Abstract
The economic theory of crime deterrence predicts that the conviction of an innocent individual (type I error) is as detrimental to deterrence as the acquittal of a guilty individual (type II error). In this paper, we qualify this result theo- retically, showing that in the presence of risk aversion, loss aversion, or type I error aversion, type I errors have a stronger effect on deterrence than type II errors. We test these predictions with two experimental studies in which partic- ipants choose whether to steal from other individuals, under alternative combi- nations of probabilities of judicial errors. The results indicate that both types of errors have a significant impact on deterrence. As predicted, type I errors have a stronger impact on deterrence than type II errors. This asymmetry is entirely explained by differences in the expected utility gains from crime, whereas no- nexpected utility factors do not play a significant role.