Herbert Hovenkamp reviews Epic Games’ lawsuits against Apple and Google for restraining users’ ability to access Epic’s offerings through third-party app stores. A comparison of the two ecosystems sheds light on what remedies would improve benefits to consumers and how the Department of Justice’s own lawsuit against Apple may fare.
Epic Games sued both the Apple and Google-Android smartphone systems in 2020, challenging their policies requiring users to purchase apps through their digital stores, including purchases of game levels, currency, and other virtual items. Apple’s digital store is the App Store, and Android’s is Google Play. In Google Play, the Ninth Circuit handed Epic a major antitrust victory, holding that Google’s policies violated prohibitions under Section 1 of the Sherman Act against anticompetitive agreements that restrain trade. The court did not find the policies to be unlawful refusals to deal under Section 2 of the Sherman Act, which concerns monopolization. It did conclude, however, that compelling Google to deal with third-party app stores could be part of an antitrust remedy for other unlawful conduct. The court required Google to supply a “catalog” of its Google Play offerings so that rivals could set up similar stores.
Epic was less successful in its similar lawsuit against Apple, losing the federal antitrust case but prevailing on a California state law issue. The California state law injunction raises an interesting question about a state’s power to impose law nationwide when it is in conflict with federal law.
Epic brought its cases against Apple and Google under both Section 1 and Section 2 of the Sherman Act. Both decisions relied almost entirely on the agreement claims under Section 1. That fact should concern the United States Department of Justice, which filed a more comprehensive suit against Apple in 2024 entirely under Section 2 for monopolization. That case has not yet gone to trial. Epic’s cases are largely about tying products together—here, tying the platform to the App Store or Google Play. Tying has historically been treated mainly as involving anticompetitive agreements rather than unilateral monopolistic acts. Why the government chose to limit its suit to Section 2 is unclear, unless perhaps it believed that a Section 2 violation might make a breakup remedy more likely.
Can the DOJ win its own case against Apple strictly on claims of monopolistic conduct? Perhaps, but there are challenges. First, the market power requirement for monopolization is higher than it is for agreements that restrain trade. Second, while there is precedent for condemning ties as monopolization, plaintiffs must walk a narrow path. Unilateral ties attack product or system design. For example, the iPhone was coded to accept sales only through its App Store. It also integrates an iMessage feature exclusive to Apple devices and disallows competitors’ wallets. However, that does not mean these ties are anticompetitive and have not been to consumers’ benefit. Smartphones also come with preinstalled cameras that consumers love but which have nearly killed off traditional digital camera sales, which have declined by 90%. Apple founder Steve Jobs was simply following iPhone photographer Chase Jarvis’ maxim that “the best camera is the one that you have with you.” One of the severe difficulties of making policy judgments in the area of tech ties is that the amount of exclusion, or competitor harm, is often proportional to the amount by which users prefer the combination.
The differences between Apple and Google
Defendants often try to recast unilateral tying cases as refusals to deal, or to sell one of the products without the other. Google attempted this unsuccessfully in the Google Play case. The rationale is clear: unilateral refusals come under the Supreme Court’s highly restrictive Trinko decision, which shields nearly all of them from antitrust challenge. In other cases that involve tying by product design, the courts impose a strict “no benefit” rule, which says that innovating a combined product is necessarily lawful unless it provides no benefit to consumers.
Here, Apple has been quite successful. Its “no benefit” argument has won in cases going back to the 2011 iTunes litigation. Apple showed that security and privacy concerns justified its policy of making its iPod player compatible only with its iTunes music database. It prevailed again in Epic Games by arguing that “we are better than Android,” because the Apple system retains tight control over secondary products. Epic’s own CEO testified that he personally used an iPhone rather than an Android because it offers “better security and privacy.” The differences create a two-tier system in which iPhone users pay more but receive a better-performing and more secure product. Opening up app store competition might be a significant improvement for consumers as well as competitive sellers, provided that concerns about security and privacy are adequately addressed. Otherwise, they might simply transform the iPhone into a virus-prone, unstable device.
It can be easy to see most of Apple’s conduct in Epic Games as unilateral because it owns the important parts of the iPhone ecosystem. Apple wrote its own code so as to exclude competing app sellers. One feature of digital products is that nearly anything that can be specified in a contract can also be implemented “unilaterally” through code. As a result, in digital markets, tying-like results, which traditionally require an agreement, must often be addressed as monopolizing conduct under the much stricter standards of Section 2 of the Sherman Act. One irrationality of antitrust law is that these unilaterally imposed ties created by code are actually more stable and enforceable than contractual ties, but the law makes it far more difficult for plaintiffs to win challenges.
In the Apple case, the court found that Apple used “both technical and contractual means” to create its closed system. Its decision was based on the contractual ones. But even if Apple’s conduct were treated as a unilateral refusal to deal, it would be a “secondary” refusal, which is a refusal to sell a secondary product other than the one in which the firm is dominant. Epic was not asking to duplicate the iPhone. Rather, it wanted to avoid the costly App Store. Secondary refusals present problems in vertical integration, similar to the traditional manufacturer who insists on selling exclusively through its own stores or franchisees. Whether they are anticompetitive depends on the firm’s market power, the impact of the exclusion, and the existence of any justification that the defendant may have for its restrictive policies.
Android, in contrast to Apple, is open source. Google controls other firms’ use of it through licensing agreements, which are contracts. Google acts more as a manager, tasked with maintaining operational quality, security, and compatibility across Android’s multiple providers. The requirement that sales be made through Google Play is imposed by licensing agreements with phone manufacturers. The court’s substantive analysis of the case concerned whether the agreements qualified as tying or exclusive dealing under antitrust law’s rule of reason. This required the plaintiff to show market power, which was aided by the court’s acceptance of the jury’s finding of an “Android-only” relevant market that excluded Apple.
By contrast, in the earlier Apple case the court found a relevant market of all mobile game transactions, which included both Android and Apple. Were those two findings inconsistent for purposes of res judicata, which may bind a party to a legal conclusion established in a previous case? Not necessarily. Sometimes a favored technology competes with a less favored technology, but not the other way around. In old commerce, for example, lower-cost Brooklyn bakeries might place competitive pressure on Manhattan bakers, but not in reverse. In any event, claims of res judicata do not always entitle a party to consistent outcomes in two cases raising different issues. The Ninth Circuit held as much in this case.
The Google Play court also required and found adequate evidence of a “significant causal connection” between the plaintiff’s harm and the violation. Causation is also relevant in other pending tech cases, including the government’s case against Google Search and, very likely, its case against Apple down the road.
Private plaintiffs and the government use two different statutes for injunctions. Private plaintiffs such as Epic Games bring actions under Section 16 of the Clayton Act, which entitles them to “injunctive relief … against threatened loss or damage” to themselves by an antitrust violation. By contrast, the DOJ obtains antitrust injunctions mainly under Section 15 of the Clayton Act, which simply authorizes federal courts in government suits to “prevent and restrain” violations. Unlike the private provision, that one has no independent causation requirement. However, under Section 1 of the Sherman Act, the government must still prove that challenged conduct “restrains trade,” or under Section 2 that it “monopolizes.” The most general meaning of these terms is that the conduct causes an output reduction or leads to higher prices. It also reaches things like agreements to reduce innovation or product or service quality. So causation of some kind is relevant in both private and government suits.
Devising remedies
The court in Google Play followed a brief written by the Federal Trade Commission and DOJ, arguing that digital markets should receive more aggressive treatment because “the particular characteristics of digital markets … can allow monopolists that achieved or maintained dominance through exclusionary conduct to perpetuate entry barriers and maintain monopoly power long after that conduct has stopped.” Neither the court nor the antitrust agencies offered empirical support for that assertion, and it seems wrong as a general matter. Indeed, the facts surrounding one of the cases it cited in support, Microsoft, suggest just the opposite. Once the court condemned Microsoft’s exclusivity requirements favoring its own internet browser, the market responded quickly.
Increasing the output of a digital product, such as an eBook, is generally far easier than increasing the output of a traditional product, such as a paper book, and there are typically no capacity constraints. In addition, networks are held together by contracts rather than ownership. Network participants are independent firms with their own competitive incentives. A corporation’s owned subsidiaries never achieve that independence short of divestiture. Even after an order terminating exclusive sales, they still own the subsidiaries. If anything, digital markets are more susceptible to rapid defection and reorganization once anticompetitive restraints are removed.
This idea that digital markets cannot respond quickly to remedies limiting exclusivity reflects an unwarranted anti-tech bias, notwithstanding its more rapid growth and innovation. Digital markets, if anything, respond more quickly to properly drafted nonstructural remedies. Given the degree of product mobility, non-exclusivity is the key. Indeed, numerous “alternative” app stores are already appearing at the time of this writing, just months after the Apple judgment. That process could take much longer in a market where new plants have to be constructed or existing plants enlarged.
On the standards for a remedy, the Ninth Circuit has equated its position with that of the D.C. Circuit’s Microsoft decision: “the available injunctive relief is broad, including to ‘terminate the illegal monopoly, deny to the defendant the fruits of its statutory violation, and ensure that there remain no practices likely to result in monopolization in the future.’”
With Microsoft as a template, the court’s Google Play remedies properly targeted exclusivity. The injunction (1) prevented Google from paying Android implementers to make the Google Play store the exclusive venue for purchasing; (2) prevented agreements to launch apps either first or exclusively on the Play Store; (3) prevented agreements to preinstall the Play Store but no other store at favored locations on Android devices; (4) prohibited the defendant from requiring Google Play Billing for all apps distributed on the Play Store; and (5) required Google to permit third-party Android app stores “to access the Google Play Store’s catalog of apps” so that they could offer a comparable library. The court also acknowledged Google’s continuing role in providing “security measures” for the Android system that are “comparable to the measures Google is currently taking for apps listed in the Google Play Store.” It held that Google may charge a “reasonable fee” for these measures.
So what might happen next, assuming no Supreme Court review? Will the Android and iPhone systems perform better as a result? The metrics for assessing an antitrust remedy are reasonably clear: Will the user experience be better, will prices be lower, or will output or innovation by some reasonable metric be higher? Google has the means to monetize its Android role other than exclusive placement on Google Play. It can still obtain revenue from other ancillary products such as Chrome or YouTube, as well as cloud and subscription services. Nevertheless, loss of Google Play exclusivity would be a substantial hit, as would Apple’s loss of an exclusive App Store. A best-case scenario is that the systems are better and cheaper for millions of users and also competing app developers, even though Google and Apple earn less.
Did the remedy in the Google Play case do enough? More aggressive ones are possible. In network cases, courts should look for remedies that can increase competition without breaking up the network, which would be disastrous. Google provides its management, security, and compatibility services vertically, from the top down. But could this also be accomplished horizontally, by agreement among participating phone producers and carriers themselves, who are individually nondominant? This need not require Alphabet to divest Android, but rather to give more control to its phone manufacturers. Challenged practices could be addressed as agreements that restrain trade. However, each phone-producing participant could individually set its own policy with respect to things such as app stores and their prices, offering them in competition with one another. The history of antitrust has an ample record of competing businesses with separate interests that collaboratively operated platforms or other shared markets. These include the Terminal Railroad case, Chicago Board of Trade, Associated Press, and NFL Properties. The members would have every incentive to make the network perform as well as possible, while competing on business aspects such as pricing, app store exclusivity and commissions, and other policies. Each could individually charge whatever commission it wished for app sales, although an agreement about them would be unlawful. Such a remedy could preserve most of the advantages of network size and variation but ensure greater competition within the network.
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.
Subscribe here for ProMarket’s weekly newsletter, Special Interest, to stay up to date on ProMarket’s coverage of the political economy and other content from the Stigler Center.
