Experiments on Social Identity

This thesis consists of three chapters examining the effectsects of social identity in an experimental framework. The first chapter uses a novel subject pool, registered members of British political parties and subjects play the Ultimatum Game. Proposers and responders make offers or state their minimum acceptable offer (MAO) conditional on the political identity of the second player. Additionally we determine how behaviour in the ultimatum game is affected when proposers have the opportunity to earn additional income prior to a system of taxation and redistribution. We find that proposer's offers are significantly larger when paired with an in-group responder and responder's MAOs are significantly lower when paired with an in-group proposer. This is robust to the inclusion of earnings and redistribution. The second chapter uses a laboratory experiment where we determine the effects of social identity on coordination and efficiency in a weakest-link game with endogenous linking. We find that endogenous linking alone increases efficiency through the exclusion of low effort players. Subjects are able to coordinate on a high effort level and maintain a highly connected network. However, the addition of social identity neither helps nor hinders this. As such we conclude that in problems of coordination with endogenous linking, social identity considerations are eclipsed by efficiency. The final chapter considers social identity as a factor in determining the level of redistribution in society. We create two social identity groups where one group is numerically and economically dominant and then compare behaviour to a baseline treatment where social identity is absent. In both treatments we elicit subjects preferences for redistribution before they behave both as candidates and as voters in an electoral setting. Additionally we vary the equity-efficiency trade-off under which subjects make their decisions. We find that elections with only minority candidates result in the highest level of redistribution. Subjects are more likely, both to vote and to propose higher levels of redistribution under these candidates. However preferences for high redistribution are greatest when the equity-efficiency trade-off is positive, that is, when more equality is more efficient.