Punishing liars—How monitoring affects honesty and trust

Abstract

Previous experiments have shown that the possibility to punish liars does not per se increase honesty in principal-agent relationships. In this study, we first establish a punishment mechanism that substantially enhances honest behavior and trust in a sender-receiver game: the possibility to impose severe sanctions that are cost-free for enforcers. Adopting this effective mechanism, we investigate how variations in the probability of detecting lies affect sender and receiver. We find that high honesty levels persist under such punishment mechanism even when the detection probability is significantly reduced. Furthermore, the relationship between monitoring and honesty does not follow a linear trend, as a moderate monitoring level proves to be less effective in enhancing honesty than high or very low levels. The punishment mechanism has an even more robust effect on receivers, showing similarly high levels of trust independently of the detection probability. Our analysis of subjects’ beliefs provides further insights into the mechanics behind these behavioral patterns.