The Google Search Decision: The Merits, Fate, and Potential Impact of the U.S. District Court’s Decision on the Question of Liability
On August 5, 2024, Judge Amit P. Mehta issued his long-awaited opinion on liability in United States v. Google.[1] Joined by the attorneys general of 11 U.S. states, the U.S. Department of Justice (DOJ) had filed its monopolization complaint,[2] under Section 2 of the Sherman Act,[3] almost four years prior, in the waning months of the Trump administration. Judge Mehta found Google liable for illegal monopoly maintenance (or monopolization) in two markets: general search service (GSS) and general text advertising.[4]
At the same time, Judge Mehta dismissed various of the states’ claims against Google, although he sustained Counts I and III of the plaintiff states’ complaint to the extent that those claims were coextensive with Counts I and III of the U.S. government’s complaint.[5] The government’s second claim was rejected, as Judge Mehta found that Google lacked the requisite market power for a Section 2 violation in the broader search advertising market.[6]
The question of remedies remains, as proceedings have been bifurcated into separate liability and remedies phases, as the plaintiffs had requested. A decision on remedies is expected by August 2025.
Judge Mehta’s 286-page memorandum opinion is detailed and carefully reasoned, but that is not to say that it is uncontroversial or that an appeal cannot find traction. Commentary on the U.S. government’s case, and on the decision of the district court, has been substantial and, like contributions to this On-Topic, divided.[7] Issues on appeal may include, among others, market definition (perhaps especially that of the general text advertising market) and various questions about the governing standard for establishing monopolization, including the question whether the government was required to show “but for” causation of competitive harm.
Google, unquestionably, has a very large share of general search in the United States. The decision says that “[b]y 2020, it was nearly 90%, and even higher on mobile devices at almost 95%.” Still, as the U.S. Court of Appeals for the District of Columbia Circuit (D.C. Circuit) noted in Microsoft, “merely possessing monopoly power is not itself an antitrust violation.”[8] What was alleged, and what Judge Mehta found, was illegal monopolization; that is, “the willful acquisition or maintenance of that power as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident.”[9]
To unpack a complex matter slightly, the DOJ alleged, and Judge Mehta found, certain contracts unlawful: Google has entered into agreements with certain device manufacturers (notably, but not exclusively, Apple) and Mozilla, which developed and markets Firefox. Under those contracts, Google is the default general search engine (GSE) preloaded in, e.g., Apple’s Safari (and hence on iPhones) and Mozilla’s Firefox browsers. These were deemed de facto exclusive-dealing contracts, although nothing in the contracts precludes access to other browsers or other search engines and it is, in fact, possible for users to switch. As such, these contracts were alleged to foreclose access to the market by actual and would-be competitors—access to the markets, that is, access to competition for a certain share of general search and, hence, allegedly, access to the advertising markets at issue.
Hence, central to the findings of illegal monopoly maintenance was the issue of “exclusionary conduct,” just as the D.C. Circuit ruled that exclusionary conduct was critical to illegal monopoly maintenance in Microsoft.
And the decision’s construction and application of the D.C. Circuit’s 2001 decision in Microsoft may appear front and center on appeal. Unlawful “exclusionary conduct” was the gravamen of the U.S. government’s complaint in Google Search, just as it had been in Microsoft.
That, in turn, raises difficult questions about the role of causation in proving monopolization. The Google Search decision recognizes that even truly exclusionary contracts are not per se illegal under U.S. antitrust law. Liability turns on the question whether the default agreements caused Microsoft’s Bing GSE, among others, to be excluded from access to the market, and what the DOJ had to show to establish causation.
Microsoft recognizes that, “to be condemned as exclusionary, a monopolist’s act must have an ‘anticompetitive effect.’ That is, [the monopolist] must harm the competitive process and thereby harm consumers. In contrast, harm to one or more competitors will not suffice.”[10] And “the plaintiff, on whom the burden of proof of course rests (. . .) must demonstrate that the monopolist’s conduct indeed has the requisite anticompetitive effect.”[11]
At the same time, Mehta reasons that causation does not require but-for proof; that is, the plaintiff is not required to show that, but for the defendant’s exclusionary conduct, the anticompetitive effects would not have followed. Such a standard would create substantial proof problems, as “neither plaintiffs nor the court can confidently reconstruct . . . a world absent the defendant’s exclusionary conduct.”[12]
And that raises the question of what it means to say that the conduct caused the competitive harm at issue if not that, but for the defendant’s exclusionary conduct, or some other intervening cause, the anticompetitive effects would not have followed.
In Microsoft, the D.C. Circuit found what it called an “underlying proof problem.” That is, for “nascent competitors,” such as Netscape, the plaintiffs (the government) could not, as a practical matter, establish the “but for” counterfactual. Under those conditions at least, faced with uncertainty about what, e.g., Netscape might have become as a competitor, the circuit court reasoned that it could infer causation under a lower evidentiary standard.
To some degree, “the defendant is made to suffer the uncertain consequences of its own undesirable conduct,” i.e., the practices that had been shown to have had anticompetitive effects. In these carefully limited circumstances, the questions were “(1) whether as a general matter the exclusion of nascent threats is the type of conduct that is reasonably capable of contributing significantly to a defendant’s continued monopoly power and (2) whether Java and Navigator reasonably constituted nascent threats at the time Microsoft engaged in the anticompetitive conduct at issue.”[13]
Was Judge Mehta right in holding that Microsoft’s burden of proof applies? Of particular interest is that Mehta explicitly distinguishes a later opinion of the D.C. Circuit; that is, the court’s 2008 opinion in Rambus,[14] where, seven years after the Microsoft decision, the court applied the but-for causation standard: “in the world that would have existed (. . .) [the] alleged deception cannot be said to have had an effect on competition in violation of the antitrust laws.”[15]
According to Judge Mehta’s opinion, the 2008 determination that “the Commission failed to demonstrate that Rambus’s conduct was exclusionary, and thus to establish its claim that Rambus unlawfully monopolized the relevant markets,”[16] was confined to the peculiarities of the standard-setting claims at issue. He reasoned that “Rambus does not establish a categorical rule that the anticompetitive effects of an exclusive agreement must be measured against a but-for world,”[17] because otherwise the 2008 Rambus decision would have repudiated Microsoft, and it does not. That is, Rambus dealt with a special case—an exception to the general rule established by Microsoft.
That view of Rambus is controversial, and it may well find traction on appeal. A discussion of the D.C. Circuit’s opinion in Microsoft, in a 2020 article by Judge Douglas Ginsburg and Koren Wong-Ervin,[18] may be instructive, even if it does not settle the question as a matter of law (or bind courts going forward). As it happens, Judge Ginsburg was the chief judge of the D.C. Circuit when the court issued its per curiam opinion in Microsoft; and he may have authored that opinion, in whole or in part. In brief, Ginsburg and Wong-Ervin argue that Microsoft is the special case, an exception to the general rule that was restated in Rambus.
On that view, in Microsoft, the standard was not lowered simply because the government could not meet a higher one. Rather, numerous findings of fact in Microsoft were taken to have established the effects of Microsoft’s conduct; these, in turn, were deemed to have kept Netscape and others, as matters of fact, from reaching the minimum efficient scale at which they could have become effective competitors to Microsoft. Under those conditions, faced with uncertainty about what Netscape might have become as a competitor, the circuit court reasoned that it could infer causation under a lower evidentiary standard.
Ginsburg and Wong-Ervin argue that Rambus underscored the general rule, and not an exception to it, in deciding “that the agency failed to prove that ‘but for’ the defendant’s conduct, there would have been harm to the competitive process.”[19]
The contributions to this On-Topic cover a range of views on the Google Search matter, from the merits of the district court opinion on liability to challenging issues in the remedies phase. Notably, the contributors—including those highly critical of the district court’s findings of liability—uniformly recognize that the opinion was carefully drafted, based on close attention to the facts and the law in the case.
Ioannis Stefatos reviews the district court’s opinion favorably, finding it grounded in “market realities” rather than theoretical possibilities. Given entry barriers, the apparent effects of the contracts at issue, and the degree of apparent foreclosure, he suggests that “the court clearly established the causal link between Google’s exclusionary practices and the ensuing anticompetitive harm.”
Walid Chaiehloudj also views the decision favorably, while reserving judgment on the ultimate impact of the decision on relevant markets and, more broadly, on U.S. antitrust law. Contrasting U.S. and European antitrust approaches, he takes it as a given that Google enjoys a dominant market position and that Google’s default agreements are exclusionary and anticompetitive. He suggests that U.S. antitrust has been too concerned with Type I errors—false positives—and that Judge Mehta’s findings were accurate and his rejection of the “but-for” causation standard articulated in Rambus was salutary. He reserves judgment on the ultimate impact of the case, given, first, the simple fact that the court has yet to impose remedies, second, that we have yet to see how the decision will fare on appeal, and, third, that future market developments may be hard to predict in any case.
Jonathan Barnett, in contrast, argues that Judge Mehta’s opinion, while carefully reasoned, falls prey to what he calls “the tyranny of the model.” That is, the opinion relies overmuch on an abstract model of “‘default bias,’ mostly as set forth in the academic literature, [paying] insufficient attention to on-the-ground competitive conditions in search markets.” As a consequence, he suggests that the court paid insufficient attention to the factual record and actual market dynamics at play, which “most likely [do] not,” in Barnett’s view, support the finding of an unlawful exclusionary practice.
My own piece, similarly, critiques the use of certain generalizations in the court’s findings of fact. Specifically, it focuses on evidence from behavioral economics regarding default bias and evidence of scale effects—and purported self-perpetuating scale advantages—in engineering general search engines. It is argued that both sorts of generalities are overstated and misapplied in the district court’s decision. Those might not be leading vulnerabilities on appeal, but they are important to the extent that similar reasoning could be applied in tech monopolization cases going forward.
Gregory Werden is similarly critical of the court’s decision, if on different grounds. Werden suggests that the bifurcation of the trial into liability and remedies phases obscures key questions about what Google should have done differently; that is, what alternative arrangements would not be de facto exclusive, such that they would have made a competitive difference? The decision of liability turns, in part, on the notion that the default agreements at issue prevent rivals from reaching the scale at which they might effectively compete (perhaps, minimally efficient scale). Werden finds the connection tenuous, first because Google achieved significant scale advantages in advance of the arrangements at issue, and second, because of Google’s share of general search on Windows machines running Microsoft’s Edge, on which Bing is the preloaded default browser. Werden argues that the proper question for the court should have been whether Google’s arrangements flunk the “no economic sense” test described in Microsoft. He suggests that they did not. Among other things, “Apple concluded that installing Microsoft’s Bing as the default in Safari would not make business sense even if Apple got every penny of Microsoft’s GSA revenue from Safari users.”
Bilal Sayyed’s contribution takes a different tack, looking forward to the remedies phase and the diverse behavioral and structural remedies proposed by the government (and advocated by some commentors). He focuses specifically on the treatment of structural remedies in exclusive dealing and refusal to deal matters, by the U.S. enforcement agencies and the courts, in the wake of the D.C. Circuit’s decision in Microsoft, where the court “indicated that ‘structural relief (. . .) require[s] a clearer indication of a significant causal connection between the conduct and creation or maintenance of the market power. Absent such causation, the antitrust defendant’s unlawful behavior should be remedied by ‘an injunction against continuation of that conduct.’” He notes that, since the Microsoft decision, “[n]either the [Federal Trade] Commission nor the Department of Justice has pursued structural relief in matters finding monopolization through exclusive dealing or discriminatory practices.” Rather, across industries and courts, “the agencies’ remedies have aligned with the requirement of Microsoft that the remedy address the conduct or agreements that allegedly maintained or enhanced the defendant’s or respondent’s monopoly power. They did not require relief that attempted to undo monopoly power not found to have been obtained unlawfully.” He argues that the factual findings in the liability phase, coupled with Microsoft’s remedy principles, “should constrain the DOJ from seeking structural relief.” And, further, that they “will undoubtedly constrain the court from ordering structural relief.”
Courtney C. Radsch & Karina Montoya also discuss potential remedies in the case. Given the findings of liability, they argue for far reaching behavioral remedies in addition to certain divestitures. They argue that the complexity of the technical space and commercial activity at issue suggest that narrow remedies will be ineffective and potentially call for utilities-style regulation.
Kai-Uwe Kühn and Miroslava Marinova critique Judge Mehta’s decision, although their critique is at least somewhat agnostic on the question whether “a deeper analysis (. . .) would have altered the outcome.” Kühn and Marinova focus on the relationship between the D.C. Circuit’s Microsoft and Rambus decisions discussed above, noting that “Judge Mehta relied heavily on the Microsoft precedent, treating it as a general rule.” They argue, to the contrary, that Rambus articulates the general standard for monopolization cases under Section 2 of the Sherman Act: “but-for” causation. Microsoft, they argue, in keeping with Ginsburg and Wong-Ervin, is a special case, fashioned to deal with the specific conduct at issue, its demonstrated effects on nascent competitors, and the difficulty of forecasting its likely effects on what competition would have become, but for the conduct at issue. By way of contrast, the Google case “involves well-established competitors like Bing in a mature market.” The difference is “crucial,” they say, given ample evidence of the competitive interactions between existing competitors. The court’s erroneous market definition effectively lowered the government’s burden of proof, distracting from the difficult—but not necessarily intractable— question whether Google’s default agreements caused harm to competition or were, in the alternative, efficient allocations of default status. They acknowledge that “Google’s agreements with device manufacturers and browsers may raise legitimate concerns about market foreclosure,” but that the proper resolution of those concerns is not clear.
On their analysis, one “should expect markets for exclusive default placement of search engines to develop to allocate such defaults efficiently. This will lead to considerable payments to browsers or device manufacturers for these default slots.” But that does not, for them, settle the question whether the specific terms of the agreements at issue did or did not have the requisite “but-for” effects.
Geoffrey Manne also critiques Judge Mehta’s reliance on Microsoft’s “edentulous” test for causation, instead of the “but-for” causation standard articulated in Rambus. Manne argues that this dramatically—and wrongly— lowered the government’s burden of proof based on a misunderstanding of “both the precedent and the nature of causation in antitrust cases.” Throughout the decision, he identifies examples of “post hoc reasoning”; that is, fallacious inferences from correlation to causation.
[1] United States v. Google, LLC, 2024 WL 3647498, at *66 (D.D.C. Aug. 5, 2024).
[2] Ibid. (Compl.) (Oct. 20, 2020.)
[3] 15 U.S.C. § 2.
[4] United States v. Google at 276.
[5] Ibid.
[6] Ibid. at 180.
[7] Compare, e.g., G. J. Werden, Harm to the Competitive Process in the Google Case, Mercatus Ctr. Antitrust and Comp. Working Papers, July 9, 2024, https://www.mercatus.org/research/working-papers/harm-competitive-process-google-case (critiquing the government’s case) with E, Hovenkamp, What Does the Google Antitrust Decision Mean and Where Will It Take Us?, ProMarket, Aug. 22, 2024, https://www.promarket.org/2024/08/22/what-does-the-google-antitrust-decision-mean-and-where-will-it-take-us (arguing that default placements had anticompetitive effects). See also, e.g., G. A. Manne, A Critical Analysis of the Google Search Antitrust Decision, ICLE White Paper 2024-08-14, Aug. 14, 2024, https://laweconcenter.org/resources/a-critical-analysis-of-the-google-search-antitrust-decision; D. A. Crane, Ranking the Big Tech Monopolization Cases, Yale J. Reg. (Notice & Comment, online), Mar. 26, 2024, https://www.yalejreg.com/nc/ranking-the-big-tech-monopolization-cases-by-daniel-a-crane; D. J. Gilman and B. C. Albrecht, Ranking the Big Tech Monopolization Cases in the Wake of the Google Search Decision, Yale J. Reg. (Notice & Comment, online), Sept. 19, 2024, https://www.yalejreg.com/nc/ranking-thebig-tech-monopolization-cases-in-the-wake-of-the-google-search-decision-perspectivesof-some-economists-and-legal-scholars-by-daniel-j-gilman-brian-c-albrecht.
[8] United States v. Microsoft, 253 F.3d 34, 51 (D.C. Cir. 2001).
[9] Ibid. at 50.
[10] Ibid. at 58.
[11] Ibid.
[12] United States v. Google at 220.
[13] Microsoft, 253 F.3d at 79.
[14] Rambus Inc. v. FTC, 522 F.3d 456 (D.C. Cir. 2008).
[15] United States v. Google at 218–219.
[16] Rambus, 522 F.3d at 467.
[17] United States v. Google at 219.
[18] D. H. Ginsburg and K. Wong-Ervin, Challenging Consummated Mergers Under Section 2, CPI North America Column, May 25, 2020, https://www.competitionpolicyinternational.com/wp-content/uploads/2020/05/North-America-Column-May-2020-4-Full.pdf.
[19] Ibid. at 4 (citing Rambus, 522 F.3d at 466).