Doctoral Dissertation Research in Economics: Knowledge of Rivals' Types and Its Effect on Auctions

University Of Wisconsin-Madison, Madison WI

Investigators

Abstract

The purpose of the doctoral dissertation project is to experimentally test the effect of bidders' information of their rivals' types in standard auctions. Auctions are used widely for trading goods and services, ranging from artwork and agricultural products to government assets such as PCS spectrum rights, timber and mineral harvesting rights. While the existing studies on auctions have made great strides, our understanding remains far from complete. In particular, how bidders' knowledge of their rivals affects their bidding behavior and the performance of various auctions is not well understood. The existing work assumes that bidders have either no information or full information about their rivals. In practice, however, bidders' knowledge about their rivals tends to be asymmetric. For instance, in a procurement auction contested by domestic firms and foreign firms, domestic firms are likely to be better informed about the technical capabilities of their domestic rivals than those of their foreign rivals. The current project will contribute to our understanding of the impact bidders' information has on their behavior and on the performance of standard auctions. Bidders' asymmetric knowledge about their rivals matters and it may affect both the bidders' behavior and the performance of alternative auction forms. The first chapter of my dissertation, written jointly with Yeon-Koo Che, has also established a number of hypotheses regarding how the equilibrium bidding behavior and the revenue performance of a given auction will change as bidders learn more about their rivals' types. Testing our hypotheses requires controlling the information flow of the bidders, which would be virtually impossible if one relied on the traditional empirical method employing secondary data. A lab experiment is ideal since the experiment designer can control the information flow precisely the way it is modelled in the theory. We plan to conduct 6 sessions of lab experiments, each of which will invite 32 subjects to participate in a series of auction games. The research will feature some innovative techniques to improve the empirical identification of the subjects' bidding behavior. While the main purpose of the experiment will be to test the specific hypotheses of the theory chapter, it will in the process contribute to the testing of the underlying Nash hypothesis, by subjecting it to more comparative static restrictions. Further, the test is expected to reveal the role bidders' risk aversion plays in auction experiments.