Just as we were ready to go on our summer holidays, or even after some of us had already left, the Commission published its draft Guidelines on exclusionary abuses (you can find them here in all official languages). This was a step that was long awaited with some anxiety. The draft Guidelines were expected a bit earlier than August but at the last minute the schedule was hijacked by President von der Leyen, who considered this an important policy document and ordered everything to cease, for as long as she had not been approved by the European Parliament.
Anyway, the draft Guidelines are now public and there is a consultation period running until the end of October. I encourage you all to participate in the consultation.
In this first (very) critical look, I stay at a relatively high level and take a step back by focussing on guiding principles and the broad picture. It is there where the draft Guidelines are intellectually disappointing. Their main purpose is not to offer in a positive way the Commission’s take of the way Article 102 TFEU should be enforced but rather act negatively, in the sense of exorcising the 2008 Guidance Paper. While abandoning the general framework and guiding principles of the latter, the draft Guidelines offer in return no guiding principle at all. Instead, they amount to a wordy and legalistic amalgam of dictums from the case law, usually taken out of context in a rather selective manner, and a casuistic and formalistic analysis. Contrary to the Guidance Paper, when reading the draft Guidelines, we do not really know what an exclusionary abuse is and, more importantly, why the law should prohibit such practices. In short, the impression is that these practices are there just because the case law has told us that they have to be treated in a certain way. This ultra-positivist approach, however, does little to explain or indeed enhance the legitimacy of the overall system of Article 102 TFEU.
I focus my criticism of the draft Guidelines on what I call as “four deadly sins”. My “accusatory” style means no disrespect to the Commission as an institution or to the persons that are behind the drafting of the new Guidelines but rather expresses the genuine disappointment of somebody who considers himself to be an Article 102 TFEU aficionado for many years.
1. First Deadly Sin: The regrettable abandonment of the principle of “anti-competitive foreclosure”
Intellectually, the most disappointing element is the abandonment of the cardinal principle of “anti-competitive foreclosure” of the 2008 Guidance Paper. That principle heralded the moving from formalism to an economic approach. It meant focusing not on protection of the commercial freedom of competitors as such, but rather on the foreclosure of competitors that leads to consumer harm. It is consumer harm that makes foreclosure “anti-competitive”. Foreclosure, as such, is a neutral term: it can be pro-competitive, when there is no consumer harm, or anti-competitive, when there is consumer harm. All this was encapsulated in the cardinal paragraph 19 of the Guidance Paper. By the way, paragraphs 19 and 20 were the most important paragraphs. If the Guidance Paper had to be limited to one page, these paragraphs by themselves would suffice. Paragraph 19 of the Guidance Paper went as follows:
“The aim of the Commission’s enforcement activity in relation to exclusionary conduct is to ensure that dominant undertakings do not impair effective competition by foreclosing their competitors in an anti-competitive way, thus having an adverse impact on consumer welfare, whether in the form of higher price levels than would have otherwise prevailed or in some other form such as limiting quality or reducing consumer choice. In this document the term ‘anti-competitive foreclosure’ is used to describe a situation where effective access of actual or potential competitors to supplies or markets is hampered or eliminated as a result of the conduct of the dominant undertaking whereby the dominant undertaking is likely to be in a position to profitably increase prices to the detriment of consumers.”
This approach proved to be anathema for the draft Guidelines. They now contain a specific section that I could call “The anti-Guidance Paper Exorcism”, that’s Section 3.3.4, which has the somewhat innocuous sounding title “Elements that are not necessary to show the capability to produce exclusionary effects”. I am sorry to say, this Section misrepresents the Guidance Paper in many respects and attempts to discredit it in an indirect way: for example when saying that “finding that a conduct is capable of restricting competition does not require establishing the profitability of the conduct at issue”, paragraph 71 of the draft Guidelines twists the meaning of paragraph 19 of the Guidance Paper. Or when paragraph 73 of the draft Guidelines says that “[t]he assessment […] does not require showing that the actual or potential competitors that are affected by the conduct are as efficient as the dominant undertaking”, again we have a twisting of what the Guidance Paper advocated. Indeed, it never suggested that we intervene only if the victims are as efficient as the dominant company… But the most disappointing of all is paragraph 72 of the draft Guidelines: “It is equally not necessary to prove that the conduct resulted in direct consumer harm, in other words that the dominant undertaking has effectively influenced, to the detriment of consumers, prices or other parameters of competition such as output, innovation, variety or quality of goods or services”. I take issue particularly with the latter part of that sentence.
In any event, the draft Guidelines did not have to go to such a regrettable policy change. I mean, I can see that the Commission now dislikes economic parlance such as “consumer welfare”, “access of competitors to supplies or markets” and “profitably increasing prices to the detriment of consumers”, but why did it have to depart from the more fundamental concept of “anti-competitive foreclosure”? Essentially, that principle means “we protect competition in order to protect the consumers and not in order to protect the competitors as such”. The draft Guidelines still contain a timid reference to some of the Court of Justice’s case law on this particular point, in paragraph 51, but this somewhat grudging reference in reality pays lip service to the more fundamental guiding principle that paragraphs 5, 19 and 20 of the Guidance Paper expressed, which are now gone.
The most regrettable thing is that the abandonment of this fundamental principle and of its intellectual clarity has not been replaced by anything. But how can we not have a proper, general framework of analysis? That general framework informs and ultimately influences the specific legal tests. It also helps new operational tests to be devised for new practices.
Needless to say, this abandonment is not consistent with the EU case law. Read, for example, paragraphs 21-22 of Post Danmark I, paragraphs 133-134 of Intel, paragraphs 277-278 of Google Android, paragraphs 349 and 351 of Qualcomm (exclusivity), paragraph 73 of Servizio Elettrico Nazionale, paragraph 37 of Unilever Italia, and paragraphs 126-127 of European Superleague. What you see is that the Court of Justice mentions this fundamental principle in the “programmatic” part of the reasoning of these judgments that is, the part with the most basic mantras. This is not a footnote but rather a guiding principle. Surely, some disparate sentences in paragraph 51 of the draft Guidelines are not enough to remedy this fundamental inconsistency with the case law. Indeed, recent case law also uses the term “anti-competitive foreclosure” very prominently (Google Android, paragraphs 299 and 643; Qualcomm (exclusivity), paragraph 414; Lithuanian Railways (GC), paragraph 98; Google Shopping, paragraph 615; Intel renvoi, paragraphs 287, 335, 457, 481, 525; etc.).
Such a pity!
2. Second Deadly Sin: From the new economic approach to the new formalistic approach
One of the accomplishments of the Guidance Paper had been that it rose above legalistic formalism and focused on a more economic approach, where the form and external characteristics of conduct were irrelevant. What counted was how certain practices function, what their likely anti-competitive effects are, in short what’s the “theory of harm” behind our intervention. By the way, there is not a single reference to the concept of “theory of harm” in these draft Guidelines. A concept that you find in most EU policy papers and guidelines, a sort of alpha and omega for competition law students, lawyers, judges and enforcers. Why was this abandoned?
What we have now is a retrogression and return to the form. The draft Guidelines, jealous of the annulled first Intel judgment which had also attempted a sort of categorisation opus magnum that failed miserably, proceeds to a categorisation of practices in three groups: a first group relates to those few practices that amount to “[c]onduct for which it is necessary to demonstrate a capability to produce exclusionary effects”, a second group relates to the vast majority of practices that are “presumed to lead to exclusionary effects”, these being exclusive dealing, exclusivity rebates, predatory pricing, margin squeeze with negative spreads and certain forms of tying (i.e. almost everything…), and finally a third group of “naked restrictions”. How does conduct fall into the “purgatory” of the second group (if the third group represents “hell”)? Well, paragraph 60(b), third sentence of the draft Guidelines is not attempting to conceal anything: “Once the factual existence of the relevant conduct is established, if need be[sorry: what does “if need” be mean?] under the conditions established in the specific legal test, its exclusionary effects can be presumed”. In other words, we are back in the boxes of the old formalistic approach: if the form and the external characteristics (“factual existence of the relevant conduct”) are there, we go into the presumption group. That’s why I speak about a return to formalism.
I will come back to the presumptions in my third “accusation”. Let me stay a bit longer on the draft Guidelines’ major categorisation into the three groups. Does such a categorisation reflect the case law? The answer is No! The only categorisation that the ever-developing case law knows of is the one between “by object” and “by effect” exclusionary abuses. This has now become even clearer in the European Superleague case. Yet, the draft Guidelines are silent on this most important distinction. Indeed, when one looks at the third group of practices, one sees that the title the Commission has given them is not “by object” abuses, as they would normally be called, but rather it uses the term “naked restrictions”. Yet this is a very strange name. In reality, only the first of the three examples in paragraph 60(c) are what we would call as “naked restrictions” à la Intel. It is strange to consider the Irish Sugar and Lithuanian Railways practices as “naked restrictions”. They are “by object” abuses! This is the proper name that the EU case law uses. By the way, let me stop here to say that I have no problem at all with the three examples of practices that paragraph 60(c) of the draft Guidelines uses. I applaud the Commission for being restrained and for making this a very exceptional category – as it should be. In fact, I am surprised the Commission did not go even further to include the AstraZeneca type of conduct. These practices could, indeed, be considered as “by object” abusive practices (as would be Lelos, AB InBev and Mondelez – all being not exclusionary but single market-related abuses).
So why aren’t the draft Guidelines follow the case law and why is there not a single reference to “by object”? Surely, one would expect the Commission to be happy with the “by object” box. The answer is simple. The draft Guidelines use the inappropriate term “naked restrictions”, because if they had used the “by object” term in paragraph 60(c), this would have meant that everything else is “by effect” – which is the truth of course! And this would have weakened the draft Guidelines’ narrative around “presumptions”. So this explains the “mystery” behind the disappearance of the “by object” language in the draft Guidelines….
3. Third Deadly Sin: The introduction of presumptions
And now I am coming to the famous presumptions. First of all, let me clarify that the very existence of “light” presumptions that somewhat ease the evidentiary burden for competition authorities in the presence of certain types of behaviour is not necessarily at variance with an effects-based approach. We can certainly live with a structured rule of reason. However, the categorisation must make sense in terms of the theory of harm and must not be based on arbitrary form-based distinctions. The problem with the draft Guidelines, as I explained above, is that the categories selected are arbitrary and form-based. For example, what makes self-preferencing so different from tying? Tying is a form of self-preferencing. And what makes rebates that are conditional on exclusivity different from retroactive rebates with respect to a fixed threshold when that threshold is equal or close to the total requirements of a customer? These are all very similar practices and cannot be subject to different tests just because of their external characteristics and form being different.
But I think the draft Guidelines use the word “presumption” in a “harder” way. And, therefore, we have a problem. Contrary to what the draft Guidelines say in footnote 133 (with no references to case law), it is simply not true that the existing case law favours a “hard” presumption approach for the first stage of the analysis under Article 102 TFEU, i.e. for whether certain conduct is likely to foreclose (the second stage being the objective justification/efficiency defence). Indeed, when the Court of Justice wants to refer to “hard” presumptions, in the sense of allocating or reversing the burden of proof, it uses the words “it is for the dominant undertaking” (“il appartient à l’entreprise occupant une position dominante” in French) and this refers solely to the evidentiary burden to prove efficiencies (see e.g. Post Danmark I, paragraph 42; UK Generics, paragraph 166; Google Android, paragraph 602). So, essentially, the draft Guidelines here “legislate” by introducing something new, which is not reflected in the case law.
True, there has been some confusion as a result of the Court of Justice’s 2017 ruling in Intel. In paragraph 138 of that ruling, the Court used the somewhat sibylline words: “However, that case-law 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”. As I have argued elsewhere, we should not be making too much out of these words. They were simply used by the Court to “develop” the case law (and not just to “clarify” it), in effect revisiting the older more formalistic case law. That’s all. Besides, the Court of Justice itself did not use words such as “presumption” or “presume” in its ruling and has avoided using such words thereafter. Only the General Court used the words “mere presumption” (“presomption simple” in French – which looks to me like a “light” presumption) in Intel renvoi (paragraph 124). Even if one could interpret paragraph 137 of Intel as introducing a presumption of foreclosure/illegality, which can be rebutted according to paragraphs 138 to 140, this would certainly be a very weak presumption, whose practical value for public enforcement is minimal. Of course, no reasonable person suggests that it would be enough for a dominant company to simply say “I contest this” for the presumption to be rebutted and for the Commission’s obligation to prove its case to kick in. In practice, all dominant companies defending themselves before an antitrust authority and – certainly – before the Commission will contest the fact that their practices are capable of restricting competition and will produce economic evidence to that effect. This economic evidence is likely to be quite sophisticated and, in a rebates case, it will most likely include an AEC test. So, practically speaking, a “presumption” is simply an academic point.
However, I repeat that the Commission’s introduction of the “presumptions” approach is at variance with existing case law (with the possible exception of pricing below average variable cost in predation cases) and will not result in any practical consequences. I have carefully gone through the footnotes with the references to case law that supposedly support the position about the existence of a presumption for each of the specific practices, and all of these references simply misread the case law.
Anyway, I do not think the Commission will be stupid enough to adopt infringement decisions that rely on “presumptions” and do not do a good job in rebutting the dominant companies’ evidence. If the draft Guidelines believe that there will be a different standard of review employed by the EU Courts depending on whether the Commission proves ab initio or rebuts something, they are in for a big disappointment…
4. Fourth Deadly Sin: The confusion around “competition on the merits”
To be fair to the Commission, the references to “competition on the merits” are not of its own making. These are references that go back to the early case law which initially referred to “normal competition” and ultimately settled on “competition on the merits”. This was the result of early Ordoliberal influences with a focus on economic efficiency and “performance competition” (Leistungswettbewerb). In the end, the “competition on the merits” reference became a mantra that was always mentioned in the case law. More recently, the Court of Justice, after one or two tribulations in cases such as Servizio Elettrico Nazionale, Unilever Italia and European Superleague, has settled on the idea that in order to have an abuse, you need to have two cumulative conditions: (i) conduct against competition on the merits and (ii) likelihood of anti-competitive effects. But what do these two elements mean and how they relate to each other? Personally, I would have preferred if the Court of Justice, in Servizio Elettrico Nazionale, had sided with Advocate General Rantos in that case, who opined that proving that a dominant company has engaged in conduct that is not consistent with “competition on the merits” is not different from proving anti-competitive effects, or at least the two elements should be examined together (paras 48-49 of his Opinion). The Court, however, did not follow him and found that (a) departure from competition on the merits and (b) exclusionary effects are two separate conditions that must both be established (paragraph 103 of the judgment).
But what is “competition on the merits”? I was hoping that the new Guidelines would shed light on this and would seize an opportunity to present certain guiding principles and also give some intellectual guidance to the Court of Justice. The draft Guidelines unfortunately are a disappointment on that point. In a very reactive way, Section 3.2.2, with the title “Relevant factors to establish that conduct departs from competition on the merits” is utterly underwhelming and confusing.
First, it shows no intellectual leadership to ask a junior lawyer to do a word search and simply mention all the instances where the Court has referred to conduct against “competition on the merits”. Look at paragraph 55 and you will see what I mean. Essentially, the Commission has taken all these references out of context and attempts to make a rule out of a casuistic list. No, we cannot take these references as a clarification of what is not “competition on the merits”. Especially because they can be read out of context and lead to huge Type I errors. We need a more principled approach and we need intellectual leadership from the Commission!
Second, this is all very confusing. Paragraph 53 says that “conduct fulfilling the requirements of a specific legal test is deemed as falling outside the scope of competition on the merits”. What about the second cumulative condition though, likelihood of exclusionary or anti-competitive effects? I thought the specific legal tests were specifying the two cumulative conditions for specific practices. Indeed, paragraph 47 correctly says: “Those specific legal tests are an expression of the application of the general principles discussed in this section to the specific conduct in question. Therefore, when a given conduct meets the conditions set out in a specific legal test, such conduct is deemed to be liable to be abusive because it falls outside the scope of competition on the merits and is capable of having exclusionary effects.” So how is paragraph 53 in line with paragraph 47?
Paragraph 56 makes things even worse: “In the case of certain pricing practices, namely predatory pricing (section 4.2.4) and margin squeeze (section 4.2.5), a price-cost test is required to establish whether conduct of a dominant undertaking departs from competition on the merits. Whenever a price-cost test is carried out to establish whether conduct departs from competition on the merits, the outcome of the test can also be relevant for the assessment of the capability of such conduct to produce exclusionary effects.” I rest my case.
5. Trying to find explanations – not justifications for these “Deadly Sins”
Why the draft Guidelines resorted to go to that direction? I think there are two reasons:
First, there is an “opportunistic” reason, mostly fuelled by the Legal Service and those lawyers internally at DG COMP that never liked the Guidance Paper because they thought that it was making their lives difficult. Less successful infringement cases meant less successful careers I guess (although it should not be like that). These views were given more credibility when the EU Courts annulled a number of Commission decisions, relying (in part) on ideas that were introduced by the Guidance Paper. However, in reality, the Commission has not fared that badly in Luxembourg. Some of the annulments were the product of luck. For example, Intel was simply the case where the Court of Justice decided to revisit its case law. This could have happened earlier, in Tomra for example. Qualcomm (exclusivity) is a different type of case – it was annulled because of grave procedural misconduct, primarily. So that only leaves us with the partial annulment in Google Android and the full annulment in Bulgarian Energy Holding. Since I was personally involved in both cases, I am not going to say anything, other than that surely, these two cases cannot amount to a systemic failure of the Commission before the EU Courts… The Commission has had countless victories in Article 102 TFEU. So the first reason for the new formalistic approach may be a good explanation but certainly cannot count as a justification.
Second, there is a more political reason that has caused the Commission to be defensive and depart from economics. Over the last few years, EU competition enforcement (and Commissioner Vestager even!) has been the target of attacks coming from the other side of the Atlantic, in particular from the neo-Brandesians, who are now in charge of US antitrust policy and enforcement. The criticism (echoed also by some European disciples) is that the Commission lacks ambition, that it is worshipping idols such as consumer welfare, which itself was part of a neo-liberal conspiracy in the EU, and so on. These critics do not understand anything about how EU competition law has moved over the years. As I have stressed elsewhere, EU competition law has never seen consumer welfare as an operational and limiting principle. Instead, the protection of an “effective competitive structure” has always been, and remains, a valid objective and is seen as a good proxy of the broader consumer welfare principle, which is the “ultimate objective” of the competition rules (as the Court of Justice has said) but does not amount to an operational‒technical test that can be applied to specific cases. At the same time, the EU case law makes it clear that consumer welfare is not just about prices but also pertains to quality, innovation and choice. Still, this line of criticism has perturbed the Commission and, in my view, explains some of the “deadly sins” of the draft Guidelines I recounted above, in particular the abandonment of the guiding principle of “anti-competitive foreclosure” (at least as a matter of semantics) and the passage from an economic to a formalistic approach.
So what’s next? In an ideal world, the Commission should think again about the harm that it’s going to inflict on sound Article 102 TFEU enforcement policy. Some of that harm, it will inflict also on itself, since I doubt the EU Courts will be clement. But if the consultation does not lead to any major changes in these draft Guidelines, the ball will be in a different court: the only and ultimate decision-maker remains the Court of Justice. I believe the real aim of the new Guidelines and their approach is to influence in the longer term the Court’s case law, as indeed the Guidance Paper did so – to the opposite direction. However, I have strong reservations that this attempt will be successful. The 2008 Guidance Paper brought a fresh approach, based on sound economics and was a text of intellectual leadership. It was these elements that made it so influential with the Court, in my view. I do not think these Guidelines will tick that box. There is no reason why the lawyers in Luxembourg will give so much weight to the legal stuff that the lawyers in Brussels have put together, to put it crudely…
Assimakis Komninos
(strictly personal views)
Citation: Assimakis Komninos, “J’accuse!” – Four Deadly Sins of the Commission’s Draft Guidelines on Exclusionary Abuses, Network Law Review, Summer 2024.
