Common value auctions with default: Theory and experimental evidence

This dissertation examines common value auctions in which bidder default is allowed. Contractual default on the part of bidders is commonly observed in many auction settings and yet has remained virtually unexplored in the literature. We develop a two bidder model of a common value auction with default. We solve for the equilibria in both the default and no-default cases. Two key theoretical results are produced by the model: (i) bidders are more aggressive in their bidding when default is allowed, and (ii) allowing default can actually lead to an increase in the auctioneer's expected revenue in certain situations. In addition to the theoretical approach, we use a laboratory experiment to study this problem. Our experimental results confirm result (i). Subjects in our experiments do indeed bid more aggressively when default is allowed. Result (ii) is rejected. We find that in our experimental auctions the auctioneer is always better off by not allowing bidder default. We also compare our experimental results to the existing literature. This approach extends the experimental literature on common value auctions to include cases of limited liability. We also show that a winner's curse exists in auctions with only two bidders.