Relaxing whilst doing Competition Law is not an Oxymoron

Archive for September 2017

More on Intel: some thoughts after the IBA Conference in Florence

with 5 comments

Santa Maria Novella

Last Friday I took part in an (ideally timed) panel on Intel at the IBA Conference in Florence. The panel was chaired by Thomas Janssens and Timur Bondaryev (thanks a lot to both, and the rest of the panellists!). I learnt a great deal.

Inevitably, I spent some time re-reading some decisions and judgments ahead of the event. In the process, I shaped my thoughts on Intel and its implications.

These are some key takeaways that I did not develop at any length in my first post, and which I believe are key:

The AEC test is a (valuable) proxy; it is not the only factor

One of the issues that came up in the conference (no surprise) is whether the AEC test can rule out, in and of itself, the existence of an abuse. As I understand the judgment, the AEC test is better understood as a proxy or filter, not as the end of the inquiry.

If the test suggests that an equally efficient competitor would not be forced to sell below cost (read: below ATC or LRAIC), the onus is on the Commission to show why the rebate scheme is capable of foreclosing equally efficient rivals. Anticompetitive harm is possible in such an instance, but it cannot simply be presumed. It has to be substantiated in light of the factors identified by the Court in Intel.

Post Danmark II is an exception; Post Danmark I is the rule

When Post Danmark II came out, some commentators claimed that the judgment made efficiency considerations irrelevant in practice. Intel suggests that an alternative reading of the judgment is probably more reasonable: as a matter of principle, Article 102 TFEU is concerned with the ability and incentive of equally efficient rivals to compete. This makes Post Danmark I the rule, and Post Danmark II the exception.

Post Danmark II is in fact a wonderful example of the sort of special circumstances that may justify a departure from the rule. In that case, Post Danmark’s position came close to a monopoly; this position was, moreover, protected by exclusive rights.

A victory for consistency and legal certainty

A strict stance towards exclusivity agreements and loyalty rebates has often been defended in the name of legal certainty. Is there something better for legal certainty, the argument goes, than a clear rule that states that X is unlawful?

I have never been persuaded by this claim. Those advancing the argument only pay attention to one side of the rule (the outcome) and lose perspective of the other side (the scope of the rule, or trigger).

And the scope of the rule in Hoffmann-La Roche (and Michelin I, and British Airways) has never been clear. Look no further than Intel: it is far from uncontroversial to say that all schemes in the case are truly conditional on customers obtaining ‘all or most’ of their requirements from the dominant supplier.

For a rule to provide legal certainty, both the scope and the outcome need to be clear and predictable: make the first vague and the only certainty is that an undefined but potentially very vast range of practices is caught by the prohibition.

The Court’s clarification in Intel (and I believe it is a clarification) is a valuable step towards legal certainty.

Why? Intel provides a uniform benchmark: companies know that a cost-based test can be confidently relied upon as a proxy across the board. They also know that the cost-based test will be, in practice, the starting point of any inquiry.

Intel also makes it clear that the legal characterisation of a rebate scheme no longer has fatal consequences: it does not really matter (or not that much) whether a scheme is qualified as a loyalty or a ‘third category’ rebate. The methodology and approach will be roughly the same in practice.

Capability = Plausibility

As I understand the law, the threshold of capability is a relatively low one (on this one, I understand I agree with the Commission’s submission). In a ‘by object’ case (both under Articles 101 and 102 TFEU), it is sufficient that anticompetitive harm is plausible.

By the same token, what a firm would need to show to rebut the presumption of harm is that anticompetitive effects are implausible.

The good thing about rebates and predatory pricing is that we know where the threshold of capability lies: if a practice does not force an equally efficient rival to sell below cost (again: ATC or LRAIC), anticompetitive effects are in principle implausible: absent other factors, an equally efficient rival would be in a position to match the prices offered by the dominant firm. Its ability and incentive to compete would not be affected.

Exclusive dealing and loyalty rebates are not hardcore cartels

A hardcore cartel worthy of the name is capable, always and everywhere, of having restrictive effects on competition – otherwise, the cartel would have no point. As Toshiba suggests, arguing that a genuine hardcore cartel is incapable of restricting competition is hopeless in the vast majority of, if not all, cases.

But exclusivity agreements and loyalty rebate schemes are not hardcore cartels. These practices are implemented even when anticompetitive harm is implausible – which is also the reason why small market players with little or no market power resort to them.

The acknowledgement of this reality in Intel is valuable: the Court makes explicit that not all restrictions by object are created equal and that the law should take this factor into account. What is true of exclusive dealing is also true, inter alia, of tying and RPM.

What about de minimis?

The Court held in Intel (para 139) that a dominant firm may challenge the capability of harm on the basis, inter alia, of the limited market coverage of the practice. Is there a contradiction between this principle and Post Danmark II, where the Court held that there is no de minimis threshold in Article 102?

I do not believe there is a contradiction. Once again, it is worth taking a look at how things work in the context of Article 101 TFEU.

We know from Expedia that a ‘by object’ infringement that affects trade between Member States is not de minimis. But we also know from Murphy that the parties to a prima facie ‘by object’ infringement can show that the practice is not capable of restricting competition. I have never heard anyone argue that there is a contradiction between the two.

The same system would operate in the context of exclusivity agreements and loyalty rebates under Article 102 TFEU. A claimant or a competition authority can build a prima facie case without showing that the practice is capable of having appreciable effects. This is the point the Court made in Post Danmark II. On the other hand, Intel clarifies that dominant firm(s) can show that the practice is incapable of having effects. The former does not preclude the latter (it should not).

Applications of Intel: intellectual property-related cases are the obvious candidates

Intellectual property rights may prevent actual or potential competition between firms. Competition between these firms, in other words, may be implausible due to the existence of intellectual property rights.

Cases involving the exploitation of these rights are thus an ideal arena in which firms are likely to rebut with success the presumption that a practice is capable of restricting competition.

Examples? Think of the disputes in pay-for-delay cases (Servier, Lundbeck), in which the disputes revolve essentially around whether patent protection made market entry by generic producers implausible .

Written by Pablo Ibanez Colomo

12 September 2017 at 2:26 pm

Posted in Uncategorized

More questions (and some answers) on, and beyond, Intel (C-413/14 P)

with 2 comments


The best, and quickest, Intel comment by the most preeminent academic of his generation –who also happens to be my co-blogger- already contains the keys to understand and make sense out of the very essence of today’s CJEU (Grand Chamber) Judgment in Intel (see the preceding post). In what follows I add my two cents on another set of issues raised by the Judgment. Read the rest of this entry »

Written by Alfonso Lamadrid

6 September 2017 at 5:43 pm

Posted in Uncategorized

Comments on Case C‑413/14 P, Intel: presumptions, effects-based analysis and open questions

with 15 comments

Intel Reports Quarterly Earnings

There was a lot of hype about the appeal judgment in Intel. It proved to be justified. The Court of Justice has set aside the ruling of the GC, and it has done so on the issues that have proved to be more controversial in the past few years: the question of whether, and to what extent, it is necessary to evaluate the effects of a system of loyalty rebates on competition.

Other grounds of appeal, including the territoriality question and the rights of defence, were rejected by the Court.

I will focus on the meaty stuff – Alfonso will jump in later and add his thoughts.

Does the judgment change the law?

Not really. The principle whereby exclusive dealing and loyalty rebates are prima facie abusive (or ‘by object’) stands (see para 137). What is new(ish) then? Well, the Court now clarifies that it is possible for a dominant firm to rebut the presumption that the rebate scheme is ‘capable’ of restricting competition. See para. 138: ‘that case-law [Hoffmann-La Roche and others] must be further clarified in the case where the undertaking concerned submits, during the administrative procedure, on the basis of supporting evidence, that its conduct was not capable of restricting competition and, in particular, of producing the alleged foreclosure effects’.

The Court makes explicit that anticompetitive harm is simply presumed in exclusive dealing and loyalty rebate cases. Accordingly, where supporting evidence is produced, the Commission must take seriously any arguments showing that the practice is not capable of having effects on competition.

Revolution? No. More of a desirable clarification, which makes a lot of sense.

Think of restrictions by object under Article 101(1) TFEU. In that context, the parties can also show that an agreement is not capable of having restrictive effects on competition, and therefore escape the prohibition (see for instance Murphy, paras 140 and 143). Does it mean that the Commission needs to show that the agreement has effects? No, it means that the parties can show that the practice cannot have effects. We now know that the same principle applies in Article 102 TFEU.

The Court’s clarification is welcome. A good legal system applies the same basic principles across the board. In EU competition law, it is now clear beyond doubt that ‘by object’ practices (read: those practices that are prima facie prohibited without the need to show effects) are treated in the same way under Articles 101 and 102 TFEU.

Efficiency counts in Article 102 TFEU, also in rebate cases: a welcome end to a controversy

We already knew from the Post Danmark saga – well, even from AKZO – that Article 102 TFEU is not inimical to efficiency considerations. On the contrary. The Court had already declared that the exclusion of less efficient rivals is a natural consequence of the competitive process, and therefore not attributable to the behaviour of the dominant firm.

The Intel judgment reiterates these principles and, by doing so, it gives them an aura of generality that is welcome for our understanding of Article 102 TFEU. In para 133, the Court holds that ‘it must be borne in mind that it is in no way the purpose of Article 102 TFEU to prevent an undertaking from acquiring, on its own merits, the dominant position on a market. Nor does that provision seek to ensure that competitors less efficient than the undertaking with the dominant position should remain on the market’.

It goes on in para 134, where the Court holds that ‘not every exclusionary effect is necessarily detrimental to competition. Competition on the merits may, by definition, lead to the departure from the market or the marginalisation of competitors that are less efficient and so less attractive to consumers from the point of view of, among other things, price, choice, quality or innovation’.

However, the most important bit comes in para 139. Post Danmark I and AKZO were about aggressive pricing. Do efficiency considerations count in rebate cases, where the concern is about distribution and access to outlets? They do. The judgment is crystal clear in this regard: when assessing the capability of harm, the Commission is ‘also acquired [sic, I assume it means required] to assess the possible existence of a strategy aiming to exclude competitors that are at least as efficient as the dominant undertaking from the market’.

This declaration was very important: I read it with relief. Saying that efficiency considerations are relevant only in one area of Article 102 TFEU but not in other areas seemed indefensible.

One important implication of the judgment: it is clear from AKZO and Post Danmark I that a practice is not capable of having exclusionary effects if it does not require equally efficient rivals to sell below cost. It follows, I would say, that a rebate scheme that does not force equally efficient rivals to sell below cost is prima facie compatible with Article 102 TFEU.

Open question: what is capability? How is the threshold of capability met?

The crucial part of the judgment (paras 129-147) is carefully crafted. Every word counts and is in the right place.

One aspect that I note is that the Court only uses the word ‘capability’. The open question then is: what is capability? Is it the same as likelihood? Is the meaning a different one? If the Commission needs to assess the capability of harm, does it mean that the tripartite division between loyalty, quantity and ‘third category’ rebates will disappear in practice?

We do not seem to have clear answers in the judgment. This said, it is not necessarily bad that the Court leaves the issue open.

On this point (capability vs likelihood), I happen to agree with the Commission submission in the case: capability and likelihood are not synonymous. They mean different things, and it would make little sense to give the same meaning to the two. The line between ‘by object’ and ‘by effect’ infringements (or between loyalty and ‘third category’ rebates) would otherwise become completely blurred. And this is not what the Court is declaring. Para 137 suggests that the tripartite division stands as a matter of principle.

The threshold of capability applies to ‘by object’ infringements, where harm is presumed (this is true of both Articles 101 and 102 TFEU). The threshold of likelihood, which is higher, applies to ‘by effect’ cases, where harm needs to be established on a case-by-case basis. To distinguish between the two: think of T-Mobile (by object, capability), and Delimitis (by effect, likelihood).

My views on the question: the assessment of capability is not and cannot be as detailed as the analysis found in ‘by effect’ cases. But capability plays a role! Just think of Post Danmark I. I discussed the difference between the two thresholds in Oxford back in June. Check my PPT here.

I am sure there will be a lot of commentary on this question in the coming months. All I can say for the time being is that, if I understand it correctly, what the Court has done makes a lot of sense. And that I will try to tease out its meaning and implications. Stay tuned, and let me know your thoughts!


Written by Pablo Ibanez Colomo

6 September 2017 at 10:52 am

Posted in Uncategorized