The very lengthy and complex Google Android decision provides us with lots of material for discussion. In the first part of this discussion the matter of the definition of the relevant market was addressed. Although it may appear that this is the only key issue in the case, the decision covers several other issues, which deserve equal (if not more) attention. Accordingly, this post addresses the tying abuse in the Android case and the (potential insufficient) effects based analysis thereof as well as the remedies imposed in this case that may bring about the extraterritorial application of EU competition law.
The anti-competitive potential of tying in two-sided markets
The Google Android case provides us with a complex scenario of tying in two sided markets. The tying abuse by Google involves namely the tying of the Google Search app with the Play store and the tying of Google Chrome with the Play Store and the Google Search app. Given the fact that all these products entail in essence platforms that interconnect two or more customer groups, the tying of these products constitutes a tying of two (or multi) sided platforms. Although it is rather tempting to assume that tying in such cases is no different than tying in more traditional settings, the economic literature on tying paints a slightly different picture.
Generally speaking, the anti-competitive potential has been considered to manifest in practice through:
- The leveraging of market power across markets and the extraction of supra competitive prices for the tying as well as the tied product or service (the leveraging theory);
- The deterrence of market entry as well as exclusionary effect in the tied product market for (potential) competitors that cannot make similar offers;
- The deterrence of market entry as well as exclusionary effect in the tying product market for (potential) competitors that cannot make similar offers;
- The deterrence of market entry in a third product market for (potential) competitors that would potentially offer a new product or service that may constitute a substitute for the combination of the tying and tied products or services.
The manifestation of such potential harms in practices depends, however, on the circumstances of each case and the market conditions present at the time of the analysis. For example, the Chicago School has shown that under specific market conditions the ability to leverage (monopoly) market power across markets and charge supra-competitive prices in both markets is not profitable (the single monopoly theorem). Later studies showed, however, that when such specific conditions are not present, the risk of leveraging and supra-competitive prices becomes realistic (see e.g. Whinston).
In the context of two-sided markets a similar pattern can be identified. Although the above-mentioned anti-competitive concerns have also been identified in the context of two-sided markets, their profitability and thus manifestation in practice is not certain. Accordingly, in the context of two (or multi) sided markets it has been shown that the profitability of tying depends on the degree of two-sidedness of the markets that are being tied (see here and here). Therefore, when assessing the anti-competitive potential of such tying practices one would expect that this factor be indeed taken into account to some extent. In the context of antitrust procedures this expectation is somewhat justified by the effects based approach promoted by the CJEU in Intel and the general obligation to take into account the entire legal and economic context of the investigated practice before finding an infringement of competition law (e.g. as said in Cartes Bancaires).
The effects based approach in Google Android – do we need to talk about network effects?
The early tying cases of Hilti and Tetra Pak left an impression that tying practices constitute per se abuses that do not require the finding of anticompetitive effects. This approach is considered to have been abandoned in the Microsoft case where the Commission and General Court addressed the effects of the tying practices in the case (see discussion here). In the Google Android decision, the Commission explicitly addressed this issue and noted that the Commission now follows an effects based approach in tying cases (par. 749). The essence of such an effects based approach to tying according to the Commission, seems to boil down to establishing whether such practices are capable of restriction competition (par. 733, 749). When going through the Commission decision one may wonder if the analysis provided by the Commission is indeed sufficient for this purpose.
In the context of the decision, the Commission repeatedly refused to address the matter of (indirect) network effects for the purpose of the analysis. In fact, the Commission explicitly stated the it is not obliged to look into the indirect network effects at play in this case (par. 776). Furthermore, the Commission noted that nothing in the Microsoft judgement requires it to address the matter of indirect network effects when analyzing the effects of tying practices (par. 854, 966). Nevertheless, the Commission did acknowledge the existence of such effects in the case of Google (par. 855). Moreover, the Commission stated it did take into account the different sides of the Android platform (par. 874), however, the decision is not entirely clear on how that was done. Given that the Google products in the Android case all concern some type of two or multi-sided platform, one may question whether this approach towards (indirect) network effects can be justified.
As mentioned above the anti-competitive potential of tying practices in two-sided markets depends greatly on the degree of two-sidedness of the tying and tied products or services. Assessing this nature of the products or services inherently requires looking into the (indirect) network effects at play in each case. It is precisely the presence of indirect network effects as well as their intensity and nature (positive / negative; multilateral / unilateral), which determines to what extent a product or service truly is two-sided. Accordingly, from an economic perspective looking at the indirect network effect at play seems rather inevitable when assessing the anti-competitive potential of tying in two or multi- sided markets. Although economic theories are by no means binding for the purpose of competition law proceedings, there may also be a legal argument in favor of including network effects in the legal analysis when such effects are indeed present and relevant. In the Cartes Bancaires case the CJEU explicitly found that once network effects are identified, a finding of an (object) restriction of competition cannot be made without taking such effects into account when these are relevant to the case (par. 74-77). Although it is not a 102 TFEU case, the requirement in the Cartes Bancaires case is nonetheless relevant- namely, when network effects that are relevant to the issue at hand are identified these must be taken into account. By explicitly choosing not to do so, in contrast to its approach in Microsoft, the Commission may have failed to take the entire economic and legal context of the matter into account when making its assessment of the tying practices. Accordingly, one may wonder whether such a position will survive a judicial review the General Court. Nevertheless, this does not mean that the findings of the Commission are mistaken but rather that the motivation for such findings may not be sufficiently rigorous.
Remedies – an extraterritorial approach?
The remedies in the Android case bring about quite some questions on the effectiveness of the remedies and their potentially extra-territorial scope. The remedies imposed with respect to the tying practices in this case can, unsurprisingly, be summarized as an obligation to untie all the products or services which Google was found to tie according to the Commission (par. 1394-1400). Although the nature of the remedies is rather predictable, their undefined scope is not. Accordingly, the obligation for Google to untie its products or services is not clearly limited to the scope of the EU and /or EEA market. The question is then whether the Commission expects Google to change its business practices with respect to all smartphones and or tablets running Android or only those which are intended to be sold in the EU and/ or EEA market comparable to its approach in Google Shopping (par. 700) and Samsung (par. 114).
The difficulty with this question is that answering it requires us to look into the proportionality and effectiveness of the remedies that serve as the guiding principles in this regard based on Article 7 of Regulation 1/2003. On the one hand, it would appear that obliging Google to terminate all its tying practices worldwide is rather disproportional. After all, if all smartphones and / or tablets that are produced in order to be marketed in the EU/EEA are not subject to the tying practices of Google one may argue that the infringement has been brought to an end with respect to the EU/EEA area. On the other hand, if the aim of the remedies is also to counter the anti-competitive effect of Google’s tying practices one my wonder if limiting the untying obligation to the EU/EEA area may constitute an effective remedy. If the idea behind the remedy is not only to terminate the infringement but also to counter some of its undesired effects by opening up competition (for licensable OS for handheld devices, android compatible appstores and mobile internet browsers) limiting the remedy to the EU/EEA are may not be sufficiently effective. If the restrictions to competition created by Google’s tying practices are only lifted with respect to the EU/EEA, it is not evident that this would be enough to make the affected markets (sufficiently) contestable for actual and potential competitors of Google. Accordingly, it is unclear whether the EU/EEA area offers the actual and potential competitors of Google sufficient scale (or in platform terms – critical mass) in order to effectively compete with Google when it comes to the product markets affected by Google’s tying practices.
This is a rather problematic outcome, as the Commission may not have the jurisdiction to impose a remedy that has a broader scope than the EU/EEA area. However, not having such a wider scope entails that the remedy may fall short in terms of effectiveness, as the negative effect on competition in the EU/EEA may not be removed. In this regard, while the operational part of the remedy is perhaps clearer than in the case of Google Shopping, the jurisdictional scope thereof is far less clear, causing it to be equally questionable in terms of effectiveness.
The Google Android decision will undoubtedly provide competition lawyers and scholars with plenty of discussion material for the foreseeable future. Despite the fact that the discussion of this case has been primarily perused with respect to the market definition, the decision gives rise to several other issues that, in the long run, may prove to be more important. While the debate on the market definition concerns primarily Google, the manner in which the legal analysis of effects should be performed and the jurisdictional scope of competition law remedies are matters that are relevant to all future cases concerning both the digital and non-digital economy.