Michelucci, F.; (2007) Essays on auction theory. Doctoral thesis , University of London.

PDF U592147.pdf Download (3MB)

Abstract

The thesis is composed of three essays on auction theory. The focus is mainly on the English auction, which is probably the format most commonly used. In the English auction, a bidder can see his opponents' activity as the price is gradually raised starting from a very low price. When only one bidder is left, such bidder is declared the winner and pays the price at which the last but one bidder exited. The vast majority of the studies assume that the price is raised continuously and that a bidder cannot outbid the current winner by a discrete amount, or to use the terminology employed in the literature to place a jump bid. Such a restriction is often unrealistic and examples of jump bidding can be found in many relevant contexts. In the first two chapters, we analyze settings in which jump bidding is allowed. Previous studies suggest that the possibility of placing a jump bid can be used to signal one's strength and induce the opponents to quit (on average) earlier than they would have otherwise. In our first paper we follow this approach. In particular, as Fishman (1988), we assume that the purpose of signalling is to discourage the opponents from acquiring finer information regarding their valuations. As opposed to Fishman (1988), we look at an environment where bidders' valuations are determined not only by a private (as he does) but also a common value element. We show that the presence of a common element makes jump bidding harder. Secondly, we prove that when the bidder placing the jump knows his total value but not the relative importance of the common value part, a jump bidding equilibrium can be sustained more often and it yields higher profits than when the bidder knows exactly such value. In the second paper, we introduce a different and new rational for jump bidding, which involves no signalling or information costs. In an interdependent value setting, where a bidder's value depends on the valuation of the other bidders, it is crucial whether a bidder's exit price is known or not. Jump bidding offers the possibility of hiding the exit value of some opponent, thus affecting the expected value of the remaining bidders and ultimately their bidding behaviour. We illustrate when hiding such information might be profitable. We also show that its effect both on revenues and efficiency is in general ambiguous. Finally, in the last paper (joint work with Hernando-Veciana) we contribute to the important issue of determining which is the most efficient way to allocate an object among a set of potential buyers. Previous studies provide both conditions under which it is possible to allocate the object to the buyer with the highest willingness to pay (First Best), and mechanisms that can implement such an allocation. We characterize the most efficient allocation when the First Best cannot be implemented. Then, we study whether the English auction can implement it. While an equilibrium of the English Auction that implements the Second Best exists, it is in general not robust if there are more than two bidders.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as JSONXMLCSV

Archive Staff Only

View Item