Objectives. We seek to investigate the determinants of corruption in authoritarian polities. We hypothesize that corruption in nondemocratic settings will be greater where the ruling group is personalistic rather than a political party or a military clique and that it will be greater where rulers expect to remain in power longer. We construct a new operationalization of the selectorate theory advanced by Bueno de Mesquita et al.

Methods. We use cross-sectional statistical analysis (OLS) to examine a sample of 40-odd authoritarian regimes as of 2000.

Results. Our results indicate that personalistic and personalistic-hybrid regimes are more prone to corruption than single-party and military regimes and also that rulers who expect to remain in power for longer are less corrupt. Corroborating previous studies, we document that the availability of natural resources and higher levels of institutionalized autocracy are associated with greater corruption and that wealthier countries experience less corruption. Our results are consistent with previous studies, including that of Bueno de Mesquita et al., but because of our reconstruction of selectorate theory in terms of real-world regime types, they are more easily interpretable.

Conclusions. Our study sheds light on why African countries are so notoriously corrupt. The personalistic authoritarian regimes that have arisen there in the postcolonial period appear especially prone to corruption, whereas military and single-party dictatorships are less corrupt.