Source: Dissertation Abstracts International, Volume: 64-10, Section: A, page: 3777.

Directors: Steven Berry; Martin Pesendorfer.

The dissertation considers the problem of identifying and measuring private information in First-Price auctions when unobserved auction heterogeneity is present.1

Previous studies that addressed this issue relied on the assumption that all the variation in bids is generated by the variation in private information therefore ignoring the possibility that unobserved heterogeneity may be present. If unobserved heterogeneity is present, then these methods may lead to biased estimates, which can substantially exaggerate the importance of private information in explaining bids variation.

I derive conditions under which private information can be identified under the presence of unobserved heterogeneity, develop a method for estimating the distributions of private information and unobserved heterogeneity components, and apply this method to data from Michigan highway procurement auctions. My results show that the bias induced by disregarding auction heterogeneity can be severe.

I modify standard model of competitive bidding in the first price auction to account for the possibility of unobserved heterogeneity. The modified model recognizes that bidders' costs may include a common component which is known to all bidders at the time they submit their bids as well as the individual component which is private information of the bidder. It is shown that model is identified if the individual components are independent across bidders and from the common component.

The results of Monte Carlo study illustrate that estimation methods based on the assumption of no unobserved heterogeneity, such as the independent private values model or the affiliated private values model, may lead to higher estimated mark-ups and biased estimate of the costs distribution function. The importance of both biases increases as the variation in the unobserved heterogeneity component increases.

Estimation method is applied to highway procurement data obtained from the Michigan Department of Transportation for the sample period 1997 to 2002. Bid functions, probability density functions of the individual and common components, the probability density function of total costs are estimated. The estimates suggest that the variation in the common component explains large part of the bid variation. They also show that under the null of unobserved heterogeneity both, the independent and the affiliated private values model tend to under-estimate the cost and over-estimate mark-ups over the bidders' costs. Reserve price chosen on the basis of the distribution estimate that ignores unobserved heterogeneity is lower than optimal.

1Unobserved auction heterogeneity may arise due to any information that is known to all bidders at the time they submit their bids but is not observed by the econometrician.