abstract. The rise of large firms in the digital economy, including Amazon, Apple, Facebook, and Google, has rekindled the debate about monopolization law. There are proposals to make finding liability easier against alleged digital monopolists by relaxing substantive standards; to flip burdens of proof; and to overturn broad swaths of existing Supreme Court precedent, and even to condemn a law review article. Frank Easterbrook’s seminal 1984 article, The Limits of Antitrust, theorizes that Type I error costs are greater than Type II error costs in the antitrust context, a proposition that has been woven deeply into antitrust law by the Supreme Court. We consider the implications of this assumption on the standard of proof. We find that, taking variants of the Easterbrook assumption as given, the optimal standard of proof is stronger than the preponderance of the evidence standard. Our conclusion is robust to how one specifies the preponderance of the evidence standard and stands in stark contrast to contemporary proposals to reduce or eliminate the burden of proof facing antitrust plaintiffs in digital markets.

Introduction

No area of modern antitrust has attracted greater debate than monopolization law—the body of antitrust law applied to a single firm’s conduct alleged to harm competition.1 These debates have been rekindled in recent years with the rise of large firms in the digital economy, including Amazon, Apple, Facebook, and Google, among others. Antitrust scholars, institutions, and politicians have offered a variety of proposals to change substantive antitrust rules applied to the digital economy.2 The proposals almost uniformly3 recommend changes that would make a finding of liability easier against alleged digital monopolists by relaxing substantive standards;4 flipping burdens of proof so that conduct is presumed unlawful;5 or overturning broad swaths of existing Supreme Court precedent.6

The tradeoff between Type I and Type II errors lies at the heart of analyses of monopolization rules and standards for digital markets. Type I errors, or “false positives,” refer to false convictions; Type II errors refer to “false acquittals.” In the antitrust context, a Type I error refers to a finding that conduct that is actually procompetitive violates the antitrust laws. A Type II error in the antitrust context refers to a failure to find antitrust liability for anticompetitive conduct. The opportunity for Type I and Type II error is especially high in a dynamic competitive enrivonment, where it is very difficult to discern the competitive effects of a firm’s conduct from observing a business practice alone. As the D.C. Circuit put it in the landmark Sherman Act section 2 case applied to digital markets:

Whether any particular act of a monopolist is exclusionary, rather than merely a form of vigorous competition, can be difficult to discern: the means of illicit exclusion, like the means of legitimate competition, are myriad. The challenge for an antitrust court lies in stating a general rule for distinguishing between exclusionary acts, which reduce social welfare, and competitive acts, which increase it.7

The design of antitrust rules to minimize costs of judicial error in digital markets is our focus.8 While economic and decision-theoretic analyses have been applied to substantive antitrust rules and sanctions, they have rarely been deployed to consider antitrust process and procedure. We take a novel approach, turning instead to considering the standard of proof used in antitrust trials as an instrument to minimize the costs associated with the two possible types of judicial error.9

Our analysis takes as given a presumption about the relative incidence of Type I and Type II error in markets, introduced by Frank Easterbrook in his seminal 1984 article The Limits of Antitrust.10 Easterbrook famously reasoned that preventing procompetitive behavior is more harmful than allowing anticompetive behavior. Easterbrook argued not that markets were perfectly self-correcting, but that incentives to enter and compete for the monopoly profits in markets impacted by anticompetitive behavior would constrain the social costs of Type II errors more than the legal system could succesfully limit the social costs of Type I errors. Easterbrook contended that “[i]f the court errs by condemning a beneficial practice [committing Type I error], the benefits may be lost for good. Any other firm that uses the condemned practice faces sanctions in the name of stare decisis, no matter the benefits. If the court errs by permitting a deleterious practice [committing Type II error], though, the welfare loss decreases over time.”11

Easterbrook’s assumption that Type I error costs are greater than Type II error costs in the antitrust context is not without controversy among academics. Undoubtedly this is at least in part because it is very difficult to test the proposition empirically. We do believe the assumption is appropriate and sound, based upon available economic theory and evidence, and offer a brief justification for it in Part II. The purpose of our analysis, however, is not to substantiate the Easterbrook assumption. After all, there is no debate that Easterbrook’s view of the relative social cost of Type I and Type II errors has been fully incorporated into antitrust law—and in particular, monopolization law under section 2 of the Sherman Act.12 We take that assumption—that the beneficial impact of procompetitive behavior is greater than the harmful impact of anticompetitive behavior—and provide a novel analysis to show that the optimal standard of proof in the antitrust context is greater than “preponderance of evidence.” These findings are counterintuitive and certainly an outlier among the dozens of proposals to lower liability standards and reduce the evidentiary burdens facing plaintiffs in monopolization cases in digital markets and elsewhere.

We describe the role of the Easterbrook assumption in modern antitrust law in Part I. Part II offers a brief and partial defense of the Easterbrook assumption on economic and empirical grounds. Part III presents our analysis of optimal antitrust standards of proof for monoplization and its application to digital platforms. Part IV concludes.

The full text of this Essay can be found by clicking on the PDF link to the left.