(Ed. Note: The “Innovation Fact of the Week” appears as a regular feature in each edition of ITIF’s weekly email newsletter. Sign up today.)
The House Judiciary Subcommittee on Antitrust, Commercial and Administrative Law will kick off a series of hearings on Thursday, February 25, focusing on large Internet platforms. The theme will be “reviving competition,” and Subcommittee Chairman David Cicilline (D-RI) will dedicate the first hearing to “proposals to address gatekeeper power and lower barriers to entry online.” This series follows the chairman’s 2020 committee report and Sen. Amy Klobuchar’s (D-MN) new bill on antitrust reforms. These endeavors are premised on a core foundational goal: More competition is needed to help other businesses, particularly small businesses. It is surprising that many Democrats now want to side with business, including small businesses, when historically what distinguished the Democratic Party from the Republican Party was its focus on average Americans, not business per se. To maintain that noble tradition, the subcommittee should focus first and foremost on ensuring that antitrust fosters consumer welfare, and they should evaluate large technology platforms on a single question: Is what they are doing good for average Americans, while not engaging in clearly anticompetitive conduct?
The chairman appears to want to focus on two potential paths for reform:
- Banning so-called “self-preferencing” by large online platforms; and
- Creating a sort of “Glass-Steagall” Act in antitrust—namely, separating platforms from any commerce offerings.
While both proposals might spur more competition and help other businesses, they would make consumers’ lives worse by raising costs, reducing convenience, stifling innovation, and even distorting competition they aim to revive.
“Self-Preferencing” Is Consumer Preferencing
Chairman Cicilline’s subcommittee report from last fall recommended implementing “rules to prevent discrimination, favoritism, and self-preferencing.” There are two issues here.
The first is whether platforms unfairly discriminate against competitors’ offerings to favor their own. For example, does a search engine demote organic search results related to a competitor’s offering? Or does an ecommerce site intentionally hide a competitors’ product in order to favor its own. The answer is that there is absolutely no evidence that platforms have done this, for the simple reason that doing so would reduce the value of its service to consumers.
But there is a second kind of self-preferencing that platforms engage in and that is not anticompetitive and is pro-consumer. In search for example, this can involve the search company putting its own offering at the top of the page (clearly labeled as such), while not changing the results in organic search. In ecommerce it can involve the company offering its own product, highlighted as recommended.
These platforms—and by the way, virtually every major goods producer in the economy from car companies to grocery stores—do this of course to earn money, but that depends on making consumers’ lives better. Be they their own products (e.g., Amazon), apps in app stores (e.g., Apple), or digital services (e.g., Google), it appears that Chairman Cicilline wants to make self-preferencing illegal but only for a handful of online platforms—commonly and fuzzily referred to as “digital gatekeepers.” The arguments against self-preferencing are misguided for a number of reasons.
First, the negatively connoted expression “self-preferencing” conveys an imprudent idea that the blamed preferencing take place at the expense of consumers. But most self-preferencing involves the platform’s private labels (whether Google Shopping, for example, or Amazon batteries), which are done to offer more competitive offerings for consumers. A case in point here is Amazon’s private label products, which in many or even most cases, are cheaper than branded products. Consumer preferences trade off brands’ reputation and prestige with lower prices of private labels. Thus, algorithm-driven self-preferencing may reveal consumer preferences toward less premium, more competitive products than traditional brands. Consequently, repeated consumer choices may lead to some private labels being ranked higher than traditional brands.
Unquestionably, however, the resulting search results’ ranking reveals consumer preferences rather than the platform self-preferencing. Indeed, to conclude that the platform is capable of extracting monopoly rents in their search results at the expense of consumers is tantamount to making the strong assumption according to which the platform is subject to no platform competition. Both third-party sellers and end-consumers can “vote with their feet” and switch to alternative platforms should the rankings adversely and unfairly affect them. In technical terms, to assert that self-preferencing illustrates the monopoly power by platforms exaggerates the process of downstream competition and ignores the competitive constraints exerted on the upstream competition. In other words, self-preferencing is the result of algorithm-inferred consumer patterns on the downstream level—where retailers and digital services providers compete to reach end consumers.
Moreover, platform competition disciplines upstream competition due to low switching costs for consumers: the competitive environment of the upstream level constrains the platforms’ ability to ignore consumer preferences. One cannot objectively appraise downstream competition without a proper assessment of the upstream competition. For instance, Amazon cannot profitably and soundly ignore consumer preferences in its product search results—i.e., the normal process of downstream competition—otherwise, consumers will switch to platform rivals such as Google Shopping, Walmart, Etsy, Alibaba, eBay, and many other competitors. Unfortunately, such ignorance of the competition process—both upstream and downstream—was famously made in the 2017 Google Shopping decision by the European Commission. It was then considered that Amazon was not a competitor to Google Shopping, thereby ignoring the upstream platform competition. Equally unfortunately, a similar mistake would be made should the current Hearings confirm the ban on self-preferencing.
Second, the ban of self-preferencing for digital platforms is tantamount to an implicit acknowledgment by elected officials to prevent consumers from the chance of getting cheaper, higher quality or more convenient offerings. The wider offering of products for the benefits of consumers follows identified consumer demands for more integrated and more competitive products and services. Indeed, the platform’s willingness to enter a downstream market follows a sensing of opportunities. In other words, it is the platform’s capabilities and the downstream retailer’s relative inefficiency which generate an opportunity for the platform to enter the downstream line of business at the benefits of consumers and the expense of the less inefficient retailers.
The key point is that practices should only be deemed anticompetitive when as-efficient rivals are artificially excluded from the downstream market by the upstream company. And it must remain the case: as-efficient rivals, and only them, unfairly demoted by the platform can make legitimate claims of anticompetitive behaviors from the platform. Nevertheless, it is yet to be evidenced that large digital platforms do not exhibit superior efficiencies or innovativeness than smaller downstream rivals: Smaller rivals can hardly be considered efficient rivals with large platforms. Accordingly, consumers benefit from the platform’s superior efficiencies and increased convenience (and often lower prices).
Consequently, to ban self-preferencing would be tantamount to prevent platforms from exploiting sources of potential efficiencies and to legitimately expand its offerings. Both commercial practices may generate consumer benefits and deliver innovation benefits for society. For instance, if Amazon succeeds in the pharmacy line of business, it can only be because it is capable of providing lower prices and/or better (delivery) services. In other words, if it succeeds, it is because consumers have enabled it to succeed. Otherwise, the entry would be unprofitable, and entry would accordingly not take place. Thus, to ban self-preferencing would not only suggest that any leveraging strategy become de facto illegal, but it would more severely imply foregone benefits for consumers. The platform and the consumers would unquestionably be losers—the only winners may end up being less efficient rivals who benefit from a less dynamic, less agile retail market.
Third, to ban self-preferencing for a handful of digital platforms creates unfair competition under the veil of spurring “fair competition.” The envisaged ban of self-preferencing would apply to a discretionarily selected online platform group—the regrettably labeled “digital gatekeepers” inspired by the European Digital Markets Act. The creation of such an uneven level playing field bears no economic and legal justification, but merely political gains to further the popular techlash. There is indeed no economic justification to deteriorate the market position of digital gatekeepers while preserving their rivals’ market position. There is no such thing as digital gatekeepers or digital markets—there are digital channels amongst many other distribution channels. Again, Amazon’s rivals are both online and offline companies, and its market share reveal such competitive constraints. Digital advertising is one of the many channels for advertisers to reach viewers. In other words, the creation of an uneven-level playing field at the expense of the most digitalized companies and the benefit of the least digitally capable rivals runs counter the pursuit of technological innovation at a time of the Covid-19 pandemic where only wide adoption of digital tools prove to be a way out of the crisis.
Furthermore, the discrimination between the digital and non-digital sectors is unfounded. The subtle distinction between large platforms (i.e., gatekeepers) and those excluded from this regulatory club is equally questionable. The hearings have notoriously excluded companies such as Microsoft, Shopify, Bookings.com, Netflix, Disney, etc. Does it mean that these companies are not large enough to be subject to potential bans on self-preferencing, thereby generating regulation-embedded unfair competition? If this is the case, the U.S. antitrust reforms may prove to embrace the narrow European definition of digital gatekeepers aimed at targeting especially American and Chinese tech companies. Ironically, the U.S. Congress may itself target American tech companies by drawing explicit inspirations from the European digital protectionist regulations. The ban on self-preferencing is thus unmanageable: it is either applied too broadly and sanctions a ubiquitous business practice, or it is applied narrowly, and it unfairly penalizes digital innovators at the benefit of less efficient rivals and the benefit of foreign powers. Should the hearings lead to the conclusion that a ban on self-preferencing applied to a handful of tech companies is a sound and legitimate legislative proposal, consumers and innovation will eventually be harmed for the reasons outlined. Also, the competition will be more unfair.
When Forced Separation Is Hampered Competition
Another direction the hearings are likely to consider is the long-time push for a separation of platform and commerce—in other words, the break-up of companies by structural remedies. The idea of such an extraordinarily unreasonable and disproportionate remedy springs up from two rationales: First, we should apply a sort of “Glass-Steagall Act” to platforms, and second, we should apply a sort of public utilities (or “platform utilities”) regulation to large tech companies. Both arguments are equally flawed.
Extrapolating the “Glass Steagall Act” (which separated commercial banking from investment banking) to digital platforms proceeds from a fundamental misconception of both the Act itself and the digital platforms. First, it is a misconception of the Glass Steagall Act that a similar regulation applies to tech platforms: Systemic risks generated by large financial institutions underpin the rationale for the Act. The interconnectedness of the financial institutions urged regulators to ensure that the whole financial system was less prone to crisis. No such interconnectedness exists in the digital sector—instead, it is a relationship of aggressive rivalry. Should Google suddenly disappear for some reason, rivals may immediately step in and replace Google the minute it disappeared. This is precisely what would happen if Google News’ withdrew from Australia: Microsoft’ Bing would enter and gain market share. Should Amazon become unable to supply products or provide cloud services, immediately grocery stores and other cloud competitors would substitute the tech platform. Should Facebook disappear, tens of social media platforms would immediately replace it. Thus, it is fundamentally erroneous to invoke the Glass Steagall Act, which is justified based on systemic financial risks to apply a similar regulation in an area where there is no such a thing as systemic risks, but instead, evidence of significant substitutabilities out of aggressive rivalry.
Second, it is a misconception of the Glass Steagall Act that such regulation may justifiably resonate in the digital industry. Under the motto “too big to fail,” financial institutions were separated to preserve their viability. As the motto suggests, the companies became too big to fail. On the contrary, the successful tech giants are not failing, but the separation would make them fail. A Glass Steagall Act for some tech companies will undeniably lead to structural break-ups, which may lead tech companies to be “too small to succeed.” For, the network effects and the necessary scale inherent to digital industries are not shared with the financial markets. A small bank can provide a competitive mortgage. Can a small search engine provide accurate information, or a small e-commerce platform provides competitive prices without sufficient scale? Obviously not. To transpose the Glass Steagall Act’s rationale to the digital sectors of the economy overlooks the fundamental legal and economic differences that put the financial system and the digital ecosystems apart. Regrettably, this is an error the hearings may be conducive to, should the influential Neo-Brandeisians are undeservingly granted blank checks on antitrust reforms as it seems to be the case.
To advocate the separation of the platform from commerce follows an influential article from an influential author for an influentially misconceived idea. U.S. House Judiciary’s Subcommittee’s Counsel Lina Khan advocated at the Open Markets Institute about the separation of platforms and commerce before co-authoring the House report. The markets which are referred to justify a separation regime in the digital industry are entirely irrelevant for this industry: they are either markets under previous state monopolies (e.g., TV networks, broadband spectrum telecommunications) or are susceptible to be “natural monopolies” (e.g., railroads). None of these categories apply to the newly created digital sector, which epitomizes healthy innovation patterns and aggressive rivalry. Consequently, the analogy for separation of platform and commerce to digital industries as a natural expansion of such separation, say, of public infrastructures such as railroads or airports, results from fundamental pitfalls in the analysis. It is commonsensical to have a unique railroad between New York and Philadelphia, as long as the carriers compete with one another—hence the separation can be justified. But what is the justification for a unique digital platform if other platforms are just “one click away” with obviously no physical jam and no infrastructural overlap? “Comparison is no reason enough,” as a French saying would suggest—here, to compare digital sectors of the economy with natural monopolies regulations appears unwise.
Finally, the arguments of the separation of platform and commerce implicitly infer that a sort of “public utilities”—here, platform utilities—regulation. Assuming that some platforms have become unavoidable and thus should be subject to the essential facilities doctrine, the advocates of public utility regulations to platforms propose common carrier antitrust rules at the expense of both competition and innovation. To apply essential facilities doctrine not only suggest that some assets such as data are “indispensable.” In contrast, they are non-rivalrous and amassed by competitors. Still, it would also suggest that a duty to deal can legitimately be imposed without due respect to the proprietary (sometimes patented) services and products the platforms may provide to consumers. For instance, to wrongly treat Google Android OS as an essential facility to manufacturers would ignore the fact that other Android OSs have emerged aside Apple’s iOS, thereby not rendering this operating system “indispensable”. It would also ignore the fact that Google has opened its proprietary assets to third-party developers as opposed to Apple’s closed iOS. To treat Google Android OS would not only harm unduly Google’s business model of OS compared to Apple’s, but it would also penalize the very openness is intended to create. More generally, applying public utility regulation to the platform ironically may deplete competition instead of fostering it. Indeed, per definition, the utility once regulated as “public” would enjoy a de facto monopoly since it would treat downstream players under the guidance of the regulator’s subjective notion of “neutrality.” Consequently, alternative platforms would face increased entry barriers. Akin to the GDPR’s effects of having entrenched market positions of large platforms at the expense of small companies, public utility regulations may deplete upstream (platform) competition at the expense of innovation dynamism and the expense of consumer welfare.
In short, the best conceivable way to entrench large platforms’ market power would be to treat them under public utility regulations so that government-enforced “search neutrality” will deter rivals from inventing rival platforms and may deter consumers from disciplining the platforms by “clicking away.” In conclusion, forced separation of platform and commerce proceeds from flawed analogies. It would hamper competition rather than boost it. It would deter innovation rather than turbo-starting it, and it would fundamentally harm consumers at a time they deserve stronger protection. The hearings are thus advised to be mindful of these risks.