Simply adding competition in the tech sector won’t solve problems like privacy abuses or discrimination. Competition is needed, but regulation is a necessary element of any tech solution.
The initial press coverage of Makan Delrahim’s speech at Tel Aviv University last week, in which he declared that antitrust, and antitrust alone, could solve all problems relating to Big Tech—from privacy abuses to censorship (or what he incorrectly called “free speech”)—was to be expected. With notable exceptions in Bloomberg and The Economist, the media largely lapped it up.
Call me a skeptic. Delrahim has changed his mind before, like that time he flipped on AT&T-Time Warner, initially asserting on BNN Bloomberg that the merger didn’t present any “antitrust problem.” His prior speeches on Big Tech have been fairly pro-platform. Maybe he’s come around on the tech platforms too.
There is little doubt that antitrust can be effective in policing abuses by dominant firms outside the firm’s boundaries, such as exclusive contracts with distributors or equipment makers. Indeed, antitrust can succeed there even in the absence of direct evidence of anticompetitive effects, as the recent FTC v. Qualcomm decision showed. It is no accident that the Qualcomm opinion is devoid of citations to the FTC’s or Qualcomm’s economic experts—a large part of what economic experts do is causally connect the challenged conduct to tangible harms (e.g., price, quantity, or quality effects), often via econometric models. In Qualcomm and in Microsoft, indirect proof of harms, based on the nature of the conduct and the extent of the market foreclosed from rivals, was sufficient for the government to prevail on the merits.
We don’t know yet whether Delrahim (a former Google lobbyist) believes that antitrust is also the right tool for policing conduct by platforms inside their boundaries—that is, conduct that does not require the participation of an input provider or customer. For example, a technological tie-in that bolts a web browser with an operating system occurs strictly inside the firm. Similarly, a platform’s search algorithm that affords extra weight for affiliated properties or merchandise (“self-preferencing”) occurs inside the firm’s boundaries. Amazon and Google, in particular, are appropriating content at the edge of their platforms and then using their platform power to steer users to the affiliated clones. (Facebook appropriates app functionality, but because it does not host independent edge providers on its platform, Facebook cannot be said to discriminate).
Delrahim was careful, presumably purposely so, not to mention any specific company, nor any specific conduct, when claiming that antitrust can rein in Big Tech. He did make clear, however, where he stands on regulation as a companion to antitrust enforcement: he doesn’t like it. This phrase from his Tel Aviv speech is noteworthy: “To quote Orrin Hatch, the legendary former Chairman of the US Senate Judiciary Committee, ‘Vigilant and effective antitrust enforcement today is preferable to the heavy hand of government regulation of the Internet tomorrow.’” The message was clear: “Back off, would-be regulators, I’ve got this.”
Delrahim’s condemnation of regulation stands in contrast to a growing number of influential voices, including prominent antitrust practitioners, who not only want to steer antitrust in a very different direction but also want a supplementary or reinforcing role for regulation. For example, Senator Elizabeth Warren’s tech-platform plan involves, in addition to divestiture (break up), nondiscrimination protections and a line-of-business restriction (barring a platform from offering edge services), both of which are regulatory in nature. Open Markets scholars Matt Stoller and Sally Hubbard are quick to note that regulation is a necessary element of any tech solution. After all, we solved competition problems in telephony markets using both.
Two prominent antitrust scholars and practitioners, Professors Fiona Scott Morton and Carl Shapiro, recently released separate reports explaining how certain practices of the tech platforms escape traditional antitrust scrutiny, thus implying the need for regulation. Dr. Shapiro explains that “The second area where antitrust enforcement has become inadequate is the treatment of exclusionary conduct by dominant firms. The fundamental problem in this area is that the Supreme Court has, over the past 40 years, dramatically narrowed the reach of the Sherman Act.” He also states that bringing a Section 2 case against a tech platform would be “difficult,” and that pursuing Amazon under the antitrust laws for discriminating in favor of its own merchandise would be “very difficult.” While Dr. Shapiro does not call for regulation explicitly, Dr. Scott Morton and her co-authors note that “a sectoral regulator is likely to be better than antitrust laws at enforcing fairness norms.”
Delrahim is having none of this. In his Tel Aviv speech, Delrahim offers the DC Circuit’s decision in Microsoft as a template for how to tame today’s dominant tech platforms using antitrust (and only antitrust). But the Microsoft court held that structural remedies were not appropriate unless anticompetitive effects could be demonstrated with confidence:
Microsoft’s concerns over causation have more purchase in connection with the appropriate remedy issue, i.e., whether the court should impose a structural remedy or merely enjoin the offensive conduct at issue. As we point out later in this opinion, divestiture is a remedy that is imposed only with great caution, in part because its long-term efficacy is rarely certain. … Absent some measure of confidence that there has been an actual loss to competition that needs to be restored, wisdom counsels against adopting radical structural relief.
Because the primary form of anticompetitive injury in Microsoft and any potential case against a modern tech platform would take the form of hard-to-measure innovation harms, securing a structural remedy via antitrust under current law would be challenging. It is not clear how to estimate a future loss in consumer choice due to exit by independents with any “measure of confidence.”
Even with respect to (non-structural) injunctive relief, the Microsoft court was unwilling to unwind the bundling of Internet Explorer and the operating system on the flimsiest of efficiency defenses, because such conduct occurred inside the firm boundaries and thus was sacrosanct:
As for the other challenged act that Microsoft took in integrating IE into Windows—causing Windows to override the user’s choice of a default browser in certain circumstances—Microsoft argues that it has “valid technical reasons.” Specifically, Microsoft claims that it was necessary to design Windows to override the user’s preferences when he or she invokes one of “a few” out “of the nearly 30 means of accessing the Internet.” … The plaintiff bears the burden not only of rebutting a proffered justification but also of demonstrating that the anticompetitive effect of the challenged action outweighs it. In the District Court, plaintiffs appear to have done neither, let alone both; in any event, upon appeal, plaintiffs offer no rebuttal whatsoever. Accordingly, Microsoft may not be held liable for this aspect of its product design.
Because Google’s and Amazon’s self-preferencing are analogous to Microsoft’s internal preferencing of its web browser over rivals, a plaintiff in a case against the modern tech platforms likely would encounter similar obstacles of (1) measuring the anticompetitive effects and (2) balancing those effects against any purported efficiencies.
Notwithstanding these legal impediments to a pure antitrust solution to Big Tech, Delrahim suggests in his speech that antitrust enforcement, presumably leading to more competition, is also a solution to current concerns that the tech platforms have eroded privacy: “In addition, diminished quality is also a type of harm to competition. As an example, privacy can be an important dimension of quality. By protecting competition, we can have an impact on privacy and data protection.”
I have my doubts about this, too. Suppose we somehow get to two Facebooks, via either legislative fiat, a structural antitrust remedy, or a policy intervention that breathed life into a social media entrant (such as data portability/interoperability). Would the Baby Facebooks be any less inclined to exploit user data and infringe on privacy relative to the mothership?
To answer this question, keep in mind that the very essence of Facebook’s business model is the exploitation (and monetization) of user data. Adding a horizontal rival won’t change how money is made in social media.
This is not all. It may not be in the second Facebook’s interests to hold itself out as the privacy savior. One strand of economic literature suggests that entrants often mimic the exploitative practices of incumbents because it is more profitable to do so. And a 2019 study from the Düsseldorf Institute for Competition Economics finds that “larger firms offer more privacy on their websites than smaller ones.” Why would an entrant want to lure away Facebook’s most privacy-sensitive customers, who are by selection the least attractive to advertisers? (It is possible that social media entrants had to give up competing on privacy because Facebook has forced other sites to copy Facebook’s low privacy terms, but that just presents another entry barrier.)
The best argument I’ve come across in support of the competition-solves-privacy thesis is the one outlined by Dina Srinivasan, who notes that “privacy was once a crucial form of competition” among social media platforms in their nascent, unconcentrated state. She points out that in 2007, when it lacked market power, Facebook introduced a program that recorded users’ activity on third-party sites (and inserted it into their News Feed) but was forced to retract it in response to public outrage. It’s quite possible that privacy would become a dimension of competition in a world with several small to medium-sized social networks. But the question is whether we can ever get back to that market structure, particularly given the power of network effects, data advantages, and users’ tendency to single-home. And if we can’t replicate a competitive market, will competition at the margins do anything to constrain Facebook’s exploitative ways?
Don’t get me wrong: More competition is a good thing, and I would cheer an antitrust suit—even one that pursued a novel theory of harm—against a platform. When it comes to policing privacy abuses, however, there is no substitute for stiffer privacy regulation.
Ditto for concerns relating to self-preferencing. Competition raises the cost of discriminating for a platform’s distribution division. But in the presence of switching costs and imperfect information, discrimination against similarly-situated edge rivals likely would still be profit-maximizing, even in the face of modest platform competition. In short, I am skeptical when it comes to policing the tech platforms’ worst abuses just by adding more competition. And I am especially skeptical when that message comes from Delrahim. At least other antitrust scholars and practitioners are honest about the limits of antitrust.
Hal Singer is a Managing Director at Econ One and an Adjunct Professor at Georgetown’s McDonough School of Business.
The ProMarket blog is dedicated to discussing how competition tends to be subverted by special interests. The posts represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty. For more information, please visit ProMarket Blog Policy.