We study endogenous group formation in tournaments employing experimental three-player contests. We …nd that players in endogenously formed alliances cope better with the moral hazard problem in groups than players who are forced into an alliance. Also, players who are committed to expending e¤ort above average choose to stand alone. If these players are forced to play in an alliance, they invest even more, whereas their co-players choose lower e¤ort. Anticipation of this exploitation may explain their preference to stand alone.JEL codes: D72, D74

No abstract

We study the incentives to share private information ahead of contests, such as markets with promotional competition, procurement contests, or R&D. We consider the cases where …rms have (i) independent values and (ii) common values of winning the contest. In both cases, when decisions to share information are made independently, sharing information is strictly dominated. With independent values, an industry-wide agreement to share information can arise in equilibrium. Expected e¤ort is lower with than without information sharing. With common values, an industry-wide agreement to share information never arises in equilibrium. Expected e¤ort is higher with than without information sharing.

We study con ‡ict between two groups of individuals. Using Schaffer's (1988) concept of evolutionary stability we provide an evolutionary underpinning for in-group altruism combined with spiteful behavior towards members of the rival out-group. We characterize the set of evolutionarily stable combinations of in-group favoritism and out-group spite and …nd that an increase in in-group altruism can be balanced by a decrease in spiteful behavior towards the out-group.