The theory of auctions has developed extensively since Robert B. Wilson's (1977) seminal paper. Due to the complexity of equilibrium strategies, however, empirical researchers have been slow to incorporate and test the most basic theoretical precepts.' Typical empirical studies estimate ad hoc bidding models, with no attempt to ascertain whether the implied behaviors are theoretically plausible. The recent paper by Stuart E. Thiel (1988) attempts to bridge this gap. Thiel obtains closed-form equilibrium bidding functions that are theoretically motivated and linear in parameters, and which facilitate empirical estimation and testing. If widely applicable, Thiel's empirical approach would constitute a major methodological breakthrough. Unfortunately, Thiel's approach applies only in special cases that are of limited practical interest. Linear bidding strategies emerge only under circumstances that are unlikely in the real world. The limited range of Thiel's approach may not be apparent to casual readers of his paper. Even when linear strategies do exist, they are not unique under the special assumptions of Thiel's model. For each Nash equilibrium in linear strategies, there exists a related family of nonlinear strategies. Thus, further justification must be found for basing empirical research on the linear specification. We also note a significant error in Thiel's work. The symmetric strategies he derives on the basis of order statistics do not constitute a Nash equilibrium. We derive proper expressions for the symmetric Nash strategies and discuss a specification error in Thiel's regression analysis that would account for the mixed results obtained in his application to the highway-construction industry.

We experimentally test an endogenous-timing investment model in which subjects privately observe their cost of investing and a signal correlated with the common investment return. Subjects overinvest, relative to Nash. We separately consider whether subjects draw inferences, in hindsight, and use foresight to delay profitable investment and learn from market activity. In contrast to Nash, cursed equilibrium, and level-k predictions, behavior hardly changes across our experimental treatments. Maximum likelihood estimates are inconsistent with belief-based theories. We offer an explanation in terms of boundedly rational rules of thumb, based on insights about the game, which provides a better fit than quantal response equilibrium.

John H. Kagel, Ronald M. Harstad, Dan Levin, Information Impact and Allocation Rules in Auctions with Affiliated Private Values: A Laboratory Study, Econometrica, Vol. 55, No. 6 (Nov., 1987), pp. 1275-1304

The Winner's Curse and Public Information in Common Value Auctions: Reply by Colin M. Campbell, John H. Kagel and Dan Levin. Published in volume 89, issue 1, pages 325-334 of American Economic Review, March 1999