Empirical determinants of corruption: A sensitivity analysis

Many variables have been proposed by past studies as significant determinants of corruption. This paper asks if their estimated impact on corruption is robust to alteration of the information set. A Global Sensitivity Analysis, based on the Learner's Extreme-Bounds Analysis gives a clear answer: five variables are robustly related to corruption. Corruption is lower in richer countries, where democratic institutions have been preserved for a long continuous period, and the population is mainly Protestant. Corruption is instead higher where political instability is a major problem. Finally, a country's colonial heritage appears to be a significant determinant of present corruption.