We reconsider the identification and estimation of Gorman‐Lancaster‐style hedonic models of demand for differentiated products in the spirit of Sherwin Rosen. We generalize Rosen’s first stage to account for product characteristics that are not observed and to allow the hedonic pricing function to have a general nonseparable form. We take an alternative semiparametric approach to Rosen’s second stage in which we assume that the parametric form of utility is known, but we place no restrictions on the aggregate distribution of utility parameters. If there are only a small number of products, we show how to construct bounds on individuals’ utility parameters, as well as other economic objects such as aggregate demand and consumer surplus. We apply our methods to estimating the demand for personal computers.

This paper surveys recent studies of internet auctions. Four main areas of research are summarized. First, we survey several studies that document and attempt to explain the frequently observed sniping, or last-second bidding behavior, in these auctions. Second, we summarize several methods proposed to quantify the distortions caused by asymmetric information in these markets, most notably due to the winner's curse. Third, we explore research about the role of reputation mechanisms installed to help combat these distortions. Finally, we discuss what internet auctions have to teach us about auction design.

We develop an approach to identify and test for bid rigging in procurement auctions. First, we introduce a general auction model with asymmetric bidders. Second, we study the problem of identification in our model. We state a set of conditions that are both necessary and sufficient for a distribution of bids to be generated by a model with competitive bidding. Third, we discuss how to elicit a prior distribution over a firm's structural cost parameters from industry experts. Given this prior distribution, we use Bayes's theorem to compare competitive and collusive models of industry equilibrium. Finally, we apply our methodology to a data set of bidding by construction firms in the Midwest. The techniques we propose are not computationally demanding, use flexible functional forms, and can be programmed using most standard statistical packages.

Linear Regression Estimation of Discrete Choice Models with Nonparametric Distributions of Random Coefficients by Patrick Bajari, Jeremy T. Fox and Stephen P. Ryan. Published in volume 97, issue 2, pages 459-463 of American Economic Review, May 2007

Recently, economists have developed methods for structural estimation of auction models. Many researchers object to these methods because they find the strict rationality assumptions to be implausible. Using bid data from first‐price auction experiments, we estimate four alternative structural models: (1) risk‐neutral Bayes‐Nash, (2) risk‐averse Bayes‐Nash, (3) a model of learning, and (4) a quantal response model of bidding. For each model, we compare the estimated valuations and the valuations assigned to bidders in the experiments. We find that the risk aversion model is able to generate reasonable estimates of bidder valuations.

We describe a two-step algorithm for estimating dynamic games under the assumption that behavior is consistent with Markov perfect equilibrium. In the first step, the policy functions and the law of motion for the state variables are estimated. In the second step, the remaining structural parameters are estimated using the optimality conditions for equilibrium. The second step estimator is a simple simulated minimum distance estimator. The algorithm applies to a broad class of models, including industry competition models with both discrete and continuous controls such as the Ericson and Pakes (1995) model. We test the algorithm on a class of dynamic discrete choice models with normally distributed errors and a class of dynamic oligopoly models similar to that of Pakes and McGuire (1994).

We survey and apply several techniques from the statistical and computer science literature to the problem of demand estimation. To improve out-of-sample prediction accuracy, we propose a method of combining the underlying models via linear regression. Our method is robust to a large number of regressors; scales easily to very large data sets; combines model selection and estimation; and can flexibly approximate arbitrary non-linear functions. We illustrate our method using a standard scanner panel data set and find that our estimates are considerably more accurate in out-of-sample predictions of demand than some commonly used alternatives.

We propose a new strategy for a pervasive problem in the hedonics literature: recovering hedonic prices in the presence of time-varying correlated unobservables. Our approach relies on an assumption about home buyer rationality, under which prior sales prices can be used to control for time-varying unobservable attributes of the house or neighborhood. Using housing transactions data from California's Bay Area between 1990 and 2006, we apply our estimator to recover marginal willingness to pay for reductions in three of the EPA's “criteria” air pollutants. Our findings suggest that ignoring bias from time-varying correlated unobservables considerably understates the benefits of a pollution reduction policy.