The Arc of Antitrust: A Text-based Measure of Antitrust Policy Beliefs and Attitudes

Abstract

There is an ongoing debate whether antitrust policy in the US has been too lax in the past decades. Despite an abundance of data detailing the actions of antitrust authorities and numerous studies that seek to address the issue, the controversy persists. Various statistical and econometric challenges complicate the use of metrics such as merger counts, litigation success rates, or market concentration levels to accurately gauge the rigor of antitrust enforcement. I introduce a text-based measure of antitrust beliefs and attitudes that complements data-based attempts to assess antitrust stringency. I employ a machine learning classifier, trained on a curated dataset of blog posts and other antitrust-related texts, to generate an index that categorizes texts along an ideological spectrum that ranges from Chicago School thinking at one end to a Neo-Brandeisian perspective at the other. I use this index to explore diverse corpora of antitrust texts, shedding light on the evolving landscape of beliefs and attitudes over time. These corpora include antitrust blogs, an official FTC blog, speeches by FTC commissioners and DOJ officials, as well as antitrust legislation. The results show the arc of antitrust in the US, that is, the shift from interventionist Brandeisian thinking until the late seventies, followed by several decades of Chicago School-inspired, more permissive antitrust, and more recently, a discernible reversal towards a more Neo-Brandeisian position.