Post-Cartesian Antitrust

Abstract

The Supreme Court’s 2018 opinion in American Express v. Ohio is among the most important – and divisive – antitrust opinions in the modern era of antitrust law. The simplest statement of Justice Thomas’s opinion for the majority is that it saw a 5 Justice majority of the Court fully embrace the relatively new economic understanding of two-sided markets. Supporters of the majority opinion almost uniformly view it as an obviously-correct application of important and generally-accepted recent development in economic theory. Those more amenable to Justice Breyer’s dissenting opinion do not necessarily reject the theory of two-sided markets but instead would treat arguments premised on this theory as a pro-competitive justification (that is, a defense) to what could reasonably be understood as potentially anticompetitive conduct under prevailing antitrust economics.

The central argument of this paper is that both perspectives miss the forest for the trees. The Court’s American Express opinion is not narrowly about whether (or how) antitrust law should embrace the theory of two-sided markets. Rather, I argue, this opinion is part of the Court’s ongoing efforts to understand how antitrust law should evaluate markets that are not neatly “horizontal” or “vertical."

I call these efforts to understand competition in markets that are not clearly horizontal or vertical “post-Cartesian” antitrust, and describe these markets as “messy markets.”

The important antitrust cases today do not fit neatly into the mold of either horizontal or vertical conduct. Simple horizontal or vertical market structures are not often the subject of antitrust litigation today. We see this in cases ranging from Apple v. Pepper (and the earlier Apple e-books litigation), to the AT&T-Time Warner merger litigation, from Leegin to Actavis, and of course in American Express. Indeed, the market structures in these cases are so different from horizontal and vertical markets that the tools developed for analysis of horizontal and vertical markets are as inapposite to them as SCP methodologies are to markets for differentiated products, markets characterized by rapid innovation, entry, or network effect, and high fixed costs.