The one about bananas and credit cards: exchanges of information as restrictions by object
We are back. This time, with a post that is not voluntarily humorous (nevermind the picture above). A good opportunity to prove ourselves, dear readers, that there is (some) room for (at least some) serious discussions in Chillin’ Competition. I have been meaning to write about the ECJ ruling in Bananas (hence the picture above). I am now out of excuses to avoid doing so given that I cannot (and really should not) keep myself busy forever discussing Alfonso’s posts.
The appeal judgment in Bananas is interesting in that it addresses the status of exchanges of information under Article 101(1) TFEU. Because the nature and the context of such exchanges may vary widely from one case to another, it is probably the practice that exemplifies better than any other the approach taken by the ECJ when drawing the line between restrictions by object and by effect. You know from previous posts that I find the relevant case law to be very sensible. More importantly, it is clear to me the Court has been consistent over time in its approach to the question.
The key arguments raised by Dole in its appeal are familiar ones. An exchange of information may not have an actual impact on future prices. For instance, as argued by Dole in the case, the employees discussing prices may lack the power to set them. It may also be the case that the nature of the exchange does not completely eliminate uncertainty about the behaviour of competitors. The Court clearly dismissed these arguments and held that the exchange at stake in the case amounted to a restriction of competition by object. What lessons can one draw from this case? A couple of them. Nothing, I am afraid, that you have not seen in one way or the other in the blog.
Formal categories are not particularly useful
Are exchanges of information restrictive of competition by object? Well, it depends. Sometimes they are, as in Bananas. Sometimes they fall outside the scope of Article 101(1) TFEU altogether, as in Asnef-Equifax. It all depends on the nature and purpose of the exchange in the context in which it is implemented. The formal category is as such not particularly useful. This has always been the problem with the so-called ‘object box’ and with similar attempts to capture the essence of the case law. Labels (‘exchange of information’, ‘price fixing’, ‘market sharing’) mean virtually nothing in themselves and can even be misleading.
Sometimes, an exchange of information is an ingredient in a cartel-like arrangement. In such a case, the exchange lacks pro-competitive virtues and is therefore deemed restrictive by its very nature. Sometimes, it can improve the functioning of markets and as such potentially beneficial for consumers and the economy. It is clear from the case law that the analysis of the nature and purpose of the agreement needs to be carried out on a case-by-case basis. This sort analysis is not about actual or likely effects (more on this below) but about the rationale behind the agreement.
Bananas provides an interesting example of the above. I welcome your views on the ruling, but it seems clear to me that the Court essentially considers whether the relevant exchange pursued a ‘legitimate objective’ (to take the expression found in Cartes Bancaires, which, unsurprisingly, is abundantly cited in the judgment). In this case, it was found that the purpose of the exchange was simply to remove uncertainty about the behaviour of rivals.
The ‘by object’ category is not a presumption of anticompetitive effects
It has become popular to present the ‘by object’ category as a presumption of anticompetitive effects. The likelihood of such effects is so high in the case of some restraints, the argument goes, that it makes sense to prohibit them irrespective of their actual impact on competition. This view is not unreasonable. The problem is that it is contradicted by the case law.
Bananas shows, again, that an agreement can be deemed to restrict competition by object even if it is not clear that it would have resulted in higher prices. In paragraph 123, the Court holds that ‘a concerted practice may have an anticompetitive object even though there is no direct connection between that practice and consumer prices’. Along the same lines, it is stated in paragraph 125 that ‘there does not need to be a direct link between [a concerted practice] and consumer prices’.
In light of the above, it seems clear to me that determining whether an agreement is restrictive by object has little to do with setting a presumption about its likely impact on competition. As explained above, this assessment is about something else. This also means that there is no real overlap between establishing the anticompetitive nature of the agreement (i.e. understanding its object) and showing its restrictive effects. The first test is about determining whether the agreement is plausibly pro-competitive; the second is about showing, in concreto, its negative effects. This is another major source of misunderstandings. I hope the clear statements of the Court in Bananas will prove helpful in this regard.
More on State aid and investments.
Remember my post on State aid and investment arbitration? The Commission has ordered Romania to recover the compensation to the investors following the award of an arbitration tribunal. More to follow, I am sure!
Chillin'Competition 
Interesting analysis Pablo, thank you!
Substantive issues aside, is ‘Bananas’ the best name for a competition case name so far?
Rodion Raskalnikov
13 April 2015 at 9:50 am
Thanks, Rodion!
Your question is a good starting point for a post on the subject!
Pablo Ibanez Colomo
14 April 2015 at 10:14 am
“The first test is about determining whether the agreement is plausibly pro-competitive” ; the main issue I have with your theory is that the Court actually never said (or even hinted) applying this test. Worse, the Court keeps on mentioning case after case the very categories you say aren’t helpful or meaningful (see “fixation horizontale des prix” in Cartes Bancaires). Finally, you say that object has nothing to do with likely effects, but how else can one interpret the fact that the “critère juridique essentiel” (§57) is that the practice has a “degré suffisant de nocivité à l’égard de la concurrence” (§49), which in turn is interpreted by the Court itself as referring to a practice “tellement suceptible d’avoir des EFFETS négatifs” (§51)?
Adrien Giraud
15 April 2015 at 1:33 pm
Hi Adrien,
Thanks for your comments. As I say, I welcome the discussion.
– The Court has applied the test I describe above systematically, not just once or twice. Consider what it did in Delimitis, Nungesser, Pronuptia or Asnef-Equifax, to name a few where the analytical framework can be clearly identified. In all these rulings, the Court starts by trying to make sense of the rationale behind the agreement. Where it concludes that the agreement is plausibly pro-competitive, it holds that it is not restrictive by object. If anything, the Court has become even more explicit in recent years about the question. Just consider Pierre Fabre. According to the Court, the question of whether a restraint restricts competition by object depends on whether it is objectively justified (paras 39-40). In Cartes Bancaires itself, the determinant factor was the fact that the contentious restraints were deemed to pursue a legitimate objective (para 75).
– In Cartes Bancaires, the Court does not refer to horizontal price-fixing as you seem to say, but to horizontal price-fixing by cartels, which is very different. When I wrote on the blog about the ruling, I already noted that the Court had been very careful in this regard. In para 51, the Court describes the likely effects of price-fixing in a concrete context, which is precisely one where the agreement has no pro-competitive virtues (its only credible aim is to restrict competition).
– There is no contradiction between this paragraph and what I say in the post. Where an agreement has no pro-competitive virtues, it can only have, if at all, anticompetitive effects. The determinant factor is in any event the absence of pro-competitive virtues, not the anticompetitive effects that are likely to derive therefrom. You have to live with Bananas and other similar cases like T-Mobile, which are quite clear onthis point: an agreement can be restrictive by object even in the absence of effects. If that is the case, then the test is necessarily about something else.
Pablo Ibanez Colomo
15 April 2015 at 2:16 pm
Pablo, thanks for the very swift reponse. Here’s a quick reaction:
– I am not saying the Court did not apply the test you centered your theory upon; I am just saying it never said it applied that test; to my knowledge the Court never actually said its review regarding anticompetitive object boiled down to looking whether the agreement or practice is plausibly procompetitive. For example, the Court in Cartes Bancaires describes its method in great details but never actualy says it is looking for a plausibly procompetitive justification. Isn’t this troubling you? Once again, I am not saying that your theory is incorrect; it may very well be the only way of making sense of decades of case law on the subject. I am just doubting (i) that the Court is actually (at least consciously!) applying it and therefore (ii) that this thoery enables one to predict future outcomes.
– In Carte Bancaires, the Court relies on “fixation horizontale des prix par des cartels”; so your suggestion is that because “par des cartels” is added, this means the Court is not using the boxes or categories you criticize? Despite this addition, this sentence looks to me like a categorization. And this feeling is reinforced by the fact that the Court refers to “certains TYPES de coordination” (§49) or “certaines FORMES de coordination” (§50) which also suggests the idea of categorization.
– I guess that an agreement/practice could be competitively neutral (i.e. neither anti nor pro competitive) but this is besides the point. I totally agree with you that an agreement can be restrictive by object in the absence of effects. But I don’t think an agreement can be anticompetitive by object if it is proven (admittedly a hard one to prove) that it CANNOT have anticompetitive effects. In such a case, the agreement cannot be considered as being “tellement suceptible d’avoir des effets négatifs” and therefore cannot have a “degré suffisant de nocivité à l’égard de la concurrence”. I therefore think that the notion of object is totally uncorrelated to that of actual anticompetitive effects, but not to that of possible/conceivable anticompetitive effects. Does that make sense?
Adrien Giraud
17 April 2015 at 8:33 am
Hi Adrien. Again, thanks a lot for sharing your thoughts.
– On your first question: it does not trouble me at all. I will say more: it is exactly what an academic is supposed to do. Courts are not always very explicit about what they do. There is sometimes a difference between what they say they do and what they actually do. The role of a legal academic (in this and other areas) is precisely to identify patterns in the case law and make explicit what is implicit therein. I have read carefully the case law and I have come to the conclusion that this is the analytical framework that best explains what the Court has done over the years. Cartes Bancaires further confirms this view. Nobody can anticipate how cases will be decided in the future, but the pattern in the case law is very clear.
– Cartels are restrictive by object. No doubt about it. If that is what you call categorisation, then we do not disagree. The question, as I see it, is that of how you define what a cartel is. Cartels can take many different forms, depending, inter alia, on the characteristics of the relevant market. What distinguishes a cartel from other forms of horizontal cooperation is precisely the lack of a pro-competitive explanation for the agreement. That is really the defining feature of a cartel, and not so much the forms cartel conduct might take in practice (e.g. price-fixing or market sharing).
– Again, I do not believe there is a fundamental disagreement between us. When an agreement does not have pro-competitive virtues, it must be the case that the parties to it want to harm competition. That is the only plausible explanation for the agreement (its object is deemed anticompetitive). It follows that the agreement must at least be capable of restricting competition (or, as you say, it could potentially or conceivably do so). The point I want to make is that the capability of an agreement to have restrictive effects is not the defining feature of by-object restrictions.
Pablo Ibanez Colomo
17 April 2015 at 2:49 pm
[…] category does not encapsulate a presumption of anticompetitive effects. When commenting on Bananas, I mentioned that an agreement such as an exchange of information can be found to restrict […]
What the Court said, and did not say, in Maxima Latvija | Chillin'Competition
10 December 2015 at 4:20 pm
[…] The one about bananas and credit cards: exchanges of information as restrictions by objectIt seems clear to me that determining whether an agreement is restrictive by object has little to do with setting a presumption about its likely impact on competition.Pablo Ibanez Colomo (Chillin’ Competition) […]
Blog o' Blogs April 2015 - Competition Policy International
5 September 2019 at 6:28 pm