Are we more honest than others think we are?

While the laws are justified on the basis of the efficiency they provide to society, policy makers and researchers focus on the reasons why people violate the law. Crimes and violations induce directly costs. But there is another indirect costs that is generally ignored : the fact that a person can violate the law (whether it does or not) can reduce trust in one's honesty. Thus, even if the economic agent is honest and respects the law, this loss of confidence, which could be unfounded, is also a source of inefficiency. We introduce in an experiment, a normative rule of "decision" in order to elicit both honesty and beliefs about honesty from subjects in the lab. There is no direct transfer of money between both part to avoid any inequality aversion or altruism aversion. The main question remains how individuals trust in the honesty of an anonymous group. Subjects are split into two groups : those who are subject to the temptation of (unverifiable) dishonesty and those who value the dishonesty of others. We inform each participant that we cannot identify defection. We find an important heterogeneity of trust in honesty through subjects. On average, subjects A suggests that participants B are more honest than they are. Moreover, we identify distortion of effective honesty and beliefs about other honesty when the environment of players A is unfavorable.