Market Definition
When asked about the key issues in the decision, Ms. Broz first highlighted the Court’s finding on market definition, citing arguments about whether the relevant market was that for “general search,” or if it was more appropriate to include other categories, such as social media and vertical site searches like those performed within Amazon and other sites. The Court decided that the relevant market was “general search,” reasoning that if an individual is searching for a product on social media, that search constitutes a separate market. In its market definition analysis, the Court considered the Brown Shoe factors, finding peculiar characteristics in the market for general search. When conducting general searches, consumers are looking not only for commercial information, but also noncommercial information. The Court concluded that running searches on social media or through Amazon were not reasonable substitutes for general search engines (“GSEs”), but rather were more often complements. In other words, a consumer is more likely to search for a product, for instance, on Amazon, and then search for the same product on Google, to help inform her purchase. Ms. Broz emphasized that these peculiar characteristics of general search drove the Court’s finding on market definition. Among the Brown Shoe factors, the court did not discuss the product’s sensitivity to price changes, because one of the major issues in this case was that Google’s search engine is, of course, a zero-price product.
Professor Backus expressed surprise that the Court did not perform a version of the SSNIP test in this decision, surmising that perhaps that was because Google search is a zero-price market. In this context, the SSNIP test may not make much sense, as we do not expect Google to “raise the price” of its general search engine product. Price, however, is just one attribute that can be used to identify substitute products. Professor Backus noted that there are ways to think about something like the SSNIP or hypothetical monopolist test to identify the relevant market: for example, rather than imagining a hypothetical price increase, one might think about a hypothetical degradation of quality—in other words, lower quality search results.
Conduct at Issue
Speaking on the specific conduct at issue in this case, Mr. Emanuelson sized up the case as one involving “good, old-fashioned exclusive dealing.” Here, the Court evaluated Google’s conduct under the standard Sherman Act Section 2 framework, and ultimately found the company liable. From a purely legal perspective, the decision applies standard antitrust law in the sense that it looks at Google’s conduct and evaluates whether that conduct creates anticompetitive effects. If it does, the onus is on the defendant, Google, to proffer procompetitive justifications for that conduct. If the defendant offers procompetitive justifications, then courts perform a balancing test: does the procompetitive justification outweigh the anticompetitive effect? This is the classic balancing test, also performed by the court in Microsoft; however, Mr. Emanuelson observed, the application of that analysis in this case is where much of the debate over the recent decision lies.
The alleged conduct in this case is akin to an exclusive dealing theory: the Court analyzed a series of de facto agreements between Google and other sellers of internet-connected devices. The Court deemed these agreements to be between Google and “search distributors,” companies that were either putting search directories on their browsers, or that had some involvement through their manufacturing of computing devices, computers, or cell phones, that allow for the distribution of “general search.” Mr. Emanuelson noted, however, that the exclusive agreements at issue in this case were far more nuanced than the standard variety, where a defendant company requires an entity to use its product and prohibits it from using other companies’ products. The basic concept underlying Google’s agreements was that of revenue sharing: when Google contracted with Apple or Android, the other companies would get monetary compensation if they engineered a preference for Google in their search engines. The idea was that of a default search engine, where users, upon opening up the web browser on their devices, would see Google search first.
The Court found that the general search market derives from these default preferences. In other words, browsers contain a default search access point, and only GSEs occupy those default positions. Consequently, the opinion describes Google’s exclusive agreements as the most efficient channels through which GSEs could get their search products out to the market, concluding that about 50% of the market was subject to them. In its analysis, the Court found that anticompetitive effects resulted from 50% of the market being subject to the default, and Google could not put forth a procompetitive justification in defense of its exclusive agreements. The Court did not reach the third step of the balancing test, which weighs the anticompetitive effects against the proffered procompetitive justifications, because the court found that there were no procompetitive justifications for Google’s conduct.
Microsoft’s Influence
Mr. Emanuelson also spoke about the influence on the Google decision of the Court’s prior decision in Microsoft, which is controlling precedent in the D.C. Circuit, and highlighted how the influence of Microsoft is evident at the heart of the Court’s holding in Google. The big debate here, he said, was whether Google’s conduct could be considered exclusionary conduct. There were a few analogues to the Microsoft case, which also dealt with the issue of a default product. For example, in Microsoft the Court found that Microsoft went to internet service providers, established Internet Explorer as a default product and knocked off Netscape, which had already been established as the premier browser. Importantly, Microsoft did not tell equipment manufacturers they could not use Netscape; it said only that they must use Internet Explorer as the default browser.
The key aspect of Microsoft’s influence, and what Mr. Emanuelson proposed would be the aspect most heavily litigated on appeal, is the relaxed standard for causation adopted by the Court in Google. Mr. Emanuelson thought the Court struggled with the fact that Google’s search technology was so much better than its rivals’ technologies, in contrast to Microsoft, where a good product (Netscape) already existed in the market, and Microsoft entered and encouraged the programming of its own browser as the default, at Netscape’s expense. Instead, here Google had a massive advantage in search technologies, with Microsoft’s Bing being the most likely to challenge Google’s position in the market, yet always trying to catch up with Google.
The Court wrestled with whether the exclusive agreements, which comprised 50% of the market, actually caused foreclosure of the market. Considering these agreements, the Court chose to apply a relaxed causation standard, in lieu of a standard “but-for” causation analysis. In support of its decision not to apply “but-for” causation, the Court wrote that courts “may infer ‘causation’ from the fact that a defendant has engaged in anticompetitive conduct that ‘reasonably appear[s] capable of making a significant contribution to . . . maintaining monopoly power,” and “such an inference is appropriate ‘when such conduct is aimed at producers or established substitutes.’” United States, et al. v. Google, LLC, No. 20-cv-3010 (APM) (D.D.C. Aug. 5, 2024) (citing United States v. Microsoft, 253 F.3d 34, 79 (D.C. Cir. 2001) and 3 Areeda & Hovenkamp, Antitrust Law ¶ 651c (1996)). Importantly, the government did not seek damages in this case, but rather sought an injunction, which, according to the panelists, may have had something to do with the Court’s analytical decision on causation.
Professor Backus spoke about the economic issues around market exclusion as applied to this case. In particular, he discussed Branch, a company the government cited as an example of an “innovative search-adjacent technolog[y]” that was thwarted by Google’s exclusionary distribution agreements. Branch is a search engine allowing users to search for mobile applications. Though the platform is not a web search engine, it has the capacity to deliver limited web search results if a user does not have a relevant mobile application on their device. Samsung and AT&T each considered adding Branch to their devices, but ultimately abandoned those efforts after concerns that doing so would affect their relationships with Google. The Court rejected the government’s argument that Branch was a nascent competitor to Google, concluding that Branch is not a GSE. The Court’s decision outlines that Branch is not a nascent competitor akin to Netscape in Microsoft, because there was no expectation or indication that Branch would directly compete with or would surpass Google. Professor Backus noted that what was particularly remarkable about the Branch discussion in the opinion was that the Court took care to document a scenario involving the exclusion of a potential competitor demonstrating innovation on the fringes of the market—a platform akin to and offering a new use case to Google’s search technology.
Anticipated Remedies
When asked about potential remedies that may follow from the Court’s finding of liability, Ms. Broz surmised that it was unlikely the Court would break up Google. She wondered instead if the Court would order a behavioral remedy regarding the exclusivity agreements, an idea she said seems like the bare minimum that should be done, given the Court’s findings on the illegal nature of those agreements. Ms. Broz went on to note that she would not be surprised if the Court attempted to order a structural remedy, saying that while she is unsure what that remedy might look like, it would certainly impact how we all engage with search engines around the world. Professor Backus spoke further about the likelihood that any remedy would address the problems the Court found. He noted that users will want a default, high-quality search engine on their phones and devices. Outlawing payments like those Google paid to entities like Apple and Samsung is doing Google a favor, in one sense, but the payments are “partly revenue sharing and partly user”—in other words, they are part of the marginal cost of selling an iPhone for Apple, or a Galaxy for Samsung. If Apple is not receiving a payment on a per-user basis in expectation of the use of Google on that device, that raises the cost of selling each iPhone, shuts down a source of revenue, and may result in higher prices for devices on the market. That, says Professor Backus, is a bad outcome—it does Google a favor, and harms device manufacturers and consumers. Alternatively, the Court may instead manage to limit Google’s market share by attacking the exclusive contracts, but again, if Google really is the superior product, delivering consumers an inferior default search engine also causes consumer harm.
Further Questions
Ultimately, the panelists agreed that any viable path forward must consider the dynamic story of where this market is going to be in five years, if the Court is going to construct a set of remedies that will be productive in preserving competition. The Court will be forced to wrestle with the question of what constitutes enough competition to be sustainable in the tech industry. Ms. Broz brought up the possibility of requiring Google to share its algorithms or its data with nascent competitors, although that raises further questions about how the law should reconcile increased competition that relies on more providers having access to more data about consumers with privacy concerns. Mr. Emanuelson said that, ultimately, any path forward must consider Google’s incentive to innovate, given that Google is an important source of innovation in the tech industry. Finally, Professor Backus noted that if we can find ways to lower barriers to entry and deal with incumbent entities, that is where there can be noncontroversial gains for consumers. Any solution should consider ways to lower entry costs for competitors. What that looks like in practice remains a difficult question to answer.