Erik Hovenkamp reviews the findings of Judge Amit Mehta’s ruling against Google for monopolizing the internet search market and discusses what the case will mean for the other ongoing Big Tech cases and the future of antitrust.
Four years after the Department of Justice (DOJ) accused Google of monopolizing the internet search market, the verdict is finally in: Google violated the antitrust laws. District Court judge Amit Mehta found that Google illegally maintained its search monopoly through a series of contracts that distorted the competitive process in Google’s favor.
Google’s contracting partners include Apple, Android device makers (e.g. Samsung), browser companies (e.g. Mozilla), and wireless carriers (e.g. Verizon). In a nutshell, the contracts stipulate that Google will pay its partners a share of its search advertising revenue in exchange for making Google Search the default search engine on their products. For example, under Google’s deal with Apple, all search queries typed into the Safari search bar are executed by Google Search by default. All told, about 70% of all search queries in the U.S. flow through “search access points” (mainly browsers) on which Google is the default search engine.
Google’s defaults don’t compel anyone to use Google Search. For example, although Google is the default search engine on Safari, an iPhone user can switch to a different search engine through the settings menu. In fact, this is pretty easy to do. This highlights a key difficulty in the case: if it isn’t very hard to switch to a different search engine, then do defaults really “restrain competition” in a meaningful way?
As a starting point, it is worth highlighting one of the most conspicuous facts in the case. The payments from Google to its contracting partners are astoundingly large. For example, it paid Apple about $20 billion in 2021 alone. Google would not shell out this kind of cash unless it expected to receive something immensely valuable in return. While not incriminating on their own, these large payments signal that Google’s defaults may be more powerful than one might guess.
There are two main things that drive the competitive effects of Google’s default agreements. (I study these effects in detail in a recent economics article.) The first is that, although defaults can be circumvented, they are nevertheless rather “sticky.” Both the DOJ’s expert and Google’s internal researchers estimated that Google would lose significant market share (especially on mobile) if its default positions were taken over by rival search engines.
What accounts for this stickiness? Defaults create “choice frictions” or switching costs that lead some users to stick with the default platform not because they necessarily like it better, but simply because it’s the default. Judge Mehta was also persuaded by arguments from behavioral economics about the influence of habit on consumer choice, and of the impact of defaults on habit formation.
The second key factor has to do with the relationship between quality and scale in the general search market. Search algorithms “learn” from user data, leading them to improve with use. This process resembles network effects in the sense that a search algorithm’s quality goes up when more consumers use it.
These network effects magnify the competitive impact of Google’s defaults. When the defaults are implemented, they shift some market share from rivals to Google, due to the aforementioned “stickiness.” Due to network effects, this shift in market share increases Google’s quality and reduces rivals’ quality. This separation in quality levels leads even more consumers to switch to Google, which further increases the quality gap, creating a sort of feedback loop.
Significantly, no one questions the fact that Google has invested heavily in creating a high-quality search engine—one that, as Mehta notes, “is widely recognized as the best.” Thus, no one should be surprised when Google Search remains the first choice of most users, even after its default contracts are struck down. This has led to much confusion, both by opponents of the decision, who say it implies antitrust liability was a mistake (it doesn’t), and also by those who want to see Google broken up, who mistakenly suggest that merely terminating Google’s contracts won’t do anything to restore competition (it will).
The fact is, Google would not pay tens of billions annually if its default contracts did not provide an immense competitive advantage—something it cannot get by simply offering a higher quality product. This advantage consists in suppressing the emergence of serious competition over the long run. Indeed, two of the three anticompetitive effects identified by Mehta involve threats to dynamic competition: depriving rivals of sufficient scale and artificially suppressing their incentive to invest in their own search engines. With Google’s contracts eliminated, this suppression of rival investment will be relieved, opening the door for meaningful competition to develop. But it won’t happen overnight.
What remedy?
Mehta’s decision did not determine a remedy. That will be established later, after a hearing. But we can form some educated guesses about what the remedy will be.
Let’s get the big one out of the way: a breakup (forcing Google to divest Android or Chrome) is very unlikely in this case. A modern court is unlikely to consider a breakup unless it believes there are no simpler and less intrusive ways to purge the defendant’s misconduct. But this case centers on vertical contracts that could be easily terminated through an injunction. Moreover, even if you think a stronger remedy is needed, data-sharing (forcing Google to share data with rivals) would surely make more sense than a breakup. A breakup would be largely punitive; data-sharing would actually stimulate competition.
The most likely remedy is an injunction terminating some or all of Google’s contracts. However, the court might permit Google to continue giving itself default status on Chrome. It might even allow Google to continue purchasing some default privileges from other firms, provided that such deals are sufficiently narrow in scope. For example, the court might permit Google to continue buying default status on Firefox, which accounts for only a small percentage of search queries. But its much larger Safari contract is destined for the chopping block.
Note that an injunction of Google’s agreements would not stop Apple or other companies from continuing to make Google the default. It just prevents Google from buying default status. Nor would an injunction prevent rival search engines from entering into default agreements. In fact, allowing small firms or entrants to acquire some limited default privileges may actually be procompetitive, as it can stimulate entry or investment in search quality. However, to avoid legal trouble, such deals would have to be narrow in scope. The logic of the Google decision suggests that no one search engine should have defaults covering a large portion of search traffic.
Some commentators argue that all smartphones and browsers should be compelled to offer users a “choice screen” during their initial setup. This would eliminate defaults by forcing users to choose their preferred search engine. However, Mehta cannot force nonparty firms like Apple to implement a choice screen; his remedy will only bind Google. Additionally, at least in my opinion, a choice screen is not the best solution, because allowing limited default agreements (particularly by small firms or entrants) can be procompetitive, as noted above.
Behavioral economics and “nudges”
As noted above, parts of the Google decision rely heavily on behavioral economics. What does this portend for the future of behavioral economics in antitrust? Behavioral work may well have a greater role to play moving forward—not necessarily because of the Google decision, but because behavioral considerations may become increasingly relevant in future cases. The antitrust agencies have their eyes set on Big Tech. And a number of controversial practices in this space resemble the “nudges” often studied in behavioral economics. They don’t force consumers to choose the defendant’s product; they merely encourage it by introducing subtle biases or frictions. Behavioral economics can help to elucidate the competitive impact of these effects.
Many practices fitting this description are described as “self-preferencing,” a catch-all term for a range of different behaviors designed to nudge consumers toward the platform’s own products. For example, Amazon has been accused of manipulating its search results to give higher placement to its own products.
However, the fact that the DOJ prevailed against Google doesn’t mean that plaintiffs will necessarily be successful in challenging other platform “nudges.” For one thing, while the government had little trouble establishing that Google was a monopolist in search, a self-preferencing platform like Amazon may not have monopoly power where it counts (namely, in the “relevant market” where rivals are allegedly being excluded). Additionally, unlike Google’s default contracts, most types of “self-preferencing” involve forms of unilateral conduct that are hard to challenge under existing antitrust law.
Some forms of self-preferencing may also be hard to police effectively. For example, suppose Amazon is sued by a competing product seller who alleges that Amazon improperly manipulates its search results, giving higher placement to Amazon’s product and lower placement to the rival’s. Implicitly, this allegation suggests that the rival’s product deserves to be ranked higher. But the question of which product brand is most “deserving” of the top ranking may be subjective or otherwise hard for a generalist judge to answer. Nor are courts well-equipped to micromanage search ranking algorithms to ensure they are sufficiently “fair.”
Implications for other Big Tech cases?
The American antitrust agencies are actively litigating several other major cases against Big Tech companies, including Facebook, Amazon, Apple, and even a second case against Google. There is lots of speculation online that the Google decision will have a huge impact on these other cases. The truth is less exciting than that.
In fact, the Google decision is unlikely to make much difference in the other Big Tech cases. First, the other cases involve very different allegations, so Google is unlikely to be a relevant precedent. Second, there is nothing especially transformative in the Google decision—nothing that might produce broad, high-level changes in the antitrust system. Rather than attempting to break new ground, judge Mehta relies very heavily on the “OG” Big Tech case: Microsoft, a 20 year-old decision that set the standard for modern monopolization cases.
Still, one aspect of the Google decision may provide a relevant preview of what is to come in the Apple case. One argument pressed by the plaintiff states (but not the DOJ) is that a Google subsidiary, SA360, declined to provide Microsoft’s advertising platform with the same features that it provided to Google’s platform. Mehta concluded that this allegation fell under antitrust’s “refusal-to-deal” doctrine, which the courts have largely killed off in recent decades. As such, Mehta rejected the states’ argument. Some aspects of the DOJ’s complaint against Apple are facially similar to the states’ allegations. For example, the DOJ alleges that Apple withheld features from third-party smartwatches and messaging apps. Apple will argue that these claims should be rejected on the same legal grounds that led Mehta to reject the states’ arguments about SA360.
Defining Platform Markets
Search engines are two-sided media platforms. They cater to two separate but interdependent groups—search users and advertisers. In American Express, the Supreme Court held that platform markets must sometimes be defined so as to include both “sides” of the platform. That can raise significant practical difficulties, particularly if the defendant’s conduct harms one side and benefits the other. However, the Court held that this rule should not apply in cases where indirect network effects are strong in only one direction. It cited ad-supported media platforms as an example of this. Thus, although Mehta’s analysis consistently considered both groups, he identified separate markets for search users and advertisers.
Many antitrust scholars argue that this approach should be the rule, not the exception. That is, they believe that courts should almost always define separate markets for the two sides, albeit while still accounting for their interdependence. The Google decision provides a nice case study of this “separate markets” approach. Hopefully courts will take note of its advantages over the single-market rule endorsed in American Express.
Author Disclosure: the author reports no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
