In a second-price auction (SPA) with private values, the bidder who submits the highest bid wins the auction and pays the second highest bid. In a SPA, bidders simultaneously submit one bid each. According to theory, the bidder's best strategy is to bid exactly what the item is worth to her. In game theory parlance this optimal bid is known as a "dominant strategy," that is, it is optimal for each bidder regardless of what others are bidding. Dominant strategies have robust properties but are extremely rare in auction environments. Past evidence from laboratory experiments has shown that bidders almost always bid above the amount corresponding to the dominant strategy in an SPA, i.e. they behave irrationally according to the theory.
This project examines the conjecture that, while bidding behavior is not rational, it does respond in a sensible and predictable way to incentives. A laboratory experiment will test whether bids above the dominant strategy increase when such overbidding is less costly, and decrease when it is more costly, even though the experiment is constructed so that these incentives do not affect the value of the dominant strategy bid itself.
Should the conjecture prove correct, the results will inform designers of selling mechanisms. Although game theory makes very clear predictions, bidders do not often behave exactly as predicted. Results from investigations that examine systematic deviations from the theory can be incorporated by auction designers who wish to maximize revenue or efficiency. Given that such auctions have been used by both government and the private sector to sell a variety of securities and physical assets, the practical applications of such research are widespread.
