Identification, Estimation, and Applications in Auction and Differentiated Products Markets

National Bureau Of Economic Research Inc, Cambridge MA

Investigators

Abstract

This project involves research on empirical methods and applications in two broad areas of industrial organization: 1) auctions and 2) demand for differentiated products. Auctions are important market institutions and provide a unique window on strategic behavior in markets with asymmetric information. Demand estimation is fundamental to a wide range of empirical work in microeconomics. The work on auctions includes methodological research on estimation of first-price auctions when unobservables affect the sample of auctions observed, the set of bidders participating, and the bids they make. We provide new identification results, a new estimation approach, and simulation studies to evaluate the methods. The project also includes an empirical examination of Federal mineral lease auctions on the U.S. Outer-Continental Shelf. Over the last 25 years there have been several major changes in technology and policy that are likely to have significant effects on these sales. We examine the effects of these changes on bidder participation, bidding, drilling, and production. We study the effects of three-dimensional imaging technology on the information structure and bidding at these auctions, and also assess the controversial royalty relief program. The work on demand for differentiated products is comprised of three components, each examining what can be learned under weaker assumptions than those typically considered previously. The first considers the case in which consumer-level characteristics and choices are observable (micro data). The model of preferences is nonparametric and distribution free, nesting random coefficients discrete choice models widely used in practice with parametric functional form and distributional assumptions. It incorporates choice-specific unobservables and a very flexible specification of how preferences vary across products and individuals. This component of the work provides a deeper understanding of the data and assumptions required to obtain reliable estimates. The second project considers the case in which one observes only market level data. This case is more challenging and requires a different type of argument; however, we again obtain positive results for less restrictive models than those previously considered. The third project examines demand estimation allowing for the fact that when firms choose the products to offer, this leads to correlation between observable and unobservable characteristics of products. We provide conditions for identification, then develop, evaluate, and apply new estimation methods. The methodological components of this work addresses issues of importance to the analysis of economic and policy questions in fields ranging from anti-trust policy to the economics of education and the allocation of public resources. Some of the empirical work addresses issues of direct importance to government policy in the U.S. The work will be disseminated broadly through research seminars, conference presentations, and publication of research articles. This work will contribution to the training of graduate students, who will serve both as research assistants and co-authors.