B.J.Pol.S. 32, 147–170 Copyright  2002 Cambridge University Press Printed in the United Kingdom Sources of Corruption: A Cross-Country Study GABRIELLA R. MONTINOLA AND ROBERT W. JACKMAN* Why is government corruption more pervasive in some societies than in others? In this article we examine public choice explanations that attribute corruption to a lack of competition in either political or economic arenas or both. The principal part of our analysis draws on recently- published data about levels of corruption for a broad cross-section of countries reported for the early 1980s. We supplement this with an additional analysis of a second dataset on corruption measured during the late 1980s. Our analyses confirm that political competition affects level of corruption, but this effect is nonlinear. Corruption is typically lower in dictatorships than in countries that have partially democratized. But once past a threshold, democratic practices inhibit corruption. However, we obtained mixed results with respect to the relationship of economic competition and corruption: government size does not systematically affect corruption, but membership of the Oil Producing and Exporting Countries (OPEC) does. Finally, corruption is more pervasive in low-income countries which tend to underpay public sector employees. Why is government corruption more pervasive in some countries than in others? This question intrigued scholars studying developing countries in the 1960s, because the presence of corruption represented an apparent anomaly. The decolonization that began after the Second World War and that culminated around 1960 was supposed to have generated many new liberal democratic states. After all, nationalism and the attack on colonialism were both legitimized in democratic terms, and the key political actors of the day equated self-government and democratic government. 1 Observed performance in many of the new states fell far short of these expectations, however, and in rapid order. Indeed, the problems of personal rule and corruption that became abundantly clear shortly after independence have often continued unabated since, as evidenced most visibly by the long-lived regimes of former presidents Mobutu and Suharto of Zaire and Indonesia, respectively. Our purpose is to offer a systematic accounting for observed cross-country differences in corruption. The principal part of our analysis draws on data about levels of corruption for a broad cross-section of countries reported for the early 1980s. We supplement this with an additional analysis of a second dataset on * Department of Political Science, University of California, Davis. We would like to thank the editors and referees of the Journal for their comments. This research was supported in part by the Academic Senate at UC Davis. 1 Rupert Emerson, From Empire to Nation: The Rise to Self-Assertion of Asian and African Peoples (Cambridge, Mass.: Harvard University Press, 1960); and Edward Shils, ‘The Fortunes of Constitutional Government in the Political Development of the New States’, in Jon H. Hallowell, ed., Development: For What? (Durham, NC: Duke University Press, 1964), pp. 103–43.