Cheating and corruption: Evidence from a household survey

Abstract

This study tests the prediction that a corrupt government reduces ethical behavior among its citizens. We integrate a standard "cheating" experiment into a broad household survey and find clear support for this prediction: respondents who perceive corruption in state affairs are more likely to cheat. Interestingly, there is a small group of non-conformers. The main relation is robust to consideration of many (largely insignificant) socio-demographic control variables. Attendance of others at the cheating experiment, thus stimulating the reputational concern to be seen as honest, reduces cheating. Again, this does not diminish the predictive role of corruption.