Pricing and Competition between Heterogeneous Auction Hosting Sites1

This paper derives the analytics of seller and buyer participation in an auction hosting site, characterizes optimal hosting site pricing and inves­ tigates the nature of competition between two auction houses or hosting sites that are differentiated in the eyes of the bidders. The auction sites earn revenue by setting positive listing fees, trading off the increased revenue per seller from higher fees with the revenue reduction from the loss of sellers. The reduction in the number of sellers participating in a site has feedback effects, as it affects the number of bidders who would choose to visit that site. The sellers set the reserve in each sale to max­ imize their revenue. Unlike most prior models of auction site competi­ tion, buyers have preferences for bidding in the two sites. These pref­ erences may be driven by interface layout, prior experience with a site (which makes it less costly in terms of transaction costs to buy from that site), customer service experience, transaction reliability and reputation of sellers, and — more generally — anything that would fall under the rubric of site loyalty. Different buyers have different preferences for site interface and different experiences, resulting in some measure of hor­ izontal differentiation between the sites. This differentiation between the two sites implies that even if they both set the exact same rules and pricing policies, some buyers would strictly prefer to purchase from one, while others would strictly prefer to purchase at the rival site. Prefer­ ences for the two sites are not necessarily symmetric: one site may at­ tract more buyers than the other even if the number of sellers in the two sites is the same. However, differentiation results in market power and positive economic profits for both sites. We start by the careful analysis of the monopoly case, in which the second of the two firms is absent. We then consider the effect of the entry of the second firm and inves­ tigate how the equilibrium pricing policies of the two sites depend on the degree of differentiation, the degree of asymmetry between the sites, and the density of sellers. Though both sets of analyses are based on parameterized models, they form a testbed that can be generalized and modified to account for different types of pricing strategies and market features.