Traditionally economic theory assumes that preferences are stable facilitating positive predictions of economic policy. While there is conflicting experimental evidence on the temporal stability of cooperation preferences in public goods provision, surprisingly little is known about their stability in different institutional settings. We contribute to this literature by testing whether social identity impacts on cooperation preferences in public goods provision. Specifically, our experiment features a within subject design based on one shot public good games in strategy method, which are carried out in random, in and out group matching protocols. Our findings indicate that cooperation preferences are not stable across these matching circumstances. Quite to the contrary, we find that when matched with in group members, subjects consistently show the preference for higher levels of conditional cooperation and thus less self serving bias than in out group matching. Additionally, while the probability to be a conditional cooperator remains stable under each treatment, we identify an elevated propensity to be a free rider when matched with individuals of a different identity. These results indicate that it can be reasonable to devise policy institutions that strengthen the feeling of belonging to a particular group in order to enhance social welfare.
public goods cooperation preferences conditional cooperation free riding social identity experimental economics