Apple and App Neutrality under EU Law

And now for a little extracurricular detour from (the usual) net neutrality, with a venture into app neutrality. I was struck by the recent reports of Apple's revising of the License Agreement for software developers on the iPhone, iPad, or iPod Touch platform. The most striking modification reads:

3.3.1 — Applications may only use Documented APIs in the manner prescribed by Apple and must not use or call any private APIs. Applications must be originally written in Objective-C, C, C++, or JavaScript as executed by the iPhone OS WebKit engine, and only code written in C, C++, and Objective-C may compile and directly link against the Documented APIs (e.g., Applications that link to Documented APIs through an intermediary translation or compatibility layer or tool are prohibited). [Italics added]

This amounts to a explicit boycott of—most visibly—Actionscript, the programming language for Adobe Flash. Apple's distaste for Flash has long been latent, and it's by no means a coincidence that Flash support has always been absent for iPhone or iPod Touch. Apple's change of the License Agreement comes at a time when Adobe is about to launch the latest iteration of its flagship software package CS, including the option to develop iPhone/iPad/iPod Touch apps using flash—again, what a coincidence.

Now, at a time when most online video runs on Flash, it is rather annoying not to have access to the Daily Show episodes on my iPhone. Besides my personal annoyance, Apple's boycott also chains software developers, and seems to have a negative effect on Flash's market position. Which begs the question: can we (myself, app developers, Adobe) turn to (European) law to do something about this? Short answer (SPOILER ALERT): unlikely so.

There are a number of possible approaches to this issue under EU law. As both Apple and Adobe are established in Europe, EU law can be invoked. The obvious approach would be European antitrust law, mainly art. 102 of the EU Treaty. It seems as if art. 102(c) is applicable in this case for instance, as Apple indeed seems to be "applying dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage." For art. 102 to apply, Apple and Adobe need not be direct competitors—which they are not at this point. Rather, abuse of dominance over potential competitors also suffices. However, under an art. 102 procedure a situation of dominance needs to be established first, which is calculated by a process of market definition. I don't want to bore anyone with the details on EU market definition (here), but bear in mind that product markets, geographic markets and temporal quality are factored in. A plausible relevant market in this case for instance would be the European market for mobile operating systems.

Case law has not established a golden rule of what constitutes dominance on a market once defined, but generally speaking anything over 40% is in trouble. Now, clearly Apple does not hold a market share of over 40%, either as a mobile OS or as a phone manufacturer. Therefore it would become hard to argue an art. 102 violation by Apple, as dominance cannot be established. Moreover, let's not forget that this case appears to concern a two-sided market: Apple functions as a platform serving both app developers and end-users. This complicates the process of market definition even further.

Let's for the sake of argument assume that Apple were dominant on the Mobile OS market. This would open up the realms of Refusal to Supply case law—as Apple would refuse to Adobe supplying flash products to end-users on its products. The ECJ has been rather strict in refusal to supply cases, and has struck down hard on firms dominant firms refusing to supply, like Commercial Solvents and United Brands. It's only in very few cases that plaintiffs in refusal cases have successfully argued for objective justifications of their conduct—Apple's rebuttal that refusing Flash will "result in sub-standard apps" and "hinder the progress of [Apple's] Mobile platform" are unlikely to hold up.

So refusal to supply would stand a chance, if Apple were in fact be dominant on the defined market for this hypothetical case—which it is likely not. So much for EU antitrust.

Next: what about European harmonizing regulation, like the telecoms Framework and the eCommerce Directive? About the telecoms framework I can be short: it does not cover content, only telecommunications on the network level. For the eCommerce directive it is the other way round: it does exclusively cover "information society services" (content), yet doesn't affect apps much—or Apple for that matter. Moreover, it is highly unlikely that a patchy Directive like eCommerce can be directly invoked by (legal) persons in a private law suit. Forgive me for not going into the details of direct effect of Directives, it's a long story.

Now please tell me if I've overlooked something, but it seems as if Apple will get a pass for its conduct under EU law. There is the fundamental rights route, with the EU Charter of Fundamental Rights codifying the freedom "to receive and impart information and ideas without interference by public authority and regardless of frontiers" (art 11). This would be a long shot though: not only would it be a hard case to make for a claimant that Apple is completely preventing app developers (or Adobe) to receive and impart information, such claims have also only rarely been upheld in private suits.

So EU law seems to be exhausted, but is this really a problem? When following a Chicago School rationale, it seems we shouldn't be too bothered with Apple's conduct. If it hampers innovation and is inefficient, even as a monopolist it will attract entry by competitors who do allow all API's on mobile app platforms. The problem with this Chicago approach is that it doesn't take network effects into account. There has in fact been a lot of entry in the mobile app field, think of Android Market. But app developers want to be where all other apps are: the amount of participants on a network increases its value for all. The Apple app store features thousands of apps, which makes it most attractive for developers to write apps for the Apple platform. Compare that with the about 800 apps currently available on Window's app Marketplace.

Apple is well aware of this situation, and naturally wants to keep it that way. So even though Apple and Adobe are not directly competing, Apple has much incentive to foreclose Adobe on its platform. As Joe Wilcox correctly remarks, flash offers developers a way around the app store—a flash platform through a browser. Apple doesn't want this of course, as the app store (1) increases the value of the iPhone/iPad/iPod Touch for consumers, and (2) generates a lot of revenue for Apple.

Concluding: it is perfectly clear why Apple is foreclosing on Adobe, Apple's conduct is likely to have an anti-competitive effect, yet European law stands idle. This re-enforces Jonathan Zittrain's call for API neutrality. The walled garden that Apple is building works very well as a stand-alone platform, in which hardware and a rich suite of software are perfectly integrated. This is satisfying to consumers (like myself), and profitable for Apple. However, in the long run walled gardens hinder innovation in a way almost similar to state planning: it adds a layer of control and bureaucracy to what otherwise would be free floating innovation by software developers.

DISCLAIMER: this is a purely (legal) theoretical point, and very much work in progress. I'm still looking into economic studies and empirical work. Please comment!

Open-access fiber in Amsterdam

Great article today in Ars Technica on the open-access Fiber project Citynet in Amsterdam. The piece is written by the CEO of Citynet, Herman Wagter.

It's a pretty elaborate (and technical!) article, and certainly stimulating for other municipalities with fiber aspirations. Also note that Citynet offers a blueprint to encourage fiber deployment in Europe, while avoiding State Aid concerns under European Law.

The key to success is an extensive preparation, a detailed design, good organization, and social engineering when dealing with people who live in the MDUs; in fact, this is much more important than trenching and putting in fiber. Only 120,000 meters of trenching was needed for the first 40,000 connections, an average of three meters per connection. Roughly 80 percent of the costs were labor costs, while 10 percent were fiber.

The payoff for the effort is that fiber, once deployed, has proven to be very reliable. So it makes sense not to cut corners and deploy a network that can be used for as many decades as possible.

Fiber and State Aid

Rudolf van der Berg's twitter feed referred me to an article about multinational cable company Liberty Global in what would be the Dutch equivalent of the Financial Times. Liberty, which owns Chellonet, Telenet and UPC in Europe, is apparently considering expanding into Eastern Europe. At the same time, its chairman denies rumors that Liberty would be interested to take over (parts of) the competing cable operator in the Netherlands, which allegedly is for sale.

What struck me in the article however is Liberty's stance on public/private partnerships to invest in fiber architecture. Freely translated, chairman Fries states:

I am worried about the intentions of governments investing in fiber. When it involves government subsidies, this will negatively affect the level playing field. When subsidies are conform a market economy investment standard this is less of an issue, but I have doubts whether this is feasible.

This comment echoes Liberty's official comments during the European Commission's consultation period concerning state aid for broadband. However, the eventual Commission guidelines on the application of state aid laws on investment in broadband that were published a few months ago, have in fact set the bar rather high for public investment into broadband to pass. The Commission has designated white areas (rural and under-served localities); black areas (two or more broadband providers already present) and grey areas (everything in between). Naturally, state aid to foster the roll-out of broadband is allowed in white areas, prohibited in black areas, and debatable in grey areas. This distinction is a bit looser concerning next generation access networks such as fiber, depending on the substitutability of existing broadband architecture. In short, if cable and DSL start lagging behind too much compared to fiber, this might justify state (market conform) investment into fiber.

Now, understandably Libery is upset about this, given that the city of Amsterdam has already successfully applied this exemption for fiber. In the Citynet case of 2008 before the Commision, Amsterdam's investment into fiber-to-the-home architecture was not deemed to concern state aid. Liberty's subsidiary UPC was one of the complainants in this case, and apparently is still sour over the Commission's decision.

So what does this tell about Liberty's incentives? Not upgrading its cable network to match fiber speeds might lead to public investment into fiber—and it shows Liberty is not too keen on that. At the same time, rolling out fiber itself carries great sunk costs. To solve this dilemma—as Rudolf (again) has accurately pointed out before—Liberty is very comfortable on markets with (DSL) incumbents averse to innovation into fiber. This allows Liberty to go sit above DSL and just below fiber speeds, which both increases its market share and keeps public investment into fiber at bay. The question is however how sustainable such a strategy is in the long run. The gap between fiber and 'old broadband' will widen eventually, and demand for fiber speeds will rise inevitably. Ergo, there will be public initiatives to roll-out fiber at some point, backed by European state aid laws. The living proof of this is Amsterdam's Citynet.

Reding v. Kroes, round one

Viviane Reding may be on her way out of DG Info Society, but it seems like she'll be bringing along some pet projects to her new Fundamental Rights Directorate. EurActiv reports today that Reding's personal priority will be updating the outdated Data Protection Directive, and sources close to Reding insinuate that her focus shall be on online privacy. Never mind that much of online privacy is being dealt with in the ePrivacy Directive, for which Reding's former Directorate is responsible.

Not coincidentally, Reding and the new commissioner at Information Society—Neelie Kroes—reportedly do not get along well. The heated debate between Kroes and Reding over functional separation of telecom operators of 2007 most clearly testified to their icy relationship.

So is a turf war to follow? Reding's approach seems to suggest that she already fired the first shot. After all, while she stresses that the Data Protection Directive has not been revised in 15 years, she does nowhere mention the 'particularizing and complementing' role of the ePrivacy Directive alongside the Data Protection laws. Moreover, the ePrivacy Directive has officially been revised only a month ago—which took place under Reding's watch.

I'm curious to see if and how Neelie Kroes will respond. Only then it can be determined whether we've got a fight on our hands...