OPINION OF ADVOCATE GENERAL
BOBEK
delivered on 5 September 2019(1)
Case C‑228/18
Gazdasági Versenyhivatal
v
Budapest Bank Nyrt.,
ING Bank NV Magyarországi Fióktelepe,
OTP Bank Nyrt.,
Kereskedelmi és Hitelbank Zrt.,
Magyar Külkereskedelmi Bank Zrt.,
Erste Bank Hungary Zrt.,
Visa Europe Ltd,
MasterCard Europe SA
(Request for a preliminary ruling from the Kúria (Supreme Court, Hungary))
(Competition — Article 101(1) TFEU — Agreements, decisions and concerted practices — Restriction of competition ‘by object’ or ‘by effect’ — Payment cards system in Hungary — Agreement on interchange fee — Participation)
I. Introduction
1. From the early days of EU competition law, much ink has been spilled on the dichotomy between restriction of competition by object and restriction by effect. (2) It may thus come as a surprise that this distinction, stemming from the very wording of the prohibition in (what is now) Article 101 TFEU, still requires interpretation by the Court.
2. The distinction is relatively easy to make in theory. Its practical operation is nonetheless somewhat more complex. It is also fair to say that the case-law of the EU Courts has not always been crystal clear on the subject. Indeed, a number of decisions given by the EU Courts have been criticised in legal scholarship for blurring the distinction between the two concepts. (3)
3. By the present case, the Kúria (Supreme Court, Hungary) invites the Court to shed further light on a dichotomy which is at the heart of Article 101 TFEU, thus allowing the Court to further develop its most recent case-law on the matter, notably CB (4) and Maxima Latvija. (5)
II. Legal framework
A. National law
4. According to Article 11(1) of the tisztességtelen piaci magatartás és a versenykorlátozás tilalmáról szóló 1996. évi LVII. törvény (Law No LVII of 1996 on the prohibition of unfair and restrictive market practices, ‘Law on unfair market practices’):
‘Agreements or concerted practices between undertakings and decisions (hereinafter collectively: agreements) by organisations of undertakings established pursuant to the freedom of association, public corporations, associations or other similar organisations of undertakings (hereinafter collectively: associations of undertakings), which have as their object or potential or actual effect the prevention, restriction or distortion of competition, shall be prohibited. Agreements concluded between undertakings not independent from each other shall not qualify as such agreements.’
III. Facts, procedure and the questions referred
5. When transactions with credit cards such as those at issue in the main proceedings take place, four main parties are typically involved: the cardholder; the financial institution that issued the credit card (‘the issuing bank’); the merchant; and the financial institution providing that merchant with services enabling him to accept the card as a means of settling the transaction concerned (‘the acquiring bank’).
6. According to the referring court, in the early 1990s, the credit card payment system in Hungary was still at an embryonic stage. During the mid-1990s, the companies Visa Europe Ltd (‘Visa’) and MasterCard Europe SA (‘MasterCard’) (together, ‘the credit card companies’) provided in their internal rules for the possibility for the amount of the interchange fee (‘IF’) to be determined jointly by the acquiring bank and the issuing bank. The IF is the amount paid by the former to the latter when a credit card transaction takes place.
7. Between 1991 and 1994, when only a few banks participated in the credit card schemes in Hungary, the banks agreed the amount of the IF bilaterally. However, in 1994, Visa invited the banks participating in its scheme in Hungary to set up a national forum with a view to, inter alia, agreeing on a local price policy for the IF. Between 1995 and 1996, the banks operating in the card sector introduced a multilateral cooperation procedure (‘the Forum’), in which they specifically discussed issues concerning the card business on which cooperation was needed.
8. Within the Forum, on 24 April 1996, seven banks — most of which were members of both companies’ credit card systems — reached agreement on the minimum level of the merchant service charge (‘MSC’) (‘the MSC Agreement’). The MSC is the fee that the acquiring bank charges to merchants that accept payments with credit cards. However, in the end the MSC Agreement never came into effect.
9. On 28 August 1996, the same group of banks adopted an agreement which introduced a uniform multilateral IF (‘MIF’) applicable to both credit card companies (‘the MIF Agreement’) with effect from 1 October 1996. The credit card companies were not present at the meeting where the agreement was concluded, but a copy of the agreement was sent to them by Kereskedelmi és Hitelbank Zrt., which acted as a point of contact. Subsequently, other banks joined the MIF Agreement and the Forum mechanism.
10. On 31 January 2008, the Gazdasági Versenyhivatal (Hungarian Competition Authority, ‘HCA’) began an investigation into the MIF Agreement. The agreement remained in force until 30 July 2008.
11. In its decision of 24 September 2009 (‘the contested decision’), the HCA concluded that, in setting a MIF and devising a uniform structure for it, and in establishing and promoting a regulatory framework for the MIF Agreement, the 22 member banks and the credit card companies had entered into an anticompetitive agreement that was in breach of Article 11(1) of the Law on unfair market practices and, from 1 May 2004, of Article 81(1) EC (now Article 101(1) TFEU). The HCA imposed fines of varying amounts on the seven banks that initially concluded the MIF Agreement and on the two credit card companies, which came to a total of 1 922 000 000 Hungarian forint (HUF).
12. In its decision, the HCA concluded that the MIF Agreement constituted a restriction of competition by object. In addition, the HCA held that the agreement also constituted a restriction of competition by effect.
13. The credit card companies and six of the banks that had been fined (‘the applicants in the main proceedings’) challenged the contested decision before the Fővárosi Közigazgatási és Munkaügyi Bíróság (Budapest Administrative and Labour Court, Hungary). That court dismissed the action.
14. The applicants in the main proceedings, with the exception of MasterCard, brought an appeal against that judgment before the Fővárosi Törvényszék (Budapest High Court, Hungary) which annulled the contested decision in part and ordered the HCA to conduct a new investigation procedure. That court found that it was not possible for conduct to constitute a restriction of competition by object and, at the same time, a restriction of competition by effect. It also held that the agreement in question did not constitute a restriction of competition by object.
15. The HCA lodged a further appeal against that judgment with the Kúria (Supreme Court, Hungary). That court, entertaining doubts as to the proper interpretation of Article 101 TFEU, decided to stay the proceedings and to refer the following questions to the Court for a preliminary ruling:
‘(1) Can [Article 101(1) TFEU] be interpreted as meaning that the same conduct can infringe this provision both because the object of the conduct is anticompetitive and also because its effect is anticompetitive, with the two cases being treated as separate grounds in law?
(2) Can [Article 101(1) TFEU] be interpreted as meaning that the agreement at issue in the proceedings, which was entered into by Hungarian banks and which establishes, in respect of the two bank card companies, MasterCard and Visa, a unitary amount for the interchange fee payable to the issuing banks for the use of the cards of those two companies, constitutes a restriction of competition by object?
(3) Can [Article 101(1) TFEU] be interpreted as meaning that the credit card companies can also be considered to be parties to an interbank agreement where they were not directly involved in defining the content of the agreement but facilitated its adoption and accepted and implemented it; or are these companies to be considered to have acted in concert with the banks that entered into the agreement?
(4) Can [Article 101(1) TFEU] be interpreted as meaning that, in view of the subject matter of the proceedings, for the purpose of finding an infringement of competition law, it is not necessary to differentiate between participation in the agreement and acting in concert with the banks that participated in the agreement?’
16. Written observations were submitted by eight of the applicants in the main proceedings (Budapest Bank Nyrt., ING Bank NV Magyarországi Fióktelepe, OTP Bank Nyrt., Kereskedelmi és Hitelbank Zrt., Magyar Külkereskedelmi Bank Zrt., Erste Bank Hungary Zrt., Visa and MasterCard), the HCA, the Hungarian Government, the EFTA Surveillance Authority (‘ESA’), and the European Commission. Those parties, with the exception of Kereskedelmi és Hitelbank Zrt., also presented oral argument at the hearing on 27 June 2019.
IV. Analysis
17. In my view, the answers to the referring court’s first, third and fourth questions are relatively straightforward. I will start by providing a succinct answer to question one, recalling that the same conduct can indeed be classified as restricting competition both by object and by effect, provided that the evidence adduced so permits (A). I then turn to what I consider to be the crux of the present case: under what conditions can an agreement, such as the MIF Agreement, be deemed to amount to a restriction by object? (B) Finally, I shall deal with questions three and four, which are interconnected, by first addressing the extent of the obligation on the competition authority to state whether the conduct in question constitutes an agreement or concerted practice (C), and concluding with the issue of participation by the credit card companies in an agreement or a concerted practice in the context of the present case (D).
A. First question
18. By its first question, the referring court asks the Court whether the same conduct of an undertaking can restrict competition under Article 101(1) TFEU both by object and by effect.
19. In my view, the answer to that question is clearly affirmative. That answer flows not only from the logic and the context of that provision, but also from by now well-established case-law of the Court.
20. The referring court is uncertain of the precise meaning of the expression ‘object or effect’. In particular, that court wonders whether the alternative character of those requirements means that a given agreement cannot be considered to restrict competition by object and by effect at the same time. In the contested decision, the HCA in fact held that the MIF Agreement constituted a restriction of competition both by object and by effect.
21. As a matter of (formal) logic, ‘or’ is normally understood as a(n) (inclusive) disjunction. A statement containing two propositions connected by an ‘or’ will be true if and only if either one or both of its component propositions are true. Thus, there can be a restriction of competition by object only, or by effect only, or by object and effect.
22. Whether logic is (or should be) a general tool for the interpretation of EU law can certainly be the subject of passionate debate, but in this particular case, understanding the ‘or’ in ‘object or effect’ as an inclusive disjunction is fully in line with the aim and purpose of Article 101(1) TFEU.
23. Article 101(1) TFEU is formulated in very broad terms. It aims at catching all forms of collusion between undertakings (‘all agreements between undertakings, decisions by associations of undertakings and concerted practices’), regardless of the aim pursued and the subject matter thereof (‘which have as their object or effect’), that may have a negative impact on competition in the European Union (‘the prevention, restriction or distortion of competition within the internal market’). Thus, any and all of the types of conduct listed are covered, regardless of whether they occur separately or, a fortiori, simultaneously.
24. In addition, agreements that are anticompetitive by object and agreements that are anticompetitive by effect are not ontologically different. From a substantive point of view, there is no difference between them: they both restrict competition in the internal market and, for that reason, are in principle prohibited. The distinction between the two concepts is based rather on considerations of a procedural nature. It is meant to indicate the type of analysis that competition authorities are required to carry out when assessing agreements in the light of Article 101(1) TFEU.
25. This understanding was emphasised by the Court as far back as in 1966 in LTM, where it pointed out that the use of the disjunctive conjunction ‘or’ in the then Article 85(1) EEC means that a competition authority should first consider the object of an agreement. Where, however, an examination of the object of the agreement ‘does not reveal the effect on competition to be sufficiently deleterious, the consequences of the agreement should then be considered’. (6)
26. That aspect also emerges clearly from more recent judgments of the Court. In CB, the Court explained that certain forms of coordinated behaviour can be regarded, by their very nature, as being harmful to the proper functioning of normal competition. In those cases, it would be ‘redundant, for the purposes of applying [Article 101(1) TFEU], to prove that they have actual effects on the market’. Indeed, ‘experience shows that such behaviour leads to falls in production and price increases, resulting in poor allocation of resources to the detriment, in particular, of consumers’. (7) Similar statements were also made in Maxima Latvija. (8)
27. Therefore, the object/effect dichotomy is, by and large, a procedural device meant to guide the competition authority on the analysis to be carried out under Article 101(1) TFEU depending on the circumstances of the case. (9) An authority is not required to carry out a fully fledged analysis of the effects of an agreement — which is often lengthier and more resource-intensive (10) — when it maintains and establishes that the agreement is anticompetitive by object.
28. However, in so far as the two types of agreement are not intrinsically different, the authority may well decide, in a specific case, to examine an agreement from the two angles at the same time, in one decision, and thus check whether both requirements are fulfilled. That practice may be justified, as argued by the Commission and the ESA, by reasons of procedural efficiency: if the anticompetitive object of an agreement is controversial, it may be ‘safer’ for the authority, in case of subsequent litigation, to demonstrate that the agreement is also anticompetitive by effect. (11) In fact, the Court has expressly accepted that undertakings can take part ‘in collusion which has the object and effect of restricting competition within the internal market within the meaning of Article 101 TFEU’. (12)
29. That said, I wish to stress an important point: accepting, as a conceptual possibility, that an agreement might amount to both types of restriction certainly does not liberate the appropriate competition authority from the requirement to, first, adduce the necessary evidence for both types of restriction and, second, evaluate and clearly subsume that evidence under the appropriate legal categories.
30. I think it is important to underline that aspect rather clearly, not because of the text of the present request for a preliminary ruling, but rather its subtext. It would hardly be sufficient, including for the purpose of subsequent judicial review of a decision, if, in its decision, a competition authority limited itself to assembling factual evidence and, without stating what inferences in terms of legal evaluation it drew from that evidence, merely suggested that certain behaviour might be this and/or that, leaving it for the reviewing court to connect the factual dots and come to a conclusion. Put simply, the existence of alternative legal boxes is no licence for vagueness, in particular when imposing heavy administrative sanctions.
31. Finally, I now turn to three additional arguments put forward by the referring court in its request for a preliminary ruling, which caused that court to express doubts as to whether it is possible for an authority to find conduct in breach of Article 101(1) TFEU on both grounds.
32. First, in so far as there is no conceptual difference between them, it is unsurprising that a competition authority might refer to the same set of facts and economic considerations to find an agreement to be anticompetitive both by object and by effect. With the caveat just outlined that the legal qualification must be clear, the difference in the analysis required of the authority in the two situations is more of degree and depth than of kind. The two types of analysis are simply different ways, in the light of the knowledge and experience acquired by the authority, to answer one and the same question: whether the agreement at issue may prevent, restrict or distort competition in the internal market.
33. Second, there is no automatic correlation between the qualification of an agreement as restrictive by object or by effect and the determination of the penalties that may be imposed upon the undertakings responsible. According to Article 23(3) of Regulation (EC) No 1/2003, ‘in fixing the amount of the fine, regard shall be had both to the gravity and to the duration of the infringement’. (13) Admittedly, agreements that are found to restrict competition by object are more likely to be considered as giving rise to serious infringements of competition law. Nevertheless, that is only the inevitable consequence of the fact that the concept of restrictions ‘by object’ is confined to forms of coordination that ‘reveal a sufficient degree of harm to competition’. (14) More importantly, it is by no means excluded that, on the one hand, certain restrictions by object might be considered, in the light of all relevant circumstances, to amount to infringements of a lesser gravity and, on the other hand, restrictions by effect might be deemed to constitute particularly serious infringements of competition law.
34. Third, the qualification of an agreement as restrictive by object or by effect also has no impact on the possibility of applying an exemption under Article 101(3) TFEU. There is nothing in the wording of that provision to suggest that exemptions can only apply to agreements that restrict competition by effect. Such a position would also be difficult to reconcile with the fact that, as explained in point 24 above, there is no conceptual difference between the two types of agreement.
35. In fact, in Matra Hachette, the General Court stated that all forms of anticompetitive practice caught by Article 101(1) TFEU may, where the relevant conditions are satisfied, be exempted under Article 101(3) TFEU. (15) That finding is not called into question by the judgment of the Court in Beef Industry Development Society and Barry Brothers. (16) Paragraph 21 of the latter judgment cannot be read as indicating that the distinction between restrictions by object and by effect is relevant under Article 101(3) TFEU. In that passage, the Court simply intended to point out that the absence of a subjective intention to restrict competition on the part of the undertakings involved in an agreement does not exclude it from being anticompetitive in nature. That is why the Court made clear that, once it is ascertained that an agreement restricts competition, whether it pursues another (allegedly lawful) objective may only be taken into account, where appropriate, for the purposes of obtaining an exemption under Article 101(3) TFEU.
36. In the light of the above, I propose that the Court answer the first question by stating that the same conduct of an undertaking can be held to infringe Article 101(1) TFEU for having both the object and the effect of restricting competition in the internal market.
B. Second question
37. By its second question, the referring court asks whether Article 101(1) TFEU is to be interpreted as meaning that an agreement such as the MIF Agreement constitutes a restriction of competition by object.
38. I do not think the Court can answer that question in the way it has been posed. In the context of the present proceedings, any substantive assessment is necessarily based on the (relatively limited) amount of information regarding the MIF Agreement and the relevant markets that has been included in the request for a preliminary ruling, or that can be gleaned from the parties’ submissions. Yet, analysing the anticompetitive nature (whether by object or by effect) of an agreement is, in most cases, no easy task. It requires a good understanding of the contractual relationship between the parties to the agreement, and an in-depth knowledge of the market in which the agreement was implemented.
39. In cases that reach the Court through the preliminary ruling procedure, inevitably only the referring court possesses that information and expertise. Therefore, rather than attempting to undertake an indirect review of a (national) administrative decision, which falls outside the jurisdiction of this Court, (17) I cannot but limit myself to suggesting some guidance and criteria as to how the referring court should conduct that review, in the light of the available information.
1. On the concept of restrictions ‘by object’
40. As the Court has emphasised in recent case-law, the concept of a restriction of competition ‘by object’ must be interpreted restrictively, and can be applied only to certain types of coordination between undertakings which reveal a sufficient degree of harm to competition and for which it is thus unnecessary examine their effects. (18) That approach is justified by the fact that certain types of coordination between undertakings can be regarded, by their very nature, as being harmful to the proper functioning of normal competition since they normally produce inefficient economic outcomes and reduce consumer welfare. (19)
41. It follows from the case-law that, in order to find an agreement anticompetitive by object, the competition authority must carry out a two-step analysis.
42. In the first step, the authority focuses mainly on the content of the provisions of the agreement and its objectives. (20) The key aim of this procedural step is to ascertain whether the agreement in question falls within a category of agreements whose harmful nature is, in the light of experience, commonly accepted and easily identifiable. (21) In that regard, experience may be understood to refer to ‘what can traditionally be seen to follow from economic analysis, as confirmed by the competition authorities and supported, if necessary, by case-law’. (22)
43. As a second step, the authority is required to verify that the presumed anticompetitive nature of the agreement, determined on the basis of a merely formal assessment of it, is not called into question by considerations relating to the legal and economic context in which the agreement was implemented. To that end, it is necessary to take into account the nature of the goods or services affected, as well as the real conditions of the functioning and structure of the markets in question. (23) In addition, although the parties’ intention is not a necessary factor in determining whether an agreement between undertakings is restrictive, that factor may be taken into account where relevant. (24)
44. It is the scope and depth of this second procedural step that is, in the eyes of the referring court, somewhat unclear. I understand those doubts: is that analysis not a de facto verification of the effects of the agreement in question? Where does the second step in the by object analysis stop and the by effect analysis begin? In particular, can any such distinction in fact be made in the context of a case where it would appear that the national competition authority conducted both types of analysis within the same decision?
45. First, why is (some) analysis of the legal and economic context required at all when an agreement appears to constitute a restriction by object? The reason is that a purely formal assessment of an agreement, completely detached from reality, could lead to condemning innocuous or procompetitive agreements. There would be no legal or economic justification for prohibiting an agreement that, despite conforming to a category of agreements that is usually considered anticompetitive, is nonetheless, because of some specific circumstances, outright incapable of producing any deleterious effect in the marketplace, or is even procompetitive. (25)
46. That is why the Court’s case-law has always been consistent on this point: the assessment of a practice under EU competition rules cannot be made in the abstract, but requires an examination of that practice in the light of the legal and economic conditions prevailing on the markets concerned. The importance of this principle is confirmed by the fact that it has been found to be valid with regard to both Article 101(1) TFEU (26) and Article 102 TFEU. (27) Not even when dealing with forms of conduct like price fixing, market sharing or export bans, which are generally recognised to be particularly harmful to competition, can the economic and legal context be totally ignored. (28)
47. On a similar note, in Toshiba, the Court stated that, with respect to agreements alleged to be anticompetitive by object, ‘the analysis of the economic and legal context of which the practice forms part may … be limited to what is strictly necessary in order to establish the existence of a restriction of competition by object’. (29) What does that mean in practical terms?
48. In my view, it means that the competition authority applying Article 101(1) TFEU must, in the light of the elements present in the case file, check that there are no specific circumstances that may cast doubt on the presumed harmful nature of the agreement in question. If experience tells us that the agreement under consideration belongs to a category of agreements that, most of the time, is detrimental to competition, a detailed analysis of the impact of that agreement on the markets concerned appears unnecessary. It is sufficient for the authority to verify that the relevant market(s) and the agreement in question do not have any special features which might indicate that the case at hand could constitute an exception to the experience-based rule. As infrequent as that may be, the possibility that an agreement may actually display such features cannot be dismissed unless the real-world context in which the agreement operates is taken into account. For example, if competition on a given market is not possible and none exists, then there is no competition that could be restricted.
49. The second step thus amounts to a basic reality check. It simply requires the competition authority to check, at a rather general level, whether there are any legal or factual circumstances that preclude the agreement or practice in question from restricting competition. There is no standard type of analysis or set level of depth and meticulousness that an authority has to adopt to carry out that verification. The complexity of the analysis required of the authority to find an agreement anticompetitive ‘by object’ depends on all of the relevant circumstances of the case. It is impossible to (or at least I am unable to) draw, in abstract terms, a bright line between (the second step of) an object analysis and an effects analysis.
50. Thus, as already suggested, the difference between the two is more one of degree than of kind. It is, nevertheless, clear that if the elements that the authority observes when looking at the legal and economic context of an agreement alleged to constitute a restriction ‘by object’ point in different directions, an analysis of its effects becomes necessary. In that case — as in any case where an agreement is not deemed to be anticompetitive by object — a fully fledged effects analysis must be carried out for the purposes of Article 101(1) TFEU. The objective of that analysis is to determine the impact that the agreement may have on competition in the relevant market. In essence, the authority has to compare the competitive structure in the market caused by the agreement under scrutiny, with the competitive structure which would have prevailed in its absence. (30) The analysis cannot, therefore, stop at the mere capability of the agreement to negatively affect competition in the relevant market, (31) but must determine whether the net effects of the agreement on the market are positive or negative.
51. Simplified to a metaphorical extreme, if it looks like a fish and it smells like a fish, one can assume that it is fish. Unless, at the first sight, there is something rather odd about this particular fish, such as that it has no fins, it floats in the air, or it smells like a lily, no detailed dissection of that fish is necessary in order to qualify it as such. If, however, there is something out of the ordinary about the fish in question, it may still be classified as a fish, but only after a detailed examination of the creature in question.
2. The MIF Agreement as a restriction by object?
52. The HCA, supported by the Hungarian Government and the Commission, considers the MIF Agreement to be inherently anticompetitive, whereas the applicants in the main proceedings dispute that position.
53. As mentioned in point 9 above, the MIF Agreement essentially introduced a uniform amount for the IF, the fee that the acquiring banks pay to the issuing banks when a credit card transaction takes place. Therefore, such an agreement is neither, as the referring court correctly points out, a typical horizontal price-fixing agreement nor, I would add, anything that could be readily qualified as a vertical resale price maintenance agreement. The MIF Agreement does not set sale and purchase prices for final customers, but merely ‘standardises’ one aspect of the cost structure of some services triggered by the use of credit cards as a means of payment.
54. In the light of this, one would have expected the parties arguing in favour of a ‘by object’ restriction to deal in particular with the following points. As a preliminary issue, the conduct alleged to constitute a restriction of competition by object should have been identified without ambiguity and its key elements explained (the parties responsible, the markets affected, the nature of the conduct in question, and the relevant time frame) (a). Next, the MIF Agreement should have been assessed in the light of that analytical framework: first, putting forward a reliable and robust wealth of experience to show that such conduct is inherently anticompetitive (b) and, second, explaining why the legal and economic context of the agreement does not call into question its presumed anticompetitive nature (c).
55. In what follows, I shall briefly dwell on each of those points, taking into account the elements brought to the attention of the Court in the course of the present proceedings, while again stressing that it is not for this Court to carry out an indirect judicial review of a national administrative decision. It will thus be for the referring court to verify that the HCA has discharged its burden of proof in the contested decision.
(a) Was the alleged infringement clearly identified and explained?
56. The conclusion that a given practice amounts to a restriction by object can only be justified when the alleged manifest infringement is clearly defined. However, the lack of precision in this regard in this case, already apparent in the written submissions, was further amplified at the hearing, during which each of the parties that argued in favour of a ‘by object’ restriction seemed to argue a somewhat different case. In particular, when asked to explain, without ambiguity, the type of competitive harm that is likely to result from agreements such as the MIF Agreement, their arguments ‘jumped’ from one market to another and from one type of harm to another, without the necessary clarity and precision.
57. At least three markets that the alleged infringement may have affected were mentioned: the market for inter-bank services related to credit card-based transactions (which the MIF Agreement directly concerned); the (partially downstream) market for services provided to merchants in relation to credit card-based transactions (which the HCA seemed most concerned about); and the (effectively upstream) market for credit card providers (which the Commission very much focused on). These three markets are clearly intertwined and their interaction cannot be ignored. (32)
58. With regard to the first market, an element of price setting was indeed introduced by the MIF Agreement. However, the harmful effects on competition identified by the HCA and the Commission do not materialise in that market. Rather, the HCA and the Commission point to harmful effects in the other two markets.
59. Starting with the market for services provided to merchants in relation to credit card-based transactions, the HCA and the Commission take the view that the IF operated, in practice, as a recommended minimum price. Indeed, acquiring banks were unlikely to charge merchants an MSC of an amount below the IF paid to the issuing bank, as that would have been uneconomical. From that perspective, a certain degree of restriction of competition is indeed plausible: the incentive for banks to compete for merchants by lowering the MSC may, in practice, be limited by the agreed MIF. Likewise, as far as the market for credit card providers is concerned, and in so far as the MIF Agreement concerned both MasterCard and Visa, it cannot be ruled out that it was capable of neutralising an element of price competition between those companies.
60. However, I am not sure that the interactions between those markets have been adequately explained. A by object analysis may further the competition authority’s task of proving the anticompetitive nature of certain conduct, but it does not relieve that authority of the requirement to identify clearly the nature of the alleged harm. More importantly, the views expressed on the interactions between those markets seem to be based on a number of assumptions, some of which are heavily contested by the applicants in the main proceedings. It is no doubt possible that the MIF Agreement might produce harmful effects, but are those effects so easily identifiable and so likely that that agreement can be considered restrictive ‘by object’?
61. A number of actors of different kinds were active in several intertwined markets, and the interaction and cross-effects between those markets do not seem to be readily apparent. Added to such composite complexity in terms of who, what and where, is the temporal aspect. The MIF Agreement lasted for more than 12 years. I doubt that the conditions on the relevant markets in Hungary remained essentially unchanged over that period. After all, it is fair to assume (and the number of participating banks might be seen as an indirect hint at that evolution) that between 1996 and 2008, the market for credit card services in Hungary as well as elsewhere in Europe changed considerably. Accordingly, what might have been useful or even necessary at a given point in time to pursue a procompetitive objective of effectively establishing a market may have ceased to be such when the conditions of competition in the market substantially changed. Such a hypothesis, if correct, could mean that it is not possible to view the entire period in exactly the same light and with the same amount of clarity for the purpose of assessing whether there was a restriction of competition by object.
62. In sum, with this enhanced complexity in terms of the number of actors active in several markets over a longer period of time comes an enhanced need for definitional clarity and precision, in particular if what is suggested is the presence of a restriction of competition by object: who was supposed to have done precisely what in which market(s) and with what consequences? Moreover, the more variables that are included in the equation in terms of structural complexity, the less likely it is, in general, that it would be possible to conclude in favour of a clear by object restriction.
(b) Is there a reliable and robust wealth of experience regarding agreements such as the one at issue?
63. Next, particularly in view of the complexity of the factual circumstances at issue in the main proceedings, I would have expected the parties arguing in favour of a restriction by object to put forward a reliable and robust wealth of experience showing that agreements such as the MIF Agreement are commonly regarded as being inherently anticompetitive. Is there a relatively widespread and consistent practice of the European competition authorities and/or of the courts of the Member States supporting the view that agreements such as that at issue are generally harmful to competition?
64. When asked at the hearing, the HCA stated that it essentially relied (only) on the Commission’s practice. In turn, the Commission stated that the inherently anticompetitive nature of agreements such as the MIF Agreement stems from the judgments of the EU Courts in MasterCard. (33)
65. I would question whether that amounts to the robust and reliable wealth of experience required to support a finding that a given form of conduct is patently and generally anticompetitive.
66. As far as the Commission’s practice is concerned, I note that, in 2002, the Commission gave exemptions under (now) Article 101(3) TFEU to certain agreements setting multilateral IF. (34) It took the view that those agreements were restrictive by effect (and not by object) but contributed to technical and economic progress as they promoted a large-scale international payment system with positive network externalities. In its decision of 19 December 2007, the Commission decided that certain decisions setting a multilateral ‘fallback IF’ constituted a restriction by effect, without taking a position on whether that agreement was also anticompetitive by object. (35) Subsequently, on 22 January 2019, the Commission took the view that certain cross-border acquiring rules applied by MasterCard, in particular with regard to interregional IF, constituted a restriction by object. (36) That is not to mention the decisions of 2010, 2014 and 2019, in which the Commission accepted commitments offered by the credit card companies to cap or lower the amount of certain types of IF. (37) Of course, commitment decisions do not involve any formal finding of an infringement of competition law.
67. It thus seems to me that the Commission’s practice can hardly be called uniform. That is not a reproach, but rather an acknowledgement of the fact that it would appear that the Commission’s own understanding of this category of agreements has gradually progressed, thanks to the experience acquired in those very cases. After all, it is clear that the concept of ‘experience’ necessarily evolves over time, in the light of the increased knowledge and experience acquired by the competent administrative and judicial authorities, the creation of more sophisticated tools of analysis, and the development of economic thinking.
68. Nevertheless, even leaving aside the much discussed temporal element, (38) I would be cautious about coming to the conclusion that a handful of administrative decisions (especially when issued by a single authority and evolving over time), which concerned similar forms of coordination, are a sufficient basis for holding that any comparable agreement can be presumed to be unlawful.
69. In addition, the practice of other national competition authorities that may have assessed agreements similar to that examined by the HCA would certainly have been of relevance, if available.
70. As regards the case-law, it is equally important to check whether the Courts of the European Union and of the Member States (39) that have dealt with this category of agreements have adopted a consistent approach in this field. (40)
71. In that connection, the validity of the aforementioned Commission decision of 2007 was indeed upheld first by the General Court, and then by the Court of Justice. (41)For what is relevant here, the key question is, however, whether those judgments suggest that the infringement of Article 101(1) TFEU ascertained by the Commission was so manifest that it could have been established without a comprehensive effects analysis. My impression is that, given the lengthy and detailed nature of the arguments developed by the EU Courts to dismiss the applicants’ claims, it is difficult to read those judgments as supporting one view or the other.
72. Finally, I am somewhat surprised that, in the submissions of the parties that argue for a restriction ‘by object’, there is no trace of studies or reports prepared by independent authors and based on methods, principles and standards recognised by the international economic community supporting their view. Indeed, whether there is a sufficient consensus among economists that agreements such as the one at issue are inherently anticompetitive would seem to me to be of the utmost importance. The concept of restriction of competition is, after all, mainly an economic concept.
73. In conclusion, the bank of experience relied on before the Court to support the view that agreements such as the one at issue in the main proceedings are, by their very nature, harmful to competition appears rather meagre. However, it will be for the referring court to verify this point in detail, in view of the arguments and documentation relied on in the administrative decision in question.
(c) Does the legal and economic context of the MIF Agreement call into question its presumed anticompetitive nature?
74. Should the referring court be convinced by the HCA’s analysis that the MIF Agreement does fall in a category of agreements that is generally regarded as being anticompetitive, the second step of its analysis should be to verify the validity of that preliminary finding by turning its attention to the legal and economic context in which the agreement was implemented. Is there any specific feature, of the MIF Agreement or of the affected markets, that may cast doubt on its harmful effect on competition? Thus, at first sight, can the proposition concerning the generally harmful nature of such an agreement reasonably be challenged in the context of the individual case?
75. The applicants in the main proceedings argue that the MIF Agreement did not have any anticompetitive object or, at any rate, that it also had some procompetitive effects.
76. First, those parties give an alternative explanation of the economic rationale of the MIF Agreement: they argue that the standardisation of the IF was necessary to ensure a proper and smooth functioning of the system, given that the credit card system in Hungary was still underdeveloped when the MIF Agreement was entered into. The agreement thus contributed, in their view, to the establishment and expansion of the credit card market in Hungary. Second, they contend that the MIF Agreement also aimed at limiting the market tendency towards an increase in the IF. That fact is supported, they argue, by the judgment of the General Court in Mastercard, (42) as well as by the fact that in many jurisdictions (including Hungary and the European Union (43)) the legislature intervened to cap the level of the IF.
77. I believe that it is not possible, in the context of these proceedings, to take a firm stance on whether or not those arguments are prima facie implausible. The information included in the case file is simply insufficient for that purpose.
78. It is for the referring court to examine those claims in order to check whether they are credible enough to warrant closer scrutiny. Should the referring court come to the conclusion that the MIF Agreement could reasonably have had some procompetitive effects and that those positive effects are not clearly outweighed by other, more profound, anticompetitive effects, that agreement cannot be classified as restrictive of competition by object. In that case, an infringement of Article 101(1) TFEU can only be established following an analysis of the effects of the agreement.
79. Thus, the yardstick should be that of a countervailing hypothesis that is not implausible at first sight and that challenges, in the context of the individual case, the general conventional wisdom. There are two key elements to this: first, the countervailing explanation must seem plausible enough at first sight to warrant further examination. Second, however, the standard is that of a reasonable countervailing hypothesis. It does not need to be fully established, argued, and proven: that is a matter for the fully fledged effects analysis.
80. In that regard, it might be added that the Court has long recognised that agreements that pursue a ‘legitimate objective’ are not necessarily caught by Article 101(1) TFEU. (44) This means that agreements that have both procompetitive and anticompetitive effects are caught by the prohibition of Article 101(1) TFEU only where the latter prevail. (45) For example, a reduction of price competition may be acceptable when it is a means to increase competition in relation to factors other than price. (46) More generally, agreements that, despite being restrictive of the parties’ freedom of action, pursue the objective of, for example, opening up a market or creating a new one, or allowing new competitors to access a market, may be procompetitive. (47) It equally follows from settled case-law that, under certain conditions, restrictions which are directly related and necessary to the implementation of a main operation, which is in itself not anticompetitive, do not constitute restrictions of competition within the meaning of Article 101(1) TFEU. (48)
81. Accordingly, any time an agreement appears to have ambivalent effects on the market, an effects analysis is required. (49) In other words, when a possible procompetitive economic rationale for an agreement cannot be ruled out without looking at the actual effects on the market, that agreement cannot be classified as restrictive ‘by object’. (50) I therefore cannot agree with the Commission when it argues that any legitimate and procompetitive effect of the MIF Agreement could only be considered under Article 101(3) TFEU for the possible granting of an exemption. Without making any statement on the MIF Agreement in particular, in general, a construction of Article 101 TFEU according to which an agreement that is on the whole procompetitive is in principle prohibited by Article 101(1) TFEU, but then can be immediately exempted under Article 101(3) TFEU, is not entirely convincing.
82. In the light of the foregoing, it is for the referring court to verify whether the MIF Agreement constitutes a restriction by object. To that end, the referring court must first examine the content and objective of that agreement to determine whether it falls into a category of agreements that, in the light of experience, is generally recognised as being harmful to competition. If the answer to that question is positive, the referring court should then verify that that finding is not called into question by considerations relating to the legal and economic context in which that specific agreement was implemented. In particular, the referring court should check whether any alternative explanation as to an allegedly procompetitive rationale of the MIF Agreement is prima facie plausible, taking account also of the time when the agreement was in operation.
C. Fourth question
83. By its fourth question, which should be examined next, the referring court seeks to know whether a competition authority, when establishing an infringement of Article 101(1) TFEU, is required to indicate expressly whether the undertakings’ conduct constitutes an agreement or a concerted practice.
84. I agree with the HCA, the Hungarian Government, the Commission and the ESA that this question should be answered in the negative.
85. In Anic Partecipazioni, the Court clarified that the concepts of an ‘agreement’ and a ‘concerted practice’ in Article 101(1) TFEU ‘are intended to catch forms of collusion having the same nature and are only distinguishable from each other by their intensity and the forms in which they manifest themselves’. Therefore, whilst those concepts have partly different constituent elements, ‘they are not mutually incompatible’. The Court explicitly recognised that infringements of Article 101(1) TFEU may often ‘involv[e] different forms of conduct [that] may meet different definitions whilst being caught by the same provision and being all equally prohibited’. On that basis, the Court held that a competition authority is not required to categorise a given form of conduct as an agreement or as a concerted practice. (51) This principle has been consistently confirmed in subsequent case-law. (52)
86. Indeed, in most cases, it would be unreasonable and unnecessary for a competition authority to try to characterise a specific form of behaviour as either an agreement or a concerted practice. The truth is that those concepts overlap to some extent, making it often hard to say where an agreement ends and a concerted practice begins. Moreover, experience demonstrates that infringements can evolve over time. They may start in one form and progressively assume the characteristics of another. (53)
87. That is why the Court has also pointed out that, regardless of the legal characterisation of a form of conduct as a ‘concerted practice’, ‘agreement’ or ‘decision by an association of undertakings’, the legal analysis to be carried out under Article 101(1) TFEU is no different. (54) In the context of the present case, it may be useful to point out that not only an agreement but also a concerted practice may be held to be anticompetitive by object. (55)
88. That obviously does not mean that the competition authority is not required to prove, to the required standard, that the conduct alleged to constitute an anticompetitive ‘agreement and/or concerted practice’ satisfies the conditions to be considered as such. (56)
89. Naturally, the undertakings charged with having participated in the infringement have an opportunity to dispute, for each form of conduct, the characterisation(s) applied by the competition authority by contending that the authority has not adduced adequate proof of the constituent elements of the various forms of infringement alleged. (57)
90. In the light of the foregoing, I am of the view that, when establishing an infringement of Article 101(1) TFEU, a competition authority is not required to categorise a given form of conduct as an agreement or as a concerted practice, provided that it adduces adequate proof of the constituent elements of the various forms of infringement alleged.
D. Third question
91. By its third question, which I shall examine last, the referring court asks the Court whether, in a situation such as that in the main proceedings, in which the credit card companies were not directly involved in defining the content of the agreement but facilitated its adoption, accepted and implemented it, they should, for the purposes of Article 101(1) TFEU, be considered to be parties to that agreement, or to have participated in a concerted practice.
92. As explained in the previous section of this Opinion, a competition authority is generally not required, when establishing an infringement of Article 101(1) TFEU, to categorise conduct as an agreement or as a concerted practice.
93. In the case at hand, it would thus be sufficient that the form of collusion or coordination that took place between the credit card companies and the banks which were parties to the MIF Agreement reaches the threshold to be considered a ‘concerted practice’ in order to hold the former responsible for the infringement alleged by the competition authority.
94. Two further issues were addressed by the parties before the Court and merit further discussion.
95. First, does the fact that the credit card companies operate in a market that is different from the one in which the agreement in question was implemented mean that those companies cannot be considered responsible for the presumed infringement of Article 101(1) TFEU?
96. The answer to that question is clearly negative. The principle underlying Article 101 TFEU is that undertakings should decide their policy on the market in an independent fashion, without engaging in any form of direct or indirect contact that may unduly affect their freedom of action. (58) To that end, the scope of Article 101(1) TFEU is, as mentioned in point 23 above, quite broad, in order to catch all forms of collusion or coordination that may lead to that result.
97. The concept of an ‘agreement’ within the meaning of Article 101(1) TFEU is not confined to so-called ‘horizontal agreements’ between undertakings active on the same market (and thus in actual or potential competition with each other). Many examples can be found in the case-law of the EU Courts in which agreements between companies active in different stages of the production chain or in neighbouring markets have been found to infringe Article 101(1) TFEU. (59) The same logic must also apply with regard to concerted practices. (60)
98. These principles were very clearly confirmed, and to some extent developed, in the recent judgment of the Court in AC-Treuhand, (61) to which the parties made numerous references in their submissions. In that judgment, the Court emphasised that it cannot be inferred from either the wording or the rationale of Article 101(1) TFEU that its scope is limited to forms of collusion that give rise to ‘a mutual restriction of freedom of action on one and the same market on which all the parties are present’. (62) The Court further stressed that its case-law has never limited the reach of that provision to undertakings operating in the affected market, or in the markets upstream or downstream of that market, or in neighbouring markets. (63) The participation of an undertaking in an agreement or concerted practice may infringe Article 101 TFEU regardless of the type of business activities carried out by that undertaking and/or the markets in which it operates, provided that it contributes to restricting competition in a given market. (64)
99. AC-Treuhand concerned a consultancy firm which provided assistance to a cartel by supplying services of an administrative nature. (65) The Court found that the very purpose of the services provided by that undertaking, on the basis of service contracts concluded with the cartel members, was the attainment of the anticompetitive objectives in question. That undertaking thus made an active contribution to the implementation and running of a cartel, while being fully aware of the unlawfulness of such activity. (66)
100. The question of whether the situation of the credit card companies may satisfy the conditions set out in AC-Treuhand regarding the liability of a ‘facilitator’ under Article 101(1) TFEU, discussed by the parties at length, is in the context of the present case a bit of a red herring. The reason is simple. In the factual and legal context of the present case, the credit card companies were not in a situation comparable to that of the company AC-Treuhand, namely that of a mere ‘facilitator’. On the facts of the case as presented by the referring court, they appear to be much more.
101. According to the information provided by the referring court, the credit card companies did more than merely ‘facilitate’ the agreement. They encouraged the banks to find an agreement and, although not formally present during the negotiations, their interests were represented in those negotiations by one bank (Kereskedelmi és Hitelbank Zrt.). In addition, the credit card companies provided for the agreement in their internal rules, were informed of the conclusion of the agreement and duly implemented it, also vis-à-vis the banks that joined the network at a later date.
102. Furthermore, unlike in the case of the company AC-Treuhand, the credit card companies had a more direct and immediate interest in the successful execution of the agreement. Indeed, they were not merely service providers hired by the banks to carry out certain specific tasks. MasterCard and Visa were suppliers of the credit cards the use of which was the subject of the MIF Agreement. The credit card companies thus did not operate in a market which was unrelated to that affected by the MIF Agreement, but in a directly concerned upstream market. The fact that they apparently did not receive any portion of the MIF directly does not detract from their interest in the successful execution of the MIF Agreement.
103. To my mind therefore, the situation in the present case falls squarely within a more ‘traditional’ vertical scenario: it has long been established that agreements or concerted practices between companies active in different stages of the production chain may infringe Article 101 TFEU. (67)
104. A second and final issue that arises from the question referred concerns the circumstances in which, in the case at issue in the main proceedings, the credit card companies may be held liable for the whole infringement, alongside the banks which were parties to the MIF Agreement.
105. The answer to that issue can equally be found in the settled case-law. In order for an authority to be able to find that an undertaking participated in an infringement and was liable for all the various elements comprising the infringement, it must prove that the undertaking concerned intended to contribute by its own conduct to the common objectives pursued by all the participants and that it was aware of the actual conduct planned or put into effect by other undertakings in pursuit of the same objectives or that it could reasonably have foreseen it and that it was prepared to take the risk. (68)
106. In its request for a preliminary ruling, the referring court points out that the credit card companies played no part in the drafting of the MIF Agreement or in determining the amount of the IF. However, as mentioned in point 101 above, the referring court considers those companies to have encouraged its formation, facilitated its adoption, and accepted and implemented it.
107. If that is indeed so on the facts of the case, which is for the referring court to ascertain, I would have little hesitation in concluding that, in the light of their role and their position vis-à-vis the banks that were parties to the MIF Agreement, the credit card companies did participate in the alleged infringement of Article 101(1) TFEU. According to settled case-law, neither the fact that an undertaking has not taken part in all aspects of an anticompetitive scheme nor that it played only a minor role has an impact on the establishment of the existence of an infringement its part. (69)
108. Consequently, I suggest that the Court answer the third question to the effect that, in a situation such as that in the main proceedings, in which the credit card companies were not directly involved in defining the content of an agreement alleged to infringe Article 101(1) TFEU, but facilitated its adoption, accepted and implemented it, which is for the referring court to verify, those companies may be considered liable for that infringement.
V. Conclusion
109. I propose that the Court answer the questions referred for a preliminary ruling by the Kúria (Supreme Court, Hungary) as follows:
– the same conduct of an undertaking can be held to infringe Article 101(1) TFEU for having both the object and the effect of restricting competition in the internal market;
– it is for the referring court to verify whether the MIF Agreement constitutes a restriction by object. To that end, the referring court must first examine the content and objective of that agreement to determine whether it falls into a category of agreements that, in the light of experience, is generally recognised as being harmful to competition. If the answer to that question is positive, the referring court should then verify whether that finding is not called into question by considerations relating to the legal and economic context in which that specific agreement was implemented. In particular, the referring court is to check whether any alternative explanation as to an allegedly procompetitive rationale for the MIF Agreement is prima facie plausible, taking account also of the time when the agreement was in operation;
– when establishing an infringement of Article 101(1) TFEU, a competition authority is not required to categorise a given form of conduct as an agreement or as a concerted practice, provided that it adduces adequate proof of the constituent elements of the various forms of infringement alleged;
– in a situation such as that in the main proceedings, in which the credit card companies were not directly involved in defining the content of an agreement alleged to infringe Article 101(1) TFEU, but facilitated its adoption, accepted and implemented it, which is for the referring court to verify, those companies may be considered liable for that infringement.