After decades out of the spotlight,1 antitrust lawsuits appear to be back in vogue.2 Though recent attempts to enjoin mergers have mostly failed,3 some experts view ongoing DOJ and FTC efforts under section 2 of the Sherman Act to rein in allegedly anticompetitive conduct by large technology companies with more optimism.4 Still, uncertainty over the agencies’ ability to show that certain firms’ activities caused anticompetitive harm persists,5 and what threshold of proof suffices to establish causation remains unclear.6 Recently, in United States v. Google, LLC,7 the U.S. District Court for the District of Columbia held that Google illegally maintained monopoly power in two online markets by forming exclusive agreements that caused anticompetitive harm.8 Ascertaining anticompetitive harm without using “but-for” evidence,9 the court clarified ambiguity over the appropriate causation standard for establishing liability, selecting a lenient one on contestable grounds and applying it in an especially lenient fashion. As a result, the DOJ’s and FTC’s ability to constrain business activity effectively expanded.
Since the 2000s, Google has maintained a leading position in the market for general search services.10 General search encompasses “the market for operating and offering a general search engine” that “crawl[s] . . . the entire (general) internet,” instead of a subset of webpages or topics, to return listed results.11 For a long time, many attributed Google’s dominance to talent and quality.12 However, antitrust agencies eventually began to ponder additional, less innocuous explanations for the firm’s success.13 After investigating for over a year, on October 20, 2020, the DOJ, with attorneys general from eleven states, sued Google in the U.S. District Court for the District of Columbia.14 According to the DOJ, Google “violated Section 2 of the Sherman Act” by forming “exclusive agreements to secure default distribution” of its search and advertising services to maintain monopolies in three online markets.15 Soon after, on December 17, 2020, thirty-eight states sued Google in the same district.16 This complaint raised additional claims of monopolization in a fourth online market, as well as of “exclusionary conduct . . . that targeted specialized vertical providers”17 (SVPs) and users of Google’s advertising management platform.18 On January 7, 2021, the court granted a motion to “consolidate[] the two cases for pretrial purposes,” before consolidating them for trial.19
After holding trial from September to November 2023,20 the court ruled largely in the plaintiffs’ favor.21 Judge Mehta stated that “Google is a monopolist, and it has acted as one to maintain its monopoly.”22 First, the court held that general search services and general search text advertising constituted “relevant product markets,”23 applying the market definition factors24 from Brown Shoe Co. v. United States.25 Second, in light of “direct and indirect evidence,” the court held that Google possessed monopoly power in these markets.26 Third, the court analyzed whether Google’s distribution contracts constituted exclusive agreements.27 Applying the exclusive dealing framework from United States v. Microsoft Corp.,28 the court held that Google’s agreements with Apple, Mozilla, and Android distributors, among others, which “establish[ed] Google as the out-of-the-box default search engine,”29 constituted exclusive dealing.30 The agreements did not expressly exclude rivals such as Bing and DuckDuckGo31 — however, they proved sufficiently exclusive in practice to the court.32
Fourth, the court held that Google’s exclusive agreements caused anticompetitive effects in the general search services and general search text advertising markets.33 In general search text advertising, Judge Mehta concluded that the agreements foreclosed 45% of the market and enabled Google to raise advertising prices, lower quality, and limit rivals’ revenues.34 In general search services, Judge Mehta also concluded that Google caused anticompetitive harm through market foreclosure, as “Google’s exclusive distribution agreements foreclose[d] 50% of the general search services market by query volume.”35 For the figure, Judge Mehta cited plaintiffs’ expert, Dr. Michael Whinston, who had found that “50% of all [general search engine] queries in the United States are run through the default search access points covered by the challenged distribution agreements.”36 As suggested by his citation to market shares, Judge Mehta relied on Microsoft’s lenient causation standard when determining the agreements’ effects.37 Microsoft allows courts to “infer ‘causation’” when “a defendant has engaged in anticompetitive conduct that ‘reasonably appear[s] capable of making a significant contribution to . . . maintaining monopoly power.’”38
In applying Microsoft, Judge Mehta declined to require evidence of but-for causation. He noted that a but-for causation standard would impose a substantial evidentiary burden on plaintiffs who cannot “confidently reconstruct . . . a world absent the defendant’s exclusionary conduct.”39 Further, a stringent standard could encourage monopolists to take earlier anticompetitive action against potential entrants whose forecasted market impact may be less measurable.40 Thus, inferring anticompetitive effects from exclusivity via Microsoft did not require “thought experiments” that compared actual outcomes to hypothetical outcomes in a “but-for world” sans Google’s default agreements.41 While Judge Mehta acknowledged that foreclosure, according to Dr. Whinston, would “ideally” be estimated against a counterfactual world without the default agreements, he concluded “the law does not require it.”42 Judge Mehta further rejected Google’s argument that Rambus Inc. v. FTC43 could be read “to support the need for a but-for world showing” when evaluating the anticompetitive effects of the default agreements.44
Fifth, the court rejected Google’s procompetitive justifications for the default agreements.45 As a result, the court concluded that Google violated section 2 of the Sherman Act.46 Judge Mehta ruled in Google’s favor on some matters, including sanctions,47 as well as the absence of a duty to deal with rivals on its own platforms.48 Still, commentators saw the ruling as a setback for Google,49 which announced its intention to appeal.50 In anticipation of the imminent remedies trial, debate over what form remedies should take has begun.51
In holding that Google’s exclusive agreements caused anticompetitive harm without requiring but-for proof, the court clarified a previously unclear standard for establishing a “causal link”52 to grant injunctive relief. Preferring Microsoft’s framework over Rambus’s stringent but-for standard,53 Judge Mehta adopted a lenient threshold requirement for establishing causation of anticompetitive harm. Still, by overlooking theoretical reasons for favoring this lenient causation standard, Judge Mehta justified dismissing the Rambus standard on rather superficial grounds. Further, by interpreting Microsoft’s causation standard leniently in Google’s distinct factual setting, Judge Mehta effectively loosened an already lenient evidentiary burden for entities that are better positioned to meet stringent standards. In all, extending an especially lenient causation standard expands antitrust agencies’ ability to enjoin business activity under section 2 of the Sherman Act.
In following Microsoft rather than Rambus, Judge Mehta installed a lenient causation framework for inferring exclusive agreements’ anticompetitive effects in a wide array of activities. Before Google, courts and experts long struggled to determine what standard of proof sufficed to establish when monopolistic conduct caused anticompetitive harm that “r[a]n afoul of Section 2 of the Sherman Act.”54 Disagreements over evidentiary standards manifested in the Microsoft and Rambus courts’ respective reasonings. In Microsoft, the D.C. Circuit was concerned that stringent causation standards would unfairly “allow monopolists free reign to squash nascent, albeit unproven, competitors” who could not reliably “reconstruct the hypothetical marketplace absent a defendant’s anticompetitive conduct.”55 That led the court to allow the inference that exclusive “conduct that is reasonably capable of contributing significantly to a defendant’s continued monopoly power” caused anticompetitive harm.56 This lenient standard would sometimes require the defendant “to suffer the uncertain consequences of its own undesirable conduct.”57 In Rambus, the D.C. Circuit reversed an FTC determination that Rambus’s misrepresentation of its patent interests to a standards-setting body violated section 2 of the Sherman Act, even though the body conferred an exclusionary advantage on Rambus by incorporating its patents into industry-wide technological standards.58 Unfazed by Microsoft, the D.C. Circuit explained that if the FTC did not show that the standards-setting body “would have standardized the very same technologies” in “the world that would have existed but for Rambus’s deception,”59 then it did not establish an anticompetitive effect from Rambus’s actions.60 Thus, before Google, some insisted that Microsoft’s “reasonably capable” standard covered most government suits under section 2.61 Others promoted Rambus’s more stringent “but-for” requirement as a default, with an exception for “nascent” competitive threats.62 By citing to market shares without requiring a more rigorous empirical inquiry into whether rivals could have grown but for Google’s default contracts,63 Judge Mehta sided with the former group.
Although rejecting Rambus clarified conflicting precedent, Judge Mehta’s reasoning for preferring Microsoft’s lenient causation standard rested on superficial factual grounds. Judge Mehta distinguished Rambus64 and noted its failure to explicitly overturn Microsoft.65 His reasoning underweighted a key theoretical rationale for relaxing causation standards: the infeasibility of producing but-for evidence.66 Both Supreme Court and D.C. Circuit precedent opined that requiring but-for proof to infer anticompetitive harm from exclusive conduct may sometimes create an insurmountable burden of proof on plaintiffs.67 Though Judge Mehta cited to these concerns when deciding to apply a “reasonably capable” standard,68 he failed to demonstrate that insurmountable evidentiary barriers justified forgoing a but-for requirement. In fact, he refused to consider the plaintiff expert’s actually produced but-for analysis on the anticompetitive effects of Google’s default agreements.69 This omission overlooked the key theoretical rationale for choosing a lenient causation requirement to infer anticompetitive harm.70
Beyond disfavoring Rambus’s stringent causation standard, Judge Mehta further relaxed causation requirements for established competitors by leniently interpreting Microsoft in Google’s distinct factual setting. In Google, Judge Mehta faithfully applied Microsoft’s first requirement that a monopolist’s conduct must not be “reasonably capable of contributing significantly to a defendant’s continued monopoly power.”71 However, he was silent on how Microsoft’s second requirement that competitors “reasonably constituted nascent threats”72 applied to Google’s facts. In Microsoft, the D.C. Circuit noted that “the District Court made ample findings” of fact before concluding that certain startups “showed potential” as viable “threats” to Microsoft’s Internet Explorer.73 How this second component maps to Google, where rivals like Bing and DuckDuckGo constitute established substitutes, not nascent threats, is unclear. One might infer a requirement that established competitors demonstrate some degree of viable capability to threaten an incumbent’s position in the absence of exclusive agreements.74 On the other hand, perhaps established substitutes are a threat simply by remaining on the market. This latter interpretation would de facto remove the second prong of the Microsoft standard for established competitors. As a result, antitrust agencies would effectively face lower evidentiary burdens when evaluating established competitors rather than nascent potential entrants, even though higher evidentiary burdens would probably be more difficult to overcome when evaluating nascent entrants.75 Thus, by favoring the latter approach, Judge Mehta further loosened an “edentulous”76 standard for inferring causation when evaluating entities that are already better positioned to satisfy more stringent standards of causal inference.
In all, the causation standard implemented in Google empowers antitrust agencies to enjoin a wider range of business activity via section 2 of the Sherman Act. Potential for error in distinguishing anticompetitive from procompetitive conduct likely explains why disagreements over the stringency of evidentiary requirements for causation have persisted.77 Requiring a stricter “but-for” evidentiary standard to infer causation would more likely ensure nonintervention in firms’ legitimate procompetitive conduct at the expense of allowing some anticompetitive conduct.78 Implementing a more lenient evidentiary standard for causation à la Microsoft would produce the inverse tradeoff: More anticompetitive conduct would be caught at the expense of chilling some permissible actions.79 Google moves antitrust doctrine toward the latter approach, yet in doing so, it arguably misallocates deferential evidentiary standards to less needy entities. Some may welcome Google’s widening of regulatory oversight over business activity as a check against anticompetitive practices.80 For others, the ruling primarily hinders procompetitive diffusion of high-quality technology.81 Either way, as more disputes between tech companies and antitrust agencies loom,82 vigorous debate over the optimal causal standard for inferring anticompetitive harm from exclusive actions will almost surely persist.