Who do you lie to? Social identity and the cost of lying

We investigate whether and how an individuals' propensity to lie is affected by the social relationship between a potential liar and her/his possible victim. We argue that a shared social identity of sender and receiver increases sender's aversion to lie by raising two types of costs: the allocative and the social costs of the lie. Allocative costs should be larger in ingroup interactions because social preferences are stronger and thus losses to the receiver are weighted more heavily while social costs should be higher in closer relationships due to stricter moral rules. In contrast to our hypothesis our experimental results from a modified three-person sender-receiver game do not provide evidence that social identity affects lying behavior. While across all treatments about half of the participants send a dishonest message, we do not observe differences in lying behavior towards ingroup and outgroup members: neither with respect to allocative nor in terms of social costs. Hence, in our experiment lying behavior is robust to social identity manipulations.