FORMATION AND CONTROL OF COLLUSION IN A SEALED OFFER MARKET: AN EXPERIMENTAL EXAMINATION.

Abstract

The main objective of this study is to examine the behavior of sellers in a sealed offer market in three distinct environments. The dissertation examines the mechanism to determine its competitive characteristics when sellers are unable to engage in overt communication. Next, the market setting is changed in order to consider the ability of sellers to form stable and effective cartels when the subjects are given the opportunity to conspire. Finally, the consequences of antitrust enforcement is considered. First, by reducing the number of communication episodes between sellers and second, by separately introducing random monetary penalties whenever the observed winning offer was greater than sellers cost. A secondary but significant contribution of this research was the use of prison inmates as subjects in economic experiments. Therefore, the financially motivated behavior of student subjects could be contrasted with prisoner subjects in a laboratory environment. This contrast proves extremely interesting when an antitrust environment, using randomly imposed monetary penalties, was introduced. The results of my research indicate, given equal constant cost suppliers, the sealed offer auction mechanism has the potential to generate very competitive behavior by sellers when communication between subjects is prohibited. However, in an identical environment, when conspiratorial opportunities exist prior to every auction, one observes very stable and effective cartels. These first two conclusions are insensitive to the subject pool under consideration (i.e., student versus prisoner subjects). When the frequency of communication is reduced (i.e., as a result of antitrust enforcement) the market is still conducive to effective collusion. However, in comparison to the experiments where subjects were allowed to communicate prior to every auction, the cartel in this environment were more unstable and less effective at extracting monopoly profit from the market. Finally, when antitrust enforcement is introduced by imposing random monetary penalties whenever the observed winning offer is greater than seller cost, the behavior of sellers is extremely sensitive to the expected returns from collusion as well as the dispersion of these returns. Also, the cartels collusive behavior indicated that students were risk averse, responded more to changes in the severity of the penalty specification (the penalty amount) while the prisoner cartels were risk takers, being more sensitive to changes in the certainty of the penalty specification (the detection level).