The four major internet platforms – Google, Amazon, Apple, and Facebook – have in recent months been subjected to increasing attacks on antitrust grounds. These attacks have been generated by various entities in the federal legislature and by the threat of individual lawsuits, some by one of the platforms against another. The subject of these attacks is the sheer size of the platforms themselves, though it is also alleged that various of the platforms engage in unfair practices that benefit their own products over those of competitors. This essay argues that these antitrust attacks ignore the character of the platforms as network industries. Network industries are different from well-known monopolies because consumers benefit, rather than are harmed, as the network expands. As a consequence, none of the attacks can demonstrate harm to consumers, the overriding standard of interpretation of the antitrust laws.
By George L. Priest1
There is a current effort in the U.S. and abroad to employ the antitrust laws to attack Big Tech firms. “Big Tech” refers to the current four major internet platforms — Google, Amazon, Apple and Facebook. These firms are variously accused of being monopolies and that on that ground should be broken up; of being monopolies illegally acquired by the acquisition of smaller competitors with the demand of divestiture of the previously acquired entities; or of engaging in practices, in particular the preference of their own or related companies through network algorithms, to the harm of competing producers.
The attack is wide-ranging. Late last year, the U.S. House of Representatives Judiciary Committee released an exhaustive report arguing that all four internet platforms should be broken up. Shortly thereafter, the U.S. Justice Department and 11 states filed an antitrust claim against Google alleging attempted monopolization. Various other states soon followed claiming that Google illegally favored its own search engines. Very recently, politicians such as Senator Amy Klobuchar have urged the new Biden Administration to break up the platforms. And even more recently, one of the platforms itself, Facebook, has been rumored to be about to file an antitrust suit against Apple, claiming that Apple gives preferential treatment in its App Stores to its own products imposing restrictions on third-party developers, including Facebook. The European Union has taken a similar approach against some of the platforms though chiefly through the imposition of enormous fines.
Although there are other criticisms of Big Tech platforms — relating to their management of user privacy and to the spread of disinformation, which are not related to antitrust questions of industrial organization — the multiple antitrust claims have an unusual character. These lawsuits do not principally derive from claims nor are based on evidence of harm to consumers. Most consumers know of the extraordinary advantages of search on Google, of shopping on Amazon, of the application developments of Apple, and of connection through Facebook. In contrast, the antitrust claims are based most simply on a distrust of the size of these firms, each of which has grown in different ways to become platforms for internet interaction as well as from a distrust of the practices in which these platforms engage taking advantage of that size.
More precisely, from an antitrust standpoint, there have been two principal claims against (or available against, since not all have faced these claims) the four platforms. The first is that the platforms have gained their dominant impact through illegal acquisitions of nascent competitors or illegal agreements with potential competitors (such as Google’s agreement with Apple) to achieve their controlling positions. This is a standard antitrust claim, invoked over the history of the antitrust laws in the U.S., most prominently against Standard Oil, American Tobacco, U.S. Steel and others, leading to the development of tight standards for the acquisition of competing companies.
There is an important difference, however, between the mergers to monopoly of the late 19th and early 20th Centuries — Standard Oil, American Tobacco, U.S. Steel — and of the creation of platforms in the modern era. Standard Oil achieved its near-monopoly over oil refining prior to the Sherman Act by acquiring — usually through profit-sharing agreements — previously competing independent refiners, first in Cleveland, creating a monopoly in that important oil refining center, then expanding it through acquisitions in Pittsburgh and the East Coast. The creation of American Tobacco was similar though more simple: a merger of previously competing tobacco manufacturers. Andrew Carnegie and J.P. Morgan created the U.S. Steel Corporation by the merger of independent steel manufacturers who had previously competed for the sale of steel.
The creation of the dominant four Big Tech platforms has been distinctly different. Their growth has been largely entirely internal or, where accelerated by acquisition, not by the acquisition of serious competitors of the growing platform. Importantly, none of the acquisitions often alleged in these complaints was challenged, when they occurred, by the Justice Department or the FTC as anticompetitive because it was recognized that these acquisitions and agreements served to expand the platform. Again, the Standard Oil, American Tobacco and U.S. Steel industries were not internet platforms, building on the gain from network benefits, but simply the agglomeration of previously independent competing companies.
The second principal antitrust claim is that these platforms have adopted algorithms that give preference — not exclusive preference, but on-screen (such as first-page or top-of-the-list) preference — to their own or related products. This discrimination claim is quite different from the other claims of monopolization. There were no general claims of discrimination in sales against Standard Oil, American Tobacco, or U.S. Steel. There have been more subtle claims in the history of the antitrust laws of differential sales agreements, such as tying arrangements, in conditions of monopoly or semi-monopoly, such as through the possession of patents, but these claims are now largely discredited because of the lack of evidence (or even theory) of consumer harm.
The first principal argument of these antitrust claims — that the size of these platforms is itself grounds for breakup — is deeply problematic. Unlike traditional monopolies such as Standard Oil, American Tobacco and U.S. Steel, the four Big Tech platforms are network industries where the benefits to consumers from participation in the network grows as the network expands. In the American Tobacco and U.S. Steel cases, consumers did not benefit as previously competing tobacco and steel manufacturers merged together eliminating competition and increasing prices. The Standard Oil experience is somewhat different because the principal motivation for Rockefeller’s initial merger to monopoly of all Cleveland refiners and his later expansion to include refineries in Pittsburgh and the East Coast was to create monopsony power against the railroads — the New York Central, the Erie and the Pennsylvania — whose charges for the transport of refined oil from the western Cleveland and Pittsburgh areas to the east were a major element (estimated at 47 percent) of the cost of shipment from the U.S. to Europe, which was then the largest market in the world for refined oil. The Standard Oil monopsony of railroads may have reduced prices to consumers, though not because of network benefits and subsequently offset by Standard Oil’s monopoly over refined oil sales.
Quite in contrast, in creating networks, the four Big Tech platforms have increased consumer benefits by expanding their size. Consumers benefit when Google expands its search algorithms and when Amazon expands the retailers made available on its networks.
As a consequence, there can be no confident expectation nor surely proof that there will be any consumer benefit or societal benefit from breaking up these platforms. There is, perhaps, some point, in theory, at which the network benefits from adding additional parties to a network begin to diminish. But no one knows what that size is or when these benefits begin to diminish, and no honest expert could testify to that fact. Even in contexts where greater network size was achieved by Big Tech acquisitions or agreements, such as the agreement between Google and Apple, this fundamental issue remains. One of the states’ complaints against Google is that the agreement with Apple resembles a typical merger to monopoly, though neither the Justice Department nor FTC has challenged it. But this claim ignores the network character of the industry where enhanced size provides clear benefits to all participants.
The second principal argument in the antitrust attack on Big Tech is that these platforms discriminate in favor of their own or related products against producers carried on their networks who compete with those products. Thus, Google’s search algorithms give preference to Google-related products over others; Amazon’s product listings give preference to Amazon products over others; the Apple App Store gives preference to Apple applications, charging a tax to applications of other producers. This alleged discrimination is often subtle. It is not that the platforms refuse to carry or advertise competing products, but that they give competing products less prominence, such as lower placement in product queues.
This subtlety itself is revealing. Why do Google, Amazon or Apple even provide access or carry competing products? Presumably, because the network benefits they provide are greater if consumers can access these competitors, even at a slightly higher cost of search to move down the list of competitors.
Preferences for a monopolist’s products have been adjudicated elsewhere in antitrust law, for example, with respect to the practice of tying arrangements. For many years, tying arrangements were prohibited by both the Sherman and Clayton Acts, though without an underlying theory as to how such practices harmed consumers. Since the late 1970s, with the change in the approach in the Supreme Court to U.S. antitrust law, the prohibition of tying arrangements has been largely relaxed requiring a showing of clear consumer harm in order to find a violation.
The question with respect to the practices of these dominant internet platforms is why should there be a prohibition of the preferences given by the platforms? Why shouldn’t these preferences — modest though they are and, as mentioned, constrained by consumer demand for choice to achieve maximum value from interconnectivity — be regarded as an appropriate return from the creation and maintenance of the platform, indeed an expected return to provide incentives for the generation of future platforms?
In older antitrust law, preferences such as tying arrangements were viewed as improper because the market power possessed by the firm with the tying product — such as from patents — was created by the government through the grant of patent authority. The Big Tech platforms are different. They have not been created by government patents but have grown basically internally from the successful creation of network benefits. Why shouldn’t the gain from these (modest) preferences be viewed as an appropriate return from the creation of the network?
To prohibit preferences of this nature and to require non-discrimination by these platforms is to convert these platforms into public utilities, such as the provision of water supply, gas and electricity, all bound to principles of regulated non-discrimination. But, here again, the Big Tech firms are different. Public utilities such as water, gas and electricity required typically local government support in the granting of rights-of-way along public streets for the delivery of their services.
The Big Tech platforms did not develop their networks through grants of public rights-of-way. Instead, they have grown as the internet has grown and as the benefit of network industries have grown.
The antitrust attack on Big Tech platforms, as a consequence, has no coherent basis. The criticisms of the size of these platforms ignores the great network benefits that they create and that they might enhance by growth into the future. The claim of discrimination ignores its modest character — constrained by consumer demand for greater internet connectivity.
The issue in these and, presumably, forthcoming lawsuits should not derive from the size of these platforms, but should be whether consumers will or will not benefit from any purported remedy.
1 Edward J. Phelps Professor of Law and Economics, Yale Law School
