Saturday, May 1, 2021

Apple raised its effective App Store commission rate in certain geographic markets to (respectively) 31.4%, 32.1%, and 35.25% in September

After this post, I'll (have to) take a break from blogging about App Store antitrust matters for a few weeks or maybe even months, as I'll explain further below. Before I do that, I'm going to share several thoughts and pieces of information in this post. You can click on any of the links below to go straight to the part you're most interested in:

  1. Effective App Store commission ("App Store tax") rate peaks at 35.25% (plus annual developer program fee plus Search Ads) -- a relative increase by 17.5%

  2. IP-related issues surrounding web apps

  3. Recent United States Senate hearing: mixed blessing for Epic's case

  4. Statement of Objections from the European Commission's Directorate-General for Competition (DG COMP) in the investigation instigated by Spotify

  5. Taking a break from commenting on app store antitrust cases


1. Effective App Store commission ("App Store tax") rate peaks at 35.25% (plus annual developer program fee plus Search Ads) -- a relative increase by 17.5%

In its proposed findings of fact and conclusions of law, Epic Games debunks Apple's claim that it has not been able to increase (or even maintain) its App Store commission rate due to competitive constraints (though in reality any reductions were motivated by antitrust-related reasons):

"92. Moreover, contrary to its claims, Apple has repeatedly increased prices after developers and consumers were locked in, including by requiring use of Apple’s IAP to process payments for in-app digital content (2009); requiring IAP for subscriptions (2011); and charging developers for search ads (2016). (Findings of Fact ¶¶ 23, 123.)"

The ability to increase prices without losing market share is characteristic of a monopolist. Sometimes, monopolies are identifiable just on that basis. In those cases where a monopolist could have increased prices, but did not do so, a SSNIP test is performed by economists on a hypothetical basis (such as by conducting a survey): Small but Significant Non-transitory Increase in Price. Generally, "small but significant" means 5%-10%, and plaintiffs often argue that they can establish even greater market power than what it takes to command a 10% increase.

Epic's examples all make sense. It's true that Apple extended the scope of applicability of the 30% commission (the Small Business Program just came recently for antitrust reasons and has nothing to do with market dynamics whatsoever) to other types of payments, as the testimony in this case confirms. Search Ads are indeed another indirect price increase, as many app developers pay for the discoverability of their apps on the App Store by promoting their apps above the organic search results that may favor a competitor--which in turn often forces the affected competitor to place Search Ads only to maintain the top spot.

But Epic's list of de facto price increases is not even exhaustive. (Can't blame them as they need to focus and Apple's sophisticated tactics raise so many issues.)

The FT's Tim Bradshaw highlighted another problem in September:

At around the same time, other people commented on it as well, and one website had to backtrack because they made it sound like Apple passed 100% of those digital services taxes on to developers. There should have been much more outrage, and in some jurisdictions developers could even have brought complaints over this particular issue. But it went almost unnoticed, probably because too few people--if any--thought it through in every detail.

Once one has thought it through, it's crystal clear: many (if not most or even all) third-party developers end up having to pay digital services taxes ("DST") only because of Apple's tying (of the payment system to the App Store as the only access route to iOS users), and wouldn't owe those taxes otherwise.

In Turkey, DST is 7.5%; in France and Italy, 3%; and in the UK, 2%. More countries will follow. In fact, President Biden is open to a global agreement on DST.

Apple treats DST like VAT (Value Added Tax) or sales tax (which is simply the same when it comes to a business-to-consumer transaction), which belongs to neither Apple nor developers, and subtracts both VAT and DST from what customers pay before splitting the income with developers. It appears that most people haven't figured out yet why that is inappropriate, unfair, and indicative of Apple's market power:

  • VAT is a concept that's about 100 years old, while DST wasn't even foreseeable when the original App Store terms were set in 2008.

  • VAT is charged on broadly defined product and service categories (such as having one rate for food, another for non-food), while DST relates to narrowly defined types of services, such as app stores and online advertising. Typically, DST does not apply to games, so even Epic's Fortnite would not be affected if they could just use a payment service of their choosing for in-app purchases.

  • The thresholds for DST are extremely high, while countries exempt companies from VAT only if their sales are below negligible de minimis thresholds. The whole idea of DST is to tax only large and rich digital gatekeepers, which is why lawmakers always "gerrymander" the thresholds. DST typically comes with a global and a domestic threshold, and applies only to those who meet both, with "local heroes" typically failing to meet the global one.

  • While both VAT and DST are charged as a percentage of sales, VAT is a consumption tax and DST is meant to be a tax on (huge) profits.

  • Therefore, VAT is meant to be ultimately paid by consumers (as an indirect tax), while DST should be paid by "GAFA" (Google, Apple, Facebook, Amazon).

Australia is a special case: it applies VAT (called "Goods and Services Tax" (GST) down under) to digital services (not just specific types of marketplaces), with a threshold of A$75,000. There are probably some developers who do not benefit from Apple's Small Business Program (worldwide revenues in excess of 1 million), but wouldn't have to charge VAT in Australia.

The situation in Turkey is different from that in France, Italy, and the UK because Apple raised consumer prices accordingly. But that doesn't mean the effective commission rate didn't increase as well--and just delivers additional proof of Apple's market power over consumers even in a market with a low iOS market share compared to Android.

If we focus--for simplicity's sake--just on the four DST jurisdictions I've already mentioned (with a combined population of roughly a quarter billion people), and on the period before Apple's Small Business Program, this means Apple raised its App Store tax to 35.25% in Turkey (30% + (70% times 7.5%)), a relative increase of 17.5% (way above the SSNIP range); to 32.1% in France and Italy (relative increase: 7%, about the middle of the typical SSNIP range); and 31.5% in the UK (relative increase: 4.67%, pretty close to the lower end of the SSNIP range.

Apple makes things look "equitable" by deducting DST, then splitting the remainder. But there are three arguments against it, any single one of which is reason enough for Apple to internalize 100% of DST:

  1. Contractual: DST wasn't foreseeable when Apple set its original App Store terms (i.e., at a time when Apple claims it didn't have market power).

  2. Policy (legislative intent): Lawmakers wanted "GAFA" companies to internalize those taxes. For example, the UK government says:

    "The measure is expected to have an impact on a small number of large multinational groups by bringing into scope of Digital Services Tax the proportion of their revenue that is derived from UK users of social media, search engines or online marketplaces." (emphases added)

    Could lawmakers have worded their DST laws more clearly to achieve that effect? Well, even if they had done a better job, companies with market power would always find a way to offload that tax burden onto those who are dependent on them, or on consumers (as Apple did in Turkey), so even the best DST law wouldn't work without effective competition enforcement.

  3. Liability: The simplest and therefore strongest point is that even Epic wouldn't pay DST on Fortnite revenues in a jurisdiction that applies it to marketplaces such as app stores if not for the mandatory honor to use Apple's payment system.

The extent to which a given developer is impacted by that de facto commission hike varies greatly. Developers who generate all or almost all of their sales in non-DST jurisdictions are not affected for the time being, though DST is getting adopted in ever more places. On the other end of the spectrum there are companies that generate all or almost all of their IAP revenues in a country like Turkey or the UK, be it because the functionality and/or content of their apps is of interest only to customers in those target markets or because they just happened to get more traction there.

To sum it up, the effective App Store tax that a developer pays on a particular IAP transaction is the percentage of ex-VAT (and ex-sales-tax) proceeds from users that the developer would additionally keep if it could charge end users directly. Apple clearly has the power to increase that rate to developers' detriment.

2. IP issues surrounding web apps

In a recent post I mentioned one of Apple's least convincing claims, which is that native IOS apps (the ones you download from the App Store) face competition from an alternative called web apps (or sometimes "progressive web apps").

While Apple's defenses against Epic are partly based on Apple wanting to be free to commercialize its intellectual property rights (in a transparent attempt to match the FTC v. Qualcomm pattern), the suggestion that developers could offer web apps instead of native apps shows that Apple has very little respect for developers' IP. Epic's proposed findings of fact and conclusions of law explain various shortcomings of web apps. I'd just like to add a couple of IP-related ones, and a commercial one, that I couldn't find in the publicly accessible part of the record:

  • Web apps are like "open source" software: you get highly human-readable code in a scripting language. Obfuscation would be theoretically possible, but practically one couldn't afford it because it would reduce the performance and bloat the file size. Web apps already start slowly because they need to be downloaded every time they're used, and if obfuscation icnreases the file size, it takes even longer.

    In theory, anything can be reverse-engineered. In practice, native apps are hugely more time-consuming to reverse-engineer.

    The IP issue facing developers is that if you essentially publish your source code, others can easily infringe your copyright, and you give up your trade secrets. The protection of software source code by trade secret was actually the only IP protection prior to the extension of copyright law to software (and software patents came even later).

  • There's also a defensive problem: source code is easily inspected, so web apps make it easy for "patent trolls" to identify targets for their infringement allegations (whether or not those would have merit).

  • The non-IP issue I wanted to raise is that--at least when I checked a few months ago--major ad networks don't support in-app advertising on a WebGL/HTML5 basis. One can display ads outside of a window in which a WebGL app runs, but that causes other problems. Also, developers can partner with web game aggregators/portals that provide APIs and sell the ad space, but then a game must be published on those third-party sites and we're no longer talking about a web app with an icon on the home screen.

3. Recent United States Senate hearing: mixed blessing for Epic's case

Last week, the Subcomittee for Competition Policy, Antitrust, and Consumer Rights of the United States Senate held a hearing on app store competition issues. There was strong bipartisan support for combating the abuse of mobile app store monopolies. From the far left (by Senate standards) to the far right (again, by Senate standards), senators are sympathetic to developers' concerns.

Apple didn't like it at all that the hearing was going to take place so close in time to the Epic Games v. Apple trial, but ultimately provided a witness.

In some ways, maybe even in many ways, that hearing was really great for Epic's purposes. But there is a potential downside:

Senator Amy Klobuchar (D-Minn.) seeks not only to overhaul U.S. antitrust law in general but also to enact some app-specific legislation, and her position is that current U.S. antitrust law (in the combination of the statutes and how the courts interpret them) is too weak to address this issue. Epic's partners in the Coalition for App Fairness, particularly Spotify, strongly agreed with her. Sen. Klobuchar had previously lamented the state of affairs of U.S. antitrust case law in Justice Amy Coney Barrett's confirmation hearing.

Spotify's written testimony (PDF) also calls on lawmakers to "enact targeted prohibitions that will stop abusive conduct by app stores." This is a typical case of conflicting goals: you obviously can't ask for new legislation and praise existing legislation as being suitable-to-task.

It would have been preferable for the app developers who testified on that occasion to state clearly that they believe Epic is going to win its case against Apple, but any litigation comes with risks and new legislation might provide a faster solution, especially with respect to Google (Epic's case against Google is trailing far behind the Apple case). Some of what was said by the #1 antitrust expert in the Senate as well as by certain witnesses could be interpreted as expressing doubts concerning Epic's chances in court against the major mobile app store operators.

It's a typical defense not only in antitrust cases to say that a plaintiff should "direct to Congress" certain complaints or concerns. Sometimes courts say so in their written opinions. There is a risk here that the judiciary will effectively refer Epic to the Capitol.

4. Statement of Objections from the European Commission's Directorate-General for Competition (DG COMP) in the investigation instigated by Spotify

It's great news that the European Commission's DG COMP yesterday announced its Statement of Objections (SO) against Apple's App Store rules for music streaming providers, further to a complaint brought by Spotify. While the scope of that particular investigation and the market definition used in that case are relatively narrow, there can be no doubt about the Commission's--and particularly Executive Vice President Margrethe Vestager's--determination to address app store issues beyond just music streaming.

The timing of that announcement added insult to injury: precisely the work day before the start of the Epic Games v. Apple trial. It's like a pretrial amicus curiae brief.

I recommend this analysis on the Platform Law Blog. Dimitrios Katsifis of Geradin Partners explains the progress this represents as well as its limitations, and encourages additional action by national competition authorities.

While Mr. Katisfis makes strong points, I actually think it is a smart strategy by the Commission to tackle the app store problem step by step, cracking one nut at a time and gradually expanding the scope of the case law. Piecemeal tactics and progressive approaches have worked in similar contexts (such as the enforcement of the GPL free and open source software license).

5. Taking a break from commenting on app store antitrust cases

Originally I intended to follow the Epic Games v. Apple bench trial by telephone (I got the dial-in number for journalists). If the day had 34 hours, not 24, I'd still do it. But it's difficult for me to set my priorities for this month with all that's going on.

I don't want to write too much about my own complaints against Apple and Google, but suffice it to say that there is a competition enforcement agency that originally gave me six weeks to reply to Apple's response to my complaint, and in order to be able to obtain relevant data from some recent filings, I requested (and was thankfully granted) a two-week extension until May 10. At the same time, there are some things going on with respect to standard-essential patents that require my attention in the coming weeks.

The Epic v. Apple trial would be fascinating to follow in some ways, but ultimately the decision will be made by the Supreme Court, which is almost certain to hear the case (and if not, then by the Ninth Circuit, but not by the trial court).

Judge Yvonne Gonzalez Rogers said at a recent pretrial conference that there won't be any surprises: the parties have already told the court what they believe their strongest legal arguments and facts are. Now it's about whether they can deliver proof.

If Epic wins, which I hope it will though I'm not going to make a prediction at this stage, it will help the developer community at large. I really feel that Epic's sacrifice for this cause would deserve a lot more credit, but that's another story.

The hurdle is very, very high (see the section on the Senate hearing). Epic made a rather ambitious jurisdictional choice. If I were in their shoes, I'd have sued in Europe, especially the largest European market (Germany), where Google's market share is so high that market definition can be won very easily--and both statutory and case law are far more favorable. But Epic wanted to go straight for the grand prize. They are trying to succeed on Broadway or in Hollywood on the first attempt. Should Apple be let off the hook in the U.S., it won't mean that Epic was wrong nor is it likely that one could blame Epic's second-to-none lawyers. One would have to attribute it to what Sen. Klobuchar said at Justice Barrett's confirmation hearing about the U.S. theoretically having broad antitrust statutes that the courts apply very narrowly.

Epic has the stronger arguments. Apple has little more than pretext to offer, and its relatively strongest point is that other digital app stores also charge 30%, though the reduction from 30% to 12% of the commission charged by the Microsoft Store shows that competition works wonders. Apple is going to basically equate the iPhone to the Xbox, despite several fundamental differences in usage patterns and the availability of other computing devices (where there is an Xbox, there is also a number of other gadgets, but sometimes an iPhone is the only device people carry with them). Apple will, as I mentioned further above, try to match the Qualcomm pattern (not their words, but basically saying: "we're just commercializing our IP and should be free to do so as we see fit"), and to benefit from the Supreme Court's American Express decision on two-sided markets. All of that is transparent based on the proposed findings of fact and conclusions of law.

I can't imagine any developer out there wouldn't want Epic to win. I believe they can prevail, even under U.S. antitrust law as it stands. But if Epic lost, it would be a huge mistake to consider Epic Games v. Apple to be the heart of the resistance against app store abuse. Arguably, Spotify is currently in the pole position, as the SO will almost certainly result in a Commission decision, which Apple will then appeal to the CJEU.

There'll be plenty of press coverage for you to get the key "soundbites" from that trial. I'll look at this at a later stage, possibly only when the district court hands down its judgment. My positions on this topic haven't changed, nor would they.

[Update on May 3] The start of my final pretrial Twitter thread (and likely my last Twitter commentary on the issue for about a month):


Share with other professionals via LinkedIn: