Market Definition - Still Needed after All These Years

Abstract

The 2010 Merger Guidelines appear to elevate game theoretic models of competitive concerns to the primary concern of merger policy, while reducing the importance of market definition. Although situations in which this change makes some sense exist, we observe that, as a practical matter, market definition is still required. Some basic understanding of the market’s contours is necessary to aid the decision maker in the choice among the relevant competitive effects models. Simply assuming that a specific theoretical analysis is correct is not scientific. Proponents of the view that markets are unnecessary rely on game-theoretic models that implicitly presume a particular competitive algorithm to create a viable alternative. Yet to displace the standard model of competitive analysis, a game theoretic model of competition requires both parameterization (estimation of the relevant coefficients) and testing (analysis of the implications of the model). Once a game theoretic model is tested, it becomes just one more Chicago-based analysis of the likely effects of a merger. Thus, theoretical considerations alone cannot eliminate the desirability of defining an antitrust market. Our review of the empirical data from Federal Trade Commission (FTC) merger investigations substantiates the potential for different market modeling structures to be applied to evaluate competition. Support for three different modeling structures (homogeneous product, static differentiation, and dynamic differentiation) is found in both a review of the court records and the FTC case studies. Choice among these modeling structures requires case-by-case analysis.

Keywords: antitrust policy, market definition, merger guidelines, critical loss

JEL Classification: K21, L40