Estimation and Identification of Demand and Supply

This paper surveys structural empirical methods used in Industrial Organization and their applications to antitrust. It brings together applied microeconomic theory, estimation technique and considerations of a practical nature. Topics covered include demand estimation for homogeneous and differentiated products, where we compare continuous- and discrete-choice models. We include a comprehensive yet accessible treatment of identification. To the demand side we then add structural models of supply, and the challenges these pose when it comes to identification and estimation. We review the alternative empirical strategies for backing out cost and conduct. The methods discussed are then compared to parsimonious alternatives for merger simulation arising in the antitrust practice such as the PCAIDS and the ALM, used by some antitrust authorities. We end by devoting a section to the practical aspects one has to consider when implementing structural models, such as data requirements and aggregation, specification tests, estimation technique and strengths and weaknesses of the different models. Our hope is to draw the connection between academics and practitioners.