THE ECONOMETRICS OF ENGLISH AUCTIONS

We characterize Bayesian Nash equilibria for asymmetric ascending auctions in which both common and private value components are accommodated, and bidders' valuations and signals are drawn from asymmetric distributions. It is shown that the equilibrium inverse bid functions in each round of the auction can be solved from a system of equations formed from bidders' expectations of the value of the object conditional on all information available during that round. The computational tractability of this characterization allows us to analytically solve for the equilibrium bidding strategies in a log-normal model. This forms the basis of an econometric model which we estimate using data from the PCS spectrum auctions run by the U.S. Federal Communications Commission.