Differentiated product competition and the Antitrust Logit Model: an experimental analysis

This paper reports a laboratory experiment designed to begin a behavioral examination of the Antitrust Logit Model (ALM), a merger simulation device that U.S. antitrust authorities use to help determine when anticompetitive problems may arise from horizontal mergers in differentiated-product markets. We find that the ALM screens out non-problematic mergers rather well, even though the ALM predicts performance in specific markets imprecisely. Further examination of the data suggests that in this context, adjustments to pre-merger deviations from the underlying Nash equilibrium, rather than the exercise of market power drive post-merger performance.