Studying Differentiated Product Industries using Plant-Level Data

This paper develops a methodology to uncover consumer preferences from a discrete-choice demand model of product differentiation using plant-level data. When prices and quantities are observed, the appropriate strategy for estimating such model is well developed. However, most plant-level data sets only report revenue (sales) and total cost, hampering initial attempts to estimate the model according to standard approaches. This paper offers a way to circumvent this problem by bringing the extra information provided by usually observable aggregate data to determine the relevant parameters. The methodology consists of solving for the demand parameter that matches the total quantity implied by the model of demand and supply to its observable counterpart. Once this parameter is determined it is possible to define welfare measures and perform counterfactual simulations. This methodology is applied to measure the economic impact of a regulatory agency imposing a monopoly break up.