IN Case C-228/18, Gazdasági Versenyhivatal v Budapest Bank (ECLI:EU:C:2020:265), the Court of Justice has provided important guidance on the requirements for an agreement to be categorised as a “restriction by object” for the purposes of Article 101(1) TFEU. It is the latest in a line of cases emphasising that the concept of restriction of competition by object is to be interpreted restrictively.

The case arose from an agreement entered into in 1996 between Hungarian banks operating within the Visa and MasterCard card payment schemes, setting uniform multilateral interchange fees (MIFs) for transactions under both schemes (MIF Agreement). Interchange fees are charged by card issuers (banks that issue cards to consumers) to card acquirers (banks that provide services to retailers to enable them to accept cards as a means of settlement), and form part of the merchant service charge levied by acquirers on retailers. The setting of MIFs within the framework of payment card schemes has been a controversial issue over the last 20 years: the Commission has brought multiple challenges against the Visa and MasterCard MIFs as an infringement of Article 101(1) TFEU and there have recently been numerous damages actions against the card schemes for breach of Article 101(1). However, the central issue in Budapest Bank was not the existence of MIFs per se, but the fact that, under the MIF Agreement, the MIFs charged under both the Visa and MasterCard schemes were fixed at the same level. In 2009 the Hungarian competition authority (HCA) adopted an infringement decision under Article 101(1) TFEU and the equivalent domestic law, imposing fines on the member banks and the credit card companies, finding that the MIF Agreement constituted both a restriction by object and a restriction by effect. Its decision was appealed and reached the Hungarian Supreme Court, which referred a number of questions to the Court of Justice. Central among these were whether an agreement could infringe Article 101(1) both because of its anti-competitive object and effect (which the court had no difficulty answering in the affirmative), and whether the MIF Agreement was correctly categorised as a restriction by object (where both the court and A.G. Bobek entertained more doubts).

Article 101(1) TFEU refers to agreements which have as their “object or effect” the prevention, restriction or distortion of competition and it is well established that “object” and “effect” are alternative bases for proving an infringement (Case 56/65, LTM [1966] E.C.R. 236, 249). The object/effect distinction assumes critical importance because once an agreement has been shown to have an anti-competitive object, it is unnecessary to prove that it also has anti-competitive effects (Case C-345/14, Maxima Latvija, EU:C:2015:784, at [17]). For a competition authority there are obvious advantages in pursuing a case as a restriction by object, as there is no need to embark on a detailed economic analysis of its effects. On the other hand, as the court in Budapest Bank made clear, a finding of a restriction by object in no way precludes a finding that the same agreement has the effect of restricting competition (at [40], [44]).

But where should the limits of the restriction by object concept be drawn? Cartes Bancaires (CB) (Case C-67/13 P, EU:C:2014:2204) confirms that the concept is to be interpreted restrictively and that it applies only to certain types of coordination “which reveal a sufficient degree of harm to competition that it may be found that there is no need to examine their effects” (at [58]); otherwise there is a risk of condemning agreements which may turn out on closer analysis not to be anti-competitive. As A.G. Wahl noted in his Opinion in that case (at [56]):

As to how the assessment should be carried out, the answer provided by the court in CB is to consider “the content of its provisions, its objectives and the economic and legal context of which it forms a part” (at [53]). For these purposes, “context” includes the nature of the products affected and the markets involved (ibid.).

Classic examples of restrictions by object include (horizontal) price-fixing, output limitation, market-sharing and (vertical) resale price maintenance. But the application of the restriction by object concept to other forms of coordination has proved more controversial. In Generics (UK) Ltd. (Case C-307/18, EU:C:2020:52), the Court of Justice recently gave a nuanced response to the question of whether pharmaceutical pay-for-delay agreements could be characterised as a restriction by object, noting that the mere fact that they involve a transfer of value by an originator to a generic is not enough (at [84]–[86]), unless it is clear that this can only be explained by the parties’ commercial interest “not to engage in competition on the merits” (at [87]). For A.G. Bobek in Budapest Bank the identification of a restriction by object required a two-step analysis. The first was to consider the content of the agreement and its objectives, to ascertain whether the agreement fell within a category whose harmful nature was, “in the light of experience, commonly accepted and easily identifiable” (at [42]). For these purposes, “experience” referred to “what can traditionally be seen to follow from economic analysis, as confirmed by the competition authorities and supported, if necessary by case-law” (ibid.). The second was to ensure that this purely formal assessment was “not called into question by considerations relating to the legal and economic context in which the agreement was implemented” including the nature of the products affected and the markets involved; in addition, the parties’ subjective intentions could be taken into account where relevant (at [43]). While this approach is pragmatic, it necessarily entails some consideration of the economic effects of an agreement and therefore risks blurring the object/effect distinction, a point acknowledged by A.G. Bobek, who admitted that he was unable to draw a bright line between this second step and a full effects analysis (at [49]); but in his view this was justified as a “basic reality check” to ensure there were no legal or factual circumstances pointing in the opposite direction (at [49]–[50]).

Applying this logic, both the A.G. and the court had doubts about whether the MIF Agreement could properly be categorised as a restriction by object. One issue for A.G. Bobek was that the MIF Agreement did not obviously entail horizontal price-fixing or vertical resale price maintenance; it merely standardised one aspect of the cost structure under both card schemes (at [53]), added to which was a lack of clarity in the competitive harm alleged by the HCA (at [56]–[60]). The court seemed less convinced, noting that it would have been sufficient that the MIF Agreement could be viewed as a form of indirect price-fixing, indirectly determining the merchant service charges (at [62]); this meant the MIF Agreement could not be ruled out a priori from being a restriction by object (at [63]). But nor could the court exclude that the real objective of the MIF Agreement was to ensure a balance between the interests of card issuers and acquirers by allowing for recovery of some costs but preventing interchange fees and merchant service charges rising to excessively high levels (at [73]). There were also suggestions from the Hungarian Supreme Court that by removing price competition between the two card schemes the MIF Agreement could have enhanced competition on parameters other than price (at [74]). Although the court noted that this was a matter for the referring court, it added that if that were the case, an infringement of Article 101(1) could only be established on the basis of a full effects analysis, by considering the counterfactual; in other words, how competition would have developed in the absence of the MIF Agreement (at [75]).

Perhaps the most powerful reason given by A.G. Bobek and the court for doubting whether the MIF Agreement could be viewed as a restriction by object stemmed from the lack of a “reliable and robust wealth of experience” showing that such agreements are commonly accepted as anti-competitive (Opinion, at [63]). The problem as articulated by A.G. Bobek (at [66]–[68]) was that the practice on which the HCA had said it was relying (essentially a series of Commission interventions against MIFs) appeared inconclusive and pointed in different directions. These included Visa International (COMP/29.373, OJ 2002 L 318 p.17) where the Visa MIF was expressly considered not to amount to a restriction by object, and was exempted; MasterCard, EuroCommerce and Commercial Cards (COMP/34.579, 36.518 and 38.580, OJ 2009 C 264 p.8) finding “fallback” (default) MIFs to be a restriction by effect, but without any conclusion on whether they amounted to a restriction by object; and Visa MIF (AT.39398, OJ 2019 C 299 p.8) accepting commitments (which necessarily entail no finding of infringement) to cap the level of various MIFs. Only in MasterCard II (AT.40049, OJ 2019 C 300 p.6) had the Commission found a restriction by object. Nor could it be said that the case law of the EU courts or of national courts assisted the debate (at [70]–71]), and, as A.G. Bobek observed, there was seemingly no consensus among economists supporting the argument in favour of a restriction by object (at [72]). Significantly, the court (at [76]) appears to endorse A.G. Bobek's reasoning on this point, noting the requirement for “sufficiently reliable and robust experience” to justify a finding of a restriction by object, adding (at [79]) that the information relied on by the HCA, the Hungarian Government and the Commission, primarily the HCA's decisional practice and EU case law, demonstrated “as things currently stand” the need to conduct a detailed effects analysis, including consideration of the counterfactual. This aspect of the case is likely to have ramifications well beyond the card payment sector, where competition authorities are using novel theories of harm to challenge types of agreement that have not previously been accepted as inherently anti-competitive. Absent evidence of a clear economic consensus, a consistent line of competition authority practice or case law precedent, Budapest Bank means that it is likely to be difficult to characterise such agreements as a restriction by object.