Policy Research Working Paper 11091 Which Data Do Economists Use to Study Corruption? A Cross-Section of Corruption Research James H. Anderson Akanksha Baidya Institutions Global Department A verified reproducibility package for this paper is March 2025 available at http://reproducibility.worldbank.org, click here for direct access. Policy Research Working Paper 11091 Abstract This paper examines the data sources and methodologies Articles in ranked journals are more likely to employ admin- used in economic research on corruption by analyzing 339 istrative and experimental data and focus on the causes of journal articles published in 2022 that include Journal of corruption. The broader dataset of 882 articles highlights Economic Literature codes. The paper identifies the most the significant academic interest in corruption across dis- commonly used data types, sources, and geographical foci, ciplines, particularly in political science and public policy. as well as whether studies primarily investigate the causes The findings raise concerns about the limited use of novel or consequences of corruption. Cross-country composite data sources and the relative neglect of research on the indicators remain the dominant measure, while single causes of corruption, underscoring the need for a more country studies more frequently utilize administrative data. integrated approach within the field of economics. This paper is a product of the Institutions Global Department. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at janderson2@worldbank.org and abaidya1@worldbank.org. A verified reproducibility package for this paper is available at http://reproducibility.worldbank.org, click here for direct access. RESEA CY LI R CH PO TRANSPARENT ANALYSIS S W R R E O KI P NG PA The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Which Data Do Economists Use to Study Corruption? A Cross-Section of Corruption Research James H. Anderson Akanksha Baidya* Keywords: governance indicators, corruption data, administrative data, survey data, JEL classification system, public administration, political science JEL codes: D73, D22, O47, O57, P50 * James Anderson (janderson2@worldbank.org; corresponding author) is Lead Public Sector Specialist for East Asia and Pacific, World Bank; Akanksha Baidya (abaidya1@worldbank.org) is a Consultant for Public Administration Global Unit, Institutions Global Practice, World Bank. This study was made possible by the support of the Korea Development Institute School of Public Policy and Management and the World Bank’s Institutions Global Practice. We are grateful to Alexandra Habershon, Till Hartmann, David Bernstein, Galileu Kim, Heather Marquette, Elizabeth David-Barrett, Elitza Mileva, and Erica Bosio for helpful comments on earlier drafts, and to Roby Senderowitsch and Cheol Liu for support in facilitating the research. 1. Introduction As a field of research, the study of corruption is booming. Having grown from modest beginnings some 30 years ago with cross-country empirical analyses of corruption and its effects, 1 large numbers of empirical studies continue to examine the relationship between corruption and a range of socioeconomic outcomes. Even within a single field of study, health, corruption has been linked to antimicrobial resistance, HIV infection, child mortality, and road fatalities (Burki, 2019; World Bank, 2019). There are studies of the impacts on macroeconomic performance (e.g., Mauro, 1995), and microeconomic behavior of citizens, firms, and public officials (e.g., World Bank and Government Inspectorate of Vietnam, 2012). For any empirical study, a researcher needs data. While health outcomes or firm performance may be relatively easily measured, and the data easily obtained, data on corruption is more challenging since those engaging in the practice often do not wish to admit to doing so. The study of corruption, and the data available for that purpose, have grown up together over the years, although innovations in measurement have arguably not kept pace with interest in the topic. As noted in Marquette & Peiffer (2022), the “governance and development field is ill- equipped to back up claims made about the success or failure of anticorruption policies because there are no instruments to accurately measure corruption.” 1 Mauro (1995) used a dataset compiled from subjective risk assessments by Business International (BI), and Knack & Keefer (1995) used data from the International Country Risk Guide (ICRG), to examine the impacts of corruption. In 1995, Transparency International released its first Corruption Perceptions Index, based on data from IMD World Competitive Report, the Political & Economic Risk Consultancy, and Business International (Transparency International, 1995). Cross-country surveys with questions on corruption expanded with the publication of World Development Report 1997: The State in a Changing World (World Bank, 1997), and further with launching of the Business Environment and Enterprise Performance Survey (BEEPS) by the EBRD in 1999. Hellman, Jones, Kaufmann, & Schankerman (2000) introduced the data and early analyses include EBRD (1999), Hellman, Jones, & Kaufmann (2003) and World Bank (2000).) In 1999, researchers at the World Bank introduced the Worldwide Governance Indicators with one dimension focused on control of corruption (Kaufmann, Kraay, & Zoido-Lobaton, 1999). 2 We examine a cross-section of published studies to help answer the question: which data do economists use to study corruption? The cross-section is for publications in the year 2022, notably those with associated Journal of Economic Literature (JEL) codes. The dataset provides a snapshot of the most popular corruption measures used for research by economists, the journals where they are published and JEL codes used, the types of studies, and the geographical foci. While other studies have catalogs of data sources, less is known about their prevalence in use by economists. The cross-section of studies we examine shows that cross-country composite indicators dominate for cross-country studies, while administrative data is more common for studies focusing on individual countries. Articles published in ranked journals were relatively more likely to use administrative data and experimental data than those published in unranked journals. Studies focused on the causes of corruption were fewer but were relatively more likely to be published in ranked journals. While most sources of data are familiar, some novel sources, not available in earlier generations of research, are also being used. Surprisingly, some well- known sources of corruption data were nearly, or entirely, absent from the studies in our sample. Section 2 looks at the literature on the use of data for research on corruption. Section 3 describes the methodology and the variables used in the present paper, and section 4 presents the basic frequencies. Section 5 presents some observations based on the findings, and section 6 suggests some directions for future research using data on corruption. 2. Literature on the use of data for research on corruption The vast literature on corruption and data is dominated by studies examining corruption as a phenomenon, rather than reviewing how data is used to study corruption, although there are a number of exceptions. Lambsdorff (1999) reviews empirical analyses of corruption and the data on which it is based, and the foundational documents for well-known composites naturally 3 examine the strengths and weaknesses of the underlying data sources (Transparency International, 1995; Kaufmann, Kraay, & Zoido-Lobaton, 1999). Studies introducing new approaches to measurement typically highlight the challenges of measurement and shortcomings of then-extant approaches. Olken (2007) provides a good example of both the introduction of a novel approach and an explanation of the importance of data for the study of corruption. Numerous critiques have been written about particular types of data. Arndt & Oman (2006), Knack (2007), Apaza (2009), Anderson (2009), and Heywood (2014) look at shortcomings of composites, as well as expert opinion indicators and surveys. Kaufmann, Kraay, & Mastruzzi (2007) argue that shortcomings of composites notwithstanding, they have value and alternatives have their own shortcomings. Azfar & Murrell (2009) and Kraay & Murrell (2016) focus on the shortcoming of surveys, in particular the problem of untruthful responses, and propose novel adaptations to survey methods to address “reticent respondents.” Subjective measures of governance generally have recently been the subject of some debate. Little & Meng (2024) argue that findings of global democratic backsliding based on subjective indicators such as those produced by V-Dem, Freedom House, and Polity are not borne out by objective indicators of democracy (e.g., incumbent performance in elections). Knutsen et al. (2023), commenting on an earlier version of Little & Meng’s study, among others, argue that measurement error can affect even seemingly objective indicators, and that subjectivity pervades all measurement enterprises. Challenges of measurement stretch beyond academic study and often generate misleading headlines. Wathne & Stephenson (2021) highlight the weak or non- existent foundations of ten global estimates of corruption and related phenomena. While the studies of corruption are many, systematic examinations of the use of data are few. To inform an ambitious program on measuring corruption, David-Barrett, Murray, Ceballos, 4 & Lee (2024) inventory the long list of existing measures, mapping the landscape of 55 different measures. To our knowledge, however, the present paper is the first attempt to systematically quantify which data are used by economists to study corruption; that is, how the landscape of indicators is used in practice by academic economists. 3. Sampling the cross-section of economics research on corruption Selecting the sample of articles To define the sample of articles, we utilized the EBSCO library, focusing specifically on the EBSCOhost platform. This platform encompasses key databases such as Business Source Corporate Plus and EconLit, chosen for their comprehensive economic and business-related literature coverage. We performed a keyword search for “corruption” within the abstracts of publications in English in 2022. The initial comprehensive search within this platform yielded well over 1,000 documents, including journal articles, papers, and case studies. To narrow and more cleanly define the universe, we focused exclusively on journal articles, yielding a longlist of 882 articles. In order to focus on the data used by economists, we further restricted the sample to articles with Journal of Economic Literature (JEL) codes either in the article itself, or as provided in the search results. The sample, therefore, focuses primarily on the economics and related literatures and does not reflect publications in journals from other fields that do not use JEL codes; the implications of this will be examined in the conclusions. The resulting sample was 339 articles. 2 2 The list of articles is included in the online reproducibility package found on the World Bank’s Reproducible Research Repository. 5 Identifying sources of data on corruption Having collected the key information about the articles, we examined each one to identify the source of corruption data, if any, that was used in the paper. In many cases, this was a straightforward exercise, as most articles included a section on data. We focused on data meant to relate to corruption, not necessarily to other factors that are often discussed in the same paper, such as measures of the rule of law or transparency. Where a specific cross-national source was identified, it was recorded by name. In cases where a smaller national or regional survey was used, these were identified as Other Survey. 3 If the paper was based on an experiment conducted as part of the research, it was identified as an Experiment. If the data represent measured quantities (such as numbers of corruption cases), as opposed to approximations from surveys or expert opinion, it was recorded as Administrative Data. In cases where there is no measure of corruption, per se, but some other measurable data is argued to proxy for, among other things, corruption, the type of that data was generally classified as Administrative Data. A number of empirical studies examined the impact of an anticorruption campaign, using an event study or difference-in-differences approach. As the dates chosen are matters of fact and public record, these were categorized as Administrative Data. Annex 1 describes the main sources of cross- national data. 3 These included custom surveys of firms or citizens or some other sample group (e.g., reporters), often designed and implemented by researchers themselves. Other surveys included the Balkans Business Barometer Survey, an SME survey in Viet Nam, the Viet Nam Provincial Competitiveness Index survey of firms, the European Social Survey, and the EBRD-World Bank Business Environment and Enterprise Performance Survey (BEEPS), which is a related component of the World Bank Enterprise Surveys. The latter is tagged as both World Bank Enterprise Survey and Other Surveys, since the BEEPS includes some questions not included or reported in the broader Enterprise Survey. 6 Other variables of interest To explore how different types of data are used for different purposes, we classified articles according to the geographical focus of the paper, the type of paper (empirical or theoretical), and whether the aim of the paper was to examine corruption as a cause or as an effect of some other thing. For the geographical focus of the studies, we categorized articles as follows: We define a study as World if it covers a large number of countries with no clear delimiting factors; Regional if it focuses on a specific region (e.g., a focus on Africa); Affinity if the selection of countries is based on some shared characteristic (e.g., BRICS countries, oil exporters, members of a particular treaty or organization); Comparative if it covers a small number of countries; and Single if it only covers a single country or a geographical area within a country. In a few cases, studies were classified in multiple categories, for example if there was both a regional and an affinity linkage. Articles were categorized as Empirical if they involved the use of data to test a hypothesis or to illustrate a pertinent point. Articles that are Theoretical offered some derivation of relationships or hypotheses, beyond just outlining the models to be tested. Some articles were both theoretical and empirical, and some were neither, as when essays discussed corruption issues but did not provide advances in theory or empirical tests. Studies that traced legal arguments were classified as Theoretical. The role that corruption played in the empirical studies was also noted. Articles were tagged according to whether they examined the impact that corruption has on something else (such as growth or firm performance) or whether the paper studies the factors that lead to corruption. In essence, the former would include studies with corruption as an independent 7 variable, and the latter would have corruption as a dependent variable. Articles for which corruption is not the variable of interest, but rather a confounder to be controlled for, were classified as being about the impact that corruption has on other things. Note that some articles could be in both categories, and some could be in neither. 4. The landscape of corruption data used in economics research Which economics journals published corruption research? The 339 articles in the basic sample described above were published in a large number of publications with varying degrees of popularity in the profession. For much of the analysis that follows, we present data for the sample of 339 articles as well as for the smaller subset of articles that were published in highly ranked journals. Two rankings were used for this purpose. The first ranking system is based on Combes & Linnemer (2010), who provide rankings of the top 600 journals in economics, making adjustments to account for citations that are often omitted from other ranking schema. The Combes & Linnemer journal ranks (CL_Index) have been used in other analyses of economic literature, including Heckman & Moktan (2020). Of the 339 articles in the basic sample, 177 were published in journals included in the CL-Index, and 35 were in journals ranked in the top 100 of the Combes & Linnemer list (CL-Index Top 100) (Table 1). The second ranking system is based on IDEAS/RePEc H-Index for Journals (IDEAS, 2024). This online ranking is based on the ratio of the number of citations to the number of items in the journal or series. The ranking covers more than 3,000 journals or publication series. Of the 339 articles on corruption in our basic sample, 282 were in journals or series included in the IDEAS/Repec H-Index rankings, and 30 were published in journals ranked in the top 100 8 (IDEAS/RePEc H-Index Top 100) (Table 1). (A third journal ranking, Google Scholar’s Top-20 economics journals, will be discussed in the paper’s conclusion.) Table 1. Articles published in highly ranked economics journals IDEAS/RePEc H- Index Top 100 CL-Index top Journals (count of 100 (count of Journal name (listed alphabetically) articles) articles) American Economic Journal: Microeconomics 1 n.a. Economic Development & Cultural Change n.a 1 Economic Journal 1 1 Economica n.a. 2 Economics Letters 3 3 Energy Economics 1 n.a. Environmental & Resource Economics n.a. 2 European Economic Review 1 1 International Economic Review 1 1 International Journal of Central Banking 1 n.a. International Tax & Public Finance n.a. 1 Journal of Banking & Finance 2 2 Journal of Comparative Economics n.a. 3 Journal of Corporate Finance 4 n.a. Journal of Economic Behavior & Organization 3 3 Journal of Economic Surveys 1 n.a. Journal of Economics & Management Strategy n.a. 1 Journal of Environmental Economics & Management 1 1 Journal of Financial Economics 1 1 Journal of Law, Economics, & Organization 1 1 Journal of the European Economic Association 1 1 Public Choice n.a. 2 Review of Economic Studies 2 2 Review of Finance 1 n.a. Review of Financial Studies 1 1 World Development 3 3 World Economy n.a. 2 Total 30 35 Notes: n.a. = not applicable. CL designations are based on Combes & Linnemer (2010). RePEc designations are based on IDEAS, 2023. 9 Which JEL codes were associated with corruption research? To investigate the topics on which the corruption research focused, we start with the Journal of Economic Literature (JEL) Codes (American Economic Association, 2023). With roots going back to 1911, the modern JEL classification system has been in use since 1990, with only incremental changes in the years since (Cherrier, 2017). JEL codes continue to be widely used for classification and to facilitate tracking the popularity of various subjects of economic research. Wagstaff & Culyer (2012) and Meng, Mu, & Tong (2022), for example, use JEL codes to examine research in health economics, and Small (2024) outlines how to use JEL codes in introductory economics courses. The JEL codes themselves are best viewed as approximations of the research topics covered in the papers that use them. Kosnik (2018) systematically examines the consistency of application of JEL codes in top economics journals, finding significant inconsistency between the codes selected by editors (and their staff) and by the authors themselves. Kosnik finds also that despite the inconsistencies, articles typically did fit within the descriptions of the JEL codes in either case. For consistency, we focus on articles for which the published versions provide JEL codes or for which EconLit search results provide JEL codes. In some cases, a published paper did not include JEL codes, although working paper versions did. (Balaeva, Rodionova, Yakovlev, & Tkachenko (2022) is one such example.) As we did not systematically seek out working paper versions, such papers are not generally included in our sample. In addition, some journals do not systematically publish JEL codes. Our universe of publications from 2022 includes six articles published in Economic Modelling, for example, although only four included JEL codes. 10 Finally, the approach of examining both final publications and the EconLit search results to identify JEL codes presented a challenge. As noted by Kosnik (2018), there are often differences in the codes selected by editors and those identified by authors. In the case of the present paper, nearly all of the articles with JEL codes present in both the search results and in the actual publication had different sets of JEL codes. Indeed, the discrepancies seem much larger than those examined by Kosnik (2018). Take, for example, the 13 articles in the sample that were published in journals ranked in the Top-20 economic journals by Google Scholar. For each one, we compared the JEL codes obtained in the search results with those in the actual publications. None of them were the same. Only one was even close, with 3 of 4 codes in the Table 2. Frequencies of JEL Categories for articles in the full sample A. general economics and teaching 2 B. history of economic thought, methodology, and heterodox approaches 2 C. mathematical and quantitative methods 38 D. microeconomics 177 E. macroeconomics and monetary economics 50 F. international economics 70 G. financial economics 71 H. public economics 76 I. health, education, and welfare 30 J. labor and demographic economics 29 K. law and economics 38 L. industrial organization 50 M. business administration and business economics ◦ marketing ◦ accounting ◦ personnel economics 34 N. economic history 6 O. economic development, innovation, technological change, and growth 147 P. political economy and comparative economic systems 51 Q. agricultural and natural resource economics ◦ environmental and ecological economics 31 R. urban, rural, regional, real estate, and transportation economics 13 Y. miscellaneous categories 2 Z. other special topics 13 Note: The table depicts the broad categories of JEL codes associated with the paper. A single paper typically has multiple codes meaning that the totals in this table are larger than the number of articles in the sample. Definitions of JEL codes are provided in American Economic Association, 2023. 11 published paper matching 3 of 6 in the EconLit search results; for several articles the two lists did not even overlap. As our dataset was assembled sequentially, the codes available in the publications themselves were generally recorded if available, and those provided with the search results were used otherwise. Kosnik’s (2018) finding that codes broadly reflect the topics of research, despite discrepancies, provides some comfort regarding discrepancies in articles focused on corruption. In the single year of research that we study, articles on corruption spanned all JEL broad categories, with the largest numbers in D (microeconomics), which includes the only JEL code that mentions corruption by name, and O (economic development, innovation, technological change, and growth) (Table 2). The high representation in the latter group largely reflects studies of the impact of corruption on growth and development, and the high representation in G (financial economics) stems from research, in part, on the impacts of corruption on financial institutions. Surprisingly, only six articles were classified as N (economic history). At a more granular level, the most common JEL code was, not surprisingly, D73 (bureaucracy ◦ administrative processes in public organizations ◦ corruption), and this was followed by O17 (formal and informal sectors ◦ shadow economy ◦ institutional arrangements), and D72 (political processes: rent-seeking, lobbying, elections, legislatures, and voting behavior) (Table 3). Which geographies were the subject of corruption research? Most of the articles in the sample were either global in nature or focused on a single country. Smaller numbers of articles focused on countries that share some characteristic (e.g., BRICS countries) or were comparative studies of a small number of countries. (Some articles fit 12 Table 3. Frequencies of the most common JEL codes for the full sample D73. bureaucracy ◦ administrative processes in public organizations ◦ corruption 140 O17. formal and informal sectors ◦ shadow economy ◦ institutional arrangements 58 D72. political processes: rent-seeking, lobbying, elections, legislatures, and voting behavior 41 K42. illegal behavior and the enforcement of law 27 G21. banks ◦ depository institutions ◦ micro finance institutions ◦ mortgages 21 O16. financial markets ◦ saving and capital investment ◦ corporate finance and governance 21 P26. property rights 21 O11. macroeconomic analyses of economic development 20 O15. human resources ◦ human development ◦ income distribution ◦ migration 19 F23. multinational firms ◦ international business 18 G32. financing policy ◦ financial risk and risk management ◦ capital and ownership structure ◦ 16 value of firms ◦ goodwill O19. international linkages to development ◦ role of international organizations 16 O43. institutions and growth 16 M14. corporate culture ◦ diversity ◦ social responsibility 15 O14. industrialization ◦ manufacturing and service industries ◦ choice of technology 15 D22. firm behavior: empirical analysis 14 H11. structure, scope, and performance of government 14 L25. firm performance: size, diversification, and scope 14 O13. agriculture ◦ natural resources ◦ energy ◦ environment ◦ other primary products 14 G34. mergers ◦ acquisitions ◦ restructuring ◦ corporate governance 13 E02. institutions and the macroeconomy 11 F14. empirical studies of trade 11 C23. panel data models ◦ spatio-temporal models 10 F21. international investment ◦ long-term capital movements 10 G38. government policy and regulation 10 Note: The table depicts only the JEL codes used by at least 10 articles. Definition of JEL codes provided in American Economic Association, 2023. in multiple categories: examples include a comparative study of investment by India and China into Africa, and one focused on Francophone Africa.) Of the studies that examined a single country, studies of China were the most common, followed by the United States, India, Viet Nam, and Nigeria. Studies of China and the United States accounted for more than half of the articles in top ranked journals. 13 Table 4. Numbers of articles by geographic focus Empirical Empirical articles in articles in top 100 Empirical ranked ranked Geography articles journals* journals* World 88 83 12 Regional 58 48 1 Affinity 25 20 1 Single 111 94 19 Comparative 11 8 1 Total* 269 234 34 All Empirical Empirical single- Empirical articles in All single- Empirical articles in geography articles in top 100 geography articles in top 100 empirical ranked ranked Geography empirical ranked ranked Geography articles journals* journals* (continued) articles journals* journals* Algeria 1 1 0 Korea, Rep. 1 1 0 Australia 1 1 0 Nigeria 6 6 0 Bangladesh 1 1 0 Oman 1 1 0 Brazil 2 2 2 Pakistan 1 1 0 Bulgaria 1 1 0 Papua New Guinea 1 1 0 Burundi 1 1 0 Peru 1 0 0 China 34 28 8 Portugal 2 2 0 Croatia 1 0 0 Romania 1 1 0 Czechia 1 1 0 Russian Federation 3 1 0 Côte d'Ivoire 1 1 0 Serbia 2 1 0 Ethiopia 1 1 0 South Africa 4 4 0 France 1 1 0 Spain 1 1 0 Ghana 1 1 1 Türkiye 1 1 0 India 8 6 1 Ukraine 3 2 1 Indonesia 5 3 0 United States 8 8 4 Iraq 2 2 0 Venezuela, RB 1 1 0 Italy 4 4 1 Viet Nam 6 5 0 Kenya 1 1 1 Zimbabwe 1 1 0 Total 111 94 19 Notes: Ranking refers to either CL-Index or RePEc rankings. For the top table, total may be less than the sum of geographical foci as a single paper may be classified in multiple categories. Studies focused on a geography within a country are included in the country counts. 14 Which indicators were used to study corruption? Out of 270 studies classified as Empirical and which used data, 4 most use either the Worldwide Governance Indicators Control of Corruption (WGI-CC) indicator or Transparency International’s Corruption Perceptions Index (TI-CPI) (Table 5). Both are composites of other indicators and offer researchers easily accessible and widely known measures covering a large number of countries and many years. This reliance on WGI-CC and TI-CPI is even more pronounced for studies that are cross country in nature, for which the need for a large number of Table 5. Data sources for studies on corruption, by geographical category World, Single regional, geography, All affinity comparative Worldwide Governance Indicators Control of Corruption (WGI-CC) 84 77 15 Transparency International Corruption Perceptions Index (TI-CPI) 72 59 15 International Country Risk Guide (ICRG) 18 16 2 World Bank Enterprise Surveys 12 8 4 Varieties of Democracy (VDEM) 6 6 0 Bayesian Corruption Index (BCI) 5 5 0 World Economic Forum Executive Opinion Survey (WEF) 4 4 0 Life in Transition Survey (LITS) 2 2 0 World Values Survey (WVS) 2 1 1 European Quality of Government Index (EQGI) 2 2 0 Basel AML Index (BASEL) 2 1 1 World Business Environment Survey 1 1 0 Afrobarometer 1 1 0 Foreign Corrupt Practices Act Clearinghouse (FCPA) 1 1 0 BBVA Research (BBVA) 1 0 1 Thomson-Reuters ASSET4/ESG (Thomson-Reuters) 1 1 0 Other Survey 34 5 30 Experiment 9 0 9 Administrative Data 57 6 51 Number 269 155 122 Note: Articles may be classified in multiple categories. For this reason, the sum of data sources is larger than the total number of articles listed at the bottom of the table. 4 Some articles that are classified as empirical use observational or qualitative techniques and do not rely on corruption data per se. 15 observations places a premium on such measures (column 2). The International Country Risk Guide (ICRG) is also used in a number of studies that are cross-country and time series, in part owing to its long time series. For single-country studies or those involving a small number of countries, administrative data, other surveys (often country-specific surveys), and experimental data are much more prominent (column 3). The overall patterns are not significantly changed if the sample is restricted to the subset of articles in ranked or top ranked journals. (Table 6.) The subset of articles in top ranked journals were somewhat less likely to use the popular composite measures WGI-CC and TI-CPI, and somewhat more likely to use measures offering long time series, such as ICRG and VDEM. Table 6. Data sources for studies on corruption, for ranked journals CL-Index CL-Index RePEc RePEc Top ranked Top 100 ranked 100 WGI 43 5 70 4 CPI 39 6 62 4 ICRG 14 3 18 2 World Bank Enterprise Survey 6 0 10 0 VDEM 6 3 6 1 BCI 4 1 5 1 WEF 2 0 4 0 LITS 2 0 2 0 WVS 0 0 2 0 EQGI 1 0 1 0 BASEL 0 0 2 0 World Business Environment Survey 1 0 1 0 AFROBAROM 1 0 1 0 FCPA 1 0 0 0 BBVA 0 0 0 0 Thomson-Reuters 1 0 1 1 Other Survey 14 2 27 3 Experiment 8 5 9 3 Administrative Data 41 12 47 14 Number 154 27 227 24 Notes: Articles may be classified in multiple categories. For this reason, the sum of data sources is larger than the total number of articles listed at the bottom of the table. CL designations are based on Combes & Linnemer (2010). RePEc designations are based on IDEAS, 2023. 16 The articles in top ranked journals were also more likely to make use of administrative or experimental data. For what purpose was the corruption data used? For the full sample of articles, 69 percent were tagged as examining the impact that corruption has on something else, such as growth or firm performance, while 24 percent examined the causes of corruption. About 6 percent did both, and about 14 percent did neither. Table 7. Data sources for studies on corruption, by direction of causality Studies examining Studies examining Either a cause corruption as a cause the causes of or an effect for something else corruption WGI 84 78 12 CPI 72 63 12 ICRG 18 14 5 World Bank Enterprise Survey 12 11 3 VDEM 6 6 1 BCI 5 4 1 WEF 4 3 1 LITS 2 0 2 WVS 2 2 1 EQGI 2 2 0 BASEL 2 2 1 World Business Environment Survey 1 0 1 AFROBAROM 1 0 1 FCPA 1 1 0 BBVA 1 1 0 Thomson-Reuters 1 1 0 Other Survey 34 28 13 Experiment 9 7 4 Administrative Data 56 41 17 Number 293 223 80 Note: Some studies examined both corruption as a cause and corruption as an effect. In addition, articles may be drawn upon multiple categories data source or none at all. For this reason, the sum of data sources is different than the total number of articles listed at the bottom of the table. Studies examining the impact of corruption on other variables were more likely to use composites such as WGI-CC and TI-CPI, while those looking at the causes of corruption were 17 relatively more likely to use administrative data, country-specific surveys, or experiments (Table 7). 5 5. Observations about the use of data on corruption The cross-section of published research from 2022 prompts several observations. First, composite indicators such as the WGI-CC and TI-CPI remain popular among researchers, especially those examining cross-country or time series data. For many studies, corruption is only one of the variables of interest, and having a large number of observations, even noisy ones, is attractive for researchers. The composites themselves have been criticized for the conceptual imprecision they bring since they combine different things in different quantities across countries and over time (Arndt & Oman, 2006; Knack, 2007; Anderson, 2009; Apaza, 2009; Heywood, 2014; Stephenson, 2023). For a response to some of the criticisms, see Kaufmann, Kraay, & Mastruzzi (2007). In the present sample, some studies using the composites cite other studies that also use them as reasons that they are acceptable measures. Their popularity is somewhat self- generating. Second, while some articles did make use of surveys such as the World Bank’s Enterprise Surveys, the number of articles that used survey data was relatively small. Particularly surprising was the fact that none of the articles in this 2022 cross-section used large-scale surveys such as Transparency International’s Global Corruption Barometer (TI-GCB), and only one used Afrobarometer. Even the World Values Survey was only used in two studies in the sample. These are untapped resources in terms of economics research on corruption, although A logical extension of the present paper, set aside for the future, would dig further into the question of 5 which causes of corruption were being explored. 18 they contribute to rich analysis in publications outside economics (e.g., Peiffer, Armytage, & Marquette, 2018; Peiffer & Rose, 2014). A possible reason that the World Bank Enterprise Survey is used more than some other surveys is that respondent-level data is easily available for researchers, while respondent-level data is not readily accessible for the TI-GCB. Firm-level data makes possible the study of firm- level dynamics that would be unavailable with country-level measures. In the sample of 2022 articles, the World Bank Enterprise Survey firm-level data was used to examine the impact of corruption at the firm level on innovation and patents (Poddar & Singh, 2022), and growth (Erhardt, 2022), among others. Firm-level data also made possible studies of the determinants of perceptions and experiences of corruption at the firm level, examining whether customs and trade barriers contribute (Kumanayake, 2022), and whether characteristics such as firm size, city size, and government size contribute to corruption (Goel, Mazhar, & Ram, 2022), among others. Third, some novel sources of data do not fit neatly into categories of surveys or expert opinions. In their study of the determinants of consumer confidence, Tjandrasa & Dewi (2022) used data from BBVA Research “based on Google Trends Big Data on searches about corruption”, a real-time, higher-frequency, measure of perceptions. As the other variables in the study are updated each month, the availability of a higher frequency corruption proxy facilitated their research. In their study of the influence of ethical failures on the cost of equity, Banerjee, Gupta, & Krishnamurti (2022) used the Thomson Reuters’ ASSET4 database for a variable on corruption based on whether firms are “under the spotlight of the media because of a controversy linked to bribery and corruption, political contributions, improper lobbying, money laundering, parallel imports or any tax fraud.” For this research, the availability of a source assessing public 19 perceptions of ethics at listed firms was essential, and familiar country-level indicators could not have served their purpose. Fourth, the popularity of administrative data, especially for single-country studies, is notable. As mentioned earlier, this category includes a somewhat heterogeneous set of data whose common characteristic is that the data are officially reported, measured or observed. The number of corruption cases or convictions was used for analysis in several studies, and media reports were also used by some. Nathan, et al. (2022) examined the effect of fiscal decentralization, government internal audit, law enforcement, and natural resources on the level of corruption in Indonesia. Al-Hadi, Taylor, & Richardson (2022) used the number of corporate corruption convictions and cases in the United States to look at the relationship between corruption and tax avoidance. Du & Heo (2022) employed data on convictions for corruption- related crimes involving abuses of the public trust by government officials, such as bribery and extortion, to examine how political corruption affects corporate investment in the United States. Alexeev & Zakharov (2022) used data on bribe-taking incidents registered by police in the Russian Federation to examine links between oil windfalls and income inequality, with corruption proxying for economic rents. James & Rivera (2022) combined data on convictions for corruption offenses and the frequency of related words in newspapers to examine whether oil-rich U.S. states experience more corruption than their oil-poor counterparts. De Andrés, Polizzi, Scannella, & Suárez (2022) used hand-compiled media reports of corruption scandals to examine why financial institutions in certain European countries disclose corruption-related information in their annual financial reports. Audit reports were employed for some studies. Stoecker (2022) used hand-collected data from official audit reports in Ghana to examine the link between partisan alignment of local 20 politicians and the incidence of political corruption. Colonnelli & Prem (2022) examined data from randomized anticorruption audits to study how anticorruption affects dependency on government relationships and public procurement in municipalities in Brazil. We included in the category of administrative data the large number of studies examining the impact of China’s anticorruption campaign. Many used time series data on firms in China and introduced dummy variables to examine the impact of the campaign on some other variable. These include a study of stock price co-movements before and after the anticorruption campaign using links between firms and politicians (Piotroski, Wong, & Zhang, 2022), a study examining the campaign’s effect on poverty (Han, Li, & Xu, 2022), a study of the effect of the campaign on the accuracy of analyst earnings forecasts (Hou, Li, Teng, & Hu, 2022), a study of the impact of the anticorruption campaign on households’ decisions to invest in the stock market (Bu, Hanspal, & Liao, 2022), and how reporting on corruption changed after the campaign (Zhuang, 2022), and many others. A feature of the many studies on the impact of China’s anticorruption campaigns is the use of data on firm-level performance and corporate governance, facilitated by public information on listed companies. (See Tong, 2022, for a broader review of research on corruption in China.) A constraining factor for the use of administrative data is that analyses are limited to countries where data is made available (Anderson, Bernstein, Kim, Recanatini, & Schuster, 2023). The assortment of studies using administrative data shows the breadth of corruption- related issues, and richness of findings, such data can facilitate. Countries could facilitate better research on corruption by making administrative data more readily available. Fifth, while most of the empirical studies used econometric techniques to examine cross- section or time series data, several employed experiments. For their study on gender differences 21 in the tendency to act corruptly and in their tolerance of corruption, Guerra & Zhuravleva (2022) recruited students for an experiment. “At the beginning of the experiment, players are matched randomly and anonymously to eight groups of four players. In each group, subjects are assigned randomly to one of the following roles: a citizen, a public official, another member of society, and a monitor. The roles remain fixed throughout the experiment. The design comprises a four- person, sequential-move game….” For their study on tax evasion and corrupt behavior, Banerjee, Boly, & Gillanders (2022) also conducted a role-playing experiment: “Subjects in a session are randomly divided into groups of three in each round. One of them is randomly assigned the role of a Public Official (PO), and the two others are assigned the role of citizens. The roles remain the same throughout the experiment. Citizens perform a real effort task and receive an income based on their performance….” For their study of proclivity to cheat, Denisova-Schmidt, Huber, & Prytula (2022) used an online coin-tossing experiment. “Over 1,500 participants were asked to make ten coin tosses and were randomly assigned to one of the three treatment groups tossing coins (1) online, (2) manually, or (3) having the choice between tossing manually or online. The study outcomes suggest that students are more inclined to cheat when they perceive the coin toss to be more ‘private.’ Moreover, the students’ attitudes toward corruption appear to matter for the extent of their cheating, while socio-demographic characteristics were less important….” Researchers may also wish to note that top journals favor papers using experimental and administrative data on corruption relatively more than do lower ranked journals. Experimental data were used in nearly 14 percent of the articles in CL-Index Top 100 journal, compared to only 1 percent of articles that are not ranked in that group. Similarly, administrative data were used in 34 percent of the articles in the CL-Index Top 100 journals, compared to only 15 percent of articles that are not ranked in that group. 22 Sixth, the study raises a number of questions about how much the availability of data is driving the topic of research, rather than the other way around. The topics in this paper— geography, type of data, and source of data—are not independent. For example, the easy availability of big composites, such as WGI-CC and TI-CPI, facilitates cross-country research, while administrative data, such as corruption convictions, is more readily available in more developed countries. Single-country empirical studies are focused on countries with availability of particular types of data: dates, enforcement or the anticorruption program, and administrative data in the case of China, convictions in the case of the United States, and availability of surveys with long time series in the case of Viet Nam. In Viet Nam’s case, annual surveys of citizens (CECODES, VFF-CRT, RTA & UNDP, 2023) and firms (Malesky, Pham, & Phan, 2024) were repeatedly drawn upon. The public good value of such surveys, and of making data available, should not be underestimated. Finally, economists undertaking research on corruption may wish to reflect on the missed opportunity when they rely exclusively on the economic literature. Studies using administrative data, often found in public administration journals, may have important policy content; for example, identifying how corruption impacts state management capacity in Lee & Liu (2022). Similarly, political topics that are typically studied in political science journals may provide insight for issues of interest to economists; for example, linking municipal-level corruption with recessions in Sanz, Solé-Ollé, & Sorribas-Navarro (2022). 6. Future directions This study focused on a particular cross-section of corruption-related research. By focusing on published articles with JEL codes, the study helps shed light, in a systematic way, on which data is used by economists to study corruption. It is worth reflecting on what is not 23 included in this cross-section. Analyses that are published in report format, or are in working paper format, are not included. Many studies published by international organizations, think tanks, civil society organizations, and others are not in the sample of academic articles. Even within the confines of academic articles, the circumscribed approach to sampling only articles with the word “corruption” in the abstract omits articles that reference related variants, such as “integrity,” “ethics,” “money laundering,” and others, unless they also mention “corruption.” Most importantly, studies in other disciplines, especially political science and public administration, are generally not included in our sample. While the approach used in this paper helps us to focus on the research carried out by academic economists, it does not provide a comprehensive picture of research using data on corruption. To put the economics literature into perspective, we reexamined the original universe of 882 English-language articles emerging from the EBSCOhost and EconLit searches. As a rudimentary approximation of the extent of data-oriented research on corruption in different disciplines, we used the “Top 20” journals according to Google Scholar for economics, political science, and public policy and administration, respectively. The number of articles on corruption published in these top 20 political science and public policy and administration journals is more than three times the number published in the top 20 economics journals; the number in these other disciplines using data on corruption is more than four times the number in the top economics journals (Table 8). 6 Even with these limitations, in a single year there were 339 published journal articles, of which 31 were published in journals ranked in the top 100 by IDEAS/RePEc H-Index, and 36 in 6 Arguably the present research design focusing on JEL codes, and therefore the economics literature, should be expanded, but we leave that for a future paper. 24 Table 8. Use of data on corruption in economics, political science, and public administration Articles in Of which, Of which, use Google Scholar’s Top 20 Journal (of journals with articles the 2022 with JEL corruption with corruption in the abstract) sample codes data Economics Journals 15 13 12 − The Economic Journal 1 1 0 − Economic Modelling 6 4 6 − Economics Letters 3 3 2 − European Economic Review 1 1 0 − Journal of the European Economic Association 1 1 1 − The Review of Economic Studies 2 2 2 − The Review of Financial Studies 1 1 1 Political Science Journals 23 0 17 − American Political Science Review 1 0 1 − Annual Review of Political Science 1 0 0 − British Journal of Political Science 2 0 2 − Comparative Political Studies 3 0 2 − Governance* 15 0 11 − Journal of European Public Policy 1 0 1 Public Policy & Administration Journals 46 5 39 − Administration & Society 1 0 1 − American Review of Public Administration 1 0 1 − Governance* 15 0 11 − International Journal of Public Administration 13 0 13 − Journal of Public Administration Research and Theory 2 0 2 − Public Administration 2 0 1 − Public Administration Review 1 1 1 − Public Management Review 2 0 2 − Public Performance & Management Review 3 0 3 − Regulation & Governance 4 4 2 − Review of Public Personnel Administration 2 0 2 *Note that Governance appears in two lists. Top 20 lists as of May 2024: Economics (subcategory of Business, Economics & Management); Political Science (subcategory of Social Sciences); Public Policy & Administration (subcategory of Social Sciences). 25 the top 100 according to the CL-Index. The sheer volume of economics research on the impacts and drivers of corruption is enormous. 7 The general patterns on the use of data were as one might expect, with a study’s purpose helping to dictate the choice of data used. At the same time, the reliance on existing indicators, especially the large composites, suggests a certain degree of opportunism in going after data that is easy, rather than data that is appropriate or novel. There is space to pay greater attention to newer data sources, including those made possible by advances in technology. 8 The rollout of e- procurement systems, e-tax systems, digital land registers, digital beneficial ownership registers, and others afford expanded opportunities for empirical corruption-related research (Alcaide, et al., 2023; Fazekas, Tóth, & King, 2016; Dávid-Barrett, Fazekas, Hellmann, Márk, & McCorley, 2020; World Bank, 2024). There are currently international efforts to move the discussion on corruption measurement forward. UNODC has organized a Task Force on Corruption Measurement, including experts from national and international organizations active in the field of corruption measurement both in developed and developing countries, to support efforts to monitor the Sustainable Development Goals, and organized an international conference on the topic (UNODC, 2023). The International Anti-Corruption Academy has likewise established a Global Programme on Measuring Corruption (David-Barrett, Murray, Ceballos, & Lee, 2024; International Anti-Corruption Academy, 2024). Efforts to improve the benchmarking of 7 Another benchmark illustrates the scale of research on corruption. Matthew Stephenson’s Bibliography on Corruption and Anticorruption, which was last updated in October 2022, stretches to 774 pages. Even with 2022 only about three-quarters complete, there were 344 entries. 8 A wide range of approaches were in evidence at a symposium exploring Data Analytics for Anticorruption, supported by the KDI School and the World Bank (World Bank, 2021) Eight papers from that symposium were published in a special collection in the journal Data & Policy (Data & Policy, 2023). 26 anticorruption policies have also moved forward with the publication of OECD’s Public Integrity Indicators (OECD, 2023), and efforts to benchmark de facto transparency have also advanced with the development of the T-Index (Mungiu-Pippidi, 2022). For its part, the World Bank and KDI, together with other partners, sponsored a Symposium on Data Analytics and Anticorruption (World Bank, 2021), and the World Bank has created a new Public Institutions unit to strengthen approaches to integrating institutional data into the World Bank’s work. Many studies point to the (mostly) harmful effects of corruption. The relatively smaller number of studies aimed at explaining what causes corruption suggests that this area of research is ripe for advancement. In addition to being more actionable and policy-relevant, such research also tends to fare better in top journals. Whereas only 22 percent of articles that were published in journals outside of the CL-Index looked at the causes of corruption, 40 percent of those that were published in CL-Index Top 100 journals did so. Economics researchers might be better advised to focus more attention on what drives corruption than on its effects. And innovating and diversifying data sources to rely less on composite country-level indicators would help. 27 References Alcaide, M. D., Anderson, J., Kramer, M., LaCascia, H., Mells, T., & Valentine, J. (2023). Reaching the Potential for the Digital Economy in Africa: Digital Tools for Better Governance. Washington: World Bank. doi:https://doi.org/10.1596/40271 Alexeev, M., & Zakharov, N. (2022). Who profits from windfalls in oil tax revenue? Inequality, protests, and the role of corruption. Journal of Economic Behavior & Organization, 197, 472-492. Al-Hadi, A., Taylor, G., & Richardson, G. (2022). Are corruption and corporate tax avoidance in the United States related? Review of Accounting Studies. American Economic Association. (2023). JEL Classification System / EconLit Subject Descriptors. Retrieved December 1, 2023, from American Economic Association website: https://www.aeaweb.org/econlit/jelCodes.php Anderson, J. H. (2009). A Review of Governance and Anticorruption Indicators in East Asia and Pacific. World Bank. Anderson, J., Bernstein, D. S., Kim, G., Recanatini, F., & Schuster, C. (2023). Understanding Corruption through Government Analytics. In D. Rogger, & C. Schuster, The Government Analytics Handbook: Leveraging Data to Strengthen Public Administration (pp. 131-148). Washington, DC: World Bank. Apaza, C. R. (2009, January). Measuring Governance and Corruption through the Worldwide Governance Indicators: Critiques, Responses, and Ongoing Scholarly Discussion. PS: Political Science & Politics, 42(1), 139-143. Arndt, C., & Oman, C. P. (2006). Uses and Abuses of Governance Indicators. Paris: OECD. Azfar, O., & Murrell, P. (2009, January). Identifying Reticent Respondents: Assessing the Quality of Survey Data on Corruption and Values. Economic Development and Cultural Change, 57(2), 387-411. Balaeva, O., Rodionova, Y., Yakovlev, A., & Tkachenko, A. (2022). Public Procurement Efficiency as Perceived by Market Participants: The Case of Russia. International Journal of Public Administration, 45(16), 1156-1167. Banerjee, R., Boly, A., & Gillanders, R. (2022). Anti-tax evasion, anti-corruption and public good provision: An experimental analysis of policy spillovers. Journal of Economic Behavior & Organization, 197, 179-194. Banerjee, R., Gupta, K., & Krishnamurti, C. (2022). Does corrupt practice increase the implied cost of equity? Journal of Corporate Finance, 73. Basel Institute on Governance. (2023). Basel AML Index 2023: 12th Public Edition. Ranking money laundering and terrorist financing risks around the world. Basel: Basel Institute on Governance. Bayesian Corruption Index. (2023). The Bayesian Corruption Index 2023 update. Retrieved from The Bayesian Corruption Index 2023 update 28 BBVA Research. (2018). Assessing Corruption with Big Data. BBVA Research. Retrieved from https://www.bbvaresearch.com/wp-content/uploads/2018/03/Eco-Watch- Corruption_vf.pdf Bu, D., Hanspal, T., & Liao, Y. (2022). Political corruption, trust, and household stock market participation. Journal of Banking & Finance, 138. Burki, T. (2019). Corruption is an 'ignored pandemic'. The Lancet, 19, 471. CECODES, VFF-CRT, RTA & UNDP. (2023). The 2022 Viet Nam Provincial Governance and Public Administration Performance Index (PAPI 2022): Measuring Citizens’ Experiences. Hanoi: A Joint Policy Research Paper by Centre for Community Support and Development Studies (CECODES), Centre for Research and Training of the Viet Nam Fatherland Front (VFF-CRT), Real-Time Analytics (RTA), and United Nations Development Programme (UNDP). Cherrier, B. (2017). Classifying Economics: A History of the JEL Codes. Journal of Economic Literature, 55(2), 545-579. Colonnelli, E., & Prem, M. (2022). Corruption and Firms. Review of Economic Studies, 89(2), 695-732. Combes, P.-P., & Linnemer, L. (2010). Inferring Missing Citations: A Quantitative Multi- Criteria Ranking of All Journals in Economics. HAL Open Science. Retrieved from https://shs.hal.science/halshs-00520325 Data & Policy. (2023). Special Collection: Data Analytics for Anti-Corruption in Public Administration. Retrieved from Data & Policy website: https://www.cambridge.org/core/journals/data-and-policy/special-collections/data- analytics-for-anti-corruption-in-public-administration Dávid-Barrett, E., Fazekas, M., Hellmann, O., Márk, L., & McCorley, C. (2020). Controlling Corruption in Development Aid: New Evidence from Contract-Level Data. Studies in Comparative International Development, 55, 481-515. David-Barrett, E., Murray, A., Ceballos, J. C., & Lee, A.-Y. (2024, June). Global Programme on Measuring Corruption: Phase I Synthesis Brief. IACA Insights Brief(09). de Andrés, P., Polizzi, S., Scannella, E., & Suárez, N. (2022). Corruption-related disclosure in the banking industry: evidence from GIPSI countries. European Journal of Finance. Denisova-Schmidt, E., Huber, M., & Prytula, Y. (2022). Perceived Anonymity and Cheating in an Online Experiment. Eastern European Economics, 540-558. Du, Q., & Heo, Y. (2022). Political corruption, Dodd–Frank whistleblowing, and corporate investment. Journal of Corporate Finance, 73. EBRD. (1999). Transition report 1999: Ten years of transition. London: EBRD. Erhardt, E. C. (2022). Prevalence and Persistence of High-Growth Entrepreneurship: Which Institutions Matter Most? Journal of Industry, Competition and Trade, 22, 297-332. 29 Fazekas, M., Tóth, I. J., & King, L. P. (2016, April 25). An Objective Corruption Risk Index Using Public Procurement Data. European Journal on Criminal Policy and Research, 22, 369-397. Foreign Corrupt Practices Act Clearinghouse. (2023). The Foreign Corrupt Practices Act Clearinghouse: About Us. Retrieved from Stanford Law School and Sullivan & Cromwell LLP: https://fcpa.stanford.edu/about-the-fcpac.html Goel, R. K., Mazhar, U., & Ram, R. (2022). Dimensions of size and corruption perceptions versus corruption experiences by firms in emerging economies. Journal of Economics & Finance, 46, 374-396. Guerra, A., & Zhuravleva, T. (2022). Do women always behave as corruption cleaners? Public Choice, 191, 173-192. Haerpfer, C., Inglehart, R., Moreno, A., Welzel, C., Kizilova, K., Diez-Medrano, J., . . . (eds). (2022). World Values Survey: Round Seven - Country-Pooled Datafile Version 5.0. Madrid, Spain & Vienna, Austria: JD Systems Institute & WVSA Secretariat. doi:doi:10.14281/18241.20 Han, L., Li, X., & Xu, G. (2022, November). Anti-corruption and poverty alleviation: Evidence from China. Journal of Economic Behavior & Organization, 203, 150-172. Heckman, J. J., & Moktan, S. (2020). Publishing and Promotion in Economics: The Tyranny of the Top Five. Journal of Economic Literature, 58(2), 419-470. Hellman, J. S., Jones, G., & Kaufmann, D. (2003). Seize the State, Seize the Day An empirical analysis of State Capture and Corruption in Transition. Journal of Comparative Economics, 31(4), 751-773. Hellman, J. S., Jones, G., Kaufmann, D., & Schankerman, M. (2000). Measuring governance and state capture: the role of bureaucrats and firms in shaping the business environment Results of a firm-level study across 20 transition economies. London: EBRD. Heywood, P. M. (2014). Measuring Corruption: Perspectives, Critiques, and Limits. In P. M. Heywood, Routledge Handbook of Political Corruption. London: Routledge. Hou, Q., Li, W., Teng, M., & Hu, M. (2022, October). Just a short-lived glory?The effect of China's anti-corruption on the accuracy of analyst earnings forecasts. Journal of Corporate Finance, 76, 1-25. IDEAS. (2024). IDEAS/RePEc H-Index for Journals. Retrieved December 18, 2023, from IDEAS: https://ideas.repec.org/top/top.journals.hindex.html International Anti-Corruption Academy. (2024, June 1). Global Programme on Measuring Corruption. Retrieved from International Anti-Corruption Academy: https://www.iaca.int/measuring-corruption/ James, A., & Rivera, N. M. (2022). Oil, politics, and "Corrupt Bastards". Journal of Environmental Economics & Management, 111, 1-26. 30 Kaufmann, D., Kraay, A., & Mastruzzi, M. (2007). The Worldwide Governance Indicators Project: Answering the Critics. World Bank Policy Research Working Paper(WPS4149). Kaufmann, D., Kraay, A., & Mastruzzi, M. (2010). The Worldwide Governance Indicators: Methodology and Analytical Issues. Rochester: SSRN Scholarly Papers. Kaufmann, D., Kraay, A., & Zoido-Lobaton, P. (1999). Aggregating Governance Indicators. World Bank Policy Research Working Paper(2195). Kim, J.-B., Lee, E., Tang, X., & Zhang, J. (2022). Collusive versus coercive corporate corruption: evidence from demand-side shocks and supply-side disclosures. Review of Accounting Studies, 28, 1929–1970. Knack, S. (2007, Sep-Dec). Measuring Corruption: A Critique of Indicators in Eastern Europe and Central Asia. Journal of Public Policy, 27(3), 255-291. Knack, S., & Keefer, P. (1995). Institutions And Economic Performance: Cross-Country Tests Using Alternative Institutional Measures. Economics and Politics, 7(3), 207-227. Knutsen, C. H., Marquardt, K. L., Seim, B., Coppedge, M., Edgell, A., Medzihorsky, J., . . . Lindberg, S. I. (2023, May). Conceptual and Measurement Issues in Assessing Democratic Backsliding. V-Dem Institute Working Paper, 2023(140). Kosnik, L.-R. (2018). A Survey of JEL Codes: What Do They Mean and Are They Used Consistently? Journal of Economic Surveys, 32(1), 249-272. Kraay, A., & Murrell, P. (2016). Misunderestimating Corruption. The Review of Economics and Statistics, 98(3), 455-466. Kumanayake, N. S. (2022). Do customs and other trade regulatory barriers lead firms to bribe? Evidence from Asia. Journal of International Trade & Economic Development, 31(3), 340-357. Lambsdorff, J. G. (1999). Corruption in empirical research: A review. Transparency International, processed. Lee, J., & Liu, C. (2022). Public Corruption and Government Management Capacity. Public Performance & Management Review, 45(2), 327-427. Little, A. T., & Meng, A. (2024). Measuring Democratic Backsliding. PS: Political Science & Politics, 57(2), 149-161. Malesky, E., Pham, T. N., & Phan, N. T. (2024). The Vietnam Provincial Competitiveness Index and Provincial Green Index Report: Promoting a Business-Enabling and Environmentally Friendly Investment Climate, 2023 Final Report. Hanoi, Vietnam: Vietnam Chamber of Commerce and Industry and United States Agency for International Development. Marquardt, K. (2023). V-Dem Methodology. V-Dem. Varieties of Democracy. Retrieved from https://v-dem.net/about/v-dem-project/methodology/ Marquette, H., & Peiffer, C. (2022). Chapter 13: Corruption and anticorruption: Uganda and South Africa as positive outliers in governance reforms? In W. Hout, J. Hutchison, & eds, 31 Handbook on Governance and Development. Cheltenham, UK: Edward Elgar Publishing. doi:https://doi.org/10.4337/9781789908756.00022 Mauro, P. (1995). Corruption and Growth. Quarterly Journal of Economics, 110(3), 681-712. Meng, X., Mu, G., & Tong, J. (2022). Health economics in Africa from 1991 to 2020: A systematic review. Journal of Public Health in Africa, 13(2027). Mungiu-Pippidi, A. (2022). Transparency and corruption: Measuring real transparency by a new index. Regulation & Governance, 17, 1094–1113. doi:https://doi.org/10.1111/rego.12502 Nathan, L., Aswar, K., Jumansyah, Mulyani, S., Hardi, & Nasir, A. (2022). The Moderating Role of Natural Resources Between Fiscal Decentralization, Government Internal Audit, Law Enforcement and Corruption: Evidence from Indonesian Local Government. Contemporary Economics. OECD. (2023). OECD website. Retrieved from OECD Public Integrity Indicators: https://oecd- public-integrity-indicators.org/ Olken, B. A. (2007, April). Monitoring Corruption: Evidence from a Field Experiment in Indonesia. Journal of Political Economy, 115(2), 200-249. Peiffer, C., & Rose, R. (2014, October 7). Why Do Some Africans Pay Bribes While Other Africans Don't? Afrobarometer , Working Paper No. 148. Afrobarometer . Peiffer, C., Armytage, R., & Marquette, H. (2018, May). Uganda’s Health Sector as a ‘Hidden’ Positive Outlier in Bribery Reduction. Birmingham: Developmental Leadership Program. Piotroski, J. D., Wong, T. J., & Zhang, T. (2022). Political Networks and Stock Price Comovement: Evidence from Network-Connected Firms in China. Review of Finance, 26(3), 521-559. Poddar, P., & Singh, S. K. (2022). Innovation And Corruption: Dissecting Causal Linkage Using Patent Application Information From India. Singapore Economic Review, 67(3), 1147- 1173. PRS Group. (2023). The ICRG Methodology. Liverpool, NY: PRS Group. Retrieved from https://www.prsgroup.com/wp-content/uploads/2022/04/ICRG-Method.pdf Quality of Government Institute. (2023). European Quality of Government Index. Retrieved from University of Gothenburg website: https://www.gu.se/en/quality-government/qog- data/data-downloads/european-quality-of-government-index Sanz, C., Solé-Ollé, A., & Sorribas-Navarro, P. (2022). Betrayed by the Elites: How Corruption Amplifies the Political Effects of Recessions. Comparative Political Studies, 55(7), 1095- 1129. Schwab, K., & Zahidi, S. (2020). The Global Competitiveness Report: Special Edition 2020. How Countries are Performing on the Road to Recovery. (Annex C. The Executive Opinion Survey: The Voice of the Business Community). Geneva: World Economic Forum. 32 Small, S. F. (2024). Bringing breadth and relevance to introductory economics courses using JEL codes. Journal of Economic Education, 55(4), 425-433. Spyromitros, E., & Panagiotidis, M. (2022). The impact of corruption on economic growth in developing countries and a comparative analysis of corruption measurement indicators. Cogent Economics & Finance, 10. Standaert, S. (2015). Divining the Level of Corruption: a Bayesian State Space Approach. Journal of Comparative Economics, 43(3), 782-803. doi:10.1016/j.jce.2014.05.007 Stephenson, M. (2023, January 31). Another Annual CPI Is Out. Yet Again, Here’s a Gentle Public Service Reminder Not to Focus on Year-to-Year Changes in Individual Countries’ Scores. Retrieved from The Global Anticorruption Blog: https://globalanticorruptionblog.com/2023/01/31/another-annual-cpi-is-out-yet-again- heres-a-gentle-public-service-reminder-not-to-focus-on-year-to-year-changes-in- individual-countries-scores/ Stoecker, A. (2022). Partisan alignment and political corruption: Evidence from a new democracy. World Development, 152, 1-15. Thomson Reuters. (2017). Thomson Reuters ESG Scores. Thomson Reuters. Retrieved from https://www.esade.edu/itemsweb/biblioteca/bbdd/inbbdd/archivos/Thomson_Reuters_ES G_Scores.pdf Tjandrasa, B. B., & Dewi, V. I. (2022). Determinants of Consumer Confidence Index to Predict the Economy in Indonesia. Australasian Accounting Business & Finance Journal, 16(4), 3-13. Tong, S. (2022, May). Corruption and anti-corruption in China: A review and future research agenda. Asian-Pacific Economic Literature, 36(1), 3-16. Transparency International. (1995). Press Release: New Zealand Best, Indonesia Worst In World Poll Of International Corruption. Retrieved from https://www.transparency.org/files/content/tool/1995_CPI_EN.pdf Transparency International. (2023). Corruption Perceptions Index Technical Methodology Note. Berlin: Transparency International. UNODC. (2023, August 21). Global Conference on Harnessing Data to Improve Corruption Measurement. Retrieved from UNODC: https://grace.unodc.org/grace/en/news-and- events/unodc-iaca-global-conference-on-harnessing-data-to-improve-corruption- measurement.html Wagstaff, A., & Culyer, A. J. (2012). Four decades of health economics through a bibliometric lens. Journal of Health Economics, 31, 406-439. Wathne, C., & Stephenson, M. C. (2021). The credibility of corruption statistics: A critical review of ten global estimates. U4 Issue Paper, 2021(4). World Bank. (1997). World Development Report 1997: The State in a Changing World. Washington: World Bank. 33 World Bank. (2000). Anticorruption in Transition: A Contribution to the Policy Debate. Washington: World Bank. World Bank. (2019). Anticorruption Initiatives: Reaffirming Commitment to a Development Priority. Washington, DC: World Bank. World Bank. (2021, October 25). Symposium on Data Analytics and Anticorruption. Retrieved from World Bank: https://www.worldbank.org/en/events/2021/10/25/symposium-on- data-analytics-and-anticorruption World Bank. (2023). World Bank Enterprise Surveys. Retrieved from World Bank: https://www.enterprisesurveys.org/en/enterprisesurveys World Bank. (2024, June 1). The importance of transparency and integrity in Public Procurement. Retrieved from ProACT Procurement Anticorruption and Transparency platform (prototype): https://www.procurementintegrity.org/about World Bank and Government Inspectorate of Vietnam. (2012). Corruption from the Perspective of Citizens, Firms, and Public Officials' Results of Sociological Surveys. Hanoi: National Political Publishing House. Retrieved from https://documents.worldbank.org/en/publication/documents- reports/documentdetail/989741468133546312/corruption-from-the-perspective-of- citizens-firms-and-public-officials-results-of-sociological-surveys World Values Survey. (2023). What We Do. Retrieved from World Values Survey website: https://www.worldvaluessurvey.org/WVSContents.jsp Zhuang, M. (2022). Intergovernmental Conflict and Censorship: Evidence from China's Anti- Corruption Campaign. Journal of the European Economic Association, 20(6), 2540– 2585. 34 Annex 1. Descriptions of cross-country indicators This annex provides descriptions of some of the cross-country measures used in studies of corruption in the cross-section published in 2022. The descriptions are drawn from the sources themselves. Composite indicators • Worldwide Governance Indicators Control of Corruption (WGI-CC): “The WGI covers over 200 countries and territories, measuring six dimensions of governance starting in 1996: Voice and Accountability, Political Stability and Absence of Violence/Terrorism, Government Effectiveness, Regulatory Quality, Rule of Law, and Control of Corruption. The aggregate indicators are based on several hundred underlying variables from various data sources. The data reflect the views on governance of survey respondents and public, private, and NGO sector experts worldwide. The WGI also explicitly reports margins of error accompanying each country estimate.” (Kaufmann, Kraay, & Mastruzzi, 2010) • Transparency International Corruption Perceptions Index (TI-CPI): The Corruption Perceptions Index (CPI) methodology “follows four basic steps: selection of source data, rescaling source data, aggregating the rescaled data and then reporting a measure for uncertainty. The calculation process also incorporates a strict quality control mechanism which consists of parallel independent calculations conducted by two in-house researchers and two academic advisors with no affiliation to Transparency International. The CPI draws upon 13 data sources which capture the assessment of experts and business executives on a number of corrupt behaviours in the public sector, including: bribery, diversion of public funds, use of public office for private gain, nepotism in the civil service, and state capture.” (Transparency International, 2023) • Bayesian Corruption Index (BCI): The Bayesian Corruption Index (BCI) “is a composite index of the perceived overall level of corruption. … It combines the information of 17 different surveys and 110 different survey questions that cover the perceived level of corruption. … Methodologically, it is most closely related to the latter as the methodology used in the construction of the BCI can be seen as an augmented version of the Worldwide Governance Indicators’ methodology.” (Bayesian Corruption Index, 2023; Standaert, 2015) • Basel AML Index (Basel): “The Basel AML Index uses a composite methodology based on 18 indicators relevant to evaluating ML / TF risk at the jurisdiction level. These are categorised into five domains in line with the five key factors considered to contribute to a high risk of ML / TF.” One of the domains is Domain 2: Corruption Risk which is based on Transparency International: Corruption Perceptions Index and TRACE: Bribery Risk Matrix. (Basel Institute on Governance, 2023) 35 Surveys • World Bank Enterprise Survey: “World Bank Enterprise Surveys (WBES) are nationally representative firm-level surveys, with top managers and owners of businesses interviewed using a globally comparable questionnaire that covers a broad range of business environment topics as well as firms’ characteristics and performance measures. All information collected through the WBES— raw granular data, the WBES indicators at the firm and economy level—are made publicly available upon completion of the surveys through our website and data portal. The website currently contains a total of 333 WBES collected through a consistent methodology across the world, 12 Informal Sector Enterprise Surveys covering 38 cities, and other surveys, along with the cross-economy databases.” (World Bank, 2023) • World Business Environment Survey: “The World Business Environment Survey, an initiative led by the World Bank Group in 1999 and 2000, collected enterprise data from more than 10,000 firms in 80 countries.” • World Values Survey (WVS): “The World Values Survey, which started in 1981, seeks to use the most rigorous, high-quality research designs in each country. The WVS consists of nationally representative surveys conducted in almost 100 countries which contain almost 90 percent of the world’s population, using a common questionnaire. The WVS is the largest non-commercial, cross-national, time series investigation of human beliefs and values ever executed, currently including interviews with almost 400,000 respondents. Moreover the WVS is the only academic study covering the full range of global variations, from very poor to very rich countries, in all of the world’s major cultural zones.” (World Values Survey, 2023; Haerpfer, et al., 2022) • World Economic Forum Executive Opinion Survey (WEF): “The Executive Opinion Survey (the Survey) ... is the longest-running and most extensive survey of its kind and provides a yearly evaluation of important aspects of socioeconomic development for which statistical data is missing because it is either impossible or extremely difficult to measure on a global scale. The aim of the Survey is to capture reality as well as possible, and business leaders are arguably the best positioned to assess the business environment in which they operate.” (Schwab & Zahidi, 2020). The survey includes a question on “irregular payments and bribes.” • Afrobarometer: “Afrobarometer provides reliable, timely data on the views of ordinary Africans to inform development and policy decision making. Afrobarometer is a pan- African, non-partisan survey research network that has been conducting public attitude surveys on democracy, governance, the economy, and society since 1999. On a two- to three-year cycle, our national partners in about 35 African countries carry out face-to- face interviews with nationally representative samples, then analyse the data and disseminate the findings.” • Life in Transition Survey. The EBRD conducts the “Life in Transition Survey (LITS) – a major survey of households and individuals in the economies where it invests – in 36 collaboration with the World Bank, in order to inform its operations. Four such surveys have been carried out so far: in 2006, 2010, 2016 and 2022-23.” The most recent round “surveyed over 37,000 households in 33 economies in the EBRD regions and four comparators,” and earlier rounds had sample sizes of 29,000 to 51,000. Expert opinion / Risk assessments • International Country Risk Guide (ICRG): “The International Country Risk Guide (ICRG) ... covers some 150 developed, emerging, frontier markets, and offshore banking centers. The ICRG methodology evaluates, scores, and ranks countries separately according to 32 political, economic, and financial risks. A composite score, derived from the risk categories, is provided for each country. The political risk ratings integrate 12 weighted metrics, covering a range of elements that could prove injurious to business and other commercial interests.” (PRS Group, 2023) • Varieties of Democracy (VDEM): “V-Dem uses innovative methods to aggregate expert judgments and thereby produce estimates of important concepts. We use experts because many key features of democracy are not directly observable. ... V-Dem typically gathers data from five experts per country-year observation, using a pool of over 3,700 country experts who provide judgment on different concepts and cases. Experts hail from almost every country in the world, allowing us to leverage diverse opinions. Despite their clear value, expert-coded data pose multiple problems. Rating concepts requires judgment, which varies across experts and cases; it may also vary systematically across groups of experts. We address these concerns by aggregating expert coded data with a measurement model, allowing us to account for uncertainty about estimates and potential biases.” (Marquardt, 2023) • European Quality of Government Index (EQGI): “The European Quality of Government Index (EQI) is the result of novel survey data regional (e.g. sub-national) level governance within the EU. The data was first gathered and published in 2010 and then repeated in 2013, 2017, and 2021. The index is based on a large citizen survey where respondents are asked about perceptions and experiences with public sector corruption, along with the extent to which citizens believe various public sector services are impartially allocated and of good quality. ... It covers all 27 EU member states, the UK before Brexit and two accession countries (Serbia and Turkey are also included in the 2013 round).” (Quality of Government Institute, 2023) Cross-country administrative data • Foreign Corruption Practices Act Clearinghouse (Stanford Law School) (FCPA): The Foreign Corrupt Practices Act Clearinghouse (FCPAC) functions as a comprehensive database, document repository, and analytics provider for enforcing the Foreign Corrupt Practices Act (FCPA). “The FCPAC comprises several unique but interconnected datasets: FCPA Matters, Enforcement Actions, Investigations, and Entities. Information 37 can be organized and presented in a number of different ways across each of these datasets. Information can be organized and presented in a number of different ways across each of these datasets. … We track all FCPA-related enforcement actions initiated by the SEC or DOJ since the statute's enactment in 1977. We also track all publicly disclosed and confirmed investigations into FCPA-related misconduct that are conducted internally by the company and/or externally by the SEC, DOJ or unspecified ‘US Authorities.’” (Foreign Corrupt Practices Act Clearinghouse, 2023) Indicators based on other sources • BBVA Corruption Research: Assessing Corruption with Big Data (BBVA): The BBVA measures corruption perceptions by constructing a Corruption Perception Index based on Google Trends data. This index covers 191 countries and has offered real-time and high- frequency data since January 2004. The methodology used by BBVA Research involves analyzing Google Trends data specifically for the topic ‘Corruption’ within the ‘Law & Government’ category. “Google Trends provided relative data: ‘numbers represent search interest relative to the highest point on the chart for a given region and time. A value of 100 is the peak popularity for the term, 50 indicates half as much popularity, and a value of 0 reflects insufficient data.’” (BBVA Research, 2018) • Thomson-Reuters ASSET4/ESG (Thomson-Reuters): “With over 150 content research analysts that are trained to collect ESG data, we have one of the largest ESG content collection operations in the world. With local language expertise and operating from different locations across the globe, we process numerous publicly available information sources with the aim of providing up to date, objective and comprehensive coverage. There are over 400 ESG measures, which our analysts process manually for each company within the Thomson Reuters ESG universe, with each measure going through a careful process to standardize the information and guarantee it is comparable across the entire range of companies.” (Thomson Reuters, 2017) A sub-indicator of the controversy score is the “number of controversies published in the media linked to business ethics in general, political contributions or bribery and corruption.” 38