Venue: KD²Lab Seminarraum

Group decision making: An experimental comparison of multilateral bargaining models

We experimentally test how useful the predictions of the models of Banks and Duggan (2000) and Compte and Jehiel (2010) are in organizing outcomes from multilateral bargaining. The basic difference between these models is how proposals are generated: endogenously through strategic choice vs. drawn from an exogenous distribution. The models lead to different predictions of which group members' opinions are important for the group's decision, conditional on decision rule and discount factor. In the experiment, we employ two types of environments: one with homegrown individual preferences based on a simple risky investment task and another one with individual ideal points induced by the experimental design. For groups of five, we systematically vary the decision rule (majority vs. unanimity) and the group members' induced discount factor. We find that both models do not perform well both in point predictions and in comparative statics across parameter constellations. Observed preference aggregation patterns are remarkably robust to varying time pressure and decision rules. In both environments and under all parameter settings, the median group member (in terms of individual preferences) has the highest impact on the group choice. Sometimes, the group members next to the median have some influence, too, while extreme group members are mostly ignored.