Files in this item
|(no description provided)|
|Title:||Four Essays on Auction Theory|
|Author(s):||Menezes, Flavio Marques|
|Doctoral Committee Chair(s):||Kahn, C.M.,|
|Department / Program:||Economics|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
|Abstract:||The first essay examines the outcome of an ascending-price multiple-object auction. Two bidders with continuous, convex demand functions participate in the auction of a certain number of divisible objects. The auctioneer starts the process by announcing an initial price and by asking both bidders to submit sealed-bids (desired amounts). The auctioneer keeps increasing the price until the total amount of bids is less than the total supply. This mechanism has been used to sell stocks of government owned companies under the Brazilian Privatization Program. We compute the outcome of this auction game under full information and conclude that this mechanism has perverse effects in terms of revenue maximization.
The second essay incorporates the existence of small bidders who cannot play strategically. A genetic algorithm is used to investigate the effects of learning and evolution in participants' behavior. Our results suggest that strategic bidders learn how to play the game in few rounds. We also conclude that as the number of players behaving strategically decreases the solution approaches the competitive outcome.
The third essay describes equilibria for a restricted version of the auction of the first essay under incomplete information. We found both separating and pooling equilibria to exist where an agreement is reached in at most two rounds. Unfortunately, these equilibria do not hold in the unrestricted game and there are also equilibria where an agreement is reached at the last round.
We investigate in the fourth and last essay the outcome of second-price sequential auctions where players face delay costs if they decide to stay for subsequent rounds. First, we consider the case where agents have to decide to stay for the next round before seeing their valuations for the good. For an arbitrary number of stochastically equivalent objects, we specify the conditions under which ex-ante expected prices are monotonically decreasing. Second, we study an alternative model where agents decide whether to stay for the subsequent round after seeing their valuations for the good. For a two-period model we describe an equilibrium involving a screening level such that only those bidders with values bigger than that amount decide to stay for the last period. We also provide an example where prices are decreasing.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1993.
|Date Available in IDEALS:||2014-12-17|