What determines the extent of electoral fraud? This paper constructs a model of the tradeoff between fraud and policy concessions (public good provision) which also incorporates the strength of the state. In addition, we parameterize the extent to which economic elites (to whom fraud is costly) and political elites (to whom fraud is advantageous) “overlap.” The model predicts that fraud will be lower and public good provision higher when land inequality is higher, the overlap between elites lower, and the strength of the state higher. We test these predictions using a unique, municipal‐level dataset from Colombia's 1922 Presidential elections. We find empirical support for all the predictions of the model.