Sustaining Group Reputation

When individuals trade with strangers, there is a temptation to renege on agreements. If repeated interaction or exogenous enforcement are unavailable, societies often solve this problem via institutions that rely on group, rather than individual, reputation. Groups can employ two mechanisms to uphold reputation that are unavailable to individuals: information sharing and in-group punishment. We design a laboratory experiment to distinguish the roles of these mechanisms when individual reputations are unobservable. Subjects are split into groups and play a trust game with random re-matching, where only the group identity of one's partner is known. Treatments differ by whether information about group members' transactions is shared and whether in-group punishment is possible. We find that information sharing encourages path dependence via group reputation: good (bad) behavior results in greater (fewer) gains from exchange in the future. However, the mere threat of in-group punishment is enough to discourage bad behavior