The U.S. Department of Justice and eight states recently sued Google, claiming it runs its digital ad business to unfairly advantage its own business, to the detriment of its customers and potential rivals. Some commentators have expressed concerns that the government would make its case using novel and untested legal models, but Salil Mehra writes that these concerns are misplaced and that the case will be primarily argued using traditional antitrust law.
In January, eight states and the U.S. Department of Justice jointly filed a lawsuit against Google for monopolization, attempted monopolization, and tying in the digital advertising market. The suit’s focus on the “Ad Tech Stack” is novel, and, before the FTC’s 2020 suit against Facebook and its parent company, Meta, the DOJ’s prayer for relief in the form of divestiture and its filing of a Sherman Act Section 2) claim—monopolization and attempted monopolization—would have been fairly novel, at least for the government. The case comes in the wake of critics expressing concern that Biden Administration antitrust officials would bring antitrust cases based on novel or untested legal theories. Whether the DOJ eventually does bring such boundary-testing cases, the Google Ad Tech Stack case is not one of them. In fact, the legal theory behind the DOJ’s lawsuit against Google is well-established.
As background, it’s worth noting that the DOJ is really positing four substantive antitrust violations: 1) that Google has monopolized the publisher ad server market in violation of Sherman Act Section 2; 2) that Google has either monopolized or attempted to monopolize the ad exchange market in violation of Section 2; 3) that Google has monopolized the advertiser ad network market in violation of Section 2; and 4) that Google has unlawfully tied its ad exchange (AdX) to its publisher ad server—known as DoubleClick for Publishers or “DFP”—in violation of Sections 1 and 2.
All four accusations depend on showing that Google has monopoly power (or for tying, significant market power) in the publisher-side ad server market, the advertiser-side ad network market, and/or the ad exchange market where these two sides meet. While the ad tech stack as a factual matter may be unfamiliar to those outside the online advertising world, the analysis is well-established and straightforward: market power can be shown either directly or, more likely in this case, indirectly. The latter depends on establishing that publisher ad servers or ad exchanges are relevant product markets without reasonable substitutes, that Google has a high market share, and that entry would be difficult, costly, or time-consuming.
Of course, monopolization does not stop there as a theory. The DOJ must show improper exclusion or, for the tying claim, conditioned sale of products with separate consumer demand. The three monopolization-related claims are built on overlapping allegations of unjustified exclusionary conduct by Google. These include Google acquiring DoubleClick and subsequently restricting its advertiser-side sales tool, Google Ads (which allows businesses to buy ads that can be seen by Google search users alongside search results), to working only with its ad exchange AdX, an intermediary between the advertiser and publisher sides of the Ad Tech Stack. These claims also allege that Google restricted real-time access to AdX exclusively to users of DFP.
Assuming that the DOJ can successfully show Google has monopoly power in the ad exchange and/or publisher ad server market, though, allegations such as the anticompetitive acquisition of a competitor and exclusion via preventing customers from using competitors have a long pedigree in monopolization cases. Of course, even were the DOJ to prove these allegations, Google would still have the opportunity to show a legitimate procompetitive justification for these actions. For example, Google might point to effects, at the margin, of these policies that make its ad stack or elements of it more attractive for advertisers, publishers, or both. It will be interesting to read Google’s forthcoming Answer to the DOJ’s complaint to get a sense of what Google will be arguing in this vein.
That said, at least one set of factual allegations in this case does involve a fairly novel phenomenon for antitrust law to grapple with: algorithmic pricing. Google matches ad buyers with publishers through an AdX automated auction. Until 2018, Google, which is also a publisher, had a “Last Look” policy which gave its exchange the option to come in at the end of an auction to win with a better bid. However, the publishers, together with non-Google publisher-side ad tech firms, had used “header bidding” to, among other things, set a higher floor price for Google’s AdX to win the auction relative to other ad exchanges to win back control of their ad inventory. In 2018, Google discontinued Last Look because it had by then created a proprietary algorithm called “Smart Bidding” that had been trained on and could continue to draw on a data trove exclusive to Google that other firms could not replicate. The DOJ alleges that Smart Bidding is able to achieve the same kind of “insider” auction advantage that Google had enjoyed with Last Look.
This algorithmic pricing allegation is not crucial to the DOJ’s success in the case as a whole—one could imagine the DOJ winning the relief it seeks without proving this story, and that may explain why it only occupies about three pages near the end of the complaint. Certainly, observers have pointed for almost a decade to the possibility that algorithmic pricing creates a novel challenge for antitrust law. Additionally, this set of facts, if proven true, might argue for compulsory access to the algorithm or data trove at issue, perhaps under fair, reasonable, and non-discriminatory (FRAND) terms. This would raise traditional concerns about the essential facilities doctrine (in which an owner of a bottleneck facility must provide access at a reasonable price), at least in a Section 2 context. As such, to the extent that the DOJ might be able to connect the algorithmic pricing allegations to its Section 1 tying claims, these concerns might be much less relevant, given that even cases like Trinko did not extend their doubts about essential facilities to the Section 1 context.
In sum, while the Ad Tech Stack will take some serious explaining, whether to a judge or, as the DOJ seeks, a jury, the actual legal theories behind the case are not the sort of novel Neo-Brandeisian-inflected stories we have been warned about. We will have to wait and see how they play out in a new and complex factual setting.
The posts represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Read about our disclosure policy here.