Posted by Social Science Research Network
How Much Does the Choice between Collusion and Unilateral Effects Matter in Merger Analysis?
By Malcolm B. Coate & Shawn W. Ulrick (Federal Trade Commission)
Abstract: In many, but certainly not all, merger investigations, the analyst faces a choice between the coordinated interaction (collusion) and unilateral effects models when organizing a competitive review. This paper applies statistical analysis to model the implications of this decision and determine if the choice appears to materially affect the outcome of the merger review process. By first modeling the combined choice of theory of concern and policy outcome, we show that the statistical properties of the error terms allow the analysis to focus on the challenge decision. Next, we estimate merger policy models for both collusion and unilateral effects variables to determine if theory affects merger policy. As the models materially differ, they are used to parameterize a decomposition analysis and evaluate the implications of the theory choice. An alternative analysis applied a matching model to explore the effect of the theory choice on the outcome of the investigation. We find that that the choice of theory has an effect in homogeneous goods market, with a movement from a collusion theory to a unilateral theory reducing the average challenge rate by 12.7-17.4 percentage points and a movement from a unilateral theory to a collusion theory raising the challenge rate by 6.1 to 11 percentage points. For differentiated products, the changes are smaller and insignificant.
