Estimating Discrete Games with Many Firms and Many Decisions: An Application to Merger and Product Variety

Abstract

This paper presents a method for estimating discrete games based on bounds of conditional choice probabilities. The bounds are probabilities that an action is dominant and that it is not dominated. Because the bounds are easy to compute, our method is scalable to models with many firms and discrete decisions. We apply the method to study the effects of a hypothetical merger on firm entry and product variety in local retail craft beer markets in California. We find that the merger induces firm entry. The net effect on product variety is ambiguous once a fixed cost efficiency is taken into account.