Saturday, March 26, 2022

Single-brand market definition is the way to go in Epic Games v. Apple as Apple's blunder and a legally and factually nonsensical sentence by the district judge show

Yesterday I pointed out a major mistake by Apple's otherwise absolutely great lawyers: in their response to Epic Games' opening brief they conceded that iOS competes with Android on app (not just in-app purchasing) security, which simply validates Epic's proposal of a smartphone operating system foremarket (and two iOS-specific aftermarkets in which Apple then has a market share of 100%, i.e., total monopoly power). I wrote that I had more questions than answers because I had to look into this again. As a result, I am not only able to explain Apple's blunder more specifically but I also came to realize that one sentence in the foremarket-related part of the district court's judgment is complete and unbelievable nonsense.

A Ninth Circuit judge said during the FTC v. Qualcomm hearing two years ago that the house of cards (based on which Qualcomm was held liable for antitrust violations, and enjoined) began to fall. From my vantage point, the basis on which the district court sided with Apple as it decided the single most important question in the case--market definition--against Epic is so fundamentally flawed that "house of cards" would be a euphemism.

For app developers and users, only a single-brand market definition can really solve the problems we face with app taxes, self-preferencing, unreasonable restrictions on advertising, and unjustified rejections (short of new legislation, of course, which would establish different criteria than traditional antitrust law--but would come too late for Epic Games v. Apple).

There's a transparent reason--but unfortunately one that has nothing to do with proper administration of justice--for which Judge Yvonne Gonzalez Rogers, whose approach to the case I described as "extremely honest" in January, presumably never wanted to adopt a single-brand market definition. Already at the TRO/PI stage she expressed her discomfort with a decision in Epic v. Apple that might also affect console makers like Sony. She later faulted Epic for allegedly failing to distinguish the iOS app tax from similar regimes on gaming consoles. In other words, she was very afraid. Very afraid of setting a precedent with unintended consequences. Very afraid of opening the floodgates. At least she was being honest about it. But it was disappointing to see her ignore what she said toward the end of the trial and what Tim Cook couldn't deny, which is essentially that Apple competes for end users, not for developers.

It's understandable that she was wondering why Epic was complaining about Apple's 30% when Sony gets a lot more money from Epic on the same basis, with even tighter restrictions, yet chose to pick a fight with Apple rather than (its own small shareholder) Sony--though quite frankly it would even be legitimate if Epic first picked a fight it can afford in terms of lost revenues. Anyway, Epic didn't have that burden of addressing hypothetical future cases involving other platforms, and it might even have met it without having it in the first place. Her primary responsibility, however, was to correctly decide the case before her. The right approach would have been to let the Sonys and Nintendos of the world do the job of distinguishing their fact patterns from Apple's, as opposed to her developing a slippery slope argument and then buying it. She should have enough confidence in those sophisticated parties and in the judiciary that those other cases would be decided appropriately. In the single most important part--the one where she rejects Epic's market definition with a smartphone operating system foremarket--she wrote something reflective of a complete misconception:

"In essence, Epic Games ignores these marketplace realities [such as competition existing between smarpthones] because, as it presumably knows, Apple does not have market power in the smartphone market."

She then clarified the last part noting that "Apple only has 15 percent of global [smartphone] market share in 2020."

Why is that passage off-base? So very flawed in fact and law?

It's factually nonsensical because the iPhone's market share among smartphones and iOS's market share among smartphone operating systems are identical. (We can ignore the potential distinction between smartphones and tablets because one would obviously always compare apples to apples: whether we're talking about smartphones or about smartphones and tablets, Apple's market share as a device maker equals its market share as an operating system maker either way.)

It's legally nonsensical (even if we assume an alternative universe with different laws of nature so that Apple could have market power in smartphone operating systems while lacking it in phones) because Kodak and Newcal--the relevant single-brand market cases the district court also cited to--do not at all require market power in the foremarket. Much to the contrary, they're about a company facing competitive constraints in the foremarket but exercising monopoly power in the aftermarket. Competition in razors, but not in blades. Competition in cars, but not in spare parts. You can have a 0.1% market share in the foremarket and still 100% in the aftermarket, and we'll get to the standard further below, but even someone with a 0.1% market share in the foremarket could theoretically get into trouble with misconduct in the aftermarket.

The Supreme Court's Kodak syllabus says this: "Kodak's theory that its lack of market power in the primary equipment market precludes-as a matter of law-the possibility of market power in the derivative aftermarkets rests on the factual assumption that if it raised its parts or service prices above competitive levels, potential customers would simply stop buying its equipment. Kodak's theory does not accurately describe actual market behavior [...]" (emphasis added)

How could the judge get that part so very wrong? The psychological explanation--fear of ripple effects (Sony etc.)--is obvious just based on her own statements. The more difficult part is to figure out how she ever came to think that Epic's "proposed foremarket is entirely litigation driven [and] misconceived" (a gratuitous attack just like the passage I quoted further above). What did she suspect Epic was trying to gain? Maybe she thought that what Epic meant to be a competitive foremarket (the judge herself wrote that "[c]onsumers should be able to choose between the type of ecosystems," i.e., she acknowledges users may choose between iOS and Android phones) was already an aftermarket: she might have suspected Epic of focusing on smartphone operating systems because then Apple has a 100% market share among operating systems for iPhones. In that case, phones would have been the foremarket, operating systems the aftermarket, and app distribution and in-app payments would have been aftermarkets of the operating system market, i.e., second-degree aftermarkets.

The ruling reflects her predilection for what Apple's expert witnesses said as compared to Epic's. I don't mean to allege a bias in the sense that she preferred them because they were retained by Apple. Not at all. It may very well be that in her unbiased eyes Dr. Schmalensee simply came across as more competent and convincing than Dr. Evans (both actually co-authored a book on platforms a while ago). But what Dr. Schmalensee said about competition for smartphones extending beyond the operating system and involving features such as battery life, durability, camera resolution etc. was just a smokescreen that such a smart judge should never have been misled by.

It doesn't even matter in the foremarket-aftermarket context why a user chooses an iPhone, like camera resolution or battery life. The choice can be made because the user thinks Apple does a better job fighting climate change than its rivals and, therefore, wants to support Apple. I just returned a bunch of old devices to an Apple Store so they can recycle them, so I do appreciate their environmental protection efforts. The customer's choice can even be made erroneously with a drunk or vision-impaired person confusing an iPhone for a Samsung Galaxy. It's all irrelevant in the foremarket-aftermarket analysis. So what really is the standard? Why should Epic--on behalf of all of us--win the market definition part? And why did Apple, through a blunder, strengthen Epic's case?

What fortunately simplifies the analysis is that even Judge YGR found that Epic met three of the four criteria for an aftermarket. This means we need to focus only on the foremarket and the one factor relating to the aftermarket that the district court said Epic failed to satisfy.


As Epic and the DOJ explained in their submissions, the district judge was wrong when arguing that there cannot be "a market for something that is not licensed or sold to anyone." That passage, by the way, only makes sense if one inserts the word "separately." Otherwise it would be legally and factually crazy. Of course end users do get an iOS license, and iOS is sold--as part of a bundle. But I don't have to reinvent the wheel after Epic and DOJ already having addressed that part. I'm pretty confident the Ninth Circuit will agree, and the appeals court can say that even Apple's army of first-rate lawyers inadvertently conceded away the foremarket part:

"Apple's iOS also performs better in this respect [i.e., on security] than its 'main competitor' in the relevant market: Android."

I'll now make good on my promise to be more specific about Apple's mistake. The devil's advocate may ask: am I taking this out of context? Actually, Apple would have to take this out of context in order to explain away its concession...

The subhead of the section that contains the quoted sentence is: "Epic Did Not Prove Viable Less Restrictive Alternatives" That's where Apple claims (rightly or wrongly, doesn't matter here) that at this stage Epic's sole remaining proposal for a remaining less restrictive alternative (to the current setup where all iOS apps must be approved by Apple for App Store distribution, subject to developers accepting its app distribution terms) is a "notarization" program "whereby native iOS apps could run on iOS only if they have been 'notarized' by Apple."

Apple then makes two points why it considers this alternative unacceptable. The second one is that the current system makes it easier and more efficient for Apple to collect its compensation. It is in the context of the first argument that Apple said iOS competes with Android on security. Apple says that it "intentionally designed a more secure ecosystem for iOS by requiring human review of iOS apps" (emphasis in original) and first explains that iOS was always meant to be more secure than macOS. After the macOS security comparison, Apple then brings up Android:

"Apple's iOS also performs better in this respect than its 'main competitor' in the relevant market: Android. [...] Google, which charges the same headline commission rate as Apple while allowing sideloading and alternative stores, has 'higher malware rates' [showing a chart from the evidentiary record] Tellingly, Google has increasingly moved toward more robust human review on its Google Play store."

How could one try to reconcile that passage with Apple's support of the district court's opinion that there cannot be a smartphone operating system foremarket, and that the "relevant market" here is digital mobile gaming transactions?

To be fair, it would take Apple's statement out of context--and put it into the wrong context--if I claimed that Apple made this concession in the market definition context. It obviously stayed on message while discussing market definition. Apple also stayed on message in various other contexts--in all but this one (least restrictive alternatives).

The word "Android" has 54 occurrences in the district court's ruling and 12 occurrences in Apple's response to Epic's appeal. That's a total of 66 occurrences. There is only one--the one I just quoted again--where Android and iOS are described as competing products. In all other contexts, it's about Android devices, the Android ecosystem, or the (Android) Play Store, but not about Android as a competitor to iOS. It's not just a minor inconsistency. It's extremely revealing. It happened because the notion that there is no smartphone OS market is so contorted and contrived that sooner or later someone trying to stay on message ends up conceding the truth. The horse is now out of the barn.

Oddly, one could not even reconcile that iOS-and-Android-compete statement with Apple's position on market definition by rephrasing it as follows:

"Apple's iOS App Store also performs better in this respect [security] than its 'main competitor' in the relevant digital mobile gaming transactions market: Android the Google Play Store."

The modified version above would focus on the two app stores, and would explicitly mention that it's about how they compete on digital mobile gaming transactions. But it wouldn't make technical sense as Apple refers to "sideloading and alternative stores" in the very same paragraph as what makes iOS more secure. Architecturally, it is not the App Store that prevents sideloading and alternative stores. It's iOS which disallows everything to that effect but the App Store.

So if we rephrased Apple's blundered sentence like that, we'd just again be pointed to the fact that iOS and Android compete on security, with Apple arguing that a prohibition of sideloading and alternative stores is good for security (which Judge YGR doubted, at least during the trial, given that more competition might also lead to greater security).

To be clear, they're not talking about the security of IAP transactions via the App Store versus the security of IAP transactions via the Google Play Store. It's about iOS apps (which can't be downloaded from alternative stores or sideloaded) versus Android apps as sideloaded apps, or apps downloaded from third-party stores, would not undergo Apple's human app review. It's about the differentiation of iOS from Android, not the App Store from the Play Store.

Just one last point before I turn to the aftermarket part: Apple's blunder can even be interpreted as acknowledging that what drives in-app purchases on iOS is Apple's success in the foremarket. If users choose iOS devices for greater security (right or wrong), they then transact via the App Store, the only game in town. If it was about which app store they trust, Apple could still just compete with sideloading and third-party app stores, and those who value security could still choose the App Store over sideloading and third-party app stores for that very reason. Epic can be criticized for all sorts of things, but they've been consistent and clear that they stand ready to let an iOS version of their Epic Games Store compete with the App Store, just that Apple wants to avoid that kind of competition.


The district judge took the position that she wouldn't even have to analyze an aftermarket because "there are no derivative markets" absent a foremarket. Nevertheless, she "addresse[d] the additional problems with Epic Games’ attempt to define the market with the confines of a single brand." While "three of the four indicators are fulfilled" even according to the decision, there's one aftermarket criterion the judge didn't see satisfied by Epic: "whether competition in the initial market suffices to discipline anticompetitive practices in the aftermarkets"

By blaming Epic for artificially alleging "market power" (through a smartphone OS rather than device market as the proposed foremarket), Judge YGR suggested a requirement that doesn't exist. In a proper foremarket, there are competitive constraints. They will be weaker in some cases and stronger in others, but there is at least some degree of competition. There can be a case A in which there's very little competition in the foremarket, but there may be dynamics that are sufficiently strong to ensure that a company exploiting its aftermarket monopoly will pay a price for this in terms of losing market share in the foremarket. Conversely, there can be a case B in which there's perfect competition in the foremarket, but there are no such dynamics, and therefore even a small player can subsequently exploit the aftermarket monopoly without losing market share in the foremarket.

In Epic Games v. Apple, there's a duopoly in the foremarket if we define it as smartphone operating systems. If we define it as smartphones, Apple's market share is the same. Epic had nothing to gain, contrary to what Judge YGR suspected. Much to the contrary, it is actually her market definition of digital mobile gaming transactions in which Apple has a higher market share thanks to the average purchasing power of its affluent customers than in the smartphone device or smartphone OS market.

The question in Epic Games v. Apple is, therefore, whether Apple will lose market share because of end users buying non-iOS devices as a result of treating developers the way it does. In order for that to happen, some developers would have to stop developing for iOS or some might invest less in iOS development. In that case, products competing with Apple's might become more attractive by virtue of a superior app offering.

There's no question that Apple could do things that would discourage iOS app development on a large scale. If it asked for the moon just for someone to be allowed to submit an app for App Store review, no one could afford it. But Apple won't do that, and it doesn't have to in order for there to be an aftermarket problem.

Judge YGR set the bar too high for Epic in this respect. I wish the DOJ had also addressed the aftermarket part, but it focused on the foremarket. After all, the DOJ clearly decided to go as far as it could in supporting Epic without reaching the point where its positions would inevitably, if adopted, enable Epic to win--in that case, the DOJ would have had to formally file in support of Epic and not "of neither party." But my hope is that the Ninth Circuit will easily identify the foremarket part of the decision as being off-base, also in light of the fact that Apple itself ended up describing iOS and Android (not just iPhone and Samsung phones) as competitors in the least-restrictive-alternative context, and that Epic's very strong aftermarket argument will then be given serious consideration.

Epic's brief convincingly explains that a shift in policy (i.e., the accused party exploited its aftermarket monopoly only later on) or definitive customer unawareness (of the challenged aftermarket business terms) are requirements that the district judge thought she had found in various other U.S. antitrust decisions but which aren't absolute requirements. If the appeals court agrees with Epic in that regard, the outcome-determinative question is then whether switching costs (between iOS and Android phones) create a certain degree of customer lock-in. Contrary to the decision that is being appealed, it doesn't take "nefarious" conduct by Apple to have a lock-in, and customer loyalty may very well be the result of great customer satisfication: it's all about whether Apple faces an indirect competitive constraint in the aftermarket because of what might happen in the foremarket. Even the district court's own judgment actually acknowledges various facts that the appeals court may consider a sufficient basis to conclude that Apple can act abusively toward app developers without losing market share. It's all about (the absence of) a feedback loop.

I'm already looking forward to the market definition part of Epic's reply brief as I believe Apple has actually weakened its position through its responsive brief.

Share with other professionals via LinkedIn: