Addressing Endogenous Product Choice in an Empirical Analysis of Merger Effects

Industrial organization economists have made significant progress on consumer demand estimation in product differentiated markets. Crucially, price is endogenized in these models. As a result researchers can predict a merger’s consequences by estimating ex ante demand parameters, changing the “identity” of the various players (i.e., combining the behavior of merging parties), allowing prices to adjust endogenously, and calculating changes in consumer welfare. Because the price effects of mergers are critical to consumer welfare, models that endogenize price play an important role in antitrust analysis. In many applications, however, this approach to evaluating mergers does not go far enough. While prices are explicitly included as choice variables of the industry participants, their product characteristics are treated as exogenous. As such, product characteristics are constrained to be fixed after the merger. This abstraction has consequences for the accuracy of a simulation; to the extent that merged firms may cull duplicate products or competitors may introduce new varieties post-merger, evaluating consumer surplus becomes more complex. If the resulting product variety in the industry exhibits more overall differentiation, price competition may be softened beyond the effect of removing one competitor. However, industry participants may choose to expand their product offerings given the new market structure. Consumer welfare can increase if products with more highly valued characteristics are available more after a merger. This paper focuses on three aspects of endogenous product choice related to merger simulation. First, we elaborate on the role of abstracting from potential post-merger product repositioning in the economics literature and selected antitrust cases. Second, we demonstrate the potential impact of the fixed product characteristics assumption on consumer welfare citing empirical, theoretical and simulation examples. Finally, we introduce an empirical framework that incorporates endogenous product choice into differentiated-product demand estimation and discuss how this and similar methods can potentially be incorporated into antitrust analysis.