Document Type


Publication Date



This is a response to a query from the Judiciary Committee of the U.S. House of Representatives, requesting my views about the adequacy of existing antitrust policy in digital markets.

The statutory text of the United States antitrust laws is very broad, condemning all anticompetitive restraints on trade, monopolization, and mergers and interbrand contractual exclusion whose effect “may be substantially to lessen competition or tend to create a monopoly.” Federal judicial interpretation is much narrower, however, for several reasons. One is the residue of a reaction against excessive antitrust enforcement in the 1970s and earlier. However, since that time antitrust has shifted very far in the other direction. Today the marginal antitrust decision is much more likely to reflect under- rather than overenforcement. A second reason is the fact that many judges obtained any antitrust training they received a quarter century or more ago. Since then, notable progress in theoretical and empirical economics has both improved our techniques of analysis and shown the need for greater enforcement, particularly in markets with a significant technological or digital component. A third is a naivete about efficiencies, which assumes that they explain many more anticompetitive practices than they do in fact. Finally, a fourth is residue of a belief, once widespread, that markets tend naturally to self-correct, resulting in a bias against enforcement. These same developments in economics indicate that this proposition is false, and that we have paid a heavy price for it in the form of lower output, unnecessarily high price-cost margins, and reduced innovation.

The federal judiciary exhibits a damaging anti-enforcement bias in some areas of antitrust. For example, the Supreme Court has been unreasonably harsh in defining the burden that plaintiffs must meet in rule of reason cases. A plaintiff’s obligation should be to provide evidence of power and a sufficiently suspicious restraint that it requires an explanation. At that point the burden of proof should shift to the defendant. Important Supreme Court decisions require plaintiffs to prove far too much at the beginning. The result is that the rule of reason has lost much of its usefulness as an enforcement tool. The Court has also been unnecessarily harsh on class actions, and more particularly on enforcement of arbitration agreements. The Federal Arbitration Act was intended to provide a different, less cumbersome forum, not to take away rights that the law grants.

When applying antitrust in digital markets direct measurement of market power, rather than by indirect inference from market share, is often superior. For example, to state that “Google and Facebook control 70% of digital advertising” tells us little unless we know the extent to which digital advertising competes with more traditional advertising media. By contrast, if we include traditional advertising in the market, we treat the two forms as perfect competitors, which understates market power. Market definition is necessarily binary: something is either inside or outside the market. In general, market definitions that include differentiated products will understate power. One unfortunate roadblock to direct measurement is the Supreme Court’s conclusion in its American Express decision that a relevant market must be defined in cases involving vertical relationships. Any statutory reform for dealing with digital commerce should make clear that market power for antitrust purposes should be measured by the best available technique for the situation, and it should avoid the error of continuously expressing market power in terms of a market share of a relevant market.

The large digital platforms do not likely have enough market power in product markets to be deemed “monopolists.” One exception may be Amazon’s large share of the eBook market. There is ample market power, however, to support many allegations of anticompetitive contract practices, including exclusive dealing, MFN clauses, and tying.

Various proposals to break up large digital platforms appear to see size itself as the wrong to be proscribed and offer little assurance that price or output will improve. The opposite is more likely. Setting aside recent mergers, breaking up highly integrated digital platforms will do serious harm to both consumers and investors. While exclusionary contract practices might be unlawful, they are not justifications for structural relief. Rather the remedy should be a prohibitory injunction plus, where appropriate, disgorgement of unlawfully obtained gains. To that end, the Seventh Circuit’s 2020 Credit Bureau decision limiting the equitable power of the FTC to obtain disgorgement is both bad policy and, in any event, inconsistent with Supreme Court precedent.

Antitrust policy should also avoid overly categorical conclusions about the impact of large platform practices on certain groups. For example, while some small businesses who compete with Amazon are undoubtedly harmed, many others are benefitted because Amazon has effectively become their internet broker. They receive access to internet distribution tools including billing and collection that they could not match on their own.

One issue in the digital economy that may require new legislation is firms who renege anticompetitively on FRAND patent licensing commitments. Many substantial networks, including telecommunications, video technologies, and autonomous vehicles, are the product of collaborative innovation. These networks require both technological compatibility and interconnection, which is facilitated by firm commitments to license their patents on fair, reasonable and nondiscriminatory terms. If one firm is able to renege on these commitments with impunity, others will certainly follow. The result will be the unraveling of a system of cooperative innovation that has made significant contributions in an area where economic growth is high. Here, much depends on the outcome of the Qualcomm litigation currently in the Ninth Circuit.

One pressing merger threat is widespread digital platform acquisitions of much smaller firms. Many of these acquisitions do not fit into the framework that the antitrust Agencies apply. First, the acquisitions of competitors typically involve firms that are too small to trigger scrutiny under existing law. Many troublesome acquisitions involve complementary products, such as when a platform acquires a technology that improves its messaging abilities or augments its product line. The principal threat is potential competition: these acquisitions foreclose the possibility that these firms will grow into significant players themselves. Analyzing platform acquisitions will require more categorical treatment of a class of practices, as antitrust currently does with its per se rule. One promising solution would be to prohibit such acquisitions broadly, but permit dominant firms to obtain nonexclusive rights in acquired technology.

Finally, I offer some thoughts about the wisdom of two-agency antitrust enforcement, as well as current policy conflicts between the Antitrust Division and the Federal Trade Commission on digital competition issues.


antitrust, monopoly, digital economy, sherman act, clayton act, FRAND, patents, market power, mergers