Abstract The purpose of this research is to explain the force of homophily and in-group bias that involve some degree of discrimination towards others, in the context of the pension game, a transfer-making setting where mutual transfers are efficient, but transfers are individually costly. We study reciprocity, in-group bias and homophily experimentally using three scenarios; first, all participants are identical, second, participants are assigned randomly to two artificial groups, and third, participants can affect their likelihood to be matched with a participant of the same group in an endogenous matching scenario. The second and third scenarios are designed to measure the effects of group identity and group preference on transfer decisions. Although numerous studies exist on reciprocity and lately homophily, the experimental setup introduces a sequential matching structure relevant in many social transfer settings, such as pension savings. Results show that participants behaved differently in different matching structures, homophily, and in-group bias affect the maintenance of a cooperative outcome. In the case of sequential transfers, participants have transferred nearly 40% more in in-group transfers compared with out-group transfers. Regarding reciprocity, in random matchings, we see that there is higher reciprocity in out-group matches compared with in-group matches, meaning higher rewards together with higher sanctions in out-group transfers, but when we allow for preferential matchings, reciprocal behaviour is replaced with an inclination to control the similarity of partners.
Göktuna et al. (Tue,) studied this question.