In a new report from the Knight-Georgetown Institute, Alissa Cooper, Jasper van den Boom, and Zander Arnao examine how to make remedies most effective in the Google Search antitrust case. They argue that restoring competition in online search requires a comprehensive package of remedies that takes into account the multiple levers by which Google Search built, maintains, and could rebuild its monopoly.
On August 5, 2024 Google was found liable for violating Section 2 of the Sherman Antitrust Act by illegally monopolizing the markets for online search and search text advertising. The trial hinged on a series of agreements through which Google paid various tech companies tens of billions of dollars to make Google Search the default search engine on their devices and applications.
The case has now entered the remedies phase, with the Department of Justice and plaintiff states set to propose a package of remedies to the court by November 20th, which reporting indicates will likely include data-sharing, a separation of the Android operating system from Google’s core products, and divestiture of Chrome. This proposal represents a historic opportunity to promote competition in key digital markets, and it is imperative that the set of proposed remedies be comprehensive and seek to ignite competition.
A comprehensive, multi-faceted remedies package is essential
For at least 15 years, Google has maintained the default status of its search engine on a dominant set of search access points and thereby monopolized key distribution channels in the markets for general search and search text advertising.
This sustained advantage has had cumulative effects: competition has been virtually eliminated as consumers formed habits around the use of Google Search and grew to associate it synonymously with online search. Google grew its user base such that most queries are directed to its search engine, providing it with rich data that it has used to improve its product in ways that competitors could not. All of this means that restoring competition in online search must go beyond simply halting Google’s conduct that has so far foreclosed its rivals.
As Steven Salop has previously argued in ProMarket, merely enabling a rival search engine to appear alongside Google or even to replace it as the default on some search access points will be insufficient to restore competition. It would take time and significant investment for Google’s rivals to build brand value, improve quality, and be perceived as compelling alternatives by consumers and advertisers, who have developed deep experience and familiarity with using Google. Inertia will continue to guide consumers and advertisers’ choice of search engine, raising barriers to switching that dampen the prospects of rivals and potential entrants.
To ensure the effectiveness of remedial interventions, the court needs to enable competing search engines to quickly achieve parity with Google’s quality and reputation.
To realize this goal, any remedy package must recognize the “interlocking, mutually reinforcing nature” of the market conditions in which Google and its rivals operate. No search engine can operate without access to a web index and algorithms to rank and return results to users, and those results cannot improve in quality without user data. Nor can they reach users who generate that data without access to distribution channels. None of this can take place without a search engine monetizing the results to make its search business viable.
The Google Search remedy package must improve distribution, data access, user choice, and public education. These remedies will be more effective in concert, and the absence of any one would undermine the overall package’s ability to restore competition.
Contractual restrictions for all search access points
Google’s exclusive contracts are central to the findings of US v. Google. As a result, a wide variety of potential remedies have been suggested to end or limit them. These contracts enable Google to share its advertising revenue with partner technology companies, through which it pays them for giving Google Search default status and other advantages. Remedy proposals notably vary in their approach to addressing this revenue sharing —with some calling for a cap while others urge an outright prohibition.
Whatever form contractual restrictions take, it is imperative that they apply across all search access points. For contractual restrictions to be effective, Google must be unable to pay or otherwise incentivize its current partners to give Google Search default status or other preferential treatment. This should extend to all access points that rely on Google Search assets, even if the product to which they provide access is not branded as a search product.
For example, if Google’s artificial intelligence models rely on its web index or search data for training, testing, retrieval-augmented generation (RAG), or other purposes, those should be considered search access points to which remedies apply. This would ensure that Google cannot engage in similar anticompetitive conduct merely because the underlying technology of its products has changed. This approach will help maintain the remedies’ effectiveness even as the user-facing experience of what we currently understand as “search” continues to evolve, and to prevent restrictions from circumvented by Google’s product changes and branding choices.
If structural remedies for Android and Chrome are adopted, they must be accompanied by behavioral remedies
Several proposals have suggested structural remedies for Google’s Android mobile operating system and Chrome web browser due to concerns that the company could leverage them to undermine the goal of ensuring competitors have equal access to channels of distribution.
For both products, any structural remedy should meet two objectives: eliminate incentives to favor Google Search and make it possible to observe attempts to evade compliance. To eliminate incentives for favoritism, however, structural remedies would need to be coupled with behavioral remedies that prevent recreating the ties between Google Search, Android, and Chrome that exist today. Structural and behavioral remedies should not be viewed as alternatives, but as a package that only works if both are adopted together.
In the proposed scenario where a remedy requires Google to divest Android or Chrome entirely, different types of behavioral restrictions would be necessary to effectively stimulate competition. Contractual prohibitions restricting Google’s agreements with a structurally separated Android or Chrome would obviously be crucial. Line of business restrictions on both entities would also be required until competition is restored so as to avoid Google moving back into the mobile operating system or browser market and the separated entity moving back into the search market.
Both types of restrictions need to be carefully crafted to account for changes underway in the search market as a result of generative AI. Prohibitions concerning search should be scoped to include products and businesses that rely on search assets even if they are not branded as search. Google is already offering preferential access to its own on-device AI model on certain handsets. Similarly, Chrome has already built in preferential access to Google’s proprietary AI models, and work is underway to broaden access to browser APIs that would drive ever more traffic to those models.
Effective data access remedies must be precise and tailored to market conditions
Finally, a wide variety of remedies have been proposed to give rivals access to search data accumulated and developed by Google, including the web index, click and query data, ranking signals, search results, and ads. Many of these proposals aim to allow search rivals to quickly achieve a higher quality search experience than has heretofore been possible as a result of Google’s illegal conduct.
Any data-access remedy that seeks to advance this goal must be precise about what data should be made available, in what formats, and at what time scales. The absence of these specifications has hindered the utility of click-and-query data sharing under the Digital Markets Act (DMA) in the European Union, which requires the sharing of ranking, query, click and view data in relation to free and paid search. The DMA’s vagueness on this requirement has been exploited by Google in numerous ways; as a matter of policy, Google excludes most queries on privacy grounds, makes the data available months after it has been collected, and charges high prices for access, limiting its utility for rivals looking to improve their own quality or indexes.
Should the court choose to adopt a query data-sharing remedy in this case, the technical committee (discussed below) should receive proposals for privacy-preservation schemes from relevant stakeholders. This body could solicit input from potential licensees of the data, experts in privacy-preserving data sharing, user and civil society representatives, and other interested parties. Rather than simply balancing privacy and competition interests, this process should seek to arrive at solutions to safeguard consumer privacy while realizing to the extent possible the competitive benefits of query sharing.
Robust enforcement must include a technical committee and performance benchmarks
The challenges of enforcing antitrust remedies effectively in the tech sector are well understood. The dynamism of digital markets, network effects, and information and expertise asymmetries between the enforcer and the tech firm have created hurdles to enforcement globally. To surmount these barriers, the court should ensure that any remedies package is modifiable and enforced by a monitor and technical committee.
A technical committee and monitoring body must be staffed and resourced to take on functions more typical of tech companies and expert regulators. These functions include obtaining stakeholder input, issuing design guidelines, adjudicating complaints, assessing compliance, and analyzing the privacy implications of data sharing.
In the past, courts in antitrust cases have focused on stopping the conduct found to be anticompetitive, which focused enforcement on the assessment of compliance. A more promising approach would be directly assessing whether the remedies are achieving the goals of the litigation. The court should adopt specific performance benchmarks, such as the share of queries being received by each competitor, against which the effectiveness of remedies is measured, and if market conditions are not found to be improving, remedies should be adjusted and augmented.
Conclusion
For too long, Google’s monopoly power has constrained how people find, access, and interact with information online. The remedies phase of this historic antitrust case presents a rare opportunity to ignite competition in search and search text advertising. With a comprehensive package of remedies, online search could enter a new era.
Authors’ Disclosures: Zander Arnao reports no conflicts of interest. Alissa Cooper is a board member of The Tor Project, Inc. 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.
