MACROECONOMICS, TRADE AND INVESTMENT E Q U I TA B L E G R O W T H , F I N A N C E & I N S T I T U T I O N S N OT E S Offshore data leaks and tax enforcement in developing countries Pierre Bachas ESSEC Business School and World Bank Matthew Collin EU Tax Observatory Tatiana Flores World Bank, DIME Thiago Scot World Bank, DIME Hao Lyu EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 1 © 2024 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. 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. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. 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 this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Cover design: Anatol Ursu, https://www.behance.net/olywebart EQU ITABLE GROWTH, FINANC E & INSTITUTIONS NOTES Offshore data leaks and tax enforcement in developing countries Pierre Bachas ESSEC Business School and World Bank Matthew Collin EU Tax Observatory Tatiana Flores World Bank, DIME Thiago Scot World Bank, DIME Hao Lyu Table of contents ACKNOWLEDGEMENTS 5 1. INTRODUCTION 6 2. BACKGROUND AND MOTIVATION 9 2.1 The scale and nature of offshore evasion 9 2.2 Policy efforts to reduce offshore evasion 10 3. DATA AND METHODOLOGY 12 4. WHAT CAN WE LEARN FROM LEAKED DATA? 15 5. MATCHING WITH ADMINISTRATIVE DATA: CASE STUDIES IN HONDURAS, ECUADOR, AND SENEGAL 20 5.1 Honduras 20 5.2 Ecuador 23 5.3 Senegal 24 6. DISCUSSION AND CONCLUSION 26 REFERENCES 28 APPENDICES 31 A Appendix Figures and Tables 32 B Tax Haven List 33 C Methods for Matching Leak Data and Administrative Records 34 4 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE ACKNOWLEDGEMENTS This note was prepared by Pierre Bachas (ESSEC Business School and World Bank), Matthew Collin (EU Tax Observatory), Tatiana Flores (World Bank, DIME), Thiago Scot (World Bank, DIME), and Hao Lyu. We express our gratitude to the Global Tax Program (GTP) for their generous support in funding this research. Additionally, we extend our appreciation to the Norwegian Agency for Development Cooperation (Norad), grant number QZA-22/0011, for providing funding, which supported a portion of the researchers’ time. We thank the SAR authorities in Honduras, Marlon Ochoa, Christian Duarte, and Alessandra Díaz, for their endorsement to conduct this research. We thank Gabriel Oqueli for his support as a research assistant. We thank Kinnon Scott (World Bank, Resident Representative LCCHN), Anne Brockmeyer (World Bank, EMFTX), Isabel Chiri (World Bank, EFI-MTI), and Yara Esquivel (World Bank, EFI-FCI) for their valuable comments. EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 5 INTRODUCTION The past decade has seen a rapid increase in the number of data leaks from tax havens, ranging from the Luxembourg Leaks to those revealed in the Panama and Pandora Papers. While these leaks often have political impact, leading to investigations of politicians with offshore accounts and calls by civil society for more transparency in corporate ownership, they can also be useful for tax authorities interested in investigating cross-border tax evasion and avoidance. To demonstrate this, this report uses leaked data published by the International Consortium of Investigative Journalists (ICIJ) in its Offshore Leaks Database (OLD) to document the relevance of the data for three countries where the authors have actively engaged with the national tax administrations: Ecuador, Honduras, and Senegal. 6 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE The ICIJ’s OLD comprises five separate data leaks, spanning roughly 2013 to 2021. Its published format enables us to use these leaks to identify links between firms (often shell companies) formed in a number of tax havens and those firms’ individual shareholders and ultimate beneficial owners. While owning or controlling an offshore shell company is not itself illegal,1 it is often considered a risk factor for engaging in some form of tax evasion or avoidance, as many tax authorities in low- and middle-income countries do not have access to information on cross-border ownership. This report presents a practical guide that tax authorities can follow for cleaning the ICIJ data ahead of merging it into their taxpayer data. We then use the ICIJ data to document the distribution of offshore company formation across jurisdictions for the shareholders and ultimate beneficial owners from each of the three countries, as well as the relative prevalence of each country in the overall set of leaks. To understand how often taxpayers appear in the leaks, we then match these data with the personal income tax (PIT) register of the three countries. The match rates vary substantially: in Honduras 89% of leak covered individuals can be matched to a registered taxpayer with a tax identification number and 86% can be matched to a registered taxpayer who has filed some form of domestic income; in Ecuador 90% can be matched to a tax identification, but we can assign a domestic income only to 56%. Finally, in Senegal only 34% of individuals can be matched to a taxpayer (and all of them have reported income). In the three countries, individuals identified in the ICIJ data are among the highest-earners, and represent a disproportionate share of the top 100 richest individuals and of the top 0.01% richest individuals (ranging from 6% in Senegal to 20% in Ecuador). In Honduras, using additional shareholder data, we also show that individuals that appear in leaked data are much more likely to own substantial shares in domestic companies. We also discuss the reasons for the disparate results in Honduras and Ecuador versus Senegal and highlight that leaked data may not always provide a complete picture of offshore ownership. This report underscores the critical need to establish formal, systematic, and regular mechanisms for information exchange between jurisdictions. This holds particular significance for tax authorities in developing countries with constrained access to cross-border ownership data. While implementing such mechanisms may require time, this practical guide can provide an interim solution. It empowers tax authorities to strengthen their capabilities to address cross-border tax evasion activities, to promote transparency, and to inform policy choices. The rest of this report proceeds as follows. Section 2 summarizes recent research on the magnitude of tax evasion using offshore structures and the impact of current policy efforts aimed at curbing it. In Section 3 we describe how we processed the ICIJ data into a format ready for linking to PIT registers and how the linking takes place. Section 4 contains a descriptive analysis of offshore ownership in our three countries held by individuals who appear as shareholders or beneficial owners in the ICIJ data. Section 5 presents the results from our matching with administrative data in Honduras, Ecuador, and Senegal. In Section 6 we provide our conclusions. 1 This is highlighted in ICIJ’s data disclaimer, which reads “There are legitimate uses for offshore companies and trusts. The inclusion of a person or entity in the ICIJ Offshore Leaks Database is not intended to suggest or imply that they have engaged in illegal or improper conduct. Many people and entities have the same or similar names. We suggest you confirm the identities of any individuals or entities included in the database based on addresses or other identifiable information. The data comes directly from the leaked files ICIJ has received in connection with various investigations and each dataset encompasses a defined time period specified in the database. Some information may have changed over time.” EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 7 2 8 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE 2. Background and motivation 2.1 The scale and nature of offshore evasion Recent evidence suggests that the use of offshore The extent to which offshore wealth is concentrated at structures to evade tax is at least as big a problem in the top varies across countries. By merging tax records low- and middle-income countries as it is in high-income with data from the Panama Papers’ leaks, Londoño- countries. Globally, Alstadsæter et al. (2023) estimated Vélez and Ávila-Mahecha (2021) found a relatively more the offshore financial wealth held by households to be concentrated distribution in Colombia, suggesting that around US$12 trillion, equivalent to 12% of world GDP in 40% of the top 0.01% evade taxes by concealing one-third 2022. of their wealth offshore. On the other hand, Leenders et al. (2023) estimated a less concentrated distribution of Empirical evidence suggests that individuals at the top of offshore wealth in the Netherlands. Exploiting an amnesty the income and/or wealth distribution evade a substantial program that allows eligible corporate and individual amount of taxes using offshore accounts. For Scandinavian taxpayers to declare offshore wealth without incurring civil countries, Alstadsæter, Johannesen, and Zucman (2019) penalties or criminal exposure, Leenders et al. estimated employed leaks data from the ICIJ and estimated that that the top 0.01% owns 7% of wealth declared through those in the top 0.01% evade 25% of their true tax liability the Dutch amnesty, while the top 90% to 99.9% group through the use of offshore financial services. In the owns around 67% of the declared wealth. United States, Guyton et al. (2021) analyzed random audit data and predict that the percentage of hidden wealth as a fraction of true income rises from 7% in the bottom 50% wealth distribution to over 20% in the top 1%. EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 9 2.2 Policy efforts to reduce offshore evasion Two major types of policy tools can help tackle offshore taxpayer information from counterpart jurisdictions. tax evasion. The first are voluntary disclosure schemes or Research has shown that signing these treaties leads tax amnesties, in which tax authorities temporarily reduce to a decline of offshore wealth held by residents of the cost of reporting past tax evasion to allow taxpayers to signatory countries (Johannesen and Zucman 2014; become compliant. The second are information exchange Johannesen 2014; Caruana-Galizia and Caruana-Galizia treaties under which tax authorities, either on demand 2016; Omartian 2017; Menkhoff and Miethe 2019). Two or automatically, agree to exchange information on factors are important when considering the impacts of assets held by their respective taxpayers in counterpart these policies. First, signing treaties triggers relocation of jurisdictions. deposits to non-cooperating tax havens after the treaties take effect (Johannesen and Zucman 2014; Caruana- Voluntary disclosure schemes To encourage the Galizia and Caruana-Galizia 2016). Second, the impact of compliance of those who have delinquent tax payments, such treaties does not seem to last in the long run. The governments take advantage of legal tax breaks such as long-term effect of treaties is undermined by tax evaders’ Voluntary Disclosure Schemes (VDS). In Norway, a VDS adaptation to them and by the emergence of new evading policy led to a 60% increase in reported net wealth, a responses (Menkhoff and Miethe 2019). 25% increase in income, and a 30% increase in tax paid (Alstadsæter et al. 2022). In the US, the Offshore Voluntary Common Reporting Standard (CRS) To combat Disclosures (OVD) program led to a sharp increase in tax evasion globally, the Organisation for Economic reported taxable capital income among participants Cooperation and Development (OECD) established an (Johannesen et al. 2020). These researchers also information standard for the Automatic Exchange of observed an increase in capital income reported through Information (AEOI) between tax authorities, known as other programs, namely the Report of Foreign Bank and the CRS. In contrast to bilateral information exchange Financial Accounts (FBAR). The increase in FBAR was due agreements, the CRS facilitates the automatic, annual to a significant number of quiet disclosers who responded exchange of information on all financial wealth and income to enforcement initiatives without admitting tax evasion held in financial institutions (bank deposits and dividend explicitly. Finally, Argentina’s tax amnesty program led to and interest payments). Since its inception in 2017, more the disclosure of wealth worth roughly 20% of the country’s than 110 countries have begun participating in the CRS. GDP, including a significant surge in reporting offshore Research has shown that the adoption of the CRS leads wealth (Londoño-Vélez and Tortarolo 2022). Disclosure to a substantial reduction in deposits held in tax havens schemes are often complemented by an increase in the tax by residents of participating countries (Casi, Spengel, and authority’s ability to observe offshore wealth, which can be Stage 2020).2 As a consequence, however, some holders enhanced through information exchange agreements. of offshore wealth have relocated their holdings to non- cooperating countries, particularly the United States (Casi, Information-exchange agreements Governments engage Spengel, and Stage 2020). In addition, some tax evaders in various forms of information exchange agreements, now use new strategies to avoid being reported on, such as including the Convention on Mutual Administrative using citizenship-by-investment schemes to hide their true Assistance in Tax Matters, a comprehensive multilateral tax jurisdiction (Langenmayr and Zyska 2023) and moving instrument that encompasses activities such as their wealth into non-reported assets, such as real estate. information exchange, including automatic exchanges, Researchers estimate that, globally, around 25% of the and recovery of foreign tax claims, among other functions. wealth that left tax havens due to AEOI was transformed Moreover, governments can establish bilateral information into real estate (Bomare and Herry 2022). exchange treaties to facilitate the specific request for 2 The Foreign Account Tax Compliance Act (FATCA) forces foreign financial institutions to automatically collect and transfer financial account information on US citizens to the Internal Revenue Service (IRS). 10 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE Slow progress in developing countries Many low- information sharing through the CRS is reciprocal (countries and middle-income countries have been slow to take must send reports to receive them), nonadopting countries up the OECD’s CRS, due to the technical difficulties in lack the same information set. Barring access to CRS data, implementing a complex reporting standard, combined leaked data may provide information that tax authorities with lower capacity. Out of the three countries in our report, can use to better understand offshore ownership by their as of 2023 only Ecuador had adopted the CRS. Since 3 tax residents. Box 1: How much tax revenue was recovered after release of the Panama Papers? Since the Panama Papers were first released in 2016, many countries have taken enforcement actions directly related to the information contained in the papers. The International Consortium of Investigative Journalists (ICIJ) has compiled instances of countries claiming to have used the papers’ revelations to collect tax revenue; it estimated that a total of $1.4 billion would be collected by 2021 (see ICIJ, 2021). The United Kingdom, Germany, Spain, France, and Australia alone claim to have recovered almost $900 million in total. The Australian Taxation Office, for example, has performed over 500 reviews and audits of Australians named in the Panama Papers.a Belgium has opened more than 200 investigations and has already collected over €15 million in taxes.b a. Information available here b. See The Brussels Times Note 3 OECD (2023), Tax Transparency in Latin America 2023: Punta del Este Declaration Progress Report, Global Forum on Transparency and Exchange of Information for Tax Purposes, OECD, Paris, https://www.oecd.org/tax/transparency/documents/tax-transparency-in-latin-america-2023.pdf EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 11 3. Data and Methodology The goal of the data exercise described in this paper is to owners for each country of interest using three steps: (i) understand the prevalence in the ICIJ’s Offshore Leaks assigning officers to countries, (ii) selecting firms, and dataset of taxpayers from the three countries we focus (iii) cleaning name lists. All code used to obtain the ICIJ on and the degree to which the data can be merged with database, to clean it, and to assign individuals to countries taxpayer registers to gauge the incidence of offshore (as well as the coded needed to replicate results with ownership along the income distribution. To do this, we publicly available data in this report) is available on Github. generated a list of names of shareholders and beneficial 12 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE The first step: Assign officers to countries. The ICIJ OLD dataset contains one database of officers (comprising shareholders and beneficial owners as well as directors and secretaries).4 We assigned these officers to specific countries (the presumptive country of tax residence), whenever possible using the ICIJ’s own assignment process. Particularly for the Pandora Papers data, ICIJ used additional variables to link officers with a country, including nationality, residency, and country of birth.5 We considered three criteria, assigning a given officer to a country if he or she complies with any of them: 1. The ICIJ database originally assigns that officer to the country. 2. The officer has at least one registered address in that country. (Officers are usually linked to one or more addresses through an ICIJ-provided identifier linking them to the OLD’s database of addresses. The addresses could be either residential or business.) 3. The individual is an officer of an entity that has a physical address in that country. The second step: Select firms. The goal of this step was to trace the firms of which the officers were beneficial owners. Thus, we assigned firms to a given country if at least one officer had been assigned to that country in the previous step. Even if a firm has a physical address elsewhere, we still flagged that firm under the country of its beneficial owners. If it has multiple beneficial owners with multiple nationalities, we linked that firm to all the countries. The final step: Create a list of names in Honduras, Senegal, and Ecuador that can be matched to taxpayers. In creating our list we followed these rules: 1. If an individual name appears to be joint tenancy (more than one individual in the same observation), we parsed the name column and kept all the names as separate officials. 2. We clean names in different patterns (e.g., names that appear in the format “last name, first name”; names including “Ms.” or “Mr.”) to facilitate matching by tax authorities. 4 For our analysis, we consider only shareholders and beneficial owners, given that these are the officers that ultimately control or benefit from the wealth held through shell companies. 5 ICIJ details how they link officers to countries: https://offshoreleaks.icij.org/pages/faq#country_links EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 13 4 14 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE 4. What Can We Learn from Leaked Data? Based on the data cleaning procedures described in Pandora Papers, and over 80% of officers from Ecuador the previous section, we built a database of individuals are connected to the Panama Papers. Data from Senegal (officers), with their country of residence, and of legal is limited and mostly evenly divided between the Panama entities located in tax havens. The descriptive statistics and the Pandora Papers. for the entire dataset, and for each of the three case study Table 1 Panel B provides the breakdown of jurisdictions countries, are presented in Table 1. of the 413,937 legal entities in tax havens documented Table 1 Panel A shows the number of individuals, their in the data, of which 182 are linked to Honduras, 33 to role with respect to the entity opened in a tax haven, and Senegal, and 686 to Ecuador (also shown graphically in the source of the leak that identified them. The entire ICIJ Figure 1). The most frequently used location for offshore dataset includes information on more than a half million legal entities is in the British Virgin Islands (BVI), with a individuals. Of those, we can assign 178 to Honduras, 30% worldwide share, followed by Malta, with an 18% only 38 to Senegal, and 939 to Ecuador. The majority of share, and Barbados, with a 10% share. The countries we individuals associated with the three case study countries study, however, have a relatively skewed composition of are shareholders or beneficial owners, with some also legal entities: 70% of the companies linked to Honduras serving in management roles (e.g., as directors), and a officials in the BVI and 7% in Panama; 60% of those smaller number of individuals filling both roles.6 The linked to Senegal individuals are in the BVI and a few most important leak in terms of coverage is the Paradise in each of Malta and the Seychelles; and 40% of those Papers (slightly over 50% of all officers are linked to the linked to Ecuador are in the BVI and 20% in Panama. This Paradise Papers). For the case study countries, however, highlights the varying geography of offshore entities and over 80% of officers linked to Honduras come from the the importance of regional patterns. 6 We note that these shares can sum to more than 100% since individuals might hold several different positions in different entities or even in the same entity. EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 15 Table 1: Descriptive Statistics for Leaked Individuals and Entities Entire Sample Honduras Senegal Ecuador Panel A: Individuals N Individuals 524,736 178 38 939 Shareholder or Beneficial Owner (%) 60.97 94.38 100 98.2 Management (Director, Secretary, Judicial) (%) 47.85 5.06 7.89 3.8 Both Shareholder/BO and Management(%) 14.67 1.12 7.89 2.64 Paradise Papers (%) 52.61 6.18 18.42 4.22 Panama Papers (%) 22.56 12.92 39.47 82.89 Pandora Papers (%) 5.89 80.9 42.11 12.25 Other Sources 18.94 0 0 0.63 Panel B: Legal Entities N entities 413,937 182 33 686 N jurisdictions 59 8 6 18 British Virgin Islands (%) 30.13 70.49 60.61 38.61 Malta (%) 18.09 1.09 12.12 1.16 Other Jurisdictions 17.57 2.19 3.03 5.81 Barbados (%) 9.64 0 0 0 Undetermined (%) 7.83 15.85 3.03 6.53 Panama (%) 5.61 7.1 9.09 22.79 Bahamas (%) 4.74 0 0 4.93 Seychelles (%) 3.24 0.55 12.12 2.18 Bermuda (%) 2.22 2.73 0 2.32 British Anguilla (%) 0.66 0 0 9.43 Nevada (%) 0.26 0 0 6.24 Source: Original calculations for publication Note: This table presents data on the distribution of individuals and legal entities across different jurisdictions based on the sample data. The data is divided into two panels: Panel A focuses on individuals, and Panel B provides information about legal entities. Figure 1: Distribution of Jurisdictions Where Firms in Leaked Data Were Incorporated 100% Share of firms incorporated 75% in each jurisdiction 50% 25% 0% World Ecuador Honduras Senegal British Virgin Islands Panama Malta Seychelles Other Source: Original calculations for publication Note: Data were retrieved from ICIJ Leaks Database in July 2022. The graph compares the distribution of jurisdictions related to each country and to the world. The authors chose four major jurisdictions in the dataset. The share of the remaining jurisdictions is summed up as “Other.” 16 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE How prevalent is the use of tax havens across countries on average high income countries have 85 individuals and how does use of these havens change with income per million inhabitants, as opposed to 22 for an low- and levels? Figure 2 shows the number of individuals named middle-income countries. The case study countries are in leaks per million inhabitants, plotted against country’s close to the average predicted by their income levels, with per capita GDP (in log). We observe a substantial increase 2 for Senegal, 17 for Honduras, and 54 for Ecuador. in the leaked use of tax havens in higher income countries: Figure 2: Individuals in Offshore Leaks Database (OLD) by Country GDP Per Capita 500 United Arab Emirates Individual in data leaks (per M. inhabitants) 400 300 Ecuador Honduras Senegal 200 100 0 6 7 8 9 10 11 log (GDP per capita) Source: Original calculations for publication Note: Data were retrieved from ICIJ Leaks Database in July 2022. The graph compares the distribution of jurisdictions related to each country and to the world. The authors chose four major jurisdictions in the dataset. The share of the remaining jurisdictions is summed up as “Other.” EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 17 Figure 3 depicts the distribution of the share of population America, Scandinavia, and Oceania countries. The map listed in the Offshore Leaks Database, represented by also highlights some interesting outliers: Mexico displays a quartiles indicated by different colors. Notably, we identify relatively low use of tax havens, while Argentina exhibits a relatively fewer individuals from African countries in the relative higher use compared to its neighbors. Overall, the leaks database, as a share of total population, compared map reveals significant disparities in tax haven use across to many countries in South America, Eastern Europe, regions, shedding light on the global dynamics of financial and Central Asia. The largest relative use of tax havens secrecy and offshore practices. identified in the Offshore Leaks Database is found in North Figure 3: Share of Population Named in the Offshore Leaks Database (OLD) Quartiles: [0.1 - 3.2] [3.2 - 17.7] [17.7 - 66.8] [66.8 - +] Source: Original calculations for publication Note: Leaks data were retrieved from ICIJ Leaks Database. The graph shows the quartile of the number of leaks per million inhabitants in each country. The red dots represent tax havens. We excluded them when calculating the quartiles because officers related to tax havens (citizens or residents) have minimal or no tax liability naturally. Gray shading indicates countries for which no data is available. Another interesting fact arises from the time patterns of around 2014. It is important to note that post-2015 data incorporation of tax haven entities. Figure 4 shows in the is limited, and the observed post-2015 fall is due to the top-left panel the worldwide trends between 1990 and nature of the data. The rising pattern over time prior to 2020. The other three panels show the incorporation years 2015 is likely to be more accurate, although it should also for entities linked to individuals in Ecuador, Honduras, and be taken cautiously, given the cross-section nature of the Senegal. Incorporation rises over time, with a peak at data leaks. 18 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE Figure 4: Year of Incorporation of Entities in Tax Havens 40000 80 35000 World 70 Ecuador 30000 60 25000 50 20000 40 15000 30 10000 20 5000 10 0 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 18 6 15 Honduras 5 Senegal 12 4 9 3 6 2 3 1 0 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Paradise Papers Pandora Papers Panama Papers Offshore Leaks Source: Original calculations for publication Note: Data retrieved from ICIJ Offshore Leaks Database on July 2022. For Honduras, the authors only selected firms that were incorporated after 1990. Six firms incorporated between 1950 and 1990 were dropped from this graph. The total number of entities related to officers is 182. The year with the largest number of firms incorporated is 2016. For Senegal, the first entity was incorporated in 2003. The total number of entities related to Senegalese officers is 33. All firms in the ICIJ Leaks Data are on the graph. The largest number of incorporations occurred in 2016. For Ecuador, the authors dropped firms incorporated before 1995 from this graph. The first entity related to Ecuador was incorporated in 1905, and the total number of entities related to Ecuadorian officers is 686. The month showing the largest number of incorporations is January 2014, with the number of incorporations intensifying after February 2007. EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 19 5 20 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE 5. Matching with Administrative Data: Case Studies in Honduras, Ecuador, and Senegal 5.1 Honduras For Honduras, we identified 178 officers in the ICIJ data. How do the characteristics of the individuals we match Starting with those 178 names, we were able to uniquely differ from the rest? One prior expectation was that match 159 of them to a taxpayer in Honduras — a 89% individuals who have offshore accounts were likely to be match rate. These unique matches were determined, high-income earners: Londoño-Vélez and Ávila-Mahecha jointly with the Tax Authority, to refer to the same (2021) documented a steep increase in the probability individual. (We provide more details about the matching of being named in the Panama Papers across the wealth process in Appendix C.3.) distribution in Colombia. We report results for a similar exercise in Honduras, where we first ranked individuals We then proceeded to match these individuals to annual by their total income and then computed the share of databases that collect all income attributed to any taxpayer individuals in each group observed as being officers of an — these include those filing personal income taxes but offshore firm in the ICIJ data. also others for whom their only income is withheld at source, such as wages, capital gains, or dividends. We also Those results are presented in Figure 5, which pools all included undistributed profits from corporations in which years in the period 2016 to 2020. Among those ranked with individuals are shareholders, that is, we attributed profits the 100 highest incomes in each of those years, almost as income to shareholders in proportion to their shares in 15% are observed in the ICIJ data. The share is still high for the company. For the period 2016 to 2020, we were able those individuals between top 500 to 200, where close to to observe some declared or assessed income for 153 1 in 20 are observed in the leaks data, while below the top individuals — the remaining six did not declare or receive 1,000 the probability quickly decreases toward zero.7 This any income that could be registered by the tax authority, is compelling evidence that the use of offshore accounts is even though they were identified as valid taxpayers. particularly common at the top of the income distribution, 7 We present a similar graph for each year separately in Figure A1. The results are noisier but broadly consistent with the pooled sample. EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 21 and therefore compliance risks related to the use of such are shareholders. Among all 159 individuals identified in instruments are likely also concentrated at the top. the leaks data and matched to domestic taxpayers, 97, or 61%, were shareholders in some corporation that filed Another stylized fact we observed about individuals taxes in 2020. Not only they were these individuals more present in the leaks data is that they were more likely likely to be shareholders in corporations, they were also to be shareholders of domestic corporations. We used a particularly likely to be shareholders in large corporations. newly constructed registry of corporate shareholders in If we only consider the largest 500 corporations in terms Honduras to document whether each individual in our of turnover in 2020, 11% of those identified in the leaks income database was a shareholder of any company.8 data were shareholders, versus less than 0.1% for the For 2020, we observed approximately 670,000 individuals remaining taxpayers; and 4% of those identified in received some income. As shown in Figure 6, the probability leaks were shareholders in one of the top 100 largest that one of these taxpayers would be a shareholder of a corporations, versus 0.02% for others.9 domestic company was close to 8%; approximately 50,000 Figure 5: Share of Taxpayers Matched to Leaks Data (Honduras) .2 Share of taxpayers in leaks data .15 .1 .05 0 Top 20,000 Top 10,000 Top 5,000 Top 4,000 Top 3,000 Top 2,000 Top 1,000 Top 500 Top 400 Top 300 Top 200 Top 100 Quantiles of total income distribution Source: Original calculations for publication Note: This figure reports the probability of a taxpayer being matched to the leaks data, by quantiles of total individual income. The data is pooled for the 2016–2020 period (i.e., the “Top 100” group refers to 500 taxpayer-year observations, those who were among the top 100 highest earners in each year). 7 We present a similar graph for each year separately in Figure A1. The results are noisier but broadly consistent with the pooled sample. 8 We define an individual as being a shareholder if we assign them as owning at least 1% of any corporation. We use that restriction to exclude defining as shareholders individuals who receive dividends from financial corporations and therefore are classified as owning shares but with tiny participations. 9 In Table A1, we show that this gap in the probability of being a shareholder is not fully explained by differences in income: even after conditioning on income, taxpayers appearing in the leaks data were much more likely to also be shareholders in a domestic company, although the gap is smaller when we focus exclusively on the highest-income individuals. 22 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE Figure 6: Share of Taxpayers as Shareholders of Corporations (Honduras) 0.61 0.6 Percentage who are shareholders 0.4 0.2 0.08 0.11 0.04 0.00 0.00 0 Not on leaks Matched to leaks Own shares in any company Own shares of top 500 corporations Own shares of top 100 corporations Source: Original calculations for publication Note: This figure reports the probability of individual taxpayers being shareholders in corporations filing income tax in 2020. We define shareholders as taxpayers holding at least 1% participation in some company. Top 100 and top 500 companies are classified based on turnover in 2020. Figure 7: Share of Taxpayers Matched to Leaks Data (Ecuador) .2 Share of taxpayers in leaks data .15 .1 .05 0 [99.95-99.99) [90-95) [95-99) [99-99.5) [99.5-99.9) [99.9-99.95) [99.99-100] Quantiles of total income distribution Source: Original calculations for publication Note: This figure reports the probability of a taxpayer being matched to the leaks data, by quantiles of total individual income. The data is pooled for the 2016–2020 period. EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 23 5.2 Ecuador For Ecuador, we identified 939 individuals in the Offshore individuals at the top of the income distribution were Leaks Database, most of them as shareholders or much more likely to be identified in the Offshore Leaks beneficial owners (98%). Of those, 854 have a registered Database. When we pool individuals for the period 2016 tax identification number based on the publicly available to 2020, approximately 20% of those in the top 0.01% of repository of taxpayers ID (RUCs). Using these tax the income distribution (around 60 individuals each year) identification numbers we were able to find a filled tax are matched to the OLD. This share falls rapidly outside the return in the administrative tax data for 530 individuals, very top of the distribution — for those just above the top thus giving us a 56% match rate to current income filers 0.1% the rate is closer to 5%, and it falls to virtually zero (but a 91% rate to tax identification numbers). Similar to our among those between the 90th and 99th percentiles.10 findings in Honduras, we also document that Ecuadorian 5.3 Senegal We only identified 38 individuals connected to Senegal in distribution.11 While the number of individuals we were the Offshore Leaks Database. Using first and last names able to identify as domestic taxpayers was very small, we of officers and taxpayer data for 2019 and 2020, we were again observed that these came disproportionately from able to match 13 of those to taxpayers in the country (a the top of the income distribution: 6 of the 13 identified 35% match rate). In Figure 8, we present the probability individuals were among the top 100 earners. that taxpayers were identified in the OLD across the income Figure 8: Share of Taxpayers Matched to Leaks Data (Senegal) .1 Share of taxpayers in leaks data (%) .05 0 Top 150,000 Top 100,000 Top 50,000 Top 20,000 Top 10,000 Top 5,000 Top 4,000 Top 3,000 Top 2,000 Top 1,000 Top 500 Top 400 Top 300 Top 200 Top 100 Quantiles of total income distribution Source: Original calculations for publication Note: This figure reports the probability of a taxpayer being matched to the leaks data, by quantiles of total individual income in Senegal. The data is pooled for the 2019–2020 period. 10 For more details on the construction of the income variable, see Brounstein (2023). 11 Individuals are ranked according to their average incomes in 2019 and 2020. For more details on the construction of income variables, see Bachas et al. (2023). 24 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE Box 2: Using leaks data to understand wealth tax evasion in Colombia Colombia is one of the few middle-income countries with a long-standing tradition of wealth taxation. In a study of how taxpayers responded to a tax amnesty that allowed them to declare previously unreported wealth, Londoño-Vélez and Ávila- Mahecha (2021) illustrate a potential use case for the ICIJ leaks in tandem with administrative tax data. Londoño-Vélez and Ávila-Mahecha studied a program implemented in 2014 that allowed taxpayers to disclose previously unreported assets and debt. In exchange for paying 10% of the disclosed net wealth in fines, the taxpayers were exempt from past income and wealth taxes on those hidden assets. Almost 12,000 taxpayers disclosed new assets in the amount of 1.7% of GDP; over 0.2% of GDP was collected in fines. Furthermore, almost 90% of assets had been concealed offshore, highlighting the importance of monitoring offshore accounts for tax compliance. Out of 1,751 individuals with names in the Panama Papers leak associated with addresses in Colombia, Londoño-Vélez and Ávila-Mahecha (2021) were able to match 1,208 to taxpayers in Colombia using their full names. They attribute this high match rate to the “naming custom of Hispanic America practiced in Colombia, which involves two given names, plus a paternal surname, followed by a maternal surname.” This is a very similar to results in Honduras and Ecuador, where an even higher match rate of 89 to 90% was attained and also illustrates why the same methodology might fail in countries with different naming conventions such as Senegal. The authors used the matched data from leaks to administrative data to perform two exercises. First, they used leaks data to validate how the probability of having hidden wealth varies across the wealth distribution. They started by documenting that individuals at the very top of the distribution are much more likely to report hidden wealth as compared to those lower in the distribution — 40% of those in the top 0.01% disclose hidden assets versus 3% of those in the top 5%. But this does not necessarily illustrate the true hidden wealth: it is possible that these individuals at the top are simply more likely to report their hidden assets in response to the reform. To assess whether that is the case, the authors performed a similar exercise and computed the probability of individuals appearing in the leaks data across the distribution. While these data do not necessarily cover the universe of offshore owners, it does not suffer from the same selection problem that affects self-declaration in response to the reform. What the authors found was a similar pattern across the wealth distribution: individuals at the top 0.01% are 13 times as likely to appear in the Panama Papers as compared to those in the top 5%. A second exercise described in the paper exploited the timing of the release of the Panama Papers to investigate how the probability of punishment affected the decision to self-report hidden assets. The disclosure of the Panama Papers in 2016 included the names of wealthy Colombians who owned offshore wealth. The disclosure occurred after the first year in which taxpayers could self-declare their hidden assets, but before the second round. The authors exploited that information shock to perform a differences-in-differences analysis: they compared the probability of disclosing assets and amounts disclosed between those appearing and not appearing on the leaks data, before and after the disclosure. What they show is that those appearing in the Panama Papers were 30 percentage points more likely to report foreign assets and paid 26% more wealth taxes in the years after the leak, as compared to those not appearing in the leaks. This provides strong evidence that being named in the Panama Papers caused taxpayers to disclose their assets — likely in response to their perception that they faced a higher likelihood of punishment after their offshore accounts were made public. EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 25 6. Discussion and Conclusion This report provides insights on how policy makers can use information from the ICIJ Offshore Leaks Database (OLD) and upgrade their tax compliance risk models. Second, the data could support the identification of or other similar financial data leaks to strengthen domestic irregular patterns of financial transactions as well tax compliance enforcement. It illustrates the use case as potential mechanisms for tax evasion specific to of this data in three countries — Honduras, Senegal, and each country. Third, the reported pattern of offshore Ecuador — documenting different challenges through the wealth owned by individuals at the very top of the process and showing the prevalence of offshore owners income distribution reenforces the need for tax among the highest earners in these economies. These administrations to focus on this segment, for example, exercises complement the World Bank’s ongoing work via units focusing on high net worth individuals.13 with its client countries to improve tax and ownership Reck and Bomare (2022) elaborate a multipronged transparency and anti-corruption efforts (Cebreiro Gomez 12 approach to enforcement measures targeting high- et al. 2021). The following are additional implications and income individuals that involves (1) enhanced cross- opportunities arising from the analyses in this report. border information sharing, (2) thorough audits, (3) active pursuit of tax disputes, (4) significant penalties 1. Undertake targeted, data-driven tax enforcement on severe evaders, (5) vigorous monitoring and measures. This data exercise is relevant for improving enacting reforms addressing aggressive tax shelters, tax enforcement and compliance measures. First, by and (6) encouragement to whistleblowers to uncover uncovering the prevalence of taxpayers involved in evasion. offshore ownership, tax administrations can update 12 Much of this work is conducted through the World Bank’s Global Tax Program. 13 Brounstein (2023) found that individuals connected to tax havens had a significant and lasting increase in domestic reporting and income taxes paid, largely driven by increases in reported capital income and independent labor income, as a result of the outflows tax in Ecuador. This suggests that the behavioral responses of individuals exposed on leaked data may differ from the responses of unexposed individuals. 26 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE 2. Identify high-risk jurisdictions that require Assets Abroad (VDP) by using data from the OECD, enhanced enforcement initiatives. Breaking down which joint with other efforts has empowered them to the jurisdictions for legal entities in tax havens was accomplish gains in revenue collection. (OECD et al. a vital aspect of this exercise. It revealed that certain 2023). Honduras signed the Multilateral Convention locations, such as the British Virgin Islands (BVI), on Mutual Administrative Assistance in Tax Matters Malta, and Barbados, are favored for creating offshore (MAAC) in 2022, which allows signatory countries legal entities and that the specific locations favored to request and provide assistance in tax-related vary across the countries analyzed (Senegal, Ecuador, matters, including through the exchange of financial and Honduras). This knowledge can empower tax information. As of May 2024, however, Honduras had administrations to develop a prioritized list of higher- yet to ratify the agreement. Authorities should keep risk jurisdictions to use in improving enforcement in mind that signing and ratifying these agreements initiatives and increasing the quantity and quality is only the first step in effectively using financial of monitoring processes and auditing efforts at the information to improve compliance: countries should transaction and taxpayer level. When possible, tax establish a comprehensive implementation strategy administrations could prioritize these jurisdictions as for these agreements, including setting up systems for they strengthen international cooperation to combat efficiently tracking, processing, and analyzing received tax evasion and financial crimes, such as through data. Moreover, participating countries’ efforts should information exchange agreements. include the developing robust legal frameworks 3. Establish information exchange mechanisms. that align with international standards to ensure the Collaboration among jurisdictions in sharing financial actionable utility of the data. Finally, the usefulness of data plays a pivotal role in revealing concealed these information exchanges relies on the quality of assets and detecting tax evaders using cross-border data provided by the sending countries. For example, operations. This is particularly significant in the a report from March 2021 by the European Court of context of territorial tax systems, such as Honduras’s, Auditors found that information collected by Member where taxation is limited to income generated States lacks quality and completeness. Most Member within national boundaries.14 Monitoring cross- States do not adequately control the reporting entities, border financial activities becomes imperative for leading to failures in fulfilling their due diligence ensuring tax compliance and preventing tax evasion obligations under the CRS (Matras 2023). in territorial jurisdictions. Among the three countries This report provides key information that can significantly we analyzed, only Ecuador has implemented the improve tax enforcement, compliance, transparency, Common Reporting Standard (CRS) agreement, and anti-corruption efforts in developing countries such although Senegal has committed to implementing it as Honduras, Senegal, and Ecuador. Armed with this by 2025. Since it signed on to the agreement in 2021, knowledge, authorities can strengthen their tax systems, the Ecuadorian Internal Revenue Service (Servicio address offshore ownership issues, and take effective de Rentas Internas, or SRI) has implemented control steps toward curbing tax evasion and illicit financial measures that efficiently leverage the financial activities. Additionally, using this information can help account information it automatically receives. For foster international collaboration to tackle tax havens and instance, SRI complemented the Voluntary, Unique, ensure fair taxation worldwide. and Temporary Tax Regime for Regularization of 14 In worldwide tax systems, residents are obligated to report and pay taxes on their global income. Failure to do so, such as by not disclosing offshore accounts or by using offshore entities to conceal income, is considered tax evasion. In territorial tax systems, however, detecting tax evasion using offshore information is more difficult due to the emphasis on domestic source income. Additional details on transaction patterns and background are necessary to determine instances of tax evasion. EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 27 References Alstadsæter, Annette, Sarah Godar, Panayiotis Nicolaides, and Gabriel Zucman. 2023. “Global Tax Evasion Report 2024.” Alstadsæter, Annette, Niels Johannesen, Ségal Le Guern Herry, and Gabriel Zucman. 2022. “Tax evasion and tax avoidance.” Journal of Public Economics 206 (February): 104587. https://linkinghub.elsevier.com/retrieve/pii/S0047272721002231 Alstadsæter, Annette, Niels Johannesen, and Gabriel Zucman. 2019. “Tax Evasion and Inequality.” American Economic Review 109 (6): 2073–2103. https://www.aeaweb.org/articles?id=10.1257/aer.20172043 Bachas, Pierre Jean, Leo Czajka-Dubois, Aissatou Diallo, Justine Simone Cecile Knebelmann, and Anaelle Binta Toure. 2023. “Strengthening Personal Income Taxation in Senegal — Case Study.” World Bank Group, June 2023. https://policycommons.net/artifacts/4248515/strengthening-personal-income-taxation-in-senegal/5057288/ Bomare, Jeanne, and Ségal Le Guern Herry. 2022. “Will we ever be able to track offshore wealth? Evidence from the offshore real estate market in the UK.” Working Paper No. 4, EU Tax Observatory, June 2022. https://www.taxobservatory.eu/wp-content/uploads/2022/06/BLGH_June2022.pdf Brounstein, Jakob. 2023. “Can countries unilaterally mitigate tax haven usage? Evidence from Ecuadorian transaction tax data.” Job Market Paper. https://jakobbrounstein.github.io/files/brounstein_jmp_taxevasionecuador.pdf Caruana-Galizia, Paul, and Matthew Caruana-Galizia. 2016. “Offshore financial activity and tax policy: Evidence from a leaked data set.” Cambridge University Press, Journal of Public Policy 36 (3): 457–88. https://www.jstor.org/stable/26775350 Casi, Elisa, Christoph Spengel, and Barbara M. B. Stage. 2020. “Cross-border tax evasion after the common reporting standard: Game over?” Journal of Public Economics 190 (October): 104240. https://www.sciencedirect.com/science/article/pii/S0047272720301043 Cebreiro Gomez, Ana, Matthew Collin, Yara Esquivel, Yasmin de Magalhaes Pinto Almeida, Emile Johannes Marie Van Der Does de Willebois, and James Anderson. 2021. “Opening Pandora’s Box : What the Latest Leaks Mean for the World Bank’s Mission.” World Bank Group, Washington, DC. http://documents.worldbank.org/curated/en/883871637828758326/Opening-Pandoras-Box-What-the-Latest-Leaks-Mean-for-the-World- Banks-Mission Guyton, John, Patrick Langetieg, Daniel Reck, Max Risch, and Gabriel Zucman. 2021. Tax Evasion at the Top of the Income Distribution: Theory and Evidence. Technical Report w28542. National Bureau of Economic Research, Cambridge, MA, March 2021. http://www.nber.org/papers/w28542.pdf Johannesen, Niels. 2014. “Tax evasion and Swiss bank deposits.” Journal of Public Economics 111 (March): 46–62. https://www.sciencedirect.com/science/article/pii/S0047272713002454 Johannesen, Niels, Patrick Langetieg, Daniel Reck, Max Risch, and Joel Slemrod. 2020. “Taxing Hidden Wealth: The Consequences of US Enforcement Initiatives on Evasive Foreign Accounts.” American Economic Journal: Economic Policy 12 (3): 312–46. https://pubs.aeaweb.org/doi/10.1257/pol.20180410 Johannesen, Niels, and Gabriel Zucman. 2014. “The End of Bank Secrecy? An Evaluation of the G20 Tax Haven Crackdown.” American Economic Journal: Economic Policy 6 (1): 65–91. https://pubs.aeaweb.org/doi/10.1257/pol.6.1.65 28 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE Langenmayr, Dominika, and Lennard Zyska. 2023. “Escaping the exchange of information: Tax evasion via citizenship-by-investment.” Elsevier, Journal of Public Economics 221 (C). https://ideas.repec.org//a/eee/pubeco/v221y2023ics0047272723000476.html Leenders, Wouter, Arjan Lejour, Simon Rabaté, and Maarten van ’t Riet. 2023. “Offshore tax evasion and wealth inequality: Evidence from a tax amnesty in the Netherlands.” Journal of Public Economics 217 (January): 104785. https://www.sciencedirect.com/science/article/pii/S0047272722001876 Londoño-Vélez, Juliana, and Javier Ávila-Mahecha. 2021. “Enforcing Wealth Taxes in the Developing World: Quasi-Experimental Evidence from Colombia.” American Economic Review: Insights 3 (2): 131–48. https://www.aeaweb.org/articles?id=10.1257/aeri.20200319 Londoño-Vélez, Juliana, and Dario Tortarolo. 2022. “Revealing 21% of GDP in Hidden Assets: Evidence from Argentina’s Tax Amnesty.” Working Paper. https://www.taxobservatory.eu/publication/revealing-21-of-gdp-in-hidden-assets-evidence-from-argentinas-tax-amnesty/ Matras, Tomasz. 2023. “Analysis of the Functioning of the Automatic Exchange of Information about Financial Accounts after Implementation of the Common Reporting Standard.” Prace Naukowe Uniwersytetu Ekonomicznego we Wrocławiu 67 (1): 106–17. Menkhoff, Lukas, and Jakob Miethe. 2019. “Tax evasion in new disguise? Examining tax havens’ international bank deposits.” Journal of Public Economics 176 (August): 53–78. https://www.sciencedirect.com/science/article/pii/S0047272719300817 OECD (2023), Tax Transparency in Latin America 2023: Punta del Este Declaration Progress Report, Global Forum on Transparency and Exchange of Information for Tax Purposes, OECD, Paris, https://www.oecd.org/tax/transparency/documents/tax-transparency-in-latin-america-2023.pdf Omartian, James D. 2017. Do Banks Aid and Abet Asset Concealment: Evidence from the Panama Papers. SSRN Scholarly Paper. Rochester, NY. October 2017. https://papers.ssrn.com/abstract=2836635 Reck, Daniel, and Jeanne Bomare. 2022. “Different from You and Me: Tax Enforcement and Sophisticated Tax Evasion by the Wealthy.” LSE Public Policy Review 2 (4, (November): 6. https://ppr.lse.ac.uk/articles/10.31389/lseppr.71/ EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 29 30 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 31 A Appendix Figures and Tables Figure A1: Share of Taxpayers Matched to Leaks Data by Year (Honduras) .2 Share of taxpayers in leaks data .15 2016 2017 2018 2019 2020 .1 .05 0 Top 20,000 Top 10,000 Top 5,000 Top 4,000 Top 3,000 Top 2,000 Top 1,000 Top 500 Top 400 Top 300 Top 200 Top 100 Quantiles of total income distribution Source: Original calculations for publication Note: This figure reports the probability of a taxpayer being matched to the leaks data, by quantiles of total individual income, for each year in the 2016–2020 period. Table A1: Correlates of being a large shareholder (1) (2) Leaks taxpayer 0.486*** 0.154*** (0.04) (0.05) Log(total income) 0.011*** 0.175*** (0.00) (0.00) Observations 635,911 100,000 R-Squared 0.03 0.12 Sample Full Top 100,000 Source: Original calculations for publication Note: This table presents results from regressions where the dependent variable is an indicator of whether the taxpayer is a large shareholder in any company, defined as having at least a 1% share in a corporation. “Leaks taxpayer” is an indicator of whether the taxpayer is identified in the ICIJ database as being an officer in an offshore firm. Column (1) presents results using the whole sample for 2020, while column (2) restricts it to the top 100,000 taxpayers with highest yearly income. Robust standard errors are appear in parentheses. 32 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE B Tax Haven List Andorra Grenada Netherlands Antilles Anguilla Guernsey Niue Antigua and Barbuda Hong Kong Palau Aruba Ireland Panama Austria Isle of Man Samoa Bahamas Jersey Saint Kitts and Nevis Bahrain Jordan Saint Lucia Barbados Lebanon Saint Martin Belgium Liberia Saint Vincent and the Grenadines Belize Liechtenstein San Marino Bermuda Luxembourg Seychelles British Virgin Islands Macao Singapore Cayman Islands Malaysia Switzerland Chile Maldives Tonga Cook Islands Malta Trinidad and Tobago Costa Rica Marshall Islands Turks and Caicos Curaçao Mauritius Uruguay Cyprus Monaco US Virgin Islands Dominica Montserrat Vanuatu Gibraltar Nauru Vatican Source: Menkhoff and Miethe (2019). EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 33 C Methods for Matching Leak Data and Administrative Records Using the leaked data structure provided by ICIJ, we matched individuals’ names with the tax administrative records. This method proved notably effective in Latin American countries, where it’s customary for individuals to possess more than a single first and last name. We also followed a conservative approach by considering as a “match” only those cases that represented a perfect or close to perfect name match and thus could be uniquely identified in the administrative tax records. For instance, if a name in the leaked dataset corresponded with multiple individuals possessing distinct Tax IDs and supplementary details, such as addresses, but failed to definitively indicate the associated taxpayer, we excluded that particular instance from the final set of matched data. C.1 Matching Method for Ecuador In the case of Ecuador, the matching method involved a fuzzy match on individual names and addresses extracted from the ICIJ leaks when comparing them with the publicly available repository of taxpayer IDs (RUCs). Individual matches that were also associated with companies or produced cosine similarity distances greater than 0.4 were identified, resulting in a total of 854 individuals.15 C.2 Matching Method for Senegal Starting with the 38 individuals identified in the OLD, our goal was to match them to potential taxpayers registered in Senegal. To match administrative data with the officers present in the leaks, we performed a fuzzy matching with Reclink based on the first and last names of officers. We consistently ignored homonyms, as we were interested in unique matches. Our match used data for 2019 and 2020, which were structured in different ways: 2020. We used the consolidated database of all taxpayers with declared income in 2020, taking together all revenue datasets available for the year. We found 10 matches, including one that we identified as a “probable match” (28%). Two officers were matched to several observations (homonyms) (5%). The remaining 24 officers were not matched (67%). 2019. Matches had to be done individually with each dataset, offering more opportunity for homonyms (for names found in two datasets, it was difficult to determine whether they corresponded to the same person). We found 11 officers matched to at least 1 dataset (30%). One officer was matched to several observations, but never uniquely (homonyms) (3%). The remaining 24 officers were not matched (67%). 15 We thank Jakob Brounstein for his support on this task. 34 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE C.3 Matching Method for Honduras Starting with the 178 individuals identified in the OLD, our goal was to match them to potential taxpayers registered in Honduras. We started by comparing the 178 individuals’ names with a list of all taxpayers in Honduras, and after matching by similarity of name characters, we retained only matches with high levels of similarity. That provided us with a potential number of matches for each name in the OLD, which we then manually cleaned based on the following criteria: We only considered taxpayers in Active status. We looked for perfect or close to perfect name matching. In cases where the name similarity was not perfect, the Tax Authority performed additional research, including, for example, checking for matching addresses or for references to the same individuals in news articles about past data leaks. Special cases in which names could not be identified included: 1. Individuals in OLD for whom we found more than one taxpayer with a similar name, making it impossible to distinguish who was identified in the leaks data. In those cases, we considered a match to be infeasible. 2. Individuals in OLD for which we found no individuals with a similar name among registered taxpayers. The cleanup resulted in 159 names (89.3% of the total) that uniquely matched specific domestic taxpayers in Honduras to the leak data; 19 names remained unmatched. EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE <<< 35 36 <<< EQUITABLE GROWTH, FINANCE & INSTITUTIONS NOTE