*** 9731025 Lawrence M. Ausubel Peter C. Cramton Project Abstract During the 1990s, auction theory has been asked to provide guidance to policy makers in two important areas: the design of Treasury auctions (and other auctions of multiple units of a homogeneous good), and the design of FCC auctions (and other auctions of heterogeneous items). Unfortunately, the received theory has largely focused on auctions of a single object, and so has had relatively little to say about these two empirically-relevant auction design problems. Recent research has begun to expand the boundaries of auction theory and analysis so as to encompass these two important multiple-item auction problems, by studying the theoretical properties of existing multi-unit auction formats, empirically evaluating the spectrum auctions, and proposing new methods and applications of auctions. The current project builds upon the principal investigators' recent work in this area, by examining the following topics: Equilibria of multi-unit auctions where each bidder's valuation depends on other bidders' information. Auctions with resale markets or pure common values. Efficient information revelation versus tacit collusion in dynamic auctions. Extent of simultaneous versus sequential sale when selling multiple items. Empirical analysis of the U.S. Treasury auctions. Continuing evaluation of the spectrum auctions as well as evaluation of electricity auctions. The project provides a balance of further theoretical development of multiple-item auctions and empirical testing of the ideas of auction theory. This blend of theoretical and empirical work, with both tied closely to auction practice, has been especially productive. ***
