New publications: Self-Preferencing (World Competition) and Anticompetitive Effects in EU Competition Law (JCLE)
The year is coming to close and, true to form, there is still plenty of fear and uncertainty in the air (even more so if, like myself, you happen to be in the UK).
The above said, looking back to the things we have done over the past twelve months always manages to bring a kernel of satisfaction amid the exceptional circumstances. I was recently reminded of a few of these when I was notified that two of my papers have been released by the journals to which I submitted them.
My paper on Self-Preferencing (see here for the pre-edited version available on ssrn) has recently come out in this year’s last issue of World Competition (see here), which is a great one. Giorgio Monti‘s paper on CK Telecoms definitely deserves a read. The editorial team was lovely and efficient (I am, in fact, proud that a paper of mine finally comes out with them).
My paper on Anticompetitive Effects in EU Competition Law (see here for the pre-edited version) is already available as an advanced article on the website of the Journal of Competition Law & Economics (see here), alongside some inspiring pieces (this one by Stavros Makris was a discovery). The team was not any less amazing (and my paper, full of figures and tables, was admittedly not the easiest one to handle).
My thanks go again to all among you who provided comments on these two pieces (I certainly welcome more). I look forward to sharing more ideas in the coming year (if anything because it is one of the things that keeps us sane).
Chillin'Competition 
Dear Pablo,
The holiday season with lock-down is a perfect opportunity to finally catch up with reading. Below some remarks on the ‘effects’ paper. I would have expected that I would finally understand what ‘effects’ are, but I am not sure I am out of the woods yet. I feel confused. Just three general remarks after reading this revised version.
First, I have the impression that several sections, in particular section 7, conflate the legal test with the standard of proof. A recent post by Andriani Kalintiri (here), that I can only recommend, warned against such approach.
It is, in my view, counter-intuitive to submit that ‘prima facie unlawful conduct’ is subject to a ‘lower’ threshold (7.2.1). I would suggest that if what you call prima facie cases are easier to prove, it is not because a lower threshold applies, but simply because experience/practice shows that they are typically capable/likely (reader’s choice) to produce anti-competitive effects. The fact that the undertakings will have more difficulties to ‘rebut’ the ‘presumptive illegality’ stems from the existence of factual/normative assumptions that such practice is restrictive (typically based on experience or normative objectives), not because a lower threshold is applied. The difference with what you call ‘standard effect analysis’ is not one of threshold, but of the operation of factual/normative assumptions and shifts in the evidential burden of proof.
Conversely, as regards the other extreme situation, what you call ‘enhanced effects’, it is still conceptually possible to have a legal test of ‘exclusion of all (effective) competition’ which must be proven according to a ‘balance of probabilities’ standard. I have found nothing in the case law suggesting that ‘certainty’ is required when that test is applied.
Second, I would have liked to see a bit more about the role of the ‘counter-factual’ when examining effects. Section 5 promises to deal with the counter-factual, but in Generics, in the section on ‘effects’, after recalling that ‘competition should be assessed within the actual context in which it would occur in the absence of the agreement in dispute’, the Court clarified, for the first time in such explicit manner, that ‘the sole purpose of the counter-factual is to establish the realistic possibilities with respect to that manufacturer’s conduct in the absence of the agreement at issue’ (para 120). It is about alternative conduct then, not more widely about market outcomes. I see no trace of this (to me, at least) important clarification in the paper, or any reference to this ‘realistic possibilities’ test (which was already mentioned, in a slightly different context, in Mastercard).
Third, it’s clear Article 102 TFEU does not ensure that competitors less efficient than the undertaking with the dominant position should remain on the market. There is no guarantee against exclusion.
It is less clear, however, if this means that then any conduct which excludes less efficient competitors escapes the prohibition of Article 102 TFEU (as your paper could be interpreted), or if this test only refers to some types of conduct. If ‘the departure of less efficient rivals would be deemed a natural and desirable consequence of the operation of the competitive process, not one that would trigger a prima facie prohibition’ (as you put it), does this mean that any means (tying? misleading representations? vexation litigation? discriminatory conditions?) are justifiable to achieve that ‘desirable’ aim? Does it depend on the means used? Is ‘competition on the merits’ relevant, as suggested by the link made in para 22 of Post Danmark (I) or 134 of Intel?
I note that para 22 of Post Danmark (I) does not state (as cited in 5.2.1) that ‘the exclusion of those firms that are ‘less attractive to consumers from the point of view of, among other things, price, choice, quality or innovation’ does not give rise to anticompetitive effects under Article 102 TFEU’. It says that ‘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’. Not sure one can even interpret much in this sentence, since ‘competition on the merits’ was generally considered lawful, anyway. The Court appears to be just saying that less efficient competitors are not protected by Article 102 TFEU against being driven out of the market by ‘competition on the merits’ (in case somebody had a doubt about it), so ‘no guarantee against exclusion’.
It seems to me that there is a significant difference between (a) stating that Article 102 TFEU does not ‘ensure that competitors less efficient than the undertaking with the dominant position should remain on the market’, or that ‘competition on the merits’ may result in excluding less efficient competitors, and (b) affirming that excluding less efficient rivals is generally (or prima facie) not prohibited (except in one single scenario, that you acknowledge) and that the causal link (which in your view is necessary) is broken if the dominant undertaking can prove that the competitor would be out of the market or marginalized anyway, despite the dominant undertaking’s own conduct.
Tintin
2 January 2021 at 1:08 pm
Thanks so much for taking the time to read the paper and share your comments with us! An author cannot ask for more
You are absolutely right to recommend Andriani Kalintiri’s writings. She is by far the top European scholar on all things standard of proof (and beyond). That is the reason I asked her to write the blog post you mention. It is also the reason why I make sure to share my papers with her (including the one you have kindly taken the time to read).
On your first comment, and the different thresholds of effects: I believe your interpretation of the case law would not be able to explain cases like T-Mobile or Bananas. These cases are interesting precisely because the practices at stake were not likely to have anticompetitive effects (if anything, experience suggests that a single meeting is unlikely to yield restrictive effects, as the referring court explained in its submission in T-Mobile).
In spite of the above, the practices in the two cases were found to be restrictive by object by the Court. It seems to me that the most reasonable conclusion to draw from this outcome (and other cases, such as AKZO) is that the threshold of effects is lower when prima facie unlawful conduct is at stake.
I find your interpretation of Generics intriguing. It would be unusual for the Court to casually reverse its consistent case law in such a way. Your interpretation of that passage (that the analysis of the counterfactual relates to the conduct alone) is contradicted by a plain reading of Budapest Bank, which came out shortly afterwards. If there was any doubt, the Court unambiguously held, in para 55 of that judgment, that:
‘Where the agreement concerned cannot be regarded as having an anticompetitive object, a determination should then be made as to whether that agreement may be considered to be prohibited by reason of the distortion of competition which is its effect. To that end, as the Court has repeatedly held, it is necessary to assess competition within the actual context in which it would occur if that agreement had not existed in order to assess the impact of that agreement on the parameters of competition, such as the price, quantity and quality of the goods or services (see, to that effect, judgment of 11 September 2014, MasterCard and Others v Commission, C‑382/12 P, EU:C:2014:2201, paragraphs 161 and 164 and the case-law cited)’.
It is not immediately obvious to see how your interpretation is tenable following this ruling. If you ask me, para 120 in Generics has to be read together with para 119. Both show that the Court is addressing a very specific point of law (whether the assessment of effects demands an analysis of the probabilities of the generic producer succeeding in patent litigation). Read in its proper context, para 120 of Generics comes across as sound, uncontroversial, and fully in line with the preceding and subsequent case law.
On less efficient competitors, it probably makes sense to focus on the law as interpreted by the Court in individual cases. First, I find that some practices are inherently against competition on the merits and are prima facie unlawful irrespective of their effects. Second, there are practices that are only abusive where they are likely to have anticompetitive effects. If they have no effects, they are valid expressions of competition on the merits.
Starting with the second, it is clear from landmark rulings such as Deutsche Telekom, TeliaSonera, Post Danmark I and Intel that the relevant question is whether the practice is likely to drive an equally efficient competitor out of the market. There is no abuse (and thus no departure from competition on the merits, or at least not in principle) where the practice would not harm an equally efficient competitor’s ability and incentive to thrive.
Other practices are prima facie abusive irrespective of their effects. Even then, the fact that they are capable of driving equally efficient rivals out of the market is an element of the analysis. This has been true pretty much from the outset (paras 71-72 in AKZO come to mind immediately).
Thanks again!
Pablo Ibanez Colomo
2 January 2021 at 5:05 pm
Thanks for your reaction.
On the first point, you mention that my first comment would not be able to explain cases like T-Mobile or Bananas. I do not see why. I do not know why a single meeting is unlikely to yield restrictive effects. In any event, I am afraid that there are even more judgments which cannot be explained by your position.
On para 120 of Generics, my point was the lack of engagement with the notion as reflected in the case law, since the word ‘counter-factual’ is very rarely used by the ECJ, but it was used in Generics and Mastercard, in the sense of alternative (realistic) courses of action. I find intriguing to devote a section to this notion without referring to the judgment(s) where the ECJ actually uses that expression. There is no suggestion in the judgments that the ECJ changed its case law, because it has not. Para 55 of Budapest Bank does not appear to refer to any notion of counter-factual, at least not explicitly or by implication. It simply refers to taking account of ‘actual context’ (uncontroversial, and I was not suggesting ‘actual context’ is irrelevant) and the assessment of the impact of the agreement ‘on the parameters of competition, such as the price, quantity and quality of the goods or services’ (other judgments, such as Intel, refer also to ‘choice’ and ‘innovation’). Yes, analysis of effects must bear in mind several things, but I was focusing on the ‘counter-factual’. By the way, I note that the paper does not assess how ‘price, quantity and quality’ play a role in the assessment of effects.
On less efficient competitors, I completely agree that ‘it probably makes sense to focus on the law as interpreted by the Court in individual cases’, and that’s why I found the quite general statements in the paper disconcerting. I note that you appear to accept that the efficiency of the competitor is only relevant for some types of abuse, not all of them (you appear to mention only price-related cases). It is interesting that you mention Intel in this context, since in Intel the ECJ consistently (for once) uses the term ‘capability’ to define the threshold of effects, the one you consider should not be used.
Tintin
2 January 2021 at 6:56 pm
Thanks very much again for reading the paper and for this discussion! It is really wonderful. Please do not hesitate to send further comments via email.
I pointed out that your first point is incompatible with, and cannot explain, T-Mobile and Bananas. The simple reason is that your approach was proposed by the College van Beroep in its reference in T-Mobile, but was flatly rejected by the Court. I am not saying it would not be a reasonable way to go about the question (it would, and the College van Beroep has a well-deserved reputation as a top specialist tribunal). It just so happens that it is simply not the approach that the Court chose to follow.
What you mean by counterfactual is not immediately clear. Evaluating the counterfactual means the evaluation of the conditions of competition as they ‘would occur in the absence of the agreement in dispute’. In other words, it means contrasting the conditions of competition with and without the agreement. This is not only the plain meaning of the word counterfactual, but also the sense attached to the notion in the the case law, including Generics (paras 118-119) and Budapest Bank (para 55).
On efficient competitors, there is nothing in the case law suggesting that the principle whereby the exclusion of less efficient competitors is unproblematic is confined to what you call price-related cases. In fact, the Court was careful to refer to various parameters of competition when laying down the principle in Post Danmark I and Intel, thereby strongly suggesting that the principle applies across the board.
Finally: Intel referring to capability is consistent with the interpretation of the case law I suggest in the paper. Since exclusive dealing is prima facie abusive irrespective of its effects, one should expect the Court to apply this lower threshold, as it does. Which, funnily enough, takes us back, full circle, to the first comment you made.
Thanks so much again! And all the best for 2021!
Pablo Ibanez Colomo
3 January 2021 at 12:17 pm
Thank you again for your reaction, which does not take us very far though.
Indeed, the ECJ did not follow the referring court in T-Mobile, but not because the ECJ suggested in any way that a single meeting is not likely to restrict competition, as you have interpreted it apparently. The ECJ applied the ‘capability’ threshold (‘apte à’ in the French version), but this cannot be interpreted, a contrario, as meaning that the ECJ thought that a higher threshold would not have been met by the practice at issue.
The general point I made is that your distinction between a ‘capability’ criterion for some practices and a ‘likely’ criterion for others is not very convincing to me. That’s all. The test could be ‘capability’ or ‘likelihood’, I left that to the reader’s choice, but I would find it odd that there are actually two distinct ones, even if the English versions may differ sometimes. The French versions (the language in which judgments are originally written and deliberated) tend to use ‘susceptible de’ for ‘likely’ which appears to be less ‘strong’, and in T-Mobile the ECJ said ‘apte à’, so one may wonder if ‘apte à’ and ‘susceptible de’ really convey fundamentally different standards, as you suggest. I note that AG Kokott in Post Danmark II (that you prominently mention in section 7.2.2) did not find two different thresholds in the case law, but only one, and tried to reconcile the language in different judgments. More fundamentally, if ‘prima facie unlawfulness’ cases would be subject to a threshold which is easier to comply with than other cases, what made the conduct in question ‘prima facie unlawful’ in the first place?
The more general point I made was also that your approach conflated the substantive legal test and the standard of proof. The allegation that ‘certainty’ is required as regards ‘refusal to supply’ is a good example.
On counter-factual, I note your notion also refers to the alternative ‘but for’ conduct (‘would occur in the absence of the agreement in dispute’). The point was that you do not engage with the ‘realistic possibilities’ test (does ‘realistic’ equal ‘likely’ or ‘plausible’?), and whether it plays a role as Generics seems to suggest.
You now suggest that the ‘less efficient’ competitor approach may apply in non-price cases. Maybe, we do not know. So far, the ECJ has not done it, and I see no reference to it outside price-related cases, so I would not assume that it plays a role if in decades the ECJ has not mentioned it. You do not give any theoretical example either, outside the price-related cases you mentioned, where a dominant undertaking would escape the prohibition of Article 102 TFEU because the competitor was less efficient. Your paper suggested that it applied across the board, and I reacted to that. Now I do not know what your position is any more. Para 22 of Post Danmark I, later recalled in other judgments, has clearly a very specific meaning (not the one you appeared to give to it), as I explained in my previous comment.
I wish you all the best for 2021 as well. I think we can keep it there for the moment, the reader can read the judgments and make up her mind about these issues.
Tintin
3 January 2021 at 2:08 pm
WIthout engaging in other aspects worth discussing and putting into other light, two aspects from my point of view should be emphasized additionally.
Since I have read the argument of “failing” remedies applied by the Commission in Art. 102 TFEU cases by referring to Microsoft I/Media Player tying case you cited in your paper now a lot of times, it is worth pointing out how successful the Commission was when implementing the remedies related to the Internet Explorer choice (which is quite similar to the Media Player case). Maybe learnings from the past, besides (in any case) never an argument against enforcement of competition rules as seemingly intended in your paper. In detail:
“The choice screen was provided as of March 2010 to European Windows users who have Internet Explorer set as their default web browser. While it was implemented, the choice screen was very successful with users: for example, until November 2010, 84 million browsers were downloaded through it. When the failure to comply was detected and documented in July 2012, the Commission opened an investigation (see IP/12/800) and before taking a decision notified to Microsoft its formal objections in October 2012 (see IP/12/1149).”
see: https://ec.europa.eu/commission/presscorner/detail/en/IP_13_196
Which means, within 9 months, 84m free decisions for pre-installed browsers.
Second point, related to the argument that Commission decisions in Art. 102 TFEU cases could hamper dynamic efficiency and incentives for innovations by undertakings concerned, as stated in your article: often used argument in this context, however, I would be keen to read a footnote ore quote relying on real world economic evidence for this normative high level argument. Real world evidence for the opposite: Microsoft had not stopped to be innovative after the Commission decisions in 2004 or 2009; neither did Google after the shopping case (etc. the list goes on).
Christian
25 January 2021 at 9:23 am