Information Disclosure and Cooperation in a Finitely-Repeated Dilemma: Experimental Evidence

A large volume of past research has suggested that making information on people’s past behaviors visible to others may enhance cooperation in finitely repeated environments. But, do people cooperate with randomly-matched peers by voluntarily revealing their past when they have an option to conceal it? This paper experimentally shows that while voluntary information disclosure does help strengthen cooperation in a random-matching environment, such disclosure does not have effects if it involves a cost because a large fraction of people does not reveal their past. The data also shows that, when subjects can choose an environment with a reputation mechanism or one without it, a stable number of subjects join the reputation community (where their past is revealed) and cooperate with others, especially when the sorting involves a cost. However, some subjects stay away from the reputation community and fail to cooperate with peers in the ‘anonymous’ community (where their past is not revealed). When there is no cost for sorting process, the reputation mechanism’s high efficiency may decrease because some subjects frequently switch between the two communities and attempt to exploit cooperators, although they could cooperate with others if they were forced to disclose without opportunities to choose.