Ankwei Chen
The US Department of Justice (DOJ) filed a complaint (“Complaint”) against Google at the US District Court for the District of Columbia on October 20, 2020, in which it alleged Google of unlawfully maintaining a monopoly in search services and advertising in relation to its search services. As this action represents a rare instance of major antitrust litigation that the DOJ has been shying away from in recent years, media commentators have already remarked on the potential for this to be the next landmark antitrust case against a tech giant in almost twenty years after US v. Microsoft. However, based on the Complaint, there are also some inherent differences that may complicate how the DOJ can establish its case. This article will provide a brief summary of the Complaint.
The DOJ alleged three Sherman Act Section 2 claims against Google in this complaint:
- Illegally maintaining a monopoly in general search services (in the US);
- Illegally maintaining a monopoly in search advertising (in the US); and
- Illegally maintaining a monopoly in general search text advertising (in the US).
Sherman Act Section 2 prohibits as a felony the monopolization or attempted monopolization of any part of trade or commerce among the US or with foreign countries. As there is no attempted monopolization allegation in the Complaint, only the jurisprudence on unlawful monopolization is relevant. The offense of monopolization under Section 2 contains two main elements: (i) the defendant has monopoly power in the relevant market, and (ii) the defendant has acquired or maintained such monopoly power through growth or development that is not based on “superior product, business acumen or historic accident.” The US courts have historically interpreted Section 2 as only condemning monopoly power that came as the result of conduct that has an “anticompetitive effect”, which is harmful to the competition in that relevant market, and by extension, the consumers of that relevant market. If a plaintiff has demonstrated both monopoly power and anticompetitive conduct, the defendant would have the burden of proof to show that the alleged anticompetitive conduct had procompetitive justifications, which, if unrebutted by the plaintiff, will be evaluated against the anticompetitive harm to see whether the benefits outweigh the harms alleged.
The DOJ cites to the same set of factual allegations for all three of its claims in the Complaint, which are structured along the aforementioned elements of a Sherman Act Section 2 claim:
Monopoly Power in the Relevant Market
In defining the relevant market for Section 2 claims, the US courts primarily looks at demand substitution in the designated region, which means that all products that are “reasonably interchangeable for the purposes for which they are produced” within that designated region. As a result, “general search services”, as the DOJ has defined it, consist only of “one-stop shop” search services online that provide access to essentially all information available on the Internet as distinguished from localized or specialized searches on Amazon (regarding products sold on its website), Expedia (travel resources), social media platforms (users and posts), or even offline search tools. For the appropriate geographic market of such general search services, the DOJ limits the scope to the United States in arguing that Google has a US domain that is accessible only to users located in the US and provides optimized results based on such location, which would eliminate Baidu and other general search engines that are generally only used in a particular country or jurisdiction from being considered as a substitute product.
To then establish monopoly power after defining the relevant market, the conventional inquiry is to determine whether the alleged entity has a dominant market share in the relevant market, and that there is significant barrier to entry (which may be caused by the alleged entity) to that market. Under the above market definition for general search services, the DOJ contends that there are only three other entities besides Google that offer general search services in the United States, namely Bing, DuckDuckGo and Yahoo! In what should be unsurprising information even to the general public, the DOJ asserts that Google is by far the predominant entity, with about 88% market share. In contrast, Microsoft’s Bing only accounts for 7%, Yahoo! (which actually uses Bing’s search engine) has about 4%, and DuckDuckGo, a search provider that advertises itself as not storing user data and caring for user privacy, has less than 2%. As for the barriers to entry, the DOJ provides a cursory argument that in the US, only Google and Microsoft have the capability to maintain a comprehensive “search index” built by their webcrawlers, due to the significant capital investment, technological level and scale required.
For the other two illegal monopoly claims, search advertising encompasses all advertising shown after a user makes a search, while general search text advertising is a subset of search advertising and generally appears in the form of “sponsored ad” text links that appear within the search results. The gist of the relevant market and monopoly power analyses is similar to that for general searches above. One key feature of search (text) advertising that the DOJ claims is particularly resistant to substitution by advertisement in other media is Google’s ability to provide targeted advertising based on customer search patterns, so the relevant market can only be online searches that are attached to search engines. When combined with Google’s alleged more than 70% market share in the search (text) advertising market, and the significant barrier to entry, because a competitor will need to construct a search engine to even serve the advertising results as well as maintain/develop it into a sufficient scale so as to make it a financially viable source of income, the DOJ thus concludes Google also have monopoly power in those relevant markets as well.
Unlawful Abuse of Monopoly Power
As mentioned above, the next element of a Section 2 claim is to establish how the accused monopoly has been engaging in conduct that has an anticompetitive effect in the relevant market. The meat of the allegations in the Complaint thus focuses on how Google has been allegedly playing a long game of using “defensive” exclusionary agreements to foreclose competitors from effectively challenging its dominance in search.
In particular, Google’s Android operating system for mobile devices is characterized in the Complaint as similar to a Trojan horse that was crucial in allowing Google to engage in conduct that has anticompetitive effects. It was offered as open source by Google in 2007 that is free for everyone to make their own versions, which are known as “forks” in open-source. Given that Google had already gained significant renown for its search engine in the early to mid-2000’s, Android was able to initially attract the attention of some major mobile distributors, which in turn started a positive cycle of attracting more developers to build apps for the platform and more end consumers looking to use Android. Google was further able to accelerate Android’s growth by offering mobile distributors a share of its search advertising revenue if they use Android for their phones (along with certain key conditions as detailed below). As a result, Google was incredibly successful in more than tripling its market share of licensable mobile operating systems to over 80% in only a few short years of effort, thereby effectively making the US mobile operating system market a two-horse race between Android and Apple’s iOS (which is not licensable) for more than a decade.
To exploit that wide user base in the protection of its search monopoly, the Complaint alleges that Google needed a way to retain some measure of control over an open source operating system that they are giving away for free. This was done by requiring mobile distributors to enter into three categories of exclusionary agreements with Google in exchange for a license to use the Android operating system for their phones and a share of Google’s search revenue: (i) Anti-forking agreements, (ii) pre-installation agreements, and (iii) revenue sharing agreements.
Anti-Forking Agreements
Anti-forking agreements are intended to limit the scope of modifications mobile distributors may make to Android, with Google allegedly having last say in determining what changes constitutes compatibility with Google’s technical standards. As case in point, the Complaint brings up the general failure of Amazon’s efforts in around 2014 to launch its Fire OS operating system as an Android fork that significantly deviates from Google’s accepted version, in particular the use of Microsoft’s Bing as the default search engine. The Complaint attributes Fire OS’s market failure to reluctance from other manufacturers to use Amazon’s Android fork for their phones, as they perceived a risk of potentially violating their own anti-forking agreements with Google.
Pre-installation Agreements
What is then so important about Google enforcing its own vision of what the “standard” Android phone should look like through the anti-forking agreements? The Complaint next alleges the exclusionary effects of Google’s pre-installation agreements (Mobile Application Distribution Agreements, “MADAs”). MADAs are used to ensure that all Android phones have Google’s six core apps appear front and center on the user interface of the device, which the Complaint lists as Google Play, Google Chrome, Google search, Gmail, Maps and YouTube, and that the consumer may not delete them like they could with third-party apps. In other words, if the distributor would like to include Google’s app store Google Play, which gives customers access to millions of other third-party Android apps, it would have to include all the other five Google apps regardless of whether the distributor or its customers would prefer otherwise. With the anti-forking agreement and the MADA, Google thus appears to have locked down virtually every single Android phone for sale to the consumer to look the way it wanted, with Google apps and Google search being permanent fixtures on the home screen. The placement of Google search widget on the home screen is allegedly an unnegotiable part of the MADA – the Complaint alleges that Google has rejected all attempts to waive the Google search widget placement requirement. As a result, competitors are prevented from effectively presenting their alternative products, since mobile distributors would have no reason to confuse their customers by simultaneously pre-installing two search tools on the home screen.
In addition, and perhaps more technically important from a mobile distributor standpoint, MADAs give access to important Google APIs (application programming interfaces, which enables apps to communicate with other apps, e.g., use the location data to provide localized information, as well a the operating system and the device hardware itself), which are intentionally kept out of Android’s source code but are essential for many apps’ basic functionality. Mobile distributors thus had a very strong fear of pushing back against Google’s pre-installation demands, and the Complaint provided examples of Google warning unnamed manufacturers that any effort to replace the Google search widget would be considered a violation of the pre-installation agreement, which would then cause the manufacturers to lose access to the aforementioned Google apps and APIs.
Revenue Sharing Agreements
Finally, as part of the incentive for mobile distributors to receive a share of the search advertising revenue, the revenue sharing agreements (RSAs) that the mobile distributors enter into with Google restrict them from using any search other than Google search, thereby covering any potential overlooked search points not already covered by the MADAs, including any new ways to search (e.g., voice searches) that does not interact with the Google search app or Chrome. However, the mobile distributor must have already executed a MADA with Google before it is eligible to enter into a RSA, and the RSAs either require all Android phones across all models sold by the distributor to comply with all the other Google-mandated exclusivity requirements, or all units within specific models. Furthermore, Google allegedly makes the RSAs have retroactive effect, so in other words, if a mobile distributor decides to not renew the RSA after a 2-3 year term, then not only will it lose out on the search revenue share for every new mobile device after the expiration of the RSA, but all devices that have already been sold to consumers; a competitor would thus have to offer more money than Google for all new devices as well as enough to offset the loss of Google revenue from all the sold devices for it to be an attractive alternative.
Outside of the Android ecosystem, Google has covered the iOS market by entering into a RSA with Apple where Google also pays Apple in exchange for making Google search the default on all of Apple’s apps, such as the Safari browser and Siri. The Complaint alleges that the billions paid by Google to Apple for such exclusivity on Apple devices again makes it virtually impossible for a competitor to dislodge Google. For searches outside of the mobile space in the US, which are predominantly browsers on PCs, other than Chrome and the aforementioned Safari, Google also pays Mozilla Firefox a share of its advertising revenue to set Google search as the browser’s default search. The Complaint further warns that Google already has plans to adapt its existing agreements to cover other searches on devices other than mobile phones or PCs – Google allegedly already has contracts with partner automobile manufacturers to not install competing search apps, and Google’s operating system (“Wear OS”, which is an Android fork itself) for wearable devices requires the manufacturer of such devices to refrain from installing any third party apps. The DOJ thus asserts that “Google is poised to ensure…that all search access points funnel users in one direction: toward Google.”
Preliminary Analysis
Even though Google is expected to contest every assertion made by the DOJ regarding Google’s monopoly power in the relevant market, it may be difficult to convince a court otherwise, assuming that the DOJ’s general search engine data comparing the market shares of Google with Bing, Yahoo and others are accurate. In the DOJ’s characterization of “general search”, an input keyword is internally assessed and compared with a database of webpages whose data were previously collected by automated “webcrawlers”, and returns a list of webpage results of what the search engine considers to be the most relevant to the keyword input. Although online searches of all kinds are based on roughly the same mechanics, the DOJ appears to be making an distinction with “general search” operators that collect information from all over the Internet and return links to information instead of information themselves or products, i.e., they can respond to “navigational queries” in addition to informational and transactional queries. Under such definition, there does not seem to be great room to argue that there are any other notable general search operators in the US that the DOJ has failed to take into consideration. Google is likely to attempt to argue that the DOJ has not actually limited general searches to navigational queries only, and even the so-called specialized search providers, especially those on major social media platforms (e.g. Twitter, Reddit), now contain sufficient navigational query functionality that outperforms Google and should thus be included in the relevant market as reasonable substitutes to better reflect how the people’s surfing habits have changed. However, it is somewhat in question whether there is persuasive data that show a sufficient change in the surfing habits of the general US public for there to be a critical mass of users that regularly use certain specialized search provider(s) for their everyday online searches instead of Google and other “general search engines”.
The strongest parts of the DOJ’s case at this stage appear to be its allegations of the extent of control Google imposes on Android through its exclusionary agreements in order to ensure Google search’s supremacy in the mobile device sector (and beyond). Here, several parallels can be clearly drawn with the anticompetitive conduct of Microsoft in the US v. Microsoft case. In that case, Microsoft was famously accused of abusing its monopoly with its Windows operating system by bundling its browser Internet Explorer along with Windows, while taking exclusionary legal and technical measures to prevent competing browsers, in particular Netscape Navigator, from being used or installed alongside in violation of Sherman Act Section 2. If it seems odd that the central issue concerned a browser when the monopoly is the operating system, it is because at the time, it was thought that browsers have their own APIs for software development and are operating system-agnostic (i.e., does not depend on the underlying OS), which may cause developers to write software using those APIs instead of Windows’ if a critical mass of users was reached, thereby eventually supplanting Windows entirely. Microsoft thus bundled Internet Explorer with Windows to increase the user base for IE while (i) prohibiting OEMs from installing competing browsers with the copies of Windows provided with their computer products “to prevent customer confusion”; (ii) integrating Internet Explorer into Windows 98 so as to make it difficult to remove without breaking certain functionality; (iii) offering Internet Explorer free of charge to “Internet Access Providers” (a term that combines internet service providers and now defunct “online services” such as America Online that provide dial-up Internet access prior to widespread broadband) and paying a bounty for each customer that signed up for such online services to use Internet Explorer; and (iv) entering into deals with software vendors for their software to use Internet Explorer to browse the Internet in exchange for preferential support or quality certification from Microsoft, as well as forcing Apple to stop installing Netscape Navigator as a standard with Mac OS. All of the above conduct was found to be anticompetitive via restricting rivals from effective distribution of its competing products, and Microsoft’s proffered procompetitive justifications were either rejected or deemed as outweighed by the anticompetitive harm alleged.
The parallels with Google’s alleged conduct are clear to see. The pre-installation agreements with Android phone distributors are extremely similar to Microsoft’s exclusionary agreements with OEMs, and the anti-forking agreements essentially makes the version of Android distributed on Android phones under Google’s control as if it had been a closed, proprietary operating system like Windows. Google’s payments to other browsers for Google search to be the default search option also echoes Microsoft’s conduct with software vendors in encouraging/mandating Internet Explorer use. However, one potential stumbling block for the DOJ is that anticompetitive effect is measured based on the level of consumer harm resulting from the restricted competition. With harm from pricing irrelevant, the DOJ characterized the harm to consumers in the Complaint as a reduction in the quality of general search services (including in terms of personal privacy) and a lessening of choice by Google denying scaling to potential newcomers. The issue is that if the performance of a search engine is based on the ability to deliver relevant results to the user, which the DOJ claims to be almost entirely due to scale, as scale improves the underlying algorithms used to assess relevancy of results versus the keyword input, then by definition the consumer cannot be said to be harmed by a search engine operator that seeks to expand its scale to the greatest extent possible; to what end would a consumer be served with additional choice in general search services if they cannot offer the most relevant results? On the other hand, Google’s poor handling of user privacy may be attributable to a lack of competition; if DuckDuckGo were to attain comparable scale to Google in a general search market with normal competition, Google may be compelled to improve its capability to maintain user privacy in order to compete (although whether DuckDuckGo could attain a comparable scale without making a tradeoff on user privacy for advertising purposes is an entirely different question). As a result, the DOJ will likely have to put in significant effort to show the court that consumers are being objectively harmed when Google’s monopoly scale is likely the reason it is providing the most relevant search results possible.
As for the procompetitive justifications that Google can raise in response to the DOJ’s anticompetitive allegations should the analysis proceed to that point, Google’s only public comment so far is that its agreements are no different from those in every other industry in which competitors have to compete for “eye-level shelf space”, and that both Bing and Yahoo are also paying Apple and others to be featured on their devices. While a widespread “everybody else is doing it” practice by big and small players alike in a relevant market is a strong indication that the practice is procompetitive, it is not clear how such argument would apply to the anti-forking agreements, which has no equivalent in the relevant market, or that Bing and Yahoo’s payments are anywhere comparable in scale or exclusionary effect as Google’s RSAs. Furthermore, in light of Microsoft, Google will have to be careful in distinguishing a justification that the conduct was for a lawful purpose versus a justification that the conduct was procompetitive. For example, a potential argument can be made that the anti-forking and pre-installation agreements were for ensuring that Android phones offer a level of desired security, usability and performance envisioned by Google that are free from potential faulty customizations made by the mobile distributor that may unfairly affect customer opinion. While these goals may be considered as lawful reasons for such exclusionary agreements, it is more questionable that they are done to strengthen Google’s competitiveness in search instead of being a pretext to prevent competing searches from taking the spotlight; in fact, Microsoft had advanced a similar argument regarding Windows in justifying its OEM licensing restrictions with mixed results. Unless the facts about the substantively change from what the DOJ is alleging, it seems that Google may encounter difficulties in proffering procompetitive justifications for its exclusionary agreements.
In an interview given a week before the submission of the Complaint, former Microsoft chairman Bill Gates remarked that he “had been naïve about government scrutiny” when Microsoft was enjoying significant commercial success in the late 1990s. Gates is not expecting the same naïveté from Google et al. in dealing with antitrust today. Even at this early stage, the DOJ’s case looks to be a formidable legal challenge to Google that has the potential to dramatically alter how Google’s core businesses will be run, and how this case is being handled once the political situation is stabilized will be highly indicative of the US government’s resolve in pursuing this wave of scrutiny against Big Tech.