Monday, November 7, 2011
Posted by D. Daniel Sokol
Joseph J. Simons, Paul, Weiss, Rifkind, Wharton & Garrison LLP and Malcolm B. Coate, U.S. Federal Trade Commission (FTC) provide A Comment on Choosing Among Tools for Assessing Unilateral Effects Analysis.
ABSTRACT: In a recent paper, Gregory Werden and Luke Froeb present a general discussion of unilateral effects analysis with a particular focus on factors that limit the applicability of the Upward Pressure on Price (UPP) model. In one of our earlier papers, we had provided simple simulations showing that the application of UPP in the broad manner suggested by Farrell and Shapiro could be used to dramatically expand the universe of mergers subject to challenge. Werden and Froeb, in contrast to Farrell and Shapiro, see application of UPP screening as constrained by Bertrand competition, an assumption that implies our simulations over-estimate the potential for the UPP model to increase enforcement. Moreover, they posit that merger simulation is well suited for use in either screening or competitive effects’ analysis, a result, which if true, would marginalize the UPP model. We disagree with their observations, noting that the UPP methodology, as described by Farrell and Shapiro, is designed to be agnostic to the underlying competitive process, and potentially applicable in a range of situations in which simulation would be infeasible. As a result, our simulations do not over-estimate the extent to which a broad application of UPP could possibly expand merger enforcement. We also stress the importance of verifying the predictions of any theoretical model (including an UPP analysis generalized to predict actual price effects) with empirical evidence as an important tool to realistically limit the analysis. Successful substantiation would be necessary to allow the analysis to survive a Daubert challenge. Under Daubert, experts must move beyond theory that is generally accepted in their fields and present evidence to the court to show that the theory fits the facts at issue and is reliably predictive.