Review 54 (1): 82–95. Howard, M. 2001. Public Sector Economics for Developing Countries. Kingston, Jamaica: University Press of the West Indies. Huizinga, H., and L. Laeven. 2008. “International Pro�t Shifting within Multination- als: A Multi-country Perspective.� Journal of Public Economics 92 (5–6), 1164–82. Huizinga, H., L. Laeven, and G. Nicodeme. 2008. “Capital Structure and Interna- tional Debt Shifting.� Journal of Financial Economics 88 (1): 80–118. Keen, M., and M. Mansour. 2008. “Revenue Mobilization in Sub-Saharan Africa: Challenges from Globalization.� Draft working paper, August 11, Fiscal Affairs Department, International Monetary Fund, Washington, DC. Klemm, A. 2009. “Causes, Bene�ts, and Risks of Business Tax Incentives.� IMF Work- ing Paper 09/21, International Monetary Fund, Washington, DC. Maf�ni, G. 2009. “Tax Haven Activities and the Tax Liabilities of Multinational Groups.� Working Paper 09/25, Oxford University Centre for Business Taxa- tion, Oxford. OECD (Organisation for Economic Co-operation and Development). 2001. Corpo- rate Tax Incentives for Foreign Direct Investment. OECD Tax Policy Study 4. Paris: OECD. Oxfam. 2000. “Tax Havens: Releasing the Hidden Billions for Poverty Eradication.� Oxfam Brie�ng Paper. Oxfam International, London. Tax Evasion and Tax Avoidance: The Role of International Pro�t Sharing 141 ———. 2009. “Tax Haven Crackdown Could Deliver 120bn a Year to Reduce Pov- erty: Oxfam.� Press release, March 13, Oxfam GB, London. http://www.oxfam .org.uk/applications/blogs/pressof�ce/?p=3912. Pak, S. J. 2007. “Capital Flight and Tax Avoidance through Abnormal Pricing in Inter- national Trade: The Issue and the Solution.� In Closing the Floodgates: Collecting Tax to Pay for Development, Tax Justice Network, 118–22. London: Tax Justice Network. http://www.innovative�nance-oslo.no/pop.cfm?FuseAction=Doc&p Action=View&pDocumentId=11607. Pak, S. J., S. Zanakis, and J. S. Zdanowicz. 2003. “Detecting Abnormal Pricing in International Trade: The Greece-USA Case.� Interfaces 33 (2): 54–64. Schott, P. K. 2008. “The Relative Sophistication of Chinese Exports.� Economic Policy 23 (53): 5–49. TJN (Tax Justice Network). 2005. “The Price of Offshore.� Tax Justice Network Brief- ing Paper, March, TJN, London. ———. 2007. Closing the Floodgates: Collecting Tax to Pay for Development. Lon- don: TJN. http://www.innovative�nance-oslo.no/pop.cfm?FuseAction=Doc&p Action=View&pDocumentId=11607. UNCTAD (United Nations Conference on Trade and Development). 1999. World Investment Report 1999: Foreign Direct Investment and the Challenge of Develop- ment. Geneva: UNCTAD. Weichenrieder, A. 2009. “Pro�t Shifting in the EU: Evidence from Germany.� Interna- tional Tax and Public Finance 16 (3): 281–97. Zdanowicz, J. S., S. J. Pak, and M. A. Sullivan. 1999. “Brazil–United States Trade: Capital Flight through Abnormal Pricing.� International Trade Journal 13 (4): 423–43. Part II Illegal Markets 5 Illicit Capital Flows and Money Laundering in Colombia Francisco E. Thoumi and Marcela Anzola Abstract In Colombia, the need to send capital abroad to avoid anti–money laun- dering (AML) controls is not particularly great. During the last few decades, the domestic risk has been low, and so has the incentive for out- flows. This phenomenon has been associated with Colombian institu- tional characteristics: the conflict between the formal (legal) and infor- mal (social) norms that characterizes the country and that serves as a breeding ground for illegal economic activities there. Geography has been a huge obstacle to economic integration and modernization and a main obstacle to effective central state presence in large parts of the coun- try. The country has a large informal economy; �nancial sector services are expensive and cover only a minority of the population; illegal eco- nomic activities are widespread; and violence is frequently used to achieve economic goals. This environment has been fertile for the establishment of the illegal drug industry and provided good support for this industry’s resiliency. Law enforcement is a constant struggle. This explains why, although the country’s AML legislation is one of the most advanced in the world, the results have been meager. In Colombia, controlling internal and external illegal capital flows presents a daunting policy problem. 145 146 Draining Development? Unfortunately, as long as the conflict between the formal (legal) and informal (social) norms persists, illegal economic activities will flourish in Colombia. This is why cultural change should be a main policy goal. Introduction Most illegal capital outflows from developing countries are associated with kleptocracy, corruption, and tax evasion and avoidance (Baker 2005). Since criminal gains need to be laundered to disguise their illegal origin, the funds are commonly channeled through scattered accounts at banks situated in offshore �nancial centers that either facilitate �nancial transac- tions without investigating the origin of the funds or do not cooperate in AML investigations. Once the money is laundered abroad, the funds can be used as legal capital anywhere, including in the country of origin. The need to send illicit funds abroad to be laundered may arise from several situations. Thus, it might arise if the country of origin has the capability to exercise some control over crime and illicit gains. In other cases, as in countries in which presidents or dictators have transferred billions of dollars to hidden accounts in the �rst world, the fear of losing the illicit money because of political change may explain the need. Coun- tries in which military forces control the territory and support authoritar- ian regimes have all experienced large illegal kleptocratic capital outflows that have skimmed from the national budget and national resources. Such has been the case in Haiti, Indonesia, Nicaragua, Pakistan, Paraguay, Peru, the Russian Federation, Saudi Arabia, and several African countries (Baker 2005). The Colombian case differs substantially from the common pattern, however. In Colombia, money laundering has mainly been associated with the illegal drug industry, in which criminal activities cross borders, and payments are commonly made abroad. The main concern for the recipients of these funds is not only to launder the money, but also to bring the money to their countries, including back to Colombia. In the case of kleptocrats, guerrillas, and common organized criminals, even though the illegal funds are, in some cases, sent abroad, they are invested or spent directly in Colombia in most cases. The amounts of funds that these actors wish to send abroad likely pales in comparison with the amounts that traf�ckers wish to bring back in. Illicit Capital Flows and Money Laundering in Colombia 147 This particular situation is reflected in Colombian AML policy and laws (UIAF 2008). Indeed, in recent decades, the regulatory framework has focused on three dimensions of the effort to combat money laun- dering in Colombia: the administrative control of �nancial transac- tions, the penalization of money laundering and illicit enrichment as separate offenses, and the development of comprehensive asset forfei- ture legislation. In this context, �nancial entities are responsible for adopting adequate and suf�cient control measures to avoid being used as instruments for concealing, handling, investing, and, in any other way, using funds or other assets originating from criminal activities or to give an appearance of legitimacy to criminal activities or transactions and funds related to such activities. Colombia has also criminalized the laundering of the proceeds of extortion, illicit enrichment, rebellion, narcotics traf�cking, arms traf�cking, crimes against the �nancial sys- tem or public administration, and criminal conspiracy (Law 365 of 1997). Additionally, Colombian law provides for both conviction-based and nonconviction-based in rem forfeiture. The academic literature on illicit money flows in Colombia has also focused on illicit inflows, particularly on the macroeconomic effects of these inflows, and has disregarded illicit outflows (Steiner 1997; Rocha 2000, 2005). This situation raises a question, however: why do Colombian recipi- ents of illicit funds have little or no incentive to send or maintain the funds abroad? The answer is quite simple: in Colombia, despite some of the most advanced AML legislation in the world, there are no real obsta- cles to invest or spend illegal revenue. Institutional weaknesses have impeded the adequate implementation and enforcement of the AML laws, and the power of the paramilitary, guerrillas, and drug traf�cking organizations, particularly at the local level in many regions, and the gap between formal norms and culturally accepted informal norms have shaped an extralegal economy that is immune to traditional mechanisms for combating illegal economic activities. More precisely, in the case of corruption and illegal drug income, the decision on where to invest the illicit revenues depends on the social and criminal networks in which the lawbreaker participates. For example, a corrupt of�cial who is not part of a criminal network and receives a large sum would be tempted to deposit the sum in an offshore center. However, 148 Draining Development? if this person is part of a group that influences political elections, whose members are friends and relatives of the political elite, and that has a strong private military arm, the corrupt of�cial would be tempted to invest locally. Indeed, doing so, the of�cial will obtain not only an eco- nomic return, as he would abroad, but also political influence. So, in this case, there is not much incentive to invest abroad.1 This explains why most funds generated through corruption are likely to remain in the country. The exceptions are the money paid to local politicians or deci- sion makers by transnational corporations that deposit the funds in for- eign accounts or unusually large bribes associated with big government contracts. Thus, in Colombia, where an important part of the economy is involved in illicit, but socially legitimate economic activities and where corruption is not centralized, but widespread, the need to send capital abroad to avoid seizures and expropriations is not particularly great. As a result, during the last few decades, the domestic risk has been low, and so has the incentive for outflows. Even though this might have changed recently, the incen- tives to invest illegal capital in Colombia discourage illegal capital out- flows. The most recent cases of corruption in Colombia have provided evidence of the existence of bank accounts and investments abroad. However, in many cases, it is possible that the related illicit payments have also been made abroad. The authorities have not yet reported any capital outflows associated with these cases. This chapter analyzes why Colombian institutions constitute a fertile environment for the laundering and concealment of the proceeds of cer- tain unlawful activities such as the illegal drug industry, kleptocracy, and organized crime, provide good support for their resiliency, and, as a result, encourage illegal capital inflows and reduce the need to launder illegal capital abroad. Using country-speci�c characteristics as explanatory variables is dis- couraged in the social sciences because doing so opens a Pandora box that explains everything through these particularities. However, Colombia has been experiencing almost constant conflict for at least seven decades. As shown below, of�cial data indicate that 9 percent of the country’s popula- tion has been internally displaced. The International Organization for Migration estimates that, in 2008, there were 4.2 million Colombians liv- ing abroad (Ramírez, Zuluaga, and Perilla 2010). This means that about 18 Illicit Capital Flows and Money Laundering in Colombia 149 percent of Colombians have been displaced internally or left the country. Colombia is the most important producer of cocaine in the world (despite the many other possible competitors), and, during the last 15 years, it has also been the largest producer of coca. Colombia also has the distinction of being among the foremost countries in the world in number of kidnap- pings, assassins for hire (sicarios), child warriors, landmines and victims of landmines, producers of counterfeit U.S. dollars and other foreign exchange, exporters of Latin American prostitutes, and so on (Rubio 2004; Child Soldiers International 2008; United Nations 2009).2 It is the country where the term desechable (throw away) was coined to refer to indigents and other undesirables who have been socially “cleansed,� where “Not for Sale� signs are used widely to prevent fraudulent transactions aimed at stripping property from the legal owners, and where contraband is a socially valid way of life to the point that, in border regions, there have been public demonstrations to demand the right to smuggle. All these facts have shaped the way in which markets and law enforce- ment organizations operate. Indeed, Colombians colloquially refer to Colombian capitalism as savage, in contrast with capitalist markets in Europe, the United States, and even developing countries such as Chile. A study of illegal capital flows in Colombia that does not take into account the particular characteristics of Colombian markets and transactions runs the risk of reaching contradictory or erroneous conclusions. The next section shows how Colombian institutions are an obstacle to law enforcement and help propagate illegal economic activities. The sec- tion that follows explains how these institutions represent an obstacle for the enforcement of AML legislation. The penultimate section describes capital flows in and out of Colombia and shows the role of the illicit drug industry. The �nal section presents a summary and conclusions. Institutional Obstacles to Law Enforcement and the Upsurge in Illegal Economic Activities The Spanish Conquest that began in 1492 was never completed in Colom- bia. Large areas of the country have still not been settled. Local loyalties are strong, and the development of a national identity has been slow and remains un�nished. An important U.S. historian of Colombia refers to the country as “a nation in spite of itself � (Bushnell 1993). 150 Draining Development? Geography has been a huge obstacle to economic integration and modernization and a main obstacle to effective central state presence in large parts of the country.3 As a result, Colombia has a large informal economy, signi�cant land concentration, and a wide gap between formal institutions (constitution and laws) and the local culture. The informal economy Colombia has a large informal economy that is a principal obstacle to the enforcement of economic laws and facilitates money laundering activities. While there is no agreement on the de�nition of economic informality, Schneider and Klinglmair (2004), using a combination of statistical procedures, estimate that the informal economy accounted for 39.1 percent of gross domestic product (GDP) in Colombia in 1999/2000. A World Bank study that involved the measurement of informality using employment classi�cations found that, between 1992 and 2005, inde- pendent workers increased their share in total employment in Colombia by 17.3 percent (Perry et al. 2007). It also found that the share of infor- mal independent workers, plus informal wage earners, in total employ- ment in Colombia in 2006 was 66.8 percent. These are workers who do not comply with many legal requirements, contribute little if at all to the social security system, and do not have many social bene�ts. A low level of penetration of banking services also creates an environ- ment conducive to economic lawbreaking and money laundering. Data for 1990–99 and 2000–05 show that, in the �rst period, deposits repre- sented 14 percent of gross national product, which increased to 22 per- cent in the second period (Rojas-Suárez 2006). These shares are about one-third of the corresponding shares in the developed countries included in the survey and about one-half of the shares in Chile. Simi- larly, the number of bank of�ces and automated teller machines (ATMs) per 100,000 inhabitants is low in Colombia: about half the number in Brazil, Chile, and Mexico. These numbers for Colombia are only about 10 percent or less of the numbers found in the developed world. A more detailed study shows that 26.5 percent of the Colombian population resides in municipalities that do not have banking services (Marulanda 2006). These municipalities cover large areas of the country that have a strong guerrilla and paramilitary presence. Data from large urban areas show that only 26.4 percent of the adult population has deposits in or Illicit Capital Flows and Money Laundering in Colombia 151 loans from the �nancial system. Upper-income groups account for a large share of this population segment. Colombia experienced a �nancial crisis in 1999 caused by a signi�- cant change in the way mortgage balances were estimated. In response to persistent high inflation rates, a constant value unit system was estab- lished in the early 1970s. In essence, this set a �xed interest rate on mort- gages, although the principal could be adjusted according to inflation. In 1999 in the midst of a �nancial crisis and high interest rates, the adjust- ment mechanism was changed, and interest rate levels were used instead of inflation. The values of the principal in mortgages ballooned, and many borrowers found themselves with mortgages that exceeded sub- stantially the value of their properties; the real estate market collapsed. In 1999, Colombia, for the �rst and only time in the postwar period, showed negative growth in GDP (−4.2 percent). After the �nancial and real estate crisis, �nancial institutions had to clean their portfolios, and their total credit exposure declined. Using annual data, Tafur Saiden (2009) shows that, during the 1990s, the ratio of credit to GDP increased, reaching 37.9 percent in 1998. Then it dropped sharply, to the 24–25 percent range in 2002–05.4 A tax on �nancial transactions that was raised from 0.002 to 0.003 percent and then to 0.004 percent was earmarked to support the troubled �nancial sector. The government, however, found it an expedient source of funds and maintained it after the crisis was over. Today, it gathers in about 1 percent of GDP (Marulanda 2006). Increases in service costs, including many hidden fees, were another response of the �nancial sector to the crisis. Bank charges now include account management fees, fees to use ATM machines, and even fees to consult balances on ATM machines and the Internet, among many others. Moreover, the interest intermedia- tion gap is probably the largest in Latin America.5 The penetration of banking services in Colombia faces other obstacles, such as the fear of physical assault following a deposit withdrawal or cashing a check. This type of crime is so common that the police are now offering protection on demand to bank customers who make large withdrawals. Land concentration and the lack of territorial control The lack of territorial control by the central state has facilitated the growth of power among local groups. Left-wing guerrillas (Ejército de 152 Draining Development? Liberación Nacional and Fuerzas Armadas Revolucionarias de Colom- bia, more well known as ELN and FARC, respectively) and right-wing paramilitary groups have controlled many municipal governments and had suf�cient influence on some governments in the country’s depart- ments to be able to extract signi�cant amounts from local budgets to fund their activities. Illegal capital has been a factor in the extreme concentration of land tenancy and the political power of local landlords who, in large parts of the country, have had close ties to the illicit drug industry or have been active in this industry.6 Illegal drug money could not be easily invested in the modern economy, and drug traf�ckers purchased large amounts of rural land (Reyes Posada 1997). This concentration has been achieved by purchases, forced purchases under threat, and forced displacements of peasants and has been facilitated by the lack of well-de�ned, defendable property rights in many regions. Moreover, in towns controlled by armed groups or local traditional landlords, public land and title registries (catastros) are outdated and incomplete, and tax assessments grossly underestimated. Municipal reg- istries that should provide information on ownership are also inade- quate, and, frequently, municipal staff refuse to cooperate with the Fis- calía (Attorney General’s Of�ce) because the staff are controlled by people whose property rights may be challenged for money laundering. Land taxes are woefully low in many regions, and landed interests con- trol municipal councils, where they block any attempts to update the registries and raise rural real estate taxes. Paramilitary organizations have stripped large amounts of land from peasants. The lack of reliable land records and the fact that many peas- ants have not had formal property rights to their plots make it impossi- ble to develop accurate estimates on the extent of the plundering. Today, but for Sudan, Colombia has the largest number of displaced citizens in the world. The Internal Displacement Monitoring Centre estimates the number of displaced Colombians, as of June 2010, at 3.3 million to 4.9 million in a total population of 45.7 million.7 The gap between formal and informal norms Geography has also been a main factor in the development of diverse cul- tures because many regions essentially represent cultural endogamies Illicit Capital Flows and Money Laundering in Colombia 153 based on homogeneous values, beliefs, and attitudes (Yunis 2003). This explains why, in Colombia, there is a wide gap and strong contrast between formal institutions (constitution, laws, decrees, and so on) and the infor- mal unwritten norms recognized and followed by many social groups. Scholars have identi�ed this gap as a main problem in establishing the rule of law in Colombia (Thoumi 1987; Herrán 1987; Kalmanovitz 1989; Mockus 1994; Yunis 2003; Puyana-García 2005).8 One of the principal problems law enforcement efforts face in Colombia is the fact that most decent Colombians break economic laws in some form or another. Some do it without violence. For example, they buy contraband or evade taxes. Others go further and may use social and political influence to get govern- ment jobs or contracts; still others may use fraud, violence, or the threat of violence to achieve their goals. The persistence and diversity of the contrast and conflict in norms, coupled with large internal migrations partially in response to forced displacement, have resulted in a signi�cant number of Colombians developing a deep amoral individualism; they have become anomic: the effects of their actions on others are simply irrelevant.9 Many Colombi- ans tend to be selective regarding the laws that they comply with and those that they break. Most Colombians simply accept that they should obey “good� laws, but that they may disobey “unjust� laws.10 The problem of Colombia is not only the weak state, which could be strengthened, but, rather, that this state is embedded within a lax society in which social controls on behavior have been enfeebled to the point that they are irrelevant for many Colombians. In the Colombian institutional environment, the line dividing legal and illegal economic activities is fuzzy. Indeed, the division of activities between legal and illegal provides only a partial picture of reality. It is more relevant to divide the economy in terms of the law into four sectors: legal and socially legitimate, that is, activities that comply with the law and are accepted and reinforced by unwritten social norms; illegal and legitimate, that is, activities that involve lawbreaking, but that are accepted socially (purchasing contra- band, for example); legal and illegitimate, that is, activities that comply with the law, but contradict social norms (abortion, for example, accord- ing to the views of conservative religious groups); and illegal and illegiti- mate, that is, activities that involve breaking the law and contravening the norms of society. Of course, the classi�cation of activities into each of 154 Draining Development? these categories varies across social groups. The issue is how large the numbers of illegal but legitimate activities are and who agrees with the assessment. As these numbers grow, the effectiveness of law enforcement agencies declines, and, if the numbers are large, law enforcement becomes ineffective and the risk associated with lawbreaking is low. The Failure of Enforcement: The Gap between Formal and Informal Norms Colombia has one of the most advanced AML legislative frameworks in the world. It is focused on attacking the assets of drug traf�ckers. How- ever, enforcement has faced huge obstacles, and the results have been meager. The Financial Information and Analysis Unit has developed a system to manage the risk of asset laundering and the �nancing of terrorism. The system has been used effectively in the �nancial sector. However, in Colombia, the �nancial sector is not the main laundering venue because of its closed and oligopolistic structure and the limited penetration of banking services.11 The unit acknowledges, for example, that contraband plays a more important role as a means of laundering, and this is why the unit is now turning its attention to the real economy.12 The unit has also detected a signi�cant increase in the number of businesses that have been established speci�cally to provide specialized money laundering services.13 The Fiscalía �nds a similar evolution, not- ing that, 10 years ago, traf�cking organizations were pyramidal. Today, they are fragmented and more collegial. Money laundering �rms have increasingly become specialized and offer their services to meet many types of needs. In the past, they were part of the main contraband traf- �cking structures; today, they are independent. Many of these �rms have licit fronts and provide diverse services. In one case, for example, a �rm had 1,500 cédulas, the Colombian national identity cards, that it used to make �nancial transactions (smur�ng) for various clients.14 Additionally, substantial logistical obstacles impede the adequate implementation and enforcement of AML legislation. The Fiscalía has a large backlog of cases. They have few prosecutors, and the Judicial Police has only 40 of�cers assigned to �ght money laundering. Prosecutorial processes are time-consuming and labor intensive. Because of the preva- Illicit Capital Flows and Money Laundering in Colombia 155 lence of testaferros (front or straw men), prosecutors have to investigate traf�ckers and all their family members. In 2004, with the encouragement and support of the United States, Colombia changed its traditional justice system to an accusatory one. The shift to the new system has not been easy, and many lawyers, judges, and prosecutors do not understand the new system. Many young judges are not well quali�ed, do not understand money laundering issues, and should be trained. High personnel turnover is another problem because it means there must be regular training programs. These have been the main reasons for the large backlog of cases. Of�cials of the United Nations Of�ce on Drugs and Crime believe that there is a pressing need to improve the quality of justice system per- sonnel.15 They think that personnel problems are behind a declining trend in AML convictions. They also think that the quality of the data available on judicial processes are poor. The government agencies involved in �ghting money laundering have provided information to the judicial system for prosecutions. However, no information is available about what goes on during the processes. The National Narcotics Directorate (DNE) is in charge of administer- ing seized assets and forfeitures, but faces great challenges. Its record in administering, managing, and disposing of seized goods has been ques- tionable at best. DNE regularly encounters problems because of the titles of seized properties. Frequently, the relevant records have disappeared. Because of these problems, DNE has opted to accept assets for manage- ment only after formal precautionary measures have been completed. DNE currently manages assets in several special categories: chemical substances; urban real estate; rural real estate; land vehicles; airplanes; boats; cash, art, and other; and businesses. DNE decides how to dispose of the forfeited assets. Both DNE and the Fiscalía have had great dif�- culty separating front men from good faith asset holders. They know that a large share of the seized assets is in the hands of front men, but this is dif�cult to prove. Another problem arises because many assets are not productive, but must be stored and maintained and therefore generate substantial cash outflows. Others require special skills to be managed properly. By the end of November 2008, DNE had received 80,860 assets to manage, while the forfeiture processes were advancing. Of these, 12,397 156 Draining Development? (15.3 percent) had been returned to their owners, who had obtained favorable court judgments. Only 7,734 (9.6 percent) had been forfeited, and the remaining 60,729 (75.1 percent) were still the subject of judicial processes. The forfeiture process is slow. Indeed, in early May 2008, DNE auctioned property seized from Pablo Escobar, who was killed on December 2, 1993. The fact that the share of the assets that has been returned to asset owners is larger than the share that has been forfeited also indicates the level of problems in the process. DNE of�cials suggest that the data are misleading because the de�nition of asset that is used does not refer to an individual piece of property, but to all properties that are seized during each seizure event and that they believe are in the possession of one individual. They claim that some of the forfeited assets include several properties. However, they do not have data to corrobo- rate this. Table 5.1 summarizes the results of AML efforts involving the seized and forfeited assets of drug traf�ckers. The of�cial data have been aggre- gated and do not provide important details. For example, there are no estimates about the value of the assets seized and forfeited. Moreover, the de�nition of an asset is vague (see above). Despite the data de�ciencies, the results are clearly not encouraging. Table 5.1. The Results of Anti–money Laundering Efforts: Number of Seized and Forfeited Assets number Returned by In judicial Asset type judicial decision process Forfeited Total Urban real estate 2,571 11,790 2,965 17,326 Rural real estate 1,282 3,659 695 5,636 Businesses 62 2,54 1 287 2,890 Cash 410 5,420 327 6, 15 7 Controlled substances 525 9,439 653 10,61 7 Land vehicles 3,823 9,198 585 13,606 Airplanes and helicopters 361 660 44 1,065 Boats 179 571 99 849 Other 3, 184 1 7,451 2,079 22, 7 14 Total 12,397 60, 729 7,734 80,860 Source: UIAF 2008. Note: The data apply to the period up to November 30, 2008. Illicit Capital Flows and Money Laundering in Colombia 157 Several public of�cials have proposed legal changes to allow the gov- ernment to sell seized properties and return the proceeds if the owners win their cases. This would eliminate the problem of managing seized property, but raises the issue of how the sale price is to be determined in environments of extremely imperfect markets. Capital Flows in and out of Colombia: The Role of the Illicit Drug Industry Colombia began to be a player in the international illegal drug industry some 40 years ago. Since then, asset and money laundering and illegal capital flows have been an important policy issue in the country. Before 1970, policy debates related to illegal capital outflows existed, but the problem did not have a high priority on the policy agenda. The outflows then were generally motivated by the foreign exchange control system and the protectionist policies of the government, the fear of a left-wing political revolution, and the desire to evade Colombian taxes. Since the illicit drug industry started to grow in Colombia, the debate surrounding illegal capital flows and the related policies have been focused on capital inflows, and the policy concern over capital outflows has been minimal. This has been particularly the case in the last two decades, since the foreign exchange control regime was discarded and the economy was opened up. Kleptocracy is primarily a local phenomenon in Colombia; it has been limited by the small size of local budgets and the productive activities from which “commissions� may be illegally collected or bene�t may be illegally drawn from the associated public contracts.16 The Colombian military have never controlled the territory; they have relatively low social status, support elected governments, and play only a marginal political role.17 Colombia has not had a caudillo (leader) similar to the ones who, at times, have been typical in other Latin American countries and who have grossly enriched themselves and taken huge amounts of wealth from the country.18 The most publicized corruption event in Colombia in the 1990s was the funding of the 1994 presidential campaign of Ernesto Samper by the Cali cartel to the tune of some US$8 million. Despite a national scandal and a detailed investigation, Samper was never accused of having personally pro�ted from the money, although some of the funds were deposited in 158 Draining Development? accounts of a close campaign associate in New York who had absconded with some money for his own bene�t. However, this situation may be changing because of the surge in cor- ruption in the past decade. During the eight-year presidency of Alvaro Uribe (2002–10), corruption increased especially in the infrastructure sector, where many large projects were left incomplete or barely begun. While media reports suggest that a large part of the funds remained in the country, a signi�cant, though unknown portion may have been invested abroad. The available information concerning this issue is frag- mented and anecdotal, and there are no de�nitive �gures permitting us to describe a pattern of behavior or to reach an acceptable conclusion. Future research might be able to throw light on this issue. Capital flows and economic policies: From a foreign exchange control regime to an open economy The literature of the early 1970s describes Colombia as a foreign exchange–constrained economy that has a strict foreign exchange con- trol system (World Bank 1972; Nelson, Schultz, and Slighton 1971). Colombia had the longest running exchange control regime in Latin America. The system was in place from 1931 to 1991. There were several traditional ways for capital to flee the country illegally, including import overinvoicing and export underinvoicing, contraband exports (coffee and emeralds to the developed world and cattle to República Bolivariana de Venezuela were favorites), and, in the case of transnational corpora- tions, transfer prices and payments for patents and royalties.19 Tax eva- sion and avoidance were not likely to be an important determinant of capital flight because one could keep one’s tax payments low through creative accounting and the frequent tax amnesties that the government periodically enacted.20 Colombia did not have large investments by foreign corporations; so, transfer pricing, another illegal capital outflow or inflow modality, was not likely to have been signi�cant, although it was probably used by for- eign investors. The situation had changed drastically by the mid-1970s when illegal drug exports of marijuana and then cocaine grew quickly. From Decem- ber 1974 until the change in the foreign exchange regime in 1991, the black market exchange rate remained signi�cantly below the of�cial one Illicit Capital Flows and Money Laundering in Colombia 159 because drug money was flooding the black market with cash. Also, the net mispricing of international trade changed so that, on average, imports were underinvoiced and exports overinvoiced to bring foreign exchange revenues into the country (Thoumi 1995). The exception to this was 1982–83, when high interest rates in the United States appar- ently induced Colombian traf�ckers to purchase U.S. Treasury Bills and other interest-bearing papers abroad. This was a time of high cocaine prices in the United States and a sharp decline of illegal capital inflows in Colombia (Thoumi 1995). Similarly, in 1982 and 1983, worker remit- tances collapsed, but, “after 1984 Colombian expatriate workers appear to have been a lot more ‘generous’ with their Colombian brethren� (Thoumi 1995, 191). After the exchange control regime was eliminated in 1991, the parallel market in foreign exchange remained alongside the of�cial one, although it has declined in Bogotá and other large cities in the last few years, and, today, it is only a couple of percentage points below the of�cial one. Sell- ing foreign exchange in the of�cial market requires that sellers account for the origin of their funds. The gap between the two exchange rates has, at times, been explained by the taxes and fees charged in the of�cial market. The gap has varied not only over time, but also according to geography. In coca-growing regions, for example, peasants are frequently paid in U.S. dollars brought in the same small planes used to export cocaine. In these cases, the peasants sell the dollars at up to a 30 percent discount relative to the of�cial rate. The close current gap between the two exchange rates suggests that there is little if any questioning of suspi- cious transactions and that the risks associated with money laundering in the country are low. In the period before 1991, the year the economy was opened and for- eign exchange controls were eliminated, capital outflows were associated with the incentives generated by the exchange control regime and politi- cal risks. During the 1990s, illegal armed groups gained strength relative to the government, and the illicit drug industry’s control shifted from the old Cali and Medellín cartels to armed groups, guerrillas, and para- military groups. The increased political uncertainty and the growing risk were clearly reflected in the outflow of legal capital. Colombian investments abroad are a recent phenomenon. Before 1990, outward foreign direct investment (outward FDI) was rare, and 160 Draining Development? most of it was the result of the implementation of entrepreneurial strate- gies under the Andean Pact framework. Starting in the early 1990s, Colombian enterprises began to invest abroad (see �gure 5.1). Such investments between 1992 and 2002 were directed mainly to Panama, Peru, the United States, and República Bolivariana de Venezuela. The most important investments in the �rst two countries—the largest investment recipients—were in the �nancial sector. In 1997, the peak year, they grew dramatically. The �nancial sector was a good venue for people to buy certi�cates of deposit and other �nancial instruments abroad. The investments had a substantial component of capital flight generated by growing political instability. Indeed, the �gures for 1997 are extraordinarily high, reaching US$687.4 million (see �gure 5.1, particu- larly the data at the bottom). The growth of outward FDI in the �nancial sector did not coincide with large investments by Colombian banks. During these years, the lion’s share of �nancial investments was in other �nancial services that included a vari- ety of contracts and activities in which individual investors could partici- Figure 5.1. Colombia: Outward Foreign Direct Investment, 1994–2010 7,000 6,000 5,000 4,000 US$, millions 3,000 2,000 OFDI total 1,000 0 OFDI financial-related services –1,000 1994 1996 1998 2000 2002 2004 2006 2008 2010 1 994 1 995 1 996 1 997 1 998 1 999 2000 200 1 2002 2003 2004 2005 2006 2007 2008 2009 2010 OFDI TOTAL 149.0 256.2 327.9 809.4 796.0 1 1 5.5 325.3 1 6. 1 856.8 937.7 1 42.4 4,66 1 .9 1 ,098.3 9 1 2.8 2,254.0 3.088. 1 6,528.7 OFDI �nancial- related services 54.7 14 1.4 273.5 687.4 292.6 1.5 225.3 –227.4 75.2 32.5 32. 1 32.9 1 84.6 745.9 346.4 204.0 1 39. 1 Source: Central Bank of Colombia. Note: OFDI = outward foreign direct investment. Illicit Capital Flows and Money Laundering in Colombia 161 pate.21 After the launch of Plan Colombia and after the guerrilla threat had died down, Colombians began to regain con�dence in the economy, and, in 2001, �nancial sector outward FDI was negative. After 2002, Colombian outward FDI shifted from the �nancial sector to industry, mining, and public utilities. Colombian enterprises, like other Latin American enterprises, tend to invest regionally, close to the home country (Ecuador, Panama, Peru, República Bolivariana de Vene- zuela, and, most recently, Brazil, Chile, and Mexico). While in the 1990s, Colombian outward FDI could be characterized as typical capital flight explained by a loss of con�dence in the country, in the last decade by contrast, it has been the result of economic growth and the internaliza- tion strategies of Colombian enterprises. Illegal capital flows associated with the illicit drug industry Most of the revenues of Colombian drug traf�ckers has come from foreign markets. A kilogram of cocaine sold per gram in Europe or the United States generates revenues 50 to 100 times the export value in Colombia. There are no data on the value of all domestic cash seizures, but most have been in U.S. dollars. Fiscalía staff describe a case of the seizure in 2008 of US$3 million in cash. They found that the bills were soiled and smelled of soil, which indicate that they had been stored underground. They interpret this as a sign that the dollars had been brought into the country and stashed away without any intention of sending them abroad. They also claim that, the few times the government has found stashes of Colombian pesos, these have belonged to the guerrillas. These pesos are likely to be the product of kidnappings for ransom and extortions. When the �ndings have been mixed, both pesos and dollars, the latter account for the vast majority of the �nd.22 Remarkably, in Bogotá between August and October 2010, US$80 million and €17 million were found and seized altogether in �ve locations. These monies supposedly belonged to El Loco Barrera, a well-known traf�cker.23 A main problem faced by traf�ckers is how to bring such resources into Colombia.24 The traditional money laundering system of placing the cash in the �nancial sector, layering to hide the origin of the funds, and then integrating the resources into the legal economy is not necessarily desired by these people. Drug revenues have been brought into the country in many ways: physical cash, real sector contraband, underinvoicing imports 162 Draining Development? and overinvoicing exports, and fake labor remittances have been impor- tant. Rocha (1997) �nds that illegal flows hidden in the current account of the balance of payments show that the value of underinvoiced imports and overinvoiced exports, plus Central Bank foreign exchange purchases of fake labor and other remittances, averaged about US$1 billion a year.25 Given that various estimates place the total value of annual illegal drug revenue brought from abroad in the 1990s at between US$2 billion and US$5 billion, it is clear that the latter methods were important.26 Physical contraband has been another important method for bringing drug money into the country. As noted above, contraband is socially acceptable for many Colombians. Every city has a San Andresito, a shop- ping center where contraband is known to be sold openly, mixed with simi- lar legal imports.27 Computers and other electronic equipment are favorite illegal imports. These and other bulky items with empty space in the inte- rior, such as household appliances, are also used to bring U.S. dollars into the country. Many of these items are imported legally. In the 1990s, cigarettes were a principal means of laundering drug money. The international tobacco industry was at least an implicit accom- plice in this process. Both British Tobacco and Philip Morris exported huge amounts of cigarettes to Aruba, Curaçao, Margarita Island, Panama, and other Caribbean locations, from which the cigarettes were smuggled into Colombia. At almost any traf�c light, one could �nd someone selling cigarettes. The case of Aruba is remarkable in this respect, as its cigarette imports were equal to 25 percent of the island’s national income (Steiner 1997). For many years, the Philip Morris advertising budget in Colombia exceeded the value of of�cial Philip Morris cigarette exports to Colombia. The of�cial Philip Morris response to Colombian of�cials who ques- tioned these advertising expenditures was simply that they were trying to increase their market share even though they had over 70 percent of the market already.28 When these facts became known, there was a public out- cry that led to negotiations between the government and the tobacco companies to suppress this contraband (Thoumi 2003). Several local gov- ernments sued the tobacco companies for tax evasion, and, in a June 2009 agreement, Philip Morris agreed to pay US$200 million without acknowl- edging any wrongdoing or accepting any liability. Cigarette smuggling was a case in which there was simultaneous capital inflow and outflow. The tobacco companies were selling their products in Curaçao, a free Illicit Capital Flows and Money Laundering in Colombia 163 port, for consumption in Colombia, thereby evading import and income taxes in Colombia. During the 1990s, the illegal drug industry evolved substantially. The area under cultivation with illegal crops grew dramatically. The govern- ment succeeded in destroying the Medellín cartel. The Cali cartel expanded, but was then also destroyed by the government. Traf�cking organizations became smaller, and their armed branches became weaker. The paramili- tary organizations that had developed through links with the illegal indus- try gained ascendancy over the traditional traf�ckers because they gained control of areas in which cocaine was re�ned, as well as of traf�cking cor- ridors. They also controlled some coca- and poppy-growing regions. Another change in the industry was caused by the increased involve- ment of Mexican traf�ckers in cocaine distribution in the United States. The pressure exerted by the United States on the Caribbean routes begin- ning in the late 1980s led to the displacement of the traf�c to Mexico and the Paci�c coast. Colombians started hiring Mexicans to smuggle cocaine into the United States, but Mexicans soon realized that they could have their own distribution networks and the power to gain greater pro�t by selling directly in the U.S. market. Furthermore, as the large Colombian cartels weakened, the relative power of the Mexican organizations increased. As a result, a signi�cant number of Colombian traf�ckers now sell their cocaine in Mexico or in Central America to the powerful Mexican cartels. This has lowered the revenues of the Colombian cocaine industry and thus also the amount of money to be laundered in Colombia. The use of the illegal revenues has also evolved. The early traf�ck- ers invested in real estate and some businesses. During the surge in the market for marijuana in the early 1970s, there was a real estate boom in Barranquilla and Santa Marta, the two cities closest to the marijuana- growing region in the Sierra Nevada de Santa Marta. When the Cali and Medellín cartels gained prominence, the cities experienced another real estate boom. Traf�ckers, mainly from the Medellín cartel, invested heavily in rural land, while traf�ckers from Cali appeared to prefer urban real estate and investments in various industries and services.29 The con- centration of rural land among armed groups has been substantial. Reyes Posada (1997) estimates that, by 1995, drug traf�ckers already controlled over 4 million hectares. This process has continued as the ter- ritorial control by armed groups strengthened (Reyes Posada 2009). 164 Draining Development? During the last decade, the control of the industry by warlords has meant that a signi�cant share of the illegal drug revenues has been used to fund the armed conflict: the purchase of weapons, payments to the armed personnel in the counterinsurgency, and the like. This presents a problem in estimating illegal capital inflows because it is dif�cult to de�ne weapons as capital. A study by the intergovernmental Financial Action Task Force concludes that “a project team investigated the links between narcotics traf�cking and terrorist �nancing, but had to conclude that this subject was less suitable for analysis through publicly accessible information� (FATF 2005, 1). This assertion might be valid in cases of international terrorist organizations that implement suicide attacks or plant bombs, but do not have armies to support. In the case of Colombia, there is no question that drugs have sustained the armed personnel and allowed the guerrillas and paramilitary groups to grow. It is remarkable that, before the expansion of the coca plantings, FARC was able to sur- vive, though it had never become an important problem or a threat to the establishment. Similarly, the paramilitary movement became strong only after coca and cocaine had come to account for important activities. Conclusion Over the last 35 years, the main issue in illegal capital flows has been the consequences of the entry of the foreign revenues of the illicit drug industry into the country. In contrast to many other countries, the main policy issue has been the control of illegal capital inflows, not the control of illegal outflows. From the 1970s onward, these revenues have been invested in the legal economy and have created a particularly grave problem in rural and urban real estate. They have also aggravated rural land concentration, which has been a factor in making Colombia a more violent country, with a displacement of people second only to Sudan in the world. There is no question that, although almost 80 percent of the population is urban, the unresolved land problem remains a key policy issue in Colombia. The government has been unable to impose the rule of law over the national territory, and various regions have developed their own norms. Other factors, including the persistently high level of violence, have compounded the problem, and civil society has become increasingly Illicit Capital Flows and Money Laundering in Colombia 165 indifferent to the rule of law. The decision about whether to respect a law has become similar to an investor’s portfolio decision in which the risks of various actions are weighed to determine what to do. Not surprisingly, law enforcement has become a constant struggle, which reflects the gap between the law and socially acceptable behavior. Many government of�cials either do not believe in the laws or are unable to enforce them. Thus, although the country’s AML legislation is one of the most advanced in the world, the results have been meager, and there is little need to take illegal capital abroad to protect it. Controlling internal and external illegal capital flows presents a daunting policy problem given that a large proportion of the population considers the law as a grey area that one may or may not respect. As long as the conflict between the formal (legal) and informal (social) norms persists, illegal economic activities will flourish in Colombia. Closing this gap requires signi�cant institutional and cultural change. Notes 1. For example, Pablo Ardila, former governor of Cundinamarca Department, who was sentenced to jail in 2007 on kleptocracy charges, had transferred sub- stantial assets abroad, but, his lifestyle in Colombia was incongruent with his reported income and caught the attention of the authorities. He regularly brought money from accounts in Panama and other offshore centers back to Colombia to support his high living standard (“Las cuentas del Gobernador,� El Espectador, November 27, 2007). 2. “Colombia is a major source country for women and girls traf�cked to Latin America, the Caribbean, Western Europe, Asia, and North America, including the United States, for purposes of commercial sexual exploitation and involun- tary servitude� (U.S. Department of State, Traf�cking in Persons Report, June 2009 [Washington, DC, 2009], 107). 3. “Because of its geography, Colombia was until the early XX century the Latin American country with the lowest per-capita international trade. Palmer (1980, p. 46) shows that as late as 1910 Colombian exports per capita were 77% of the second lowest country (Honduras), 67% of those of Peru, 52% of those of República Bolivariana de Venezuela, 12% of those of Argentina and 9% of those of Uruguay. Only the development of the coffee industry modi�ed this condi- tion� (Thoumi 1995, 18). 4. These data are consistent with a graph in Marulanda (2006, 50) that, unfortu- nately, does not provide numbers. In the graph, the ratio of deposits to GDP follows a similar path. 166 Draining Development? 5. One of the authors of this chapter, for example, closed a savings account in 2008 that had a balance of about US$7,000 because the interest received was less than the monthly cost of the account. 6. Rural land property has always been a key issue in Colombian society. Land has been a symbol of wealth and power and remains so for many Colombians despite the rapid modernization that the country has experienced (Zamosc 1986; Fajardo 1986). La Violencia of the 1940s and 1950s, during which at least 200,000 Colombians (about 1.8 percent of the population) were killed, was, to a great degree, a �ght over land, and it generated large displacements of people because armed bands were killing peasants and forcing many others off their land (Guzmán, Fals Borda, and Umaña Luna 1962; Fals Borda 1982). 7. The website of the center is at http://www.internal-displacement.org/countries/ colombia. 8. It is remarkable that this gap was the main issue in the campaign for the 2010 presidential election. 9. Over 50 years ago, Ban�eld’s pathbreaking book on southern Italy (1958) referred to the society there as “amoral familism.� Today, in Colombia, even this has weakened to the point that we have “amoral individualism.� 10. These beliefs are not new. A good example is Augusto Ramírez-Moreno’s col- umn in El Siglo on March 20, 1936, cited by Acevedo-Carmona (1995, 153): “The Liberal regime has declared civil war on Colombians. . . . There is a need to disobey. Citizens are relieved from obeying the wicked laws and the illegitimate authorities in power.� Remarkably, the reference was to the duly elected govern- ment and Congress. During La Violencia, some Catholic priests, such as Monsi- gnor Builes of Santa Rosa de Osos, incited his flock by preaching that killing Liberals was not sinful. 11. Banks have taken strong measures. Banco de Bogotá, for example, has a team of 50 professionals, each of whom visits 50 clients a month to examine how their businesses operate. 12. The unit is working with Fenalco (the Colombian Retailers Association) and with the Chamber of Commerce of Bogotá. They all support and encourage investigative journalism into issues related to money laundering. 13. Author interviews with staff of the Financial Information and Analysis Unit, March 2009. 14. Author interviews with Fiscalía staff, March 2009. 15. Author interviews with of�cials of the United Nations Of�ce on Drugs and Crime, March–April 2009. 16. For example, indexes of Transparencia por Colombia show that entities in the municipalities and departments are at substantially greater risk of corruption than the central government agencies that have much larger budgets. See Trans- parencia por Colombia, http://www.transparenciacolombia.org.co/LACORRUP CION/tabid/62/Default.aspx. 17. Since the new Constitution was enacted in 1991, for example, Congress has always had more members who have formerly been guerrillas than members who have formerly been in the armed forces. Illicit Capital Flows and Money Laundering in Colombia 167 18. For example, see Naím (2005), who covers the main cases of kleptocracy in devel- oping countries; Colombia is not included. Jorge (2008) shows the signi�cant wealth stashed away by Montesinos in Peru during the Fujimori administration. 19. The International Coffee Agreement was signed in 1962 and modi�ed several times. Until 1994, it provided for export quotas for the producing countries, and this induced contraband. 20. Tax amnesties became almost routine as part of the tax reforms enacted by every administration. They were implemented by the López (1974), Turbay (1979), Betancur (1982), and Barco (1986) administrations (Thoumi 1995). 21. Disaggregated data on these investments are not public, but well-informed �nan- cial sector professionals agree that this was a period during which people sought ways to take capital out of the country and that it was possible to do so legally through outward FDI in �nancial services. 22. Author interviews with Fiscalía staff, March 2009. 23. “Más de US$ 200.000 millones hallados en caletas en Bogotá, a manos del Estado,� El Espectador, March 31, 2011. 24. The physical weight and size of the dollars from sales abroad exceed the volume and weight of the cocaine itself. 25. Since the mid-1970s, the Central Bank has frequently changed its foreign exchange purchasing policies. These changes were particularly important before 1991 when the country had strict foreign exchange controls. Since then, under a system in which there is legal parallel market, a bank policy of easing purchasing require- ments has lost importance. 26. Estimates of Colombian GDP in U.S. dollars vary. In the early 1990s, they were around the US$100 billion level. The estimates have increased substantially in recent years. 27. The origin of this term is consistent with the gap in Colombia between legal and social norms. In the mid-1950s, during his presidency, General Rojas-Pinilla wished to develop the Archipelago of San Andrés and Providencia in the Carib- bean near Nicaragua, but he decided against a program to attract foreign tourists and promoted tourism by Colombians instead, given that the vast majority could not afford vacations in other Caribbean resorts. To make it attractive for Colom- bians to visit San Andrés and Providencia, Rojas-Pinilla allowed Colombians who spent a few days there to purchase certain articles tax free while vacationing. This was a time of strict foreign exchange controls, high tariffs, import quotas, and prohibitions. San Andrés became a de facto smuggling-tourism center. Store owners underinvoiced their sales, and tourists brought back expensive and pro- hibited goods, particularly home appliances: thus, the name San Andresitos. 28. Interview with Miguel Fadul Jr., director of the Colombian Commercial Of�ce in Washington, DC, August 1998. 29. The leaders of the Cali cartel were known for their investments in a large drug store chain, real estate management agencies, several factories, and, at one point, a bank. 168 Draining Development? References Acevedo-Carmona, D. 1995. La mentalidad de las élites sobre la violencia en Colombia (1936–1949). Bogotá: IEPRI and El �ncora Editores. Baker, R. W. 2005. Capitalism’s Achilles Heel: Dirty Money and How to Renew the Free- Market System. Hoboken, NJ: John Wiley & Sons. Ban�eld, E. C. 1958. The Moral Basis of a Backward Society. With the assistance of L. F. Ban�eld. New York: Free Press. Bushnell, D. 1993. The Making of Modern Colombia: A Nation in Spite of Itself. Berke- ley, CA: University of California Press. Child Soldiers International. 2008. Child Soldiers: Global Report 2008. London: Child Soldiers International. Fajardo, D. 1986. Haciendas, campesinos y políticas agrarias en Colombia, 1920–1980. Bogotá: Centro de Investigaciones para el Desarrollo. Fals Borda, O. 1982. Historia de la cuestión agraria en Colombia, 1982. Bogotá: Carlos Valencia Editores. FATF (Financial Action Task Force). 2005. “Money Laundering & Terrorist Financ- ing Typologies, 2004–2005.� Report, June 10, FATF, Paris. http://www.fatf-ga�. org/dataoecd/16/8/35003256.pdf. Guzmán, G., O. Fals Borda, and E. Umaña Luna. 1962. La violencia en Colombia: estudio de un proceso social. Bogotá: Tercer Mundo Editores. Herrán, M. T. 1987. La sociedad de la mentira. 2nd ed. Bogotá: Fondo Editorial CEREC– Editorial la Oveja Negra. Jorge, G. 2008. “La experiencia de Perú: el caso ‘Fujimori-Montesinos’.� In Recuper- ación de activos de la corrupción, ed. G. Jorge, 231–51. Buenos Aires: Editores del Puerto. Kalmanovitz, S. 1989. La encrucijada de la sinrazón y otros ensayos. Bogotá: Tercer Mundo Editores. Marulanda, B. 2006. “Una nueva política para un mayor acceso a los servicios �nan- cieros en Colombia.� In La Extensión del crédito y servicios �nancieros: obstácu- los, propuestas y buenas prácticas, ed. Secretaría General Iberoamericana, 49–60. Madrid: Secretaría General Iberoamericana. Mockus, A. 1994. “An�bios culturales y divorcio entre ley, moral y cultura.� Análisis Político 21. Naím, M. 2005. Illicit: How Smugglers, Traf�ckers, and Copycats Are Hijacking the Global Economy. New York: Anchor Books. Nelson, R. R., T. P. Schultz, and R. L. Slighton. 1971. Structural Change in a Develop- ing Economy: Colombian Problems and Prospects. Princeton, NJ: Princeton Uni- versity Press. Palmer, D. S. 1980. Peru: the Authoritarian Tradition. New York: Praeger Publishers. Perry, G. E., W. F. Maloney, O. S. Arias, P. Fajnzylber, A. D. Mason, and J. Saavedra- Chanduvi. 2007. Informality: Exit and Exclusion; Building Effective and Legitimate Institutions. Latin America and Caribbean Studies. Washington, DC: World Bank. Illicit Capital Flows and Money Laundering in Colombia 169 Puyana-García, G. 2005. ¿Cómo somos? Los Colombianos: reflexiones sobre nuestra idiosincrasia y cultura. Bogotá: Panamericana Editorial. Ramí rez, C., M. Zuluaga, and C. Perilla. 2010. “Per�l Migratorio de Colombia 2010.� Report, June, International Organization for Migration, Bogotá. http:// publications.iom.int/bookstore/free/PERFIL_27ABRIL_IMPUESTAS.PDF. Reyes Posada, A. 1997. “La compra de tierras por narcotra�cantes en Colombia.� In Drogas ilícitas en Colombia: su impacto económico, político y social, ed. F. E. Thoumi, 279–96. Bogotá: Editorial Planeta. ———. 2009. Guerreros y campesinos: el despojo de la tierra en Colombia. With the assistance of Liliana Duica-Amaya. Bogotá: Friedrich Ebert Stiftung and Grupo Editorial Norma. Rocha, R. 1997. “Aspectos económicos de las drogas ilegales en Colombia.� In Drogas ilícitas en Colombia: su impacto económico, político y social, ed. F. E. Thoumi, 141–51. Ariel ciencia politica. Bogotá: Editorial Planeta. ———. 2000. La economía Colombiana tras 25 años de narcotrá�co. Bogotá: Siglo del Hombre Editores. ———. 2005. “Sobre las magnitudes del narcotrá�co.� In Narcotrá�co en Colombia: economía y violencia, ed. A. Rangel, 145–288. Bogotá: Fundación Seguridad & Democracia. Rojas-Suárez, L. 2006. “El panorama regional.� In La extensión del crédito y servicios �nancieros: obstáculos, propuestas y buenas prácticas, ed. Secretaría General Iberoamericana, 17–41. Madrid: Secretaría General Iberoamericana. Rubio, M. 2004. “Kidnapping and Armed Conflict in Colombia.� Paper presented at the Center for the Study of Civil War’s workshop, “Techniques of Violence in Civil War,� Peace Research Institute Oslo, Oslo, August 20–21. Schneider, F. G., and R. Klinglmair. 2004. “Shadow Economies around the World: What Do We Know?� IZA Discussion Paper 1043 (March), Institute for the Study of Labor, Bonn. ftp://repec.iza.org/RePEc/Discussionpaper/dp1043.pdf. Steiner, R. 1997. Los dólares del narcotrá�co. Cuadernos de Fedesarrollo 2. Bogotá: Fedesarrollo–Tercer Mundo. Tafur Saiden, C. 2009. “Bancarización: una aproximación al caso colombiano a la luz de América Latina.� Estudios Gerenciales 25 (110): 13–37, Universidad Icesi, Cali, Colombia. Thoumi, F. E. 1987. “Some Implications of the Growth of the Underground Economy in Colombia.� Journal of Interamerican Studies and World Affairs 29 (2): 35–53. ———. 1995. Political Economy and Illegal Drugs in Colombia. Boulder, CO: Lynne Rienner. ———. 2003. Illegal Drugs, Economy and Society in the Andes. Baltimore: Johns Hop- kins University Press. ———. 2005. “Why a Country Produces Drugs and How This Determines Policy Effectiveness: A General Model and Some Applications to Colombia.� In Elusive Peace: International, National and Local Dimensions of Conflict in Colombia, ed. C. Rojas and J. Meltzer, part 3. New York: Palgrave Macmillan. 170 Draining Development? UIAF (Financial Information and Analysis Unit). 2008. “Description of Typologies Published by the UIAF, 2004–2008.� Report, UIAF, Bogotá. United Nations. 2009. “Assistance in Mine Action: Report of the Secretary-General.� Document A/64/287 (August 12), United Nations General Assembly, New York. World Bank. 1972. Economic Growth of Colombia: Problems and Prospects. World Bank Country Economic Report. Washington, DC: World Bank; Baltimore: Johns Hopkins University Press. Yunis, E. 2003. ¿Por qué somos así? ¿Qué pasó en Colombia? Análisis del mestizaje. Bogotá: Editorial Temis. Zamosc, L. 1986. The Agrarian Question and the Peasant Movement in Colombia: Struggles of the National Peasant Association, 1967–1981. Cambridge Latin American Studies. Cambridge, U.K.: Cambridge University Press. 6 Human Traf�cking and International Financial Flows Pierre Kopp Abstract Little attention has been paid to the �nancial flows generated by human traf�cking. Nonetheless, adding the many different flows of money linked to human traf�cking together into one big number generates a meaningless �gure. The socioeconomic consequences of each monetary flow are closely linked to the characteristics of the speci�c component of traf�cking the flow reflects. Following the money flows can provide use- ful information on the heterogeneity of human traf�cking. The flow of money entering in the source countries and linked to human traf�cking is negligible if it is compared, for example, with total remittances. Most of the pro�ts made by petty traf�ckers stay in the host countries. The only important flow of money internationally results from criminal organizations. It is almost impossible to distinguish the pro�ts of human traf�cking from those associated with the core busi- nesses of criminal organizations (prostitution, illegal drugs, counterfeit- ing, corruption, and so on). The research described in this chapter, even if preliminary, suggests that money flows entering developing countries and linked to human traf�cking are merged into broader flows of criminal money. There is no 171 172 Draining Development? evidence that the flows have a particular, speci�c impact on the countries exporting the humans. Whatever the source of the criminal money, the impact is mostly the same. Introduction Human traf�cking has gained wide attention in recent years. Important debates have focused on the magnitude of the problem, the involvement of criminal organizations, and the share of sexual exploitation in the global phenomenon. However, little attention has been paid to the �nan- cial flows generated by this traf�cking. This is all the more regrettable because it seems that these flows should be easier to bring to light than the traf�cking patterns themselves, which, for obvious reasons, remain shrouded in secrecy. Analysis of the monetary flows is thus likely to pro- vide a precious tool for the representation of a traf�c that is dif�cult to chart directly. Indeed, pursuing the flows of money provides useful information on the design of human traf�cking patterns worldwide. Furthermore, it offers an excellent ground for overturning the overly simplistic image of these activities that portrays them as entirely con- trolled by transnational criminal organizations. Instead, this analysis allows us to construct an image of a world that is deeply fragmented and heterogeneous. A re�ned understanding of �nancial flows can also help build the structure necessary to analyze the impact of the traf�c on the developing countries that are home to most of the individuals who are implicated in the traf�c. Indeed, if formerly traf�cked individuals develop sources of revenue, they often channel part of this revenue back to their countries of origin. On a purely macroeconomic level, the impact of these flows of money is not sensitive to the origin of the victims. Incoming money flows improve the balance of payments; outgoing flows cause a deterio- ration in the external position of the country. However, tracking the socioeconomic impact of the inflows of money cannot be accomplished with any accuracy if the origin of the victims of traf�cking and of the money is not taken into account. The socioeconomic consequences of monetary flows are known to be determined by the characteristics of the activities that generate them. Social relations or a country’s reputation can be affected by illegal activities even if these don’t account for signi�- Human Traf�cking and International Financial Flows 173 cant economic transactions and have almost no macroeconomic effect. This might be even truer at the provincial or city levels. For these rea- sons, the �nancial repercussions of the trade in human beings must be analyzed with an eye to the speci�cities of the various activities for which compensation is being received. Only a nuanced account of �nancial flows can provide the information necessary to distinguish the various types of impacts human traf�cking have on source countries. This chapter aims to structure the elements of a methodological framework for a nuanced analysis of the �nancial flows of human traf- �cking. It argues that the methods used to establish global estimates are seriously flawed in that they conflate sources in such a way as to produce nonsensical results. If estimates of the illegal drug trade are added to the turnover associated with heroin and with cannabis, the result fails to take into account the fact that the traf�cking networks and the source countries are different. Such problems are compounded if it is a matter of estimating human traf�cking: whereas, in the case of drugs, at least the products are clearly identi�ed and the wholesale and retail prices more or less well known (Kilmer and Reuter 2009); in the case of human traf�cking, the modes of distribution do not allow for pertinent price observations and make it dif�cult to identify the various components in the turnover. The contribution of this chapter will be more analytical than empiri- cal. Our key contention is that more light on how criminal networks are organized and how they run their various activities will allow us to improve our understanding of both the origins and the destinations of the flows of money. This, in turn, will furnish two types of bene�ts. First, knowing where the money comes from and where it goes to illuminates the various strategies of those people who are active in human traf�ck- ing. A clearer understanding of how traf�ckers conduct their business might provide new suggestions to the authorities who seek to control the trade. Second, identifying the patterns of �nancial flows can provide a better understanding of their socioeconomic impact on the source countries. The next section will give some insight into the main theoretical con- troversies that have surfaced around the de�nition of human traf�cking. The subsequent section describes how the hopes of individuals to �nd a promised land pushes, more or less voluntarily, millions of people into 174 Draining Development? the hands of criminals. Monetary flows are generated both by the com- pensation demanded of people for help with illegal border crossings and by the revenue of those criminals who exploit the slaves, prostitutes, and so on that these people become after they have crossed the borders. This is examined in the penultimate section. The last section details the design of the criminal organizations involved in this traf�cking and considers their respective needs for money laundering. This section concludes with a taxonomy of the illegal flows generated by human traf�cking. Conceptual Issues Human traf�cking is a part of the shadow economy (Fleming, Roman, and Farrell 2000). The term traf�c in economics explicitly refers to the idea of transporting merchandise. However, if the merchandise is human, there are speci�c dif�culties in establishing de�nitions. A major dif�culty emerged when United Nations agencies were charged with establishing of�cial de�nitions of smuggling and traf�cking. Nonethe- less, we contend that the conceptual problem with which the legal de�ni- tions grapple is not necessarily the most relevant angle for studying traf- �cking from an economic standpoint. Consider the de�nitions provided by the United Nations Protocol against the Smuggling of Migrants by Land, Sea, and Air and the United Nations Protocol to Suppress and Punish Traf�cking in Persons, Espe- cially Women and Children. The Protocols are supplements to the Pal- ermo convention, the United Nations Convention against Transnational Organized Crime, adopted by the United Nations General Assembly on November 15, 2000 (see UNODC 2004). The smuggling protocol de�nes smuggling as “the procurement, in order to obtain, directly or indirectly, a �nancial or other material ben- e�t, of the illegal entry of a person into a State Party of which the person is not a national or a permanent resident� (article 3), and the traf�cking protocol de�nes traf�cking as “the recruitment, transportation, transfer, harbouring or receipt of persons, by means of the threat or use of force or other forms of coercion, of abduction, of fraud, of deception, of the abuse of power or a position of vulnerability or of the giving or receiving of payments or bene�ts to achieve the consent of a person having con- trol over another person, for the purpose of exploitation� (article 3). The Human Traf�cking and International Financial Flows 175 main forms of exploitation are prostitution, forced labor, slavery, and the removal of organs. Buckland (2008) recalls that those who gathered in Vienna to negoti- ate the traf�cking protocol were united in mandate, but sorely divided when it came to choosing the road ahead. The major cleavage became apparent around the issue of de�nitions. Is traf�cking only about prosti- tution? Is all prostitution traf�cking? Are men traf�cked, too? The nego- tiators’ answer (according to Buckland) reflects a certain ambivalence in the fundamental conception of traf�cking. While the protocol includes a broad range of activities in the de�nition, both the title and the statement of purpose cast traf�cked people as victims and make clear that the vic- tims of traf�cking whom it seeks to protect are, �rst and foremost, women and children in forced prostitution. The obvious consequence of overem- phasizing the victim status is that it allows an arti�cial and tendentious distinction between smuggled people (painted largely as criminal by international institutions) and traf�cked people. As studies have shown, however, smuggled people may be trapped in debt bondage or subject to exploitative working conditions, and it therefore seems problematic to maintain conceptual categories that cast these people simply as criminals rather than victims (Andreas and van der Linden 2005). Given that intentionality is key to determining the nature of acts in legal terms, it is not surprising that those who drafted the of�cial de�ni- tions felt compelled to take a stance on the issue. Our concern, however, is less with ascribing legal responsibility than with reaching an accurate description of this illegal global traf�c. For our purposes, it seems that distinguishing between voluntary and involuntary immigration raises more problems than it solves. It seems clear that, if an organ is taken from a dead body and sent to another country to serve as a transplant without consent, this is an involuntary traf�c. Likewise, a clandestine migrant whose passport is con�scated and who �nds himself enslaved has not chosen this situation. Yet, grouping these two movements together under a label of involuntary (or voluntary) obscures the differ- ent sorts of traf�c involved in these two cases. Furthermore, the distinction between voluntary migration in search of opportunity and involuntary traf�cking of human laborers may be conceptually satisfying, but applying this distinction in practice is close to impossible. While the conditions of their migration may, indeed, seem 176 Draining Development? shocking, it is far from clear that all migrants who are treated as goods to be exploited are acting involuntarily. What of migrants who turn them- selves into transportable commodities because they expect to obtain a better price for their work in a target country? Classic economics theory would hold that such a decision may be deemed rational within a nor- mal economic framework because no one is better placed than the indi- vidual to know whether what he is doing is in his self-interest no matter whether or not his decision turns out, in the end, to have been a good calculation. Tens of thousands (perhaps even hundreds of thousands) of clandestine migrants who cross the Mexican–United States border each year purchase the services of smugglers and become part of the flow of smuggled human beings. The persistence of the flow and the fact that the same individuals may recross the border several times, depending upon the state of the employment market (and the riskiness of the crossing), can be taken to indicate that the decision is voluntary. And, yet, those who might initially be thought merely to be purchas- ing the services of smuggling networks can, in effect, become subject to traf�cking. Massive illegal immigration and the strengthening of con- trols (for example, with the fence along extensive sectors of the Mexico– United States border) have created situations in which border crossing has become an industry. To improve their chances of successfully cross- ing the border, candidates for illegal immigration appeal to underground networks for human smuggling. Crossing the border with a network enables one to bene�t from the network’s investments (trucks, corrup- tion, tunnels, and so on), but such a choice also exposes one to great danger. To avoid endangering their investments, or simply to skimp on costs, smugglers are sometimes prepared to dispose of their human cargo in the sea or in the desert. Furthermore, the condition of absolute dependence into which the individual is placed if he has put his destiny into the hands of people running a network exposes him to all sorts of risks, sometimes dramatic. Clandestine migrants can be robbed of their papers and money; they can be raped and killed. In such circumstances, it becomes disingenuous to maintain that they are not being treated as merchandise. If one considers, furthermore, that the basis on which the decision to migrate is made may already be determined in part by the smuggling networks, the category of voluntary migration becomes even more prob- Human Traf�cking and International Financial Flows 177 lematic. To sustain their activity, smugglers attempt to recruit new can- didates for crossings and develop their activities in source countries. While they seem the best source of information, the smugglers, by pre- senting a false image of the conditions under which illegal emigration can take place, increase the number of their clients and maintain a sup- ply effect. Through this dynamic, the offer of smuggling contributes to an activation of the demand (Jewell and Molina 2009).1 Insofar as it is created at least in part by false information, even the initial demand for smuggling cannot simply be considered entirely voluntary. Assessing the voluntary or involuntary character of migration is thus extremely dif�cult. More importantly, the attempt produces misleading categorizations of activity at least as regards the economic reality. Yet, the principle of this distinction, inherited from legal debates, underlies the conceptual categories used by the international agencies that monitor or assess these activities. In this chapter, we use the term traf�cking in a broad sense to cover both smuggling and traf�cking as de�ned by the United Nations. Indeed, we adopt a broad de�nition of traf�cking that includes all means of smuggling, prostitution, organ traf�c, slavery, and so on.2 Some of these activities are clearly coercive; some are not; and some are partially coer- cive. However, all must be considered if we are to establish a taxonomy of the different flows of money linked to human traf�cking. To underline the stakes in choosing this broad de�nition, we now briefly consider estimates that have been made of international human traf�cking using other de�nitions so as to emphasize how poor de�ni- tions can lead to problematic numbers. It is exceedingly dif�cult to estimate the scale of human traf�cking. Evaluating illegal markets is always dif�cult. Their clandestine nature makes them dif�cult to observe, and the absence of a uniform interna- tional standard of illegality renders the comparison and collation of observations a perilous exercise. But the task of evaluating human traf- �cking is even more dif�cult than evaluating other illegal markets, such as the market for drugs. In the case of the drug market, the best estimates of quantities and revenues come from national population surveys (for example, Abt Associates 2001). In the case of human traf�cking, there is no comparable mass consumer market that can provide demand-side estimates. 178 Draining Development? In 2001, there were an estimated 175 million people living outside their countries of birth, a number that had doubled since 1975 (United Nations 2002). Most immigrants were living in Europe (56 million), fol- lowed by Asia (50 million), and North America (41 million), but most of these came legally. The International Organization for Migration (IOM) is unable to provide �gures that accurately reflect the amplitude of the phenomenon of human traf�cking as a whole.3 IOM’s global database on human trade contains information on about 13,500 victims who have bene�ted from IOM assistance. This �gure in no way reflects the global reality. Until IOM can collect data through a reliable and system- atic method, the following estimates—which must be considered as crude guesses—are being used by IOM to evaluate the amplitude of the global phenomenon of the trade in humans. According to a handbook for parliamentarians, human traf�cking affects 2.5 million individuals every day (UNODC and IPU 2009). This represents a low-end estimate of the number of victims of the trade at any given moment.4 The estimate is based on analysis of publicly avail- able data published between 1995 and 2004 and presented for the �rst time in 2005. It has not been updated since. IOM considers that 800,000 victims of the trade in humans cross international borders every year or are authorized to leave or enter the territory of a country and are thereafter exploited.5 This is a high-end estimate. It was established by the U.S. government based on data col- lected in 2006. Soon, these estimates were no longer being mentioned in the annual report of the U.S. Department of State (2010) on human traf�cking in the world. In their place, the report mentioned estimates from the Inter- national Labour Organization (ILO) on forced labor, according to which 12.3 million people are victims of forced labor, servitude, forced child labor, or sexual servitude at any given moment in the world (ILO 2009). This is an estimate established by ILO in 2005 and represents stocks of forced labor on a worldwide scale. Among these 12.3 million people, 2.5 million are considered victims of human traf�cking. According to ILO, the rest of the people are victims of forced labor who are not considered victims of the trade in humans (that is, their exploitation was unrelated to their entry into the destination country). ILO also indicates that the amount of lost wages from forced labor is US$21 billion. Human Traf�cking and International Financial Flows 179 The United Nations Of�ce on Drugs and Crime has estimated the monetary pro�ts from the worldwide trade in persons at US$7 billion annually (UNODC 2009). About US$3 billion is associated with human traf�cking involving around 150,000 victims (which means approxi- mately US$20,000 per victim), and over US$6 billion is associated with the smuggling of migrants, which involves 3 million people (US$2,000 per person). The United Nations Children’s Fund estimates the pro�ts at US$10 billion (UNODC 2009; ILO 2009). ILO estimates that the total illicit pro�ts produced by traf�cked forced labor in one year at slightly less than US$32 billion (ILO 2009). Others have estimated the pro�t from commercial sexual exploitation at above US$30 billion a year. None of these estimates is convincing. The estimates on the numbers of persons involved in human traf�cking fluctuate between two extremes. At one extreme, the estimates reflect only the number of people known to have been harmed; in this case, the estimates fail to account for all the victims who are reluctant or incapable of reporting their situation. At the other extreme, the estimates include all illegal migrants, thereby fail- ing to discount for those who have not been traf�cked in any sense. The estimates of the dollar flows generated by human traf�cking are often produced for advocacy purposes to underline the magnitude of the problem. However, adding flows together that have nothing in common other than the units in which they are expressed does not produce mean- ingful values. One of the reasons many of the available data are too flawed to be useful is that they are often produced with an eye to isolat- ing human traf�cking from the illegal activities the traf�cking facilitates. Human traf�cking is, of course, a crime in itself, but its economic raison d’être is to supply a workforce for criminal or illegal activities. In this sense, human traf�cking is a means and not an end; it is one step in a complex economic subsystem. If a political agenda drives the desire to separate out this element, it is almost impossible to develop good data. The dearth of international data is not offset at the national level by good �eld studies. We know little about national �gures on human traf- �cking. Staring (2006) is an exception; he provides some data on the Netherlands. He makes a distinction between human smuggling and human traf�cking. Human smuggling in the Netherlands increased between 1994 and 2002 from 4 cases per year to 201, while human traf- �cking increased from 63 to 201. Staring notes both that the data are 180 Draining Development? scarce and that, if they are available, the methods used for data collection are unclear. Slavery and prostitution are other activities that derive from human traf�cking. One might be tempted to distinguish between criminal activities that violate universal social norms (malum in se) and activities that may be (temporarily) prohibited by law, such as homosexuality or cannabis smoking (malum in prohibitum). However, this distinction does not make a great deal of sense from a legal policy standpoint insofar as both sorts of activity trigger law enforcement. It therefore seems rea- sonable to consider that all legally punishable activity involved in human traf�cking is criminal. These activities do not all cause identical harm to society. Criminal codes reflect these differences by modulating the severity of the punish- ments associated with various offenses. Obtaining false papers for an illegal migrant or harboring a clandestine person are often only �ned, while slavery or prostitution is more severely punished. Other compo- nents of human traf�cking are even closer to the core activity of classical organized crime. Indeed, human traf�cking networks are sometimes intertwined with networks for the transportation of stolen goods or ille- gal merchandise (stolen cars, drugs). Persons involved in one type of network easily pass to another. The respective shares of the different components of human traf�ck- ing are unknown. In the public imagination, human traf�cking is often linked to the sex industry, but labor traf�cking is probably more wide- spread, given the simple fact that the world market for labor is far greater than that for sex (Feingold 1998, 2005).6 A detailed study by ILO (2005) �nds that, of the estimated 9.5 million victims of forced labor in Asia, fewer than 10 percent are traf�cked for commercial sexual exploitation. Worldwide, less than half of all traf�cking victims are traf�cked as part of the sex trade, according to the same report. Can we describe the market for human traf�cking? If we consider that organs or human beings are the merchandise exchanged, several inter- connected markets for human merchandise would have to be taken into account: a prostitution market, an organ market, and a market for illegal manpower. As in any market, the suppliers put the merchandise on the Human Traf�cking and International Financial Flows 181 market, and the purchasers buy it. The overall market would have the speci�city that, occasionally, an individual may, if he has recovered some autonomy, put himself on the market and become both supplier and merchandise. The market for human beings is not an ordinary market. The usual public policy recipes will not work in the same way they work on a clas- sical market. Criminalization, for instance, directly affects the supply of and the demand for the merchandise (prostitution, organs, slaves, and so on), but it only has a limited and delayed impact on the flow of individu- als who cross borders, whether on their own initiative or under the dom- ination of a criminal group. The important point is that the supply of immigrants is not directly affected by the repression of a market seg- ment. Thus, it takes time before repression on the markets of rich coun- tries affects the supply of migration. Numerous buffers slow the trans- mission of incentives. The Hope of a Promised Land The value of accounting is to make goods that are not homogeneous into goods that are homogeneous. A dollar adds to a dollar, but what is there in common between a dollar paid by the clandestine immigrant to his smuggler and the dollar spent by a prostitute’s customer? To understand clearly the nature of the various flows of money, we must examine and compare the corresponding components. Figure 6.1 summarizes the functioning of human traf�cking. Figure 6.1 shows that the supply of and demand for services (forced or not) in the markets for sex, illegal X-rated movies, organ transplants, slaves, and so on can be provided by a direct connection between the suppliers and the consumers in exchange (or not) for money. The sup- pliers of the criminal services can provide the connection. These crimi- nal organizations or public employees corrupted by individual criminals can offer help to the illegal workers and services to consumers. In com- pensation, they receive money, which will be laundered if the sums are big, the flows are repeated, or the activity is connected to a criminal organization. 182 Draining Development? Figure 6.1. Human Traf�cking: Supply and Demand Supply Individuals supplying (with or without coercion) their work, bodies, or internal organs Money Services Services Money Supplier of criminal services Corruption, administrative papers, border crossing, housing, running of bordellos, or providing services to the X-rated movie industry Services Money Demand Clients for sex, illegal X-rated movies, slaves, or organs for transplants Source: Author compilation. The motivations of illegal migrants An act of free will may sometimes determine the decision to seek the help of smugglers to cross a border illegally, the use of coyotes (a smug- gler of immigrants), for instance, at the Mexico–United States border (Spener 2009). In such cases, illegal immigrants momentarily place their destiny in the hands of the people who run a smuggling network and expect to recover their freedom on the other side of the border. There is a demand on the part of the individual immigrant and a supply of ser- vices through a criminal network. The decision of the immigrants may also be forced if the prospective immigrants are recruited or abused by the networks of smugglers and remain dependent on the networks once they have crossed the border. In this case, the individual becomes mer- chandise; the demand arises from those who want to import human beings, and the supply is provided by the network. Human Traf�cking and International Financial Flows 183 Thus, the supply and demand must be located differently according to whether the decision to cross the border is free or forced. Most often, the smugglers vaunt the bene�ts they offer. Often exaggerated and romanticized, the descriptions of the bene�ts are tailored to the expecta- tions or desires of the prospective immigrants. The economy of illegal immigration is based on the existence of two differentials (Cornelius et al. 2010). The �rst differential separates the wage levels in the source country and the host country. The second dis- tinguishes the standards of living in the two countries. First, the illegal immigrant provides manpower at a price that is lower than the price on the market in the host country or �lls a gap in man- power of a particular type. This differential persists over time if anti- immigration regulations are suf�ciently strict to remain a threat to the illegal immigrant. In the absence of legal barriers to entering the employ- ment market, the wages of the illegal immigrant will tend to become aligned with the wages of the local employee. Recourse to illegal man- power is also a means of exerting pressure on local employees. Histori- cally, illegal manpower is less unionized than the legal workforce, and contractors have often used the supply of illegal immigrants to force down the average union wage rate. The second differential concerns the living conditions of illegal immi- grants and of the local workforce. The fact that the illegal immigrants accept less than average wages normally means that they also live in more straitened circumstances. As long as the illegal immigrants accept or must accept such a differential, the initial characteristics remain intact (the lower wage rate). Over time and in some countries (the United States, but not Saudi Arabia, for example), the illegal immigrants will escape at least some of the initial conditions and will gain professional quali�cations and bargaining power that will enable them to gain access to better employment status than they had on arrival. The more transient the entry barriers, the more illegal immigrants can hope to melt into the legal workforce and make the individual tra- jectory of their jobs converge with that of the average employee in the country. Occasionally, an individual may escape the initial conditions rela- tively quickly. Marriage remains an important possibility. Thus, a demo- graphic gender imbalance can create a greater demand for one gender 184 Draining Development? group (generally women, as in some agricultural areas in Europe). The online dating services that offer to organize dates and marriages between rural French men and women from Asian countries often operate on the borders of legality. Moreover, some occupations seem always to require more manpower, enabling illegal immigrants to bene�t from a certain rent of scarcity. Maids from the Philippines bene�t from being Catholic and speaking English, which are valuable characteristics in the eyes of some European employers. These individuals manage to obtain hourly wages that are about 20 percent above the minimum wage. They do not have the social security advantages of the local workforce, but they can sometimes bene�t from free health care coverage (in France and some Nordic countries). Orrenius (2001), in an informative paper, notes that the earliest mod- els of migration emphasize wage differentials as the impetus to mobility. Massey et al. (1987) and Massey (1987) break down the cost of migration into direct monetary costs, information and search costs, opportunity costs, and psychic costs. They demonstr ate that access to migrant net- works reduces cost in all four categories. In the case of Mexico, access to a network gives access to more reliable coyotes, lowers the probability of apprehension, facilitates settlement, and shortens the job search. It thus appears that those who become involved in human traf�cking in the wide sense we are using in this chapter are sometimes forced to do so, but may often be considered clients who voluntarily use illegal ser- vices to immigrate. The expectation of quick improvement in earnings and the distorted bundle of information available through different channels—media, friends, family, former migrants, and smuggler net- works—provide strong incentives for migration. During their journey or once in the host country, immigrants will, perhaps, recover their power to make free decisions. The longer they stay under the influence of a net- work, a criminal organization, or an employer that forces them to work for almost nothing, the longer they will be part of human traf�cking. In the best possible con�guration, their connection with traf�cking will not last longer than the hours necessary to cross the border. Monetary compensation The monetary flows that accompany the flows of people we have described can be divided into three channels. Human Traf�cking and International Financial Flows 185 First are the immediate payments that occur at the border. These pay- ments may be spread among several individuals. For example, a corrupt employee at the consulate of a rich country located in a poor country who delivers a tourist visa will receive a bribe. Most often, visa traf�cking is the act of small groups of corrupt embassy employees. The amounts thus collected generally represent fringe bene�ts for low-level consular employees. Occasionally, the amounts are used to make small invest- ments in the home country of the employees. More often, these amounts are spent locally or enable the satisfaction of vices or double lives. These flows therefore usually remain in the country that is the source of the manpower. Globally, they are insigni�cant.7 Second, clandestine immigrants can bene�t from the paid help of a network to ease the border crossing. Snakeheads (the term used for smugglers in China) or coyotes (at the Mexican border) are most often members of a network. Sums collected by these networks represent criminal revenue. They are typically used in the country that is the source of the immigration. As a general rule, police control over the monetary flow is weak, and the money is used directly by the criminal organization without having to develop a sophisticated laundering strategy (Amuedo- Dorantes and Pozo 2005). Third, illegal immigrants may be required to stay in contact with criminal networks after their arrival in host countries. It seems plausible that, particularly in the case of Chinese networks, illegal immigrants are forced to accept the continued collection of payments after their entry and during employment. Our conclusion is highly speculative, but it appears illegal Chinese immigrants may often be forced to pay during longer periods. One explanation is that organizing the trip from China to the United States is more complicated than organizing the trip from Mexico. Hence, the price demanded for the service is higher, and the number of pay- ments larger. It may also be that Latino workers are more prone to inte- grate into the general workforce, while Chinese immigrants are more likely to be working for businesses owned by people of Chinese ethnicity. In the �rst case, the workers are illegal, but are free to choose their employers, while, in the second case, the workers are restricted to a lim- ited set of businesses. The latter workers often have to give a share of their wages to the criminal organization or repay business loans incurred 186 Draining Development? with the criminal organization. Furthermore, the language barriers in the host country are relatively greater for speakers of the many idioms of China than for Spanish-speaking immigrants. If a new immigrant cannot �nd a job independently, the connection with a criminal network remains strong. In such situations, the share of the payment made by the immigrant that is transferred to the source country is probably important. Indeed, the criminal organization may have reason to move its income as far as possible from the place where the criminal activity is ongoing. Another, more modest source of regular income for criminal organizations is the assistance provided to illegal immigrants in obtaining new identity papers or in gaining access to medical services. Flows of money generated by human traf�cking partly vanish in the host country if they are one-time payments to acquire a speci�c, time- limited service. The more personal and speci�c the illegal service, the more expensive it is. It is also most probably rendered by a criminal organization and not by an individual or small group. Additional pay- ments are scheduled over time. The money ends up in the possession of a more sophisticated criminal organization, and a good part of it is probably channeled out of the host country and sheltered in the country of origin of the criminal organization. Human Traf�cking Is Also a Business Human traf�cking is not only driven by the desire of some inhabitants of poor countries to emigrate to rich countries. The flow must be con- sidered, in part, the creation of a demand in the rich countries. Further- more, the illegal networks that engineer border crossings and organize the related manpower-intensive criminal activities in rich countries have shown the ability to innovate. Motivated by the desire to increase their pro�ts, they develop new products and services. Indeed, some criminal organizations specialize in the production of criminal goods and services. Criminal organizations are heterogeneous. Some have originated in the source countries and undertake operations in host countries. An example is the Chinese Triads. The Salvadoran Maras or Mara Salvatru- cha, in contrast, originated in the United States and then spread to the Human Traf�cking and International Financial Flows 187 home country of many of the members. The methods of these groups, particularly their management of �nancial flows, vary greatly. This het- erogeneity makes it impossible to establish a strict correlation between a given criminal activity and the criminal �nancial flows. The criminal portfolio From a legal perspective, any group of more than two people who are committing crimes together can be deemed a criminal organization. We will, however, use the term only to describe substantially larger groups working together to achieve criminal ends. Several decades have passed since Schelling (1967) and Buchanan (1973) analyzed the capacity of organized crime to impose its monopoly on criminal activities. Since then, many scholars and law enforcement agencies have conceived of organized crime groups as hierarchical, cen- tralized, and bureaucratic and have come to consider their structure analogous to the structure of a modern corporation; the Ma�a remains a paradigm (Abadinsky 1990; Cressey 1969). Pioneer work has argued against the idea that the monopoly prevails as the natural structure for large-scale criminal activity. Studying the structure of the drug market, Reuter (1983), Reuter, MacCoun, and Murphy (1990), and Kleiman (1989) have underlined the dif�culties that criminal organizations face in effectively imposing the barriers to entry necessary to preserve a monopoly. Reuter has challenged the myth of the American Ma�a, arguing that, in the 1980s, the Ma�a was only able to enjoy high levels of pro�t because of its prestige, but that it no longer possessed suf�cient power to maintain the organization at the level of its reputation. According to Reuter, the Ma�a is weakly central- ized and has high coordination costs, which are structural consequences of illegality. Insuf�cient information and a weak de�nition of property rights of each of its subgroups deprive the Ma�a of some of the pro�ts that one might otherwise expect. The more recent view argues that the concept of organized crime should be abandoned in favor of a criminal enterprise model (Block and Chambliss 1981; van Duyne 2007). Zhang and Chin (2002) note that the latter model suggests that networks are flexible and adaptive and that they can easily expand and contract to deal with the uncertainties of the criminal enterprise. Criminal organizations respond to varying market 188 Draining Development? demands by narrowing or expanding their size; their organizational structures are not so much predesigned as shaped by external social and legal factors. Certain criminal groups, such as those involved in the con- struction or gambling businesses, may require an elaborate hierarchy and a clear division of labor. These are necessary organizational prereq- uisites to remaining in the illicit business. In contrast, criminal groups such as those involved in pawnshops and the fencing of stolen goods must respond rapidly to changing demand in the streets and therefore require little or no formal organizational structure. Prostitution, porno- graphic �lms, the traf�c in organs, and the supply of illegal manpower are often monopolized by criminal organizations. These organizations respond to an existing demand, but they also create new demand. The innovative capabilities of criminal organizations enable them to convert individual preferences into commercial demand. Some data con�rm that, if human traf�cking is a big business, it is mostly a disorganized criminal business that depends on individuals or small groups linked on an ad hoc basis. According to Feingold (2005), there is no standard pro�le of traf�ckers. They range from truck drivers and “village aunties� to labor brokers and police of�cers. Traf�ckers are as varied as the circumstances of their victims. Although some traf�ck- ing victims are literally kidnapped, most leave their homes voluntarily and slip into being traf�cked on their journey. By the same token, some traf�ckers are as likely to purchase people as they are to transport them. Participation in a criminal organization requires speci�c talents. Not everyone can be a smuggler or a flesh-peddler. Not everyone is prepared to risk prison or violence. The talent is therefore uncommon and attracts a price. Smugglers. There is a scholarly consensus that most smugglers are not members of a criminal organization. Zhang and Chin (2002) and Zhang, Chin, and Miller (2007) write that Chinese human smuggling is domi- nated by ordinary citizens whose family networks and fortuitous social contacts have enabled them to pool resources to transport human car- goes around the world. Orrenius (2001) has studied the illegal immigra- tion along the Mexico–United States border and emphasizes the evi- dence for easy entry into the traf�cking industry. In theory, any migrant who has undertaken an illegal border crossing can use the experience to Human Traf�cking and International Financial Flows 189 work as a coyote. We disagree with Andreas (2000) for whom the expense involved for the traf�ckers and smugglers has led to a parallel process of centralization as smaller, poorer, and less sophisticated operators are forced out of the market. Many specialized roles have emerged in response to the particular tasks required in smuggling operations such as acquiring fraudulent documents, recruiting prospective migrants, serving as border crossing guides, driving smuggling vehicles, and guarding safe houses. These roles correspond to the successive stages of a smuggling operation: the recrui- ter is active at the beginning, and the debt collector at the end. Most of these activities take place in the host country and do not involve large amounts of money. The income they provide tends to stay in the host country and is probably quickly reinjected into the host economy through consumer spending. Smuggling activities are normally carried out following negotiations between the smugglers and their clients that leave few opportunities for others to intervene. Smuggling operations consist in secretive and idio- syncratic arrangements known only to those directly involved. Further- more, the contacts between the organizers of a smuggling operation and their clients are mostly one-on-one. This special business arrangement serves two crucial functions in an illicit economy: (1) it maximizes the smuggler’s pro�t by monopolizing the services critical to speci�c aspects of the smuggling process, and (2) it minimizes the potential exposure to law enforcement. Smugglers working along the border are the least skilled people in the traf�cking chain. The risk is limited, and the reward is small. There is not a lot of evidence to document the risk. Former border agents consider that one in four illegal migrants is arrested, but many of these migrants make multiple attempts (Slagle 2004). The rates of apprehension by bor- der patrols in Texas show large variations from one period to another (Orrenius 2001). According to Orrenius (2001), in real terms (1994 U.S. dollars), the median reported coyote price per individual for a single crossing fell from more than US$900 in 1965 to about US$300 in 1994. Higher post- 1994 prices (still below US$400) are consistent with the impact of greater enforcement on smuggler fees.8 Large pro�ts come only with the large numbers of people that can amortize �xed costs (such as regular bribes, 190 Draining Development? safe houses, and so on). Although the value of the time devoted to smug- gling should be taken into account, most of the revenue is pro�t. Low as the prices of single border crossings are, if they are multiplied by the number of illegal immigrants, they generate revenues on the order of US$700 million annually, which is roughly equivalent to the revenue generated by 30 tons of imported cocaine (300 tons are imported annu- ally).9 In other words, importing illegal immigrants is a criminal busi- ness equivalent in size to 10 percent of the business of importing cocaine. The flow of money generated by this business is, nonetheless, negligible compared with the remittances sent by Mexicans living abroad, mostly in the United States, to their families in Mexico. These remittances are a substantial and growing part of the Mexican economy: in 2005, they reached US$18 billion (World Bank 2006). A coyote’s income does not create wealth in the way drug traf�cking does. At a couple of hundred dollars per smuggled person, coyotes make too little to save. This income is mostly a transfer between individual A and individual B. The money is quickly reinvested in the economy with- out having a speci�c effect. Whether it is spent by the migrant or by the coyote, the money has the same socioeconomic impact. Violence and other costs of traf�cking. By contrast, other activities of criminal organizations create major costs and expose participants to the risk of arrest and other harms. Revenue from criminal activity should not systematically be interpreted as pro�t. We know little about the costs that criminal organizations face, which makes it dif�cult to differentiate between gross income and pro�t. We also generally confuse accounting pro�t and economic pro�t. This shows up in our omissions of important cost elements accruing to criminal organizations. For example, violence is a speci�c characteristic of crimi- nal activity. Criminals are exposed to a risk of arrest, but also to being attacked, injured, or killed by their clients or by other criminals. This is a factor in their costs. Violence directed at migrants is not a necessary part of the illicit mar- ket. It is attenuated by the social networks, cultural norms, and contrac- tual relationships in which the market is embedded (Kyle and Scarcelli 2009). Criminal organizations seem much less violent in human traf- �cking than in the drug market. Perhaps the reason is the absence of Human Traf�cking and International Financial Flows 191 territoriality. The �nal distribution of drugs requires the control of a territory in a way that is not required in human traf�cking.10 Further- more, the fact that drugs concentrate great value in a small volume cre- ates a singular opportunity because it is possible to steal the stock of drugs from a dealer. It is impossible or, at least, more dif�cult to steal a stock of organs or clandestine immigrants. Violence is therefore fairly rare among the criminal organizations involved in human traf�cking. We may thus legitimately consider that the costs are mainly reduced to the costs associated with logistics and corruption. The less competitive a market is, the less it will be innovative. Drug markets are conservative, with little product innovation relative to legal market counterparts.11 However, if a new drug is introduced on the mar- ket, waves of violence are observed. We might expect the same for crimi- nal activities based on human traf�cking, but what would constitute innovation in this �eld? The pornographic �lm industry provides a good illustration of inno- vation without (much) violence. Note that it is an extremely competitive market: production costs are low, and entry in the business is easy.12 Ini- tially, the X-rated industry offered poor-quality �lms. Little by little, the quality improved. The scenarios became more sophisticated, and the shots improved. Progressively, the industry covered all sexual tastes. Now, the catalogue of X-rated offerings allows the consumer a wide choice of sexual orientation and diverse sexual practices. The supply is thus extremely varied. Having accomplished this range, the X-rated industry has also been able to renew itself by offering home movies and then hidden cameras �lming without the knowledge of the participants. There have also been �lms involving actual violence and murder. The X-rated industry’s capacity for innovation is impressive. The industry requires a lot of labor, which is frequently imported ille- gally (Mahmoud and Trebesch 2009). Indeed, illegal immigrants often agree readily to perform in the �lms. The need for money laundering Transnational human traf�cking is a business without territory. Unlike gambling and prostitution, which tend to serve regular clients and involve activities tied to speci�c neighborhoods, human traf�cking mostly involves one-time transactions. Additionally, although many people want 192 Draining Development? to enter the United States (for example), only a limited number are eligi- ble because smugglers usually select clients based on ability to pay. Each of the payments that feed the revenue of traf�cking organiza- tions is typically small (usually only a couple of hundred dollars), but the revenue can accumulate quickly. The total income is proportional to the number of clients. Some people cross the border many times and spend a lot of money hiring coyotes, but most clients only cross a border once. The relative share of individuals using the service of a coyote only once is certainly higher than the relative share of individuals who use other criminal services, such as prostitution, only once. The structure of the revenue flows associated with traf�cking within a criminal organization depends on the speci�c activity, whether slav- ery, the traf�c in organs, the sex industry, and so on. In the case of organs, a piece of merchandise can only be sold once. The criminal organization involved carries out most of its activities on the territory of other countries besides the host country. The removal of the organs is usually undertaken in poor countries, and the organs are imported underground into developed countries where they are sold and trans- planted. The payments are thus made in the host country, while an important part of the crime is committed abroad. This separation between the location of the crime and the location of the payment makes legal prosecution more dif�cult. While prosecution for organ traf�cking is possible in the host country, it is much more dif�cult to extend the prosecution to murder even if this is thought to be the source of the organs. We do not address this issue in depth here, but we note that, in this case, the flow of money is collected in the host country. The criminal organization will then either try to launder the money on the territory of the host country or send it back to the source country, which is often also the home country of many of the members of the criminal organization (Kopp 2004). Prostitution is typically linked to organized crime (Zhang 2009; Zaitch and Staring 2009). The activity takes place in the host country, and the payments are collected there. The money is laundered on the spot or in the country of origin of the traf�ckers. The logistics of traf- �cking for prostitution and the sex industry are different from the logis- tics of organ traf�cking. Smuggling dead organs and smuggling living beings present different problems. Human Traf�cking and International Financial Flows 193 The activities linked to living beings imply that the smugglers must maintain contact with the people who are traf�cked because these peo- ple must be lodged and cared for. Given that they must establish this sort of connection with their clients, criminal organizations take measures to separate manpower and money flows. Police raids and the arrest of pros- titutes are almost inevitable; so, the criminal organization must protect its revenue by separating the merchandise (that is, the people) and the money. The money paid by the client is thus usually passed on immedi- ately to a local manager, who then promptly gives it to a collector. In contrast, organs are discretely removed in a poor country, while the purchasers usually live in rich countries. The transplantation may be carried out in a poor country with good medical infrastructure. We know little about this type of traf�c. It is plausible that an organ removed in India, for example, might be transplanted into a Swiss patient who has traveled to Romania for the purpose. In such a scenario, the flow of money is particularly dif�cult to chart. Slavery requires different �nancial circuits. Most often, the individual is offered assistance in entering a host country and then, after the entry, is deprived of his papers. He must then work for a family or, more rarely, a company.13 The migrant pays the criminal organization. Sometimes, the bene�ciary of the labor also pays a fee. The flows are irregular unless the bene�ciary wishes to change personnel. We know little about the laundering techniques used by the criminal organizations involved in human traf�cking. It is possible that these organizations have extremely large sums of money and launder them successfully. It is also possible that they possess much more sophisticated laundering systems than anything we have seen in the domain of drugs (Zaitch 1998; Reuter and Truman 2004). The taxonomy of the flows We note above that the modes of circulation of the revenue flows created by human traf�cking are heterogeneous. It is thus almost impossible to distinguish the sums that correspond to the human traf�cking segment of the operations of criminal organizations from the sums that corre- spond to the pornographic industry, prostitution, or labor. Nonetheless, in table 6.1, we attempt to classify the various flows of money generated by human traf�cking according to the most signi�cant characteristics. 194 Draining Development? Table 6.1. The Financial Flows Generated by Human Traf�cking Provider of the money Recipient Service Frequency Income Laundering Source country Illegal migrants Border patrol Visas, One time Pro�t No passports Criminal organizations Government of�cials Facilitate the Repeated Pro�t Depends on and bureaucrats traf�c the scale Border Illegal migrants Local police Passage One time Pro�t No Illegal migrants Coyotes Passage One time Pro�t No Illegal migrants Criminal organization Passage One time Pro�t, Yes less cost Host country Illegal migrants Employers Free service Repeated Value No (no wages) Illegal migrants Service providers Protection Repeated Pro�t, Depends on (flesh-peddlers, housing, services less cost the scale administrative papers) Illegal migrants Government of�cials Protection One time Pro�t No Service providers Government of�cials Protection Repeated Pro�t Depends on the scale Criminal organizations Government of�cials Protection Repeated Pro�t Depends on the scale Clients and illegal Sex workers, other Services, Repeated Pro�t, No employers workers workforce less cost Source: Author compilation. In the table, we �rst indicate the location of the monetary flow. It can be located in three places: the source country, the border, and the host country. Sometimes, �nancial flows may transit through one or more third countries. Most of the time, however, this occurs during the laun- dering process and does not appear in the table. The provider and the recipient indicate who is paying and who is receiving the money. The service gives an indication about what is compensated by the monetary flow. Frequency denotes whether the transaction takes place repeatedly or only once. Income describes whether the flow of money is pure pro�t or whether the recipient must deduct costs to calculate pro�t. The flows of money to the source country, whether originating there or returning from elsewhere, are mainly generated by corruption. Human Traf�cking and International Financial Flows 195 Whether they will or will not be laundered depends on the level of scru- tiny of local authorities. Political stability also plays a role. If the govern- ment is weak, corrupt government of�cials and bureaucrats may try to place their assets in other countries where they might plan to go if their security is threatened. These flows are repetitive if they are oriented toward high-level government of�cials. Most of the time, local bureau- crats can only be corrupted over a short period of time before they are removed from of�ce. There is almost no cost for those who receive the money. The only real cost is the risk of being prosecuted, which is not a monetary cost. This diminishes the economic pro�t, but not the account- ing pro�t. As long as corrupt of�cials are not prosecuted, the bribes they collect are both income and pro�t. Corruption may occur regularly at border crossings over long periods without this implying that any single individual maintains an income from this corruption for long. Corrupt of�cials will usually be rotated out of their positions. While their successors may be equally amenable to corruption, and thus the flow of immigrants may be maintained, the window of opportunity for any single of�cer to take advantage of the demand for illegal border crossings is relatively brief. This implies that individual recipients of bribes do not accumulate suf�ciently large sums to be considered to represent a flow requiring laundering. Instead, the sums, which generally amount simply to additional income for local border guards and policemen, are usually spent on consumption or invested with little precaution. Only if a criminal organization estab- lishes an ongoing relationship with a relatively high-level of�cial are the bribes likely to accumulate into sums large enough to warrant a sophis- ticated laundering scheme. In the host country, there is a missing monetary flow, which repre- sents the work of those who have, in fact, become slave laborers. It appears in the table as a nonmonetary flow that corresponds to a free service. Another flow represents the services provided by intermediaries who offer protection, housing, and new documents to the illegal work- ers. These payments are regular. The pro�t of the intermediaries is the same as the payments they receive as long as they do not have to spend money to provide the service. If they do incur costs, these must be deducted from the income to establish their pro�t. Generally, the larger the scale of operations, the more likely that intermediaries will incur 196 Draining Development? costs in providing services. Some illegal workers can obtain the services they need directly by paying government of�cials or bureaucrats. In such cases, the payment is usually made only once, though the corrupt of�- cials and bureaucrats may receive money regularly from many people. Most of the time, the corrupt of�cials do not have to pass on part of their pro�ts to other of�cials above them in a hierarchy, though it may be that numerous corrupt of�cials have become involved and require payment for the service activity. Criminal organizations may pay corrupt of�cials directly for visas or to turn a blind eye toward the border crossings. They include this service in the package they sell to the migrants, and it is simply a cost that must be taken into account in calculating pro�t. The breakdown of the pro�t among the various levels of the criminal organization is unclear. The protection bought from corrupt of�cials may be used in many activities (illegal drugs, human smuggling, prostitution, and so on). We lack infor- mation on the level of diversi�cation in the criminal organizations involved in human traf�cking. The flows between the criminal organiza- tions and the government of�cials are repeated. If a contact is caught and prosecuted, the criminal organization identi�es a new contact. The last flow is composed of the direct payments of clients to illegal workers (prostitutes; employers of maids, housekeepers, and gardeners; and so on). These flows are permanent. The flows compensate the ser- vices that are provided by the workers. Conclusion In this chapter, we provide a methodological contribution to the dif�cult task of assessing the impact of human traf�cking in economic terms. First, we underline the reasons to resist the temptation to add together all the flows of money that are more or less linked to human traf�cking. If the results of any analysis are to be meaningful, they must be grounded on a realistic understanding of the complex web of criminal activities that are involved in human traf�cking. The money flows associated in human traf�cking arise through distinct activities that have different implications. It is important to relate the various money flows to the speci�c criminal activities that generate them. Human Traf�cking and International Financial Flows 197 Second, we point out that, in certain circumstances, it is plausible that human traf�cking may only generate a single flow of money within a single local setting for the bene�t of a single criminal organization. However, in other circumstances, for instance, if it is mostly based on geographical competitive advantage (villages close to the border) or if it is associated with a job that involves frequent border crossings (truckers, waterfront workers, and so on), human traf�cking is a pro�table, but weakly organized activity. In this case, most of the criminal income is spent in consumption, and the remainder is invested without speci�c precautions. The monetary income from the criminal activity is diffused in the economy without noticeable effects. Third, large criminal organizations that undertake traf�cking in ille- gal workers, including women, are not necessarily involved in prostitu- tion. The biggest pro�ts, however, are derived from exploiting illegal migrants in sweatshops or houses of prostitution. Traf�cking is also nec- essary to provide a labor force for criminal activities. Traf�cking is not a particularly pro�table business per se. Most of the pro�t is made in sell- ing criminal services to end users, not in extorting money to help migrants cross borders. Fourth, we assume that the flow of money entering the source coun- tries and linked to human traf�cking is negligible in comparison with remittances sent back by the illegal immigrants to their home countries. Most of the pro�ts made by petty traf�ckers stay in the host countries. The only signi�cant flows of money back and forth to source countries and host countries are generated by large criminal organizations. It is reasonable to believe that there are considerable flows of criminal money circulating internationally, but it is almost impossible to distinguish within these flows the pro�ts generated by traf�cking and the pro�ts generated by the core businesses of large criminal organizations (prosti- tution, illegal drugs, counterfeiting, corruption, and so on). Fifth, developing countries are already aware of the negative socioeco- nomic impact of criminal money flows. We do not believe that, of this money, the small portion produced by human traf�cking generates any unique sort of harm. Good governance and civil society are harmed by the total flows of money generated by criminal activities. Distortion, lack of competiveness, and bad incentives are the well-known consequences 198 Draining Development? of criminal money. Whatever the source of such money, the impact is likely to be similar. Our research is preliminary, but it already points to a somewhat mod- est conclusion, as follows. The money flows linked to human traf�cking that enter developing countries are merged into bigger flows of criminal money. There is no evidence that these flows have a speci�c unique impact on these countries. Notes 1. Orrenius (2001) and Orrenius and Zavodny (2005) have documented the supply effects of networks; see also Mckenzie and Rapoport (2007); Sana and Massey (2005). 2. Organ traf�cking is included if it refers to killings that are motivated by the desire to sell the organs internationally. The Palermo convention and the U.S. Depart- ment of Justice do not consider organ smuggling a part of human traf�cking. 3. See the website of the IOM, http://www.iom.int/jahia/jsp/index.jsp. 4. It is unclear whether this includes only those who are enslaved or also those who use the services of smugglers. 5. Such an estimate uses traf�cking in the narrow sense, that is, only those who are exploited. 6. Feingold (2005) reports that statistics on the end use of traf�cked people are often unreliable because they tend to overrepresent the sex trade. For example, men are excluded from the traf�cking statistics gathered in Thailand because, according to national law, men cannot qualify as victims of traf�cking. 7. In Manila, an employee at the consulate of a French-speaking country demanded US$500 to issue a tourist visa to an ineligible person. According to our source, he spent three years at his desk providing not more than 20 visas per month. This last number is an approximation deduced from the waiting time for the visas. The employee’s justi�cation for a longer wait (three months) than normal for the visas for the ineligible people was the necessity to merge the visas with a substantial flow of real tourist visas. Imagine that �ve employees had under- taken the same activity: the flow per year would be US$600,000. In Moscow, a French diplomat was questioned about the origin of the money he used to buy two apartments in the most sought-after location in the city. He resigned from his job at the embassy. 8. The estimate provided by the United Nations Of�ce on Drugs and Crime is higher, close to US$2,000 (UNODC 2010). 9. At US$300 for each of 2 million entrants (allowing for some individuals, say 10 percent, to make two or three entries) yields approximately US$700 million. At Human Traf�cking and International Financial Flows 199 an import price of US$23,000 per kilogram of cocaine, 30 tons of imported cocaine would yield US$700 million. 10. Mexican drug violence is not generated by retail activities, but by the �ght for territorial control, which is a broader notion than routes and gross delivery. 11. Alcohol and tobacco are extremely innovative. Products, marketing, and pack- aging are in permanent evolution. 12. The market is less competitive at the distribution level. (Note that distribution over the Web is not the core of the pornographic �lm industry.) 13. We concentrate on human traf�cking toward rich countries and exclude cases of servitude in low-income developing countries, forced labor in war zones, and so on. References Abadinsky, H. 1990. Organized Crime. Chicago: Nelson-Hall Publishers. Abt Associates. 2001. “What America’s Users Spend on Illicit Drugs, 1988–2000.� Of�ce of National Drug Control Policy, Executive Of�ce of the President, Wash- ington, DC. Amuedo-Dorantes, C., and S. Pozo. 2005. “On the Use of Differing Money Transmis- sion Methods by Mexican Immigrants.� International Migration Review 39 (3): 554–76. Andreas, P. 2000. Border Games: Policing the U.S.-Mexico Divide. Cornell Studies in Political Economy. Ithaca, NY: Cornell University Press. Andreas, P., and M. N. J. van der Linden. 2005. “Designing Traf�cking Research from a Labour Market Perspective: The ILO Experience.� International Migration 43 (1–2): 55–73. Block, A. A., and W. J. Chambliss. 1981. Organizing Crime. New York: Elsevier North- Holland. Buchanan J. 1973. “A Defense of Organized Crime.� In The Economics of Crime and Punishment, ed. S. Rottenberg, 119–32. Washington, DC: American Enterprise Institute for Public Policy Research. Buckland, B. S. 2008. “More Than Just Victims: The Truth about Human Traf�cking.� Public Policy Research 15 (1): 42–47. Cornelius, W. A., D. FitzGerald, P. Lewin Fischer, and L. MuseOrlinoff, eds. 2010. Mexican Migration and the US Economic Crisis: A Transnational Perspective. La Jolla, CA: Center for Comparative Immigration Studies, University of California– San Diego. Cressey, D. R. 1969. Theft of the Nation: The Structure and Operations of Organized Crime in America. With an introduction by J. Finckenauer. New York: Harper- Collins. 200 Draining Development? Feingold, D. A. 1998. “Sex, Drugs and the IMF: Some Implications of ‘Structural Readjustment’ for the Trade in Heroin, Girls and Women in the Upper Mekong Region.� Refuge 17 (5): 4–10. ———. 2005. “Human Traf�cking.� Foreign Policy, 150 (September–October): 26–32. Fleming, M. H., J. Roman, and G. Farrell. 2000. “The Shadow Economy.� Journal of International Affairs 53 (2): 387–409. ILO (International Labour Organization). 2005. “A Global Alliance against Forced Labour.� Report of the Director-General. International Labour Conference, 93rd Session, Report I(B), International Labour Of�ce, Geneva. ———. 2009. “The Cost of Coercion.� Report of the Director-General. International Labour Conference, 98th Session, Report I(B), International Labour Of�ce, Geneva. Jewell, R. T., and D. J. Molina. 2009. “Mexican Migrations to the US: A Comparison of Income and Networks Effects.� Eastern Economic Journal 35 (2): 144–59. Kilmer, B., and P. Reuter. 2009. “Doped: How Two Plants Wreak Havoc on the Coun- tries That Produce and Consume Them and Everyone in Between.� Foreign Policy 175 (November/December): 2–6. Kleiman, M. A. R. 1989. Marijuana: Costs of Abuse, Costs of Control. Contributions in Criminology and Penology. Westport, CT: Greenwood Press. Kopp, P. 2004. Political Economy of Illegal Drugs. Routledge Studies in Crime and Economics. London: Routledge. Kyle, D., and M. Scarcelli. 2009. “Migrant Smuggling and the Violence Question: Evolving Illicit Migration Markets for Cuban and Haitian Refugees.� Criminal Law and Social Change 52 (3): 297–311. Mahmoud, O., and C. Trebesch. 2009. “The Economic Drivers of Human Traf�cking: Micro-evidence from Five Eastern European Countries.� Kiel Working Paper 1480, Kiel Institute for the World Economy, Kiel, Germany. Massey, D. S. 1987. “Understanding Mexican Migration to the United States.� Ameri- can Journal of Sociology 92 (6): 1372–403. Massey, D. S., R. Alarcón, J. Durand, and H. González. 1987. Return to Aztlan: The Social Process of International Migration from Western Mexico. Studies in Demography 1. Berkeley, CA: University of California Press. Mckenzie, D., and H. Rapoport. 2007. “Network Effects and the Dynamics of Migra- tion and Inequality: Theory and Evidence from Mexico.� Journal of Develop- ment Economics 84 (1): 1–24. Orrenius, P. M. 2001. “Illegal Immigration and Enforcement along the U.S.–Mexico Border: An Overview.� Economic and Financial Policy Review 2001 (Q1): 2–11. Orrenius, P. M., and M. Zavodny. 2005. “Self-Selection among Undocumented Immigrants from Mexico.� Journal of Developments Economics 78 (1): 215–40. Reuter, P. 1983. Disorganized Crime. Cambridge, MA: MIT Press. Reuter, P., R. J. MacCoun, and P. Murphy. 1990. Money from Crime: A Study of the Economics of Drug Dealing in Washington, D.C. With A. Abrahamse and B. Simon. Santa Monica, CA: Rand Corporation. Human Traf�cking and International Financial Flows 201 Reuter, P., and E. M. Truman. 2004. Chasing Dirty Money: The Fight against Money Laundering. Washington, DC: Institute for International Economics. Sana, M., and D. Massey. 2005. “Household Composition, Family Migration, and Community Context: Migrant Remittances in Four Countries.� Social Science Quarterly 86 (2): 509–28. Schelling, T. C. 1967. “Economics and Criminal Enterprise.� Public Interest 7 (spring): 61–78. Slagle, J. W. 2004. Illegal Entries. Bloomington, IN: AuthorHouse. Spener, D. 2009. Clandestine Crossings: Migrants and Coyotes on the Texas-Mexico Border. Ithaca, NY: Cornell University Press. Staring, R. 2006. “Controlling Immigration and Organized Crime in the Nether- lands: Dutch Developments and Debates on Human Smuggling and Traf�ck- ing.� In Immigration and Criminal Law in the European Union: The Legal Mea- sures and Social Consequences of Criminal Law in Member States on Traf�cking and Smuggling in Human Beings, ed. E. Guild and P. Minderhoud, 241–69. Vol. 9 of Immigration and Asylum Law and Policy in Europe. Leiden, the Netherlands: Martinus Nijhoff Publishers. United Nations. 2002. “Secretary-General Stresses ‘Clear Need’ for International Cooperation on Refugee, Migration Policy.� Press Release SG/SM/8522, Novem- ber 22. http://www.un.org/News/Press/docs/2002/SGSM8522.doc.htm. UNODC (United Nations Of�ce on Drugs and Crime). 2004. “United Nations Con- vention against Transnational Organized Crime and the Protocols Thereto.� UNODC, Vienna International Center, Vienna. ———. 2009. Global Report on Traf�cking in Persons. Vienna: UNODC. ———. 2010. The Globalization of Crime: A Transnational Organized Crime Threat Assessment. Vienna: UNODC. http://www.unodc.org/unodc/en/data-and -analysis/tocta-2010.html. UNODC (United Nations Of�ce on Drugs and Crime) and IPU (Inter-Parliamen- tary Union). 2009. “Combating Traf�cking in Persons: A Handbook for Parlia- mentarians.� Handbook 16, UNODC, Vienna. U.S. Department of State. 2010. Traf�cking in Persons Report, 2010. Washington, DC: Of�ce To Monitor and Combat Traf�cking in Persons, U.S. Department of State. http://www.state.gov/g/tip/rls/tiprpt/2010/. van Duyne, P. C. 2007. “Virtue and Reality: An Introductory Tale of Two Cities.� In Crime Business and Crime Money in Europe: The Dirty Linen of Illicit Enterprise, ed. P. C. van Duyne, A. Maljevic, M. van Dijck, K. von Lampe, and J. Harvey, 1–14. Cross-Border Crime Colloquium 7. Nijmegen, the Netherlands: Wolf Legal Publishers. World Bank. 2006. Global Economic Prospects 2006: Economic Implications of Remit- tances and Migration. Washington, DC: World Bank. http://go.worldbank.org /0ZRERMGA00. Zaitch, D. 1998. The Dutch Cocaine Market in European Perspective. Amsterdam: Uni- versity of Amsterdam. 202 Draining Development? Zaitch, D., and R. Staring. 2009. “The Flesh Is Weak, the Spirit Even Weaker: Prostitu- tion Clients and Women Traf�cking in the Netherlands.� In Prostitution and Human Traf�cking: Focus on Clients, ed. A. Di Nicola, A. Cauduro, M. Lombardi, and P. Ruspini, 67–121. New York: Springer Science+Business Media. Zhang, S. X. 2009. “Beyond the Natasha Story: A Review and Critique of Current Research on Sex Traf�cking.� Global Crime 10 (3): 178–95. Zhang, S. X., and K. L. Chin. 2002. “Enter the Dragon: Inside Chinese Human Smug- gling Organizations.� Criminology 40 (4): 737–68. Zhang, S. X., K. L. Chin, and J. Miller. 2007. “Women’s Participation in Chinese Transnational Human Smuggling: A Gendered Market Perspective.� Criminol- ogy 45 (3): 699–733. Part III To What Extent Do Corporations Facilitate Illicit Flows? 7 Transfer Price Manipulation Lorraine Eden Abstract Multinational enterprises (MNEs) are powerful actors in the global economy. Transfer price manipulation (TPM) is one of the bene�ts of multinationality. By over- and underinvoicing intra�rm transactions, multinationals can arbitrage and take advantage of differences in gov- ernment regulations across countries. This chapter explains the motiva- tions of �rms that engage in TPM, illustrates the ways in which MNEs can arbitrage government regulations, and reviews the empirical esti- mates on TPM for developed and developing countries. We conclude that the strongest and clearest evidence of TPM comes from transaction- level studies of U.S. intra�rm import and export prices and from �rm- level studies using Chinese tax data. No data set is perfect; so, the various estimates are flawed. Still, the balance of the evidence suggests that income shifting does occur through the manipulation of transfer prices. What is needed is greater accessibility to transaction-level data on cross- border export and import transactions and on MNE income statements and balance sheets. This would enable scholars to shine more light in the dark corners of TPM and provide more accurate assessments of TPM impacts on developing countries. 205 206 Draining Development? Introduction World Investment Report 2011 contains the estimate that there are now 103,786 MNE parent �rms and 892,114 foreign af�liates (UNCTAD 2011). These numbers have grown enormously since the �rst United Nations Conference on Trade and Development (UNCTAD) estimate of 35,000 parents and 150,000 foreign af�liates (UNCTAD 1992). Not only are there ever-growing numbers of MNEs, their relative size as a share of the global economy is also growing. Of the world’s 100 largest econo- mies, 42 are MNEs, not countries, if one compares �rm revenues with country gross domestic product (GDP). The value added by MNEs con- stitutes about 11 percent of world GDP. Multinationals likewise bulk large in terms of international trade flows. For example, World Investment Report 2010 estimates that foreign af�liate exports are now one-third of world exports (UNCTAD 2010). If trade occurs between related parties (that is, between af�liated units of an MNE), the transactions are referred to as intra�rm trade. Statistics on intra�rm trade are scarce because most governments do not require MNEs to report their crossborder intra�rm transactions separately from their trade with unrelated parties. The United States is one of the few countries to report intra�rm trade statistics; the U.S. Census Bureau (2010) estimates that 48 percent of U.S. exports and 40 percent of U.S. imports represent trade between related parties. The Organisation for Economic Co-operation and Development (OECD) provides prelimi- nary estimates that related-party trade represents 7–12 percent of world merchandise trade and 8–15 percent of OECD trade (OECD 2010). The price of an intra�rm transfer is called a transfer price, and transfer pricing is the process by which the transfer price is determined. Transfer pricing, once an obscure area studied only by a few academics such as Hirshleifer (1956, 1957), Horst (1971), and Rugman and Eden (1985), has now become front-page news because of recent attention to TPM, that is, the over- or underinvoicing of transfer prices by MNEs in response to external pressures such as government regulations (for instance, taxes, tariffs, exchange controls). For example, Forest Laboratories, a U.S. pharmaceutical company, was pro�led in Bloomberg Businessweek for using TPM to shift pro�ts on Lexa- pro, an antidepressant drug, from the United States to Bermuda and Ire- Transfer Price Manipulation 207 land. The headline, “U. S. Companies Dodge US$60 Billion in Taxes with Global Odyssey,� compared transfer pricing with the corporate equivalent of the secret offshore accounts of individual tax dodgers (Drucker 2010a). Google was similarly pro�led in October 2010 for using transfer pricing and a complex legal structure to lower its worldwide tax rate to 2.4 percent (Drucker 2010b). Politicians have also become involved. The Ways and Means Committee of the U.S. House of Representatives held a public hearing on transfer pricing in July 2010. One report for the hearing, by the Joint Committee on Taxation, explored six detailed case studies of U.S. multinationals using transfer pricing to reduce their U.S. and worldwide tax rates.1 Sikka and Willmott summarize several recent cases involving transfer pricing and tax avoidance, concluding that “transfer pricing is not just an accounting technique, but also a method of resource allocation and avoidance of taxes that affects distribution of income, wealth, risks and quality of life� (Sikka and Willmott 2010, 352). Several coinciding forces have raised the visibility of transfer pricing from the academic pages of economics journals to the front pages of major newspapers. All three key actors in the global economy—govern- ments, MNEs, and nongovernmental organizations (NGOs)—now view transfer pricing as critically important. Government authorities, the �rst set of key actors, have long recog- nized that transfer prices can be used by MNEs to avoid or evade national regulations. For example, by setting a transfer price above or below the market price for a product and shifting pro�ts to an af�liate taxed at a lower rate, an MNE can reduce its overall tax payments and achieve a higher after-tax global pro�t relative to two �rms that do not have such an af�liation arrangement. Most governments of industrialized nations now regulate the transfer prices used in the calculation of corporate income taxes (CITs) and cus- toms duties. The worldwide regulatory standard is the arm’s-length standard, which requires that the transfer price be set equal to the price that two unrelated parties have negotiated at arm’s length for the same product or a similar product traded under the same or similar circum- stances with respect to the related-party transaction (Eden 1998). The arm’s-length standard, �rst developed in the United States, became an international standard when the OECD issued transfer pricing guide- lines that were later adopted by OECD member governments (OECD 208 Draining Development? 1979). The purpose behind the guidelines is to prevent undertaxation and overtaxation (double taxation) of MNE pro�ts by national tax authorities (Eden 1998). Starting in the early 1990s, the legal landscape for transfer pricing changed dramatically. In 1994, the U.S. Internal Revenue Service’s major revisions to its transfer pricing regulations became law. In 1995, the OECD published revised transfer pricing guidelines, which have been regularly updated since then (OECD 1995). While only a few govern- ments had detailed transfer pricing regulations attached to their CIT laws before the 1990s, now more than 40 tax jurisdictions around the world—including all the OECD countries and, therefore, the bulk of world trade and foreign direct investment flows—are covered and have highly technical and sophisticated transfer pricing regulations attached to their tax codes (Eden 2009; Ernst & Young 2008). In 1995, UNCTAD surveyed national governments about current developments in accounting and reporting and asked a few questions about TPM as part of the survey. The results for 47 countries were circu- lated in an unpublished working paper and analyzed in Borkowski (1997). In the survey, 60 percent of government respondents stated that MNEs in their country engaged in TPM and that it was a signi�cant problem. Almost 80 percent of respondents believed that MNEs were using TPM to shift income, and 85 percent thought this was a serious problem. This belief held regardless of income levels: 11 of 13 low-income, 16 of 18 middle-income, and 13 of 16 high-income country governments believed that TPM was being used to shift income. TPM was seen as a problem because it led to distorted competitiveness between local �rms and MNEs, enabled MNEs to withdraw funds from the country, and reduced tax rev- enues. Governments perceived income shifting by foreign MNEs as more frequent and larger in magnitude relative to home country MNEs. In the late 1990s, the OECD also launched a major push to deter abu- sive tax practices, focusing on abusive tax havens (OECD 1998; Eden and Kudrle 2005). While tax havens and TPM are two separate topics and should be treated separately, there are overlaps. Tax havens can encour- age abusive transfer pricing practices, for example, by creating CIT dif- ferentials between countries that offer pricing arbitrage opportunities to MNEs. Secrecy havens provide opportunities for parking MNE pro�ts away from the eyes of national tax authorities (Kudrle and Eden 2003). Transfer Price Manipulation 209 The second key actor in the global economy—MNEs—have long seen transfer pricing as an important international tax issue (Ernst & Young 2010). From an international tax perspective, tax avoidance (tax plan- ning that complies with the law) is viewed by MNE executives and the tax planning industry as both legal and morally acceptable (Friedman 1970).2 Because the goal of the �rm is to maximize shareholder wealth and because transfer pricing can raise the MNE’s after-tax pro�ts on a worldwide basis, transfer pricing is a valued activity for the MNE. Trans- fer pricing has also become an increasingly important issue for MNE executives because government regulations have become more complex and the number of governments that regulate transfer pricing, require documentation, and levy penalties continues to grow (Eden 2009; Ernst & Young 2010). In addition, recent U.S. legislative changes such as Sarbanes-Oxley and FIN 48 have made transfer pricing important for MNE executives from a corporate �nancial and reporting perspective (Ernst & Young 2008). Lastly, with the collapse of Enron, WorldCom, and others in the early 2000s and bankruptcies or near bankruptcies among many huge multi- nationals (for example, General Motors) during the current interna- tional �nancial crisis, NGOs are now paying more attention to corporate fraud, in particular to abusive �nancial behaviors that may be related to the global �nancial crisis (for example, see TJN 2007; Christian Aid 2009; Sikka and Willmott 2010). Transfer pricing has been speci�cally attacked by NGOs. For example, Christian Aid has published reports arguing that transfer pricing is tax dodging, cooking the books, secret deals, or scams that rob the poor to keep the rich tax-free, thereby strip- ping income from developing countries (Christian Aid 2009). Thus, the third key actor in the global economy—NGOs—is now also paying much more attention to transfer pricing. The purpose of this chapter is to review the literature on empirical estimates of TPM, focusing, where possible, on developing countries. We also situate this chapter within the context of the work on illicit flows of funds out of developing countries, the theme of this volume. In the next section, we explore the reasons why MNEs engage in TPM. The follow- ing section examines expected manipulation patterns in response to par- ticular forms of government regulation. The penultimate section reviews the empirical evidence on TPM. The last section concludes. 210 Draining Development? Why Engage in Transfer Price Manipulation? The primary reason why governments have developed the arm’s-length standard is that they believe MNEs do not set their transfer prices at arm’s length, but rather engage in widespread TPM for the purpose of avoiding or evading government regulations. TPM is the deliberate over- or underinvoicing of the prices of products (goods, services, and intan- gibles) that are traded among the parent and af�liates of an MNE. Over- pricing inbound transfers and underpricing outbound transfers can be used to move pro�ts out of an af�liate located in a high-tax jurisdiction or from a country that does not allow capital remittances. Differences in CIT rates across countries (which are exacerbated by tax havens and tax deferral) create pro�table opportunities for MNEs to engage in TPM. Below, we explore the bene�t to the MNE of engaging in TPM and the different ways that MNEs respond to these external motivations. One of the less well known bene�ts of multinationality is the ability to arbitrage differences in government regulations across countries (Eden 1985). One bene�t of internalizing transactions rather than using the open market is that the goals of �rms change, and this change makes a big difference. The goals of parties to an intra�rm (intracorporate) transaction are cooperative (the purpose is to maximize joint MNE pro�t), whereas the goals of arm’s-length parties are conflictual (to max- imize their individual pro�ts), that is, units of an MNE collude rather than compete in the market. This gives them the ability to reduce overall tax payments and avoid regulatory burdens by under- or overinvoicing the transfer price, the price of the intra�rm transaction. The fact that two related parties (parent and subsidiary or two sister subsidiaries) can collude in setting the price gives an MNE the ability, which unrelated �rms do not legally have, to choose a price that jointly maximizes their pro�ts, the pro�t-maximizing transfer price. The deter- mination of a pro�t-maximizing transfer price is a complex decision- making process because the MNE must take into account both internal motivations (the costs and revenues of the individual MNE af�liates) and external motivations (the existence of external market prices and government regulations such as taxes and tariffs) that can affect the opti- mal transfer price (Hirshleifer 1956, 1957; Horst 1971; Eden 1985). Where no external market exists, the MNE should set the transfer price Transfer Price Manipulation 211 equal to the shadow price on intra�rm transactions; generally, this is the marginal cost of the exporting division. Ef�cient transfer prices for ser- vices and for private intangibles have similar rules; they should be based on the bene�t-cost principle, that is, each division should pay a transfer price proportionate to the bene�ts it receives from the service or intan- gible (Eden 1998). Where external market prices exist, they should be taken into account in setting the transfer price, that is, divisions should be allowed to buy and sell in the external market. In each case, the pur- pose is ef�cient resource allocation within the MNE group. While economists focus on the TPM bene�ts in using the pro�t- maximizing transfer price, managers of MNEs are more likely to focus on practical considerations. First, there may be cases where the MNE has no internal reasons for setting a transfer price; transactions may be small in volume, dif�cult to value, or occur with extraordinary rapidity. Con- ventional accounting practice, for example, generally defers valuation of intangible assets until there are arm’s-length purchases or sales, creating the balance sheet item “goodwill,� which measures the excess purchase price over the fair value of the assets acquired. In such cases, the MNE may ignore the issue altogether and not set transfer prices. However, in the typical case, the MNE has multiple internal motiva- tions for setting a transfer price. Some of these internal motivations include the performance evaluation of pro�t centers, motivating and rewarding the managers of foreign subsidiaries, preventing intersubsid- iary disputes over intermediate product transfers, more ef�cient track- ing of intra�rm flows, and so on (Borkowski 1992; Cravens 1997; Tang 1993, 1997, 2002). TPM in these situations involves balancing the incen- tives, reporting, and monitoring activities associated with setting trans- fer prices for internal ef�ciency. MNEs typically also have a variety of external motivations for setting transfer prices. In a world of CIT differentials and tariffs, the MNE has an incentive to manipulate its transfer prices to maximize its global net- of-taxes pro�ts. The bene�ts from TPM of real flows, as opposed to �scal transfer pricing, must, however, be traded off against internal distor- tions. The primary purpose of setting transfer prices now becomes global after-tax pro�t maximization; such transfer prices, however, may or may not look like the regulatory methods outlined by national gov- ernment authorities. 212 Draining Development? Transfer Price Manipulation and Government Regulations Corporate income taxes and transfer price manipulation The most well known external motivation for manipulating transfer prices is the differences in CIT rates between countries or between states within countries. For evidence on tax-induced motivations for TPM, see Li and Balachandran (1996); Eden (1998); Swenson (2001); and Bartels- man and Beetsma (2003). There are several ways to engage in TPM in such cases: • The MNE can overinvoice tax-deductible inbound transfers into high-tax countries and underinvoice them into low-tax countries. This shifts corporate pro�ts from high-tax to low-tax jurisdictions. Examples of inbound transfers include imported parts and compo- nents, payments for engineering and consulting services, and royalty payments for intangibles. • The MNE can underinvoice taxable outbound transfers from high- tax countries and overinvoice them from low-tax countries. This shifts corporate pro�ts from high-tax to low-tax jurisdictions. Exam- ples of outbound transfers include exports of �nished goods, charges for the provision of services to other parts of the MNE network, and licensing and royalty payments for outbound intangible transfers. • If the home government allows deferral of CITs on MNE foreign source income, the MNE can avoid the home country CIT on foreign source income by not repatriating foreign source earnings to the home country. The funds can either be reinvested in the host country or moved to another country in the MNE network. • If the host country levies a withholding tax on the repatriated pro�ts of foreign af�liates and the withholding tax is not fully creditable against the home country tax, not repatriating foreign source income avoids the tax. In these cases, the MNE can use the rhythm method to time its repatriated earnings only in tax years when the withholding tax is fully credited against the home tax (Brean 1985). In other years, no pro�ts are remitted to the home country. • If withholding taxes vary according to the form of repatriation (for example, management fees, royalty and licensing payments, and divi- dends are typically subject to quite different withholding tax rates), the MNE can move the funds out in the form that incurs the lowest Transfer Price Manipulation 213 withholding tax. Note that if the MNE receives a full foreign tax credit against the home country tax for the withholding tax, the form by which the MNE moves the funds out becomes irrelevant since the tax is fully credited. • Tax holidays can also be a motivation for TPM, particularly if the holi- day is conditional on the pro�ts earned by the foreign af�liate. In China in the 1980s, the government offered a tax holiday for foreign �rms as long as they did not show a pro�t. Not surprisingly, the foreign af�liates did not show pro�ts until after the law was changed (see below). • Some forms of intra�rm transfers are more fungible than others and therefore more susceptible to TPM. Management fees are particularly notorious because the MNE parent charges each af�liate for the costs of services provided by the parent to the af�liates, and these charges are dif�cult to measure. Many host country tax authorities have speci�c rules limiting the deductibility of the management fees charged by an MNE parent to its foreign subsidiaries because governments see these deductions as a method to eviscerate the host country’s national tax base. Governments also levy withholding taxes, in addition to CITs, if foreign af�liates repatriate income to their parent �rms. The withhold- ing taxes on management fees are often in the range of 30–35 percent. • Some forms of intra�rm transfers are easier to misprice simply because there is no open market for the product. (The product is never exchanged between arm’s-length parties; an example is highly sophisticated new technologies.) So, arm’s-length comparables are impossible to �nd. Payments for intangible assets are particularly sus- ceptible to TPM because there are often no outside transfers available to determine an arm’s-length comparable. Trade taxes and transfer price manipulation Trade taxes provide a second external motivation for manipulating trans- fer prices. For evidence on tariff-induced motivations for TPM, see Eden (1998), Vincent (2004), Goetzl (2005), and Eden and Rodriguez (2004). Some examples of TPM motivated by trade tax include the following: • If customs duties are levied on an ad valorem (percentage) basis, the MNE can reduce the duties paid if it underinvoices imports. Speci�c or per-unit customs duties cannot be avoided by over- or underinvoicing. 214 Draining Development? • If export taxes are levied on an ad valorem basis, the MNE can reduce the export taxes paid if it underinvoices exports. Speci�c export taxes cannot be avoided through TPM. • Rules of origin within free trade areas offer another potential arbi- trage opportunity. Most rules of origin are on a percent-of-value basis. For example, products qualify for duty-free status if 50 percent or more of the value added is derived from inside one of the countries that is a partner in the free trade area. By overinvoicing the value added, the MNE can more easily meet a rule-of-origin test and qualify for duty-free entry for its products into another country in the free trade area. Foreign exchange restrictions and transfer price manipulation A third external motivation for TPM is foreign exchange restrictions. For evidence on foreign exchange rate motivations for TPM, see Chan and Chow (1997a), who �nd that foreign MNEs were engaged in TPM to shift pro�ts out of China not because of CIT differentials (in fact, Chinese tax rates were lower than elsewhere), but to avoid foreign exchange risks and controls. Examples of TPM in response to exchange rate restrictions include the following: • If the host country’s currency is not convertible so that the MNE can- not move its pro�ts out, the MNE can, in effect, move its pro�ts out despite the nonconvertible currency if it overinvoices inbound trans- fers and underinvoices outbound transfers. • If there are foreign exchange restrictions on the amount of foreign currency that can be bought or sold in a particular time period, using overinvoicing of inbound transfers and underinvoicing of outbound transfers enables the MNE to move more funds out than would be permissible with currency controls. Political risk and transfer price manipulation Another area susceptible to TPM is political risk. For evidence on the capital flight and political risk motivations for TPM, see Gulati (1987); Wood and Moll (1994); Baker (2005); and de Boyrie, Pak, and Zdano- wicz (2005). Examples of TPM in response to various types of political risk include the following: Transfer Price Manipulation 215 • If the MNE fears expropriation of its assets in a host country or, more generally, if political risk is great, overinvoicing of inbound transfers and underinvoicing of outbound transfers can be used to shift income out of the high-risk location. • More generally, policy risk, that is, the risk that the government may change its laws, regulations, or contracts in ways that adversely affect the multinational, also provides an incentive for MNEs to engage in TPM. Policy risks, as discussed in Henisz and Zelner (2010), are opaque and dif�cult to hedge, and there is typically no insurance. Moreover, the authors estimate that policy risks have risen substan- tially since 1990. Income shifting through TPM may well be a rational response to policy risk. • If the host country currency is weak and expected to fall, the MNE can underinvoice inbound transfers and overinvoice outbound trans- fers to shift pro�ts out of the weak currency. • Another form of political risk is the requirement that foreign af�liates must take on a forced joint venture partner. Foreign af�liate pro�ts earned in a country with this rule in effect suffer a tax because pro�ts must be split between the MNE and the joint venture partner. In these cases, the MNE may engage in income shifting to move funds out of such a high-tax location. Empirical Evidence on Transfer Price Manipulation Let us now turn to empirical work on TPM. The evidence that multina- tionals engage in TPM to arbitrage or avoid government regulations does exist, but is fragmented and often backward induced, that is, esti- mated indirectly rather than directly. The extent and signi�cance of TPM, especially for developing countries, is not clear, although, by extrapolating from existing studies, we now have a better understanding of the circumstances under which TPM is likely to occur. Below, we review the empirical evidence on TPM. Evidence from developed countries By far, the bulk of empirical research on TPM has been done using U.S. data sets, and almost all studies have been done on U.S. multinationals with controlled foreign corporations overseas. 216 Draining Development? Income shifting studies. Perhaps the largest number of empirical studies has involved estimates of income shifting from high-tax to low-tax juris- dictions. In these studies, either foreign direct investment or a pro�t- based measure such as return on assets or return on sales is the depen- dent variable that is used to test whether MNEs shift income to locations with lower CIT rates (for example, see Bartelsman and Beetsma 2003; Grubert and Mutti 1991; Grubert and Slemrod 1998). Because TPM affects MNE pro�ts at the country level, this approach focuses not on prices, but on pro�t shifting. These partial and general equilibrium models use national CIT differentials and custom duty rates to predict TPM and to estimate the income moved in this fashion. These studies are more appropriately identi�ed as �scal TPM because they focus on income manipulation and not on the pricing of products per se. Changes in the form of pro�t remittances from royalties to dividends and the excessive padding of management fees, are examples. A few of the key studies are reviewed below. In one of the earliest tests, Grubert and Mutti (1991) used a data set of manufacturing af�liates of U.S. MNEs in 33 countries to examine how tax differentials and tariffs generate income transfers because of TPM. The authors aggregated country-level data from the U.S. Bureau of Eco- nomic Analysis on foreign af�liates of U.S. multinationals and regressed the pro�t rates of the af�liates against host country statutory CIT rates. The authors concluded that the empirical evidence was consistent with MNE income shifting from high- to low-tax jurisdictions. Harris, Morck, and Slemrod (1993), based on a sample of 200 U.S. manufacturing �rms over 1984–88, �nd that U.S. MNEs with subsidiar- ies in low-tax countries pay less U.S. tax, while those with subsidiaries in high-tax countries pay relatively more U.S. tax per dollar of assets or sales. Income shifting by the largest MNEs is, they argue, primarily responsible for these results. These studies provide, however, only indi- rect evidence of TPM. For example, the results of the authors can be explained by MNEs shifting income from high- to low-tax locations, but also by cross-country differences in the intrinsic location-speci�c pro�t- ability of MNE subunits. The authors are aware of this possibility, but show evidence that does not support this interpretation. Grubert (2003) uses �rm-level data on U.S. parents and their foreign controlled af�liates to test for evidence of income shifting. He regresses Transfer Price Manipulation 217 pretax pro�ts against host country statutory tax rates, while controlling for parent and subsidiary characteristics and �nds evidence supporting income shifting, particularly among �rms with high ratios of research and development to sales. McDonald (2008) expands on Grubert (2003) by attempting to sepa- rate out income shifting arising because of tangibles, research and devel- opment, marketing intangibles, and services. In particular, he addresses income shifting through cost-sharing arrangements. He concludes that the empirical results from his tests are “not inconsistent with the existence of possible income shifting� and that there is some evidence that foreign af�liates whose U.S. parents engage in cost-sharing arrangements may also “engage in more aggressive income shifting� (McDonald 2008, 30). Clausing (2009) uses U.S. Bureau of Economic Analysis data for 1982– 2004 to test whether U.S. multinationals engage in income shifting to low- tax locations. She argues that MNEs can engage in both �nancial and real types of tax avoidance. Financial avoidance is estimated by comparing tax differentials across countries and foreign af�liate pro�t rates, while real avoidance is estimated by comparing tax differentials and foreign employ- ment. She �nds that a host country statutory CIT rate 1 percent below the U.S. rate is associated with a 0.5 percent higher foreign af�liate pro�t rate; using these estimates, she argues that US$180 billion in CIT had been shifted out of the United States. The losses as measured by real responses to income shifting, however, were about half that: about US$80 billion less in pro�ts and 15 percent less in tax revenues. The most recent paper on the topic, by Azémar and Corcos (2009), uses a sample of Japanese MNEs and �nds greater elasticity of invest- ment to statutory tax rates in emerging economies if foreign af�liates are wholly owned by research and development–intensive parent �rms. The authors argue that this is indirect evidence of TPM. TPM trade mispricing studies. A second approach to estimating TPM has involved examining individual transactions using huge databases of transaction-level import and export data (for example, see Clausing 2003; Swenson 2001; Eden and Rodriguez 2004). The focus of these authors is trade mispricing, that is, the under- or overinvoicing of imported and exported goods in response to CIT differences, tariffs, foreign exchange restrictions, and so on. Where the researcher focuses on TPM, we call 218 Draining Development? such studies TPM trade mispricing studies. The idea behind this research is to compare arm’s-length comparable prices with the reported intra�rm prices to determine the extent of over- or underinvoicing of related-party transactions. In the trade mispricing papers, the estimates of trade mispricing are typically done using regression analysis. It may be helpful for the reader if we explain briefly how these models work. If Pijkt is the transaction price of product i imported by �rm j from country k at time t, the researcher regresses the price (the dependent variable) against a vector of independent variables (product and �rm characteristics and tax and tariff rates), a dummy variable for related- party trade (1 for transactions between related parties, 0 if the �rms are unrelated), and a set of control variables for other possible explanations. The regression equation takes the following form: Pijkt = a + b Xijkt + c I FT + d Xijkt × IFT + Z, (7.1) where Xijkt is the vector of independent variables, IFT is the intra�rm trade dummy variable, and Z is a vector of control variables. The IFT variable is interacted with the independent variables to see whether there are statistically signi�cant differences between arm’s-length and related-party trade if one of the independent variables changes, for example, whether IFT is sensitive to differences in CIT rates or to cus- toms duties. All variables except IFT are normally logged so that the regression coef�cients are elasticities, showing the responsiveness (per- centage change) in the import price to a percentage change in the inde- pendent variables. It may be helpful to also see this equation as a graph. Figure 7.1 shows a simple regression that relates trade prices to the volume of trade, hold- ing other influential variables constant (for example, product and indus- try characteristics). By examining and adding up the outliers, one can estimate the extent of mispricing. The key issue is therefore to determine what is an outlier. Under U.S. Code section 6662 on CIT, transfer pricing misevaluation penalties apply if MNEs set their transfer prices signi�- cantly outside the arm’s-length range of acceptable prices. The arm’s- length range, according to the section 482 regulations, is determined by the transfer pricing method that gives the most reliable measure of an Transfer Price Manipulation 219 Figure 7.1. Estimating Trade Mispricing from International Trade Data Outliers (outside of interquartile range) Upper bound (75th percentile) transaction price Fitted regression Lower bound (25th percentile) X (product, firm, and country characteristics) Source: Author. arm’s-length result. In establishing the range, the bottom 25 percent and top 25 percent of observations are normally discarded, leaving the inter- quartile range (between 25 and 75 percent) as the acceptable arm’s- length range. If the MNE’s transfer price falls within this range, no sec- tion 6662 penalties apply. If it is outside the range, unless the �rm has demonstrated a good faith effort to comply with the section 482 regula- tions (for example, by �ling complete, contemporaneous documenta- tion of its transfer pricing policies), penalties do apply (Eden 1998; Eden, Juárez, and Li 2005). We have therefore used the interquartile range to mark outliers in �gure 7.1 because these outliers would not normally fall within the arm’s-length range. There are, of course, problems with the TPM trade mispricing approach, not the least of which is, �rst, that it is critically important to identify which transactions occur within the MNE and which are arm’s- length transactions. Too often, the studies attribute all trade mispricing to the MNE, without having the data to determine whether the trade moved within the MNE or not. Second, the key to TPM trade mispricing studies is that the data set must include both arm’s-length and intra�rm interna- tional transactions, with a clear marker that distinguishes transactions 220 Draining Development? between related parties and transactions between arm’s-length �rms. The huge advantage of TPM trade mispricing studies over income shifting studies is that the data are transaction-level (not �rm-level) data, but the problem is that the marker may be wrong or missing. Two data sets have been used in the U.S. studies of trade mispricing: U.S. Bureau of the Census data and U.S. Bureau of Labor Statistics (BLS) data. The census data tapes are raw tapes of U.S. export and import transactions. Export �gures are reported directly to the Census Bureau; import �gures come from the U.S. Customs Service. The data, while extraordinarily rich, suffer from several problems, the most important of which for TPM trade mispricing studies is that the intra�rm trade marker is highly problematic. If the data are taken from shipping docu- ments reported to customs authorities, such markers are widely recog- nized to be unreliable (Eden and Rodriguez 2004). As a result, empirical studies of trade mispricing using the census data tapes cannot reliably argue that trade mispricing estimates are also estimates of TPM. For example, Pak and Zdanowicz (1994) use the census data on monthly merchandise export and import prices to look for outliers; they estimate that the U.S. government lost US$33.1 billion in tax revenues because of unreported taxable income. Unfortunately, the authors can- not identify individual transactions as arm’s-length or intra�rm transac- tions; so, they should not (but do) attribute the tax loss to TPM. Moreover, simply using U.S. Census Bureau data on U.S. merchandise import and export transactions can be quite problematic not only because the intra�rm trade marker may be missing or wrong, but also because prices may be reported for different quantities, leading to spuri- ous estimates of TPM. If the comparability of units cannot be con�rmed, it is possible to �nd huge variations in prices and attribute this to trade mispricing, though the variations are simply caused by differences in units (Eden and Rodriguez 2004; Kar and Cartwright-Smith 2008; Nitsch 2009). For example, comparing the price of a single boxed tooth- brush with the price of a freight load of toothbrushes could lead a researcher to conclude that U.S. imports from the United Kingdom were overstated by US$5,655 per unit if the unit in each case were reported as a box (Pak and Zdanowicz 2002). Without the ability to clean the data set to ensure comparability of quantities, such estimates must be regarded as problematic. In addition, the method is only as good as the Transfer Price Manipulation 221 all-else-being-equal variables, that is, the controls used to ensure compa- rability between the controlled and uncontrolled transactions. As we stress above, to measure TPM, one must know what the arm’s-length price is for comparison purposes, which requires close comparables. U.S. import data from the U.S. Census Bureau have also been used by Swenson (2001) at the product level by country to test for evidence of TPM over 1981–86. Swenson constructs prices by dividing reported cus- toms values by reported quantities. She �nds that a 5 percent fall in for- eign CIT rates caused a tiny rise in U.S. import prices. However, she is unable to separate intra�rm from arm’s-length trade; so, her study suf- fers from the same problem as the Pak and Zdanowicz studies. Recently, Bernard, Jensen, and Schott (2008) used U.S. export transac- tions between 1993 and 2000 to examine the wedge between arm’s-length and related-party transactions and determine how the wedge varies with product and �rm characteristics, market structure, and government pol- icy. Their data set is compiled from U.S. export data and U.S. establish- ment data from the U.S. Census Bureau, as well as U.S. Customs Bureau data. Because they use census data, their results also suffer from unreli- able coding for related-party transactions. The price wedge is measured as the difference between the log of the averaged arm’s-length prices and the log of the related-party prices at the �rm-product-country level. The authors �nd that export prices for arm’s-length transactions are, on aver- age, 43 percent higher, all else being equal, than prices for intra�rm exports. The wedge varies by type of good and is smaller for commodities (8.8 percent, on average) than for differentiated goods (66.7 percent, on average). The wedge is larger for goods shipped by larger �rms with higher export shares and in product-country markets served by fewer exporters. The wedge is negatively associated with the foreign country’s CIT rate and positively related to that country’s import tariff. A 1 percent drop in the foreign tax rate increases the price wedge by about 0.6 per- cent. Lastly, they �nd that a 10 percent appreciation of the U.S. dollar against the foreign currency reduces the wedge by 2 percent. In two situations, however, the quality of the intra�rm trade marker should be quite high, as follows: (1) if government agencies keep data on speci�c products that are of high salience or (2) if the government is concerned about related-party transactions. In these cases, more than normal attention is paid to data quality and therefore to whether the 222 Draining Development? transactions are intra�rm or arm’s length. The most well known exam- ple of the �rst situation (government agencies keeping data on speci�c products) is the data set on crude petroleum imported into Canada and the United States from the 1970s through the 1980s. The most well known example of the second situation (concern about related-party transactions) is the data set on U.S. export and import transactions col- lected by BLS, which are used to compute the U.S. export and import price index. We look at each in turn. Five studies have used con�dential data from the Canadian and U.S. governments on crude petroleum imports to test for TPM. Bernard and Weiner (1990, 1992, 1996) �nd weak evidence of TPM in Canadian and U.S. import prices, which may have been partly related to CIT differen- tials. Rugman (1985), looking only at the Canadian data, concludes that there was no TPM in Canadian oil import prices in the 1970s. Bernard and Genest-Laplante (1996) examine oil shipments into Canada and the United States over 1974–84 and �nd that the six largest Canadian af�li- ates pay the same or lower prices for crude oil imports relative to the prices of third-party transactions; they argue this is evidence of TPM because Canada was a low-tax country relative to the United States over this period. The second source of high-quality data on intra�rm trade is BLS data on U.S. import and export merchandise transactions. BLS �eld econo- mists work with approximately 8,000 �rms across the United States to identify the most representative export and import transactions, appro- priate volumes, and key characteristics, including whether transactions involve related or unrelated parties. Once the transactions have been identi�ed, the �rms voluntarily provide transaction-level data on a monthly or quarterly basis to BLS, which cleans the data and uses them to construct U.S. export and import price indexes. As a result, the BLS data are of much better quality for trade mispricing studies in general and for TPM trade mispricing studies in particular. However, there is a sample selection bias problem given that participation in the BLS pro- gram is voluntary, and �rms can and do refuse to participate. Clausing has tested the links between CIT differentials and TPM using con�dential monthly BLS export and import price data for January 1997– December 1999. She �nds a strong relationship indicating tax avoidance: a “tax rate 1% lower in the country of destination/origin is associated with Transfer Price Manipulation 223 intra�rm export prices that are 1.8% lower and intra�rm import prices that are 2.0% higher, relative to non-intra�rm goods� (Clausing 2003, 16). There are some problems with Clausing’s analysis. Her test period most likely underestimates TPM because BLS did not include non–market- based transfer prices until April 1998, and the CIT rate she uses is not the theoretically preferred rate for TPM (the statutory rate adjusted for tax preferences: see Eden 1998; Grubert and Slemrod 1998). The little evidence we have involving a comparison of the BLS and census data sets can be found in Eden and Rodriguez (2004), who use BLS and U.S. Census Bureau data for January–June 1999 to estimate the responsiveness of U.S. import prices to CIT differentials and tariffs. The authors �nd that the gap between the BLS- and census-based price indexes widen by 1.3 percent for every 10 percent increase in the share of intra�rm trade and that these results are even stronger where govern- ment trade barriers encouraged TPM. Their study suggests that using census data to estimate TPM is problematic; the BLS data set provides superior estimates. Evidence from developing countries TPM has long been seen as a way to move funds out of developing coun- tries. The Committee of Eminent Persons at UNCTAD, for example, issued a 1978 report surveying studies of MNEs that use TPM to extract resources from developing countries (UNCTAD 1978). However, the empirical evidence is much more sparse for developing countries than for developed countries (especially in comparison with U.S. studies), which probably is to be expected because careful studies require data sets that often do not exist in developing countries. The early studies. Most of the research has been in the form of country case studies; for example, Ellis (1981) looked at TPM in Central America; ESCAP and UNCTC (1984) in Thailand; Lecraw (1985) in the countries of the Association of Southeast Asian Nations; Natke (1985) in Brazil; Vaitsos (1974) in Columbia; and Lall (1973) at TPM involving developing countries in general. Some of these studies were highlighted in UNCTAD (1978), Murray (1981), Rugman and Eden (1985), and Plasschaert (1994). UNCTAD (1999) and Eden (1998) provide more recent overviews of this literature. 224 Draining Development? Some researchers have compared intra�rm prices for selected imports directly with world or domestic prices for the same products. Vaitsos (1974), for example, concluded that foreign MNEs overinvoiced intra�rm imports into Colombia to avoid Colombia’s foreign exchange controls. Natke (1985) found that MNEs were overinvoicing imports into Brazil to avoid Brazil’s extensive regulations, which included price and credit controls, pro�t repatriation restrictions, and high CIT rates. Lecraw (1985) concluded that tariffs, relative tax rates, price and foreign exchange controls, and country risk were signi�cant variables explaining transfer pricing behavior in the Association of Southeast Asian Nations. There are also studies of speci�c industries, such as petroleum and phar- maceuticals, where pro�t ratios are quite high so that TPM can be an important determinant of income flows among countries. For a recent study on the offshoring of business services, see Eden (2005), which focuses on call centers. The China studies. China is perhaps the country where TPM has been studied the most frequently, certainly in recent years, primarily because the Chinese government has provided researchers access to �rm-level balance sheets, income tax �lings, and customs valuation data. Impor- tant in the studies is the fact that China has offered tax holidays and other incentives to �rms that locate in special economic zones and research parks. A few of the more well known studies are reviewed below. Chan and Chow (1997a) provide the �rst evidence on TPM in China. In 1992–93, they collected 81 tax audit cases from city tax bureaus and analyzed the cases for differences across industries, according to �rm size, by nationality, and so on. They argue that low tax rates and tax holi- days in China should encourage income shifting into China, but the for- eign exchange controls and risks would work in the opposite direction. The tax cases, however, involved income shifting out of China (perhaps because that attracted the attention of tax authorities). The six minicases discussed in the article each involved underinvoicing of exports to related parties, creating losses on exports. The authors estimated the extent of underinvoicing by comparing the pro�ts on export sales rela- tive to the pro�ts reported on domestic sales or to industry averages; these are crude measures, but the results are consistent with other case studies of TPM in developing countries. Transfer Price Manipulation 225 Chan and Chow (1997b) review the various external motivations that �rms in China face and that might lead them to engage in TPM. The authors again argue that, while tax rates are low, which should encourage income shifting into China, foreign exchange controls are an important counterweight encouraging income shifting out of China. Assuming that exchange controls are more important than tax motivations, the authors argue that TPM by foreign �rms in China should be visible in their overinvoicing of imports and underinvoicing of exports relative to domestic �rms for the same or similar products, that is, foreign �rms should be shifting funds out of China through transfer pricing to avoid the foreign exchange controls. To test this hypothesis, the authors sub- tract the trade transactions of foreign �rms from total trade transactions in 1992 to estimate trade by Chinese �rms. Comparing foreign with domestic trade patterns, the authors estimate a rate of 11 percent in overinvoicing of imports and 12 percent in overinvoicing of exports among foreign �rms relative to Chinese �rms. Because the authors �nd evidence of overinvoicing on both exports and imports, other factors besides foreign exchange controls may be important in this case. An additional factor affecting transfer pricing in China was the �ve- year tax holiday for foreign manufacturing af�liates provided by the Chinese government in the 1990s (Chow and Chan 1997a; Chan and Mo 2000). If a foreign af�liate made losses, the �rm was exempt from the Chinese CIT. The normal CIT rate was 30 percent, but foreign af�liates in Special Economic Zones were taxed at 15 percent. The tax holiday was structured so that it took effect only after the af�liate declared a pro�t. For the �rst two pro�table years, the exemption continued, and then the CIT rate was reduced by 50 percent for three more years. After the �ve- year period, the regular CIT rate would apply. Therefore, while the tax holiday was attractive to foreign MNEs, it was temporary; meanwhile, as long as the foreign af�liate made losses, the �rm would be exempt from tax. Foreign MNEs therefore had a strong incentive to underreport prof- its in China because of the peculiar structure of the tax holiday. Chan and Mo tested this argument by examining 585 tax audits in 1997; the authors concluded that foreign MNEs “in the pre-holiday posi- tion often manipulate taxable income by exaggerating losses before their �rst pro�tmaking year, so as to extend their tax holiday� (Chan and Mo 2000, 480). What the authors discovered is that the �rms were using TPM 226 Draining Development? (mostly by inflating the cost of sales) to prolong the period of time dur- ing which they declared losses. The authors recommended that the Chi- nese government should limit the preholiday window on tax losses; the government did change the policy and implemented a �xed time period for allowable losses, closing off this particular arbitrage opportunity. Two other studies of transfer pricing in China report on interviews with �nancial controllers of 64 foreign af�liates in 2000 (Chan and Chow 2001; Chan and Lo 2004). Both studies �nd that the perceptions of managers on three environmental variables affect their choice of transfer pricing methods. The more important the interests of local partners and the need to maintain a good working relationship with the Chinese government, the more likely the foreign MNE is to select a market-based transfer pricing method. The more important are foreign exchange controls, the more likely that cost-based prices are used. The series of Chan and Chow studies are summarized in Chow (2010). She argues that the policy environment in China in the 1990s and 2000s has had a major impact on the transfer pricing policies of foreign MNEs in China. In the 1990s, the combination of tax incentives for foreign MNEs, plus weak regulatory enforcement by inexperienced Chinese tax auditors created a low-tax environment that encouraged TPM. In 2008, China adopted a flat CIT rate of 25 percent that applied to both Chinese and foreign MNEs. The CIT rate remains below the rates of most of China’s trading partners except for Hong Kong SAR, China and for Singapore; Chow therefore argues that tax motivations for shifting pro�ts into China still exist. CIT differentials, however, can be offset by other policy environment factors, such as tariffs, foreign exchange controls, political risk, and forced equity joint venture partner- ships. Chow discusses each of these factors, concluding that policy liber- alization had signi�cantly reduced the impact of the factors by the time of her study, leaving tax incentives as the major factor affecting transfer pricing policies by foreign MNEs in China as of 2010. Taking a different perspective on TPM, Lo, Wong, and Firth (2010) investigate the impact of corporate governance on earnings manage- ment through TPM, using a sample of 266 Chinese �rms listed on the Shanghai Stock Exchange in 2004. They measure the manipulation of earnings through TPM by comparing the gross pro�t margin on related- party transactions with the corresponding margin on arm’s-length Transfer Price Manipulation 227 transactions. The authors �nd that Chinese MNEs with higher-quality governance structures (for example, a higher percent of independent directors and �nancial experts on their audit committees) are less likely to engage in TPM. Their results suggest that improving corporate gover- nance of MNEs in emerging markets may lead to less earnings manipu- lation through transfer pricing. Looking across these studies of income shifting and trade mispricing, the evidence supports the presumption of TPM in China. Kar and Cartwright-Smith (2008) suggest that China may be the single most important developing country to experience TPM. Perhaps this explains why the Chinese government is now paying so much attention to the development of sophisticated transfer pricing regulations and why Bei- jing and Shanghai are two of the most rapidly growing transfer pricing locations for the Big Four accounting �rms. Conclusion In this chapter, we attempt to contribute to the debate on illicit flows and developing countries through the lens of TPM. Based on our literature review, we conclude that empirical evidence for TPM exists, but is not overwhelming. The small number of empirical studies is perhaps not surprising given the �ne-grained—individual transactions identi�ed as related-party or arm’s-length transactions—and highly con�dential nature of the data needed to test for TPM. The best empirical estimates come from transaction-level data collected and processed by statistical agencies that have paid close attention to intra�rm trade, such as the U.S. data set on the prices of crude petroleum imports and the BLS data set used to compute U.S. export and import price indexes. In terms of devel- oping countries, the best and largest number of currently available stud- ies are the Chinese studies. No data set is perfect, and, as a result, all the various estimates are flawed. Still, the balance of the evidence suggests that income shifting does occur through the manipulation of transfer prices. What would be helpful is greater accessibility to transaction-level data on crossborder export and import transactions and MNE income state- ments and balance sheets. This would enable scholars to shine more light in the dark corners of TPM, providing scholars and policy makers with 228 Draining Development? the ability to construct more accurate estimates of TPM and its impacts on both developed and developing countries. The Chinese government may be the most well placed to move in this regard, and the rapidly growing number of empirical studies using data on Chinese trade trans- actions and tax audits suggests we may soon have more reliable estimates of TPM for at least one developing country. Notes 1. Staff of the Joint Committee on Taxation, 2010, “Present Law and Background Related to Possible Income Shifting and Transfer Pricing,� Report JCX-37–10, submitted to the House Committee on Ways and Means, U.S. Congress, Wash- ington, DC, July 20. 2. See the decisions by Judge Learned Hand in Helvering v. Gregory, 69 F.2d 809 (CA-2, 1934) and Commissioner v. Newman, 159 F.2d 848 (CA-2, 1947). References Azémar, C., and G. Corcos. 2009. “Multinational Firms’ Heterogeneity in Tax Respon- siveness: The Role of Transfer Pricing.� World Economy 32 (9): 1291–318. Baker, R. W. 2005. Capitalism’s Achilles Heel: Dirty Money and How to Renew the Free- Market System. Hoboken, NJ: John Wiley & Sons. Bartelsman, E. J., and R. M. W. J. Beetsma. 2003. “Why Pay More? Corporate Tax Avoidance through Transfer Pricing in OECD Countries.� Journal of Public Eco- nomics 87 (9–10): 2225–52. Bernard, A. B., J. B. Jensen, and P. K. Schott. 2008. “Transfer Pricing by U.S.-Based Multinational Firms.� Discussion Paper 08–29 (September), Center for Eco- nomic Studies, U.S. Bureau of the Census. Bernard, J.-T., and E. G. Genest-Laplante. 1996. “Transfer Pricing by the Canadian Oil Industry: A Company Analysis.� Applied Economic Letters 3 (5): 333–40. Bernard, J.-T., and R. J. Weiner. 1990. “Multinational Corporations, Transfer Prices, and Taxes: Evidence from the U.S. Petroleum Industry.� In Taxation in the Global Economy, ed. A. Razin and J. Slemrod, 123–60. Chicago: University of Chicago Press. ———. 1992. “Transfer Prices and the Excess Cost of Canadian Oil Imports: New Evidence on Bertrand Versus Rugman.� Canadian Journal of Economics 25 (1): 101–19. ———. 1996. “Export Pricing in State-Owned and Private MNEs: Evidence from the International Petroleum Market.� International Journal of Industrial Organiza- tion 14 (5): 647–68. Borkowski, S. C. 1992. “Organizational and International Factors Affecting Multina- tional Transfer Pricing.� Advances in International Accounting 5 (1): 173–92. Transfer Price Manipulation 229 ———. 1997. “The Transfer Pricing Concerns of Developed and Developing Coun- tries.� International Journal of Accounting 32 (3): 321–36. Brean, D. 1985. “The Financial Dimensions of Transfer Pricing.� In Multinationals and Transfer Pricing, ed. A. M. Rugman and L. Eden, 149–64. Croom Helm Series in International Business. London: Croom Helm. Chan, K. H., and L. Chow. 1997a. “An Empirical Study of Tax Audits in China on International Transfer Pricing.� Journal of Accounting and Economics 23 (1): 83–112. ———. 1997b. “International Transfer-Pricing for Business Operations in China: Inducements, Regulation and Practice.� Journal of Business Finance and Account- ing 24 (9): 1269–89. ———. 2001. “Corporate Environments and International Transfer-Pricing: An Empirical Study of China in a Developing Economy Framework.� Accounting and Business Research 31 (2): 103–18. Chan, K. H., and A. W. Y. Lo. 2004. “The Influence of Management Perception of Environmental Variables on the Choice of International Transfer-Pricing Meth- ods.� International Journal of Accounting 39 (1): 93–110. Chan, K. H., and P. L. L. Mo. 2000. “Tax Holidays and Tax Noncompliance: An Empirical Study of Corporate Tax Audits in China’s Developing Economy.� Accounting Review 75 (4): 469–84. Chow, L. 2010. “Transfer Pricing Environment in Mainland China.� Chinese Manage- ment Review 13 (1): 1–17. http://cmr.ba.ouhk.edu.hk/cmr/webjournal/v13n1/ CMR258E09.pdf. Christian Aid. 2009. “False Pro�ts: Robbing the Poor to Keep the Rich Tax-Free.� Christian Aid Report, March, Christian Aid, London. Clausing, K. A. 2003. “Tax-Motivated Transfer Pricing and US Intra�rm Trade Prices.� Journal of Public Economics 87 (9–10): 2207–23. ———. 2009. “Multinational Firm Tax Avoidance and Tax Policy.� National Tax Journal 62 (4): 703–25. Cravens, K. S. 1997. “Examining the Role of Transfer Pricing as a Strategy for Multi- national Firms.� International Business Review 6 (2): 127–45. de Boyrie, M. E., S. J. Pak, and J. S. Zdanowicz. 2005. “Estimating the Magnitude of Capital Flight Due to Abnormal Pricing in International Trade: The Russia- USA Case.� Accounting Forum 29 (3): 249–70. Drucker, J. 2010a. “Companies Dodge US$60 Billion in Taxes, Even Tea Party Con- demns.� Bloomberg Businessweek, May 14. ———. 2010b. “‘Dutch Sandwich’ Saves Google Billions in Taxes.� Bloomberg Busi- nessweek, October 22. Eden, L. 1985. “The Microeconomics of Transfer Pricing.� In Multinationals and Transfer Pricing, ed. A. M. Rugman and L. Eden, 13–46. Croom Helm Series in International Business. London: Croom Helm. ———. 1998. Taxing Multinationals: Transfer Pricing and Corporate Income Taxation in North America. Toronto: University of Toronto Press. 230 Draining Development? ———. 2005. “Went for Cost, Priced at Cost? An Economic Approach to the Transfer Pricing of Offshored Business Services.� Transnational Corporations, 14 (2): 1–53. ———. 2009. “Taxes, Transfer Pricing, and the Multinational Enterprise.� In The Oxford Handbook of International Business, 2nd. ed., ed. A. M. Rugman, 591– 622. Oxford: Oxford University Press. Eden, L., L. F. Juárez, and D. Li. 2005. “Talk Softly But Carry a Big Stick: Transfer Pric- ing Penalties and the Market Valuation of Japanese Multinationals in the United States.� Journal of International Business Studies 36 (4): 398–414. Eden, L., and R. Kudrle. 2005. “Tax Havens: Renegade States in the International Tax Regime.� Law and Policy 27 (1): 100–27. Eden, L., and P. Rodriguez. 2004. “How Weak Are the Signals? International Price Indices and Multinational Enterprises.� Journal of International Business Studies. 35 (1): 61–74. Ellis, F. 1981. “Export Valuation and Intra-Firm Transfers in the Banana Export Industry in Central America.� In Multinationals beyond the Market, ed. R. Mur- ray, 61–76. New York: John Wiley and Sons. Ernst & Young. 2008. “2007–2008 Global Transfer Pricing Survey: Global Transfer Pricing Trends, Practices, and Analyses.� Ernst & Young, London. ———. 2010. “2010 Global Transfer Pricing Survey: Addressing the Challenges of Globalization.� Ernst & Young, London. ESCAP (United Nations Economic and Social Commission for Asia and the Paci�c) and UNCTC (United Nations Center for Transnational Corporations). 1984. “Transnational Corporations and Transfer Pricing: A Case Study of the Phar- maceutical Industry of Thailand.� Working paper, ESCAP/UNCTC Joint Unit on Transnational Corporations, New York. Friedman, M. 1970. “The Social Responsibility of Business Is to Increase Its Pro�ts.� New York Times Magazine 33 (September 13): 122–26. Goetzl, A. 2005. “Why Don’t Trade Numbers Add Up?� ITTO Tropical Forest Update 15/1, International Tropical Timber Organization, Yokohama. Grubert, H. 2003. “Intangible Income, Intercompany Transactions, Income Shifting and the Choice of Location.� National Tax Journal 56 (1, part 2): 221–42. Grubert, H., and J. Mutti. 1991. “Taxes, Tariffs and Transfer Pricing in Multinational Corporation Decision Making.� Review of Economics and Statistics 73 (2): 285–93. Grubert, H., and J. Slemrod. 1998. “The Effect of Taxes on Investment and Income Shifting to Puerto Rico.� Review of Economics and Statistics 80 (3): 365–73. Gulati, S. K. 1987. “A Note on Trade Misinvoicing.� In Capital Flight and Third World Debt, ed. D. R. Lessard and J. Williamson, 68–78. Washington, DC: Peterson Institute for International Economics. Harris, D., R. Morck, and J. Slemrod. 1993. “Income Shifting in U.S. Multinational Corporations.� In Studies in International Taxation, ed. A. Giovannini, R. G. Hub- bard, and J. Slemrod, 277–308. Cambridge, MA: National Bureau of Economic Research; Chicago: University of Chicago Press. Transfer Price Manipulation 231 Henisz, W. J., and B. A. Zelner. 2010. “The Hidden Risks in Emerging Markets.� Har- vard Business Review 88 (4): 89–95. Hirshleifer, J. 1956. “On the Economics of Transfer Pricing.� Journal of Business 29 (3): 172–83. ———. 1957. “Economics of the Divisionalized Firm.� Journal of Business 30 (2): 96–108. Horst, T. S. 1971. “The Theory of the Multinational Firm: Optimal Behavior under Different Tariff and Tax Rates.� Journal of Political Economy 79 (5): 1059–72. Kar, D., and D. Cartwright-Smith. 2008. “Illicit Financial Flows from Developing Countries, 2002–2006.� Global Financial Integrity, Washington, DC. http://www .g�p.org/storage/g�p/economist%20-%20�nal%20version%201-2-09.pdf. Kudrle, R. T., and L. Eden. 2003. “The Campaign against Tax Havens: Will It Last? Will It Work?� Stanford Journal of Law, Business and Finance 9 (1): 37–68. Lall, S. 1973. “Transfer-Pricing by Multinational Manufacturing Firms.� Oxford Bul- letin of Economics and Statistics 35 (3): 173–95. Lecraw, D. 1985. “Some Evidence of Transfer Pricing by Multinational Corpora- tions.� In Multinationals and Transfer Pricing, ed. A. M. Rugman and L. Eden, 223–40. Croom Helm Series in International Business. London: Croom Helm. Li, S. H., and K. R. Balachandran. 1996. “Effects of Differential Tax Rates on Transfer Pricing.� Journal of Accounting, Auditing and Finance 11 (2): 183–96. Lo, A. W. Y., R. M. K. Wong, and M. Firth. 2010. “Can Corporate Governance Deter Management from Manipulating Earnings? Evidence from Related-Party Sales Transactions in China.� Journal of Corporate Finance 16 (2): 225–35. McDonald, M. 2008. “Income Shifting from Transfer Pricing: Further Evidence from Tax Return Data.� OTA Technical Working Paper 2 (July), Of�ce of Tax Analy- sis, U.S. Department of the Treasury, Washington, DC. Murray, R., ed. 1981. Multinationals beyond the Market: Intra-�rm Trade and the Control of Transfer Pricing. London: John Wiley and Sons. Natke, P. 1985. “A Comparison of Import Pricing by Foreign and Domestic Firms in Brazil.� In Multinationals and Transfer Pricing, ed. A. M. Rugman and L. Eden, 212–22. Croom Helm Series in International Business. London: Croom Helm. Nitsch, V. 2009. “Trade Mispricing and Illicit Flows.� Paper presented at the World Bank’s conference, “The Dynamics of Illicit Flows from Developing Countries,� Washington, DC, September 14–15. OECD (Organisation for Economic Co-operation and Development). 1979. “Trans- fer Pricing and Multinational Enterprises: Report of the OECD Committee on Fiscal Affairs.� OECD, Paris. ———. 1995. “Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations.� OECD, Paris. ———. 1998. “Harmful Tax Competition: An Emerging Global Issue.� Report, OECD, Paris. 232 Draining Development? ———. 2010. “Intra-�rm Trade: A Work in Progress.� Document STD/TBS/WPTGS (2010)24 (September 8), Working Party on International Trade in Goods and Trade in Services Statistics, Statistics Directorate, OECD, Paris. Pak, S. J., and J. S. Zdanowicz. 1994. “A Statistical Analysis of the US Merchandise Trade Data Base and Its Uses in Transfer Pricing Compliance and Enforcement.� Tax Management: Transfer Pricing Report 3 (1): 50–57. ———. 2002. “U.S. Trade with the World: An Estimate of 2001 Lost U.S. Federal Income Tax Revenues Due to Over-Invoiced Imports and Under-Invoiced Exports.� Unpublished working paper, School of Graduate Professional Studies, Penn State University–Great Valley, Malvern, PA. Plasschaert, S. R. F. 1994. Transnational Corporations: Transfer Pricing and Taxation. Vol. 14 of United Nations Library on Transnational Corporations. London: Routledge. Rugman, A. M. 1985. “Transfer Pricing in the Canadian Petroleum Industry.� In Mul- tinationals and Transfer Pricing, ed. A. M. Rugman and L. Eden, 173–92. Croom Helm Series in International Business. London: Croom Helm. Rugman, A. M., and L. Eden, eds. 1985. Multinationals and Transfer Pricing. Croom Helm Series in International Business. London: Croom Helm. Sikka, P. N., and H. Willmott. 2010. “The Dark Side of Transfer Pricing: Its Role in Tax Avoidance and Wealth Retentiveness.� Critical Perspectives on Accounting 21 (4): 342–56. Swenson, D. 2001. “Tax Reforms and Evidence of Transfer Pricing.� National Tax Journal 54 (1): 7–26. Tang, R. Y. W. 1993. Transfer Pricing in the 1990s: Tax and Management Perspectives. Westport, CT: Quorum Books. ———. 1997. Intra�rm Trade and Global Transfer Pricing Regulations. Westport, CT: Quorum Books. ———. 2002. Current Trends and Corporate Cases in Transfer Pricing. Westport, CT: Quorum Books. TJN (Tax Justice Network). 2007. Closing the Floodgates: Collecting Tax to Pay for Development. London: TJN. http://www.innovative�nance-oslo.no/pop.cfm? FuseAction=Doc&pAction=View&pDocumentId=11607. UNCTAD (United Nations Conference on Trade and Development). 1978. “Domi- nant Positions of Market Power of Transnational Corporations: Use of the Transfer Pricing Mechanism.� Document TD/B/C.2/167, United Nations, New York. ———. 1992. World Investment Report 1992: Transnational Corporations as Engines of Growth. Document ST/CTC/130. Geneva: UNCTAD. ———. 1999. “Transfer Pricing.� Document UNCTAD/ITE/IIT/11 (vol. I), UNCTAD Series on Issues in International Investment Agreements, UNCTAD, Geneva. ———. 2010. World Investment Report 2010: Investing in a Low-Carbon Economy. Geneva: UNCTAD. ———. 2011. World Investment Report 2011: Non-Equity Modes of International Pro- duction and Development. Geneva: UNCTAD. Transfer Price Manipulation 233 U.S. Census Bureau. 2010. “U.S. Goods Trade: Imports & Exports by Related-Parties 2009.� U.S. Census Bureau News CB10–57 (May 12), U.S. Department of Com- merce, Washington, DC. http://www.census.gov/foreign-trade/Press-Release/ 2009pr/aip/related_party/rp09.pdf. Vaitsos, C. V. 1974. Intercountry Income Distribution and Transnational Enterprises. Oxford: Clarendon Press. Vincent, J. R. 2004. “Detecting Illegal Trade Practices by Analyzing Discrepancies in Forest Products Trade Statistics: An Application to Europe, with a Focus on Romania.� Policy Research Working Paper 3261, World Bank, Washington, DC. Wood, E., and T. Moll. 1994. “Capital Flight from South Africa: Is Under-Invoicing Exaggerated?� South African Journal of Economics 62 (1): 28–45. 8 The Role of Transfer Pricing in Illicit Financial Flows Carlos A. Leite Abstract As multinationals have become a larger �xture in international trade and the weight of intracompany transactions has increased, the practice of transfer pricing has come under greater and often strident scrutiny. In this chapter, we review the current practice of transfer pricing and its impact on the ability of developing countries to collect tax revenue. After an overview of the arm’s-length principle (ALP), the key con- cept in modern-day transfer pricing regulations, we discuss the com- plexities involved in determining prices for transactions between related parties in a context of competing interests and unique transactions. By means of examples drawn from speci�c real-world cases, we (1) illus- trate the dif�culties in simultaneously assuring that companies are not double taxed and that competing tax authorities collect their fair share of tax revenue and (2) distinguish between tax avoidance and tax eva- sion and examine the relationship of each to the illegality of �nancial flows. Unfortunately, the existing literature on the impact of transfer pricing abuses on developing countries provides insuf�cient evidence on the extent of income shifting and the role of transfer price manipulation and 235 236 Draining Development? few insights into the possible illicit nature of the underlying transac- tions. We conclude by noting that (1) expanding the use of the ALP is the best option for a theoretically sound and globally consistent approach to transfer pricing, and (2) it is crucial for developing countries to adopt a speci�c transfer pricing regime and to engage in a higher degree of information sharing among tax authorities. Introduction Multinationals are the most successful form of business organization largely because of inherent advantages in the access to crucial knowledge and processes and in the ability to secure economies of scale. Over the past three decades, the number of multinationals has risen 10-fold, from a modest 7,258 in 1970 to nearly 79,000 by 2006 (UNCTAD 2008). Against the background of falling barriers to international commerce, these corporations have played a key role in boosting trade and foreign direct investment (FDI) flows, and, as their far-flung operations have expanded, so has the importance of intra�rm trade.1 It is estimated that, globally, multinationals account for approximately two-thirds of trade flows, and fully half of these are intra�rm transactions.2 These devel- opments and trends have distinct consequences for international taxa- tion, and some observers have linked weaknesses in tax collection in developing countries with the propensity of multinationals to use �nan- cial arti�ces, including favorable transfer pricing practices on intra�rm trade, to shift income (illegally) around the world. As a result, transfer pricing has been raised from the status of an arcane subject solely in the purview of tax administrators and practitio- ners to a ranking as the most important issue in international taxation in surveys of business people. It has also become a key point of discussion among politicians, economists, and nongovernmental organizations. Technically, transfer pricing refers to the pricing or valuation of intra�rm transactions (a complex task in practice) and the consequent allocation of pro�ts for tax and other internal purposes (a complicated balancing act of competing national interests and public policy choices). On one side of these transactions, multinationals aim, �rst, at avoiding double taxation (a strong deterrent to international trade) and, second, at mini- mizing the tax burden (potentially a source of competitive advantage). The Role of Transfer Pricing in Illicit Financial Flows 237 Meanwhile, the tax authority of the entity on the selling side of the trans- action attempts to raise the price to maximize its own tax collection (on which its performance is judged), and, �nally, the tax authority of the entity on the purchasing side of the transaction attempts to lower the price so that its own tax collections can be higher (on which its effective- ness is evaluated). The complexity of the resulting to-and-fro necessi- tates a speci�c set of concepts and rules (including transfer pricing regu- lations and tax treaties) and even a speci�c interjurisdictional dispute resolution mechanism (the Mutual Agreement Procedure and Compe- tent Authority process). Following a summary introduction to the building blocks of transfer pricing and a brief discussion of some of the complexities of implement- ing the ALP, the stage is set for a discussion on the signi�cant differences between tax avoidance and tax evasion and how these concepts intersect with the practice of transfer pricing. Next, we consider the existing lit- erature on the impact of transfer pricing abuses on the tax revenues col- lected by developing countries; unfortunately, this literature tends to provide, at best, indirect evidence on the extent of income shifting and the role of transfer pricing, but few insights into the motivation or the possible illicit nature of the underlying transactions. We then discuss the importance of developing countries adopting a speci�c transfer pricing regime; we note, in particular, that a globally consistent tax regime and higher degree of information sharing among tax authorities on the application of transfer pricing policies will help to ensure consistency of application, make the rules clearer for taxpayers, and increase the ability of tax authorities to act in a fair and coordinated manner. We conclude by noting that the ALP and ongoing efforts of the Organisation for Eco- nomic Co-operation and Development (OECD) to expand the use of the ALP are the best option for a theoretically sound and globally consis- tent approach to transfer pricing. Throughout the chapter, we illustrate some of the implementation and administration complexities faced by multinationals and tax author- ities by discussing three cases taken from actual exchanges between tax- payers and tax authorities. In addition to tax-minimizing choices by management, these cases also illustrate the lack of coordination among tax authorities that sometimes risks compromising solid business strate- gies and valuable commercial relationships. 238 Draining Development? What Is the Arm’s-Length Principle? A simple theoretical construct In a world of multijurisdictional taxation and multinational enterprises (MNEs) operating across borders, intercompany transactions have important tax implications, primarily because the associated transfer prices determine the distribution of pro�ts and tax liabilities by jurisdic- tion.3 Crucially, however, such transactions are not open-market trans- actions, which leaves unresolved the issue of the appropriate pricing mechanism (see box 8.1).4 For these cases, the pricing mechanism used most often is the ALP initially adopted by the OECD in 1979. The appeal of the ALP resides in its theoretical simplicity: simulate the market mechanism and set the transfer prices as if the transactions had been carried out between unre- lated parties, each acting in its own best interest. This simple construct provides a legal framework that easily accommodates the two overriding objectives of an internationally consistent system: governments should collect their fair share of taxes, and enterprises should avoid double taxa- Box 8.1. A Trading Intermediary: A Case of Management Foresight? Background: A large natural resource company has vertically integrated operations in an industry comprised of a relatively few similarly structured �rms and with sales agreements primarily in the form of long-term supply contracts. Business arrangement: An intermediary subsidiary in a low-tax jurisdiction buys 1 00 percent of the company’s production under a long-term contract and resells the product under a mix of long-term arrangements and spot contracts. The intermediary receives �nancing from the com- pany, legally holds inventory, and assumes the business risks associated with the negotiating and reselling function. Market developments: Because of unforeseen circumstances, the market changes (in what amounts analytically to an outward shift of the supply curve), and prices drop. It is during this time that the intermediary is set up. As the reasons for the original supply shift abate and as stronger economic growth boosts demand, the market changes again, and prices rise to historical highs. Financial results: With long-term purchase contracts in place, the intermediary’s �nancial results bene�t signi�cantly from the positive developments in the spot market, and, as some of the long- term sales contracts come up for renewal, its �nancial results improve additionally. Bottom line: Should this outcome be attributed to management foresight (or perhaps fortune)? Should it be considered, ex post, as an example of abusive transfer pricing? The Role of Transfer Pricing in Illicit Financial Flows 239 tion on their pro�ts. Procedurally, effective application of the ALP in a way that minimizes crossborder disputes requires a system of bilateral tax treaties, together with mutual agreement and competent authority procedures that encourage principled negotiations and timely resolu- tions between the tax authorities on different sides of each transaction. The ALP was formally de�ned in article 9 of the OECD Model Tax Convention on Income and on Capital and, later, in the OECD’s transfer pricing guidelines, as follows: [When] conditions are made or imposed between . . . two [associated] enterprises in their commercial or �nancial relations which differ from those which would be made between independent enterprises, then any pro�ts which would, but for those conditions, have accrued to one of the enterprises, but, by reason of those conditions, have not so accrued, may be included in the pro�ts of that enterprise and taxed accordingly. (OECD 1995, I-3) From this de�nition, the tax authority retains the right to adjust the transfer prices in a way that restores market prices. In adopting the ALP, a number of countries have gone further in laying out a framework of transfer pricing rules, including the following: • Speci�c provisions for recharacterization and adjustments of transac- tions (that is, repricing carried out by the tax authority) • Documentation requirements (typically, on a contemporaneous and transactional basis) • Penalty provisions for noncompliance (typically focused on penaliz- ing egregious failures of compliance rather than good faith errors) • Advance pricing arrangements (which allow an ex ante agreement between taxpayers and tax authorities on the treatment of speci�c transactions) Less than an exact science at the implementation level Notwithstanding the broadbased agreement on the underlying theoreti- cal principle and an extensive framework for settling disputes, the appli- cation of the ALP tends to be fact intensive and judgment-based, which poses challenges and opportunities for tax authorities and taxpayers alike (box 8.2). In other words, while the more popular discourse centers on potential transfer pricing abuses by MNEs, there is also no guarantee 240 Draining Development? of a fair application of the ALP by tax authorities. During competent authority negotiations, for example, of�cers and representatives of sepa- rate tax authorities may be more concerned about protecting their respective tax bases than about a straightforward application of the ALP.5 Additionally, the formal dispute resolution mechanisms do not mandate an outcome that necessarily avoids double taxation or even a timely con- clusion, despite the emphasis on principled negotiations. Consider, for example, the recent case of a company with well-known brands that were initially developed in the United States and that are legally owned by the U.S. parent, but that are marketed globally. How much of a royalty for the use of the brand should the U.S. parent charge its Canadian subsidiary, or, equivalently, how much of the operating pro�t of the Canadian subsidiary should be attributed to the value of these brands? The U.S. parent continues to expend substantial amounts Box 8.2. A Trading Intermediary: Redux Accounting treatment: The trading intermediary records all the revenue from its contracts, the cost of purchasing the goods from the parent company, and the cost of subcontracting the parent company for administrative and other business functions (at the equivalent cost, plus a market-related markup). The remaining pro�t is booked by the intermediary. Possible taxpayer’s position: The trading intermediary should be compensated for the business and market risks it has assumed. It trades on its own account, signs purchasing and sales contracts, takes legal possession of inventory, and has suf�cient working capital to �nance its own operations. Management had the foresight to set up the business structure at an opportune time (and, if mar- ket developments had gone in the other direction, the intermediary would have incurred losses). Possible tax authority’s position: The trading company is recording larger than market-related pro�ts because it is overpricing the risks assumed, or, equivalently, it is underpricing the unique contribution of the strategic guidance provided by the parent company, particularly on the nego- tiation of contracts. Current debate: The ability of taxpayers to modify their business structures in conformity with the ALP is currently a topic of much debate. In this case, the question could be framed around the suf�ciency of the economic substance in the trading company effectively to take on all the risks that the parent company ascribes to its subsidiary, where, in the opinion of some tax authorities, the risks necessarily follow the assets and functions. In this case, the taxpayer could argue that the negotiating functions are more routine (based quite simply on the widely available spot price and market trends), and some observers would argue that the tax authority is simply adapting its own conceptual framework to ensure short-term revenue maximization (which is generally not good public policy). The Role of Transfer Pricing in Illicit Financial Flows 241 on research and development, and it both develops and guides market- ing campaigns around the world. Its transfer pricing policy speci�es a royalty of 4 percent of sales. The U.S. tax authority looked at the value of the brand and indicated that, based on analysis of comparable transac- tions, a more appropriate charge would be in the range of 7–11 percent (thus increasing the U.S. reported income). In looking at the same trans- action, the Canadian tax authority analyzed company functions in Can- ada and indicated that it would disallow any crossborder royalty; it indi- cated that the marketing campaign in Canada, together with other activities and investments of the Canadian entity, created local brand awareness and equity. In this dispute, the MNE is essentially neutral: the Canadian subsid- iary is 100 percent owned, and internal transfers are a zero-sum game, plus the tax rates in the two jurisdictions are broadly similar, and the business environment is suf�ciently comparable. Still, it now faces a conundrum on how to set its transfer prices as it is pulled by one tax authority and pushed by the other; failure to comply immediately with both demands could mean large penalties and late-interest payments.6 Ultimately, the company could simply let the dispute go to competent authority, let the tax authorities decide how to benchmark the royalty charge, and focus on avoiding double taxation. However, there is no cer- tainty of a settlement that fully eliminates double taxation, and the pro- cess of appeals and arbitration is a costly and time-consuming exercise.7 In other words, despite the focus in the public discourse on the potential for transfer pricing abuses, there is a flip side that places MNEs at risk of being caught between competing tax authorities and ending up with double taxation (a signi�cant cost if corporate tax rates are in the range of 30–35 percent in Canada and the United States) and large interest payments, as well as heavy penalties (which could amount to 10 percent of any adjustments proposed by a tax authority). This case highlights the fact that the theoretical simplicity of the ALP is matched by judgment and complexity in implementation, but it also illustrates some of the tough choices faced by MNEs in navigating the sometimes choppy waters of a multijurisdictional taxation system. In this case, the underlying transaction involves a unique intangible asset and wide-open questions on the existence, ownership, and localization of the asset. In practice, the absence of exactly comparable market transactions 242 Draining Development? from which to extract a market price necessitates approximations from either a range of similar transactions (adjusted for key comparability dif- ferences) or the use of valuation techniques that fundamentally rely on subjective assessments of economic contributions and company-speci�c �nancial forecasts. In such cases, the transfer pricing analysis tends to generate a range of prices, and any point within that range may be con- sidered an arm’s-length price. Transfer pricing is not an exact science, and the OECD guidelines explicitly state as follows: There will be many occasions when the application of the most appropri- ate method or methods produces a range of �gures all of which are rela- tively equally reliable. In these cases, differences in the �gures that make up the range may stem from the fact that in general the application of the arm’s-length principle only produces an approximation of conditions that would have been established between independent enterprises. (OECD 1995, I-19) Furthermore, the OECD guidelines specify the following: A range of �gures may also result when more than one method is applied to evaluate a controlled transaction. For example, two methods that attain similar degrees of comparability may be used to evaluate the arm’s-length character of a controlled transaction. Each method may produce an out- come or a range of outcomes that differs from the other because of differ- ences in the nature of the methods and the data, relevant to the applica- tion of a particular method, used. Nevertheless, each separate range potentially could be used to de�ne an acceptable range of arm’s-length �gures. (OECD 1995, I-19) To guide the process of selecting appropriate comparable transac- tions, transfer pricing reports include extensive functional analyses of a company’s assets employed, functions undertaken, and risks borne.8 Even if the product transacted is a commodity for which market prices are widely available, differences in circumstances and timing issues likely necessitate comparability adjustments. Consider, for example, a hypo- thetical multinational oil company selling a cargo of oil from Cabinda, Angola. Starting with the published price of Brent crude, you need to adjust (at least) for differences in chemical characteristics between the two types of crude and for transportation costs.9 As the type of product The Role of Transfer Pricing in Illicit Financial Flows 243 or service becomes less of a traded commodity and more of the interme- diate good that is often the subject of intra�rm transactions or even a more unique service or intangible, transfer pricing comes to rely even more on (1) comparability adjustments and ranges of prices obtained from a group of similarly comparable transactions and (2) pro�t-based methods, rather than product-based prices.10 In practice, derived ranges for arm’s-length pricing can be quite broad. In a cost-markup situation, for example, it is not uncommon for a range with a central point (or average) at 10 percent to have a minimum of less than 5 percent and a maximum of around 15 percent. The range arises from differences in results across comparable companies and may also be affected by adjustments for differences in comparability between the tested party (the company for which the range is being calculated) and the comparable companies (which provide the market evidence). In addition, to account for possible differences in the business and economic cycles, a three- or �ve-year average of results is typically used as the market benchmark (OECD 1995). The OECD guidelines recog- nize that, particularly in applying pro�t-based transfer pricing methods, it is necessary to take account of business cycles and special circum- stances, such as the startup of a business, that cannot be fully addressed through comparability adjustments. Absent this option, a pro�t-based method might indicate that a transfer pricing adjustment is called for (simply) because the tested party is at the bottom of its business cycle when the uncontrolled comparables are at the peak of theirs. Finally, differences in approach across tax authorities (for example, Canada with its general reliance on form and the United States with its general reliance on economic substance) create interstate frictions. Moreover, inconsistencies in domestic tax codes also create domestic frictions and additional opportunities for tax planning; for example, Canada actually blends an application of legal form for general anti- avoidance purposes with a reliance on economic substance for transfer pricing regulations.11 These frictions can become either a source of con- cern for multinationals aiming to comply with differing provisions or an opportunity for tax arbitrage in the presence of creative tax planning and a favorable fact pattern. At the least, these frictions make MNEs base their decisions on transfer pricing issues not only on the underlying facts and the legal requirements, but also on the degree of aggressiveness (and 244 Draining Development? capacity of monitoring) of each of the tax authorities. In other words, the decision framework becomes a risk-based compliance exercise driven by a fact-intensive and judgment-based analysis. In summary, application of the ALP is more complex in practice than suggested by the relatively elegant theoretical construct. In recent years, the higher propensity for intra�rm trade to involve services, as well as intangible and �nancial assets, has raised particularly complex issues. Tax Avoidance and Tax Evasion Tax avoidance and tax evasion are fundamentally different The courts �rst established the difference between tax avoidance and tax evasion in IRC v. Duke of Westminster ([1936] 19 TC 490). In this case, pay- ments were made by the taxpayer to domestic employees in the form of deeds of covenant, but which, in substance, were payments of remunera- tion. The House of Lords refused to disregard the legal character (form) of the deeds merely because the same result (substance) could be brought about in another manner. Lord Tomlin commented, as follows: “Every man is entitled, if he can, to order his affairs so that the tax attaching under the appropriate Acts is less than it otherwise would be.� In other words, taxpayers are entitled to arrange their tax affairs in the most tax ef�cient way, and this is the principle behind tax planning and tax avoidance. Tax evasion, in contrast, involves an element of illegality and clearly gives rise to illicit �nancial flows.12 Those participating in tax evasion may underreport taxable income or claim expenses that are nondeduct- ible or overstated.13 Tax evaders generally face criminal prosecution, and criminals are sometimes charged with tax evasion (given their usual reluctance to report income from criminal activities) if there is insuf�- cient evidence to try them on the underlying criminal action. In the con- text of a transfer pricing system based on the ALP, income shifting could be characterized as illicit to the extent that it is not commensurate with a shift in assets, functions, and risks. In recent years, the public debate has focused on activities that may be said to fall in a grey area, or situations open to an interpretation of the law. In this context, some commentators have described selected avoid- ance tactics as unacceptable because they may take unfair advantage of tax rules or run counter to the intentions of the legislators. The behavior The Role of Transfer Pricing in Illicit Financial Flows 245 of corporations in pursuing these more aggressive forms of tax avoid- ance has sometimes been labeled contradictory to the notion of corpo- rate social responsibility. Others argue that the evenhanded administra- tion of taxation is a matter of statutory interpretation of the tax statutes and should not be open to appeals to the intentions of legislators or vague notions of corporate social responsibility. Business restructuring as a form of tax avoidance: An old concept In a world of mobile factors of production, signi�cant differences in tax rates across countries will inevitably lead to tax-based locational advan- tages, and pro�t-maximizing global companies will be at the forefront of securing the bene�ts of structuring their supply chain in a manner con- sistent with business objectives and tax minimization. However, the idea of structuring business operations in a way that minimizes tax liabilities is not, by any stretch of the imagination, a new construct. In ancient Greece, sea traders avoided the 2 percent tax imposed by the city-state of Athens on imported goods by using speci�c locations at which to deposit their foreign goods; in the Middle Ages, Hanseatic traders who set up business in London were exempt from tax; and, in the 1700s, the Ameri- can colonies traded from Latin America to avoid British taxes. Securing tax-based advantages will sometimes mean relocating pro- duction, development, and even administration centers to minimize the tax burden in a legally consistent manner. The resulting shift in assets, functions, and risks will inevitably also shift pro�ts to those countries that offer an environment more consistent with pro�t maximization. Such a shift in pro�ts, as long as it is accompanied by a commensurate shift in assets, functions, and risks, is a legally acceptable strategy; this is the essence of tax planning and tax avoidance. The related �nancial flows could not be considered illicit. However, shifting pro�ts without a business justi�cation (that is, in a manner not commensurate with the assets employed, the risks borne, or the functions provided) is clearly illegal and should be considered illicit. For corporations engaging in such manipulative practices (which effec- tively means that the transfer prices employed are outside of the relevant arm’s-length interval), the penalties can be costly. For example, Glaxo SmithKline announced, in September 2006, that they had settled a long- running transfer pricing dispute with the U.S. tax authorities, agreeing 246 Draining Development? to pay US$3.1 billion in taxes related to an assessed income adjustment because of improper transfer pricing and also agreeing not to seek relief from double taxation (which is available through the bilateral tax treaty between the United Kingdom and the United States). A case of windfall pro�ts Consider the case of a large petroleum company doing business in a high-risk country. During a typically turbulent period (in a recess from open conflict and in the run-up to a contentious election), the govern- ment solicits the assistance of the company’s local subsidiary in building an electricity generating station (not normally an activity that the com- pany would choose) and agrees to pay in barrels of oil (from government stocks) valued at the spot-market price (�xed over the term of the con- tract).14 In the circumstances, the level of conflict escalates; the current government is removed from power; and the world price of oil rises con- siderably. For the company, the contract terms yield a windfall: it is able to sell the in-kind payment for roughly three or four times the �xed value set in the contract. In the normal course of business, the local sub- sidiary sells its local oil production to a foreign-related party (the group’s trading subsidiary), which onsells to oil traders or re�neries. In a country with ALP transfer pricing regulations, the difference between the price set by the contract and the arm’s-length price for local crude would likely be attributed to the local subsidiary as income. In turn, the trading subsidiary would be accorded a normal trading pro�t. In other words, given that the key value-driver in generating this income was the local contract and the related business risks assumed by the local subsidiary, an ALP-consistent allocation of pro�ts would call for the bulk of the total pro�t to be declared in the source country, and the trad- ing operation would earn a return consistent with comparable compa- nies (that provide similar services to third parties). Additional considerations are as follows: • During this time period, the source country had no speci�c transfer pricing legislation, a marginal corporate tax rate of 60 percent, and a generally weak tax administration. • By contrast, the OECD country where the trading subsidiary resides had a highly competent tax administration (which would certainly The Role of Transfer Pricing in Illicit Financial Flows 247 have pressured the MNE into an ALP-consistent treatment if the country roles were reversed), a substantially lower marginal corporate tax rate of 35 percent, and a (legal) scheme that allowed a recharacter- ization of non–ALP-consistent inbound transactions (through a prac- tice known as informal capital rulings). • The local subsidiary was generally worried about the high-risk envi- ronment (including the possibility of expropriation) and a binding limit on its ability to repatriate pro�ts. In the end, the MNE managed to assign the (bulk of the) windfall pro�ts to the trading subsidiary, which used an informal capital ruling to achieve a highly favorable tax treatment in the OECD country (likely at a tax rate of 1 percent for what was considered a capital contribution instead of taxable income). Should this be considered a case of tax avoidance or tax evasion? In the absence of a legal requirement for related-party transactions to be carried out at arm’s length, the local subsidiary honored its contract with the government, assumed the risk of a �xed price compensation scheme, bene�ted from the subsequent rise in oil prices, and onsold the oil to its related party at the same price it negotiated with the government. In the tax system of the source country, there was nothing that would label this transaction as illegal.15 Suppose, for instance, that the local subsidiary (which only has exper- tise in exploration and not in trading) had presold the oil receivable under the contract to its trading subsidiary at the time that it initially entered into the contract in an attempt to hedge its position. The impact on the assignment of pro�ts would have been similar to a later transfer at the contract price. Should this transaction be labeled as a cautious business approach, or should it be characterized, ex post, as income shifting and, possibly, tax evasion? Global tax planning and transfer pricing practices in developing countries The case of the large petroleum company and its allocation of trading pro�ts from the country of exploration to a low-tax OECD jurisdiction raises important questions about the tax system in developing countries and global tax cooperation. This discussion is particularly signi�cant for 248 Draining Development? resource-dependent economies that tend to have a large proportion of national production under the control of MNEs. Currently, the centerpiece of international tax coordination is the bilateral tax treaty, which provides for speci�c tax treatment for taxpay- ers having sources of income in both countries and, possibly, crossbor- der transactions and generally includes certain obligations in terms of information sharing. Some of the key objectives for entering into such a treaty are the reduction of barriers to trade and investment, the reduc- tion of double taxation (under the principle that the same source of income should only be taxed once), and elimination of tax evasion (including through information sharing). By specifying the tax treat- ment of certain transactions and instituting mechanisms for dispute resolution, tax treaties generally improve certainty for taxpayers and tax authorities. Under these treaties, speci�c mutual agreement procedures are used to resolve international tax grievances, including double taxa- tion issues arising out of a transfer pricing transaction. The government agency responsible for implementing these dispute resolution proce- dures (and negotiating with other national tax authorities) is the compe- tent authority of the country, which has the power to bind its govern- ment in speci�c cases. While the network of tax treaties is fairly extensive among OECD countries, most developing countries have relatively few tax treaties in place. This severely restricts the ability of their tax authorities to share information and coordinate enforcement across borders. In addition, while internationally consistent transfer pricing regulations have been almost universally adopted throughout the OECD, relatively few devel- oping countries have speci�c transfer pricing regulations. This helps to explain the recourse in some developing countries to speci�c arrange- ments for individual projects and individual MNEs (through, for exam- ple, individually negotiated production-sharing contracts, which effec- tively lay out a complete legal, tax, and customs regime speci�c to individual projects), although, given the general lack of negotiating capacity and sector-speci�c expertise, it would seem more appropriate to lay out one general regime by which all related projects would abide. However, the practice of bilateral company-by-company negotiations raises the level of influence of speci�c bureaucrats and politicians. This situation often leads to practices that are judged as corrupt in OECD The Role of Transfer Pricing in Illicit Financial Flows 249 countries, and companies operating in such corrupt environments face thorny issues on the legality and licit nature of certain transactions. In countries with highly corrupt regimes or stringent exchange controls (the two often go hand in hand), companies may consider the manipula- tion of transfer prices—which is particularly easy in the absence of spe- ci�c transfer pricing regulations—as an acceptable method of pro�t repatriation. In the case above of the large petroleum company, this con- sideration was likely important given the state of the business environ- ment and the existence of severe restrictions on pro�t repatriation at the time. It is questionable whether such a technically illegal attempt to shift income out of a corrupt regime should be considered an illicit flow of funds. More recently, with the substantial reduction in exchange controls across the world, the concern of tax authorities, particularly in OECD countries, has shifted to how easy it has become for taxpayers to conceal foreign-based income on which residents should pay national tax. By accessing foreign jurisdictions offering low or zero tax rates and failing to report the proceeds to their home authorities and then not reporting the foreign-based income, these taxpayers evade their tax obligations with respect to their country of residence. Required reporting under exchange controls has provided information that is helpful in tracking the flows of funds and, therefore, in assessing tax obligations. Without transfer pricing regulations or exchange controls in the country in which the income is earned, taxpayers may �nd it particularly easy to evade tax obligations through a combination of transfer price manipulation and access to a tax haven.16 Empirical Evidence on Illicit Financial Flows from Developing Countries The wide variations in tax rates across the globe certainly lead MNEs to respond to the incentives created by these differences. We have good evi- dence that recorded pro�ts of MNEs shift among countries after a change in tax rates.17 Of course, MNEs are likely to respond in a similar fashion to comparable arbitrage opportunities created by the different mix of skills and costs among different locations within countries and across countries. After all, their performance is judged by shareholders 250 Draining Development? on their ability to maximize after-tax pro�ts, and economic ef�ciency is generally raised by the search for lower costs. What is less clear is how much of this pro�t shifting is associated with tax evasion, and, within that category, how much is due to transfer pricing. In other words, given that transfer pricing analyses typically provide a range of prices or pro�t levels consistent with the ALP, is it the case that MNEs simply select a point within the range that legally minimizes their tax liability, or is it, instead, the case that MNEs effectively choose a point outside the ALP range? The role of transfer pricing To date, there is no direct evidence on the role that transfer pricing plays in tax evasion (box 8.3). It is likely true, however, that MNEs routinely choose prices within the ALP range in a tax ef�cient manner, but this would be classi�ed, at most, as tax avoidance. Two recent studies, Bartels- man and Beetsma (2003) and Clausing (2003), provide direct evidence on the related issue of pro�t shifting associated with transfer pricing. Their �ndings are broadly consistent with tax minimization strategies, but neither of these studies distinguishes between tax avoidance and tax evasion. Both studies �nd that a 1 percentage point increase in a coun- try’s tax rate leads to a decline in reported taxable income of roughly 2.0–2.5 percent. These �ndings are consistent with the interpretation that a lower corporate tax rate tends to encourage �rms to anchor their Box 8.3. A Measure of Tax Evasion in the United States In what is likely the most rigorous exercise of its kind, the U.S. tax authority estimated that the gross tax gap for the U.S. taxation year 2001 was close to US$350 billion, of which US$30 billion was due to the corporate sector (on account of non�ling, underreporting, and underpayment).a It is not clear how much of this was associated with transfer pricing issues. In any case, this estimate is also likely to be an overestimate given that the U.S. tax authority would not have taken account of the reverse situation in which a United States–resident corporation overreports income in the United States (because the other jurisdiction is the high-tax country). a. The gross tax gap is de�ned as “the difference between what taxpayers should have paid and what they actually paid on a timely basis�; it excludes the impact of penalties and late interest collected by the tax authority. In the U.S. study, the bulk of the tax gap arose from underreporting on individual tax returns (close to US$200 billion) and on employment returns (close to US$55 billion). The Role of Transfer Pricing in Illicit Financial Flows 251 transfer pricing policy at different points within the ALP range or else to restructure their businesses in a way that increases reported pro�ts in the taxing country. For developing countries, Kar and Cartwright-Smith (2008) and Christian Aid (2009) provide estimates of trade misinvoicing practices by which pro�ts are shifted out of developing countries. The former study suggests that, between 2002 and 2006, approximately US$370 bil- lion in pro�ts were shifted out of the developing world every year, and the latter study suggests that the �gure for the period between 2005 and 2007 is more than US$1 trillion. The headline number in these reports is the estimated tax revenue loss, which is calculated at close to US$120 billion per year. (This compares with global flows of of�cial develop- ment assistance of roughly the same magnitude in 2008.) These �gures are not widely accepted, and the academic literature has raised important questions with respect to the methodology and the strength of the conclusions.18 For our purposes, these studies do not allow any conclusions on the role of transfer pricing because they do not use company-level data; given that transfer pricing is intrin- sically fact-dependent, the extent of non-ALP transfer pricing can only be determined by analyzing transaction-level data to ensure that apparent income shifting in the aggregate data is not caused by (1) rea- sonable comparability differences, (2) variations in product quality, or (3) a reflection of shifts in assets, functions, and risks.19 In fact, these studies do not even distinguish between transactions involving unre- lated and related parties. Impact on local economies and tax collections What is also less clear is the impact on the economies and on tax collec- tions of developing countries. In fact, during the recent globalization drive that included a substantial increase in the number of MNEs and in the share of intra�rm transactions in trade flows, corporate tax revenues remained stable across the world. So, why are there differences in the perceived widespread effects of transfer pricing practices and the avail- able empirical analyses? First, the existing estimates are derived from studies with substantial methodological shortcomings, and, second, the evidence of these studies runs counter to the observed response of FDI to tax rates and the recent buoyancy in corporate tax revenues. 252 Draining Development? As recently reviewed by Fuest and Riedel (2009), the existing studies are purely descriptive, employ the marginal tax rate to calculate the tax gap, and do not account for the reverse effects of pro�ts shifted into developing countries, some of which do have signi�cantly lower corporate tax rates.20 In addition, recent work on the demand for tax haven operations suggests that tax havens may not have purely negative consequences (Desai, Foley, and Hines 2006). This work indicates the following: • Multinational parent companies in industries that typically face low foreign tax rates, those that are technology intensive, and those in industries characterized by extensive intra�rm trade are more likely than others to operate in tax havens. • Firms with growing activity in high-tax countries are most likely to initiate tax haven operations, and the reduced costs of using tax havens are likely to stimulate investment in nearby high-tax countries. The same studies that provide estimates of large tax gaps created in developing countries by illicit flows from tax-induced manipulations also need to provide answers to puzzles related to (1) the sensitivity of FDI to tax rates, (2) the broadly similar levels of tax rates across coun- tries, and (3) the buoyancy of tax collections in the 1990s. Consistent with economic theory, we have good empirical evidence that FDI responds to statutory tax rates and other features of the tax sys- tem.21 Meanwhile, it is popularly thought that MNEs are �nding more and more ways to shift income (including by optimizing �nancing struc- tures, choosing advantageous accounting treatments, and using offshore �nancial centers and tax havens).22 Taken together, these facts suggest that the sensitivity of FDI to tax factors should be decreasing over time given that the increasing availability of income shifting practices suggests that the burdens of high host country rates can be more easily avoided (with- out the need for costly reallocations of real investment). However, the sensitivity of FDI to tax factors appears, if anything, to be increasing over time.23 This is consistent with tax avoidance rather than tax evasion. Stöwhase (2006) �nds that, in addition, the elasticity of FDI to tax incentives in the European Union differs strongly according to economic sector. While taxes are insigni�cant as a determinant of FDI in the pri- mary sector, tax elasticity is substantial in manufacturing, and it is even greater in service industries.24 The increasing differentiation of products The Role of Transfer Pricing in Illicit Financial Flows 253 and the growing reliance on service transactions and intangibles as one moves from the primary to the secondary and tertiary sectors would make it easier to shift pro�ts simply by manipulating transfer prices, without the need for costly reallocations. While this �nding, as indicated by Stöwhase, amounts to (indirect) evidence of income shifting, the responsiveness of FDI is evidence that pro�ts are moving in a way com- mensurate with FDI flows (that is, with shifts in assets, risks, and func- tions). In other words, this evidence is more consistent with legal tax planning and tax avoidance strategies rather than with tax evasion. We also know that corporate tax rates are broadly similar in developing countries and developed countries, notwithstanding differences in tax rates and tax structures across countries (World Bank 2004).25 For transfer pricing, this implies that manipulations that are purely tax motivated should involve a tax-evading higher import price into a developing coun- try and a tax-inducing higher export price from a developed country as much as a tax-evading higher import price into a developed country and a tax-inducing higher export price from a developing country. In this con- text, it is not clear why MNEs would consistently engage in tax-driven practices geared strictly to stripping out pro�ts from developing countries. Yet, many studies consider only one direction in mispricing transactions, which results in a signi�cant overestimation of the net tax revenue losses; this includes estimates prepared by tax departments of OECD countries. Finally, some commentators point out that the revenue collected tends to be lower in developing countries than in developed countries, notwithstanding broadly similar tax rates. Data of the World Bank’s World Development Indicators suggest that the average ratio of tax rev- enue to gross domestic product in the developed world is approximately 35 percent compared with 15 percent in developing countries and only 12 percent in low-income countries.26 This gap, it is further argued, is a sign of a nexus of phenomena, such as encroaching globalization and corporate pro�t shifting, that work against the ability of developing countries to collect needed tax revenue (generally dubbed the harmful tax competition argument). Developments during the 1990s do not appear to support the argu- ment of a race to the bottom, however. As the pace of globalization picked up in the 1990s, there was a substantial increase in the number of MNEs and a signi�cant extension of the set of options available for the 254 Draining Development? manipulation of tax liabilities. Concurrently, there were signi�cant increases in the share of intra�rm transactions in trade flows, but either slight increases or, at worst, stability in corporate tax revenues across the world (despite falls in marginal tax rates).27 Thus, clear evidence on the tax motives of transfer pricing abuses requires a disaggregated empirical analysis using evidence on how com- panies actually set their transfer pricing policies and a comparison with arm’s-length price ranges by transaction. In addition, to assess whether any related tax gap could be remedied by a tax-system response would also require evidence on the extent to which tax considerations influence transfer pricing decisions. Limited by the availability of data, the result- ing analyses now tend to be conducted at too high a level of aggregation and without a suf�cient theoretical foundation to provide any direct evi- dence on the extent of the problem or any insight into the motivation of the participants.28 Most careful analysts now explicitly recognize that care must be taken in interpreting the results of the typical regression analysis of subsidiary pro�tability on tax rates.29 On the ground, the buoyancy of corporate tax revenues across the world during a period characterized by a substantial increase in the share of intra�rm transac- tions suggests that transfer pricing did not play a signi�cant role in the manipulation of tax liabilities. A Plan of Action for Developing Countries Developing countries need primarily to become more effectively inte- grated into the international taxation regime.30 In particular, these coun- tries should do the following: • Strengthen tax administration and enforcement (institutions matter) • Improve the exchange of tax information between governments (reflecting the globalized state of business transactions and MNE operations and the consequent need for international coordination among tax authorities)31 In the area of transfer pricing, they should adopt the following: • Globally consistent regulations for transfer pricing (implement best practice) The Role of Transfer Pricing in Illicit Financial Flows 255 • Contemporaneous documentation requirements (to facilitate enforcement) • The use of advance pricing agreements, which hold the promise of simplifying the negotiation and monitoring of transfer pricing prac- tices, including by generating relatively quickly a much-needed body of locally relevant precedents32 The call to strengthen enforcement is a common refrain, but it is nonetheless of primary importance. Low tax collection rates occur more often in countries where fundamental issues of tax administration and institutional weaknesses tend to complicate (severely) the enforcement of the existing tax regime, regardless of choices on tax policy. The opera- tional challenges include not only questions of organizational design (encompassing accountability mechanisms, the relationship between policy and administration, and risk management practices), but also capability in basic tools such as auditing and a sound knowledge of legal principles. The fact that these economies tend to be economically less diversi�ed actually enhances the ability to make quick progress toward effective implementation of the ALP. MNEs operate in relatively few sectors in these economies, and tax auditors require a clear understanding of only a few types of business models, that is, by strengthening their degree of familiarity with the assets, functions, and risks deployed in only a few sectors, tax auditors can typically cover a signi�cant proportion of the tax clientele that uses transfer prices in determining tax liabilities (for example, see the case of Zambia in box 8.4). The other recommendations (exchange of information and a globally consistent transfer pricing regime) are crucial because taxpayers tend to view transfer pricing (primarily) from a risk-management point of view. As a result, there is a natural tendency for companies to undertake more efforts to comply with the requirements of those countries that have more effective regulations (including documentation requirements) and enforcement capacities (so as to minimize the possibility of compliance penalties and negative publicity). By the adoption of clear and globally consistent transfer pricing regulations, developing countries should be able to nudge MNEs into a more balanced risk assessment, which would permit the revenue authorities to direct their limited resources to 256 Draining Development? Box 8.4. Zambia: Tax Collection in the Copper Sector A recent boom in the copper sector in Zambia resulted in a substantial increase in copper exports, but only minimal increases in tax revenue: mining royalties received by the Zambian government in 2006 represented only 0.6 percent of sales. As identi�ed by advocacy group Southern Africa Resource Watch, the culprits were the �scal terms in agreements signed by the government dur- ing the slump experienced in the copper sector in the late 1990s. Faced with falling levels of investment, a collapsing state-owned mining conglomerate, and low global prices, the govern- ment agreed to renegotiate the original contracts and include sector-speci�c incentives designed to make mining investments more attractive (Horman 2010). As a result, the new mining tax regime proposes to raise tax rates, reduce incentives, and require that transfer prices be based on the price of copper quoted at the London Metal Exchange or the Shanghai Stock Exchange. For purposes of transfer pricing, the commodity nature of unprocessed copper (as exported by Zambia) makes the use of market-quoted prices the undisputed choice for a transfer pricing methodology (possibly with an adjustment for differential transportation costs). In other words, the simple adoption of the ALP would bring about the desired result with relatively little need for upfront capacity building (market quotes are widely available) and would also avoid an ad hoc sector-speci�c approach to the problem. enforcement areas where they are most needed.33 At the same time, the enhanced opportunities for international cooperation and access to a wider set of knowledge resources will tend to strengthen efforts at capac- ity building. Consistent with these priorities, OECD has focused its tax evasion work on the widespread adoption of the OECD transparency and exchange of information standard. In the circumstances, the number of tax information exchange agreements, a key platform of this approach, doubled to over 80 during the �rst six months of 2009. This should make it easier for tax authorities to share information and cooperate on cases of tax evasion, and it may even lead to the increased use of simultaneous examinations (by different tax authorities of the same taxpayers).34 Conclusion In discussing the role of transfer pricing in illicit �nancial flows in devel- oping countries, we start with a review of the current OECD-supported system based on the ALP, and we examine the differences between tax avoidance and tax evasion. Although an elegant and simple concept as a theoretical construct, the ALP becomes a fact-intensive and judgment- The Role of Transfer Pricing in Illicit Financial Flows 257 based system at the implementation level. Often, the benchmarking of comparable transactions and companies provides a range of acceptable prices and the conclusion that any point inside this range is consistent with arm’s-length pricing. In this context, selecting a lower (or higher) price within the arm’s-length range of prices, whether in response to tax objectives or otherwise, cannot be construed as anything other than a legal practice or, at most, as tax avoidance. Similarly, shifting income from one country to another in a manner that is commensurate with a shift in assets, functions, and risks cannot be construed as anything other than a legal way of business restructuring. Alternative proposals for a transfer pricing system, including grandi- ose schemes involving a global tax authority and global apportionment formulas, have been proposed and debated. Overall, these alternatives not only lack theoretical rigor, but are also less likely to win international agreement (which is crucial to ensuring that rules are consistently and fairly applied). Therefore, we argue that the current framework provides a construct that should help corporations avoid double taxation and should also help tax administrations receive a fair share of the tax base of multinational enterprises. While any set of rules can be potentially manipulated, the OECD-led initiative to develop and extend the reach of transfer pricing guidelines based on the ALP represents the best alterna- tive for a system that is theoretically sound and amenable to being imple- mented in an internationally coordinated and consistent manner. We also review the empirical evidence on the impact of transfer pric- ing in illicit flows from developing countries. Essentially, there is no direct evidence of such an impact. The majority of available studies are conducted at an aggregate level, and even the studies that provide direct evidence on the extent of income shifting arising from transfer prices do not distinguish between ALP-consistent price changes (tax avoidance) and non–ALP-consistent changes (tax evasion). There is also a dearth of analytically rigorous evidence on the impact of income shifting on developing countries and their ability to collect taxes. Finally, we discuss the importance of introducing a speci�c transfer pricing regime in developing countries (rather than relying on ad hoc measures and general antiavoidance provisions). This would ensure that tax structures and policies are more well integrated with international best practice, and, for the private sector, it would make the rules of the 258 Draining Development? game more globally consistent and transparent. However, it is clear that, at the implementation level, developing countries face signi�cant chal- lenges in adopting the OECD-style transfer pricing regime, owing not only to the malleability of the ALP concept, but also the lack of access to appropriate benchmarks by which to assess transfer prices. The impact of these dif�culties is likely to be more acute for those countries that are highly dependent on revenues from natural resources, a sector that tends to be dominated by multinationals and intra�rm trade. Notes 1. Intra�rm trade is de�ned here as crossborder trade among af�liates of the same multinational corporation. 2. The coverage of data on intra�rm trade is not comprehensive. For the United States, the best estimates now suggest that intra�rm trade accounts for approxi- mately 50 percent of all imports, and transactions involving other countries of the Organisation for Economic Co-operation and Development (OECD) are signi�cantly more likely to involve related parties. 3. For a complementary discussion, see chapter 7 by Lorraine Eden in this volume. 4. The ability to collect tax revenues from the corporate sector is particularly important for developing countries, which tend to rely relatively more on cor- porate taxes as a source of revenue. Developing countries tend to collect about half the tax revenue, as a share of gross domestic product, relative to developed countries and to rely much more heavily on indirect and trade taxes (Tanzi and Zee 2001). 5. Competent authority refers to a national body entrusted with interpreting and implementing tax conventions. The competent authority deals with situations in which taxpayers are subject to taxation not in accordance with the provisions of the relevant tax convention, including situations of double taxation. 6. In Canada, transfer pricing penalties effectively amount to 10 percent of the resulting adjustment to the transfer prices. This is a non–tax-deductible expense. 7. The start of this process involves a tax audit that may occur up to �ve or six years after the relevant �scal year and that may take years to complete. Thereafter, settlements from the competent authority process typically take years to con- clude, and the negotiation process requires multiple stages of written submis- sions. The complexity of this process means that, in most cases, the taxpayer employs highly specialized tax advisers over the course of the tax audit and the deliberations by the competent authority. 8. The types of risks that must be taken into account include market risks; risks of loss associated with the investment in and use of property, plant, and equip- ment; risks associated with the outcome of research and development activities; The Role of Transfer Pricing in Illicit Financial Flows 259 risks of collection of accounts receivable; risks associated with product liability; and �nancial risks such as those caused by currency exchange rate and interest rate variability. 9. Adjustments routinely undertaken in transfer pricing analyses account for dif- ferences in accounting practices for inventories, intangible assets, or deprecia- tion; level of working capital; cost of capital; level of market; treatment of foreign exchange gains; and impact of extraordinary items. The total of these adjustments can sometimes produce a substantial cumulative adjustment to the underlying market price, in which case the OECD guidelines question whether the degree of comparability between the selected companies and the tested party is suf�ciently high, but, in practice, there are no speci�c criteria by which to make this determination. Of course, in the absence of a more ade- quate method or more comparable data, the analysis in question may remain the most feasible, and, in any case, it may still produce a relatively tight range of arm’s-length prices. 10. The OECD views pro�t-based methods (such as the transaction net margin method or, in the United States, the comparable pro�t method) as generally less reliable than the traditional transaction methods such as the comparable uncon- trolled price, which is a product comparable method. The latter, however, requires the identi�cation of transactions involving the same product and third parties (to be used as the comparable uncontrolled transaction by which to price the related-party transaction), and, in practice, companies tend not to transact with a third party on the same terms as a related party. Differences in transac- tions with related and unrelated parties include (a) volume or other size or scale factors, (b) level of the market, (c) temporary versus permanent relationship, (d) primary sourcing versus secondary sourcing, and (e) after-sales service, war- ranty services, and payment terms. Thus, either comparability adjustments are necessary (if a potential comparable uncontrolled transaction can even be iden- ti�ed) or pro�t-based methods must be used. 11. In this context, form means that the legal setup or structure of the transactions takes precedence in terms of its characterization (and treatment) for tax pur- poses, whereas economic substance implies that the underlying business or eco- nomic rationale is paramount (regardless of the legal form). Where the eco- nomic substance concept applies, for example, tax authorities may look through intermediate transactions and merely consider the aggregate effect for purposes of tax treatment. In the speci�c case of Canada and the United States, this has promoted the use of hybrid transactions that are characterized differently in each country (so that the same transaction, for example, may be viewed as a dividend receipt by one jurisdiction and an interest payment by the other jurisdiction). 12. As de�ned in Wikipedia (http://en.wikipedia.org/wiki/Tax_avoidance_and_tax_ evasion [accessed August 31, 2009]), “tax avoidance is the legal utilization of the tax regime to one’s own advantage, in order to reduce the amount of tax that is 260 Draining Development? payable by means that are within the law. By contrast tax evasion is a felony and the general term for efforts to not pay taxes by illegal means.� Consistent with these de�nitions, the OECD has been focusing on combating tax evasion (OECD 2009). 13. Concerns with egregious cases of tax evasion have recently come to the fore with the widespread publicity accorded to the 2008 case involving LGT Bank of Liech- tenstein. After the initial case became public in Germany, a U.S. Senate investiga- tions committee extended the inquiry to possible abuses by American taxpayers; other countries followed suit; and the issue of tax evasion and tax havens has since become a key agenda item at various international policy forums. The report by the U.S. committee estimated that offshore tax havens supported tax evasion schemes that reduced tax collections in the United States by about US$100 billion per year (compared with annual gross domestic product of around US$14.2 tril- lion, annual federal tax revenue of close to US$2.5 trillion, and annual federal spending on Medicare and Medicaid of some US$675.0 billion) (see Permanent Subcommittee on Investigations, 2008, “Tax Haven Banks and U.S. Tax Compli- ance: Staff Report,� Report, U.S. Senate, Washington, DC, July 17). 14. In this region of the world, government routinely collects royalty and tax pay- ments in kind. At the time the contract was signed, the price of the commodity was near historical lows. 15. In the absence of a speci�c transfer pricing regulation, there was no speci�c authority to enforce compliance with the ALP even though the tax laws included general antiavoidance provisions (which are typically vaguely worded, dif�cult to interpret without legal precedents, and cumbersome to implement). 16. Tax authorities in some OECD countries are combating the lack of disclosure that is part of these tax evasion schemes by accessing a wider set of information. In the United States, for example, the Internal Revenue Service has started rely- ing on records from American Express, MasterCard, and Visa to track the spend- ing of U.S. citizens using credit cards issued in Antigua and Barbuda, The Baha- mas, and Cayman Islands. As a direct result of such information, the accounting �rm KPMG admitted that its employees had criminally generated at least US$11 billion in phony tax losses, often routed through Cayman Islands, which cost the United States US$2.5 billion in tax revenue, and, in 2007, the drug maker Merck agreed to pay US$2.3 billion to the government to settle a claim that it had hid- den pro�ts in a Bermudan partnership. 17. For example, Maf�ni and Mokkas (2008, 1) �nd that “a 10 percentage points cut in the statutory corporate tax rate would increase multinationals’ measured [total factor productivity] by about 10 percent relative to domestic �rms, con- sistent with pro�t-shifting by multinationals.� 18. For a review, see Fuest and Riedel (2009). Even for the United States, where more disaggregated and higher-quality data are available, the U.S. Government Accountability Of�ce recognizes that the usefulness of the available estimates The Role of Transfer Pricing in Illicit Financial Flows 261 and our ability to interpret the results crucially “[depend] on how many of the important nontax characteristics have been included in the analysis� (GAO 2008, 5). This is because the statistical relationships exploited by most studies are liable to be considerably affected by the researchers’ ability to control for signi�cant industry- and corporate-speci�c effects. The importance of these controls is highlighted by the U.S. Government Accountability Of�ce (GAO 2008), which indicates that, for the United States, foreign-controlled domestic corporations tended to report lower tax liabilities than United States–controlled domestic corporations between 1998 and 2005 (by most measures), but also that the percentage reporting no tax liability was not statistically different after 2001. Most importantly, these two types of corporations differ signi�cantly in age, size, and industry, and companies in different sectors are likely to have dif- ferent �nancial characteristics (including levels of assets) and therefore different levels of receipts and pro�tability. The U.S. Government Accountability Of�ce does not attempt to explain the extent to which such factors affect differences in reported tax liabilities (GAO 2008). 19. There is, for example, empirical support for the notion that price differences within product groups are mostly related to quality differences, where “devel- oping countries tend to export low-end/low-price products whereas devel- oped countries export high-end products with higher prices� (Fuest and Rie- del 2009, 33). 20. For more detailed considerations, see Fuest and Riedel (2009) and our discus- sion elsewhere in the text on the incentives for mispricing on both the import side and the export side. 21. The empirical evidence also con�rms that FDI is responsive to other locational advantages such as country risk, public inputs, and skill availability. 22. For a more scholarly summary, see Altshuler and Grubert (2006). For a more populist treatment, see recent publications by the Tax Justice Network, Christian Aid, and Oxfam, among others. 23. See, for example, Mutti and Grubert (2004) and, for a brief summary, Zodrow (2008). 24. The results of this research are summarized in Haufler and Stöwhase (2003), as follows: • FDI in the primary sector (consisting of agriculture, �shing, mining and quarrying) has a tax elasticity of around zero, implying that FDI is not driven by tax incentives. • Investment in the secondary sector (manufacturing) is negatively and signi�- cantly affected by an increase in effective taxation. The tax elasticity is around −2, implying that a 1 percent increase in the tax rate of the host country decreases FDI by roughly 2 percent, all else being equal. • Compared with the secondary sector, FDI in the tertiary sector (consisting of investment in service industries such as transport, communication, and 262 Draining Development? �nancial intermediation) is even more strongly affected by an increase in tax rates, and the tax elasticity for this sector is around −3. • By weighing sector-speci�c elasticities with the sector’s share of total FDI, we obtain an average tax elasticity of −2.5. This average tax elasticity is compa- rable to the results derived in more aggregated studies. 25. In fact, the statutory tax rates in the most economically signi�cant developing countries tend to be lower than the average in OECD countries (Gordon and Li 2009). 26. See World Development Indicators (database), World Bank, Washington, DC, http://data.worldbank.org/data-catalog/world-development-indicators/. Econ- omists contend that these differences in revenue collection can be explained by both (a) the fact that demand for public services increases more than propor- tionally as income rises and (b) weaknesses in the ability of developing countries to raise the revenue required for the provision of adequate public services. On the latter, the World Bank (2004) emphasizes the narrowness of the tax base and problems in tax administration. 27. World Bank (2004) reports that the only exception to this outcome was the region of Central and Eastern Europe and Central Asia, where revenues fell because of privatization and a general contraction in the size of the state. 28. Fuest and Riedel formulate the issue as follows: Attempts to estimate the amount of tax avoidance and tax evasion there- fore have to build on concepts which exploit correlations between observ- able and statistically documented variables and evasion. These data prob- lems may explain why there is very little reliable empirical evidence on tax avoidance and evasion in developing countries. The existing studies mostly rely on highly restrictive assumptions and have to make use of data of mixed quality. (Fuest and Riedel 2009, 6) 29. For example, see Fuest and Riedel (2009); GAO (2008); McDonald (2008). 30. We adopt the view that, in the presence of an international tax regime, countries are effectively constrained to operating in the context of that regime and not free to adopt any international tax rules they please. This view is espoused in Avi- Yonah (2007). The basic norms that underlie this international tax regime are the single tax principle (that is, that income should be taxed once, no more nor less) and the bene�ts principle (that is, that active business income should be taxed primarily at the source and that passive investment income should be taxed primarily at residence). 31. For a discussion on the role of information sharing in international taxation, see Keen and Ligthart (2004), who consider that this form of administrative coop- eration may be a feasible substitute for the dif�cult-to-achieve target of the coordination of tax systems. 32. For a strong argument on the role of advance pricing agreements and tax trea- ties in improving the current international tax regime, see Baistrocchi (2005). The Role of Transfer Pricing in Illicit Financial Flows 263 33. This behavior by taxpayers and the suggested policy response are broadly con- sistent with economic models of tax evasion (following the work on the eco- nomics of crime by Nobel laureate economist Gary Becker and the tax evasion model of Allingham and Sandmo), which contend that the level of evasion of income tax depends on the level of punishment provided by law. The suggestion herein extends this notion to the practice of tax avoidance. 34. For some developing countries, a key obstacle to effective implementation of information sharing is the lack of the capacity required to determine what infor- mation should be requested from other countries and also to make proper and timely use of the information received. References Altshuler, R., and H. Grubert. 2006. “Governments and Multinational Corporations in the Race to the Bottom.� Tax Notes 110 (8): 459–74. Avi-Yonah, R. S. 2007. “Tax Competition, Tax Arbitrage and the International Tax Regime.� Working Paper 07/09, Oxford University Centre for Business Taxa- tion, Oxford. Baistrocchi, E. 2005. “The Transfer Pricing Problem: A Global Proposal for Simpli�- cation.� Tax Lawyer 59 (4): 941–79. Bartelsman, E. J., and R. M. W. J. Beetsma. 2003. “Why Pay More? Corporate Tax Avoidance through Transfer Pricing in OECD Countries.� Journal of Public Eco- nomics 87 (9–10): 2225–52. Christian Aid. 2009. “False Pro�ts: Robbing the Poor to Keep the Rich Tax-Free.� Christian Aid Report, March, Christian Aid, London. Clausing, K. A. 2003. “Tax-Motivated Transfer Pricing and US Intra�rm Trade Prices.� Journal of Public Economics 87 (9–10): 2207–23. Desai, M. A., C. F. Foley, and J. R. Hines Jr. 2006. “The Demand for Tax Haven Oper- ations.� Journal of Public Economics 90 (3): 513–31. Fuest, C., and N. Riedel. 2009. “Tax Evasion, Tax Avoidance and Tax Expenditures in Developing Countries: A Review of the Literature.� Report, Oxford University Centre for Business Taxation, Oxford. GAO (U.S. Government Accountability Of�ce). 2008. “Tax Administration: Com- parison of the Reported Tax Liabilities of Foreign- and U.S.-Controlled Cor- porations, 1998–2005.� Report GAO-08–957 (July 24), GAO, Washington, DC. http://www.gao.gov/products/GAO-08-957. Gordon, R., and W. Li. 2009. “Tax Structures in Developing Countries: Many Puzzles and a Possible Explanation.� Journal of Public Economics 93 (7–8): 855–66. Haufler, A., and S. Stöwhase. 2003. “Taxes as a Determinant for Foreign Direct Invest- ment in Europe.� CESifo DICE Report 4 (2): 45–51, CESifo Group, Munich. Horman, Chitonge. 2010. “Zambia’s Development Agreements and the Soaring Cop- per Prices.� Resource Insight 11 (September 29), Southern Africa Resource Watch, Johannesburg. 264 Draining Development? Kar, D., and D. Cartwright-Smith. 2008. “Illicit Financial Flows from Developing Countries, 2002–2006.� Global Financial Integrity, Washington, DC. http:// www.g�p.org/storage/g�p/economist%20-%20�nal%20version%201-2-09.pdf. Keen, M., and J. E. Ligthart. 2004. “Incentives and Information Exchange in Interna- tional Taxation.� CentER Discussion Paper 2004–54, Center for Economic Research, Tilburg University, Tilburg, the Netherlands. Maf�ni, G., and S. Mokkas. 2008. “Transfer-Pricing and Measured Productivity of Multinational Firms.� Unpublished working paper, Oxford University Centre for Business Taxation, Oxford. McDonald, M. 2008. “Income Shifting from Transfer Pricing: Further Evidence from Tax Return Data.� Technical Working Paper 2, Of�ce of Technical Analysis, U.S. Department of the Treasury, Washington, DC. Mutti, J., and H. Grubert. 2004. “Empirical Asymmetries in Foreign Direct Invest- ment and Taxation.� Journal of International Economics 62 (2): 337–58. OECD (Organisation for Economic Co-operation and Development). 1995. “Trans- fer Pricing Guidelines for Multinational Enterprises and Tax Administrations.� OECD, Paris. ———. 2009. “An Overview of the OECD’s Work on Countering International Tax Evasion.� Background Information Brief, August 28, Centre for Tax Policy and Administration, OECD, Paris. Stöwhase, S. 2006. “Tax-Rate Differentials and Sector Speci�c Foreign Direct Invest- ment: Empirical Evidence from the EU.� FinanzArchiv: Public Finance Analysis 61 (4): 535–58. Tanzi, V., and H. Zee. 2001. “Tax Policy for Developing Countries.� Economic Issues 27 (March), International Monetary Fund, Washington, DC. UNCTAD (United Nations Conference on Trade and Development). 2008. World Investment Report 2008: Transnational Corporations, and the Infrastructure Challenge. Geneva: UNCTAD. World Bank. 2004. World Development Report 2005: A Better Investment Climate For Everyone. New York: World Bank; New York: Oxford University Press. Zodrow, G. 2008. “Corporation Income Taxation in Canada.� Canadian Tax Journal 56 (2): 392–468. 9 Accounting for the Missing Billions Richard Murphy Abstract This chapter considers, �rst, whether substantial transfer mispricing by major corporations that contributes to a loss of at least US$160 billion a year to developing countries is plausible in the context of total likely corporate pro�ts earned worldwide in a year.1 Second, it tests whether corporate activities in developing countries and in the extractive industries in particular might be especially prone to this abuse. Third, it considers whether this sum could be hidden from view within the accounts or �nancial statements of the multinational corpo- rations (MNCs) that might be perpetrating the mispricing. Fourth, it explores the possibility that the MNCs might use secrecy jurisdictions to assist in hiding these transactions from view. Some of the data used in this chapter have been researched for the Mapping the Faultlines Project of the Tax Justice Network (TJN), �nanced by the Ford Foundation, the results of which are being published at http://www.secrecyjurisdictions.com/. The author is research director of that project. Additional tables and �gures from this chapter can be found at http://go.worldbank.org /N2HMRB4G20. 265 266 Draining Development? In each case, it is suggested that the behavior described is plausible and that, as a consequence, losses of the estimated amount are also plau- sible, although not proven to exist as a result of this work. Introduction Since 2000, a body of literature, mainly emanating from nongovernmen- tal organizations (NGOs), has developed suggesting that systematic transfer mispricing is taking place within the world’s MNCs. This lit- erature asserts, in the �rst instance, that the flows in question abuse international standards on transfer pricing implicit in the arm’s-length pricing rules of the Organisation for Economic Co-operation and Devel- opment (OECD) and that the abuse results in signi�cant loss in revenues to developing countries.2 Estimates of the loss vary. Christian Aid’s estimate (2008) �nds that transfer mispricing and related abuses result in the loss of corporate tax revenues to the developing world of at least US$160 billion a year, a �g- ure that, it notes, is more than 1.5 times the combined aid budgets of the rich world, which totaled US$103.7 billion in 2007. Raymond Baker (2005) proposes that total annual illicit �nancial flows (IFFs) might amount to US$1 trillion, asserting, in the process, that these flows pass illegally across borders aided by an elaborate dirty money struc- ture comprising tax havens, �nancial secrecy jurisdictions, dummy corpo- rations, anonymous trusts and foundations, money laundering techniques, and loopholes intentionally left in the laws of western countries.3 Of this amount, Baker estimates that some US$500 billion a year flows from devel- oping and transitional economies. He suggests that at least 65 percent of these flows may be accounted for by transfer mispricing. Baker’s �ndings have been endorsed by Kar and Cartwright-Smith (2008), who estimate that, in 2006, developing countries lost US$858.6 billion to US$1.1 trillion in illicit �nancial outflows. Note that the term illicit is appropriately used in this chapter. The Oxford English Dictionary de�nes illegal as contrary to or forbidden by law, but illicit as forbidden by law, rules, or custom. The distinction is important. Transfer mispric- ing is illicit; it is contrary to known rules or customs, but, in many of the transactions of concern, it is not illegal because, as noted later in this Accounting for the Missing Billions 267 chapter, double tax agreements (DTAs) that would make it so are not in place, nor are relevant local laws. NGOs and civil society organizations are not the only source of litera- ture proposing that transfer mispricing might be used to reallocate prof- its across jurisdictions. There is a substantial body of academic literature asserting that this practice is commonplace, although this literature does not focus on developing and transitional economies in coming to its conclusion. For example, Dischinger and Riedel (2008) state that intan- gible assets such as patents and trademarks are increasingly seen as the key to competitive success and as the drivers of corporate pro�t, but also constitute a major source of the opportunity for pro�t shifting in MNCs because of the highly nontransparent transfer pricing process. They argue that this provides MNCs with an incentive to locate intangible property in jurisdictions with relatively low corporate tax rates. They �nd evidence to support this activity, showing that the lower a subsid- iary’s tax rate relative to other members of a multinational group, the higher the subsidiary’s level of intangible asset investment. Harry Huizinga (2009), in a wide-ranging review, indicates that MNCs can relocate pro�ts in a number of ways. First, they could change real activity, that is, where they locate; second, they could manipulate transfer prices; third, they could choose the location of intangible assets and associated income such as royalties (as Dischinger and Riedel 2008 emphasize); or, fourth, they could choose the location of their headquar- ters to create possible favorable international double taxation conse- quences. Having reviewed the current literature within Europe on this issue, Huizinga concludes that international pro�t shifting erodes the corporate tax base in Europe, that the best approach to tackling the problem is to eliminate incentives for �rms to shift pro�ts, and that international policy cooperation is necessary to achieve this. Notably, the academic studies, in contrast to those by NGOs, �nd that the scale of income shifted is relatively small, only a few percentage points, at most, of the tax base. For example, in chapter 4 in this volume, Fuest and Riedel posit that Baker (2005), Kar and Cartwright-Smith (2008), Christian Aid (2008), and others all overstate the extent of trans- fer mispricing, although all these authors reject the assertion because Fuest and Riedel make a fundamental error by assuming that overpriced inflows and underpriced outflows may be netted off for the purposes of 268 Draining Development? assessing resulting tax losses, when, in fact, these sums should be aggre- gated for this purpose. In chapter 10 in this volume, Nitsch also ques- tions the same group of estimates, suggesting that the forging of docu- mentation relating to IFFs may not be motivated by taxes. It is a curious argument if the consequence is tax loss whatever the motive. There are, however, reasonable methodological grounds that explain why this low level of transfer mispricing is reported by academic studies, all these grounds deriving from signi�cant methodological weaknesses in the studies. First, most academic studies on this issue use database information supplied by agencies that summarize accounting data; as a result, they do not use the accounts themselves as their data source, losing considerable vital information on tax as a consequence (see below). Second, these data sets tend to result in the use of pretax pro�t as a proxy for taxable income—which is rarely appropriate—and in the use of pro�t and loss account tax charges as a proxy for taxes paid. This use of the pro�t and loss account tax charge as an indicator of taxes paid is almost always inappropriate. This is because these charges are invariably made up of two parts. The �rst is the current tax charge. This is the tax that a company actually believes will be paid within 12 months of a period end as a result of the pro�t arising during the accounting period. It is, therefore, a reasonable measure of the tax liability accruing during the period. The second component of the tax charge in a pro�t and loss account is described as the deferred tax charge. Deferred tax is not tax at all. Rather, it represents an accounting entry relating to taxes that may potentially (but will not de�nitely) be paid in future periods as a conse- quence of transactions that have occurred in the current period. So, for example, if the tax relief on the acquisition of �xed assets in the period exceeds the charge for depreciation on these assets in the pro�t and loss account, there is a potential deferred tax charge in future periods if the situation were to reverse. The liability is provided in the accounts, even though it may never arise simply because the situation may not reverse. Similarly, if transfer mispricing can defer the recognition of pro�t in a high-tax parent-company location by the current relocation of the prof- its in question to a secrecy jurisdiction where tax is not currently due on the pro�ts, then a deferred tax provision can be made for the potential liability arising on the eventual remittance of the pro�ts to the parent company, but there is no guarantee that this sum will ever be due. As a Accounting for the Missing Billions 269 consequence, many consider deferred tax charges as accounting �ction and as unreliable, and they are, curiously, the only liability included in the set of accounts whether or not there is any prospect of their settle- ment. In this case, to include them in the tax charge in considering the real taxation paid by corporations is seriously misleading and under- mines many existing studies of effective taxation rates among MNCs. Thus, the methodology used in these studies is inappropriate for apprais- ing the transfer mispricing that these studies survey. Using accounting data from the companies, Murphy (2008) shows that the largest 50 companies in the FTSE 100 increased their net deferred tax provisions (that is, the cumulative provisions that are made by a company for deferred taxation arising over the period of trading to date and that are shown as a long-term liability on balance sheets) from £8 billion to £46 billion between 2000 and 2006. This hints at the existence of signi�cant tax avoidance that database information on the taxation charges in company accounts is unlikely to reveal. The room for disagreement is thus substantial on whether there is, or not, a major transfer mispricing issue that might have particular rele- vance to developing nations and in which secrecy jurisdictions may play a signi�cant part in a way that policy change may need to address.4 It is not the purpose of this chapter to resolve whether the mispricing takes place or not. The chapter has another purpose, which is approached from an accounting and auditing perspective. An audit tests the credibil- ity of reported data. This is not a test of whether the variable is right or wrong: �nancial audits do not offer an opinion on this to their users. Instead, the audit seeks to test whether the variable may be true and fair. In seeking to prove or disprove the credibility of data, auditors have, for some time, realized that a microapproach, that is, veri�cation based on transactional data alone, is unlikely to provide all (and, in some cases, any) of the data needed to determine the likely credibility of the overall stated position. The alternative approach involves verifying data by test- ing the credibility of sums in total. This testing can take a number of forms. For example, do the data fall within the known range of plausible outcomes based on the third-party data that are available? Alternatively, are the data within the likely pat- tern of outcomes that may be observed within the entity that is being tested? And are the total data plausible in that they are consistent with other known totals? 270 Draining Development? Importantly and, in the current case, crucially, all such tests must take into consideration the broader commercial, regulatory, legal, and risk environments in which the transactions or balances being considered occur. An auditor is not allowed to consider numerical data in isolation; the use of such data must be contextualized. This chapter seeks, �rst, to test the hypothesis according to which transfer mispricing by major corporations that gives rise to tax losses amounting to at least US$160 billion a year among developing countries is plausible in the context of the total likely worldwide corporate pro�ts in a year during the same period of reference (that is, pre-2008). Second, it tests whether corporate activities in developing countries and, particularly, in the extractive industries may be especially prone to this abuse. Third, it considers whether this sum may be hidden from view within the accounts and �nancial statements of the MNCs that might be perpe- trating the mispricing. Fourth, it explores the possibility that MNCs may use secrecy juris- dictions to assist in hiding these transactions from view. The rest of the chapter is divided into �ve sections. The next section explores the tax rates that MNCs have and are likely to face; it shows that much of the existing literature on this subject offers misleading indica- tions of likely effective tax rates. In the subsequent section, the state of transfer pricing regulation and practice within the extractive industries is explored on the basis of a range of sources on which the author has worked over a number of years. The following section looks at the way in which MNCs are structured and how this structure interacts with the corporate and tax law of the locations that host some of these MNCs; it also contrasts these relation- ships with the requirements of International Financial Reporting Stan- dards (IFRSs), which are the standards that govern the �nancial report- ing of most such entities now that U.S. Generally Accepted Accounting Principles are converging with the standards issued by the International Accounting Standards Board. The penultimate section considers the nature of secrecy jurisdictions. Data on the use of such locations by MNCs are presented. The role of the Big Four accounting �rms is touched upon. The evidence is drawn together in the concluding section. Accounting for the Missing Billions 271 Multinational Corporations and Their Tax Rates Data sources This part of the chapter seeks to compare the corporate tax rates offered by a wide range of jurisdictions over time. The basic data source for the time series data on tax rates used in this report is the annual corporate tax rate reviews published by KPMG, a Big Four �rm of accountants and business advisers. KPMG has been publishing the reviews since 1996, and, although the jurisdictions reviewed have varied slightly over the period, the data have always covered between 60 and 70 jurisdictions in each annual report up to and through 2008, all of which have been included in the current survey. The KPMG data consistently cover 30 OECD countries and the 15 preenlargement members of the European Union (EU). The other jurisdictions surveyed vary widely. A consistent feature is that few of the places are recognized secrecy jurisdictions. The data on populations and gross domestic product (GDP) used in this part of the chapter were extracted from the CIA World Factbook in July 2009.5 As a result of the omission of tax rate data for secrecy jurisdictions, an alternative data source for these locations has been used. Given that the KPMG data were, without doubt, produced for marketing purposes, another, similar source has been sought. The source the author of the chapter has settled on is a data set downloaded from OCRA Worldwide.6 These data, which were extracted in November 2008, relate to the corpo- rate tax rates of secrecy jurisdictions as applied to foreign source income. This is relevant for the purpose of this review because one is concerned with transfer pricing, and the income that will pass through these loca- tions in connection with intragroup trades is likely to be considered for- eign source income in these jurisdictions; the tax rates provided by OCRA will therefore be the appropriate tax rates to consider. Methodology The KPMG data are summarized in table 9A.1.7 The data are categorized according to the following characteristics for the purposes of the analysis: 1. Whether or not the jurisdiction was a member of the OECD 2. Whether or not the jurisdiction was one of the 15 EU preenlargement states (the EU15) 272 Draining Development? 3. Whether the jurisdiction was large or small (for these purposes, a large jurisdiction has a population over 15 million, which splits the data into two broadly equal parts) 4. Whether the jurisdiction had a high or low GDP per head (for these purposes, a high GDP is above US$25,000 in 2009, which splits the data into two roughly equivalent parts) 5. Whether or not the jurisdiction had a high proportion of govern- ment spending in relation to GDP (in this case, 30 percent govern- ment spending as a proportion of GDP indicates high spending) The categories have been chosen broadly to reflect developed and less-developed nations (the developed nations are the OECD members) and large and small jurisdictions. Categorizing according to a high or low tax spend also broadly reflects the effectiveness of the tax system in the jurisdiction because it is likely that those jurisdictions with low spending had ineffective tax collection systems given that this is com- monplace in developing countries. The average data for each year have then been calculated for each of these groups. The resulting data set is reproduced in table 9A.2.8 Initial results The initial results are best presented graphically. An overall summary of the data is shown in �gure 9.1. There is a strong and, in almost all cases, persistent downward trend in nominal corporate tax rates over the period under review. This, how- ever, does not reveal much of the subtlety inherent in what is happening, which greater exploration of the data reveals. First, with regard to the OECD countries (of which the EU15 are an effective subset), there has been a signi�cant decline in notional rates that has seen them converge with the rates offered by the other, non- OECD countries. Rather surprisingly, in 2008, OECD country rates fell, on average (weighted by the number of countries), below the average rate of non-OECD countries. Given the substantial differential of more than 6 percentage points in 1997, this is a remarkable change. The play- ing �eld, much talked about in OECD circles over many years, appears to have been leveled. Accounting for the Missing Billions 273 Figure 9.1. Corporate Tax Rates, Initial Results, 1997–2008 40 Average tax rate 38 Weighted average based on GDP of country 36 Average tax rate in the EU 15 Average tax rate in the OECD 34 Average tax rate in non OECD 32 Average tax rate in large countries percent 30 Average tax rate in small countries 28 Average tax rate in countries with high GDP per head 26 Average tax rate in countries with low GDP per head 24 Average tax rate in countries with high state spending Average tax rate in countries 22 with low state spending 20 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Source: Author compilation, based on data of Tax Tools and Resources (database), KPMG, Zug, Switzerland, http://www .kpmg.com/Global/en/WhatWeDo/Tax/tax-tools-and-resources/Pages/default.aspx. This story is not repeated elsewhere. The data for large and small jurisdictions do, for example, show a persistent gap in tax rates between these two groups, the extent of which marginally increases during the period. The trend in rates seems clear: smaller jurisdictions (those with populations of less than 15 million people) would appear to wish to cre- ate competitive advantage by having lower corporate tax rates. This trend is also found by comparing jurisdictions with high GDP per head and jurisdictions with low GDP per head. In this case, it is, however, clear that the margin is closing: a level playing �eld is being created between these sets of jurisdictions. This is not surprising: there is signi�cant overlap between the high-GDP jurisdictions and the OECD countries. 274 Draining Development? A comparison of jurisdictions with high and low state spending in pro- portion to GDP shows an even more marked contrast. Quite surprisingly, by 2008, the jurisdictions with high state spending (in excess of 30 percent in proportion to GDP) were offering lower notional corporate tax rates than jurisdictions with a lower (under 30 percent) proportion of GDP going to state spending. The implications of the change would appear clear: this is the consequence of a shift in the tax burden in high-spending jurisdictions from corporations to individuals that was rapid and marked. These �ndings are signi�cant. They con�rm what other literature (of which there is a considerable body) has also shown, that is, that corpo- rate tax rates are steadily falling. However, unlike most surveys of such rates, which are concentrated most often solely on EU or OECD coun- tries, these data show that the issue of changing corporate tax rates is more complex than one may suspect at �rst sight.9 There are strong dif- ferences in the trends that a simple analysis does not reveal. Even so, the analysis noted here does not show the whole picture with regard to notional tax rates because a simple averaging methodology has been used to present the data. This means that the tax rates surveyed have been totaled over the set of jurisdictions and divided by the number of jurisdictions within the set. This standard methodology is flawed. To assume that all jurisdictions stand equal in the assessment of changes in average tax rates is inappropriate: it would seem that weighting should be an essential part of any analysis. There are a number of ways to weight these data. The usual method is simply to attribute an equal weight to each jurisdiction. This would be misleading. The method gives an undue emphasis to the tax rates of small economies with limited populations. Because such economies are frequently associated with secrecy jurisdiction activity, there is an obvi- ous risk of inherent bias. It is precisely for this reason that an alternative weighting has been used in the current exercise, resulting in a perspec- tive that is different from the one usually provided by analyses of this sort, which are undertaken, in most cases, by members of the accounting profession, who have an inherent bias (as noted elsewhere below) toward secrecy jurisdiction activity. As a necessary alternative, a different weight- ing has been used here. We have weighted by the GDP of the jurisdiction offering the tax rate. There is a good reason for this choice: GDP is a reasonable indication of the size of national markets, and this is a good Accounting for the Missing Billions 275 proxy for the capacity to generate pro�ts. Because this exercise is focused on the taxation of pro�ts in particular locations, this weighting is likely to indicate the effective tax rates that companies are seeking to avoid by transfer pricing activity given that the taxation of pro�ts is the motive for using transfer pricing. The only other viable method that might indi- cate the impact of pro�t shifting through transfer mispricing is to weight tax rates by the population of the jurisdictions: such an exercise would indicate a shift from locations with high populations, requiring taxation revenue to service local need, to locations with low population, requir- ing little taxation to service the needs of the population (this is also, of course, a characteristic of secrecy jurisdictions). This second weighting method is also considered in the analysis offered below. Weighting by GDP supplies a different picture of average tax rates, as shown in �gure 9.2. The weighting in �gure 9.2 has been carried out using notional 2009 GDPs expressed in U.S. dollars as a consistent rank- ing mechanism that is unlikely to have introduced distortion because relative positions are unlikely to have changed materially for this pur- pose over time. Average tax rates fall within the ranges noted within the standard economic literature mentioned above. This literature gives, however, a misleading view. Because of the presence of the biggest economy in the Figure 9.2. Corporate Tax Rates, Weighted by GDP, 1997–2008 40 35 percent 30 25 20 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Average tax rate Weighted average based on GDP of country Source: Author compilation using 2009 GDP data taken from World Factbook (database), Central Intelligence Agency, Washington, DC, https://www.cia.gov/library/publications/the-world-factbook/. 276 Draining Development? world (the United States), which has one of the highest effective tax rates on corporate pro�ts in the world (notionally stated, as are all rates used here) if federal and state taxes are combined, and many other major countries that combine high absolute GDPs and high tax rates, the average corporate tax rate weighted by GDP is much higher than the apparent simple average; in 2008, the difference was about 5 percent. This is signi�cant. Murphy (2008) �nds that the effective current rates of corporation tax paid by the largest 50 companies in the FTSE 100 fell steadily, from 26 percent in 2000 to about 22 percent in 2006, both com- pared with a headline rate of 30 percent. This closing rate differential of 8 percent looks substantially more signi�cant, however, if it is compared with the worldwide weighted average tax rating of about 34 percent in 2007 calculated on the basis noted above. A 12 percent differential, assuming (as is likely) that many of these companies have signi�cant sales in the United States, seems high indeed. Including data from the OCRA data set on some secrecy jurisdictions— only those secrecy jurisdictions on which OCRA provides data have been used in the survey discussed hereafter, and only those on which OCRA provides speci�c rate data have been included in the exercise—changes the perspective on these data once again.10 Data are only available for 2008; so, trend analysis cannot be offered, but, even so, the position is quite differ- ent. The data are summarized in table 9A.3.11 The summary in table 9.1 is based on the 90 jurisdictions used for computational purposes. The equiv- alent data, excluding secrecy jurisdiction locations, are shown in table 9.2. Table 9.1. Summary Data: Corporate Tax Rate, Including Secrecy Jurisdiction Locations, 2008 percent Description Rate Notional simple average of corporate tax rates 20.30 Corporate tax rate weighted by GDP 32.10 Corporate tax rate weighted by population 29.90 Source: Author compilation based on data of Jurisdiction Centre (database), OCRA World- wide, Isle of Man, United Kingdom, http://www.ocra.com/jurisdictions/index.asp; Tax Tools and Resources (database), KPMG, Zug, Switzerland, http://www.kpmg.com/Global /en/WhatWeDo/Tax/tax-tools-and-resources/Pages/default.aspx; World Development Indicators (database), World Bank, Washington, DC, http://data.worldbank.org /data-catalog/world-development-indicators/ (for GDP). Accounting for the Missing Billions 277 Table 9.2. Summary Data: Corporate Tax Rate, Excluding Secrecy Jurisdiction Locations, 2008 percent Description Rate Notional simple average of corporate tax rates 26.77 Corporate tax rate weighted by GDP 32. 1 0 Corporate tax rate weighted by population 29.90 Source: Author compilation based on data of Jurisdiction Centre (database), OCRA World- wide, Isle of Man, United Kingdom, http://www.ocra.com/jurisdictions/index.asp; Tax Tools and Resources (database), KPMG, Zug, Switzerland, http://www.kpmg.com/Global /en/WhatWeDo/Tax/tax-tools-and-resources/Pages/default.aspx; World Development Indicators (database), World Bank, Washington, DC, http://data.worldbank.org /data-catalog/world-development-indicators/ (for GDP). The secrecy jurisdictions all have 0 percent tax rates. They therefore do not change the weighted data, but they do signi�cantly reduce the simple average data. It is obvious that excluding secrecy jurisdiction data from a sampling of average corporate tax rates, as has been conventional in most academic reviews to date, makes a substantial difference in the presentation of information. If the simple averaging method is used and secrecy jurisdictions are excluded from review, a quite misleading per- spective on current likely effective tax rates is presented. There is one computation to note. For several of the jurisdictions in the KPMG data set, OCRA notes that a differential tax rate is available for foreign source income. For these jurisdictions alone and only if OCRA could indicate the alternative rate that was available, we have undertaken further analysis using this rate for foreign source income. The jurisdictions for which this has been done are Hong Kong SAR, China; Hungary; Ice- land; Israel; Luxembourg; the United Kingdom, where limited liability partnerships are tax transparent; and the United States, where limited lia- bility companies offer the same �scal transparency. Singapore and Swit- zerland offer differential rates, but we do not restate them here because OCRA does not indicate the alternative rate, which varies according to the circumstances. Our results are shown in table 9.3. The resulting tax rate data weighted by GDP look much closer to the tax rate actually found if one examines effective corporate tax rates declared by companies, as noted by Murphy (2008), than to any tax rate data one may present by undertaking calculations weighted simply by 278 Draining Development? Table 9.3. Summary Data: Tax Rate on Foreign Source Income, Selected Secrecy Jurisdiction Locations, 2008 percent Description Rate Notional simple average of corporate tax rates 18.48 Corporate tax rate weighted by GDP 20.70 Corporate tax rate weighted by population 27.20 Source: Author compilation based on data of Jurisdiction Centre (database), OCRA World- wide, Isle of Man, United Kingdom, http://www.ocra.com/jurisdictions/index.asp; Tax Tools and Resources (database), KPMG, Zug, Switzerland, http://www.kpmg.com/Global /en/WhatWeDo/Tax/tax-tools-and-resources/Pages/default.aspx; World Development Indicators (database), World Bank, Washington, DC, http://data.worldbank.org /data-catalog/world-development-indicators/ (for GDP). the number of jurisdictions based on the notional tax rates of all juris- dictions that are not secrecy jurisdictions. Ring-fences increase the num- ber of jurisdictions offering low tax rates. In combination with the weighted and secrecy jurisdiction data, this understanding adds a new approach to analyzing these data. The developing-country perspective A �nal dimension to this issue needs to be noted. The KPMG data include some transition countries, but few developing countries, and Africa is, for example, seriously underrepresented. Keen and Mansour (2009) provide some data to correct this omission, although only in graphical form, as shown in �gure 9.3. The statutory corporate income tax rate plot relating to tax rates is relevant here. Simple averaging of the tax rates in Africa (with all the inherent faults in this process, which is used here, as elsewhere in most economic literature on this issue) shows that corporate income tax rates in Africa fell from 44.0 percent, on average, in 1990 to 33.2 per- cent in 2005. The rate of 33.2 percent might compare favorably (only barely) with the weighted average rate noted above, but it does not compare well with any other. The reality is that, on the basis of simple weighted averages, Africa has high corporate tax rates on pro�ts, at least 5 percent above the KPMG simple weighted average for the same year and much higher than the weighted average, including secrecy jurisdictions, in 2008. Accounting for the Missing Billions 279 Figure 9.3. Corporate Income Tax, Africa, 1980–2005 10 50 9 45 8 40 7 35 6 30 % of GDP percent 5 25 4 20 3 15 2 10 1 5 0 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 CIT Revenue (left axis) CIT Base (left axis) Statutory CIT rate (right axis) Source: Author compilation based on Keen and Mansour (2009) and data of International Bureau for Fiscal Documentation and International Monetary Fund. Note: CIT = corporate income tax. Revenue excludes oil, gas, and mining companies. These �ndings are important. It is likely that the incentive to avoid taxes through transfer pricing is based on two considerations. The �rst is the differential between the tax rates of the countries through which ownership of the goods might pass as part of the transfer pricing supply chain. The greater the differential in tax rates across jurisdictions in the intragroup supply chain, it may be suggested, the greater the prospect that a group of companies will pro�t from transfer mispricing. The sec- ond consideration likely to feature signi�cantly in the decision process on mispricing is the chance that the activity may take place without dis- covery within the intragroup supply chain. The �rst issue is considered here; the second, in later parts of this chapter. The literature that �nds evidence for substantial IFFs relating to transfer mispricing also �nds evidence that this process is facilitated by the existence of secrecy jurisdictions, most or all of which offer no taxa- tion on foreign source income (Baker 2005; Christian Aid 2008; and so on). As the data noted above show, this 0 rate tax offering has, over time, had a slightly diminishing impact as simple weighted average tax rates have fallen, but, with the rates still hovering at around 27 percent in many of the measures noted, the differential remains large. 280 Draining Development? The differential is even more signi�cant if comparison is made with the simple weighted average tax rates in Africa. The incentive to tax avoid through transfer mispricing from that continent would appear to be strong, a theme to which we return below. What does need to be assessed is whether the incentive based on tax rate differentials is suf�cient to justify the claimed US$160 billion or more of transfer mispricing that is said to take place each year. Analytical review techniques such as are found in auditing are used here with the objective of testing whether the claimed audit outcome (in this case, that transfer mispricing results in losses to developing countries of US$160 billion a year) is within the plausible range of data outcomes that popu- lation information implies may be likely. It is proposed that, in analytical review terms, the estimates of trans- fer mispricing would be considered plausible if the total estimated amount of tax lost to transfer mispricing were materially less than the amount of tax lost as measured by the differential between weighted average headline tax rates (which would indicate tax due if corporations were tax compliant) and the tax likely actually to be paid calculated using likely cash paid tax rates, as noted above. In this context, tax compliance is de�ned as an effort to pay the right amount of tax (but no more) in the right place at the right time, where right means that the economic sub- stance of the transactions undertaken coincides with the place and form in which they are reported for taxation purposes. Data on tax rates have already been developed above. The other data needed are as follows: • A measure of world GDP, because the trade that is transfer mispriced is a proportion of world GDP, and that proportion must be credible • An indication of corporate pro�ts as a proportion of world GDP, because, ultimately, it is pro�ts that are shifted as a result of transfer mispricing These data, if combined, will generate the following: • A measure of worldwide corporate taxable pro�ts • If weighted by the tax rate data noted above, the measure of world- wide corporate taxable pro�ts will give measures of both the likely tax paid and the tax lost because tax rates weighted by GDP are not paid by major corporations, as noted above Accounting for the Missing Billions 281 Worldwide GDP was approximately US$61 trillion in 2008.12 The IFFs that give rise to the estimated tax loss calculated by Christian Aid (2008) amount to between US$850 billion and US$1 trillion a year, rep- resenting, in some cases, sums as high as 10 percent of GDP (Kar and Cartwright-Smith 2008). These calculations are based on actual trade data. Illicit flows of this scale are considered plausible for the purposes of this chapter. As Murphy (2010a) notes, the U.K. tax authorities estimate that 15 percent of total value added tax due on U.K. gross sales and imports in 2008–09 was not collected (HMRC 2010). The part not collected was associated with what might reasonably be described as IFFs using the de�nitions offered by Kar and Cartwright-Smith (2008). By extrapola- tion, the total tax loss over the entire U.K. economy was at least 5 percent of GDP in that case. The United Kingdom is considered a well-regulated economy with low levels of tax evasion, although the domestic tax authority acknowledges that one-sixth of total gross commercial �nan- cial flows are outside the tax system. It is widely accepted that a signi�- cant amount of lost value added tax arises as a result of international unrecorded trading flows and illicitly recorded transactions, a matter that has been of considerable concern in the United Kingdom and the rest of the EU for many years. In this circumstance, the estimates of IFFs in developing countries that Kar and Cartwright-Smith (2008) provide appear to be well within the plausible data range based on comparison with alternative reliable information. Total corporate added value in the U.K. economy represented by the gross operating surplus of corporations is in excess of 20 percent of GDP.13 In 2008, this gross operating surplus of corporations amounted to £339 billion. Yet, this sum is stated before charges such as interest are offset against pro�t. Total corporation tax revenues in the same year were £46 billion.14 On gross value added, this implies an effective tax rate of 13.6 percent. However, Murphy (2008) states that an effective average tax rate of about 22 percent should be anticipated. If this latter rate were applied to tax paid, the gross added value after the interest offset is reduced to £210 billion, a sum that was 14.5 percent of U.K. GDP in 2008. Data on the same ratio in the United States for 2006, 2007, and 2008 imply that corporate pro�ts were lower, less than 11 percent of GDP and on a downward trend (12.0 percent in 2006, 10.9 percent in 2007, and 282 Draining Development? 9.4 percent in 2008).15 Nonetheless, this excludes the pro�ts of private companies (which are included in the U.K. data). In this case, one may presume that the two ratios are largely consistent, and a total pro�t rate of approximately 10 percent of GDP might be considered to be associ- ated with large entities likely to undertake multinational trade. This is probably a reasonable estimate of global corporate pro�ts given that the United Kingdom and the United States are the two largest centers for the location of MNCs in the world. Assuming that this rate may be indica- tive of worldwide rates, this leads one to suspect that, based on world- wide GDP of US$61 trillion, corporate pro�ts may be conservatively estimated at a total of approximately US$6 trillion in 2008. We have used secondary sources to check this conclusion. Global data on pro�ts are remarkably dif�cult to source. However, McKinsey has published a review of the after-tax pro�ts of the top 2,000 companies in the world in 2006 and established that the after-tax earnings of this group in that year were US$3.2 trillion (Dietz, Reibestein, and Walter 2008). Murphy (2008) �nds that declared tax rates for companies regis- tered in the United Kingdom and quoted on the stock exchange for that year averaged 25.8 percent (this rate reflecting their worldwide rates and not merely the rate applicable in the United Kingdom), which suggests that the worldwide pretax pro�ts of the McKinsey sample may be US$4.3 trillion. This �gure is, of course, somewhat lower than US$6 trillion, but a sample of 2,000 companies is also somewhat lower than the total pop- ulation of MNCs. There are, for example, more than 2,200 companies quoted in the United Kingdom alone (London Stock Exchange 2010). That 2,000 companies might represent 72 percent of total global pro�ts likely to be of concern does seem plausible, however, and this is consid- ered strong supporting evidence that the estimates of global pro�ts are reasonable. On this global pro�t estimate of US$6 trillion, the tax due at weighted average tax rates based on GDP noted above (32.1 percent) might be about US$1.9 trillion. If secrecy jurisdictions and ring-fences are consid- ered, this might fall to a tax �gure as low as US$1.2 trillion at a rate of 20.1 percent, as noted above. The rates are selected from among those noted above because, �rst, it is suggested that the average tax rate weighted by GDP indicates the taxes that would be due if companies Accounting for the Missing Billions 283 were tax compliant, that is, if they paid the right amount of tax (but no more) in the right place at the right time, where right means that the economic substance of the transactions undertaken coincides with the place and form in which they are reported for taxation purposes and that pro�ts arise where third-party sales (for which GDP is a proxy) occur. Second, the rate, including secrecy jurisdictions and ring-fences, reflects the reality of the tax system that, this chapter notes, MNCs can actually exploit, that is, the rate is used in recognition of the fact that secrecy does allow the relocation of pro�ts and that ring-fences for foreign earnings might be used to apply low tax rates to pro�ts reallocated by transfer mispricing techniques behind the veil of secrecy. It is plausible that the differential in taxes paid, based on the varying assumptions, of some US$700 billion might include transfer mispricing effects of US$160 billion that have been claimed (Christian Aid 2008). Transfer mispricing abuse would represent 22.8 percent of the tax defer- ral on this basis, and one may note that, because some of this deferral may be represented as deferred tax provisions in the accounts of MNCs, the actual impact on the tax charge in the accounts of MNCs may be lower than this implies without diminishing the cash loss to developing countries. A ratio of this proportion allows ample margin for numerous other factors that may reduce declared tax rates, including advanced capital allowances (by far the biggest claimed factor in tax deferral noted in Murphy 2008), nonremittance of pro�ts (perhaps transfer mispricing induced), tax holidays, and so on. If private company pro�ts were to be included in the pro�t base on which the calculation has been under- taken, the margin for these other factors would obviously be higher still. What is clear is that these data suggest that US$160 billion in transfer pricing abuse affecting developing counties is plausible within the frame- work of the world economy given what we know of corporate pro�ts, corporate tax rates, and the opportunities for corporate tax planning. The risk of being caught If it is plausible that transfer mispricing of the suggested order may have taken place, we must then ask if the transfer mispricing could have taken place without detection or sanction. 284 Draining Development? First, this would not happen if all corporations sought to be tax com- pliant (see above). While some corporations are risk averse, by no means all are.16 In this case, it is unlikely that all corporations are tax compliant. The likelihood of noncompliance is increased because of the limited application of the OECD arm’s-length pricing rules, which are meant to govern transfer pricing in international trade (see above). For the pur- poses of this chapter, transfer mispricing is considered a breach of these rules. Because these rules usually only apply if legislation requires or if a DTA is in existence requiring the trade between two jurisdictions to be priced in accordance with these principles, it is likely that transfer mis- pricing is commonplace. In EU and OECD countries, which have been used as the basis for much of the published research on transfer mispricing, such DTAs usu- ally exist, as do the resources to monitor the application of these agree- ments. As a result, it is now often said at tax conferences that mispricing in the trade in goods is rare or almost unknown, although the same is not said of intangibles. Such comments, however, ignore the fact that this is a select sample base that gives little indication of the opportunity for abuse in much of the world. For example, even a brief review shows clearly that DTAs are notable by their absence in Africa. While South Africa has an impressive range of DTAs, other countries are in a different position.17 Botswana has 10.18 Zambia has 17, but most are old, and none has been signed since 1989.19 The Democratic Republic of Congo has only two DTAs.20 Angola is far from alone in having no DTAs. The lack of progress in developing new agreements implies that the resources devoted to monitoring the issue are limited in Africa. Especially since the G-20 summit in April 2009, DTAs have been sup- plemented by new tax information exchange agreements. These, how- ever, are limited to information exchange issues, as their name implies, and not the regulation of transfer prices (in other words, they exclude standard article 9 of the OECD Model Tax Convention; see OECD 2003). Moreover, as research by Tax Research LLP and the Tax Justice Network (TJN) has shown, as of November 2009, Brazil, China, India, Japan, most of Africa, and almost all developing countries were notable absentees from the list of states that had signed tax information exchange agree- ments with secrecy jurisdictions (Murphy 2009b). The implication is Accounting for the Missing Billions 285 obvious: the places most likely to be subject to transfer mispricing abuse are also the places least likely to enjoy protection from such abuse. It is accepted that other regulation, such as controlled foreign com- pany rules, might limit the opportunity for such abuse, but, as secondary protection, they do not do so ef�ciently. First, they do not restore correct pricing between the parties that initiated the trade; so, tax remains inap- propriately allocated to jurisdictions given that the application of these rules to transfer mispricing only gives rise to an additional tax payment in the jurisdiction in which the ultimate parent company is located, not in the jurisdiction that lost out initially. Second, the abuse has to be dis- covered in the parent company jurisdiction. For the reasons noted below, this can be dif�cult. Consequently, the chance that transfer mispricing will take place without being detected is high. The Particular Problem in Developing Countries The issue of enforcing transfer pricing rules in developing countries is particularly acute, as many published reports have shown.21 Global Witness has published a report on the operations of Mittal Steel (now Arcelor Mittal) in Liberia. The report provides commentary on the tax provisions of Mittal’s mineral development agreement, noting that “probably the single biggest problem with this agreement is that it gives the company [Mittal] complete freedom to set the price of the iron ore, and therefore the basis of the royalty rate� (Global Witness 2006, 7). There were no restrictions at all in the original agreement between Mittal Steel and Liberia on the transfer prices the company could use. As a result, while there is no suggestion of impropriety, the possibility that transfer mispricing occurred was increased by the absence of any regula- tion intended to prevent it. In this case and as a direct result of the work of Global Witness, the contract was revised. The changes were noted in a new commentary issued by Global Witness, which reported that “under the amended agreement the [transfer] price is set under the arms length rule, which means that it will be based on the international market price of the ore� (Global Witness 2007, 1). It would be pleasing to report that all such risks of transfer mispricing have been eliminated so speedily, but the evidence is clear that this is not the case. Problems with transfer mispricing have been found after similar 286 Draining Development? NGO studies of the extractive industries in many countries. For example, in Zambia, Christian Aid has stated that “in his budget speech in Febru- ary 2006, the minister of �nance estimated that the government was likely to receive less than US$11 million from royalty payments in 2006: that’s 0.1 per cent of the value of production in 2005� (Christian Aid 2007, 24). Christian Aid believes that this is in no small part caused by transfer mispricing, which has an impact, in this case, on both royalty payments and declared taxable pro�ts. This is unsurprising. The Invest- ment Act 1993 of Zambia, like its predecessor, the 1991 Investment Act, does not address the issue of transfer pricing (Mwenda 1999). Nor, it seems, do many of the mineral development agreements that have been negotiated in Zambia. This is a situation that may have been addressed by amendments in the Zambian Income Tax Act, passed by parliament in April 2008, which stipulated that royalties are to be calculated based on the average monthly cash price on the London Metal Exchange, Metal Bulletin, or any other metals exchange as agreed with the government (Open Society Institute et al. 2009). The impact may be limited, however: most Zambian mineral development agreements have stability clauses exempting them from the effects of any changes in tax law for up to 20 years (Christian Aid 2007). In the logging sector in the Democratic Republic of Congo, Green- peace notes as follows: Internal Danzer Group documents show in great detail the price �xing arrangements between the Group’s Swiss-based trading arm Interholco AG and the parent �rm’s logging subsidiaries in the DRC and the Repub- lic of the Congo. The DRC-based Siforco sells its wood to Interholco at an of�cial price below the true market value of the wood. The shortfall is made up through unof�cial payments into offshore bank accounts in Europe. (Greenpeace 2008, 3) A review undertaken for this report found no evidence that issues related to transfer pricing were addressed in �ve mineral development agreements signed from 1994 to 2007 between the government of Tan- zania and companies mining gold in Tanzania. Royalty rates were funda- mental to anticipated government revenues from royalties and, ulti- mately, from pro�ts in each case, but on no occasion was the basis speci�ed for setting the price of exports. The tax base on which royalties Accounting for the Missing Billions 287 were to be charged was thus capable of discretionary determination by the company liable to make the payment of the taxes due. In addition, all the agreements include �scal stability clauses, and most specify that the basis of the pricing of gold (even though unspeci�ed) should not be unilaterally changed, presumably by the government of Tanzania. For this reason, recent changes to Tanzanian law introducing transfer pricing regulation are likely to have little or no impact in this critical Tanzanian export sector. The problem has also been found within the Extractive Industries Transparency Initiative. In a review of the �rst audit of the initiative in Ghana, Murphy (2007, 10) notes that the audit objective to “ascertain the appropriateness of payments made with regards to mineral royalties; ground rent; dividends; taxation on pro�ts and for mineral rights� had not, in the opinion of the reviewer, been ful�lled, largely because of the use of indeterminate prices unrelated to veri�able benchmarks and the use of apparently uncorroborated exchange rates for valuing gold exports, both clear indications that proper transfer pricing controls were not in operation. The evidence appears to be telling: there is a pattern of transfer pric- ing abuse or at least the risk of such abuse in developing countries. The evidence from mineral development agreements implies there is no change in the prospects in this area even if legislation to introduce arm’s- length pricing rules is enacted because companies in the extractive industries are almost entirely immune to legislative changes affecting the way in which their tax liabilities are computed for periods of up to 30 years after signing mineral development agreements. It is also important to note another key feature emphasized by this work: the transfer pricing abuse in these cases is highly unlikely to be motivated by taxes on corporate pro�ts alone. The abuse is likely to extend to royalties, sales and purchase taxes, dividends, abuse of pro�t- sharing agreements, and more. The incentives to abuse are high, indeed; the consequences of not tackling the issue are considerable; and the prospects for tackling the abuse within current legislative and contrac- tual constraints are not good. Even if arm’s-length transfer pricing rules do exist in developing countries, there appear to be almost insurmountable problems in enforc- ing them. As one of the rare cases of suggested transfer pricing abuse 288 Draining Development? ever brought to court in Africa has shown (Unilever Kenya Ltd v. Com- missioners of Income Tax, Kenya Income Tax Appeal 753, 2003), the absence of the accounts of the related party with whom trade was being undertaken in the destination jurisdiction is a signi�cant cause of the failure to prove that pro�t was being shifted through the transfer mis- pricing of goods; this is what the Kenyan authorities were seeking to prove had occurred using the OECD arm’s-length principle. Although the accounts in question were necessarily available to the group of com- panies the Kenyan subsidiary of which made the appeal in this case, they were not made available to the court. It is likely that the withholding of this accounting data, albeit entirely legal, had a material impact on the resulting decision of the court in Kenya. Secrecy prevented the proper determination of a transfer pricing issue in this case, whether rightly or wrongly. This is a recurring theme of work in this area, as is the persistence of the assertion that developing countries are particularly vulnerable to the effects of secrecy. If this is the case, it is important that the mechanisms for creating this secrecy that permits transfer mispricing to take place undetected, unchallenged, or uncorrected be considered. Unless it can be shown that corporations can make use of secrecy to achieve this outcome, then it remains implausible that transfer mispricing of the alleged scale takes place. If, in contrast, signi�cant secrecy is available to corporations, then corporations have the opportunity to transfer misprice, as some believe is taking place. How Multinational Corporations Exploit Secrecy The modern MNC is a complex entity. This is not the place to explore all aspects of the nature of the MNC or the motivations for creating some of the structures MNCs use, but, without consideration of the interac- tion of the corporation, jurisdictions, corporate law, and tax law, testing the proposition that transfer mispricing can take place within MNCs and be hidden from view within the accounts and �nancial statements of MNCs is not possible. The MNC is almost invariably headed by a single company, the parent entity, which is almost always a limited liability corporation. The parent entity comprises a number of other, usually similar limited liability cor- porations spread over one or more other jurisdictions. For example, in Accounting for the Missing Billions 289 the course of researching this report, we noted that United Kingdom– based BP plc has more than 3,000 subsidiaries in over 150 jurisdictions. Subsidiaries need not be limited liability corporations. As noted in International Accounting Standard 27 (IAS 27, 2009), a subsidiary is an entity, including an unincorporated entity such as a partnership, that is controlled by another entity (known as the parent).22 Subsidiaries can therefore also be limited liability partnerships in any form, whether trusts, charities, or other arrangements, but limited liability corpora- tions are the most common by far. The key is that the parent company controls the subsidiary. Control is widely de�ned by IAS 27, but most commonly means that the parent has direct or indirect control of a majority of the equity shares. However control is established, if a parent entity governed by the IFRSs (which, in this respect, operate almost identically to the U.S. Gen- erally Accepted Accounting Principles) has subsidiary entities, then IAS 27 requires that consolidated accounts and �nancial statements be pre- pared. These are the �nancial statements of a group of entities presented as if they were those of a single economic entity. It is immediately apparent that, within this requirement, there is an obvious conflict of interest. The parent entity of such a group may be required to present the group’s accounts as if the group were a single entity, and, yet, in practice, the group may be made up of thousands of entities that are under the control, but not necessarily the sole owner- ship, of the parent entity. Thus, in substance, each entity within the MNC may remain legally distinct, and each may be subject to the rules of accounting, taxation, and disclosure of the jurisdiction in which it operates. In a real sense, each subsidiary is therefore without obligation to the other members of the group, barring the duty the directors and managers of the entity may owe to the owners to whom they report under the law of the jurisdiction in which the entity is incorporated, and this duty varies widely from place to place. Curiously, according to the one nearly constant assumption in com- pany law, the shareholders and the owner do not have the right to man- age the entity: that right belongs to the directors. Of course, the share- holders may have the right to appoint the directors, but, it is important to note, the assumption underpinning group accounts pierces the veil of incorporation that, in turn, underpins the notion of the limited liability 290 Draining Development? entity. The dichotomy is that the assumption accomplishes this in the effort to reinforce the division inherent in the act of incorporation through the presentation of the group as one undertaking distinct from the owners, who obtain only the limited information the directors may wish to supply, subject to legal constraints. The inherent conflict in reporting results is exacerbated by a number of other factors. First, the de�nition of control used for accounting may be different from the de�nition used by some jurisdictions for tax. So, some entities that are within the group for tax purposes in some loca- tions may not be within the group in other locations. For example, tax may require 75 percent control, while accounting requires 50.01 percent in most cases. Therefore, entities that, for accounting purposes, may be related parties requiring inclusion in a common set of consolidated accounts and �nancial statements may not be so treated for transfer pricing purposes. Second, note that some entities are deliberately structured to exploit the rules on consolidated �nancial reporting. In particular, the �nancial services industry has become expert at creating orphan entities. These are companies that are created by a parent organization and that are deliberately structured by the parent entity so that they are off the bal- ance sheet; thus, the assets or liabilities that the orphans own are excluded from consolidation in the parent entity’s accounts and �nancial state- ments, as are the results of the trading of the orphans. A common way to engineer this outcome is to create a company to which are transferred the off–balance sheet assets and liabilities the par- ent entity wishes to hide from view. The new entity is owned by a chari- table trust, for example. As such, it is not considered to be under the ownership or control of the parent entity. This is why it is described as an orphan; it has become parentless, although it is utterly dependent on the parent entity. These entities are hard to spot, but commonplace. While the entities are used to exclude liabilities from accounts, the rules that permit this will, in most cases, also allow them to be used for transfer mispricing, which may pass undetected, subject to the caveat that the proceeds must then be used for purposes that the group may wish to keep at quasi- arm’s length. This purpose may be fraudulent. Accounting for the Missing Billions 291 Third, it is widely assumed that consolidated accounts and �nancial statements are created by adding together all the accounts of the indi- vidual entities that make up the group and then eliminating all the intra- group transactions and balances. This is a simplistic, but not wholly inappropriate view of what should happen. The reality is that MNCs can deliberately obscure the relationship between the underlying accounts and �nancial statements of the subsidiary companies and the group accounts in ways that make it almost impossible to detect what is really happening within the group. The �rst method to achieve this is the use of different accounting year-end dates for group companies. IAS 27 (section 26) says this should not occur and that any differences should be explained, but the reality is (as the author has frequently witnessed) that noncoterminous year-ends are commonplace and almost never disclosed or commented upon. If noncoterminous year-ends are used, transfer mispricing may then be relied on to shift pro�ts (and losses) around the group almost at liberty and almost entirely undetected. Next, nonstandard accounting policies may be used in different places to recognize transactions even though, according to IAS 27, this should not occur. This is now commonplace. The parent company might account using IFRSs, but local entities may well account using local Gen- erally Accepted Accounting Principles, and there is nothing to prevent this. Some signi�cant transactions have different tax treatments depend- ing on the accounting standards used. There are, for example, conflicts between the United Kingdom and the International Accounting Stan- dards on �nancial derivatives for tax purposes. These differences and conflicts have been exploited by international banks. In addition, entries can be made in the consolidated accounts alone, without ever appearing anywhere in any of the accounts of the underly- ing entities. In principle, this should not occur because the accounts can then be said not to reflect the underlying books and records of the MNC that is publishing them, but, in practice, if the entries are considered a nonmaterial adjustment in the assessment from the point of view of the user of the �nancial statements, which both the International Account- ing Standards Board and the U.S. Federal Accounting Standards Board de�ne as “a provider of capital to the company,� then no auditor is likely 292 Draining Development? to object. This can, however, disguise radically different presentations of pro�t on trades between related undertakings in group consolidated accounts and �nancial statements and individual entity accounts and �nancial statements, especially if the tax implications are considered, so that the bene�t may be hidden from view in the accounts of the group as a whole. Interview-based evidence indicates that this practice may be commonplace. It will never be discovered by tax authorities because it is the accounts and �nancial statements of the individual subsidiary enti- ties that are used to determine tax liabilities in each jurisdiction in which these entities trade. The consolidated accounts and �nancial statements are deemed to have no tax interest to tax authorities (although it is not clear that this is true), and, as such, in jurisdictions such as the United Kingdom, the tax authorities are not entitled to ask questions about the entries that make up these published accounts. This last point is, perhaps, of the greatest signi�cance because the exact entries that are eliminated from view when the consolidated accounts and �nancial statements are prepared are those same transac- tions that will always have the potential to give rise to transfer pricing disputes. For this reason, the most useful evidence that consolidated accounts and �nancial statements could provide to tax authorities—the data relating to transfer pricing issues, which are the data on the “most contentious issue in tax,� according to a poll of U.S. tax directors in 2007—is denied to the tax authorities who need it.23 This omission is exacerbated by a number of other practices, all endorsed by the IFRSs and all of which make it easier to hide transfer pricing abuse. First, in the individual accounts and �nancial statements of the subsidiaries that make up an MNC, it would seem obvious that the transactions with other group companies should be highlighted if only to indicate that they will disappear upon consolidation. This is theoretically required by another International Accounting Standard, IAS 24, on related-party transactions. Broadly speaking, IAS 24 (section 9) de�nes a party as related to an entity if one party directly or indirectly controls the other, but associates, joint venturers, and some other arrangements are also included in the de�nition. IAS 24 de�nes a related-party transaction as a transfer of resources, services, or obligations between the related par- ties regardless of whether a price is charged (section 9). As a result, it would seem that all transactions among group companies must be dis- Accounting for the Missing Billions 293 closed in a group’s �nancial statements because a long list of disclosable transactions of this sort is included in the standard. Unfortunately, the prospect of disclosure that this requirement cre- ates is then dashed. IAS 24 proceeds to state that, while the disclosure must be made separately for the parent company, subsidiaries, and other identi�ed categories of related parties, the information within each such category “may be disclosed in aggregate except when separate disclosure is necessary for an understanding of the effects of the related party trans- actions on the �nancial statements of the company.� As a result, all trad- ing by a subsidiary with all other subsidiaries can be aggregated into one number in most cases, and no indication need be given of the other party in a trade, what has been traded, or on what terms and where the other side of the transaction might be recorded. The result is obvious: the accounts and �nancial statements end up providing no meaningful information at all on transfer pricing issues. The information is excluded from consolidated accounts and �nancial statements because related-party trades between parents and their sub- sidiaries and between fellow subsidiaries are always excluded from these accounts, while the disclosure requirement on individual group mem- bers is so limited that forming a view on transfer pricing is almost impossible in most cases: it is rare for the other party to any transaction to be disclosed, especially within a large and complex group. This might be thought an accident. Regrettably it is not, as IAS 24 makes clear. In the latest version of the standard, introductory note IN7 states that “discussions [in the standard] on the pricing of transactions and related disclosures between related parties have been removed because the Standard does not apply to the measurement of related party transactions.� This is an extraordinary statement. Accounts prepared under IFRSs and their U.S. equivalents are the basis of corporate taxation in a great many countries in the world, but the International Accounting Standard that is responsible for their promulgation says that it is not the purpose of the standard to assist in the measurement of matters related to trans- fer pricing. Moreover, it offers no suggestion on how such matters should be considered or measured. What is clear is that it is not the intent in the standards that the only other part of the IFRS environment that might provide information on 294 Draining Development? the issue—IFRS 8, on segment—do so. IFRS 8, by default, usually only applies to companies quoted on a stock exchange in a jurisdiction that has adopted IFRSs (as all countries in the EU have done, for example). It de�nes an operating segment as a component of an entity: • that engages in business activities from which it may earn revenues and incur expenses; • the operating results of which are reviewed regularly by the entity’s chief operating decision maker so as to make decisions about the resources to be allocated to the segment and assess its performance; and • for which discrete �nancial information is available. IFRS 8 requires an entity to report �nancial and descriptive informa- tion about the entity’s reportable segments. These are operating segments or aggregations of operating segments that account for more than 10 per- cent of the revenues, pro�ts, or assets of the entity. Smaller segments are combined until ones of this size are created, supposedly to reduce infor- mation overload. In practice, this might, of course, hide necessary detail. Required disclosure by reportable segments targets trading data, including pro�t and loss, assets and liabilities, and limited geographical analysis. Such a summary does not, however, show the true level of prob- lems inherent within IFRS 8, which also allows segment data to be pub- lished using accounting rules that are not the same as those used in the rest of the accounts and �nancial statements, meaning, as a result, that segment data may be formulated in a way harmful to the appraisal of transfer mispricing. In addition, IFRS 8 does not require that segments cover all of the MNC’s activities, meaning some information may be omitted, providing more opportunity for transfer mispricing to be hid- den from view. As a consequence, considerable support has developed for an alter- native form of segment accounting called country-by-country report- ing, created by the author of this report (Murphy 2009a). Country-by- country reporting would require an MNC to disclose the name of each country in which it operates and the names of all its companies trading in each country in which it operates. Currently, these data are usually unavailable. Country-by-country reporting would then require publi- Accounting for the Missing Billions 295 cation of a full pro�t and loss account for each country in which the MNC operates, plus limited cash flow and balance sheet information. Radically, the pro�t and loss account would break down turnover between turnover involving third parties and turnover involving group entities. Costs of sale, overhead, and �nance costs would have to be bro- ken down in the same way, while a full tax note would be required for each country, as is presently necessary for IFRSs. In addition, if the company operates within the extractive industries, one would also expect to see all those bene�ts paid to the government of each country in which the MNC operates broken down across the cate- gories of reporting required in the Extractive Industries Transparency Initiative. As Murphy (2009a, 18) notes, “country-by-country reporting does not [stop transfer mispricing]. What it does do is provide data that . . . tax departments . . . can use to assess the likely risk that exists within the accounts of a multinational corporation. They can do this by • “Assessing the likelihood of risk within the group structure; • “Reviewing the overall allocation of pro�ts to countries within the group to see if there is indication of systematic bias towards low-tax jurisdictions; • “Assessing whether the volume and flows of intragroup trading dis- closed by country-by-country reporting suggests that this outcome is achieved as a result of mispricing of that trade; • “Using that information to assess where that abuse is most likely to occur so that an appropriate challenge can be raised.� So far, the International Accounting Standards Board has only indi- cated willingness to consider this issue with regard to the extractive industries, and current indications are that, despite the considerable lob- bying, there is little prospect of an advance on this issue. The conclusion is inescapable: as one board member said when the issue of country-by- country reporting was being discussed by the International Accounting Standards Board, “this looks like it deals with the issue of transfer pric- ing, and we do not want to go there.�24 The comment is succinct and neatly summarizes the design of current accounting standards, which seem purpose-made to hide the subject of transfer pricing from view. 296 Draining Development? The Role of Secrecy Jurisdictions The literature that alleges substantial transfer mispricing abuse by MNCs also �nds that tax havens play a signi�cant role in the process. The term tax haven is, however, so widely misunderstood that this chapter does not use it, preferring instead to use the term secrecy jurisdiction. For a more detailed consideration of the term, the nature of a secrecy jurisdic- tion, and the economic signi�cance in the matters under consideration here, see chapter 11 in this volume. A list of the places currently consid- ered signi�cant secrecy jurisdictions is available in table 9A.4.25 The term secrecy jurisdiction is considered more appropriate for the purposes of the current analysis because, although the process of trans- fer mispricing to which this chapter refers seeks to secure a tax advantage (by way of reduced tax payment) for those who pursue the activity, this advantage is not normally available unless the abuse giving rise to the advantage—the arti�cial relocation of activities to one or more secrecy jurisdictions—can be hidden from view behind a veil of secrecy. Secrecy jurisdiction opacity is often, of course, linked to low tax rates (see above). The combination of low tax rates and secrecy has obvious attractions for those seeking to transfer misprice. However, to demonstrate whether or not MNCs actually use secrecy jurisdictions and which ones they might use if they do, TJN has coordinated, under the direction of the author of this chapter, a survey of the locations of MNC subsidiaries, paying particular attention to the secrecy jurisdictions TJN has identi- �ed. The results of the study by the U.S. Government Accountability Of�ce of December 2008 (GAO 2008), “Large U.S. Corporations and Federal Contractors with Subsidiaries in Jurisdictions Listed as Tax Havens or Financial Privacy Jurisdictions,� have been used as part of the survey. Because the U.S. survey excludes data on the Netherlands, the United Kingdom, and the United States, these locations have also been excluded from the TJN survey. Austria (for practical rather than meth- odological reasons), Belgium, and Madeira (because of dif�culties in isolating data independently from Portugal) have also been excluded. The total sample of MNCs surveyed is shown in table 9.4. It should, however, be noted that the data selection has been prag- matic: the U.K. data should have been the entire FTSE 100, that is, the 100 largest companies in the United Kingdom, designed to match the U.S. Accounting for the Missing Billions 297 Table 9.4. Multinational Corporations Surveyed by the Tax Justice Network number Country MNCs sampled France 39 Germany 28 Netherlands 23 Switzerland 20 United Kingdom 33 United States 100 Total 243 Source: Mapping Financial Secrecy (database), Tax Justice Network, London, http://www.secrecyjurisdictions.com/index.php (accessed December 12, 2009). sample. In practice, although all United Kingdom–quoted companies are legally required to publish the names, places of incorporation, and per- centage of the holdings for all their subsidiary companies annually either in their audited accounts and �nancial statements or as an appendix to their annual declaration made to the U.K. Registrar of Companies, only 33 of the FTSE 100 companies did so. Enquiries found that no company had ever been prosecuted for failing to �le this information. It is a curi- ous example of the United Kingdom’s opacity (see table 9A.5).26 It should also be noted that substantial problems were encountered with all other samples. The French data undoubtedly underreport the number of subsidiaries because they only relate to principal subsidiaries, not all subsidiaries. German companies do not always make the distinc- tion between subsidiaries and associates clear. The Dutch and Swiss data have been taken from databases, not original documentation, which implies that there are inconsistencies in approach, particularly about whether dormant subsidiaries are counted or not, and so on. All such issues do, however, reveal one consistent theme: it is immensely dif�cult to determine the composition of an MNC. Detailed analysis of the regulatory requirements of the 60 secrecy jurisdictions surveyed by TJN highlights the issues. Of the 60 jurisdic- tions surveyed, accounts of companies were available on easily accessible public record in only six.27 298 Draining Development? Table 9.5. Secrecy Jurisdiction Locations of Multinational Corporation Subsidiaries number Rank Secrecy jurisdiction MNC subsidiaries 1 Cayman Islands 1, 130 2 Ireland 920 3 Luxembourg 824 4 Switzerland 771 5 Hong Kong SAR, China 737 6 Singapore 661 7 Bermuda 483 8 Jersey 414 9 Hungary 252 10 British Virgin Islands 244 11 Malaysia (Labuan) 177 12 Mauritius 169 13 Bahamas, The 156 14 Guernsey 151 15 Philippines 126 16 Panama 125 17 Isle of Man 99 18 Costa Rica 85 19 Cyprus 69 20 Netherlands Antilles 68 21 Uruguay 67 22 Malta 60 23 United Arab Emirates (Dubai) 58 24 Israel 56 25 Gibraltar 54 26 Barbados 51 27 Latvia 40 28 U.S. Virgin Islands 37 29 Monaco 35 30 Liechtenstein 32 Source: Mapping Financial Secrecy (database), Tax Justice Network, London, http://www.secrecy jurisdictions.com/index.php (accessed December 12, 2009). Accounting for the Missing Billions 299 The situation was worse for the bene�cial (as opposed to nominal) ownership information on public record. Only Monaco requires that these data be available. In all other cases, nominee ownership may be recorded, or there is simply no requirement to record data on public record at all.28 It is readily apparent, as a consequence, that, unless data are required from MNCs on the companies making up or not making up their group and the operations of each, as shown by the audited accounts, then the current legal requirements for data registration within secrecy jurisdic- tions ensure that the information required to assist in the appraisal of MNC activities, including those relating to transfer mispricing, will sim- ply be unavailable to most people and, quite possibly, to many tax authorities if that is the MNC’s wish, as it will be if the MNC is seeking to hide transfer mispricing activity. This would not be an issue if MNCs did not use secrecy jurisdictions. The reality is that they do use secrecy jurisdictions extensively. Of the European and U.S. samples, 97.2 and 84.0 percent, respectively, had secrecy jurisdiction subsidiaries as de�ned by the TJN. Table 9.5 indi- cates the number of subsidiaries by location according to the TJN survey. (Data on the 24 additional, smaller jurisdictions have been ignored; these jurisdictions are immaterial for the purposes of this chapter.) It is readily apparent that some locations stand out, but the data make a lot more sense if they are plotted against two control variables: popula- tion and GDP (�gure 9.4). The data in �gure 9.4 are ranked by the number of subsidiaries by GDP in billions of U.S. dollars. In most cases, the correlation with a ranking by the number of subsidiaries per head of population is clear. These weighted data give a much better view of the relative importance of these secrecy locations. It is apparent that some show extraordinary amounts of activity relative to their size. There is only one explanation for this: the secrecy jurisdictions are not creating entities for use by the local population, but, as the de�nition of these jurisdictions in this chap- ter suggests is likely, for the use of people resident elsewhere. The com- panies that are registered in these jurisdictions do little or nothing in these locations. 300 Draining Development? Figure 9.4. Top 20 Subsidiaries by Secrecy Jurisdiction, Population, and GDP 700 Subsidies per 1,000 people 600 Subsidies per GDP (US$ billions) 500 400 300 200 100 0.0 in nds Na s all uru Be nds a er y Gi sey of r s Ca Mo n Ne C s Isl o la Is s s m s An rg Ba illa Lib s au a Ba tius os e ta d th oo and U. nds land rg ille xe nd a M eri ud Gu rse a c u m an ad M ico na gu Isl al bo n Vi Isla a Lu Isla ri rm Vi Ant ha Je br Isl Isl rb sh an er k in rg sh ym ar Ca S. M s& iti rk Br Tu Source: Mapping Financial Secrecy (database), Tax Justice Network, London, http://www.secrecyjurisdictions.com/index. php (accessed December 12, 2009). That this must be true is indicated by the number of subsidiaries active in �nancial services as a proportion of the total working population in the secrecy jurisdictions. As TJN has noted, this exceeds 20 percent in Cayman Islands, Guernsey, Isle of Man, and Jersey and 10 percent in Ber- muda, Liechtenstein, and Luxembourg.29 Financial service sectors of this size crowd out the possibility of any other signi�cant economic activity taking place. The overlap between this list (�gure 9.4) and the locations with the most MNC subsidiaries (table 9.5) is obvious, and the implica- tion is clear: these locations do not create value. Their sole raison d’être is the provision of corporate and �nancial service structures that may record value, but do not generate it. Of course, one way in which this value may be relocated into these places is through transfer mispricing. Accounting for the Missing Billions 301 The existence of the Big Four �rms of accountants—Pricewater- houseCoopers, Deloitte Touche Tohmatsu, Ernst & Young, and KPMG— in many secrecy jurisdictions in which their location cannot possibly be justi�ed solely by the needs of the local populations reinforces this view. As Murphy (2010b) shows, the Big Four �rms are signi�cantly overrep- resented in small secrecy jurisdictions (those with less than 1 million population) compared with other locations of this size (see table 9A.6).30 As he also shows, these locations of the Big Four have an average GDP per head that is approximately four times greater than the average GDP in similar locations in which the Big Four are not present. Of course, cause and effect cannot be proven based on this circumstantial evidence, but the possibility exists, especially in the smallest of such locations, that the income in question is not earned in these places, but is transferred in through transfer mispricing, and, in that case, the Big Four �rms, as Murphy suggests, facilitate the structures that allow the transfer mispric- ing to occur. Conclusion This chapter has sought, �rst, to test the credibility of the claim that at least US$160 billion a year may be lost by developing countries as a result of transfer mispricing by MNCs. Second, it has sought to demonstrate that developing countries and, particularly, the extractive industries may be especially prone to this abuse. Third, it has considered whether this sum may be hidden from view within the accounts or �nancial statements of the MNCs that may be perpetrating the mispricing. Fourth, it has explored the possibility that these MNCs may use secrecy jurisdictions to assist in hiding these transactions from view. As the chapter indicates, the incentive to transfer misprice is great. This is because, �rst, the differential in effective tax rates between the jurisdictions in potential supply chains is higher than the existing litera- ture on tax rates suggests would normally be the case. Second, it is because the structure of international tax regulation at present suggests that the chance of detection of transfer mispricing is low. 302 Draining Development? The chapter then shows that the risk that transfer mispricing may go undetected within the extractive industries that supply much of the exter- nal earning capacity of many developing countries is great. As the chapter also notes, transfer mispricing in this sector is unlikely to be motivated by a desire to avoid taxes on corporate pro�ts alone. The abuse is likely to extend to royalties, sales and purchase taxes, dividends, abuse of pro�t- sharing agreements, and more. The incentives to abuse for all these rea- sons are substantial, and the consequence of not tackling the issue consid- erable, but, as the chapter shows, the prospects of tackling the issue are limited given the current legislative and contractual constraints. The chapter also reviews a number of IFRSs, in particular the standards addressing consolidated accounts and �nancial statements, related-party transactions, and segment reporting: the three standards most likely to relate to disclosure of intragroup trade within the accounts of MNCs. The analysis shows that these standards are not designed to, and are not capa- ble of, leading to the disclosure of these transactions and the associated transfer pricing. It proposes that this failure may be intentional. The chapter de�nes secrecy jurisdiction and argues that secrecy juris- dictions are deliberate constructs. Secrecy jurisdictions are used dispro- portionately by MNCs as measured relative to local populations and GDP. In the secrecy jurisdictions most popular with the MNCs, there is little prospect of any real added value arising because, as the chapter shows, the economies of these secrecy jurisdictions are largely dedicated to the supply of �nancial services, part of which activity is administered by the MNC subsidiaries. This pattern of use does, however, accord with the proposed de�nition of a secrecy jurisdiction, suggesting that the pur- pose of these jurisdictions is to disguise the true nature of activity under- taken elsewhere so that compliance with the regulations of other states can be avoided or evaded. Transfer mispricing may be only one abusive activity that may be occurring. What overall conclusions may one draw from this evidence? The fol- lowing conclusions are proposed. First, it is apparent that there is a stronger incentive to transfer mis- price than the existing literature has suggested. Second, developing countries are particularly susceptible to this activity. Third, this activity can be hidden from view in accounts. Fourth, secrecy jurisdictions pro- vide an additional layer of opacity to disguise this activity. Fifth, the Accounting for the Missing Billions 303 combination of the secrecy inherent in accounting rules and secrecy jurisdiction legislation provides a deep opacity that limits the possibility of discovering transfer mispricing activity. As a result, the chapter �nds that transfer mispricing may be taking place undetected. In this case, is it also plausible that the quantum of the loss to devel- oping countries may be as much as US$160 billion per annum? As noted in the chapter, this seems to be quite plausible on the basis of data veri- �ed from a variety of sources and therefore considered a credible basis for the analytical review techniques common in auditing methodology. Thus, on the basis of the methodologies noted, the chapter concludes that substantial transfer mispricing by major corporations contributing to a loss of at least US$160 billion a year to developing countries is plau- sible in the context of the total likely corporate pro�ts earned worldwide in a year. Notes 1. Transfer mispricing occurs if two or more entities under common control that are trading across international borders price transactions between them at rates designed to secure a tax advantage that would not be available to third par- ties trading in the same goods or services across the same borders. 2. The arm’s-length principle is the international standard that OECD member countries have agreed should be used to determine transfer prices for tax pur- poses. It is set forth in article 9 of the OECD Model Tax Convention, as follows: where “conditions are made or imposed between the two enterprises in their commercial or �nancial relations which differ from those which would be made between independent enterprises, then any pro�ts which would, but for those conditions, have accrued to one of the enterprises, but, by reason of those condi- tions, have not so accrued, may be included in the pro�ts of that enterprise and taxed accordingly� (Centre for Tax Policy and Administration, “Annex 3: Glos- sary,� OECD, Paris, http://www.oecd.org/document/41/0,3343,en_2649_33753 _37685737_1_1_1_1,00.html, accessed October 13, 2011). See also chapter 7 in this volume. 3. IFFs are de�ned by Kar and Cartwright-Smith (2008) as “the proceeds from both illicit activities such as corruption (the bribery and embezzlement of national wealth), criminal activity, and the proceeds of licit business that become illicit when transported across borders in contravention of applicable laws and regulatory frameworks (most commonly in order to evade payment of taxes).� 4. Secrecy jurisdictions intentionally create regulation for the primary bene�t and use of nonresidents in the geographical domain. The jurisdiction regulation is 304 Draining Development? designed to undermine the legislation or regulations of another jurisdiction. The jurisdictions create a deliberate, legally backed veil of secrecy that ensures that those from outside who use the jurisdiction regulations cannot be identi- �ed by others. 5. See World Factbook (database), Central Intelligence Agency, Washington, DC, https://www.cia.gov/library/publications/the-world-factbook/. 6. Jurisdiction Centre (database), OCRA Worldwide, Isle of Man, United King- dom, http://www.ocra.com/jurisdictions/index.asp. 7. The table may be found on the website of this volume, http://go.worldbank.org /N2HMRB4G20. 8. The table may be found on the website of this volume, http://go.worldbank.org /N2HMRB4G20. 9. Devereux, Lockwood, and Redoano (2002) are typical of the other surveys in that they also entirely ignore the tax haven–secrecy jurisdiction issue. 10. Jurisdiction Centre (database), OCRA Worldwide, Isle of Man, United King- dom, http://www.ocra.com/jurisdictions/index.asp. 11. The table may be found on the website of this volume, http://go.worldbank.org /N2HMRB4G20. 12. Based on International Monetary Fund–World Bank data summarized at http:// en.wikipedia.org/wiki/List_of_countries_by_GDP_(nominal) (accessed March 3, 2010). In each case, the data are in the range of US$60 trillion to US$61 trillion. Original sources are linked through the website noted. 13. Data of the U.K. Of�ce for National Statistics. 14. HMRC (HM Revenue & Customs), 2009, “HM Revenue and Customs Annual Receipts,� data sheet, December, London, http://www.hmrc.gov.uk/stats/tax _receipts/table1-2.pdf. 15. Data of the Bureau of Economic Analysis, U.S. Department of Commerce. 16. Anecdotal evidence based on conversations with high-ranking of�cials of HM Revenue & Customs leads us to believe that approximately 35 percent of all large companies in the United Kingdom are considered to have little appetite for tax- ation risk, while at least 40 percent are considered to have exposed themselves to high risk in the management of their taxation affairs. 17. See South African Revenue Service, http://www.sars.gov.za/home.asp?pid=3906 (accessed August 17, 2009). 18. Lowtax, Global Tax and Business Portal, “Botswana: Double Tax Treaties.� http:// www.lowtax.net/lowtax/html/botswana/jbo2tax.html (accessed August 17, 2009). 19. Zambia Revenue Authority, “Tax Treaties.� http://www.zra.org.zm/Tax_Treaties _Agreemnet.php (accessed August 17, 2009). 20. See United Nations Conference on Trade and Development, http://www.unctad .org/sections/dite_pcbb/docs/dtt_Congo_DR.PDF. 21. It is appropriate to note that the author of this chapter has acted as adviser on transfer pricing or contractual issues to the authors of most of the reports referred to in this section. Accounting for the Missing Billions 305 22. See Standards (IFRSs) (database), International Accounting Standards Board, London, http://www.ifrs.org/IFRSs/IFRS.htm. 23. For the citation, see “US Tax Directors’ Poll: TP Is Most Contentious Issue,� TP Week, December 12, 2007, http://www.tpweek.com/Article.aspx?ArticleID= 1786798. 24. Reported by Richard Murphy, September 2006. 25. The table may be found on the website of this volume, http://go.worldbank.org /N2HMRB4G20. 26. The table may be found on the website of this volume, http://go.worldbank.org /N2HMRB4G20. 27. See Mapping Financial Secrecy (database), Tax Justice Network, London, http:// www.secrecyjurisdictions.com/index.php (accessed December 12, 2009). 28. See Mapping Financial Secrecy (database), Tax Justice Network, London, http:// www.secrecyjurisdictions.com/index.php (accessed December 12, 2009). 29. See Mapping Financial Secrecy (database), Tax Justice Network, London, http:// www.secrecyjurisdictions.com/index.php (accessed December 16, 2009). 30. The table may be found on the website of this volume, http://go.worldbank.org /N2HMRB4G20. References Baker, R. W. 2005. Capitalism’s Achilles Heel: Dirty Money and How to Renew the Free- Market System. Hoboken, NJ: John Wiley & Sons. Christian Aid. 2007. “A Rich Seam: Who Bene�ts from Rising Commodity Prices?� Christian Aid Report, January. Christian Aid, London. ———. 2008. “Death and Taxes: The True Toll of Tax Dodging.� Christian Aid Report, May. Christian Aid, London. Devereux, M. P., B. Lockwood, and M. Redoano. 2002. “Do Countries Compete over Corporate Tax Rates?� CEPR Discussion Paper 3400, Centre for Economic Pol- icy Research, London. Dietz, M., R. Reibestein, and C. Walter. 2008. “What’s in Store for Global Banking?� McKinsey Quarterly, January. Dischinger, M., and N. Riedel. 2008. “Corporate Taxes and the Location of Intangible Assets within Multinational Firms.� Discussion Paper in Economics 5294, Department of Economics, University of Munich, Munich. GAO (U.S. Government Accountability Of�ce). 2008. “Large U.S. Corporations and Federal Contractors with Subsidiaries in Jurisdictions Listed as Tax Havens or Financial Privacy Jurisdictions.� Report GAO-09–157 (December), U.S. Gov- ernment Accountability Of�ce, Washington, DC. Global Witness. 2006. “Heavy Mittal? A State within a State: The Inequitable Mineral Development Agreement between the Government of Liberia and Mittal Steel Holdings NV.� Report, October, Global Witness, Washington, DC. 306 Draining Development? ———. 2007. “Update on the Renegotiation of the Mineral Development Agree- ment between Mittal Steel and the Government of Liberia.� Update, August, Global Witness, London. Greenpeace. 2008. “Logging Sector Review: Conning the Congo.� Report, July. http:// www.greenpeace.org/raw/content/international/press/reports/conning-the -congo.pdf. HMRC (HM Revenue & Customs). 2010. “Measuring Tax Gaps 2009.� Report, March (revised), HMRC, London. http://www.hmrc.gov.uk/stats/measuring-tax-gaps .pdf. Huizinga, H. 2009. “Pro�t Shifting Activities in Europe.� Paper presented at the Tax- ation and Customs Union’s Brussels Tax Forum 2009, “Tax Systems in a Chang- ing World,� European Commission, Brussels, March 30–31. http://ec.europa .eu/taxation_customs/resources/documents/taxation/gen_info/conferences /taxforum2009/pres_Huizinga.pdf. Kar, D., and D. Cartwright-Smith. 2008. “Illicit Financial Flows from Developing Countries, 2002–2006.� Global Financial Integrity, Washington, DC. http:// www.gfip.org/storage/gfip/economist%20-%20final%20version%201-2-09 .pdf. Keen, M., and M. Mansour. 2009. “Revenue Mobilization in Sub-Saharan Africa: Challenges from Globalization.� IMF Working Paper 09/157, International Monetary Fund, Washington, DC. London Stock Exchange. 2010. “Market Statistics.� Main Market Fact Sheet, Febru- ary, London Stock Exchange, London. http://www.londonstockexchange.com /statistics/historic/main-market/february-10.pdf. Murphy, R. 2007. “Ghana’s EITI: Delivering on the Promise? A Review of the First Report on the Aggregation/Reconciliation of Mining Bene�ts in Ghana.� Report, March, Tax Research LLP, Downham Market, U.K. ———. 2008. “The Missing Billions: The UK Tax Gap.� Touchstone Pamphlet 1, Trades Union Congress, London. ———. 2009a. “Country-by-Country Reporting: Holding Multinational Corpora- tions to Account Wherever They Are.� Report, June 17, Task Force on Financial Integrity and Economic Development, Washington, DC. ———. 2009b. “The TIEA Programme Is Failing.� Richard Murphy on Tax and Eco- nomics, November 27, Tax Research UK, London. http://www.taxresearch.org. uk/Blog/2009/11/27/the-tiea-programme-is-failing/. ———. 2010a. “Tax Justice and Jobs: The Business Case for Investing in Staff at HM Revenue & Customs.� Report, March, Tax Research LLP, Downham Market, United Kingdom. http://www.taxresearch.org.uk/Documents/PCSTaxGap.pdf. ———. 2010b. “Where 4 Art Thou? A Geographic Study of the Big 4 Firms of Accountants and the Implications of Their Location for Illicit Financial Flows.� Report, Tax Research LLP, Downham Market, U.K. http://www.taxresearch .org.uk/Documents/Where4ArtThou.pdf. Accounting for the Missing Billions 307 Mwenda, K. 1999. “Legal Aspects of Foreign Direct Investment in Zambia.� Murdoch University Electronic Journal of Law 6 (4), Murdoch University, Perth, Australia. http://www.murdoch.edu.au/elaw/issues/v6n4/mwenda64.html. OECD (Organisation for Economic Co-operation and Development). 2003. “Arti- cles of the Model Convention with Respect to Taxes on Income and on Capital.� January 28, OECD, Paris. http://www.oecd.org/dataoecd/52/34/1914467.pdf. Open Society Institute, Third World Network Africa, Tax Justice Network Africa, Nairobi Action Aid International, and Christian Aid. 2009. “Breaking the Curse: How Transparent Taxation and Fair Taxes Can Turn Africa’s Mineral Wealth into Development.� Report, March, Open Society Institute of Southern Africa, Johannesburg. 10 Trade Mispricing and Illicit Flows Volker Nitsch Abstract A potential vehicle to move capital unrecorded out of a country is the mis- invoicing of international trade transactions. Exporters may understate the export revenue on their invoices, and importers may overstate import expenditures, while their trading partners are instructed to deposit the bal- ance for their bene�t in foreign accounts. Aiming to quantify the extent of trade mispricing, studies have analyzed asymmetries in matched partner trade statistics or examined price anomalies in transaction-level price data. This chapter critically reviews these empirical approaches and briefly describes an alternative methodology. Overall, the accuracy and reliability of estimates of illicit �nancial flows (IFFs) based on trade mispricing are questioned. In particular, it is argued that estimates of trade mispricing are critically dependent on assumptions on how to interpret observed asym- metries in trade statistics. For instance, various reasons for discrepancies in bilateral trade statistics are discussed, and incentives for faking trade invoices other than capital flight are highlighted. Also, aggregate trade data may mask considerable variation in trade discrepancies at the transaction level. Most notably, the importance of trade mispricing as a method for the unrecorded crossborder transfer of capital is generally unclear. 309 310 Draining Development? Introduction The (in)accuracy of international trade statistics has recently (once more) become an issue of much debate. Trade data are often critically reviewed (more than other economic statistics) for at least three reasons. First, data on international trade are of considerable economic relevance. Crossborder shipments of goods and services often have sizable effects on a country’s economic activity. Also, for some countries, taxes on international trade constitute a signi�cant share of government revenue. Second, a country’s trade data allow, in principle, for full cross-checks with data from other countries. Individual trade transactions are recorded separately by both trading partners so that it should be quite easy to compare these records. As Naya and Morgan (1974, 124) note, “comparable double accounts are not usually available for domestic eco- nomic transactions.� Finally, recent research has reemphasized that trade activities are subject to criminal behavior. Fisman and Wei (2009) show that the level of smuggling of cultural property is related to the level of corruption in the country of origin. Baldwin (2006) discusses the effect of value added tax (VAT) fraud on intra-European trade �gures. A feature of international trade statistics that has frequently attracted considerable attention is the potential asymmetry in partner country trade statistics arising from mispricing. More speci�cally, it has been argued that the faking of trade invoices is a commonly used method to move money out of developing countries. If trade declarations are manipulated such that the stated value of imports exceeds the actual value (overinvoicing) or the stated value of exports is below the actual value (underinvoicing), �nancial resources are implicitly transferred abroad without any of�cial record of this having taken place. Conversely, it is assumed that, to proxy for such IFFs (often labeled capital flight), trade statistics may provide some useful empirical indication. For instance, Kar and Cartwright-Smith (2008) explore gaps in mirror trade statistics; also see Bhagwati, Krueger, and Wibulswasdi (1974) for an early contribution. De Boyrie, Pak, and Zdanowicz (2004) examine the variation of trade prices at the transaction level. This chapter provides a detailed discussion of issues associated with trade mispricing. It reviews various incentives for faking trade invoices, Trade Mispricing and Illicit Flows 311 critically examines empirical approaches to quantify the extent of mis- pricing (thereby also highlighting other potential reasons for discrepan- cies in bilateral trade statistics), and analyzes differences in mirror trade statistics at the product level (which may be associated with mispricing behavior). It also assesses how trade mispricing may be reflected in capi- tal flight and illicit flow estimates. The plan of the chapter is as follows. The �rst section reviews evi- dence on asymmetries in international trade statistics. IFFs are typically generated by underinvoicing exports and overinvoicing imports. How- ever, dozens of other reasons for the over- and underinvoicing of trade activities have been identi�ed in the literature; these reasons include overinvoicing of exports to bene�t from export subsidies and underin- voicing of imports to avoid payment of import tariffs. As a result, some of these misinvoicing activities may offset each other. In addition, there are other reasons for incorrect trade invoices and, thus, discrepancies in of�cial trade �gures. These manipulations may be the result of intended (criminal) behavior such as smuggling or carousel fraud (explained below). Similarly, however, they could simply reflect inaccuracies in compiling trade �gures (for example, because of the Rotterdam effect) and thus result from unintended behavior. Based on the discussion of incentives for misreporting trade flows, the chapter next analyzes empirical approaches to quantifying the extent of mispricing. The trade-based approach in Kar and Cartwright-Smith (2008) and the price-based work by Pak and Zdanowicz (with various coauthors) are critically reviewed. The chapter then examines how trade mispricing �ts into the illicit flow estimates. Kar and Cartwright-Smith (2008) use the gap in trade statistics (along with other approaches) to provide detailed estimates of IFFs by country. This section examines the robustness of their �ndings. For instance, results are compared with estimates derived from the product-level approach of Fisman and Wei (2009). Similarly, it may be useful to relate Kar and Cartwright-Smith’s estimates of illicit flows to other country characteristics that have been found to be associated with trade mispricing, such as corruption. The �nal section concludes. 312 Draining Development? Discrepancies in Trade Statistics Conceptually, country i’s exports to country j are equivalent to j’s imports from i. In practice, however, recorded trade �gures differ in of�- cial matched partner country trade statistics. The reasons for asymme- tries in trade statistics are manifold, but they can be broadly grouped into two categories: legitimate statistical reasons and intended misdecla- rations. For illustration, this section briefly summarizes some of the main explanations for the discrepancy in of�cially reported export and import �gures. Morgenstern (1950), in a classical paper, provides a more detailed account. Statistical reasons The most notable source of discrepancy between the exports of one country and the imports of the other is the conceptual difference in val- uation. Exporting countries report the value of goods at the initial point of departure (free on board [f.o.b.]), while import values refer to the value at the point of �nal destination, thereby including cost, insurance, and freight (c.i.f.).1 As a result, the c.i.f.–f.o.b. ratio has been frequently used in the literature as a measure of transportation costs. Limão and Venables (2001) provide a recent application of this approach; see Hum- mels and Lugovskyy (2006) for a detailed critique. Apart from the different treatment of shipping costs, there are other methodological dif�culties in matched partner trade statistics. For instance, the correct identi�cation of the source or destination country may be a problem. If the country of �nal destination is not known at the time of exportation, the exporter declares the country of last shipment; the country of �nal destination, in contrast, classi�es its imports by country of origin. Herrigan, Kochen, and Williams (2005) provide an illustration of this “Rotterdam effect� in trade statistics whereby import and export statistics are distorted because of the transit through inter- vening countries. They note as follows: Crude oil is imported from the UK via Portland, Maine in the USA, and then sent by pipeline to Canada. This is recorded in Canada as [an] import from the UK, whereas the UK records an export to the USA. The USA do not record either flow, as it is simple transit. Thus the UK records more oil exports to the USA [than] they record importing, and [records] less oil Trade Mispricing and Illicit Flows 313 exports to Canada [than] they record importing. (Herrigan, Kochen, and Williams 2005, 9) Another potential issue of importance is timing. Because there are often notable time lags between the departure and arrival of a shipment (for example, arising from long-distance sea cargo, a delay in customs declaration, or temporary storage in a warehouse), trade may be recorded in different calendar years. More importantly, statistical of�ces in the source and destination countries may value goods at different prices or exchange rates. Finally, recorded trade at the commodity level may differ because of the omission of individual transactions in one of the partner countries (for example, because of varying thresholds for the low value of trade across countries), the exclusion of certain product groups in a country’s trade statistics (such as military material or repair trade), or differences in commodity classi�cations (for example, a regrouping of a transaction into chapter 99, which covers “items not elsewhere classi- �ed,� for reasons of con�dentiality). Fraudulent trade activities Discrepancies in trade statistics may also result from intended misdecla- ration of trade activities. Transactions may be hidden completely (so that of�cial statistics underreport trade), purely imaginary (so that trade is overstated in the data), or mispriced in trade invoices (with, a priori, unknown effects on trade statistics). Unreported trade. For unreported trade activities (that is, smuggling) to affect asymmetries in partner country trade statistics, the transactions have to be recorded by one of the partners. This is the setup that Fisman and Wei (2009) have in mind. For antiques and cultural property, there are often strong export restrictions such that there is an incentive to smuggle the goods unrecorded out of a country. Imports, however, are properly declared since there are generally no constraints for the entry of such goods, and there is even the risk of seizure if there is a false declaration. Fictitious trade. An example of imaginary trade transactions (where of�cial trade �gures are arti�cially inflated) is missing trader VAT fraud in intra–European Union (EU) trade. For intra-EU trade, for which no 314 Draining Development? barriers to trade exist, and, thus, there are no customs declarations, trade statistics rely on the VAT system; that is, �rms declare to �scal authori- ties, as part of the VAT return, trade activities with customers and sup- pliers in EU member countries. Trade statistics are then affected by two types of VAT fraud. In acquisition fraud, goods are regularly imported VAT free and then sold on the home market (for example, to the next trader) at a price that includes VAT. Instead of paying over the VAT to tax authorities, however, the importer disappears. As a result, because of the missing VAT declaration, imports also remain unrecorded. A more elab- orate version of this form of criminal attack on the VAT system is carou- sel fraud whereby, after a series of sales through home companies, the imported goods are reexported to the country of origin (or any other EU member country) and, thus, move in a circular pattern. The exports are properly declared, while the imports are not captured in trade data, which may lead to substantial asymmetries in partner country trade sta- tistics. When, in 2003, the United Kingdom’s Of�ce for National Statis- tics made corrections to trade �gures for VAT fraud, real growth in gross domestic product for previous years was lowered by up to 0.2 percentage points.2 Misreported trade. Finally, trade may be recorded, but invoices are faked such that the declared value of a trade transaction deviates from the true value. A plausible explanation for trade mispricing is capital flight. If there are exchange restrictions, overinvoicing of imports and underin- voicing of exports are popular methods for the unrecorded movement of capital out of the country. However, there are other reasonable expla- nations for mispricing. Some of these explanations work in a similar direction. For instance, underreporting of exports allows �rms to acquire foreign exchange that is not disclosed to national authorities; the foreign currency can then be freely used by exporters without complying with controls and regulations (for example, a potential option may be the sale of foreign currency in the parallel exchange rate market). Furthermore, authorities may use information on the export activities of �rms to infer the production of these �rms. As a result, �rms that seek to hide output (for example, to evade domestic taxes) will automatically also seek to hide exports. Dabla-Norris, Gradstein, and Inchauste (2008) provide a description of informal activities by �rms. Trade Mispricing and Illicit Flows 315 Others work in the opposite direction. For instance, if there are import restrictions, there is an incentive to underinvoice imports; the misdeclaration of cargo is an obvious solution to circumvent these trade restrictions. Bhagwati (1964) provides an early empirical assessment of such activities. Similarly, to bene�t from export subsidies, exports can be overinvoiced. Celâsun and Rodrik (1989) argue that a sizable share of the increase in Turkish exports after 1980 is caused by a change in invoic- ing practices among domestic entrepreneurs (to take advantage of gen- erous export subsidies). Summary The �nding that of�cial trade statistics may suffer from misreporting and faked declarations is a fact well known not only to statisticians of international trade. Bhagwati (1964, 1967) provides an early economic discussion of incentives for misinvoicing in trade; Bhagwati and Han- sen (1973) develop a trade model to examine the welfare effects of smuggling. Using trade statistics to quantify the extent of misreporting, however, appears to be dif�cult. Misreporting can work in either direction, so that some activities may offset each other in aggregate trade statistics. Also, the extent to which transactions are reported at all (and thus show up in the trade statistics of at least one of the trade partners) may vary. Increased surveillance of trade transactions have apparently little mea- surable effect.3 Finally, discrepancies in trade statistics may simply arise from statistical factors. A quantitative assessment of such factors in a bilateral context has been provided for trade flows between Australia and the EU; reported trade �gures differ, on average, by about 10 percent between the two trade partners. Table 10.1 summarizes the results. As shown, the largest source of discrepancy between recorded exports and imports is the difference in (c.i.f.–f.o.b.) valuation, which inflates Euro- pean import data by, on average, about 9 percent. In contrast, goods that are imported by Australia and are subsequently reexported to the EU may have initially arti�cially lowered the discrepancy (since they are recorded in Australia’s exports, but not in the EU’s imports). In view of all these dif�culties, the EU, though aiming to reduce the declaration burden on businesses, still refrains from using mirror (single-flow) trade statistics.4 316 Draining Development? Table 10.1. Asymmetries in Australia–European Union Trade: A Statistical Practitioner’s Assessment Indicator 1992 1993 1994 1995 1996 1997 a. Australian exports (f.o.b.) 7,711 7,476 7,247 8,007 8,381 8,678 ($A, millions) Adjustments Country classi�cation Australian reexports −400 −41 7 −327 −45 1 −394 −333 EU indirect imports 1 01 1 33 1 20 1 37 242 1 85 EU reimports 215 195 240 245 275 315 Exchange rate −25 −4 7 −1 1 −28 Ships 80 — — — — — Adjusted Australian exports 7,682 7,383 7,286 7,937 8,504 8,815 b. EU imports (c.i.f.) ($A, millions) 8,72 1 7, 159 7,979 8,8 13 8,526 9,570 Adjustments Valuation −728 −605 −658 −737 −708 −795 Timing −93 25 81 54 62 23 Exchange rate −41 13 −7 — −6 –15 Ships — — — — 236 — Nonmonetary gold 63 484 126 −4 119 — Adjusted EU imports 7,922 7,077 7,520 8,124 8,228 8,782 c. Discrepancy Unadjusted 1,010 −316 732 806 1 45 892 percent 13 −4 10 10 2 10 Adjusted 240 −306 235 187 −276 −34 percent 3 −4 3 2 −3 — Source: ABS 1998. Note: — = not available. Trade Mispricing and Illicit Financial Flows IFFs are, by de�nition, unobservable in of�cial statistics. However, based on the assumption that trade mispricing is an important method for the unrecorded movement of capital out of a country, a number of papers have examined trade statistics to provide a rough empirical indication of the magnitude of IFFs. In the following, two of these approaches are discussed in detail: the analysis of asymmetries in trade statistics and the analysis of price anomalies in transaction-level trade data. Trade Mispricing and Illicit Flows 317 Conceptual issues It may be useful to address briefly some conceptual issues on the de�ni- tion of IFFs. In describing capital flows, one often applies different con- cepts to classify capital transactions. For instance, as noted by the Inter- national Monetary Fund (IMF), there is an occasional tendency to identify any capital outflow from a country as capital flight (IMF 1992). Kar and Cartwright-Smith appear to follow this (broadest possible) approach for capital outflows from developing countries; they argue that “the term flight capital is most commonly applied in reference to money that shifts out of developing countries, usually into western economies� (Kar and Cartwright-Smith 2008, iii). While this approach may be applicable, a possible shortcoming of this general de�nition is that it also covers all standard (or normal) crossbor- der capital transactions. Therefore, a number of authors prefer to take a more restrictive approach that includes only a subset of capital move- ments and also justi�es the negative connotation (capital flight). Walter (1987), for instance, emphasizes the importance of motivations for flight (such as macroeconomic mismanagement or fear of con�scation) by arguing that flight capital is capital that flees. Cumby and Levich (1987) focus on the type of transaction by excluding all freely organized legal transactions from their de�nition of capital flight. It might thus seem reasonable to distinguish capital transactions along various dimensions, such as the source of capital, the method of transfer, and the motivation for the transaction. A similar reasoning may also apply to the de�nition of IFFs. Kar and Cartwright-Smith (2008, iv) provide a comprehensive approach, treat- ing all unrecorded capital transfers as illegal and note, more generally, that “illicit money is money that is illegally earned, transferred, or uti- lized.�5 Another plausible de�nition, by contrast, also takes into account the motivation for such behavior. IFFs may then be de�ned as any capital transaction that intentionally moves capital out of a country in a man- ner that is not recorded; trade mispricing is therefore one of (potentially many) possible conduits for such conduct.6 Examining asymmetries in trade data A prominent approach to quantifying the extent of misinvoicing is the analysis of matched partner country trade statistics. Based on the principle 318 Draining Development? of double counting in trade statistics, a country’s exports to a partner are compared to what the partner reports as its imports (mirror statistics). The difference may then provide a reasonable indication of illicit flows that occur through mispricing. A recent application of this approach is Kar and Cartwright-Smith (2008); an early contribution is Bhagwati, Krueger, and Wibulswasdi (1974). Trade asymmetries mispricing. Any estimate of IFFs derived from asymmetries in trade statistics has to deal explicitly with three types of problems. The most notable issue is related to the accuracy of trade sta- tistics in general. As noted above, partner country trade data typically differ for various reasons. As a result, trade �gures have to be adjusted before any possible remaining statistical asymmetry can be reasonably interpreted as deriving from mispricing. At a minimum, trade data have to be corrected for differences in valuation, which is most likely to be the main reason for a discrepancy in mirror statistics. However, the infor- mation on transportation costs that is needed for the conversion from f.o.b. to c.i.f. values is rare; for precisely this reason, some trade econo- mists have used the c.i.f.–f.o.b. ratio as a proxy for transportation costs.7 Two approaches have frequently been applied in the literature to deal with this issue. First, the analysis may focus particularly on episodes in which f.o.b. export values exceed the corresponding c.i.f. import values. Because the latter include additional price components (transportation costs) and, therefore, should, by de�nition, be larger than the former, such perverse statistical �ndings may indicate that mispricing has occurred; this argument was �rst made by Bhagwati (1964) in an empir- ical analysis of the underinvoicing of imports. Capital outflows through trade mispricing, however, work exactly in the opposite direction; these activities are associated with overinvoicing of imports and underinvoic- ing of exports, which tend to inflate the observed difference between c.i.f. and f.o.b. values so that this approach is of little help. Second, a number of papers apply a flat c.i.f.–f.o.b. conversion factor. Typically, with reference to conventions by international organizations, a 10 percent difference between c.i.f. and f.o.b. values is assumed. For instance, in using partner data to supplement their trade database, the IMF generally applies a c.i.f.–f.o.b. factor of 1.1. Any discrepancy in mir- ror statistics that exceeds this correction might then be attributed to Trade Mispricing and Illicit Flows 319 mispricing; for instance, see Bhagwati, Krueger, and Wibulswasdi (1974). However, this approach provides, at best, only a crude empirical indica- tion of the potential presence of misinvoicing because the assumption of a �xed conversion factor that varies neither over time nor among trading partners is clearly challenging. Mispricing illicit flows. To the extent that discrepancies in trade statis- tics indeed reflect mispricing, over- and underinvoicing of trade trans- actions may be completely unrelated to IFFs (narrowly de�ned). An obvious reason to underinvoice trade is to avoid the payment of trade taxes; similarly, overinvoicing allows a trader to bene�t from trade sub- sidies. More speci�cally, if export duties are ad valorem, the motive to reduce the effective tax rate by underinvoicing exports is indistinguish- able from a flight-motivated capital outflow in the analysis of mirror statistics in international trade. In similar fashion, the overinvoicing of imports (for other reasons than capital flight) may occur if a �rm seeks to reduce its before-tax pro�ts (and, hence, the effective tax on pro�ts) by overstating the cost of imported inputs. As a result, an assumption has to be made about the extent to which trade mispricing is, indeed, a channel for illicit flows. Illicit flows trade asymmetries. Conversely, it is questionable whether the extent of illicit flows is detectable from asymmetries in aggregate trade data. Various incentives to fake trade invoices that work exactly in the opposite direction relative to the capital flight motives can be identi- �ed. The effects of these explanations on trade (that is, underinvoicing of imports and overinvoicing of exports) have been widely documented. For instance, Celâsun and Rodrik (1989, 723) note that the introduction of export subsidies in Turkey in the early 1980s led to substantial overin- voicing: “Turkish entrepreneurs, never too shy in exploiting arbitrage opportunities, used the wedge [between the pro�tability of manufac- tures exports and the pro�tability of other means of earning foreign exchange] to their advantage.� Celâsun and Rodrik conclude (1989, 729– 30) that, “once �ctitious exports are eliminated, the average growth rate of Turkish exports . . . is not nearly as spectacular as [that] . . . calculated from of�cial statistics.� The existence of obvious incentives for underin- voicing imports has been recently documented empirically by Yang 320 Draining Development? Table 10.2. Motives for Mispricing in International Trade Trade Overinvoicing Underinvoicing Exports Capturing export subsidies Capital flight, avoiding export taxes Imports Capital flight, lowering domestic pro�ts Evading import duties Source: Adapted from Dornbusch and Kuenzler (1993). (2008), who examines an increase in enforcement in Philippine customs that targets a speci�c method of avoiding import duties. He nicely illus- trates that increased enforcement (by lowering the minimum value threshold for preshipment customs inspection for shipments from a subset of origin countries) reduced the targeted duty-avoidance method, but caused substantial displacement in an alternative method (shipping via duty-exempt export processing zones). In sum, there are various motives for mispricing in trade invoices. Table 10.2 lists some of these motives. As these effects partly work in opposite directions, they may easily cancel each other out at the aggre- gate trade level. For instance, if a shipment is underinvoiced in the exporting country to move capital unrecorded out of the country, and the shipment carries the same mispriced invoice in the importing coun- try to evade import tariffs, no discrepancy in mirror trade statistics will occur.8 Reviewing the various motives, Bhagwati, Krueger, and Wibuls- wasdi (1974) argue that underinvoicing of exports, rather than overin- voicing of imports, is more often used as a vehicle of capital flight, also because export controls are less restrictive. Gulati (1987) even �nds that, if trade invoices are faked, the underinvoicing of imports more than out- weighs the underinvoicing of exports so that, for a number of countries, (illicit) capital inflows are observed. Kar and Cartwright-Smith (2008). In view of these dif�culties, it is inter- esting to review recent estimates by Kar and Cartwright-Smith (2008) of capital flight arising because of trade mispricing.9 These authors analyze international trade statistics, address the above-mentioned relevant issues in a consistent and transparent way, and estimate, based on this analysis, the volume and pattern of IFFs. While their efforts have provoked useful debate, it should also be clear that the reliability of the estimates crucially hinges on the assumptions made and the conventions chosen. Trade Mispricing and Illicit Flows 321 A �rst set of methodological issues in Kar and Cartwright-Smith (2008) relates to the interpretation of observed asymmetries in matched partner trade statistics. As noted above, trade asymmetries may arise for a variety of reasons and, therefore, do not necessarily reflect trade mis- pricing. Kar and Cartwright-Smith (2008) acknowledge the dif�culty in properly identifying mispricing from trade data by providing a de�nition for trade asymmetries of interest. Speci�cally, after a careful discussion of the limitations of various models in estimating illicit flows, they provide two sets of estimates of trade mispricing by country. In their baseline model, they apply a gross excluding reversals method. This method is based on the assumption that episodes of bilateral export underinvoicing and import overinvoicing reflect capital outflows, while they argue that episodes of pair-wise trade gaps in the opposite direction (possibly repre- senting capital inflows) are spurious because of data issues (and therefore simply set to zero). As a result, if a country reports low export �gures rela- tive to recorded imports in the corresponding partner country, this dis- crepancy is de�ned, according to Kar and Cartwright-Smith, as mispric- ing. In contrast, a statistical discrepancy similar in magnitude that results from the same country’s relatively large exports relative to another part- ner’s imports is, by de�nition, ignored. This selective interpretation of trade asymmetries appears to inflate estimates of IFFs arti�cially. Indeed, if offsetting capital flows are additionally taken into account, a country’s net position is obtained, which is often sizably lower than the estimated gross excluding reversals result. With this extension, many aggregate country estimates even change signs. However, Kar and Cartwright- Smith argue that results derived from such an approach are distorted and unrealistic and focus instead on their method. Kar and Cartwright-Smith’s (2008) interpretation becomes particu- larly troublesome if one also takes the treatment of transportation costs in their analysis into consideration. A major source of discrepancy between exports and the corresponding imports is the difference in val- uation concepts, so that c.i.f. imports must be converted to f.o.b. values. Kar and Cartwright-Smith apply a c.i.f.–f.o.b. conversion factor that is �xed (at 1.1) across all bilateral country pairs (irrespective of trade distance and the commodity composition of trade) and over time. In practice, however, c.i.f.–f.o.b. ratios in international trade statistics often lie outside a reasonable range of variation. According to Hummels and 322 Draining Development? Lugovskyy (2006), roughly half of all observations in the IMF’s Direc- tion of Trade Statistics database lie outside a (1, 2) range (which would be consistent with ad valorem transportation costs between 0 and 100 percent); the remaining observations contain substantial errors in lev- els.10 Still, Kar and Cartwright-Smith classify, after a flat 1.1 correction, any c.i.f.–f.o.b. ratio larger than 1 as evidence of trade mispricing, while ratios below this threshold (often covering substantial fractions of trade) are treated as noise.11 Another set of issues refers to the dif�culty of identifying mispricing and estimating IFFs from aggregate trade data. Mispricing occurs at the level of the individual trade transaction, whereas Kar and Cartwright- Smith (2008) examine trade asymmetries at the country level. Their focus on aggregate trade, however, is likely to produce inconsistent results. Assume, for instance, that trade gaps at the commodity level can- cel each other out at the aggregate level; these trade gaps remain uncap- tured by Kar and Cartwright-Smith, potentially leading to aggregation bias. More realistically, if trade gaps are concentrated in a few product categories with serious dif�culties in properly reporting trade flows (for example, crude oil), the resulting trade asymmetries at the country level are taken as evidence of trade mispricing (and, thus, IFFs). Moreover, any pooled estimate of asymmetries in aggregate trade data may mask considerable variation in export and import behavior. Therefore, if one computes trade gaps for a large set of countries, it appears particularly helpful to distinguish between and report results separately for the underinvoicing of exports and the overinvoicing of imports. In principle, there should be no difference between the two methods: both activities involve faking of trade declarations, and both activities may be instruments to facilitate illicit capital flows. In practice, however, there are typically large differences in the observed degrees of export underinvoicing and import overinvoicing; see, for instance, Gulati (1987). Bhagwati, Krueger, and Wibulswasdi (1974) argue that, especially, the incentive to overinvoice imports as a vehicle of capital flight is often overcome by other motives to fake trade invoices. As a result, there is an asymmetry of conduit behavior for IFFs that should be documented in empirical �ndings to aid the proper interpretation of the results. Trade Mispricing and Illicit Flows 323 Examining transaction-level price data In a series of papers, Simon Pak and John Zdanowicz analyze price data in transaction-level trade statistics. In a typical analysis, such as de Boyrie, Pak, and Zdanowicz (2005), the authors examine information from indi- vidual export declarations and entry forms in U.S. external trade. The data set is huge; it contains about two million records per year.12 Speci�- cally, Pak and Zdanowicz are interested in the product code, the partner country involved in the trade, and the price quoted in the trade docu- ment. With this information, it is possible to analyze, for each product- country pair, the range of prices recorded in trade transactions. More notably, based on this price range, transactions with abnormal prices can be identi�ed; Pak and Zdanowicz de�ne prices that fall outside the inter- quartile range as abnormal. These improperly priced transactions are then assumed to constitute illegal capital flows, and the capital outflow is determined by the dollar value of the over- or underinvoicing of a trans- action based on its deviation from interquartile prices. What is a product? Although Pak and Zdanowicz’s idea of using micro- level trade data is generally intuitive, their approach is not without dif- �culties. An obvious issue is the de�nition of a product. Pak and Zdano- wicz rely on the harmonized classi�cation of products in international trade statistics. At the most detailed level (for U.S. external trade), this commodity code system contains about 20,000 product categories. It is unknown, however, whether these products are, indeed, homogeneous; there may be considerable differences in product characteristics (such as quality) and, thus, in product prices within categories. As a result, the price range within a category may be wide so that mispriced transactions remain undetected. Alternatively, a transaction may be mistakenly iden- ti�ed as improperly priced if most transactions within a category are in low-value products, while a single transaction is in an expensive high- quality product. The relevance of product de�nitions has been recently highlighted by Javorcik and Narciso (2008), who argue that there is broader scope for fak- ing invoices in differentiated products because it is dif�cult to assess the quality and thus the price of such products. Examining trade gaps in mir- ror trade statistics for German exports to 10 Eastern European countries, 324 Draining Development? they �nd that the responsiveness of the trade gap to the level of tariffs is greater for differentiated products than for homogeneous goods. In addition, it should be mentioned that the product categories are de�ned for customs purposes. This implies, for instance, that some products are properly and tightly de�ned, while there are plenty of prod- uct codes that simply collect all other types of products. For illustration, consider the description of the 2009 harmonized tariff schedule.13 The tariff schedule lists 28,985 product categories (including subheadings), of which about half (12,581 categories) contain the catchall “other� in the description. Examples are shown in table 10.3. As indicated, the �rst category of heading 4901 covers printed material in single sheets. This material is categorized into “reproduction proofs� and all “other� single- sheet material; the other single-sheet material may range from plain black-and-white leaflets to expensive art prints on special paper, and there is no further quali�cation in the customs statistics. The category “art books� is divided into subcategories by price alone: less than US$5 or more than US$5 for each book. However, it is obvious that art books that cost more than US$5 may still vary considerably in style, format, and quality (and therefore also price). Finally, the unit of quantity for “tankers� is the number of ships; differences in the size and equipment of such ships may imply considerable price differences per unit. Abnormal prices. Another challenge facing the Pak and Zdanowicz approach is the arbitrary de�nition of abnormal prices.14 The focus on the interquartile price range appears to be taken from U.S. legal regula- tions aimed at detecting transfer price manipulation in trade between related parties. However, there is no economic reason to describe prices above or below a certain threshold level as abnormal a priori. In fact, even for products with minor variations in prices, the method may, under speci�c circumstances, identify improper pricing. More importantly, any analysis that is based on the distribution of prices appears to be sensitive to the number of observations. For exam- ple, transactions involving products for which a similar range of traded prices is observed may be classi�ed as normal or abnormal depending on the number of trade transactions. Alternatively, the occurrence of an additional data point may lead to a reclassi�cation of a transaction from normal to abnormal or vice versa. In general, the distribution of prices Trade Mispricing and Illicit Flows 325 Table 10.3. Examples of Product Categories in the U.S. Tariff Schedule, 2009 Heading/subheading, statistical suf�x Description 4901 Printed books, brochures, leaflets and similar printed matter, whether or not in single sheets 4901.10.00 In single sheets, whether or not folded 20 Reproduction proofs 40 Other Other 4901.91.00 Dictionaries and encyclopedias, and serial installments thereof 20 Dictionaries (including thesauruses) 40 Encyclopedias 4901.99.00 Other 10 Textbooks 20 Bound newspapers, journals and periodicals provided for in Legal Note 3 to this chapter 30 Directories Other 40 Bibles, testaments, prayer books, and other religious books 50 Technical, scienti�c, and professional books Art and pictorial books 60 Valued under US$5 each 65 Valued US$5 or more each Other 70 Hardbound books 75 Rack size paperbound books Other 91 Containing not more than 4 pages each (excluding covers) 92 Containing 5 or more pages each, but not more than 48 pages each (excluding covers) 93 Containing 49 or more pages each (excluding covers) 8901 Cruise ships, excursion boats, ferry boats, cargo ships, barges and similar vessels for the transport of persons or goods 8901.10.00 00 Cruise ships, excursion boats and similar vessels principally designed for the transport of persons; ferry boats of all kinds 8901.20.00 00 Tankers 8901.30.00 00 Refrigerated vessels, other than those of subheading 8901.20 8901.90.00 00 Other vessels for the transport of goods and other vessels for the transport of both persons and goods Source: Author compilation based on data of United States International Trade Commission, “Of�cial Harmonized Tariff Schedule,� http://www.usitc.gov/tata/hts/index.htm. 326 Draining Development? should become more representative, the greater the number of price observations available. For many products, however, there are typically few bilateral trade transactions. Price variations over time. Prices above or below a certain threshold are mechanically classi�ed as faked, but the results are potentially dif�cult to interpret. Differences in prices at the transaction level may arise for var- ious reasons. These reasons may be trivial, such as mistakes in �lling the form. However, there may also be differences in prices across markets (within the United States or internationally). Also, if variations in prices are analyzed over a longer time horizon, transaction dates may matter. For instance, seasonality may lead to considerable price fluctuations.15 Quantity faking. An issue that remains untouched by the Pak and Zdan- owicz approach is misinvoicing of quantities. Instead of faking prices or unit values, a trader may fake the invoice by misstating the quantity shipped, that is, the container holds a quantity that is different from the invoiced quantity. Although the method is perhaps more detectable than unit value faking, Bhagwati (1981, 417) notes that this is a “rather com- mon form of illegality.� In addition, there are other ways of misinvoicing, including, for instance, the omission of invoiced spare parts. Trade pairs. Since Pak and Zdanowicz’s results are exclusively based on data from U.S. customs, it is unclear to what extent they can be general- ized. Anecdotal evidence suggests that the faking of trade invoices occurs especially in trade with a main trading partner that is not necessarily the United States. For instance, Celâsun and Rodrik (1989) �nd strong evi- dence of misreporting only in Turkey’s trade with Germany. For African countries, trade with South Africa, the regional economic and �nancial center, may be the preferred target for misdeclaration. Trade Asymmetries: An Empirical Analysis For illustration, this section briefly examines trade asymmetries at the commodity level.16 This empirical exercise relies on the United Nations Comtrade database for export and import data at the 4-digit (harmonized Trade Mispricing and Illicit Flows 327 system [HS]) product level.17 The database contains detailed (annual) trade statistics reported by statistical authorities of close to 200 countries or territories and standardized by the United Nations Statistics Division. Among the records of shipments to the �ve largest importing nations in the world (China, Germany, Japan, the United Kingdom, and the United States), there are more than 1,200 product categories at the 4-digit level. We use the most recent commodity classi�cation (HS 2002); the data are available for �ve years, covering the period from 2002 to 2006. We begin by exploring the full sample of annual country pair-speci�c trade gaps at the 4-digit product level, that is, we compute, for each country pair and product, the percentage difference between reported imports and corresponding exports. Table 10.4 lists the �ve largest dis- crepancies in bilateral trade by importer, along with the exporter and the 4-digit product code. A few empirical regularities emerge from this rough tabulation. For instance, most experiences where recorded import values greatly exceed corresponding exports appear to be concentrated in a single product category, “petroleum oils, crude� (HS code 2709). As Yeats (1978) notes, this discrepancy is often associated with problems in valuing petroleum and the frequent diversion of petroleum exports from their original destination during transit. For other product catego- ries, in contrast, the export values (even though transportation costs are disregarded) are considerably larger than the imports in mirror statistics. These categories include “other aircraft (for example, heli- copters, aeroplanes), spacecraft� (8802), “cruise ships, excursion boats, ferry-boats, cargo ships, barges and similar vessels for the transport of persons or goods� (8901); and “gold (including gold plated with plati- num)� (7108). A possible explanation is that, especially for bulky items characterized by low-frequency trading, the time lag between exporta- tion and importation may be particularly important. Also, to the extent that there is any geographical pattern in misreporting, the overinvoicing of exports appears to be a more frequent problem in trade with neigh- boring countries. To analyze the geographical pattern in misreporting, we examine dif- ferences in trade gaps across countries in more detail. In particular, we aim to identify countries that consistently understate their exports (and, thus, appear to be particularly prone to trade mispricing or smuggling). 328 Table 10.4. Largest Trade Gaps, 2004 a. Underreporting of exports percent Draining Development? Importer United States Germany China United Kingdom Japan Exp. Prod. Gap Exp. Prod. Gap Exp. Prod. Gap Exp. Prod. Gap Exp. Prod. Gap SAU 2709 23.8 LBY 2709 22.0 PHL 8542 22.4 BWA 71 02 2 1.5 SAU 2709 23.4 VEN 2709 23.7 GBR 8803 2 1.2 AGO 2709 22.3 SAU 271 0 20.6 QAT 2709 22.4 NGA 2709 23.5 DNK 9999 2 1.2 SAU 2709 22.3 KWT 271 0 20.4 IDN 271 1 22.3 IRQ 2709 22.9 SAU 2709 20.7 OMN 2709 22.2 PHL 8542 20.3 KWT 2709 22. 1 AGO 2709 22.2 SYR 2709 20.7 IRN 2709 22.0 EGY 2709 1 9.7 ARE 2711 21.6 b. Overreporting of exports percent Importer United States Germany China United Kingdom Japan Exp. Prod. Gap Exp. Prod. Gap Exp. Prod. Gap Exp. Prod. Gap Exp. Prod. Gap DEU 8901 −1 9.9 CHN 8901 −20.3 HKG 8703 −20.8 USA 8803 −21.2 SWE 8802 −1 8.9 FIN 8901 −1 9.9 BEL 0803 −1 9.8 HKG 4101 −1 9.3 DEU 8802 −21.0 SGP 2204 −1 8.5 PRT 8802 −1 9.2 AUT 8901 −1 9.7 HKG 7108 −1 9.2 HKG 7108 −20.8 SGP 2208 −18.0 MEX 8602 −1 9.1 DNK 2716 −1 9.1 JPN 7108 −1 8.6 CAN 7108 −20.7 NZL 2709 −17.9 KOR 8901 −1 8.7 BLR 2709 −1 8.7 ARE 9999 −1 8.5 USA 8802 −20.7 BHR 7604 −1 7.6 Source: Author computation based on data of UN Comtrade (United Nations Commodity Trade Statistics Database), Statistics Division, Department of Economic and Social Affairs, United Nations, New York, http://comtrade.un.org/db/. Note: Exp. = exporter. Prod. = 4-digit HS product code. Gap = percentage difference between reported imports by the country listed in the top row and the corresponding exports reported by the country listed in the columns. Trade Mispricing and Illicit Flows 329 In a �rst exercise, we compute, for each exporter, the average trade gap across all products. Because there may be sizable product-speci�c differ- ences in reported trade values between the exporting and the importing country, taking the arithmetic mean of these reporting gaps over often hundreds of products is a simple way to identify (hopefully) country- speci�c differences in trade reporting. Table 10.5 lists the �ve countries with the largest average percentage share of missing exports by importer. As shown, we �nd, indeed, a strong and consistent mismatch in interna- tional trade statistics, with continuous underreporting, for instance, by Equatorial Guinea, Indonesia, and the Philippines. More importantly, reviewing the full distribution of exporting countries, we �nd that the extent to which countries tend to misreport exports is broadly similar across trade destinations. The correlation of exporter-speci�c average trade gaps across importing countries is astonishingly high, on the order of about 0.9. Table 10.6 reports a set of simple bivariate correlation coef�cients; Spearman rank correlations (unreported) provide similar results. These consistent patterns of misreporting in trade appear to pro- vide a useful basis for further research. In Berger and Nitsch (2012), for instance, we use regression analysis to examine the association between observed trade gaps and country- speci�c corruption levels. Holding constant a variety of other determi- nants of discrepancies in trade statistics, we �nd that the reporting gap in bilateral trade is, indeed, strongly associated with the level of corrup- tion, especially in the source country. In countries with corrupt bureau- cracies, it seems easier (and perhaps even common practice) to ignore legal rules and procedures. To the extent that this misbehavior also affects international trade transactions, our �ndings suggest that report- ing gaps in of�cial trade statistics partly reflect illegal activities for which the illicit movement of capital may be one motivation. Conclusion A potential vehicle to move capital unrecorded out of a country is the misinvoicing of international trade transactions. Exporters may under- state the export revenue on their invoices, and importers may overstate import expenditures, while their trading partners are instructed to deposit the balance for their bene�t in a foreign account. 330 Draining Development? Table 10.5. Underreporting of Exports by Country, 2002–06 percent Importer United States Germany China United Kingdom Japan All �ve Exporter Gap Exporter Gap Exporter Gap Exporter Gap Exporter Gap Exporter Gap Libya 14.5 Equatorial 12.1 Equatorial 14.9 Indonesia 12.1 Iraq 16.3 Equatorial 1 2.6 Guinea Guinea Guinea Lesotho 13.5 Indonesia 11.9 Congo, Rep. 13.5 Lao PDR 11.5 Equatorial 1 4.0 Indonesia 12.4 Guinea Indonesia 13.3 Ukraine 11.4 Congo, 1 2.7 Myanmar 1 1.5 Western Sahara 1 3.7 Philippines 1 1.6 Dem. Rep. Philippines 12.7 Philippines 11.2 Chad 11.8 Falkland Islands 11.2 Indonesia 13.2 Iraq 11.4 Iraq 12.5 Serbia and 1 1.0 Rwanda 1 1.8 Philippines 1 1.5 Botswana 1 2.8 Western Sahara 1 1.3 Montenegro Source: Author computation based on data of UN Comtrade (United Nations Commodity Trade Statistics Database), Statistics Division, Department of Economic and Social Affairs, United Nations, New York, http://comtrade.un.org/db/. Trade Mispricing and Illicit Flows 331 Table 10.6. Correlation of Exporter-Speci�c Average Trade Gaps Country United States Germany China United Kingdom Japan All United States 1.0000 Germany 0.9245 1.0000 China 0.8357 0.7824 1.0000 United Kingdom 0.9368 0.9571 0.7986 1.0000 Japan 0.90 15 0.8572 0.8963 0.8582 1.0000 All �ve importers 0.9700 0.9494 0.9 120 0.9564 0.9548 1.0000 Source: Author computation based on data of UN Comtrade (United Nations Commodity Trade Statistics Database), Sta- tistics Division, Department of Economic and Social Affairs, United Nations, New York, http://comtrade.un.org/db/. Note: The table is based on 202 observations. This chapter critically reviews empirical approaches to quantify the extent of trade mispricing. There are at least two sorts of problems. First, mispricing behavior is hard to identify. The analysis of discrepancies in bilateral trade statistics appears to be of generally limited value because gaps in trade statistics also typically arise for reasons unrelated to mis- pricing. The second set of issues refers to the motivations for fraudulent trade behavior. Even if mispricing is properly identi�ed, there are incen- tives for faking trade invoices other than the desire to transfer capital. Overall, the accuracy and reliability of estimates of IFFs based on trade mispricing are questioned. This �nding is in line with Bhagwati (1967, 63), who argues that, “whereas it is easy to establish the conditions under which the faking of trade values . . . will occur, it is in practice extremely dif�cult to set about determining whether such faking is actually occur- ring. It is further impossible to �nd out how much faking is going on.� Notes 1. This difference is based on agreed guidelines for international trade statistics as published by the United Nations (1998). Some countries, however, also report imports on an f.o.b. basis (for example, Australia). 2. Ruffles et al. (2003) provide a more detailed description. 3. Winston (1974, 64) argues that, “regardless of the sincerity of efforts, it is virtu- ally impossible to control overinvoicing considering the myriad ways it can, in fact, be done.� 4. For an early attempt, see the European Commission’s initiative Simpler Legis- lation for the Internal Market, which is documented at http://ec.europa.eu/ internal_market/simpli�cation/index_en.htm. 332 Draining Development? 5. Kar and Cartwright-Smith (2008, iii) argue that “by far the greater part of unre- corded flows are indeed illicit, violating the national criminal and civil codes, tax laws, customs regulations, VAT assessments, exchange control requirements and banking regulations of the countries out of which unrecorded/illicit flows occur.� 6. For instance, for some transactions, the avoidance of trade taxes rather than capital transfer may be the primary motivation. 7. For an early critical assessment of this approach, see Moneta (1959). 8. A more fundamental issue revolves around the fact that illicit flows through trade mispricing may not necessarily show up in trade invoices (for example, if there are unof�cial agreements). 9. Kar and Cartwright-Smith (2008) also apply other methods (such as the World Bank residual model) to derive estimates of IFFs. Since one term in the compu- tation of the World Bank residual is the current account balance, which may be distorted downward by trade mispricing, both results are added to obtain an aggregate estimate. 10. Direction of Trade Statistics (database), International Monetary Fund, Washing- ton, DC, http://elibrary-data.imf.org/FindDataReports.aspx?d=33061&e=170921. 11. Hummels and Lugovskyy (2006) note that, for New Zealand and the United States, episodes in which f.o.b. exports exceed corresponding c.i.f. imports con- stitute about one-third of the IMF data. 12. The data are obtained from the U.S. Merchandise Trade Databases; see USA Trade Online (database), U.S. Department of Commerce, Washington, DC, https://www.usatradeonline.gov/. 13. See the archive section at United States International Trade Commission, “Of�- cial Harmonized Tariff Schedule,� http://www.usitc.gov/tata/hts/index.htm. 14. See also the discussion in Fuest and Riedel’s contribution to this volume (chap- ter 4). 15. Gopinath and Rigobon (2008) examine monthly price data for approximately 20,000 imported goods and �nd that the (trade-weighted) median price dura- tion in the currency of pricing is 10.6 months. 16. This section draws on Berger and Nitsch (2012). 17. UN Comtrade (United Nations Commodity Trade Statistics Database), Statis- tics Division, Department of Economic and Social Affairs, United Nations, New York, http://comtrade.un.org/db/. References ABS (Australian Bureau of Statistics). 1998. “Bilateral Merchandise Trade Statistics Reconciliation: Australia and the European Union, 1992 to 1997.� International Merchandise Trade, Australia 5422.0 (September): 10–21, ABS, Canberra. Baldwin, R. E. 2006. “The Euro’s Trade Effects.� ECB Working Paper 594 (March), European Central Bank, Frankfurt. Trade Mispricing and Illicit Flows 333 Berger, H., and V. Nitsch. 2012. “Gotcha! A Pro�le of Smuggling in International Trade.� In Illicit Trade and the Global Economy, ed. C. Costa Storti and P. De Grauwe. CESifo Seminar Series. Cambridge, MA: MIT Press. Bhagwati, J. N. 1964. “On the Underinvoicing of Imports.� Oxford Bulletin of Eco- nomics and Statistics 27 (4): 389–97. ———. 1967. “Fiscal Policies, the Faking of Foreign Trade Declarations, and the Bal- ance of Payments.� Oxford Bulletin of Economics and Statistics 29 (1): 61–77. ———. 1981. “Alternative Theories of Illegal Trade: Economic Consequences and Statistical Detection.� Weltwirtschaftliches Archiv 117 (3): 409–27. Bhagwati, J. N., and B. Hansen. 1973. “A Theoretical Analysis of Smuggling.� Quar- terly Journal of Economics 87 (2): 172–87. Bhagwati, J. N., A. Krueger, and C. Wibulswasdi. 1974. “Capital Flight from LDCs: A Statistical Analysis.� In Illegal Transactions in International Trade: Theory and Measurement, ed. J. N. Bhagwati, 148–54. Studies in International Economics. Amsterdam: North-Holland. Celâsun, M., and D. Rodrik. 1989. “Debt, Adjustment, and Growth: Turkey.� In Country Studies: Indonesia, Korea, Philippines, Turkey, ed. J. D. Sachs and S. M. Collins, 615–808. Vol. 3 of Developing Country Debt and Economic Performance. National Bureau of Economic Research Project Report. Chicago: University of Chicago Press. Cumby, R. E., and R. M. Levich. 1987. “On the De�nition and Magnitude of Recent Capital Flight.� In Capital Flight and Third World Debt, ed. D. R. Lessard and J. Williamson, 26–67. Washington, DC: Peterson Institute for International Economics. Dabla-Norris, E., M. Gradstein, and G. Inchauste. 2008. “What Causes Firms to Hide Output? The Determinants of Informality.� Journal of Development Economics 85 (1–2): 1–27. de Boyrie, M. E., S. J. Pak, and J. S. Zdanowicz. 2004. “Money Laundering and Income Tax Evasion: The Determination of Optimal Audits and Inspections to Detect Abnormal Prices in International Trade.� Journal of Financial Crime 12 (2): 123–30. ———. 2005. “The Impact of Switzerland’s Money Laundering Law on Capital Flows through Abnormal Pricing in International Trade.� Applied Financial Economics 15 (4): 217–30. Dornbusch, R., and L. T. Kuenzler. 1993. “Exchange Rate Policy: Issues and Options.� In Policymaking in the Open Economy: Concepts and Case Studies in Economic Performance, ed. R. Dornbusch, 91–126. EDI Series in Economic Development. Washington, DC: World Bank; New York: Oxford University Press. Fisman, R., and S.-J. Wei. 2009. “The Smuggling of Art, and the Art of Smuggling: Uncovering the Illicit Trade in Cultural Property and Antiques.� American Eco- nomic Journal: Applied Economics 1 (3): 82–96. Gopinath, G., and R. Rigobon. 2008. “Sticky Borders.� Quarterly Journal of Economics 123 (2): 531–75. Gulati, S. K. 1987. “A Note on Trade Misinvoicing.� In Capital Flight and Third World Debt, ed. D. R. Lessard and J. Williamson, 68–78. Washington, DC: Peterson Institute for International Economics. 334 Draining Development? Herrigan, M., A. Kochen, and T. Williams. 2005. “Edicom Report: Analysis of Asym- metries in Intra-Community Trade Statistics with Particular Regard to the Impact of the Rotterdam and Antwerp Effects.� Statistics and Analysis of Trade Unit, HM Revenue & Customs, Southend-on-Sea, U.K. Hummels, D., and V. Lugovskyy. 2006. “Are Matched Partner Trade Statistics a Usable Measure of Transportation Costs?� Review of International Economics 14 (1): 69–86. IMF (International Monetary Fund). 1992. Report on the Measurement of Interna- tional Capital Flows. Washington, DC: Research Department, IMF. Javorcik, B. S., and G. Narciso. 2008. “Differentiated Products and Evasion of Import Tariffs.� Journal of International Economics 76 (2): 208–22. Kar, D., and D. Cartwright-Smith. 2008. “Illicit Financial Flows from Developing Countries, 2002–2006.� Global Financial Integrity, Washington, DC. http:// www.g�p.org/storage/g�p/economist%20-%20�nal%20version%201-2-09.pdf. Limão, N., and A. J. Venables. 2001 “Infrastructure, Geographical Disadvantage, Transport Costs, and Trade.� World Bank Economic Review 15 (3): 451–79. Moneta, C. 1959. “The Estimation of Transportation Costs in International Trade.� Journal of Political Economy 67 (1): 41–58. Morgenstern, O. 1950. The Accuracy of Economic Observations. Princeton, NJ: Prince- ton University Press. Naya, S., and T. Morgan. 1974. “The Accuracy of International Trade Data: The Case of Southeast Asian Countries.� In Illegal Transactions in International Trade: Theory and Measurement, ed. J. N. Bhagwati, 123–37. Studies in International Economics. Amsterdam: North-Holland. Ruffles, D., G. Tily, D. Caplan, and S. Tudor. 2003. “VAT Missing Trader Intra- Community Fraud: The Effect on Balance of Payments Statistics and UK National Accounts.� Economic Trends 597 (August): 58–70. United Nations. 1998. “International Merchandise Trade Statistics: Concepts and De�nitions.� Document ST/ESA/STAT/SER.M/52/Rev.2, Studies in Methods, Statistics Division, Department of Economic and Social Affairs, United Nations, New York. Walter, I. 1987. “The Mechanisms of Capital Flight.� In Capital Flight and Third World Debt, ed. D. R. Lessard and J. Williamson, 103–28. Washington, DC: Peterson Institute for International Economics. Winston, G. C. 1974. “Overinvoicing, Underutilization, and Distorted Industrial Growth.� In Illegal Transactions in International Trade: Theory and Measure- ment, ed. J. N. Bhagwati, 49–65. Studies in International Economics. Amster- dam: North-Holland. Yang, D. 2008. “Can Enforcement Back�re? Crime Displacement in the Context of Customs Reform in the Philippines.� Review of Economics and Statistics 90 (1): 1–14. Yeats, A. J. 1978. “On the Accuracy of Partner Country Trade Statistics.� Oxford Bul- letin of Economics and Statistics 40 (4): 341–61. Part IV Policy Interventions 11 Tax Havens and Illicit Flows Alex Cobham Many citizens of developing (and developed) countries now have easy access to tax havens and the result is that these countries are losing to tax havens almost three times what they get from developed countries in aid. If taxes on this income were collected billions of dollars would become available to �nance development. —Jeffrey Owens, Director, Centre for Tax Policy Administration, Organisation for Economic Co-operation and Development, January 2009 We will set down new measures to crack down on those tax havens that siphon money from developing countries, money that could otherwise be spent on bed nets, vaccinations, economic development and jobs. —Gordon Brown, Prime Minister, United Kingdom, March 2009 We stand ready to take agreed action against those jurisdictions which do not meet international standards in relation to tax transparency. . . . We are committed to developing proposals, by end 2009, to make it easier for developing countries to secure the bene�ts of a new cooperative tax environment. —G-20 Declaration, April 20091 Additional tables and �gures from this chapter can be found at http://go.worldbank.org /N2HMRB4G20. 337 338 Draining Development? Abstract Only limited research has been carried out on the links between tax havens and developing countries. This chapter provides a brief, critical survey of the state of knowledge on the impact of tax havens on develop- ment and then uses existing data to extend that knowledge by examining bilateral trade and �nancial flows between havens and developing coun- tries to identify the exposure of developing countries of different types. Three key results emerge. First, the chapter shows that the exposure of developing countries to tax havens is on a par with, if not more severe than, that of high-income countries of the Organisation for Economic Co-operation and Development (OECD). This supports efforts to ensure that developing countries bene�t from initiatives to require greater transparency, particularly in terms of international tax coopera- tion. Second, the differences in developing-country exposure across dif- ferent income groups of countries and regions are substantial, and rec- ognition of this must lead to more detailed and careful study and, over time, appropriate policy responses. Finally, the research has been repeat- edly blocked by a lack of high-quality, internationally comparable data. An agenda for research and for data collation and dissemination is pro- posed that would allow greater certainty of the scale of the impact of tax havens on development. Introduction In the wake of the �nancial crisis, a consensus emerged that interna- tional measures were required to limit the damage caused by tax haven secrecy and that developing countries must be included to ensure that these countries also bene�t. Research demonstrating the damage caused by havens, especially to developing countries, remains limited, however. If international policy efforts are to be well directed and, ultimately, to be effective in removing obstacles to development, then further work is needed. It is regrettable that policy research at multilateral institutions has almost completely neglected these issues until now. The research that has been carried out, at least until recently, when these issues began moving rapidly up the policy agenda, has largely been conducted by academics and civil society researchers. Tax Havens and Illicit Flows 339 This chapter sets out the key questions on this important research agenda that has now become an urgent one also for policy. There are three main areas requiring further analysis: the de�nition of tax havens, the broad development impact of tax havens, and the speci�c impact on particular developing countries in different regions and at various income levels. An additional issue that should be addressed urgently is the lack of data to examine these questions. The chapter addresses these questions in turn. We begin by consider- ing three related, but distinct concepts—tax havens, offshore �nancial centers, and secrecy jurisdictions—and the various de�nitions that have been offered for each. The lack of clarity in de�nitions has been a signi�- cant barrier to greater understanding, and it emerges from our discus- sion that the fundamental feature of these concepts relates not to the provision of low(er) tax rates, nor to the provision of �nancial services to nonresidents, but to the secrecy associated with each provision. The �nancial secrecy index (FSI) is therefore identi�ed as a potentially valu- able tool for further research. We then briefly survey the literature on the development impact of secrecy jurisdictions. This serves to highlight two main points. First, the key damage appears to stem from the functions attributed to secrecy jurisdictions, rather than to the functions of the alternatively de�ned entities discussed. Second, the literature identi�es a range of powerful arguments about the impacts, but does little to differentiate among developing countries and tends to treat the experience of these countries as homogenous. Greater evidence of direct causality is also required. The next section takes the �rst steps toward providing a more differ- entiated analysis of development links. By considering a range of bilat- eral flows and stocks in trade and �nance, we identify the key areas in which different types of developing countries are most exposed to the damage that relationships with secrecy jurisdictions can engender. The claim sometimes made that developing countries are less at risk than richer economies �nds no support in terms of the recorded scale of developing-country exposure. In addition, important differences in exposure emerge among developing countries in different regions and at different income levels. This chapter does not set out to prove causal links between the expo- sure to secrecy jurisdictions and damage to development. It does, how- 340 Draining Development? ever, close by setting out a clear agenda for future research that would address the need for more information on these links. Without such research by academics and policy researchers not only at nongovern- mental organizations, but also at international �nancial institutions, there can be no guarantee that the well-meaning efforts of the G-20 or other international policy coordination groups will generate the poten- tially substantial bene�ts for development. The requirement for more research must be addressed without delay. Tax Havens: Offshore Financial Centers or Secrecy Jurisdictions? To undertake serious analysis of the impact of tax havens, the analyst must assert a speci�c de�nition that is objectively quanti�able and directly related to the harm believed to occur. This is necessary if the impact of those jurisdictions designated as tax havens is to be evaluated statistically. On this issue, there is a broad literature focusing on aca- demic interests, advocacy, and of�cial policy positions and encompass- ing a range of views. De�nitions The most common term—tax haven—is probably also the most prob- lematic. As long ago as 1981, the Gordon Report to the U.S. Treasury found that there was no single, clear objective test that permits the iden- ti�cation of a country as a tax haven (Gordon 1981). While originally intended, presumably, to indicate a jurisdiction with lower tax rates than elsewhere, the term came to be used to cover jurisdictions with a great range of functions, many largely unrelated to taxation. Eden and Kudrle (2005), for example, draw on the literature to identify two categories of havens: one based on type of taxation, and one based on activity. The �rst category, following Palan (2002), separates havens into “countries with no income tax where �rms pay only license fees (e.g., Anguilla, Bermuda), countries with low taxation (e.g., Switzerland, the Channel Islands), countries that practice so-called ‘ring-fencing’ by tax- ing domestic but not foreign income (e.g., Liberia, Hong Kong), and countries that grant special tax privileges to certain types of �rms or operations (e.g., Luxembourg, Monaco)� (Eden and Kudrle 2005, 101). Tax Havens and Illicit Flows 341 The second category, following Avi-Yonah (2000) and Kudrle and Eden (2003), distinguishes among production havens, which relocate real value added (for example, Ireland in the 1990s); headquarters havens, which provide incorporation bene�ts (for example, Belgium and Singapore); sham havens, which provide little more than addresses for �nancial companies in particular (for example, Cayman Islands); and secrecy havens, the main advantage of which is opacity (and which include most sham havens). Analysis under the heading tax haven tends to focus, understandably, on tax aspects. This view is most commonly associated with the OECD. While an earlier report (OECD 1987) focused on reputation (“a good indicator that a country is playing the role of a tax haven is where the country or territory offers itself or is generally recognised as a tax haven� [cited in OECD 1998, 21]), there is somewhat more precision in the 1998 report. Speci�cally, the 1998 report emphasizes no or only nominal taxes as the starting point for the identi�cation of a tax haven, but allows, in addition, lack of an effective exchange of information, lack of transpar- ency, and no substantial activities as further key factors. The overarching rationale for the existence of tax havens that emerges from this approach is the provision of relief to businesses or individuals from the rates of tax that apply elsewhere. If real economic activity (in substance) is not moved to a new location from the original jurisdiction, then taxing rights have to be transferred by other means (manipulation of the form). This may involve taking advantage of genuine legal differ- ences (for example, the distinction between the tax liabilities of corpo- rate headquarters and the tax liabilities associated with locations where real economic activity takes place), of the absence of coordinated inter- national tax policy (for example, exploiting differences between the calculation of domestic and foreign tax liabilities in two or more juris- dictions), or of asymmetric information across jurisdictions (for exam- ple, hiding information about the true ownership of assets or income streams and, therefore, responsibility for the associated tax liabilities). The second frequently used term is offshore �nancial center (OFC). This term is preferred, for example, by the International Monetary Fund (IMF), the mandate of which is more closely aligned to issues of interna- tional �nancial regulatory oversight and stability than to issues of tax. 342 Draining Development? Palan explores some of the dif�culties of consistent de�nition in this case, as follows: In �nancial literature, . . . offshore is used . . . to describe unregulated international �nance. . . . Rather confusingly, however, the International Monetary Fund and the Bank for International Settlements consider only tax havens as Offshore Financial Centres, though the City of London, which does not qualify as a tax haven, is considered the hub of global offshore �nance. (Palan 1998, 64) Palan goes on to distinguish between spontaneous OFCs, which have grown up as entrepôts over time, such as the City of London and Hong Kong SAR, China, and international banking facilities that have been more recently created as part of a deliberate strategy, such as New York and Singapore. An important IMF Working Paper by Ahmed Zoromé (2007) dis- cusses the de�nitional issues in some detail and proposes new criteria (discussed in the following subsection). After surveying most of the key references, Zoromé concludes that “three distinctive and recurrent char- acteristics of OFCs have emerged from these de�nitions: (i) the primary orientation of business toward nonresidents; (ii) the favorable regula- tory environment (low supervisory requirements and minimal informa- tion disclosure); and (iii) the low- or zero-taxation schemes� (Zoromé 2007, 4). This shifts the focus onto speci�c actions taken by jurisdictions, whereas other de�nitions tend to emphasize speci�c results. Zoromé’s subsequent application of these criteria offers one solution to Palan’s point about the City of London: Zoromé classi�es the United Kingdom as an OFC. The third main term used—increasingly so since it was promoted by Murphy (2008)—is secrecy jurisdiction. The focus remains on speci�c actions taken, but is more explicit in emphasizing the legal steps. In this, it follows the logic of Palan (2002), who discusses the commercialization of sovereignty: the decision by certain jurisdictions to obtain economic advantage by allowing selected political decisions (over, for example, the taxation of nonresidents) to be dictated by likely users (for example, �nancial, legal, and accounting practitioners). The term jurisdiction is therefore used, rather than, for example, center or some other less spe- Tax Havens and Illicit Flows 343 ci�c term, because it is precisely the legal system that is the locus of the speci�c actions. The emphasis on secrecy is necessary, Murphy argues, so that nonresi- dents can take advantage of favorable changes in the jurisdiction’s legal framework with the con�dence that they will not fall foul of the legal system in the place where they reside. There are thus two key character- istics that de�ne a secrecy jurisdiction: • “The secrecy jurisdiction creates regulation that they know is primarily of bene�t and use to those not resident in their geographical domain� • “The creation of a deliberate, and legally backed, veil of secrecy that ensures that those from outside the jurisdiction making use of its regulation cannot be identi�ed to be doing so.� (Murphy 2008, 6)2 By focusing on what makes them attractive, the secrecy jurisdiction concept therefore relies, above all, on an assessment of the comparative advantage of the jurisdictions in question. The route the secrecy juris- dictions have chosen to attract (the declaration of) foreign economic or �nancial activity is the provision of relatively favorable terms to users. In effect, this indicates a reliance on regulatory arbitrage (potentially, but not necessarily including tax regulation).3 Because some arbitrage is inevitable in the absence of comprehensive global standards, this also runs the risk of excessive breadth; the damaging element relates to arbi- trage that is effective because it frustrates the intentions behind regula- tory standards in other jurisdictions in regard to activities that remain substantially in those other jurisdictions. Criteria and approaches The success of attempts to identify the jurisdictions of concern has been mixed. Table 11A.1 reproduces a set of lists compiled by Murphy in chapter 9.4 The table outlines most of the main efforts by academics and institutions over the last 30 years. For the most part, these have drawn on de�nitions that, while generally made explicit, lack a basis in objective, measurable criteria. To consider how such criteria might be created and applied, it is use- ful to start with Zoromé (2007). Having established key characteristics of OFCs, Zoromé proposes a de�nition, as follows: “an OFC is a country or jurisdiction that provides �nancial services to nonresidents on a scale 344 Draining Development? that is incommensurate with the size and the �nancing of its domestic economy� (Zoromé 2007, 7). He proposes to identify such OFCs by examining the ratio of net �nancial service exports to gross domestic product (GDP) from IMF balance of payments data and by looking at jurisdictions with especially high values.5 In practice, the limited data availability on these flows means that the data search must be supplemented with data interpolated from stock variables using proxy indicators created from data on vari- ables reflecting crossborder holdings of portfolio investment assets (the IMF Coordinated Portfolio Investment Survey) and international invest- ment positions (using data from International Financial Statistics).6 The key difference between the IMF’s existing list and Zoromé’s �nd- ings can be seen by comparing the columns denoted “IMF (2000)� and “Zoromé (2007),� respectively, in table 11A.1. Most of the de�nitions coincide (for those jurisdictions for which Zoromé had data), but an important addition to the latter’s list is the United Kingdom. While there are some issues with the proxy indicators and the interpolation from stock to flow data, the approach illustrates clearly the value of using objective criteria: a level playing �eld (including politically uncomfort- able �ndings) may be more likely to emerge. In keeping with the emphasis on the OFC discourse, Zoromé (2007) relies on the relative intensity of the provision of �nancial service to nonresidents by scaling for jurisdictional GDP. Alternative criteria include the following: • Openness to international trade in �nancial services (that is, taking the sum of exports and imports as a ratio to GDP, which might cap- ture more about the role of jurisdictions as conduits) • Net exports of �nancial services (exports, minus imports as a ratio to GDP), as an indicator of specialization • Netted trade in �nancial services (the lower of exports and imports, as an indicator only of the extent to which the jurisdiction acts as a conduit) • The absolute contribution to the global provision of �nancial services to nonresidents (that is, each jurisdiction’s provision of �nancial ser- vices to nonresidents as a ratio to the total global provision of services to nonresidents across all jurisdictions, rather than as a ratio to the jurisdiction’s own GDP) Tax Havens and Illicit Flows 345 If we are looking for a measure to capture the relative importance of jurisdictions so as to consider the appropriate policy response at a global level, this last criterion may be appropriate. Table 11A.2 com- pares the results of the use of the relative approach of Zoromé and the use of the absolute approach, showing the top �ve jurisdictions identi- �ed by each approach.7 Taking global contribution rather than relative intensity in the provision of �nancial services to nonresidents leads to quite a different picture: per 2007 data, the former criterion points to Cayman Islands, Luxembourg, Switzerland, the United Kingdom, and the United States, while the latter points, instead, to Bermuda, Cayman Islands, Guernsey, Jersey, and Luxembourg. Finally, however, we may consider the implications of the secrecy jurisdiction approach. While the relative or absolute approach to off- shore specialization may capture something about the success of juris- dictions in following this path, it has nothing directly to say about the secrecy or otherwise with which this has been achieved. The ideal objective criteria for the identi�cation of secrecy jurisdic- tions might therefore contain two separate components: one reflecting each jurisdiction’s importance in the global provision of �nancial ser- vices to nonresidents (the absolute contribution approach) and one reflecting each jurisdiction’s (objectively measurable) performance against one or more key indicators of secrecy. A �nal advantage of such a composite approach is that it would allow users to step away from reliance on lists, which, by necessity, dictates that a given jurisdiction either does or does not meet certain criteria. More powerful both for policy making and for research might be a measure that falls on a sliding scale. A method of measuring progress that is more nuanced than a blacklist may eventually produce more positive responses; see, for example, Kudrle (2008, 16), who, on the limitations of blacklist- ing, writes that “the evidence does not suggest that blacklisting made an important systematic difference.� The FSI jointly produced by Christian Aid and the Tax Justice Net- work reflects this analysis.8 The FSI combines a secrecy score based on objectively measurable criteria that reflect the secrecy or otherwise of the jurisdiction in key areas (for example, banking secrecy, the exchange of tax information) with a quantitative measure of the absolute contribu- tion as discussed above, here labeled the global scale weight (table 11A.3).9 346 Draining Development? By combining measures of secrecy and measures of the importance to global offshore �nance, the index presents a picture that is different from the one commonly envisaged. First, many of the usual suspects do not feature as signi�cantly as they do in policy discussions. For example, Guernsey does not make the top 20, while Monaco ranks only 67th. Among developing countries, many oft-named Caribbean jurisdictions also rank low; thus, for example, Antigua and Barbuda ranks 58th, and Grenada 61st. At the same time, many other usual suspects do feature at the top of the index; for example, Switzerland, Cayman Islands, and Luxembourg occupy the top three places, in that order. However, the top �ve also include the United States, despite powerful statements in recent years about the damage done by �nancial opacity. The United Kingdom and the United States each account for around a �fth of the global total of �nancial service provision to nonresidents, but the United Kingdom ranks only 13th because of a substantially better secrecy score. It seems uncontroversial that a measure of the global damage caused by �nancial secrecy would include measures equivalent to both the secrecy score and the global scale weight. Less likely, however, is agree- ment on the relative importance of the two in establishing a �nal rank- ing. The value of the FSI resides less in providing a categorical �nal ranking than in focusing the attention of policy makers on the two ele- ments, instead of on the opacity measures alone, which has historically supported a perhaps unjusti�ed concentration on smaller, less econom- ically and politically powerful jurisdictions. From the global perspec- tive, strong opacity in a major player may do more damage than com- plete secrecy in a tiny one. A particular concern is that the FSI may exclude the possibility that no jurisdictions legitimately offer high-quality banking, accounting, and legal services, but also rely on a transparent regulatory framework. The United Kingdom is often cited by those who have such concerns because it is said to have a genuine comparative advantage that does not depend on lax regulation. This chapter does not offer the space to explore this issue in detail, but the role of the United Kingdom’s network of tax havens in the development of London as a key �nancial center is cer- tainly worth considering (Shaxson 2011).10 Establishing appropriate counterfactuals about the development of �nancial centers, absent opac- Tax Havens and Illicit Flows 347 ity, especially in the presence of signi�cant agglomeration bene�ts in �nancial services, is inevitably challenging. In the penultimate section of this chapter, we use the existing list approach to identify links between developing countries and secrecy juris- dictions. However, we believe the FSI has the potential to provide valuable new insights on the issue of links, and the relevant research is under way. Econometric analysis of the differential impact of commercial and �nan- cial links with more or less extreme secrecy jurisdictions may offer much more powerful and illuminating results, and we explore some suggestions here. Additionally, an anonymous referee has suggested the value of creat- ing sub-FSI indexes that reflect the relative importance of jurisdictions in speci�c types of �nancial services, which would allow analysis of the ben- e�ts or damage attributable at this more granular level. Data constraints are a major obstacle, but such an analysis, if possible, would be revealing. The terms tax haven and secrecy jurisdiction are used interchangeably hereafter. Development Impact Before assessing the extent of the links between development and secrecy jurisdictions, we survey the literature on the potential development impact. Our intention is to establish the key relationships that require empirical testing. The literature is unfortunately somewhat limited; it has two general characteristics: a focus on developed-country impact, accompanied by the study of havens, and a general neglect of tax as a development policy issue (see Cobham 2005a on the latter). A recent assessment by an expert commission reporting to the Norwegian government represents a useful step forward in outlining the key claims and much of the existing evi- dence.11 Seven major claims are considered, the �rst three of which (at least) will affect all countries, while the other four may be especially rel- evant to developing countries: 1. Tax havens increase the risk premium in international �nancial markets. 2. Tax havens undermine the functioning of the tax system and public �nances. 348 Draining Development? 3. Tax havens increase the inequitable distribution of tax revenues. 4. Tax havens reduce the ef�ciency of resource allocation in developing countries. 5. Tax havens make economic crime more pro�table. 6. Tax havens can encourage rent seeking and reduce private incomes in developing countries. 7. Tax havens damage institutional quality and growth in developing countries. Of these, 1 relates to systematic global �nancial risk, 2–4 relate pri- marily to the damage done to effective tax systems, 5 relates to the dam- age done through crime, and 6 and 7 relate to distortions to economic activity, including corruption. We treat these in turn and also the major positive claim that tax havens can encourage investment in developing countries. Aside from the systemic stability issue, the Norwegian commission’s breakdown is in line with the most widely quoted analysis of illicit glo- bal �nancial flows, which suggests that the volume of outflows from developing countries reaches US$500 billion–US$800 billion a year (Baker 2005; subsequently revised upward by Kar and Cartwright-Smith 2008). Of this amount, 60–65 percent is estimated to be associated with commercial tax evasion (overwhelmingly in the form of mispricing and misinvoicing in trade both between unrelated parties and within multi- national groups), 30–35 percent to the laundering of criminal proceeds (for example, drug and human traf�cking), and around 3 percent to the corruption of public of�cials. The Norwegian commission summarizes these important estimates and a good deal of additional literature that deals with the general scale of illicit capital flows. We do not treat this broader literature in more detail here, but concentrate only on those contributions that speci�cally address the role of tax havens. The subsections below survey the litera- ture on the main areas of development impact, considering efforts at quanti�cation and broader arguments about the channels of impact. Systemic �nancial risk Christian Aid (2008) advances the argument that the �nancial crisis was, effectively, a capital account liberalization bust, following a classic boom Tax Havens and Illicit Flows 349 (see Williamson and Mahar 1998). Over a quarter of a century or more, the richer economies engaged in competitive deregulation of �nancial markets. Secrecy jurisdictions took this to an extreme, exploiting the gaps left by the national regulation of global �nance. The key regulatory arbitrage occurred around the regulation of banks and other �nancial institutions, particularly regulation that, to protect depositors from undue risk, limits the amount of assets banks may acquire as a proportion of their own capital base. The Basel II capital accord allows assets of US$1,250 for each US$100 of capital. In merely one example of an excess, Stewart (2008) has discovered that the Irish holding company of the now-collapsed U.S. bank Bear Stearns held US$11,900 in assets for each US$100 of capital. The late-2011 collapse of broker-dealer MF Global revealed a ratio in excess of 80:1 (Christian Aid 2011). The crisis was driven by two key factors that allowed an unsustainable expansion in credit. One was the complexity of the new �nancial instru- ments, which confused investors and regulators about the true ownership of assets and liabilities. The other was the opacity of the shadow banking system, that is, �nancial activities outside the traditional banking system, from hedge funds and private equity to the structured investment vehicles and other conduits of investment banks and others, all taking advantage of regulatory arbitrage to operate through secrecy jurisdictions. U.S. Treasury Secretary Timothy Geithner (then president of the New York Federal Reserve Bank) has estimated that only a portion of these activities exceeded the total assets of the entire U.S. banking system: “Financial innovation made it easier for this money to flow around the constraints of regulation and to take advantage of more favorable tax and accounting treatment� (Geithner 2008, 1). The genesis of the crisis was undoubtedly complex, and attempts to strictly apportion blame are likely ultimately to be futile. It does seem clear, however, that the role of secrecy jurisdictions was nonnegligible at the least. An important policy response to the speci�c problems posed by reg- ulatory arbitrage was the decision by G-20 �nance ministers in September 2009 to call for “the setting of an overall leverage ratio, which will take into account off-balance-sheet activities, in order to limit the amount of bor- rowing conducted by institutions outside the formal banking regime.�12 While researchers will continue for many years to reexamine the gen- esis of the crisis and the long period of deregulation and liberalization 350 Draining Development? that underpinned the crisis, a successful and sensible attempt to quantify the role of tax havens is dif�cult to imagine. Yet, without such a quanti- �cation, a credible estimate of the cost of tax havens to developing coun- tries through increased systemic �nancial risk is also unlikely. It does not follow, however, that this channel of impact should be ignored. Tax systems The most high-pro�le and clearly quanti�ed claims made about the damage done by tax havens to developing countries relate to the undermining of tax systems, both in the ability of these systems to deliver the revenues needed for the provision of public services and for growth-enhancing investment in, for example, infrastructure and high-quality administration and in their ability to meet social prefer- ences for redistribution. The importance of tax systems for development is becoming more widely recognized in policy-making circles, although much remains to be done. Ultimately, tax can make a great contribution to the four Rs: most obviously, revenues and redistribution, but also the repricing of economic and social “bads� and political representation (Cobham 2005b; on the important links between taxation and governance, see Bräutigam, Fjeldstad, and Moore 2008). There are multiple reasons why tax systems in developing countries fail to deliver these contributions, as follows: • A damaging tax consensus pushed by donors and the IMF, in particu- lar (Heady 2004; Cobham 2007) • Lack of capacity in domestic tax administrations • Lack of engagement by civil society to hold governments to account for tax policies and practices • The opacity of corporate accounting (on the scale of pro�t shifting within Europe, see Huizinga and Laeven 2008; on the scale of trade mispricing and revenue losses in the developing world, see Christian Aid 2009; on a key policy option, see Richard Murphy’s contribution, • chapter 9, to of volume) The opacity this secrecy jurisdictions in facilitating pro�t shifting, but also in facilitating the underdeclaration of income or assets by individuals13 Tax Havens and Illicit Flows 351 Note that these issues are largely independent of the additional prob- lem of the smaller per capita tax base in developing countries because of lower per capita GDP, while the problems associated with large informal sectors are primarily a result, rather than a cause, of these other issues (see the discussion on compliance below). There have, as yet, been no serious attempts to estimate the cost of secrecy jurisdictions in terms of forgone redistribution, repricing, or representation. Useful attempts have, however, been made to address the revenue effects. The systemic impact of tax havens on development was �rst addressed not by researchers at international �nancial institutions, but by a non- governmental organization, which relied on academics and haven insid- ers to build its analysis. The report, by Oxfam (2000), included an esti- mate that developing countries were losing around US$50 billion a year to tax havens. This estimate draws on global �gures for foreign direct investment and the stock of capital flight, combining these with esti- mated returns to investment and interest income, along with estimated tax rates: the sum is around US$35 billion in untaxed foreign direct investment and US$15 billion in untaxed personal income. Subsequent work by the Tax Justice Network (TJN 2005) suggests that the global revenue loss to the untaxed savings incomes of individuals may be as much as US$255 billion annually, and Cobham (2005b) notes that proportions equivalent to the shares of world GDP would imply a loss to developing countries of around US$50 billion a year. As Fuest and Riedel (2009) clarify, this would imply a potential tax haven total of US$85 billion a year, which is close to the total of aid received for the period in question. A subsequent Oxfam study puts the estimated loss on individual income alone at US$64 billion–US$124 billion (Oxfam 2009). More recently, Ann Hollingshead (2010) has estimated that US$98 billion to US$106 billion is the developing-country tax loss on some form of the trade mispricing that Christian Aid assesses and notes that her Global Financial Integrity study therefore well supports the work of Christian Aid. This is the extent, currently, of attempts to quantify the revenue dam- age done by havens to developing countries. Two main criticisms can be made. First, the assumptions involved are necessarily nontrivial. Global estimates of asset return, for example, are open to criticism because 352 Draining Development? opacity shields both the likely motivations and the actual portfolio choices of those holding assets in secrecy jurisdictions. Criticisms of this nature have not, however, been made, in general, by scholars offering superior alternative approaches. The second major criticism of these approaches is that no reasonable counterfactual is put forward. Imagine, for example, that there are two main channels for illicitly reducing tax liabilities in developing coun- tries. If the second is, say, 95 percent as effective as the �rst, then shutting down the �rst is likely to yield only a small revenue bene�t. In a world with multiple channels of illicitly reducing tax liabilities, it is unrealistic to expect a de�nition of an appropriate counterfactual to estimate the bene�t of closing down any particular channel. In effect, however, this criticism applies not to the quanti�cation approaches, but to a popular interpretation that the resulting estimates represent readily available revenues. In addition, such assessments of the impact of secrecy jurisdictions do not do justice to the impact on tax compliance more broadly. The experimental economics literature strongly suggests that compliance is heavily influenced by the perceptions of the compliance of others, so that high-pro�le noncompliance—fueling, for example, the perception that all rich citizens or multinational companies are hiding their incomes and pro�ts in havens—is likely to have substantial multiplier effects on compliance and revenue throughout a whole economy and over the long term, given how slowly compliance is observed to shift in practice (Bosco and Mittone 1997; Mittone 2006). While re�nements in these approaches can be envisaged, the key con- tribution of the literature on revenue losses has been to establish that there exists an obstacle to the development of suf�cient scale in esti- mates to warrant serious attention from civil society and from policy makers. It is not immediately clear at this stage what additional value could be offered to the policy debate by alternative estimates of global scale; more nuanced work, as suggested in the sections below, may be more valuable now. Few would argue that the United Nations Development Programme’s human development index is a truly appropriate measure of human development in all its complexity. The index has, however, undeniable bene�ts in shifting the emphasis in media reporting and policy discus- Tax Havens and Illicit Flows 353 sions on poverty away from unhelpfully narrow GDP per capita mea- sures and toward a broader and more human-level analysis. A similar argument may apply to Oxfam’s seminal 2000 report. The value of the report in drawing attention to the issue is certain, not least by catalyzing the emergence of the Tax Justice Network, which has played a central role in raising awareness among policy makers, civil society, and the public. A number of potential approaches to quantifying the scale of revenue losses on a country-speci�c basis are worth consideration, although they suffer from certain weaknesses also. One approach would involve pursu- ing the Huizinga and Laeven method (2008) using data from multina- tional companies on the performance of their subsidiaries to estimate the extent of pro�t shifting on a global basis, rather than only in Europe. The data have been questioned by those more familiar with corporate accounting (see, for example, Richard Murphy’s contribution in this vol- ume, in chapter 9), but are standard in the corporate �nance literature; so, they may at least provide some benchmark estimate of the revenue losses occurring speci�cally because of the behavior of multinational companies. The other avenue available is to use data on the �nancial positions of secrecy jurisdictions with regard to speci�c developing countries and, given the relative secrecy and administrative capacity of each, to make estimates about the likely extent of returns and income declarations. Current political processes through the G-20 and bilaterally mean that more data will become publicly available about the secrecy jurisdiction location of hidden assets and income streams (of richer economies at least) and, importantly, about the actual forgone tax yields. These data will provide a practical backstop to ensure that estimates for developing- country revenue losses are demonstrably anchored in reality. Neither approach seems entirely satisfactory from an academic stand- point. However, as long as secrecy is the issue under study, any impact assessment will be subject to considerable margins of error. In the longer term, a more promising avenue may be to work toward a better understanding of the nature of the problem, rather than speci�c quanti�cations of the scale in revenue terms alone. In particular, if a consistent indicator is available to capture the secrecy of jurisdictions, then it will be possible to use panel data regression analysis to estimate 354 Draining Development? the various impacts of economic and �nancial links between secrecy jurisdictions and developing countries. (This is explored in greater detail in the following sections.) Potentially, this would allow the assessment not only of revenue effects, but, for example, also of the impact on the extent of redistribution, channels of political representation, standards of governance, and so on. An additional reason to consider broader impacts, even if the focus is revenue, is a point stressed by the Norwegian commission: tax treaties between havens and developing countries are typically negotiated with the former in a position of strength, and, as a result, the havens receive a transfer of taxing rights from the developing countries, which hope that greater investment (and subsequent economic bene�ts) will follow. The three hypotheses that make up this argument should be the sub- ject of empirical scrutiny. First, do revenue patterns reflect such effects as a result of newly signed treaties? Second, do investment patterns respond as predicted? Third, does such investment deliver economic bene�ts? (If so, it would be interesting to consider whether the bene�ts are suf�cient to offset any revenue losses discovered in the shorter term.) Governance We will not revisit the range of arguments on the importance of effective taxation for improvements in governance and channels of political rep- resentation (see Christian Aid 2008 and chapter 3, by Everest-Phillips, in this volume for more details, including considerable empirical evidence). Nor do we consider the generic arguments about the contribution of illicit flows to making crime pay (see chapters 6 and 12, by Kopp and Levi, respectively, on aspects of money laundering and crime). The spe- ci�c role of tax havens with regard to the laundering of the proceeds of crime is relatively well explored, not least because it has been the focus of some of the more effective international coordination efforts; for exam- ple, see Hampton and Christensen (2002) for a discussion of the estab- lishment, in 1989, of the intergovernmental Financial Action Task Force and its subsequent impact in terms of promoting anti–money launder- ing legislation at least in havens.14 There is, however, little academic analysis of the links between the opportunities provided by havens and the damage done to developing countries. Perhaps the main innovation offered by the Norwegian com- Tax Havens and Illicit Flows 355 mission is in this area. Torvik (2009) goes beyond the generic arguments about tax and governance to make a speci�c case against havens by drawing parallels with the well-established literature on the paradox of plenty (the poor economic performance often associated with a coun- try’s natural resource wealth). Torvik argues that havens distort develop- ing countries, above all, by changing incentives. First, the balance between productive activity and rent-seeking behavior is pushed toward the latter because of the availability of secrecy, which allows the capture of higher returns to rent seeking. This process can occur in any context, but may be particularly potent in the presence of natural resources. The secondary effects may be even more powerful because they relate to the institutional and political context. Most narrowly, the availability of haven services can support the undermining or disestablishment of institutions by politicians so as to conceal corruption and �nancial mal- feasance generally. Furthermore, the incentives of politicians to achieve a shift toward political systems that make elite capture easier (that is, presidential rather than parliamentary systems) will also be strength- ened by haven opportunities. Finally, if havens support the returns to narrow self-interest over more broadly based progress, Torvik argues that they may increase the chances of conflict and contribute to the weakening of democratic processes. Torvik summarizes his conclusions thus: The negative effects of tax havens are greater for developing countries than for other countries. There are many reasons for this. Reduced gov- ernment income will have a greater social cost for developing countries than for industrialized countries. In addition, other mechanisms make themselves felt in countries with weak institutions and political systems. (Torvik 2009, 188) While the theoretical arguments set out are compelling, what remains unavailable at this point is empirical veri�cation of the haven links in the series of hypotheses that connect natural resource wealth, political sys- tems, and institutional quality with particular outcomes. Research should build on the valuable insights of Torvik (2009) and test a range of hypoth- eses in two directions: to assess the impact of natural resource wealth, political systems, and institutional quality on the extent of economic links with tax havens and to assess the impact of economic links with tax havens 356 Draining Development? on the growth-reducing effects of natural resource wealth on the develop- ment of political systems and on institutional quality. Haven bene�ts A common argument made by defenders of offshore �nance is that the key service provided by secrecy jurisdictions is not secrecy, but a tax- neutral platform from which investors based in a range of jurisdictions, with potentially quite different tax treatments, can join to make an investment in another, particular jurisdiction. The tax neutrality is gen- erally claimed to represent a safeguard against double taxation, rather than a tool for avoiding or evading taxation altogether. Internet searches for “tax-neutral platform� identify those jurisdic- tions that most commonly make this claim, but a somewhat more ana- lytical approach is offered by the Big Four accounting �rm Pricewater- house Coopers, which proposes a set of criteria for an investment platform (in a brie�ng recommending Ireland as such), including that it have an “attractive tax regime for holding and �nancing activities� and “an attrac- tive treaty network and low or zero statutory withholding tax rates on interest and dividend flows� (Arora, Leonard, and Teunissen 2009, 1).15 Leaving aside the question of whether such a platform is likely to facilitate or encourage the avoidance or evasion of single taxation, the hypothesis that follows from such claims is that the existence of havens makes investment in other jurisdictions more pro�table. The Orbis database on the accounts of the subsidiaries of multinational companies would allow a partial testing of this hypothesis using the Huizinga and Laeven approach (2008), but this would miss the phenomenon of �nan- cial vehicles relying on multiple investors in different jurisdictions.16 Because the opacity of such jurisdictions and �nancial vehicles mili- tates against direct testing in this case, a less direct hypothesis may be needed. If pro�tability is, indeed, raised by the use of haven platforms, then the level of investment should respond accordingly and be higher where havens are involved. One paper claims to have tested this hypothesis: Desai, Foley, and Hines (2006a). Using results from their longer paper (2006b), the authors offer a different interpretation, which, they claim, shows that “careful use of tax haven af�liates permits foreign investors to avoid some of the tax burdens imposed by domestic and foreign authorities, thereby main- Tax Havens and Illicit Flows 357 taining foreign investment at levels exceeding those that would persist if tax havens were more costly� (Desai, Foley, and Hines 2006a, 223). The model is dismissed by the Norwegian commission. The model is based on “the assumption that an investor can make real investments with a real level of activity in a tax haven,� the commission states on page 72 of its report.17 “In fact, foreigners who use the preferential tax regime are not permitted to invest locally, have local employees or use the coun- try’s currency. The Commission accordingly takes the view that the assumption underlying the analysis is based on ignorance of investor regulations in tax havens.� The empirical results can be criticized because they are based on a lim- ited sample of U.S. multinationals and therefore, as indicated above, may not capture the main bene�ts claimed. Certainly, the results are insuf�- cient to support the claims made for the bene�ts of tax havens. As with the claims about damage discussed above, they require closer scrutiny. Further research should assess more carefully whether any causal relationship can be established between haven origin investment and (1) the total level of inward investment and (2) the total level of invest- ment from any source (because round-tripping may lead to arti�cially inflated declarations of inward investment with no necessary associated increase in actual investment levels). Additional research should also examine whether investment received through havens has different eco- nomic and political impacts relative to other investment flows. Economic and Financial Links There are three main issues in the emerging research agenda on the links between havens and development: �rst, that assessments of the extent of these links are lacking; second, that empirical veri�cation is needed of the key claims surveyed above about the impact of these links; and, third, that little work has been done to analyze whether different developing countries are affected differently, which seems to be a key issue, not least for policy makers, in prioritizing responses. This section lays out some evidence relating to the �rst issue, the extent of the links, and highlights additional sources that can be tapped for future research. The data pre- sented will also form the basis for econometric analysis to address the second issue, the extent of the development impact of the links. 358 Draining Development? The third issue is perhaps the most interesting. One aspect of dealing with it would involve assessing whether the links vary quanti�ably across developing countries and whether there are systematic differences accord- ing to income group or region. Some attention is given to this question here. Another aspect, which will remain for future research, is whether the strength of the impact of the links differs from country to country, whether, for example, the impact of haven access in undermining gover- nance is strongest in low-income countries, natural resource countries, countries with presidential systems, South Asian countries, and so on. Two broad sets of competing hypotheses could usefully be examined: �rst, whether lower-income countries are generally less closely linked to and less affected by havens because their institutions, including their tax systems, are suf�ciently weak that haven structures are not required to facilitate rent seeking, for example, by reducing tax payments (or avoid- ing other regulation), and, second, whether lower-income countries are more closely linked to and affected by havens because they are less able to respond to the associated challenges. According to a more nuanced and interesting hypothesis, there are certain groups of (probably developing) countries in the middle of the income distribution that are the most closely linked to and affected by havens, while the poorest countries and the developed economies may be relatively less affected for different reasons (respectively, the simplic- ity of �nancial abuse in the former and the capacity of tax authorities and the relative political power of the latter). The following subsection discusses data issues in examining bilateral links between developing countries and secrecy jurisdictions, while the subsequent subsection presents initial �ndings. Data Ideally, to capture the full range of economic links, data would be avail- able on a bilateral basis on each relevant (trade) flow and (�nancial) stock, as follows: • Trade in goods. UN Comtrade provides bilateral data, on the basis of a detailed commodity breakdown if necessary, for most countries and for up to around 30 years.18 Here, we take data for the period 1995– 2008, by Standard International Trade Classi�cation (SITC), revision Tax Havens and Illicit Flows 359 3.19 Limitations include the absence of certain reporting jurisdictions of interest, for example, Channel Islands, which are counted with the United Kingdom for overseas trade purposes. This data set consists of some 580,000 observations dealing with total trade only (that is, no commodity breakdown below the SITC 3 total). Future work could expand this data set by combining totals reported on different SITCs and other bases and, eventually, by exploring patterns in particular commodity codes (for example, natural resource codes). By coding each reporter and partner country individually and according to groups (for example, regions, income groups, or various secrecy jurisdiction lists), one may obtain values for the share of trade carried by each country with members of other groups. To present the data, we have partially aggregated the jurisdictions into groups by income level and region, according to World Bank classi�- cations. The key information is the share of trade carried out with secrecy jurisdictions. • Trade in services. The United Nations also provides bilateral data on services trade, but this database is relatively recent and currently has much weaker coverage. We do not use it here, in part because of the especially weak coverage of bilateral reporting by developing coun- tries. Future work should investigate in more detail the availability and value of these data, possibly with a view to encouraging reporting. • Portfolio investment. The IMF Coordinated Portfolio Investment Sur- vey reports stocks of the following categories of portfolio investments on an annual basis: equity securities (for example, shares, stocks, par- ticipations, and similar documents such as American depositary receipts, where these denote the ownership of equity), long-term debt securities (for example, bonds, debentures, and notes, where these generally give unconditional rights to a �xed future income on a given date and have an original maturity greater than a year), and short- term debt securities (for example, treasury bills, commercial paper, and banker acceptances, with an original maturity of less than a year).20 For 2007, 74 jurisdictions reported; these were the most recent data available when this chapter was presented, and we use them here. Data are provided on the assets and liabilities of jurisdictions, but based only on reporting from the creditor side. Because each reporter therefore provides data for all other jurisdictions in which they have 360 Draining Development? investment positions, investment is reported in a total of 237 jurisdic- tions. There are relatively few developing countries among the report- ing jurisdictions, but a good many smaller secrecy jurisdictions do report. In contrast to the goods trade data, there is better direct cover- age of links originating in secrecy jurisdictions, but less direct cover- age of links originating in developing countries. No data set examined offers comprehensive coverage from both angles. As with the goods trade data, the 2007 IMF Coordinated Portfolio Investment Survey was coded to allow summary of the links by group. • Foreign direct investment. The United Nations Conference on Trade and Development provides bilateral data on foreign direct investment positions, but at greater expense than was possible for this research.21 Future work should seek to obtain these data and assess their coverage and quality. • Claims on foreign banks. The Bank for International Settlements pro- vides data on the extent of foreign claims on banks of reporting juris- dictions. These data are available on a quarterly basis with a short lag. We use the December 2008 data, the most recent con�rmed data available when this chapter was presented (BIS 2008). Sadly, these data are available only on a bilateral basis for consolidated, rather than locational, statistics. This means that apparent links between a developing country and a particular secrecy jurisdiction must be interpreted with caution, because, for example, Argentine deposits in a Bermudan bank operating out of Buenos Aires would be recorded as a Bermudan liability. Until locational statistics are made available by the Bank for International Settlements, which would be a valuable contribution to transparency, this is the only possibility; so, this strong caveat should be borne in mind. Main �ndings As noted elsewhere above, it would be preferable if we could judge juris- dictions using an indicator of secrecy on a sliding scale, thereby allowing us to measure gradual change where appropriate. For the moment, how- ever, we make do with the range of existing lists. To present the results simply, we report trade shares for two groups: the larger group of secrecy jurisdictions identi�ed by the Tax Justice Network (see table 11A.1 and the note to table 9A.4) and a narrower group closer in composition to the Tax Havens and Illicit Flows 361 usual suspects. The latter are de�ned as those jurisdictions that feature on a majority of the 11 lists shown in table 11A.1 (that is, on six or more).22 Table 11A.4 and �gures 11.1 and 11.2 present the main results.23 If the results for narrow and broad secrecy jurisdiction lists are examined and different stocks and flows are compared, a number of interesting pat- terns emerge. There is considerable variation among developing country groups both by income and by region. Analyses that fail to take this into account will be incomplete. We highlight a few other key points below. First, the trade flows of developing countries are generally more exposed (as a share of total economic links) to the effects of secrecy jurisdictions than are the trade flows of high-income OECD countries, which we take as a benchmark throughout. This is particularly pro- nounced in developing-country exports and in East and South Asia most of all. The differences in exposure between narrow and broad secrecy Figure 1 1.1. Intensity of Economic Exposure: Implied Ratios of Secrecy Jurisdiction Shares to GDP, by Income Group 50 low income upper-middle income 45 lower-middle income high-income OECD countries 40 35 30 % of GDP 25 20 15 10 5 0 narrow SJ broad SJ narrow SJ broad SJ narrow SJ broad SJ narrow SJ broad SJ exports (goods) imports (goods) portfolio investment bank claims importance of economic and financial links with SJs Sources: Trade data: UN Comtrade (United Nations Commodity Trade Statistics Database), Statistics Division, Department of Economic and Social Affairs, United Nations, New York, http://comtrade.un.org/db/. Portfolio investment data: IMF, “Coordinated Portfolio Investment Survey,� Washington, DC, http://www.imf.org/external/np/sta/pi/cpis.htm. Data on bank claims: BIS 2008. Note: SJ = secrecy jurisdiction. “Narrow� and “broad� refer to shorter or longer lists of secrecy jurisdictions (see the text). 362 Draining Development? Figure 1 1.2. Intensity of Economic Exposure: Implied Ratios of Secrecy Jurisdiction Shares to GDP, by Region 50 East Asia and Pacific Middle East and North Africa 45 Europe and Central Asia South Asia 40 Latin America and the Caribbean Sub-Saharan Africa 35 30 % of GDP 25 20 15 10 5 0 narrow SJ broad SJ narrow SJ broad SJ narrow SJ broad SJ narrow SJ broad SJ exports (goods) imports (goods) portfolio investment bank claims importance of economic and financial links with SJs Sources: See �gure 11.1. Note: SJ = secrecy jurisdiction. “Narrow� and “broad� refer to shorter or longer lists of secrecy jurisdictions (see the text). jurisdiction lists are often driven, as in the case of Latin American trade, by the inclusion of the United States in the latter list. Striking examples emerge from the underlying country-level data. One example is highlighted in �gure 11.3, which shows the total value of Zambian trade from 1995 to 2008, as well as the shares attributable to only one secrecy jurisdiction, Switzerland. Note that the share of declared Zambian exports to Switzerland in the total declared exports of Zambia (that is, exports declared at and subse- quently reported by Zambian customs) rise from a fraction of 1 percent in 1995–98 to more than 50 percent in 2008. The growth in exports over the period reflects Zambia’s emergence as a major copper producer. However, copper is not a product for which particular demand or pro- cessing capacity is noted in Switzerland, although relatively opaque metal trading does take place there. This highlights a major concern about the data and emphasizes the need for transparency. If one were to assume Tax Havens and Illicit Flows 363 Figure 1 1.3. Zambian Trade: Swiss Role, 1 995–2008 6,000 total exports (left axis) 60 total imports (left axis) exports: Switzerland (right axis) 5,000 50 imports: Switzerland (right axis) 4,000 40 US$, millions % of total 3,000 30 2,000 20 1,000 10 0 0 1995 1996 1 997 1 998 1 999 2000 2001 2002 2003 2004 2005 2006 2007 2008 year Source: UN Comtrade (United Nations Commodity Trade Statistics Database), Statistics Division, Department of Economic and Social Affairs, United Nations, New York, http://comtrade.un.org/db/. that Zambian copper is of the same quality as any copper in the Switzer- land trade data in the same detailed commodity categories (for example, coils of copper at thicknesses of less than 0.35 millimeters), so that Zam- bian copper exports to Switzerland should be priced at the same level as the declared Swiss trade in the same items, then the implied illicit capital flow out of Zambia is in the region of 30 times the declared export value. Clearly the impact on Zambian GDP would also be dramatic. However, the declared Swiss imports from Zambia and the declared total exports from Zambia are a tiny proportion of the Zambian-declared exports to Switzerland; so, such an estimate of the illicit flow cannot be supported on this basis. The use of Switzerland for transit in commodi- ties in this way is well known, but remains highly opaque (for example, see Berne Declaration 2011). What is absolutely clear is that declared Zambian copper exports, on which the country depends economically, effectively disappear once they leave the country. The potential for this opacity to undermine the country’s development is equally apparent. 364 Draining Development? The second broad result is that developing countries in general are more exposed to (more reliant on) secrecy jurisdictions in portfolio investment stocks. East and South Asia are especially exposed, reflecting the regional location of particular secrecy jurisdictions (for example, Hong Kong SAR, China; Singapore). To take an arbitrary example, the U.K. Crown Dependency of Jersey was responsible, in 2007, for more than 30 percent of the declared portfolio investment in six countries: Algeria, Guinea, and the Syrian Arab Republic (between one-third and one-half of declared investment), and Djibouti, Libya, and Turkmeni- stan (99.5 percent or more). Third, the picture is somewhat more mixed if bank claims are con- sidered. With respect to the narrow secrecy jurisdiction list, this is the one set of results in which high-income OECD countries are notably more exposed than each developing country group. On the broad secrecy jurisdiction list, however, developing countries are a little more exposed, particularly countries in South Asia and Sub-Saharan Africa. An arbitrary example: more than 20 percent of the revealed foreign claims of Eritrea and neighboring Ethiopia at the end of 2008 were against banks in one country, Switzerland. However, this may be the result of a Swiss bank operating out of, for example, Addis Ababa and taking deposits. The available data do not allow us any certainty about the locational aspects. Table 11A.5 presents the results on a different basis: the implied per- centage of GDP.24 We use additional data on economic aggregates and then extrapolate on the assumption that the shares of bilaterally reported trade (or other flows and stocks) that are associated with secrecy juris- dictions are equal to the shares of these jurisdictions in total trade (or other measures). This may be somewhat heroic, and it would clearly be preferable to have full data available. Nonetheless, for this exercise, which involves making broad comparisons across different country groups, there seems no particular reason to expect that our method would intro- duce a systematic distortion. The main results are twofold. First, the trade exposure of developing countries to secrecy jurisdictions relative to the total size of the econo- mies of these countries is notably greater than that of the high-income OECD countries. Second, the reverse pattern holds with regard to port- folio investment and foreign bank claims. It is important to understand Tax Havens and Illicit Flows 365 the difference that allowing for GDP makes: developing countries gener- ally have a much lower ratio of portfolio investment to GDP; so, while secrecy jurisdictions are responsible for a much higher proportion of the investment received (table 11A.4), the secrecy jurisdiction investment as a proportion of GDP (table 11A.5) is much lower in developing coun- tries than in high-income countries. As might be expected, the interna- tional �nancial integration of developing countries is broadly correlated with the per capita income levels of these countries. The results in table 11A.4 show that the intensity is not signi�cantly different (that is, the share of secrecy jurisdictions in the existing level of �nancial integra- tion), but the lower levels of integration mean that the total secrecy jurisdiction exposure through these channels is less. Next steps Subsequent work should extend these �ndings in three ways. First, researchers (and research funders) should prioritize the construction of a consistent data set containing as much information as possible on bilateral trade and �nancial flows and stocks. Major actors such as the IMF, the Bank for International Settlements, the OECD, and perhaps even the World Bank should aim to emulate, for various �nancial flows, the trade transparency that the United Nations has achieved. The absence of detailed results in this �eld can largely be blamed on the lack of transparency not of individual jurisdictions, but of the major play- ers in the economic system. An effort to achieve a greater understand- ing of the global economy, as well as of the major elements that form illicit flows, calls for an economic transparency initiative of the type being pursued by the Task Force for Financial Integrity and Economic Development.25 This would facilitate more detailed comparisons of this sort, but it would also help ful�ll the second requirement: rigorous econometric work to establish whether there are, indeed, systematic impacts on devel- oping countries because of their exposure to secrecy jurisdictions, such as impacts, in particular, on tax revenues, on economic growth, on inequality, and on governance outcomes. Finally, this work should be extended along the lines argued in this chapter, that is, to include the FSI or equivalently nuanced tools for capturing the relative opacity of jurisdictions. 366 Draining Development? Conclusion This chapter has drawn together the literature on the impact of tax havens on development and considered the need for further research in this neglected area. It is clear that the data on bilateral trade and �nancial flows, while frustratingly limited in some areas, do provide a suf�cient basis on which to draw quite clear conclusions about the relative exposure of developing countries to tax havens or secrecy jurisdictions, speci�cally: in general, developing countries do not appear to be any less exposed than high-income OECD countries to secrecy jurisdictions, and, in some cases, they are much more highly exposed. Therefore, the G-20 and others are correct in demanding that developing countries be included in any mechanism to require increased transparency from secrecy jurisdictions. In addition, there are clearly important differences in the exposure of developing countries across different income groups and regions. This warrants further investigation and analysis, which is likely, ultimately, to lead to the identi�cation of distinct policy recommendations. Although this brief summary of the state of research makes clear that there are more questions than answers, there can be no doubt that secrecy jurisdictions are an important issue for developing countries and that the damage done by them to development is also potentially large. For researchers, research funders, and policy makers, this is an area of work that must be given an increasingly greater priority. Notes 1. See, respectively, Owens (2009, 2); Gordon Brown, quoted in Christian Aid (2010, 22); and Declaration on Strengthening the Financial System, G-20 Lon- don Summit, April 2, 2009, 4–5, http://www.g20.org/Documents/Fin_Deps _Fin_Reg_Annex_020409_-_1615_�nal.pdf. 2. The second criteria is related to the characteristic that the OECD’s Towards a Level Playing Field Initiative seeks to measure through the consistent assessment of jurisdictions on a range of legal and regulatory features (OECD 2006). 3. See Christian Aid (2008) for a discussion of how this process of regulatory arbi- trage caused secrecy jurisdictions to play an important role in the unfolding of the �nancial crisis. 4. The table may be found on the website of this volume, http://go.worldbank.org /N2HMRB4G20. Tax Havens and Illicit Flows 367 5. For the IMF data, see http://www.imf.org/external/np/sta/bop/bop.htm. 6. For the Coordinated Portfolio Investment Survey, see http://www.imf.org /external/np/sta/pi/cpis.htm. For International Financial Statistics, see Interna- tional Financial Statistics (database), IMF, Washington, DC, http://elibrary-data .imf.org/FindDataReports.aspx?d=33061&e=169393. 7. The table may be found on the website of this volume, http://go.worldbank.org /N2HMRB4G20. 8. See Financial Secrecy Index, Tax Justice Network, London, http://www.�nancial secrecyindex.com/index.html. 9. The table may be found on the website of this volume, http://go.worldbank.org /N2HMRB4G20. 10. See also “The British connection� and “Chart 2: The British Empire by Secrecy Scores,� Financial Secrecy Index, Tax Justice Network, London, http://www .�nancialsecrecyindex.com/signi�cance.html#british. 11. Norway, Minister of the Environment and International Development, 2009, Tax Havens and Development: Status, Analyses and Measures, Oslo: Government Commission on Capital Flight from Poor Countries, http://www.regjeringen .no/upload/UD/Vedlegg/Utvikling/tax_report.pdf. 12. Financial Times, “Hurdles Remain for G20 Pact,� September 7, 2009; emphasis added. 13. This opacity is also used to hide the origin or bene�cial ownership of the pro- ceeds of crimes unrelated to tax, including other forms of corruption, but also, for example, the traf�cking of drugs and people, the �nancing of terrorism, and so on. Some implications of these activities for developing countries are treated elsewhere in this chapter; see also the chapters by Kopp and Levi in this volume (chapters 6 and 12, respectively). 14. See Financial Action Task Force, http://www.fatf-ga�.org/. 15. A casual investigation shows that The Bahamas, British Virgin Islands, Cayman Islands, and Isle of Man feature prominently in the Internet searches on claims about a tax-neutral platform. 16. For the Orbis data, see Bureau van Dijk, http://www.bvdinfo.com/Products /Company-Information/International/Orbis. 17. Norway, Minister of the Environment and International Development, 2009, Tax Havens and Development: Status, Analyses and Measures, Oslo: Government Commission on Capital Flight from Poor Countries, http://www.regjeringen. no/upload/UD/Vedlegg/Utvikling/tax_report.pdf. 18. See UN Comtrade (United Nations Commodity Trade Statistics Database), Sta- tistics Division, Department of Economic and Social Affairs, United Nations, New York, http://comtrade.un.org/db/. 19. See United Nations, “Detailed Structure and Explanatory Notes, SITC Rev.3 (Standard International Trade Classi�cation, Rev.3),� United Nations Statistics Division, http://unstats.un.org/unsd/cr/registry/regcst.asp?Cl=14. 368 Draining Development? 20. See IMF, “Coordinated Portfolio Investment Survey,� Washington, DC, http:// www.imf.org/external/np/sta/pi/cpis.htm. 21. See Data Extract Service, Division on Investment and Enterprise, United Na- tions Conference on Trade and Development, Geneva, http://www.unctad.org /Templates/Page.asp?intItemID=3205&lang=1. 22. Tables 9A.4 and 11A.1 may be found on the website of this volume, http://go .worldbank.org/N2HMRB4G20. 23. Table 11A.4 may be found on the website of this volume, http://go.worldbank .org/N2HMRB4G20. 24. Table 11A.5 may be found on the website of this volume, http://go.worldbank .org/N2HMRB4G20. 25. See the Task Force on Financial Integrity and Economic Development, Wash- ington, DC, http://www.�nancialtaskforce.org/. References Arora, P., B. Leonard, and O. Teunissen. 2009. “Ireland Emerging as an Onshore Investment Platform.� PwC Alternatives 4 (1), PricewaterhouseCoopers, New York. http://download.pwc.com/ie/pubs/IrelandEmergingAsAnOnshore InvestmentPlatform.pdf. Avi-Yonah, R. S. 2000. “Globalization, Tax Competition, and the Fiscal Crisis of the Welfare State.� Harvard Law Review 113 (7): 1573–676. Baker, R. W. 2005. Capitalism’s Achilles Heel: Dirty Money and How to Renew the Free- Market System. Hoboken, NJ: John Wiley & Sons. Berne Declaration. 2011. “Commodities: Switzerland’s Most Dangerous Business.� Berne Declaration, Zurich. http://www.evb.ch/cm_data/2011_09_19_Berne _Declaration_-_Commodities_-_English_Sample.pdf. BIS (Bank for International Settlements). 2008. BIS Quarterly Review, December 8, BIS, Basel, Switzerland. http://www.bis.org/publ/qtrpdf/r_qt0812.htm. Bosco, L., and L. Mittone. 1997. “Tax Evasion and Moral Constraints: Some Experi- mental Evidence.� Kyklos 50 (3): 297–324. Bräutigam, D., O.-H. Fjeldstad, and M. Moore, eds. 2008. Taxation and State-Building in Developing Countries: Capacity and Consent. Cambridge, U.K.: Cambridge University Press. Christian Aid. 2008. “The Morning After the Night Before: The Impact of the Finan- cial Crisis on the Developing World.� Christian Aid Report, November, Chris- tian Aid, London. http://www.christianaid.org.uk/Images/The-morning-after -the-night-before.pdf. ———. 2009. “False Pro�ts: Robbing the Poor to Keep the Rich Tax-Free.� Christian Aid Report, March, Christian Aid, London. http://www.christianaid.org.uk /Images/false-pro�ts.pdf. Tax Havens and Illicit Flows 369 ———. 2010. “Blowing the Whistle: Time’s Up for Financial Secrecy.� Christian Aid Report, May, Christian Aid, London. http://www.christianaid.org.uk/images /blowing-the-whistle-caweek-report.pdf. ———. 2011. “ ‘My Word Is My Bond’: Responsible Finance and Economic Justice.� Accounting for Change 3 (November), Christian Aid, London. http://www .christianaid.org.uk/images/accounting-for-change.pdf. Cobham, A. 2005a. “Taxation Policy and Development.� OCGG Economy Analysis 2, Oxford Council on Good Governance, Oxford. ———. 2005b. “Tax Evasion, Tax Avoidance and Development Finance.� QEH Working Paper 129, Queen Elizabeth House, University of Oxford, Oxford. http://www3.qeh.ox.ac.uk/pdf/qehwp/qehwps129.pdf. ———. 2007. “The Tax Consensus Has Failed! Recommendation to Policymakers and Donors, Researchers and Civil Society.� OCGG Economy Recommenda- tion 8, Oxford Council on Good Governance, Oxford. Desai, M. A., C. F. Foley, and J. R. Hines Jr. 2006a. “Do Tax Havens Divert Economic Activity?� Economics Letters 90 (2): 219–24. ———. 2006b. “The Demand for Tax Haven Operations.� Journal of Public Econom- ics 90 (3): 513–31. Eden, L., and R. Kudrle. 2005. “Tax Havens: Renegade States in the International Tax Regime.� Law and Policy 27 (1): 100–27. Fuest, C., and N. Riedel. 2009. “Tax Evasion, Tax Avoidance and Tax Expenditures in Developing Countries: A Review of the Literature.� Report, Oxford Univer- sity Centre for Business Taxation, Oxford. http://ec.europa.eu/development /services/events/tax_development/docs/td_evasion_riedel_fuest.pdf. Geithner, T. F. 2008. “Reducing Systemic Risk in a Dynamic Financial System.� Speech to the Economic Club of New York, New York, June 9. http://www .newyorkfed.org/newsevents/speeches/2008/tfg080609.html. Gordon, R. A. 1981. Tax Havens and Their Use by United States Taxpayers: An Over- view; A Report to the Commissioner of Internal Revenue, the Assistant Attorney General (Tax Division) and the Assistant Secretary of the Treasury (Tax Policy). Publication 1150 (4–81) (January 12). Washington, DC: U.S. Department of the Treasury. Hampton, M., and J. Christensen. 2002. “Offshore Pariahs? Small Island Economies, Tax Havens, and the Recon�guration of Global Finance.� World Development 30 (9): 1657–73. Heady, C. 2004. “Taxation Policy in Low-Income Countries.� In Fiscal Policy for Devel- opment: Poverty, Reconstruction and Growth, ed. T. Addison and A. Roe, 130–48. Studies in Development Economics and Policy. Helsinki: World Institute for Development Economics Research; Basingstoke, U.K.: Palgrave Macmillan. Hollingshead, A. 2010. “The Implied Tax Revenue Loss from Trade Mispricing.� Global Financial Integrity, Washington, DC. http://www.g�p.org/storage/g�p /documents/reports/implied%20tax%20revenue%20loss%20report_�nal.pdf. 370 Draining Development? Huizinga, H., and L. Laeven. 2008. “International Pro�t Shifting within Multination- als: A Multi-country Perspective.� Journal of Public Economics 92 (5–6): 1164–82. Kar, D., and D. Cartwright-Smith. 2008. “Illicit Financial Flows from Developing Countries, 2002–2006.� Global Financial Integrity, Washington, DC. http://www .g�p.org/storage/g�p/economist%20-%20�nal%20version%201-2-09.pdf. Kudrle, R. T. 2008. “Did Blacklisting Hurt the Havens?� Paolo Baf� Centre Research Paper 2008–23, Paolo Baf� Centre on Central Banking and Financial Regula- tion, Bocconi University, Milan. Kudrle, R. T., and L. Eden. 2003. “The Campaign against Tax Havens: Will It Last? Will It Work?� Stanford Journal of Law, Business and Finance 9 (1): 37–68. Mittone, L. 2006. “Dynamic Behaviour in Tax Evasion: An Experimental Approach.� Journal of Socio-Economics 35 (5): 813–35. Murphy, R. 2008. “Finding the Secrecy World: Rethinking the Language of ‘Off- shore’.� Report, Tax Research LLP, Downham Market, U.K. http://www.tax research.org.uk/Documents/Finding.pdf. OECD (Organisation for Economic Co-operation and Development). 1987. “Inter- national Tax Avoidance and Evasion: Four Related Studies.� Issues in Interna- tional Taxation 1, Committee on Fiscal Affairs, OECD, Paris. ———. 1998. “Harmful Tax Competition: An Emerging Global Issue.� OECD, Paris. ———. 2006. Tax Co-operation 2006: Towards a Level Playing Field; 2006 Assessment by the Global Forum on Taxation. Paris: OECD. Owens, J. 2009. “Tax and Development: Why Tax Is Important for Development.� Tax Justice Focus 4 (4): 1–3. http://www.oecd.org/dataoecd/10/22/43061404.pdf. Oxfam. 2000. “Tax Havens: Releasing the Hidden Billions for Poverty Eradication.� Oxfam Brie�ng Paper, Oxfam International, London. http://www.taxjustice .net/cms/upload/pdf/oxfam_paper_-_�nal_version__06_00.pdf. ———. 2009. “Tax Haven Crackdown Could Deliver 120bn a Year to Reduce Pov- erty: Oxfam.� Press release, March 13, Oxfam GB, London. http://www.oxfam .org.uk/applications/blogs/pressof�ce/?p=3912. Palan, R. 1998. “The Emergence of an Offshore Economy.� Futures 30 (1): 63–73. ———. 2002. “Tax Havens and the Commercialization of State Sovereignty.� Inter- national Organization 56 (1): 151–76. Shaxson, N. 2011. Treasure Islands: Tax Havens and the Men Who Stole the World. London: Bodley Head. Stewart, J. 2008. “Shadow Regulation and the Shadow Banking System.� Tax Justice Focus 4 (2): 1–3. http://www.taxjustice.net/cms/upload/pdf/TJF_4-2_AABA _-_Research.pdf. TJN (Tax Justice Network). 2005. “The Price of Offshore.� Tax Justice Network Brief- ing Paper, March, TJN, London. ———. 2008. “Creating Turmoil,� pp. 9–217 in Written Evidence to the UK House of Commons Treasury Select Committee on Offshore Financial Centres, London: HMSO, http://www.parliament.uk/documents/upload/OFCWrittenEvidence .pdf. Tax Havens and Illicit Flows 371 Torvik, R. 2009. “Appendix 1: Why Are Tax Havens More Harmful to Developing Countries Than to Other Countries?� In Tax Havens and Development: Status, Analyses and Measures, ed. Norway, Minister of the Environment and Interna- tional Development, 155–94. Oslo: Government Commission on Capital Flight from Poor Countries. http://www.regjeringen.no/upload/UD/Vedlegg/Utvikling/ tax_report.pdfpp. Williamson, J., and M. Mahar. 1998. “A Survey of Financial Liberalization.� Princeton Essays in International Economics 211 (November), International Economics Section, Princeton University, Princeton, NJ. Zoromé, A. 2007. “Concept of Offshore Financial Centers: In Search of an Opera- tional De�nition.� IMF Working Paper WP/07/87, International Monetary Fund, Washington, DC. http://www.imf.org/external/pubs/ft/wp/2007/wp0787.pdf. 12 How Well Do Anti–Money Laundering Controls Work in Developing Countries? Michael Levi Abstract Much of the attention to the link between corruption and money laun- dering understandably has focused on the developed world, given that it is the businesses of the developed world that mostly pay the bribes and that it is the banks and other institutions of the developed world that receive and recycle the bribes. However, this is not the whole picture. Substantial sums arising from corruption are often retained or even imported into developing countries to pay off political rivals, satisfy obligations to families, friends, and clans, and invest in local businesses and property. Kleptocrats usually wish to retain power and therefore require such funds. Historically, the principal bene�t of anti–money laundering (AML) regimes in developing countries has been the international political and economic acceptance and the avoidance of sanctions that they bring. This has to be set against the substantial costs of introducing such regimes in human resource–poor countries. If nations are to be required to have AML measures, one component of optimizing the bene�ts involves using these measures to combat corruption, especially corrup- tion at a high level. Namibia and Nigeria excepted, there do not appear 373 374 Draining Development? to be any cases in which internal AML reporting within developing countries has led to major cases against in-favor politically exposed per- sons (PEPs). However, suspicious activity reports (SARs) can be useful against out-of-favor PEPs and in helping to accumulate information about assets held in the actual or bene�cial ownership of in-favor and out-of-favor PEPs. If no recording systems are in place, �nancial investi- gators in the aftermath of criminal cases would face greater dif�culty in identifying assets. Unless the national �nancial intelligence unit (FIU) is believed to be both discreet and independent of the government, potential whistle- blowers might be afraid of exposure as sources of information. If suspi- cions are communicated to the FIU, what could the FIU plausibly do with the report? This highlights the importance of the relationship between FIUs and effectively independent investigation and prosecuto- rial bodies, which may be needed if AML regimes in developing countries are to have a signi�cant impact on domestic grand corruption. Local and foreign investigations running in parallel might produce bene�ts; so, too, might better domestic coordination with anticorruption bodies. Achiev- ing this level of competence and independence is a major task. Introduction AML is a major element in the standard list of interventions with appar- ent potential to reduce the flow of illicit funds both into and out of developing countries and into developed ones, usually for the purpose of storing wealth or for expenditure on conspicuous consumption. It is the �rst such intervention that has been put into place on an almost univer- sal basis. This chapter represents an assessment of what is known about the effects of AML regimes within developing countries in dealing with internal money laundering and kleptocratic wealth. The international AML system is over 20 years old, but, prior to 2001, it was primarily focused on the countries of the Organisation for Eco- nomic Co-operation and Development (OECD) and on �nancial havens (OECD or not) in which funds from drugs and fraud were laundered. Since 2001, it has been implemented in most countries, partly with the aim of protecting the West from the export of terrorism, but also to deal with grand corruption and organized crime, often arising from crimes How Well Do Anti–Money Laundering Controls Work in Developing Countries? 375 committed outside a jurisdiction, but also from crimes within it. In many developing countries, the system is new. For example, most Afri- can countries have only passed AML laws since 2005, usually under pres- sure from the developed world, and some jurisdictions have adopted such laws irrespective of the insigni�cance of their �nancial service sys- tems domestically or internationally. Examples of the latter include Malawi, Namibia (the Financial Intelligence Act 2007), and the formerly blacklisted (by the Financial Action Task Force [FATF]) South Paci�c island of Nauru, where, in the last few years, more than a third of the laws passed have related to AML, even though Nauru is no longer any sort of �nancial services center, and there is nothing left there to regulate that could plausibly cause serious harm.1 The fact that a country is not a �nancial center does not, of course, mean that there is no money laun- dering there. All countries have domestic crime and the possibility of both transnational and domestic bribery in contract negotiations and may be conduits for terrorism and the �nancing of terrorism. Given the broad de�nitions of laundering generally in place, this creates scope for AML. The question is: How much impact can we expect in such coun- tries to justify the expenditure of scarce resources that AML requires? Given the recency of these laws in many parts of the developing world, it would be unreasonable to expect much evidence of effectiveness yet. Moreover, the empirical links between reductions in serious crime for gain and any measures against serious crime (including AML) are dif�- cult to demonstrate (Levi and Maguire 2004; Levi and Reuter 2006). The arguments about the effective implementation of AML can be tautologi- cal in both the developed and the developing worlds. The waters are fur- ther muddied by the fact that most of the grand kleptocracy cases that are known occurred before the signi�cant implementation of AML in developing countries in Africa, Asia, or Central and South America, so it would be improper to deduce that AML failed in these regions. This chapter, then, has an important speculative element with respect to effectiveness. For example, it examines the plausibility of a klepto- cratic regime allowing a local FIU to be effective and ful�lling any other requirement to enable this effectiveness to be translated into broader preventive or criminal justice action. It briefly describes the historical evidence that anticorruption agencies that have had signi�cant success (such as in Kenya, Nigeria, and Zambia) have generally been dismantled 376 Draining Development? either by the governments they have attacked or by successor govern- ments (Ribadu 2010). This reflects the political economy reality of AML: it requires the support of the powerful who, in a corrupt country, have the most to fear from its effectiveness. The chapter also examines the role of domestic AML in dealing with (1) the proceeds of corruption and (2) those proceeds of crime and licit activities that are used to �nance corruption within developing coun- tries. It suggests that dictators and elected kleptocrats usually plan on staying in power and expect to do so. For these purposes, they require funds, as well as career positions to distribute to those who are useful in maintaining their power and gaining rent-seeking opportunities. They may also purchase vehicles and properties in their domestic jurisdictions in excess of their of�cial incomes. So, opportunities exist for AML inter- ventions locally, as well as in the rich world. The Goals of AML Policies in Developing Countries Looked at historically, the growth of AML measures has been motivated largely by the desire of rich countries to reduce the impact of crimes upon their own citizens, residents, and businesses.2 The theory is that AML will demotivate offenders, shorten criminal careers, and disrupt or prevent the loose construct known as organized crime. It does this by incapacitating organizational growth and stripping those assets under the control of criminals that offenders cannot demonstrate they have acquired by legitimate means.3 Another aim of AML is to increase public satisfaction with criminal justice and administrative regulation and to reduce the attractiveness of criminals as role models by showing that crime does not pay.4 Note, however, that past lifestyle expenditure can- not, in practice, be recovered. The applicability of this model to grand corruption is taken for granted in the literature, and, in particular, it is assumed that, if PEPs and their close families cannot deposit large sums in rich countries, they will not commit corruption. The mechanism for achieving these goals is to develop an interna- tional chain with no weak link, making it impossible for people to retain the proceeds of multifarious crimes both outside and inside their own jurisdictions. Though most mutual legal assistance is requested by rich countries from other rich or from poor countries, this also tends to be How Well Do Anti–Money Laundering Controls Work in Developing Countries? 377 true, paradoxically, in most grand corruption cases, though rich coun- tries do not always help as much as they could. AML is also available for purely domestic cases, in which offenders try to conceal their criminal income within the local �nancial sector. Whether from corruption or not, some alleged proceeds of crime can make a signi�cant difference to welfare within developing countries, though it may be countered that—as with some Russian oligarch émigrés in the West—people are seeking to buy their immunity from extradition by courting popular support. Kobi Alexander, an Israeli technologist whose extradition from Namibia has been sought by the United States since 2006 on charges of stock options securities fraud related to his �rm Comverse Technology (and who agreed to a U.S. Securities and Exchange Commission civil settlement in 2010 for US$54 million), brought sub- stantial, if disputed funds with him to Namibia and has invested or given away millions to �nance low-cost high-technology housing projects and endow educational funds for Namibians. Alexander transferred 120m. Namibian dollars (N$16.9m.) from Israel to Namibia and was granted a two-year work permit on a pledge to spend N$300m. in the country. . . . He had invested N$11m. in the country, including N$3.4m. in low-cost housing in the western town of Swakop- mund, because of demand created by the opening of the Langer Heinrich uranium mine. . . . Alexander invested another N$5.9m. in an auto-body shop business in Windhoek and bought a N$3.5m. house on a golf estate. He also enrolled his children in the 275-pupil Windhoek International School, where fees last year ranged between N$18,000 and N$118,000 a year. . . . “I have invested responsibly and with a view to engaging previously disadvantaged Namibians,� Alexander said.5 There is no extradition treaty between Namibia and the United States, and, as of November 2011, Alexander is out on bail of N$10 million pending a High Court challenge to the United States. In addition to governmental pressures, AML has, in recent times, been regularly supported by large international banks. These have the capacity to implement the required sophisticated control systems. They have successfully called for an extension of such controls to their smaller commercial competitors (including money transmission businesses), 378 Draining Development? driving up the latter’s costs of doing business and the cost of remittances by expatriates to poor countries. However, cooperation over tax issues by some banks (and some governments) remains an item of serious con- tention (Chaikin 2009; Sharman 2011; Shaxson 2011). In regard to the methods of laundering, the conventional wisdom since the 1980s is that they entail a three-stage process of (1) the place- ment of funds into the �nancial or other sector, (2) layering steps to make the money trail harder to follow, and (3) integration of the cleansed capital into the legitimate economy after the illegitimate origins have been obscured. This process does not necessarily involve international money flows, especially not in cases of self-laundering (see below). Nonetheless, the existence of international flows of funds, criminal pre- cursors (such as chemicals), and international illicit markets (such as drugs, people smuggling, and human traf�cking) provides leverage for peer pressure between states in the form of mutual evaluations (Levi and Gilmore 2002). Such evaluations have long been extended to corruption through the Council of Europe’s Group of States against Corruption and, more tentatively, to those nations that have rati�ed the United Nations Convention against Corruption (UNCAC). A different form of expert evaluation occurs within the framework of the OECD Conven- tion on Combating Bribery of Foreign Public Of�cials in International Business Transactions (more well-known as the OECD Anti-Bribery Convention). To implement AML controls, a broad network of FATF-Style Regional Bodies has been developed; AML legislation has been adopted; and an ever-widening range of private sector businesses and professions has become involved. These last include car dealers, jewelers, money exchanges and transmitters, real estate vendors, and, in Europe, but in few places elsewhere, lawyers who are required to make reports to their national FIUs on their subjective suspicions about their clients. Much less commonly, they also include objective cash receipts that exceed a modest level.6 The privacy of high net worth individuals with regard to their own governments, foreign governments, and the risk of criminal extortion is a matter of great political sensitivity. Thus, the person who runs an FIU (for example, typically, the central bank, the attorney general, or the police) makes a great difference in the extent to which the FIU is trusted by those it regulates and with whom it has to cooperate.7 This also makes How Well Do Anti–Money Laundering Controls Work in Developing Countries? 379 a difference in how ef�cient an FIU is in follow-up action. As of July 2010, there were 127 members of the Egmont FIU Group, including the richest jurisdictions and poorer countries. FIUs are absent altogether only in the poorest countries, and some of these countries are not admit- ted to Egmont or have been suspended from the group because, it is felt, they have not reached a suf�cient standard of competence and indepen- dence. Hitherto, most of the attention has been paid to laundering and the �nancing of terrorism because these affect developed countries (and their associated commercial and security interests worldwide). The interest of developed countries in developing countries has mainly taken the form of reviewing the ways in which the tolerance or alleged toler- ance of the latter is inhibiting the success of efforts to control drugs, organized crime, human traf�cking, and terrorism. There are signs of increased systematic action against grand corrup- tion and laundering. The G-20 Seoul Summit Action Plan of November 2010 requires an annual follow-up.8 These corruption-related reputa- tional issues and collateral harms and bene�ts arguably have become part of the goals of AML. Swiss regulators and investigative judges might have expected even greater credit for their ingenious use of their powers and creative applications of the law from Marcos onward (for a more skeptical view, see Chaikin and Sharman 2009). The Swiss use this as political capital to compare themselves favorably with the United King- dom, the United States, and the latters’ “satellite states,� while continuing to advertise Swiss banking secrecy unless criminal investigations yield suf�cient evidence to generate mutual legal assistance. Measurement and Patterns of Laundering Relative to Corruption One of the dif�culties faced by any study in this area is that the de�nition of corruption may be narrowly interpreted to include only bribery (as the criminal law and criminal statistics do) or—as is now more common in analytical discourse—may be more broadly interpreted to include the misuse of of�ce to steal funds or to acquire assets for private gain at sig- ni�cantly below true market value. If a PEP has a substantial measure of control over a country’s treasury, there is no need to bribe treasury staff: in addition to the custom of obedience, threats—including the threat of 380 Draining Development? unemployment or homicide or any form of harm to relatives—can be suf�cient to induce compliance. The PEP may bypass domestic controls anyway, for example, when Marcos instructed the Philippine Central Bank to generate large disbursements from its account at the New York Federal Reserve, rather than using the bank’s Manila of�ce (Briones 2007). Analytically satisfactory estimates are affected by the de�nitional problems that bedevil white-collar crime studies generally: acquitting former President Chiluba in August 2009, the Zambian court appeared to accept the defense that Zambian law did not lay down any upper limit on the gifts the president could accept.9 The Zambian authorities did not proceed with an appeal against this ruling, but this would mean that what to some might look like kleptocracy does not constitute criminal corruption. Unless one is imposing an objective standard of corruption that disregards the legal requirements of the mental element to prove crime, one presumably should drop such sums from the total in esti- mates. Such considerations do not appear to be incorporated in the cor- ruption estimates embodied in the governance indicators developed by Kaufmann (2005) or others. To focus only on criminal corruption or laundering cases that have passed through the �nal stages of appeal would be too strict a test. On the other hand, to allow anyone to impose their own judgment of whether any given conduct was corrupt would likely generate many different estimates without any clear way of adjudi- cating among them. This is not an appropriate place to rehearse corrup- tion de�nitions, but depending on whether or not the private sphere is included in the victim category, corruption might be the use of illegal means for private gain at public expense or the misuse of public trust or authority in private enterprises or nonpro�t organizations. Sums transferred might be referred to as funds rather than bribes to express the possibility that large tranches of pro�t might arise (1) from transfers for �ctitious or partly �ctitious goods and services (that is, false accounting) or (2) from undervalue privatizations (including land sales) to nominees of the of�cials or to apparently independent third parties as part of patron-client relationships. Noncash payments could take many forms, including (1) electronic funds, which are less anonymous than cash because, excepting some prepaid cards for which there is no serious know-your-customer effort, they leave audit trails; (2) jewels, real estate, vehicles, and so on regulated under the European Union’s third anti– How Well Do Anti–Money Laundering Controls Work in Developing Countries? 381 money laundering directive and other legislation; or (3) education or health payments to family and friends, which can be substantial and can be paid from corporate or third-party funds, usually overseas. An impor- tant question is what proportion of bribes is retained internally (from domestic sources) or imported (from external sources); conversely, what proportion is exported (from domestic sources) and retained overseas without being imported (from external sources)? Unfortunately, data are unavailable to examine this, and it is unclear what proxies could defensibly be used instead. It is conventional to start with a discussion of the extent to which the proceeds of corruption are laundered, but, because the issue is dealt with elsewhere in this volume, I make only cursory reference here. Thus, Goredema and Madzima (2009), Chaikin and Sharman (2009), and Shaxson (2011) reiterate the estimated global �gure of US$1 trillion in bribes, which itself derives from Baker (2005) and Kaufmann (2005).10 This �gure is at risk of becoming a “fact� by repetition, but, irrespective of arguments about its validity, the numbers are big enough to matter to developing countries. Most of it may well go to routine low-level bribes, incidental to the hazards of everyday interactions with of�cials in many poor countries; these rarely lead to a demand for money laundering ser- vices, but simply help to �nance the modest daily lives of low-level of�- cials. It is the high-level corruption that is salient to AML measures. It is often asserted that large sums have been exported by kleptocratic public of�cials, but much smaller sums are usually detected in overseas accounts, and even smaller sums are recovered. This is true even if no deals are done, as they were with the Abacha family, thereby allowing family members to keep many millions, while they agreed to return larger sums to the national treasuries (see Maton and Daniel’s chapter 13, in this volume). Some of the attrition may result from sophistication in the layering of bene�cially owned corporate secrecy vehicles, as well as from inadequacies or corruption in the pursuit of the money trail. In addition to secrecy over the value of exports in the extractive industries, cases in different parts of the world indicate that the nontransparency of contracts in the security and intelligence areas offers great scope for cor- ruption and the nondetection of corruption in the awarding of contracts both for goods and for protection services, overriding whatever normal internal audit controls may exist (Daniel and Maton 2008; Salonga 2000; 382 Draining Development? Wrong 2000, 2009). The same is true of the privatization of public assets at below value.11 The 18 contracts that were awarded by the Kenyan gov- ernment to Anglo-Leasing totaled US$751 million (Wrong 2009). Many of the contracts were for illusory supplies, but the literature does not discuss what proportion of the bribes (or the marginal difference between the purchase price and the proper market price) reached the domestic parties locally.12 Some de�nitional discussion may be helpful in illuminating the dimensions of the phenomenon. There are at least two signi�cantly dif- ferent ways in which people use the term money laundering. The �rst way refers to the act of hiding the illicit origins of funds so as to make tainted wealth look suf�ciently legitimate to withstand signi�cant scrutiny (which it may never receive). This plausibly is what most people who encounter the term would expect money laundering to mean. The sec- ond way refers to the act of acquiring, possessing, using, or disposing of the proceeds of crime and covers all acts that fall within the criminal laws and regulations against money laundering. These laws and regula- tions are intentionally framed broadly to stimulate business, �nance, and the professions to develop and report their suspicions, thereby mak- ing it more dif�cult for criminals to legitimize their wealth in the �rst sense above. It also penalizes broadly de�ned self-laundering by those who commit the primary money-generating crimes (in this case, giving, receiving, or extorting bribes, locally and transnationally) to such an extent that almost anything the offenders do with the proceeds consti- tutes laundering. This can generate some legal complexities if, as is com- mon in developing and many developed countries, successful mutual legal assistance and prosecution depend on an act being criminal in both countries in which the money laundering process occurs.13 A �gure one might consider using is the total assets available for asset recovery, which would be a combination of the saved proceeds of cor- ruption, the total value of the contract (under English law, for example), and the criminal and regulatory �nes imposed on the �nancial interme- diaries (though such �nes are not usually available as compensation for the developing country). Here, to avoid confusion, we examine only the total proceeds of corruption. We divide these into PC (E), representing the total paid and stored or spent outside the developing country, and PC (I), representing the total internally stored or spent, though PC (I) How Well Do Anti–Money Laundering Controls Work in Developing Countries? 383 may originally have been paid externally. PC (I) might be anticipated to vary positively with the desire and the expectation of kleptocrats to remain in power and need the funds so stored to sustain political and military support. One component of the AML regime involves steps to prevent people from integrating the proceeds of crime into the legitimate economy; though this may act as a deterrent to the involvement of people in crime or shorten the length of criminal careers, it may also prevent people from “going legit� (becoming actually respectable rather than only apparently respectable). It is a tenable explicit policy aim to show crimi- nals that, however much they may reform, they can never enjoy the fruits of their crimes, including the fruits derived from legitimate investments made possible through crime (the doctrine of the fruit of the poisoned tree). This would nullify the personal ambition to become socially inte- grated of persons such as the bankruptcy fraudster who was interviewed by Levi (2008) and observed that he only wanted to make enough money to be able to afford to be honest.14 The netting out of harms lies outside the scope of this chapter, but the interviews conducted for this report with a convenience sample of public of�cials in some developing coun- tries (including countries in the former Soviet ambit) since 1988 suggest that some of the of�cials are privately ambivalent about whether it is a bad thing to accept tainted funds into their economies. Such thoughts might invite negative reactions if expressed in of�cial meetings and are normally suppressed for that reason.15 Irrespective of personal greed and bene�ts, politicians may obtain political and social credit from inward investment, whatever the source of the funds, if it creates employment and a multiplier effect from expenditure. In major �nancial centers in the developed world, it is dif�cult to acknowledge inflows of criminal funds as an acceptable economic bene�t, and the OECD Anti-Bribery Convention expressly excludes economic interests as justi�cations for not prosecuting transnational bribery. Informal value transfer systems, including hawala-type money move- ments, have received some signi�cant interest, especially post-9/11 (Pas- sas 2005; Rees 2010).16 Most of this interest has focused on the transmis- sion of terrorist �nance to operational sites from Muslims in developing and developed countries and on terrorist �nance and the proceeds of organized crime remittances from developed countries to developing 384 Draining Development? countries. However, the signi�cance of informal value transfer systems for the proceeds of corruption has received scant attention. Perhaps this is because grand corruption payments from multinationals seem unlikely to be sent in this way—though there is no reason, in principle, why they cannot be so sent—and partly because they seem an implau- sible conduit for payment into the large international banks in the West on which the substantial legal actions and popular protests have focused. There is no evidence from published investigations such as the case of the French Elf-Aquitaine frigates purchased by Taiwan, China, and simi- lar cases in which informal transfer channels have been used (beyond the cash smuggled in diplomatic bags and containers or on private planes). Large corporations would have to unlaunder the money, turning it from white to black money, prior to sending it over informal channels (as they might also have to do with cash bribes). To the extent that informal value transfer is unattractive for grand corruption, its signi�cance for money laundering generally is reduced. One reason this area is important is the fact that, to maintain power, corrupt leaders may have to engage in patron-client deals to purchase support locally or regionally (Wrong 2000). There is no motive, in prin- ciple, why such formal or informal political funding should have to come from the money laundering circuit in the developed or the �nancial haven world. What can we learn from cases such as Abacha, Marcos, Mobutu, Montesinos, and Suharto about this local distribution and about the extent to which the existence of local FIU regimes and legisla- tion targeting the proceeds of crime might be drivers for international outflows of corrupt �nance now or in the future? Logically, for the inter- nal proceeds of corruption regimes to drive funds to leave countries because these countries reject accounts or actively share information, the regimes would have to be more draconian than the regimes existing in money centers elsewhere.17 This chapter focuses on the flows of corrupt funds into and circulat- ing around developing countries, rather than the more usual issue of estimates of the amount of money stored overseas in major �nancial centers. One component is the question of how funds obtained through corruption are spent. This has not received much systematic attention in research; the inference in high-pro�le cases is that corrupt funds are spent on affluent lifestyles in the West, including homes and so on, and How Well Do Anti–Money Laundering Controls Work in Developing Countries? 385 are stored in American, British, and Swiss banks as a part of a strategy, presumably, to avoid con�scation should there be an unfriendly regime change at home. However, because many regimes involve political clien- telism, it seems likely that some of the funds, even if they originate in Western �nancial institutions, are recycled domestically in developing countries (Warutere 2006; Wrong 2000, 2009; ongoing allegations in Nigeria). Although Mobutu (like other high-living suspect African, Asian, Caribbean, and Latin American leaders) undoubtedly required a great deal of money to purchase and maintain properties overseas, Wrong (2000) makes a convincing case that he was more interested in power than in money. Mobutu needed large amounts of cash in what is now the Democratic Republic of Congo, as well as the ability to confer well-remunerated government posts on the grands légumes (big vegeta- bles: the elite) who sought to feed at his table without actually doing anything other than not opposing him. The redistributed bribes gave Mobutu the ability to defuse threats and monitor and act effectively against internal abuses by potentially rival rent seekers. It may be hypothesized that some institutional variables may affect the relationship between �nancial institutions and the vulnerability to corruption, though a report of the Eastern and Southern Africa Anti– Money Laundering Group discusses the variables as a static preordained classi�cation rather than as a dynamic process that can change over time and place.18 It seems plausible that an additional cultural-personality variable is whether the bribe recipients are—in the typology of the Knapp Commission investigation into police corruption in New York in 1973—meat-eaters (actively in search of rent-seeking opportunities) or merely grass-eaters (passively accepting the bribes offered to them). Another important dimension of the visibility of the corruption- laundering link is the origin of the corrupt funds. Large blocks of funds from bribes in the awarding of large contracts to transnational corpora- tions might be more visible than smaller, more frequent insider con- tracts to non–arm’s-length companies. Import monopolies or privileges among family members or party members might be one important route for rent seeking. There might, however, be nothing wrong, on the surface at least, with the awarding of national or transnational contracts to the relatives of senior public of�cials directly or through companies that may be known to be associated with such of�cials or more obscurely 386 Draining Development? held.19 Whereas, in other cases, bribes, shares, or charitable donations may be given in return for permission to engage unmolested in illegal enterprises such as drug traf�cking, and these bribes may sometimes, ultimately, help �nance terrorism, as in Afghanistan (U.S. Senate 2009), the privileges assigned to family members as in the cases described above may, at the other ideological extreme, merely �nance conspicuous con- sumption, private education (schools and universities), and health care overseas. How and in what countries and currencies are the latter paid for? A typology is proposed by the Eastern and Southern Africa Anti– Money Laundering Group (see table 12.1). This typology leaves out the fact that �nancial institutions may sim- ply be afraid of or feel loyal to the national leadership—who may belong to the same ethnic group—and may bene�t from the trickle down sym- bolized by the �rst part of the title of Wrong’s book (2009) It’s Our Turn to Eat: The Story of a Kenyan Whistle-Blower. Therefore, the bank may continue to launder money for its friends or may avoid doing so out of a fear of capricious closure or other economic detriment caused by those in power. This typology also assumes that the �nancial institution has detected the laundering (or is in a position to do so with modest effort). If the Table 1 2.1. Vulnerabilities in Financial Intermediation Characteristics of a Category �nancial intermediation institution Comment A The institution involved is corrupt from The institution is primed to be a vehicle for inception. money laundering; its use to launder the proceeds of crime is thus inevitable. B The institution is corrupted by subsequent Although not at �rst intended to be a changes in ownership or management or vehicle for money laundering, it eventually in the operating environment. drifts in that direction. C The institution has corrupt employees who The institution could be separated from the provide money laundering on an ad hoc corrupt employees. and noninstitutionalized basis. D The institution unwittingly facilitates money Although the institution may not bene�t laundering because it has no mechanisms from money laundering, it sabotages the to detect money laundering. AML system. Source: Goredema and Madzima 2009. How Well Do Anti–Money Laundering Controls Work in Developing Countries? 387 funds are prelaundered by placement in a real business venture (rather than a shell company, the bene�cial ownership of which may not be transparent), then how they may be identi�ed as proceeds of corruption or any other form of crime is not obvious. This is yet another note of caution: it would be unwise to assume that the within-country identi�- cation of the proceeds of corruption is easy whether the country is a poor one or a rich one. It may be that the families of corrupt politicians in developing countries �nd it easier than corresponding families in rich countries to integrate the proceeds of corruption (including political and economic extortion) into their businesses. This issue will require more analysis. Alternatively, one might adapt Levi’s typology (2008) of bankruptcy (and other) frauds to the involvement of individuals and businesses in money laundering. The typology is as follows: • Preplanned frauds, in which the business scheme is set up from the start as a way of defrauding victims (businesses, the public sector, individuals) • Intermediate frauds, in which people start out obeying the law, but consciously turn to fraud later • Slippery-slope frauds, in which deceptions spiral out of control, often in the context of trying—however absurdly and overoptimistically— to rescue an essentially insolvent business or set of businesses, thereby escalating the losses of creditors By analogy, some schemes are set up principally for the placement, layering, or integration of the proceeds of crime; some existing busi- nesses knowingly hire themselves out or are pressured into selling out to would-be launderers; and others drift into occasional part-time laun- dering by assisting large corporations or full-time criminal enterprises if they need the business to keep going, but never intending to become full-time launderers. Given the absence of systematic data collection and the low probability that laundering schemes will be detected and acted against, it is plausible that intermediate and slippery-slope laundering will occur more often during economic downturns. In short, the motiva- tion to defraud and the motivation to launder are heterogeneous rather than a single phenomenon. 388 Draining Development? Situational opportunity models of corruption It is illuminating to examine the different conditions under which situ- ational opportunities for corruption are created and exploited. Public corruption cases typically involve either (1) a legal or natural person outside the government (the corruptor) who pays a person inside the government (a corrupt PEP) in exchange for a government bene�t, or (2) a PEP who (with or without active collusion) embezzles public funds. In the second type, corruptors outside the government are not usually involved in the scam, though they may set up corporate or other non- transparent vehicles and act as bankers or other types of value transfer- ors. The price extracted for the corrupt bene�t may be modest in the view of the corruptor, but signi�cant in the view of the recipient, or it may be signi�cant for both. The potential funds transfers are a function of this and other expenditures. The development of land for tourism— the alleged source of corruption that, not for the �rst time, led to the suspension of local government in the Turks and Caicos Islands and has been the subject of a major criminal investigation by the U.K. govern- ment (Auld 2009)—may generate substantial rewards for all parties, though only if the infrastructure actually is developed.20 Gordon (2009) argues that, in almost all grand corruption cases, the corrupt proceeds are already within the �nancial system at the time of their generation (that is, no cash is involved), while, in two cases, much of the proceeds may not have been in cash. This is not surprising given that the sources of corrupt proceeds are primarily kickbacks on govern- ment contracts and bribes to receive a government bene�t. The corrup- tor might therefore normally be a businessperson who receives the gov- ernment bene�t in the form of checks or wire transfers directly from the state. If—before or after the award of the contract, depending on the level of trust and other factors—the corrupt government contractor withdraws cash and transfers it to a corrupt PEP, this would break the chain of bank records connecting the two. However, a new problem is generated because the corrupt PEP may be required to deposit cash back into the �nancial system. If the problems associated with this redeposit of cash by the corrupt PEP are considered to outweigh the bene�ts of breaking the chain of �nancial records, the PEP and the “donors� often attempt to mask their association with the funds by hiding the identity of the bene�ciary.21 The donor—who may be at risk of being discovered How Well Do Anti–Money Laundering Controls Work in Developing Countries? 389 by prosecutors because of offenses connected with the OECD Anti- Bribery Convention—and the corrupt PEP need to (1) hide the owner- ship and control of the assets, (2) disguise the fact that the assets are illegitimate, or (3) do both. One of the most effective ways of accom- plishing this is to interpose accounts under different names or separate legal persons or other secrecy vehicles or between the origin of the illegal funds and the ultimate control by the corrupt PEP. Alternatively expressed, the successful corruptor simply utilizes the bene�t granted to him (for example, contract, franchise, license, permit, and so on) in a legal way, including tax payments that lawful tax planning requires, while the corrupted of�cial (or bribe-demander) may receive bene�ts the legitimacy of which will appear dubious if a serious, high-quality investigation occurs and must therefore launder the bene�ts or have them prelaundered. Of course, if the corrupted (or extorting) of�cial is bypassed because the briber pays the of�cial’s associates or political sup- porters on the of�cial’s behalf, then the connection with the corrupt bene�t may be dif�cult to prove. We now consider case studies of grand corruption, irrespective of whether or not the AML framework might have made a signi�cant con- tribution to the prevention, investigation, prosecution, or recovery of the proceeds. To the extent that the AML framework lacks utility, our understanding of the conditions under which th