In order to be more effective in networked markets, the US should adopt a version of the EU’s “abuse of dominance” standard that could be directed against network monopolists’ anticompetitive conduct toward smaller network participants. A properly constrained abuse of dominance standard would enable networks to perform in a competitive environment that is neither dominated nor deprived of the features that make networks valuable.
Section 2 of the Sherman Act is sweeping in its coverage. In a single sentence, it prohibits every firm who shall “monopolize,” or attempt, or conspire to monopolize, without any limitations. The term “monopolize” is itself interesting because it had so few relevant antecedents. Prior to 1890, when the Sherman Act was passed, the legal meaning of monopoly was an exclusive right granted by the government. Important Supreme Court monopoly decisions, such as the Charles River Bridge case (1837) and the Slaughter-House Cases (1872) involved the existence or scope of exclusive government grants to engage in certain economic activity. Debates over the proper scope of patents or other intellectual property laws were also dominated by concerns about monopoly, but these were also government-created rights.
Section 2 reaches firms that had become dominant or threatened dominance without an exclusive grant from the government. That phenomenon was neither well-established legally nor well-understood economically at the time. Judges searched for sources of law. The earliest antitrust cases that considered what it meant to monopolize were heavily driven by tort law and focused largely on conduct—things such as acquiring rivals and shutting them down, predatory pricing, secret rebates or other discriminatory pricing, or exclusive deals with suppliers.
The rapid rise of industrial organization economics in the early twentieth century switched the focus more to the study of markets and structures. By World War II, the law of monopolization came to be dominated by questions about structure. These concerned such things as the relationship between the number of firms in a market and its competitiveness, and whether some market structures were more prone to monopolization than others. For a period, both prominent writers and some courts flirted with the idea of “no fault” monopolization—or the idea that having a large market share was all that the offense required. Congress never passed such a “no fault” provision, and the courts have never adopted one on their own.
No fault monopolization rules are no longer realistically on the table, although some people continue to call for breaking up very large firms without regard to their conduct. There are good reasons for not doing that. Maximum sustainable market output benefits both consumers and labor by offering lower prices, higher quality, and greater innovation. At the same time, however, it may require large firms. Simply breaking them up hurts not only consumers, who will pay higher prices and face lower quality and less innovation, but it also harms workers, whose output is proportional to product output. The better approach is to scrutinize the conduct of dominant firms carefully.
Section 2 of the Sherman Act could be made more effective than it is today. As currently formulated, US monopolization law contains two requirements. The first is a showing of monopoly power, or dominance in a market, which typically requires a market share of 60 percent or more. In addition, the firm must have committed at least one anticompetitive act that has served to create, perpetuate, or strengthen its monopoly.
Those requirements have generally worked well in traditional markets, where each firm is expected to stand on its own bottom and develop or contract for its own needs. The dramatic rise of collaborative networks has changed that. Antitrust, which is not a form of command-and-control regulation, generally takes markets as it finds them. Today, networking is a fact of life. Collaborative networks are more dependent on interconnection and other interactions among multiple firms, including information sharing and agreed upon product standards. In these situations, very large firms often play an outsize role.
Many networked markets contain significant amounts of path-dependence over shared technologies, and this entails that dominant firms that go rogue can do significant economic damage. The Ninth Circuit Court of Appeals was insufficiently mindful of this in its 2020 Qualcomm decision. The court rejected an antitrust claim against a firm that was defecting from commitments it had made under a FRAND system for cross-licensing of technology in a networked market. In a FRAND system, firms have voluntarily obligated themselves to share their patents with other participants on a nondiscriminatory basis and at reasonable royalties. Such arrangements are contractual, and antitrust law should not be contract enforcement by another name. Nevertheless, when a dominant firm’s failure to honor such commitments causes significant competitive harm, neither is the existence of a contract a defense.
One feature that limits the reach of Section 2 of the Sherman Act is that it requires a likelihood of monopoly in each market to which it applies. On this issue, US law differs from that of many other countries. For example, the EU describes the analogous offense as “abuse of a dominant position.” This formulation permits EU law to reach conduct that causes harm in collateral markets even if it does not present a significant threat of monopoly in the second market.
By contrast, Section 2 of the Sherman Act does not permit claims of harm falling short of creating a second monopoly in collateral markets. For example, in the Microsoft decision, the government won most of its case based on monopolization of the Windows operating system, where Microsoft’s monopoly power was established, but not for the attempted monopolization of the Internet Explorer browser. The browser was actually the sword that Microsoft employed against rival browser Netscape. In another case involving Alaska Airlines (1991), the court refused to condemn dominant carrier United’s manipulation of a shared airline reservation system so as to discriminate against small airlines because there was not a realistic threat that the defendant would drive the smaller carriers out of business but only that it would harm them.
In the mid-twentieth century, the Supreme Court was more sympathetic to claims of injury in collateral markets. In a 1948 case involving Griffith Amusement Company, the owner of about 150 theaters followed the customary practice of employing a single agent who bargained with film distributors for exhibition licenses for all of them. Because Griffith was a large company, it obtained better rates than smaller owners who had only one or two theaters. The Supreme Court accepted the government’s claim that the defendant unlawfully used its monopoly power in some towns to obtain an unfair advantage in more competitive towns, even though monopolization of those towns was unlikely. This so-called monopoly “leveraging” theory was accepted for nearly a half century, until it was repudiated by the Supreme Court in 1993 and again in 2004.
This history suggests two things. First, in order to be more effective in networked markets, the United States should adopt a version of the “abuse of dominance” theory that could be directed against network monopolists’ anticompetitive conduct. Smaller firms in networked markets are more dependent on the cooperation of others. Second, however, overreaching is a serious risk unless the offense is carefully limited. Some alleged anticompetitive abuses, like the multi-theater bargaining in Griffith, are nothing more than cost-saving practices that can favor firms that operate in multiple markets. Others, such as Alaska Airlines, are more threatening to competition.
Proposed legislation in New York would adopt an abuse-of-dominance standard for that state’s first ever monopolization provision. But the proposed legislation unwisely provides that “evidence of pro-competitive effects shall not be a defense.” That is a mistake that could end up denying users much of the benefit that can result from networked markets. An abuse of dominance standard should enable networks to perform in a competitive environment that is neither dominated on the one hand nor deprived of the features that make networks valuable on the other. As a result, it must be made to distinguish beneficial from harmful practices. A better formulation is the one contained in the Clayton Act, which condemns practices where the effect “may be substantially to lessen competition.” That would require courts to identify some harm to competition that results from non-monopolistic abuses of dominance.
Learn more about our disclosure policy here.