Unilateral Effects Session 18.2.04 Froeb Remarks

One of the best things about this job is being able to finish the work of my predecessor. In particular, the enforcement data collection began under Dave Scheffman and I want to acknowledge the work he has done, and thank him for leaving me to enjoy the fruits of his labor.

By 1990 game theory had displaced the structure-conduct-performance paradigm in Industrial Organization economics. Thinking on pricing and output coordination in oligopolies had evolved considerably from the view that fewness made coordination almost inevitable. While economists never entirely rejected coordinated effects theories for mergers, they did reject exclusive reliance on them, and they had more plausible theories for many cases. Thus, it was not surprising that unilateral effects analysis appeared prominently in the 1992 Horizontal Merger Guidelines, which were jointly promulgated by the FTC and the Department of Justice.

The unilateral effects analysis satisfied the attorneys' demand for a simple intuition they could understand and explain to a judge, and at the same time it satisfied the economists' demand for rigorous analysis. It wasn't long before economists began using structural, game-theoretic models to make quantitative predictions of unilateral competitive effects. In a price setting model, price goes up; in an auction models firms don't have to bid as aggressively to win; in a bargaining model the merged firm gains bargaining power; and in a quantity setting model ... well we learned that mergers aren't usually profitable in quantity setting models.

The controversy surrounding unilateral effects analysis has focused on the application of structural game theoretic merger models to individual cases. Their main virtue is that they force assumptions to be made explicit, and provide a mapping from the facts of a case to the effects of a merger. The main shortcoming is that the models are necessarily unrealistic, and abstract away from important features of an industry. As such, their results may be quite misleading. Unfortunately, we have little evidence on whether these models can accurately predict the effects of real mergers. Instead, we are left with controversy about what is the best way to analyze unilateral effects theories. In what follows we will hear from Valerie Rabassa, Joseph Kattan, Greg Werden, Tad Lipsky, and Greg Leonard.

1. (Tad, Joe, Greg W.) The basic economics of unilateral effects from horizontal mergers goes back over a century, but was not a major factor in merger enforcement prior to the release of the 1992 Horizontal Merger Guidelines. Since then, the Department of Justice and the Federal Trade Commission have often relied on unilateral theories in challenging mergers, but few cases have been tried, and there is little case law support for unilateral effects theories. Do you believe a court would or should be reluctant to accept a unilateral effects theory?

2. (Tad, Joe, Greg W.) It is possible to allege both unilateral and coordinated effects in a merger case. They could be plead in the alternative or as cumulative effects. Do you think it is likely to be a good strategy for a plaintiff to allege both theories, or does each theory tend to undermine the other? In defending a merger case in which both theories were alleged, how might you use evidence in support of one theory to undermine the other?

3. (Valerie) Section 2.211 of the Horizontal Merger Guidelines twice refers to a combined share for the merging may of 35%. Do you understand this language to create a safe harbor? Do you believe there should be a special safe harbor for differentiated products mergers? If you had to craft such a safe harbor would it be based on the combined share of the merging firms or the change in the HHI, and would it be more or less inclusive than the 35% rule?

4. (Greg L.) Merger simulation-the formal use of structural, game-theoretic models to make quantitative predictions of unilateral competitive effects--offers the prospect of quantification and can make clear on what predictions depend. Do you believe a court would be receptive to merger simulation? Do you believe the economic models used in merger simulation are too simplistic to be of much use?

BIG PICTURE

5. (Greg L. Valerie) Where does the general consensus in the application of unilateral effects analysis lie and where and what are the "growing edges". Where should more research be done? What are the key questions?

6. (Joe, Tad, Greg W.) What criticisms do you have of the Merger Guidelines discussion of unilateral effects, and what positive suggestions do you have for improvement?

7. (Joe, Tad, Greg W.) What criticisms do you have of the Agencies' application of unilateral effects analysis and what positive suggestions do you have for improvement?