Journal of Public Economics 76 (2000) 399–457www.elsevier.nl/locate/econbaseThe causes of corruption: a cross-national study*Daniel TreismanDepartment of Political Science,University of California,Los Angeles, 4289Bunche Hall,Los Angeles,CA90095-1472,USAAbstractWhy is corruption — the misuse of public office for private gain — perceived to be morewidespread in some countries than others? Different theories associate this with particularhistorical and cultural traditions, levels of economic development, political institutions, andgovernment policies. This article analyzes several indexes of ‘perceived corruption’compiled from business risk surveys for the 1980s and 1990s. Six arguments find support.Countries with Protestant traditions, histories of British rule, more developed economies,and (probably) higher imports were less ‘corrupt’. Federal states were more ‘corrupt’. Whilethe current degree of democracy was not significant, long exposure to democracy predictedlower corruption.  2000 Elsevier Science S.A. All rights reserved.Keywords:Corruption; Bribery; Government; Federalism; DemocracyJEL classification:D72; D73; H11; H77; K421. IntroductionWhy is corruption — defined here as the misuse of public office for private gain— perceived to be more widespread in some countries than others? Understandingthis is important for several reasons. Corruption has been blamed for the failures ofcertain ‘developing’ countries to develop, and recent empirical research confirms alink between higher perceived corruption and lower investment and growth(Mauro, 1995; World Bank, 1997). Political scandals in countries across the globe*Tel.: 1 1-310-825-3274; fax: 1 1-310-825-0778.E-mail address:[email protected] (D. Treisman)0047-2727/00/$ – see front matter  2000 Elsevier Science S.A. All rights reserved.PII: S0047-2727(99)00092-4400 D.Treisman / Journal of Public Economics76 (2000) 399–457have sparked public outrage against corruption in recent years, and in dozens ofcountries discredited governments have been forced out of office. At the sametime, corruption is viewed as one of the main obstacles that post-communistcountries face in attempting to consolidate democratic institutions and open,market economies (Shleifer, 1997).Yet very little is known for sure about what causes corruption to be higher inone place than another. While theories abound, and while numerous case studieshave examined the details of corruption in particular countries or regions, cross-national comparative empirical research is much rarer. The difficulty of measuringlevels of relative corruption in different countries has presented a major obstacle.Recently, however, economists and political scientists have begun to analyzeindexes of ‘perceived’ corruption prepared by business risk analysts and polling1organizations, based on survey responses of businessmen and local residents.While such ratings are by definition ‘subjective’, there are compelling reasons tobe interested in the patterns they reveal. First, such cross-national ratings tend tobe highly correlated with each other and highly correlated across time. Differentorganizations using different techniques derive ratings that are similar and that donot change much from year to year. Furthermore, indexes of relative corruptionconstructed from surveys of business people operating in specific countries turnout to be highly correlated with at least one cross-national poll of the inhabitantsof those countries. This reduces the chance that one is analysing not perceptions ofcorruption but the quirks or bias of a particular monitoring organization. Second,as empirical work confirms, whatever the objective characteristics of a country’spolitical and social system, subjective evaluations of corruption do themselvesappear to influence investment decisions, growth, and the political behavior ofcitizens (Mauro, 1995). The main perceived corruption index used in this paperalso correlates positively with the size of the unofficial economy, as estimated byJohnson et al. (1998)This paper uses three annual indexes of perceived corruption (for 1996, 1997and 1998) prepared by the organization Transparency International (TI) to assessthe explanatory power of various theories of the causes of corruption. The TI index¨constitutes a ‘poll of polls’, compiled by a team of researchers at GottingenUniversity using information from up to 12 individual surveys and ratings. As willbe shown, country scores on this index correlate closely across years and alsocorrelate quite highly with two other available indexes constructed in the previous1For a few examples of use of perceived corruption indexes, see Mauro (1995), La Porta et al.(1997a, 1999), Easterly and Levine (1997), Ades and Di Tella (1999). La Porta et al. (1999) examinehistorical, cultural, and economic determinants of a variety of indicators of government quality,including corruption. This paper, which developed in parallel and has circulated in draft form since1997, confirms some of their findings, suggests an alternative interpretation of one, and adds severalnew results.D.Treisman / Journal of Public Economics76 (2000) 399–4574012decade. As a robustness check, I also examine an index of perceived corruptioncompiled by the organization Business International (BI) for the early 1980s.Multiple regressions are used to analyze the data, with variables capturing a broadrange of theoretically plausible determinants included simultaneously in the hopeof reducing omitted variable bias. Results are tested for robustness, and problemsof endogeneity are, where possible, investigated.While the complexity of the issues and the weakness of available statisticaltechniques makes it essential to be cautious, the analysis does suggest someinteresting results. I find strong and robust support for five arguments and slightlyweaker support for a sixth. First, I replicate in this dataset the findings of La Portaet al. (1999) that countries with Protestant traditions and those with moredeveloped economies have higher quality governments. Both factors are sig-nificantly and robustly associated with lower perceived corruption. Evidence inthis paper suggests that causation runs from economic development to lowercorruption as well as from corruption to slower development. I find in addition thatcountries with a history of British rule were robustly rated ‘less corrupt’. La Portaet al. (1999) argue