Chillin'Competition

The Android judgment was widely awaited, not only because of its implications for the Commission’s policy, but because the underlying legal issues were intriguing and, in many respects, completely new. It was anything but a run-of-the-mill tying case.

In its decision, the Commission ventured where it does so only rarely: the core of a company’s business model. It did not question a peripheral aspect of the company’s strategy, but the central mechanism through which it appropriated the value generated by its activity.

In addition, there were some practices (such as the so-called AFA) for which there were no obvious precedents (at least not in the context of Article 102 TFEU enforcement).

The General Court’s ruling had thus the potential to clarify (and address for the first time) a number of issues. The impression one gets is that only some of them have definitely become clearer, and confirm the trend of the past few years. Others (perhaps the most novel and exciting ones) might require additional fine-tuning.

The points where the judgment brings clarity (if only because it confirms past trends) include the following:

• Tying is subject to an effects analysis (at least sometimes).
• The judgment confirms what an effect is not.
• We know that the coverage of a practice must be significant for an effect to exist (and this is different from appreciability and de minimis).

The points that remain to be clarified, in my view, comprise:

• The point at which an anticompetitive effect can be said to exist.
• The benchmark against which effects are assessed.
• The analysis of the attributability of anticompetitive effects.

What is clearer after the Android judgment

Tying is sometimes subject to an effects analysis

If I had been asked, before the Android judgment, whether tying is a ‘by object’ or ‘by effect’ practice, I would have replied that it is abusive irrespective of its impact on competition (that is, by object). The cases that made it all the way to the Court of Justice may be old, but they state what the law is.

Now, I would be more cautious when making statements about the legal status of tying. If the Court of Justice were to be asked about it (and it may not take very long for this to happen), it is more likely than not that it will hold that an effects analysis is a precondition for a finding of infringement (at least so in some instances).

The General Court openly acknowledges that, in the specific circumstances of the Android case, the anticompetitive impact of the tying conduct could not simply be presumed, and had to be established in light of the relevant economic and legal context (para 295, where it concludes that the Commission had been right to assess the effects of the practice it its decision and draws analogies with Microsoft).

The judgment confirms what an effect is not

The past decade gave us clarity about a key aspect of the analysis under Article 102 TFUE: an anticompetitive effect is more than a competitive advantage (or disadvantage) and more than a limitation of a firm’s freedom of action. In this sense, paras 280-282 of the judgment are particularly instructive.

These paragraphs are consistent with the case law, which has long suggested that an exclusionary effect exists where the impact of a practice is such that it hinders the ability and incentive of rivals to compete (and ‘thus allow[s] that [dominant] undertaking negatively to influence, to its own advantage and to the detriment of consumers, the various parameters of competition, such as price, production, innovation, variety or quality of goods or services‘, as explained in para 281).

This clarification marks a welcome departure from the tying aspects of the Microsoft case (and one that seemed inevitable, given the evolution of the case law). In that ruling, the General Court famously held that the mere fact that tying gives a competitive advantage to the the dominant firm is sufficient to establish foreclosure.

The Android judgment confirms that Microsoft does not reflect the case law. Thus, an authority would have to dig deeper to establish foreclosure to the requisite legal standard. How much deeper, you may be asking? This is precisely one of the points where we need more clarity. At least in relation to tying, drawing the line between a mere competitive advantage and an anticompetitive effect remains an exercise for which we have few tools.

My sense is that we need a set of workable proxies to spot anticompetitive effects in relation to several non-pricing behaviours, including tying. A meaningful assessment of the impact of a practice can only be conducted in light of proxies like coverage, which are both clear and administrable. In their absence, the whole evaluation is shrouded, by and large, in mystery. Which takes me to my next point.

The coverage of a practice must be significant for an effect to exist (and no, this is not about de minimis)

As was true in the Intel renvoi, the General Court ruled that the Commission decision had failed to appropriately consider the coverage of the exclusivity-like practices in Android (the so-called RSA restraints). The judgment, in fact, is incredibly valuable in this respect.

The General Court concludes that, properly considered, the coverage of RSAs was far from significant, in that they merely encompassed [10-20]% of the national markets for general search services. Such figures would be insufficient to limit rivals’ ability and incentive to compete.

At this point, you may be asking: is the requirement of significant coverage not an endorsement of the de minimis doctrine, which the Court expressly (and rightly) rejected in Post Danmark II? If so, is the General Court deviating from the case law? I do not believe so, for the reasons explained, a while ago, in this paper.

What Post Danmark II held is that, where an anticompetitive effect is established in an Article 102 TFEU case, such an effect will by definition be appreciable. Since the de minimis doctrine has always been about the market power of the firms involved in a practice, it has no place in scenarios that presuppose a substantial degree of market power.

What Post Danmark II did not hold, however, is that anything that dominant firms do amounts to an effect. This is where the confusion lies.

Holding that the effects, if established, will be appreciable, does not say anything about what an effect is. The latter is a separate question, and one that precedes the issue of de minimis. Thanks to the Intel saga, Android and Qualcomm, we know that (i) not everything is an effect and that (ii) if the coverage is insignificant an effect does not exist (and therefore appreciability does not even come into the picture).

I will follow up with a second part, since this post is already too long. In the meantime, please let me know your thoughts (nothing to disclose, by the way).