Determinants of corruption

Abstract

In this paper we study determinants of the level of certain types corruption in a society. To that end, we apply the simplest technique used in Evolutionary Game Theory, namely, the replicator dynamics with two types of agents, corrupted, and not corrupted. Through a learning interpretation of that technique, we obtain the main determinants of corruption are the initial proportion of corrupted people, and the relative pecuniary gain of being corrupted, relative to the pecuniary gain of being not corrupted. The model applies to all types of corruption for which the larger the number of corrupted people is, the larger the expected payoff of being corrupted will be.