The Senate Judiciary Subcommittee on Competition Policy, Antitrust, and Consumer Rights will hold a hearing today on “Reining in Dominant Digital Platforms: Restoring Competition to Our Digital Markets.” This hearing is expected to set the stage for an antitrust discussion inspired by the flawed October 2020 House Antitrust Report, which sought to make U.S. competition policy target innumerable policy and social goals – including many in mutual tension with complicated tradeoffs – rather than continuing to pursue the objective of maximizing consumer welfare. The Senators must remember that targeting objectives other than consumer welfare implies causing a decline in consumer welfare. In the absence of a tradeoff, there would be no need for a new law, and the additional objectives could be achieved with a shift in enforcement practices.
The discussion may include the introduction of bills that would implement such a policy change, possibly modeled on bills from the last Congress like the American Innovation and Choice Online Act (AICOA). Economists have been very critical of bills implementing such a policy change, both due to the fundamental tradeoff against consumer welfare inherent to any such bill, and because the specific bills in question have often been poorly designed. Such bills often appear much more focused on punishing large technology companies and than pursuing any broader benefits for society. Likewise, economic research suggests that the costs of such bills are likely large and directly felt by consumers, while the benefits are likely small and accrue mostly to a handful of competitors. Economists’ recent research and commentary on bills like AICOA illustrates these points well.
University of California, Berkeley economist Carl Shapiro’s recent study questioned the underlying logic behind purported competition bills that mostly target a handful of technology and digital services firms. Shapiro noted that he was “especially concerned that Congress is focused on harm to competitors, not harm to competition.” The difference is key: robust competition from leading firms can harm competitors while benefiting consumers. However, it could be inaccurately regarded as anti-competitive under bills inspired by the October 2020 House Antitrust Report.
Shapiro noted that the October 2020 House Antitrust Report “served as the evidentiary basis for the bills relating to competition and digital platforms that were voted out by the House Judiciary Committee.” Shapiro noted central flaws in Section V.C (pp. 207-277) of the Report, including:
- “The  Report takes a negative view of conduct that generates consumer benefits, evidently because that conduct puts competitive pressure on rivals. I thought we learned 50 years ago that such conduct is pro-competitive. Has that basic lesson been forgotten?”
- “Increases in output driven by marketing innovation and low but above-cost pricing should be seen as pro-competitive. Period. I regret to say that the reasoning used in the  Report in support of its claim that Amazon has engaged in below-cost pricing involves basic errors of economics that would earn a failing grade in my MBA classes.”
- “I am concerned that the  Report takes a negative view of actions by the Big Tech firms that enable them to gain or extend their competitive advantage over their rivals by lowering their costs or expanding their product offerings. Stopping market leaders from better serving customers does not promote competition; it reduces competition and harms consumers.”
- “Well-designed rules can promote competition and protect business users and end users from abusive conduct without hindering the large platforms themselves from competing. Misguided regulatory interventions may do the opposite, harming end users and stifling innovation.”
Shapiro’s critiques are incisive and point to the flawed reasoning behind all of the proposed competition policy changes downstream of the Report, including bills like the AICOA.
Shapiro’s colleague at Berkeley, the economist Richard Gilbert, published a paper noting the flaws in AICOA in particular: “the proposed American Innovation and Choice Online Act (AICOA)  goes too far. Among its other provisions, the Act would make it unlawful to engage in conduct that would ‘unfairly preference the covered platform operator’s own products, services, or lines of business over those of another business user on the covered platform in a manner that would materially harm competition on the covered platform.’ If the AICOA becomes law, antitrust authorities and courts would have to determine enforcement standards for ‘unfairly preference’ and ‘materially harm competition’. Experience with the 1950 Celler-Kefauver Amendment to the Clayton Act raises the possibility that courts would interpret these terms narrowly, at least for the short run, with adverse consequences for consumers and the economy.”
Gilbert expressed concern that the need for courts to interpret such terms could lead to a repeat of the classic Brown Shoe and Von’s Grocery episodes in antitrust enforcement, where behavior that was not anti-competitive was nonetheless halted by the courts:
- “Something similar could happen if the proposed American Innovation and Choice Online Act becomes law. Unfair preferencing and material harm to competition are terms that do not have defined meanings from either antitrust jurisprudence or economics. Courts could point to the congressional debate preceding the proposed Act, which emphasized mostly non-specific concerns about abusive behavior by the major tech platforms, to justify a narrow interpretation that condemns any self-preferencing by a dominant platform.”
- “It would then be unlawful under the Act: (i) for Google to launch only Google maps in response to a query for ‘Italian restaurants near me’ or place any Google service at the top of a search results page unless it is accompanied by all possible rival services; (ii) for Amazon to showcase its branded products or favor third-party products that use its fulfillment service for its buy-box; or (iii) for Apple to supply prominent app search results for its own apps, even if they are rated above competing apps.”
Gilbert warned that bills like AICOA would effectively “endorse the European standard for unfair conduct and material harm to competition. I believe this is the wrong policy, for several reasons.” Gilbert concluded that “antitrust enforcement should not condemn conduct that merely promotes a new product, service, or line of business but instead should condemn conduct only if it prevents consumers from using a competing product without an efficiency justification,” as in the Second Circuit’s “Berkey Photo decision.”
A number of economists and lawyers at a recent Technology Policy Institute panel identified additional problems with such bills. Michael Katz pointed out that AICOA and related bills singled out a small number of firms for coverage, which made little sense as a general principle: “If the conduct’s bad, how come it’s only bad when it’s done by these firms? What’s the principle that says the same conduct by a smaller firm doesn’t count?” Katz’s critique implied that AICOA would constrain targeted firms from competing in markets where they were not dominant or even currently present, which could reduce competition: “Particularly when we’re talking about [targeting firms based on] total size of the firm without any reference to markets. […] I think that from the perspective of economics makes no sense.” Katz likewise observed that bills like AICOA seemed both scattershot in terms of objectives and poorly designed to actually achieve those objectives, noting “it’s not clear that antitrust is the right tool for a bunch of ills that people diagnose.”
Former FTC Acting Chairwoman Maureen Ohlhausen expanded on Katz’s points, stating that “AICOA would have prohibited conduct in markets in which there was no allegation that the platform was dominant” once a platform was designated. Ohlhausen observed that AICOA “was an attempt to not name certain companies but only have it be aimed at certain companies” and that this was reflected in the bill design. Ohlhausen also noted that by trying to target too many objectives, AICOA had potentially undermined some of those objectives, as in the example of forcing covered platforms to share access and data with potentially hostile foreign state-owned actors:
- “Some of the [AICOA] provisions were very much in tension with some of the other concerns that were being raised, particularly about privacy […] and data security and security of platforms.” Platforms “having the obligation to […] share access to your platform with all comers […] and the burden was on the platform if they tried to keep someone out for security reasons,” including “foreign state-owned enterprises.”
University of Chicago Professor Randy Picker likewise expressed concern with “the actual terms of the legislation.” With respect to AICOA, Picker noted that “there are things at risk through the rules that consumers have told us through their choices they value.” Picker also observed that there were “no hearings on whether the substance of this [AICOA] legislation would have addressed the problems that I think they see.” In other words, AICOA had many objectives, and did not clearly advance them.
Both Katz and Picker connected the poor design of AICOA at least in part to the attempt to target just a few firms up front. Katz pointed out that the coalition behind AICOA included “strange bedfellows” with questionable motives, and stated that “firm specific legislation I think almost inevitably leads to bad legislation.” Picker echoed Katz, observing that “you change the political dynamics of this type of legislation when you narrow the focus to a handful of firms.”
Recent economic research is also informative. The work of Riitta Katila and Sruthi Thatchenkery, including their HBR article, highlights how antitrust actions against tech firms and digital platforms can harm third party app developers and other “complementors” who build complementary services that work with the platforms. Using the example of prior antitrust actions against digital platforms, they found that platforms eliminated assets used by third parties in order to comply with antitrust settlements with behavioral remedies. Complementary services invested significantly in internal innovation to replace those assets, but they found that the result was rarely commercialized successfully, and some complementary services were harmed by the removal of the platform’s assets from the market. “By eliminating some of [a platform’s] complementary assets offerings, the [antitrust] settlement we studied may have inadvertently eliminated rival complementors’ paths to commercialization.”
Katila and Thatchenkery’s work implies that bills like AICOA imposing significant behavioral restrictions on digital platforms could harm many third parties whose commercialization benefits from assets or service offerings that could be curtailed by such remedies.
Research directly analyzing bills like AICOA has echoed the implications of Katila and Thatchenkery’s work. The “Tools to Compete” report by the CCIA Research Center and Engine found that bills like AICOA could impact large digital services’ ability to offer many valued services currently used by startups for free or at low cost. If startups had to replace those free or low-cost services, or find next-best alternatives, they would have to spend about $3,000 per employee per year on such replacements, which could amount to up to 8% of a small, established startup’s budget. This would increase barriers to entry, as startups have limited funds and take significant time to develop a positive free cash flow.
Additional research shows how bills like AICOA could dramatically reduce efficiencies and increase costs for targeted firms, with implications for their business users and end users. For example, one study found that bills resulting in structural separation for targeted digital platforms would impose nearly $320 billion in costs and lost efficiencies on those platforms, with significant pass-through to business users and end users. Another study found that bills like AICOA would impact firms far beyond the digital platforms by the 2030s, causing over $1 trillion in added costs and lost efficiencies throughout the U.S. economy, which would cost 27.9 million government employees including teachers, ﬁre ﬁghters, and nurses almost $4,000 per person in lost retirement benefits through their pension plans.
In short, competition bills like AICOA that target objectives other than maximizing consumer welfare are likely to impose significant costs throughout the economy with few benefits to offset those costs. Senators should keep in mind these costs as they consider endorsing competition policies that would reduce consumer welfare.