As most readers are likely aware of, on 10 September 2021 Judge Yvonne Gonzalez Rogers (“YGR”) issued her Rule 52 Order after trial on the merits in Epic Games, Inc. v Apple Inc. Based on the trial record, Judge YGR could not ultimately conclude that Apple is a monopolist under federal or state antitrust laws, hence most of Epic Games failed. Even so, the Judge found that Apple’s anti-steering provisions are anticompetitive under California’s competition law and issued a US-wide injunction barring Apple from prohibiting developers to steer users to out-of-app purchasing options.
This is a significant judgment which is expected to be intensely discussed across the world. While at first sight it may give the impression that Apple has scored a key victory, the reality is much more nuanced. We explain below why.
By way of background, it is reminded that Apple obliges app developers considered as selling “digital” goods or services to use its IAP in-app payment solution to accept user payments and pay Apple a related commission of up to 30%. In August 2020, Epic Games (the developer of Fortnite) secretly activated a “hotfix” which resulted in Fortnite bypassing IAP and offering users the ability to purchase V-Bucks (a virtual in-game currency) directly from Fortnite at a lower price. Apple reacted immediately and removed Fortnite from the App Store.
The next day Epic Games filed a lawsuit against Apple in the U.S. District Court for the Northern District of California alleging a series of anti-competitive restraints and monopolistic practices in the markets for (i) the distribution of apps on iOS devices and (ii) the iOS in-app payment processing market. Epic Games also sought a temporary restraining order (TRO) and a preliminary injunction (PI) to have Fortnite reinstated in the App Store, both of which Judge YGR denied (the Judge did grant a TRO and PI with respect to Epic Games’ Unreal Engine, a game engine).
In May 2021, a 16-day trial took place where top executives of both companies testified before the court. Last Friday, Judge YGR finally handed down her judgment.
The relevant market
As in most antitrust cases, the definition of the relevant market was a key sticking point between the litigants. Epic Games put forward a foremarket/aftermarket theory per Eastman Kodak, the foremarket comprising smartphone operating systems (where iOS is present), from which two single-brand aftermarkets flowed, namely (i) an iOS app distribution market; and (ii) an iOS in-app payment processing market. Apple, on the other hand, claimed that the relevant market is one of digital game transactions, comprising game transactions on all platforms, thus including Xbox and PlayStation. Judge YGR disagreed with both parties and eventually opted for a market of “digital mobile gaming transactions” comprising game transactions on iOS and Android.
(i) The foremarket/aftermarket put forward by Epic Games
Judge YGR rejected the foremarket/aftermarket theory of Epic Games on two grounds. First, she held that the identification of a foremarket for iOS was artificial, since iOS is not licensed to anyone; competition exists for smartphones, not the operating system. [One could query whether this issue would have been overcome had Epic Games gone for a foremarket for smartphones, instead of smartphone OS]
Second, the Judge held that the Eastman Kodak conditions for finding a single-brand aftermarket (namely, an iOS-specific aftermarket) were not met, as Epic Games did not prove that competition in the foremarket (between iOS and Android) failed to discipline Apple’s position in the aftermarket.
- On switching costs and lock in: The Judge discounted internal Apple emails where top Apple executives discussed how they could make their products stickier so that consumers would not switch, holding that they were not evidence of switching costs themselves. The Judge then dismissed entirely Dr. Susan Athey’s expert testimony, holding that she did not provide evidence on the magnitude of the switching costs, and failed to examine whether low switching was the result of customer satisfaction or lock in. Based on evidence put forward by Apple, the Judge eventually held that switching between iOS and Android is low, but this is the result of satisfaction with existing devices, rather than any lock in (page 51). Judge YGR thus concluded that Epic Games did not prove that users are locked in or would not switch to Android devices in response to a significant change in game app prices, availability, or quality (Id.).
- On substitutes: the Judge held that the data provided by the parties on user substitution patterns was not conclusive and proceeded without resolving the issue on the record. The Judge was also critical of Dr Evans’ (expert witness of Epic Games) use of a SSNIP test to confirm that iOS app distribution is a relevant aftermarket.
Against that background, the Judge held that the App Store is in the same relevant market with Android app stores with respect to game transactions (more on which below).
As for the aftermarket for iOS in-app payment processing, its identification was conditional on the identification of an aftermarket for iOS app distribution (hence it was also rejected by the Judge).
This part of the judgment is not without controversy, and I would not be surprised if competition authorities and other courts were to arrive to different conclusions.
Indeed, when faced with a similar argument put forward by Apple earlier this year (namely that competition in the primary market for smartphones was sufficient to deprive it of a dominant position in the secondary market for app distribution), the French Autorité de la concurrence reached the opposite conclusion, noting that the conditions laid down in the Pelikan/Kyocera decision of the European Commission were not met; a consumer’s choice of a mobile device is primarily influenced by its price, not app expenditure (see paragraphs 113-115 of the Autorité’s decision). The Autorité also held that “it is unlikely that iOS users will switch to an alternative offering on the primary market, in the event of price rises or a degradation of the quality ofapps in the App Store, since such a change would imply, for some iOS users, the loss of their investment in Apple’s ecosystem.”
Against this background, it is rather peculiar (and inappropriate dare I say) how Judge YGR makes some sweeping statements later in the judgment, such as that the App Store competes for both app users and developers with Google Play (page 132). But the Judge offered no basis for this broad statement; it simply proclaimed it in an axiomatic manner. Where is the evidence of such competition? When Google raised a similar argument in the context of the Google Android investigation, the European Commission rejected it as unsubstantiated; it held that any competition between Apple and Google at the level of app users and developers was insufficient.
Now, I understand that the outcome in Epic Games is different, in that Epic Games had the burden of proving that Apple’s position in app distribution was not constrained by competition in the primary market for smartphones. Yet saying that this particular plaintiff failed to discharge its burden of proof is one thing (it means that based on the evidence presented by the parties, the Judge could not accept the single-brand aftermarket on the basis of Eastman Kodak); going as far as proclaiming that Apple competes with Google for app users and developers in the absence of any evidence is another thing, and an overreach in my view.
(ii) The market put forward by Judge YGR (mobile game transactions)
Judge YGR went for a market for mobile game transactions, which is close to (but significantly narrower than) Apple’s proposed market for digital game transactions.
Taking as her starting point the fact that the App Store is a two-sided transaction market, the Judge went on to narrow down the scope of the market to game transactions (as opposed to non-game transactions), explicitly excluding subscription apps such as Spotify and Netflix. According to the Judge, such a focus on game transactions was justified by the fact that game transactions overall accounted for a very significant part of Apple’s App Store revenues; thus, “in most economic ways, and in particular with respect to the challenged conduct, the App Store is primarily a game store and secondarily an ‘every other’ app store” (page 124). Judge YGR also cited certain additional reasons why game transactions should be considered as belonging to a separate market.
This part of the judgment is also not without controversy; following the logic of the Judge, one would have to define separate markets for different categories of apps, despite the fact that all categories of apps are affected by the same restrictions of Apple.
The next question was whether the market comprised all digital game transactions, as Apple argued. The Judge disagreed, holding that there is a distinct submarket for mobile gaming [note: the Judge also referred to three other submarkets, namely PC gaming; console gaming (comprising gaming on consoles, namely Xbox, PlayStation and Nintendo Switch); and cloud-based game streaming]. Judge YGR relied, among others, on internal Apple documents viewing mobile gaming as distinct, expert opinions submitted to the Court, as well as views of industry participants such as Microsoft.
The Judge did not further segment the market according to the OS, most likely because she rejected the foremarket/aftermarket theory of Epic Games.
Having defined the market as one for mobile game transactions, the Judge estimated Apple’s market share fluctuated between 52% and 57% over the three years in evidence. Judge YGR held that Apple had considerable market power, which nevertheless did not amount to monopoly power per Section 2 of the Sherman Act, even when taking into account direct and indirect evidence proffered by Epic Games. Apple was thus held to be “near the precipice of substantial market power, or monopoly power, with its considerable market share” (page 139). Now, this is not surprising, considering how Judge YGR defined the market as comprising both the App Store and Android app stores. In such a setting, it is unlikely that one of the two duopolists will be found to have monopoly power.
In its complaint, Epic Games had made a series of claims based on Sections 1 and 2 of the Sherman Act, as well as California law.
Before examining the validity of these claims, Judge YGR discussed the anticompetitive effects and the business justifications of the impugned conducts, i.e., Apple’s restrictions on app distribution and in-app payment processing on iOS devices.
(i) App distribution restrictions
Judge YGR identified the following anticompetitive effects of Apple’s app distribution restrictions:
- Foreclosure of competition: the evidence suggests that Apple’s restrictions foreclose competition for large game developers who have well-known games, as these developers “would likely, and have the resources to, open their own stores to forego Apple’s ‘fees, rules, and review.’”
- Increased prices: Apple’s claim that “the 30% rate is commensurate with the value developers get from the App Store” is unjustified. Absent competition, it is indeed “impossible to say that Apple’s 30% commission reflects the fair market value of its services” and “Apple has provided no evidence that the rate it charges bears any quantifiable relation to the services provided.” Apple’s restrictions have thus increased prices for developers (the evidence on whether prices have increased for consumers was mixed).
- Decreased innovation: Apple’s restrictions reduce innovation in core game distribution services. “Apple’s slow innovation stems in part from its low investment in the App Store” and “a third-party app store could put pressure on Apple to innovate by providing features that Apple has neglected.”
The Judge also considered Epic Games’ claim that Apple’s restrictions result in decreased output, but held that the evidence was inconclusive due to lack of an appropriate counterfactual.
On the other hand, Judge YGR identified the following business justifications for Apple’s app distribution restrictions:
- Security, privacy and reliability: The “app distribution restrictions increase security both in “the “narrow” sense, primarily by thwarting social engineering attacks, as well as in the “broad” sense by allowing Apple to filter fraud, objectionable content, and piracy during app review while imposing heightened requirements for privacy”. Even so, the Judge found persuasive that “app review can be relatively independent of app distribution” and “alternative models [notably notarization, the model deployed on Mac, coupled with human review] are readily achievable to attain the same ends even if not currently employed.”
- Intellectual property: While accepting that “Apple is entitled to be paid for its intellectual property”, the Judge noted that “the record is devoid of evidence that Apple set its 30% commission rate as a calculation related to the value of its intellectual property rights.” She thus rejected Apple’s argument as pretextual with respect to the 30% commission rate specifically.
(ii) In-app payment restrictions
The Judge’s analysis of the anticompetitive effects and the business justifications of Apple’s in-app payment restrictions is rather short and superficial.
As far as the anticompetitive effects are concerned, Judge YGR primarily referred to her analysis of the anticompetitive effects of the app distribution restrictions. Thus, the Judge seems to consider that Apple’s in-app payment restrictions can also adversely affect competition, prices and innovation with respect to in-app payment processing (assuming a “but-for” world where third party app stores could compete with the Apple App Store, among others on in-app payment processing).
Judge YGR then examined three business justifications, namely (1) payment security; (2) commission collection; and (3) value of intellectual property.
The Judge agreed that “decentralization may decrease security in some instances” but she seemed to dismiss Apple’s argument for IAP security whereby it can verify digital good transactions, noting that Apple did not show how its process is “any different than other payment processors, and any potential for fraud prevention is not put into practice.” While Judge YGR considered Apple is entitled to a commission, and that IAP provides an efficient method for collecting it, she again observed that Apple “has not adequately justified its 30% rate”, noting that “merely contending that its commission pays for the developer’s use of the App Store platform, license to Apple’s intellectual property, and access to Apple’s user base only justifies a commission, not the rate itself.”
In concluding her analysis, YGR observed that “looking at the combination of the challenged restrictions and Apple’s justifications, and lack thereof, the Court finds that common threads run through Apple’s practices which unreasonably restrains competition and harm consumers, namely the lack of information and transparency about policies which effect consumers’ ability to find cheaper prices, increased customer service, and options regarding their purchases.”
(iii) The application of the law to the facts
As noted above, based on the definition of the relevant market as one for “mobile gaming transactions”, Judge YGR found that “while the Court can conclude that Apple exercises market power in the mobile gaming market, the Court cannot conclude that Apple’s market power reaches the status of monopoly power in the mobile gaming market. That said, the evidence does suggest that Apple is near the precipice of substantial market power, or monopoly power, with its considerable market share.”
Since Apple was not found to hold monopoly power in the relevant markets, the Judge dismissed all of Epic Games’ monopolization claims based on Section 2 of the Sherman Act, including Count 1 (monopoly maintenance of iOS app distribution market), Count 4 (monopoly maintenance of iOS in-app payment solutions market), and Count 2 (denial of essential facility).
Judge YGR also examined Epic Games claims based on Section 1 of the Sherman Act under a “rule of reason” analysis, balancing the anti-competitive effects of Apple’s restrictions with its business justifications.
Apple’s unlawful restraint of the iOS app distribution market (Count 3)
Starting with Apple’s app distribution restrictions, the Judge found that such restrictions do have some anticompetitive effects (as discussed above), but Apple has offered valid procompetitive justifications based on security and the need to protect its IP rights. As noted above, when discussing the facts Judge YGR seemed receptive to less restrictive alternatives (e.g., notarization). However, she ultimately rejected Epic Games’ claim on the ground that it “has not met its burden to show that its proposed alternatives [to centralized app distribution] are ‘virtually as effective’ as the current distribution model and can be implemented ‘without significantly increased cost.’”
Apple’s unlawful restraint of the iOS in-app payment processing market (Count 5)
Moving on to Apple’s in-app payment restrictions, the Judge similarly found that “Epic Games has presented some direct and indirect evidence showing that Apple’s IAP functionality has had anticompetitive effects”, while Apple has proffered several procompetitive justifications for such restrictions. The Judge eventually dismissed Epic Games’ claim, holding that Epic Games “has identified no suitable less restrictive alternative for Apple’s use of IAP based on the current record.”
Finally, Judge YGR also rejected Epic Games’ Section 1 claim that Apple engaged in a tie between app distribution, on the one hand, and IAP on the other (Count 6) as the requirements imposed by the case-law for such claims to succeed (and in particular the need for two separate products) were not met.
California’s unfair competition law
Having rejected all Epic Games’ claims based on US federal antitrust law, Judge YGR nevertheless recognized that Apple’s anti-steering provisions breached California’s Unfair Competition Law (“UCL”), which prohibits business practices that constitute “unfair competition.” Relying on the so-called “tethering test” developed in the case-law, the Judge found that “although Epic Games has not proven a present antitrust violation, the anti-steering provisions ‘threaten an incipient violation of an antitrust law’ by preventing informed choice among users of the iOS platform.” Judge YGR notably observed that:
- Apple’s own records revealed that “two of the top three ‘most effective marketing activities to keep existing users coming back’ in the United States, and therefore increasing revenues, are ‘push notifications’ (no. 2) and ‘email outreach’ (no. 3)” and that “Apple not only controls those avenues but acts anticompetitively by blocking developers from using them to Apple’s own unrestrained gain.”
- “In the context of technology markets, the open flow of information becomes even more critical” since “information costs may create ‘lock-in’ for platforms as users lack information about the lifetime costs of an ecosystem.”
- The anti-steering provisions “can be severed without any impact on the integrity of the ecosystem and is tethered to legislative policy.”
The Judge also held that the anti-steering provisions violate the UCL under the so-called “balancing” test, in that “the harm to users and developers who are also quasi-consumers, is considerable” as the trial “has exposed numerous anticompetitive effects,” and the rsulting harm outweighs its benefits.
Against this background, the Judge issued the following US-wide permanent injunction (to take effect in ninety days):
“Apple Inc. and its officers, agents, servants, employees, and any person in active
concert or participation with them (“Apple”), are hereby permanently restrained and enjoined from prohibiting developers from (i) including in their apps and their metadata buttons, external links, or other calls to action that direct customers to purchasing mechanisms, in addition to In-App Purchasing and (ii) communicating with customers through points of contact obtained voluntarily from customers through account registration within the app.”
Note that the language used is almost identical to the one found in the App Store Review Guidelines (see Sections 3.1.1 and 3.1.3). In effect, Apple will be forced to drop its anti-steering provisions, so that developers can direct users to purchasing mechanisms other than IAP. Importantly, and as already noted, this does not mean that developers will be able to directly integrate inside their apps third-party payment solutions alongside IAP. Instead, developers will be able to add links or buttons (or “other calls to action”; this would presumably cover promotional messages such as “become a premium member on our website”) that when clicked on will direct users to the developer’s website, where the user may then complete her purchase, possibly at a lower price (as there will be no 30% owed to Apple). Developers may also contact users through points of contact obtained voluntarily through account registration within the app (think of email or phone number) to inform them about the ability to purchase online, instead of the App Store.
A significant point is that the injunction is not limited to gaming apps, but instead extends to all categories of apps (including subscription apps such as Spotify and Netflix – note that Apple has already accepted to ease its anti-steering provisions on a global basis for so-called “reader” apps to settle an investigation from the Japan Fair Trade Commission). Judge YGR noted that she could not discern “any principled reason for eliminating the anti-steering provisions to mobile gaming only. The lack of information and transparency extends to all apps, not just gaming apps.”
Conclusion – the aftermath of the judgment
Apple has touted the judgment of YGR as a “huge win” and proof that it is not a monopoly. Is that so, and does this mean that Apple’s App Store troubles in the rest of the world are over? We doubt this is the case.
Even assuming that the judgment will not be overturned on appeal, it will have limited precedential value in other parts of the world on the crucial issue of market definition. After all, much depended on the evidence (or lack thereof) put forward by the litigants. The fact that Epic Games failed to provide evidence to support its market definition does not mean that a competition authority in the EU or the US cannot. In addition, the market was defined to exclude all non-game apps, hence it is rather uninformative when it comes to ongoing antitrust investigations into Apple’s App Store practices with respect to other categories apps (think e.g., of the European Commission’s investigation).
On the other hand, the judgment does mean trouble for Apple in several respects. First, Apple will find it even harder to justify its 30% commission rate. Judge YGR observes repeatedly that Apple’s 30% commission rate was never set by reference to competition, costs, or value of the services provided. It was a “historic gamble” that proved incredibly profitable, apparently allowing Apple to reap “supracompetitive operating margins” over 70%, while Apple has been slow to innovate and invest in App Store-specific IP.
Second, the judgment provides for a US-wide injunction that will loosen Apple’s control of the App Store, in that all developers (and not just game developers) will be able to direct users to out-of-app purchasing channels. This will have major implications.
Third, the judgment may have the unintended consequence of bolstering legislators in the US and the EU to pass legislation curbing Apple and Google’s control of their app stores (such as the Open App Markets Act or the Digital Markets Act). Indeed, the judgment could be seen as a further example of courts narrowly interpreting US antitrust law (or simply doing it poorly), hence giving ammunition to those calling for legislative reform (this was essentially the remark of US representatives Nadler and Cicilline in response to the judgment).
While only time will tell, this might turn into a Pyrrhic victory for Apple.
[Disclosure: I am the outside antitrust counsel of the Coalition for App Fairness, and advise a range of app developers. All views expressed in this post are, however, enterely mine.]