As readers of this blog probably know, last week saw Google present before the General Court its case for annulling the Google Android decision of the European Commission (famous for the staggering amount of the fine imposed – EUR 4.3 billion). The case number is T-604/18 – Google and Alphabet v Commission.
I had the pleasure (and honour) of attending all days of hearing, and I would like to share some thoughts based on my experience during this week.
My purpose is not to repeat or summarize everything that was said during the hearing – there are ample news stories on that – but rather dwell a bit on the issues of principle raised by the case, and provide my take on what direction the General Court may take. There is no denial that this is a very important case, and its consequences could reverberate across the world (including the US, where the Department of Justice has sued Google for practices that are also the subject of the Google Android decision). If I am not mistaken, this is likely the first judgment to consider in detail the application of Article 102 TFEU on tying practices deployed by digital platforms since the 2007 landmark Microsoft judgment of the then Court of First Instance. More broadly, the judgment may also reveal the General Court’s current attitude on the application of competition rules in the digital economy, hence it will be eagerly awaited by the antitrust community (it is only a shame that it will likely take quite a bit of time before the judgment is delivered).
By way of reminder, the Commission’s decision in Google Android found that Google has infringed Article 102 TFEU by entering into a number of agreements with Original Equipment Manufacturers (“OEMs”) and Mobile Network Operators (“MNOs”), namely: (1) the Anti-Fragmentation Agreements (“AFAs”), which OEMs and MNOs had to sign before being eligible to distribute Google apps on their smartphones, and which prohibited them from supporting Android versions (known as “forks”) not approved by Google; (2) the Mobile Application Distribution Agreements (“MADAs”), which required OEMs/MNOs wishing to pre-install Google Play on their devices to also pre-install other Google apps, including Google Search and Chrome; and (3) the portfolio-based Revenue Sharing Agreements (“RSAs”), according to which Google provided payments to OEMs/MNOs in return for having the Google Search app exclusively pre-installed on a given portfolio of smart mobile devices.
The Commission took the view that Google’s conduct was ultimately aimed at shielding its dominant position in general search. According to the decision, Google realized early on that the shift to mobile would be an opportunity but also a risk for its search business, as it could allow (new or existing) rivals to gain scale on mobile and threaten its position.
The Commission defined the following relevant markets: (1) a market for licensable smart mobile OSs (where Google was held dominant with Android, and which notably did not include Apple’s iOS); (2) a market for Android app stores (where Google was held dominant with Google Play, and which notably did not include the Apple App Store); (3) a market for general search services (where Google was held dominant with Google Search); and (4) a market for non-OS-specific mobile web browsers (where Google is present with Chrome). It then considered that Google had committed the following abuses:
- First, Google tied its Google Search app with the Google Play app store as part of the MADAs;
- Second, Google tied its Chrome mobile web browser with the Google Play app store and the Google Search app as part of the MADAs;
- Third, Google made the licensing of the Google Search app and the Google Play app store conditional on the OEMs/MNOs entering into the AFAs; and
- Fourth, Google granted payments to OEMs/MNOs in return for exclusive pre-installation of the Google Search app within a given device portfolio per the RSAs.
If we want to categorize these practices in competition law terms, we would say that the first three practices were forms of tying (the only difference being that in the third practice there was no tied product but instead a “supplementary obligation,” namely the AFA obligation) while the fourth practice was a type of exclusivity payment (which immediately brings up to mind the CJEU’s judgment in Intel, as well as the Commission’s decisions in Qualcomm and Broadcom).
Google raised six pleas in law to contest the decision. By its first plea, Google challenged the decision’s assessment of market definition and dominance. By its second, third, and fourth pleas, Google challenged the decision’s evaluation of the MADAs, the AFAs, and the RSAs respectively. The fifth and the sixth pleas concerned Google’s rights of defence and the Commission’s alleged error in imposing a fine and calculating such fine.
During the hearing in Luxembourg Google and its supporters elaborated on these pleas, while the Commission’s Legal Service and its supporters defended the decision. The discussion was made much richer thanks to the insightful questions of the sharp Judge Rapporteur Nielsen.
As noted in the outset, I shall not endeavour to summarize everything that was said during all these days. Instead, I will focus on a couple of points I consider interesting and/or raising issues of principle.
On paternalism and limits to platform governance: In my view, one of the most important issues of principle raised during the discussion of the AFAs before the General Court concerned the role of a platform operator in protecting its ecosystem and the potential limits to such role.
By way of reminder, in its decision the Commission considered that the AFAs were capable of restricting competition from Android forks not approved by Google (“incompatible” forks), in that they prevented OEMs from supporting devices running on such forks. Amazon’s failure to license Fire OS (an incompatible fork) was cited as an example of a fork developer that was prevented from finding distribution channels by reason of the AFAs. Besides having effects on the market for OSs, the AFAs were also held to restrict competition in the market for general search services, in that they hindered the development of incompatible forks, which could in turn form alternative channels for the distribution of rival search apps.
On its part, Google claimed that the AFAs were necessary and proportionate to prevent the fragmentation of the Android ecosystem and all the negative consequences this would entail (namely, increased developers’ costs and reduced user experience, which could result in a “death spiral” similar to that which led to the demise of Symbian).
During the hearing, the Commission repeated that it is legitimate for Google to take measures to prevent fragmentation with respect to Android devices where Google apps are pre-installed. Google was nevertheless not justified in going beyond that and imposing AFA obligations for Android devices not featuring Google apps. Google responded that its conduct was necessary to protect the Android ecosystem as a whole, since there is a risk that apps in incompatible forks would crash, and users would associate this with Google’s Android. Time and again Google stressed during the five-day hearing how it is the “sponsor and steward” of the Android platform and how it cares about its success even when its apps are not preinstalled on Android devices. Google also argued that without the AFAs there was a risk fork developers could convince OEMs to work with them, resulting in several incompatible forks and thus increasing developers’ costs. Google was worried about all OEMs “jumping off the cliff” and deserting “Androidville.”
The Commission replied that this is precisely what competition is about and dismissed Google’s arguments as paternalistic. It is not (and should not be) the role of Google to dictate how competition should take place; for instance, if app developers wish to develop for an Android fork, they should be able to do so. Market forces will eventually decide whether an Android fork will carry the day and attract OEMs, developers, and users. A lawyer for the Applications Developers Association (which had intervened for Google) accepted that, whether an app developer supports a fork or not is ultimately a commercial decision for the developer to take; however, he claimed that the individual commercial decisions of app developers risked generating wider network externalities that would threaten the Android ecosystem as a whole.
This is not the first time Google justifies its conduct by portraying itself as a sort of neutral operator that carefully balances the interests of various stakeholders to secure the long-term integrity of the ecosystem. This is a recurrent theme in Google’s advocacy. For instance, a 2020 study commissioned by Google and submitted to the Australian Competition and Consumer Commission claimed that, by virtue of its role as a search engine and a vertically integrated ad tech provider, Google had the incentive to balance the interests of various stakeholders to ensure the long-term viability of the ecosystem and avoid network externalities that could result in a “tragedy of the commons” – a view Damien and I sharply criticized in our own submission to the ACCC.
However, this is likely the first time that such a justification is actually put to the test before the EU courts, hence my interest on this topic. From what I gather from the hearing, the General Court seems reluctant to follow Google’s reasoning. On the one hand, the Judge Rapporteur observed that the AFAs were perfectly understood inside the “Androidville” (in other words, Google was justified in setting rules to for its platform). On the other hand, the Judge remarked that the AFAs should not have the effect of preventing rival Androidville-like ecosystems from emerging; this goes too far.
On nudges and tying in the digital economy: Another interesting point is how the General Court will address Google’s claims that the pre-installation conditions in the MADAs were not likely to restrict competition, in that users are free to download rival apps. In its decision, the Commission noted that pre-installation results in a “status quo bias” (a concept from behavioural economics), i.e. consumers are likely to stick to what is on the device. When questioned by the Judge Rapporteur, Google’s lawyers refused to acknowledge or contest the concept of “status quo bias,” albeit they seemed critical of the view that consumers may act irrationally; if consumers prefer another search engine, they will simply download it, so the argument goes. If they do not, this proves they prefer Google Search.
This is perhaps the first time the effects of the “status quo bias” will be addressed by the General Court (note that in Microsoft (tying), the focus was on the technical obstacles users faced for downloading rival media players). But I see no reason for the Court to depart from the Commission’s view on this point. There is rich literature on the effects of choice architecture and nudges in consumers’ decisions (see e.g., the classical book of Thaler and Sunstein). In fact, nudging is at the core of Google’s business model (and its vision for how society should function, as explained in detail in Zuboff’s The Age of Surveillance Capitalism). As noted in the hearing by one intervener, Google deploys hundreds of behavioural economists to influence consumer decisions. To shut one’s eyes and simply say that users have the theoretical ability to download rival services would be simply wrong.
On the As Efficient Competitor (AEC) test: Another point concerns the Commission’s application of the AEC test with respect to the RSAs, which was the subject of harsh criticism by Google. Google argued that an AEC would, among others, have captured far beyond a 12% market share in general search (which is the figure the Commission took as its basis in its decision). The Commission, on the other hand, noted that in light of the structure of the market in general search and the importance of scale, it was unlikely (if not impossible) that an AEC could ever emerge in practice. Following the CJEU’s judgment in Intel, the General Court is bound to examine Google’s criticism of the AEC test applied by the Commission.
The Commission does make an important point, namely that in certain cases the structure of the market may make the emergence of an AEC impossible. While the Court of Justice has accepted this in a setting involving a statutory monopoly (Post Danmark II), it will be interesting to see whether the General Court would accept this in the context of a digital market featuring strong economies of scale. On the other hand, it does strike me as a bit odd that the Commission went on to apply an AEC test, despite considering that an AEC could not emerge in practice.
Market definition: A final (yet no less important) issue is that of market definition with respect to smart mobile OSs and app stores. During the hearing, Google’s lawyers asserted that the Commission’s decision ignored all commercial reality in excluding Apple’s iOS and App Store from the relevant markets for smart mobile OSs and app stores respectively. [As a reminder, the Commission took the view that iOS was not in the same market with Android, as it is not licensed to OEMs; it thus examined the relevant market from the point of view of OEMs. However, it also examined the alleged constraint by Apple at the level of users and developers, holding that such “indirect” constraint was insufficient]. Google argued it competes fiercely with Apple for both consumers and developers, hence any degradation in Android’s quality would result in users shifting to iPhones and developers prioritizing iOS. The Google Android OS and Google Play were said to compete together as a “system” against Apple’s iOS and the App Store. Unsurprisingly, reference was also made (although in passing) to the recent ruling of the US District Court (N.D. Cal.) in Epic Games v Apple.
The Commission, on the other hand, pointed out how Apple seems to be cooperating with, rather than competing against, Google, referring inter alia to the multi-billion Safari deal and the internal quote that Apple and Google should work together as “one company.”
Needless to say, the General Court’s ruling on this point could be very important for the Commission’s ongoing investigation into Apple’s App Store rules (where, based on the press release, it appears the Commission has excluded Google Play from the relevant market where the App Store is active).
As readers of this blog are aware, Damien and I have argued that the Google Play and the Apple App Store are not in the same relevant market (note also that the French Autorité de la concurrence earlier this year dismissed the argument about “system competition” between Apple and Google, holding that the Pelikan/Kyocera conditions are not satisfied). The view that users will shift to an Android device in response to a price increase on iPhone apps (or the reverse) is simply implausible, not least because users exhibit strong loyalty and face switching costs that dissuade them from changing ecosystems. Developers, on the other hand, have to be present across both ecosystems to maximize their reach (this is particularly the case for developers relying on network effects). As a result, they are unlikely to switch to one of the two ecosystems in response to a price increase in their fees.
My overall assessment
No one can tell with certainty whether the General Court will side with Google or the Commission. However, based on my own experience from the hearing and the questions from the Judge Rapporteur – and this is my subjective assessment, so please take it with a (large) grain of salt – the General Court did not seem receptive to Google’s arguments, with one exception: the arguments concerning the RSAs (and in particular the application of the AEC test, which the General Court is bound to scrutinize post-Intel). Interestingly, on the fourth day of hearing the Court invited comments on what consequences it should draw in case it partially annuls the Commission’s decisions, for example with respect to the RSAs. Considering how this was an issue brought up by the Court itself, this does make me think the Court could do just that, i.e. partially annul the Commission’s decision as regards the RSAs. On the other hand, the Commission appeared much stronger on all other points – market definition, dominance, the MADAs, and the AFAs. Guess we will have to wait and see if this will indeed be the case! In any event, this was a fascinating hearing, and one I will remember.
[Disclosure: I have advised clients against Google, albeit not for the issues covered by the Google Android decision. The views expressed in this post are solely mine and should not be attributed to any client.]