Antitrust mixed logit model

Abstract

This paper presents the Antitrust Mixed Logit Model (AMLM), a novel methodology that shows how to calibrate the parameters of a mixed-logit demand model and simulate the competitive effect of horizontal mergers. The major advantage over the simpler Logit version (the Antitrust Logit Model, ALM,developed by Werden and Froeb,1994) is flexibility, resulting in more plausible elasticities and consequently more precise predictions about merger effects. Moreover, unlike the econometric approaches, the AMLM shares with the ALM the attributes that are particularly appealing to antitrust agencies, given time and data constraints they usually face: low data requirement and high computational speed. This model is applied to simulate mergers in the U.S. ready-to-eat cereal industry.