The various antitrust complaints the Department of Justice and the Federal Trade Commission have brought against Google, Amazon, and Facebook are based on monopolization claims under Section 2 of the Sherman Act. Herbert Hovenkamp explains why the government should also have relied on Section 1 of the Sherman Act and Section 7 of the Clayton Act to support their Big Tech cases.
The antitrust enforcement agencies—the Department of Justice Antitrust Division and the Federal Trade Commission—have filed several complaints against Alphabet (Google), Amazon, and Meta (Facebook), and may yet file one against Apple. While the facts are diverse, legally they have one thing in common. The Agencies rely almost exclusively on the monopolization provision, Section 2 of the Sherman Act, even though most of the challenged conduct is either an agreement or a merger, which means that it could also be challenged as a restraint of trade under Section 1 of the Sherman Act, or an unlawful acquisition under Section 7 of the Clayton Act. Both of those statutes apply more aggressive conduct standards and they do not require monopoly power or a dangerous probability of creating it. Under either statute it is usually easier for the government to prove a violation, particularly if there is doubt about monopoly power.
Section 1 of the Sherman Act and Section 7 of the Clayton Act Provide Better Avenues for the Government’s Complaints
Section 2’s monopolization provision reaches nearly all types of exclusionary conduct addressed by the other antitrust statutes. This could include practices such as exclusive dealing, tying or other forms of exclusive contracting, or anticompetitive acquisitions. There are some differences when these practices are addressed under the law against monopolization. First, cases brought under Section 2 are typically less fussy about doctrinal details. For example, tying law has developed technical hurdles, such as its requirement of “separate” tying and tied products in order to screen out harmless cases. Monopolization goes straight to the consequences by querying whether the conduct enables a dominant firm to exclude competition unreasonably. Second, Section 2 generally requires exclusion, so practices such as resale price maintenance that raise prices without excluding anyone are hard to fit into its coverage. Third, Section 2 requires monopoly power or the dangerous probability of obtaining it.
By contrast, under the Merger Guidelines, Section 7 of the Clayton Act reaches mergers producing a 30% market share, or even less if the market is sufficiently concentrated. It also reaches both anticompetitive exclusionary mergers and acquisitions that raise prices, but without exclusion. Clayton Act merger law also has a “probabilistic” reach that permits it to go further than the Sherman Act. Finally, Section 1 of the Sherman Act can condemn per se violations with no market power requirement at all, or on much lower market shares for rule of reason offenses. Condemned agreements under Section 1 can include both those that raise prices without excluding (e.g., price fixing) and those that exclude anticompetitively (e.g., tying).
The Federal Rules of Civil Procedure encourage challengers to bring all of their factually related causes of action against one defendant in a single complaint, or else they may be lost. In some cases it may not be too late to amend these complaints, but there is a limit. So why did the FTC not challenge Facebook’s acquisitions of Instagram and WhatsApp under Section 7 of the Clayton Act, and why were Amazon’s various tying arrangements, most-favored-nation clauses (MFNs), or other allegedly exclusionary agreements not challenged under Section 1 of the Sherman Act? Payments for exclusion are classic grist for Sherman Act Section 1’s prohibition of agreements that restrain trade.
By basing claims exclusively on Section 2, the Agencies risk defeat if questions about market definition do not go their way. That is unlikely for Google Search, which has a market share exceeding 90% and for which there is little dispute about market definition. The Agencies could lose, however, in situations such as Facebook and Amazon, where market definitions are more doubtful. For Facebook it depends on questions such as whether X (formerly Twitter), TikTok, YouTube or others are included in the same market. For Amazon it depends on whether the company competes with offline stores (“brick and mortar” businesses), as well as its market shares of individual products. Some of its products, such as streamed digital content, ebooks, and cloud services, exist solely online. By contrast, Amazon sells numerous “tactile” products that compete over a broader range. Small appliances, tools, hardware and much clothing compete closely with offline stores, including Walmart, as well as an array of online stores. For other products and services, such as groceries and try-on clothes, it faces an uphill struggle with offline stores. Amazon also must contend with the fact that customer-switching costs are generally lower on the internet than they are in offline stores. The person who does not see the toaster or sneakers she wants at Walmart must drive to a different store. On the internet she can switch away from Amazon with the click of a mouse. These are all monopoly power issues that this lawsuit will likely address.
On Amazon’s market share, a theory of “cluster” market definition occasionally warrants grouping noncompeting products, although it is not alleged in the FTC’s current complaint. A store is not a cluster market simply because it sells many products. There must be something that holds them together and facilitates the exercise of market power over the aggregation. For example, a hospital can be a cluster market of the noncompeting services it offers because the hospital is necessary to provide them. Facebook is arguably a cluster market because its customers expect to be able to use its services, such as messaging and photo/video sharing, in any combination. Facebook’s services are not as tightly bound together as, say, surgery and anesthesiology in a hospital, but that difference presents a fact question to be resolved. Merger cases such as Staples II recognize a cluster market because the object of the purchase is the entire store, not its individual products. By contrast, Amazon faces varying amounts of competition in its individual products. Customers select what they want, and the average order on Amazon is a little under $50. That does not meet the definition of “cluster market” in the 2023 Merger Guidelines (§4.3.D.4). In sum, the Amazon complaint would likely stand a better chance of success if it were pled under Section 1 of the Sherman Act for the specific products where harm to competition is likely.
Understanding the Motivation for Section 2 Claims
So why are the Agencies emphasizing Section 2 so strongly, while neglecting other, apparently more aggressive and viable antitrust provisions? One possibility is that this is a populist appeal to a public that often thinks these firms are “monopolies” simply because they are big. Another is that the Agencies are seeking structural relief—that is, some kind of breakup. That argument makes no sense for acquisitions. The federal courts clearly have the power to undo unlawful mergers in actions under the Clayton Act, even without monopoly. Indeed, the preferred remedy for Section 7 violations is an undoing of the acquisition, assuming it has been completed. If not, the most common remedy is an injunction preventing it from occurring. Why the FTC did not include a Section 7 count in its complaint is a mystery.
What about Section 1 of the Sherman Act? We may think of Section 2 as a more obvious vehicle for obtaining structural relief. But that argument is weak. First, there is absolutely no warrant for it in the statutes, which do not link the right to structural relief to any particular provision. The principal Justice Department remedy statute empowers the courts to “prevent and restrain violations,” but says nothing about the method and refers to all of the antitrust laws collectively. As a matter of precedent, older decisions used dissolution to remedy unreasonable combinations in restraint of trade. The Northern Securities case involved a corporate merger via a New Jersey holding company, but the court condemned it as both an unlawful combination in restraint of trade and as unlawful monopolization, and ordered dissolution.
The case law in government actions challenging tying and exclusive dealing indicates that the decision to use Section 2 is not driven by a wish to break firms up. For example, in Dentsply the government successfully challenged a dominant firm’s exclusive dealing under Section 2 of the Sherman Act, but the requested relief was an injunction, not a breakup. Microsoft was a challenge to a number of exclusive contract practices under both sections of the Sherman Act. The government won on several of these. It had requested structural relief, which the D.C. Circuit rejected on the existing record. The final decree did not include a breakup. The one distinctive Section 1 issue was tying, in which the D.C. Circuit declined to apply the per se rule.
The rationale for nearly exclusive reliance on Section 2 in these recent government complaints against Big Tech is difficult to comprehend. First, it diminishes the Agencies’ chances of success in any situation involving serious disputes about market definition or market dominance. The better approach is to take the Federal Rule of Civil Procedure seriously and include in a complaint every statutory count that applies to the same facts and has a realistic chance of success. Second, assuming that the intent was to increase the chances that a court would approve a structural remedy, there are not good reasons for linking these prospects to the particular statute selected. The one exception is the merger provision, which makes it unlawful to “acquire,” indicating that undoing of an acquisition is a preferable remedy. One might say the same thing of unlawful “combinations,” indicating that the remedy is to break up the combination. But that arises under Section 1 of the Sherman Act, not Section 2.
Second, Section 2 was designed as a device for targeting monopoly. We trivialize the idea of monopoly when we apply it to situations that do not involve or threaten monopoly. The status of being a monopoly applies to products, not to firms. For example, Microsoft arguably has an antitrust monopoly in its Windows operating system, but not in its Bing search engine. Alphabet almost certainly has a monopoly in Google Search, although not in Gmail or Google Docs. Whether Facebook has a monopoly of anything remains to be seen. The extent of Amazon’s monopoly is much more dubious. The strongest claims are its position in ebooks, where it has a market share of about 67%, assuming that ebooks is a relevant market. Cloud storage is unlikely, given a declining market share of about 30%. At the other end of the spectrum, it struggles in the grocery market with only an approximate 2% market share even including its Whole Food sales, which are mainly offline. For other products monopoly would have to be proven, and there are not very many.
Third, and relatedly, pursuing these firms under Section 2 indiscriminately mixes situations in which monopoly is realistically threatened and those where it is not. Rule of reason practices can harm competition on large market shares but not on small ones. One example is MFNs, under which a supplier might promise Amazon that it will not sell its product to rivals at a lower price or on more favorable terms. MFNs can be anticompetitive when a dominant firm uses them to protect its position in a particular product by imposing higher costs on rivals. When a firm does not have a dominant position in that product, however, the principal function of MFNs is to increase distribution. For example, a retailer with a small position in a product may insist that suppliers promise not to discriminate against it in setting prices and terms. The question of dominance in that case is product specific. Some might object that Amazon sells 12 million products and there is no way to analyze all of them. But that is really the point. MFNs might be anticompetitive for some small number of the products that Amazon sells, and the enforcer should then single these out. If MFNs are analogized to exclusive dealing, that would require a minimum market share of 30-40% in the covered product.
Promiscuous use of monopolization claims against big tech also trivializes and undermines any industrial policy argument favoring growth and development. First is the general problem of targeting Big Tech, among the fastest growing sectors in the economy, for hostile treatment. Good industrial policy should support its winners. Second, although we may be legitimately concerned about monopoly because it results in lower output or restrained innovation, it must be identified correctly.
The Agencies should challenge antitrust violations against all who commit them, including large tech firms. They can engage in anticompetitive acquisitions, price-fixing or other contracts in restraint of trade, or unreasonable exclusionary practices, For example, the Facebook acquisitions of Instagram and WhatsApp are prima facie challengeable, whatever the hurdles the litigation may face after these many years. The Alphabet (Google) payments for default status for Google Search certainly seems worthy of investigation and challenge. Amazon’s MFNs could be challengeable for products where its market share is substantial. There could be others. But “targeting” Big Tech for adverse treatment is counterproductive, particularly when it pushes the boundaries of “monopoly” beyond any reasonable definition.
Articles represent the opinions of their writers, not necessarily those of ProMarket, the University of Chicago, the Booth School of Business, or its faculty.
