Empirical and theoretical analysis of overlapping electronic markets

Date of Completion

January 2006


Business Administration, Management




Today's web based auctions have enabled mechanism level interactions or inter-dependency of outcomes, which were not feasible in physical auctions. One such set of interactions happens due to the existence of timely overlapping heterogeneous auction formats (having different winner and price determination rule) for the same goods at the same site. These interactions and their impact on consumer behavior and the auctioneer revenue have not been studied in the literature. Most of the research considers the auction mechanism as an independent and isolated environment for buying and selling goods. However, Internet technologies have drastically reduced the spatial and temporal constraints, and enabled the consumers to easily access to concurrently available online auctions and this makes the interaction or inter-dependency of multiple auctions possible. In this dissertation, I study the theoretical and empirical properties of overlapping auctions, with sequential and simultaneous auctions being interesting special cases. We conceptualize and examine the impact of market level factors such as the price information revealed from prior auctions, degree of overlap, and the overall market supply on a given auction's price. We expand the bidder taxonomy proposed by Bapna et al. (2004), to identify institutional bidders that tend to actively bid on a large number of auctions. Based on the empirical regularities observed from a real world data set, a theoretical framework is developed to explain how a given auction's price is influenced by the properties of the mechanism, other overlapping auctions, market factors and bidder behavior. We use the insights to analyze the optimal multiple overlapping auction market design strategy. ^