- 05 Jul 2006
- Working Paper
Information Dispersion and Auction Prices
Executive Summary — How can auctions be used most effectively? Government and industry traditionally use auctions to price and allocate assets and contracts with high but unknown value. Millions of people use Internet auctions for goods that are often of unknown value (e.g., used goods, unknown brands). This paper asks: Do bidders behave in the way auction theory predicts they should? And, what are the effects of different types of information on prices? To answer these questions, Yin combined theory, econometric modeling, and survey data. Key concepts include:
- Reputation lends credibility to the information about the product; both are important when analyzing auction prices.
- Information in auctions is dispersed among many participants. In analysis, this dispersion is a more important source of uncertainty than any information asymmetry between buyer and seller.
Do bidders behave as auction theory predicts they should? How do bidders (and thus, prices) react to different types of information? This paper derives implications of auction theory with respect to the dispersion of private information signals in an auction. I conduct a survey of non-bidders to construct a measure of information dispersion that is independent of bidding data. This permits joint tests of Bayesian-Nash equilibrium bidder behavior and information structure (common vs. private value) in a sample of eBay auctions for computers. The measure also allows me to separately estimate the price effects of seller reputation and product information. eBay prices appear consistent with Bayesian-Nash common value bidding behavior. Uncertainty about the value of goods due to information dispersed over auction participants plays a larger role than uncertainty about the trustworthiness of the sellers, but both are significant drivers of price. Thus, seller reputation complements, rather than substitutes for, information provided in the auction descriptions by lending credibility to that information, creating an incentive for sellers to reduce uncertainty in their auctions. (JEL C42, D44, D8, D82, L14, L15, L86)