Can groups solve the problem of over-bidding in contests?

This study reports an experiment that examines whether groups can better comply with theoretical predictions than individuals in contests. Our experiment replicates previous findings that individual players significantly overbid relative to theoretical predictions, incurring substantial losses. There is high variance in individual bids and strong heterogeneity across individual players. The new findings of our experiment are that groups make 25% lower bids, their bids have lower variance, and group bids are less heterogeneous than individual bids. Therefore, groups receive significantly higher and more homogeneous payoffs than individuals. We elicit individual and group preferences towards risk using simple lotteries. The results indicate that groups make less risky decisions, which is a possible explanation for lower bids in contests. Most importantly, we find that groups learn to make lower bids from communication and negotiation between group members