Open source but anti-competitive? A discussion of the EC’s case against Google

Genna Robb

On the 20th of April 2016, the European Commission issued a Statement of Objections explaining its preliminary view that Google has behaved anti-competitively (abused its dominance) in the EU in respect of its Android operating system (OS) and apps. The Statement of Objections does not constitute a finding against Google; it is only a preliminary view and the Commission will investigate further before coming to a final conclusion. In the meantime, it is interesting to unpack the EC’s objections and consider some of the possible counter arguments that Google might make.

So what is the EC concerned about? In a case which has parallels with antitrust enforcement against Microsoft in the early 2000s, the EC appears to be worried that Google is trying to make it harder for competitors to introduce rival apps, services and even OSs, by forcing device manufacturers to pre-install the full range of Google apps, and restricting their ability to experiment with rival OSs developed using the open source version of Android (so-called “Android forks”). The Statement of Objections describes three specific allegations of abusive conduct:

1. Google obliges manufacturers who wish to pre-install the Google Play Store on their devices to also pre-install Google Search and set it up as the default search app on those devices, and to also pre-install the Google Chrome browser. Manufacturers don’t have a lot of options in terms of alternative Android app stores, and the EC argues that this policy effectively ensures that “the significant majority” of devices sold in the European Economic Area (EEA) come with Google Search and Google Chrome pre-installed. The EC’s view is that this conduct amounts to an abuse of dominance with anti-competitive effects on competitors and consumers. By tying Play with Search and Chrome, it believes Google has limited manufacturers’ freedom to choose which bundle of apps they pre-install on their devices which has the effect of protecting and strengthening Google’s dominant position in the search market.

2. If a manufacturer wishes to pre-install Google proprietary apps including Google Play and Google Search on any of its devices (which, as discussed in the previous point, they almost certainly do), Google requires it to enter into an Anti-Fragmentation Agreement that commits it not to sell devices running on Android forks. The EC argues that through doing this, Google has denied consumers access to smart phones with alternative and potentially superior versions of the Android OS and which could have had potential to become a credible alternative to Google’s Android OS. This in turn may make it more difficult for rivals to introduce competing apps and services which could be pre-installed on Android forks (without backing by major handset manufacturers, relatively small numbers of Android forks are going to end up in the hands of consumers).

3. Finally, Google offers OEMs and Mobile Network Operators (MNOs) incentives for exclusively pre-installing Google Search. The EC does not go into detail on this point, but presumably considers that the incentives could lead to the foreclosure of competing search apps from the market by preventing them from being pre-installed on devices.

When considering such behaviour, it is important to keep in mind that both bundling together the sale of different products, and the provision of incentives to distributors, are widespread business practices, which, in the vast majority of cases, have no anti-competitive intent or effect. In order for conduct of this nature to be anti-competitive, the firm engaging in the behaviour must be dominant, and there must be a coherent theory of harm which explains why the (already dominant) firm would have the incentive and ability to engage in conduct which will result in harm to competition. The behaviour may also be justified by benefits to customers or efficiency gains which cannot be achieved in any other way, and these justifications must be weighed against any likely anti-competitive effects.

So it is not enough to observe behaviour which could make competitors’ lives difficult, there needs to be a careful consideration of the economics of the behaviour. So far, the EC has only released a short memo on the investigation, so we are confined to a very limited set of information. The memo clearly states, however, that the EC does believe Google to be dominant in each of the relevant markets, and that its behaviour has constrained the development of competing products. We can think of at least a few objections which Google may raise to this.

Google states that it makes money from advertising and other income associated with its apps, and not from the OS which manufacturers can use for free[1]. This implies that Google may need to insist on the distribution of Play and Search with the OS in order to make its overall offering viable. In addition, the markets under discussion here are two-sided markets – that is, there are two sets of customers being served by the same product whose demand is interrelated. In terms of Search for example, Google “sells” search services to consumers (although these services are priced at zero), but also sells advertising through its search platform to companies. The more consumers that use Search, presumably the more Google can charge companies to advertise with it. Currently Google’s share of US mobile advertising revenue is 33.2%, down from 37.7% in 2013.[2] This implies that Google’s incentive should be to ensure that its proprietary apps are distributed and used as widely as possible, not to refuse manufacturers access to these apps. Again, this tends to suggest that the bundling together of the OS with the apps is required in order to make Google’s business model work. An alternative interpretation is that Google refuses to give OEMs who run forks access to its apps so as to raise the barriers to entry into the OS market, as prospective competitors to Google would also need to design a competing app store and ecosystem of apps which can compete with Google’s. But this could also be a short-sighted strategy. It seems credible that OEMs might actually do this, and bypass Google completely (an example of this which already exists is Amazon’s Kindle Fire) which could be much worse for Google in the long term.

Google itself makes an additional argument, which is that the Anti-Fragmentation Agreement ensures that apps work across all versions of Android, thus preserving a positive user experience.[3] It also points out that Apple, Microsoft and other mobile ecosystems also come preloaded with “baseline” apps, and iOS devices in particular come with a much larger number of Apple apps pre-installed than is the case for Google apps pre-installed on Android devices. Of course, by the EC’s calculation, neither Apple nor Microsoft have a dominant position in the relevant markets, but Google’s argument seems to be that Android devices need to be shipped with these apps in order to compete with these other devices.

An additional question which should be asked is, even if Google did have anti-competitive intent in enforcing these restrictions, would this be likely to foreclose competition and harm consumers? This begs the question of whether rivals with a really good browser or search app could be successful in spite of Google’s behaviour. From a technical perspective, it seems there is no real reason why not. A new browser or search app could also come preloaded with most devices (although may not be set as the default) or could be downloaded by consumers at a later stage from the Google Play app store. The EC argues, however, that this is unlikely: “the Commission’s analysis has shown that consumers rarely download applications that would provide the same functionality as an app that is already pre-installed (unless the pre-installed app is of particularly poor quality)”. We don’t know what they based this analysis on, but it seems quite a pessimistic view. It may take time, but if a new app offers users a better experience, it seems likely it will develop a following. Also, by the EC’s own admission, Google Search is currently pre-installed on the majority of handsets in the EEC, and thus another interpretation is that the Google apps which currently come pre-installed are the best products available, and that’s why people don’t download alternative apps.

Then we come to the possibility of developing a whole new OS, app store and app ecosystem. How possible and likely is this? Some suggest that this would be extremely difficult due to the range of alternative products that would need to be developed simultaneously.[4] But is it impossible? Google had to develop all of these things from scratch when it entered the market (although possibly not simultaneously) and one would imagine this wouldn’t be beyond some of Google’s larger competitors (Microsoft for example).

We have talked a lot about the bundling elements of the case, but what about the incentives offered to OEMs for exclusively loading Google apps? From an economic perspective, this depends on how effective those incentives have been, and thus what proportion of devices are foreclosed to competitors. If this is a large proportion, a further question is whether potential competitors could match the incentive payments. However, the exclusivity still only applies to the pre-loading of apps, so again the behaviour should not lessen competition if consumers are willing to download additional or alternative apps.

It will be interesting to see the EC’s final decision on this matter, which will make it possible to engage in more detail with the evidence against Google. These markets are subject to network effects and scale economies which can lead to a significant advantage for first-movers, and which make life difficult for second-comers to the market. In this context, and where there are large dominant firms, it is important to ensure that competition is not unfair. However, it is also important to base decisions on a detailed economic analysis which focusses on the incentives of the dominant firm and the likely or demonstrated effects of its behaviour. Over-enforcement in fast-moving technology markets can dampen innovation by lowering the perceived payoffs from competing successfully. Not to mention that in such markets, disruptive innovation can completely alter the competitive landscape in a short space of time, potentially rendering competition interventions obsolete.

Google’s market share in operating systems was zero in 2008[5], so its alleged dominance of that market is a very recently acquired position, and its rapid rise to prominence suggests that it provided the market with a superior product which consumers wanted. Figure 1 shows that as recently as 2010, Android was one of the smaller OS’s in the smartphone market globally, and overcame significant competitors – Microsoft, Blackberry, Apple and Symbian – to become the most widely-used OS in the market.

Figure 1: OS market shares by worldwide smartphone sales to end users, 2010 – 2016

 

Whilst Google’s position in the market does look daunting in 2016, it is plausible that it has attained it largely through providing market-beating products. Who is to say that another revolutionary product or bundle of products isn’t just around the corner? The EC has a difficult line to tread.



[1] Google Europe Blog: Android’s model of open innovation. 20 April 2016. Available: http://googlepolicyeurope.blogspot.co.za/2016/04/androids­model­of­open­innovation.html

[2] U.S. Digital Advertising Landscape And Key Players (Part 2). Forbes, 9 July 2015. Available: http://www.forbes.com/sites/greatspeculations/2015/07/09/u­s­digital­advertising­landscape­and­key­players­part­2/#7b35c0e673ed

[3] Google Official Blog: Android has helped create more choice and innovation on mobile than ever before. April 15 2015. Available: https://googleblog.blogspot.co.za/2015/04/android-has-helped-create-more-choice.html

[5] “Europe v Google: Android attack” The Economist. 23 April 2016. Available at: http://www.economist.com/node/21697193