Micro-Directives and Computational Merger Review

Abstract

AI technologies can improve upon the current system of merger notification and review. Predictive technologies—such as supervised machine learning—combined with unprecedented growth in data will provide antitrust agencies with the opportunity to better refine the law and the review process. Such technologies will build upon how antitrust agencies already model and predict the likely consequences of mergers. Here, we explore how such predictions can reduce both the over-inclusiveness and under-inclusiveness inherent in the current system of merger notification and review. We explore the possibility of a more automated system of merger review. We argue that the greatest hurdle to the adoption of such a system is not feasibility, technological limitations, or the availability of data. Rather, the greatest hurdle is the difficulty in pinning down a precise and translatable ex ante objective that such an algorithm would optimize.