It Takes Two to Tango: Two-Sided Markets and the Appeals in Cartes Bancaires and MasterCard

By Hans Vedder

The Court has recently decided on the appeals in two seminal cases: MasterCard MIF (MasterCard) and Groupement des Cartes Bancaires (CB). Both cases result from Commission decisions that found Article 101 TFEU to have been infringed by the decisions taken within those schemes with regard to fees that form part of the working of these payment systems. To understand both cases it is necessary to first set out the background to the MasterCard and CB systems. After that we will examine the procedure and finally the judgments themselves. This will reveal essentially three interesting issues:

  1. the object-effect dichotomy,
  2. the relation between the exclusion of competitors and the object category, and
  3. the possibility to take into account redeeming features.

Bank, Debit and Credit Cards as Two-Sided Markets

Both cases deal with bank/debit cards or credit cards and thus represent two-sided markets. This follows from the fact that the attractiveness of such cards depends on the number of places where they can be used to pay or withdraw cash, which in turn depends on the number of consumers that have such a card and that in turn depends on the scale and density of the network where such cards can be used. This intuitive definition of a two-sided market therefore essentially turns on a network effect that connects two separate, but connected markets: one for consumers that are issued such cards by their banks and one for ‘card acceptance points’. The card acceptance points are the various ATMs, but also the merchants with whom the consumers do business. The judgments refer to these sides as the issuing side (or activity) and the acquiring side (or activity).

Needless to say, both sides involve contracts between the banks and the consumers/merchants and these contracts result in different costs, profits and risks. The contract between an issuing bank and the consumer will reflect the risk of that consumer not being able to pay the credit he has taken out and the interest rate that the bank will charge. The contract between the acquiring bank and the merchant is less interesting and boils down to a promise that money will be transferred, without a credit risk and concomitant interest rate. There may, however, be other financial incentives at stake that influence the costs / profitability of such a contract for the merchant and bank involved. This means that there are different incentives for an issuing and an acquiring bank. For the issuing bank there is the (attractive) prospect of consumers buying on credit and thus having to pay interest rates. Assuming that acquiring activities are less profitable, a rational bank, therefore, would be more interested in the issuing side. This would translate into a reduced incentive to invest in the acquiring side of the market and thus reduce the network of merchants that will accept the card. Alternatively, a card system may be so attractive that consumers may be willing to pay the merchant for being able to use it. In that case, the incentive for the banks may be to become an acquiring bank. It all depends on the attractiveness of a certain card system for the banks involved.

Whatever may be of the financial flows involved, the fact remains that in a two-sided market there is a need for the card organization to ensure that both sides do their best to contract as many customers and merchants. Whether it is phrased as a solidarity problem (between those banks that issue more cards and those that have more contracts with merchants) or as a free riding problem, the fact remains that the card organization has an incentive to keep banks active on both markets. In a financial world, this incentive is likely to take the form of a financial transaction; a fee to be paid by the issuing banks to the acquiring banks. This interchange fee has a different name in the two cases (multilateral interchange fee (MIF) in MasterCard and Merfa in CB) and the modalities differ as well, but this is what it boils down to. In addition to this interchange fee, the CB case also involved a reform of the membership fee and the fee to wake up dormant members. Both fees serve a sensible purpose, inter alia to encourage members to issue more cards, but at the same time also raise costs.

Interchange fees and Competition Law

So why is this a competition case? From this very short explanation of the ways in which such payment systems work, the relevance of competition law may not be obvious. However,knowing that the interchange fees end up inflating the prices charged by the banks to their customers (whether the merchants or the card holders) changes the situation. Given that these fees are set by the card organization that is set up by and closely affiliated to the banks, it is just a small step to concluding that we have a decision by an association of undertakings that raises the prices charged by banks for the use of a card system. In CB the idea seems to have been to keep entrants off the market (by raising their costs, as entrants are likely to be more active on the issuing than the acquiring side) and to raise prices for the issuing of cards. In MasterCard the fee was argued to have inflated the costs charged from the merchants, which would then be passed on to consumers. So we have something that arguably restricts market entry and looks, smells and barks like a price fixing agreement, but is it also a hard core price fixing cartel?

According to the Commission it was. It classified the facts in both cases as a decision by an association of undertakings that restricts competition. In CB, the Commission found a restriction by object and in MasterCard the restriction was one by effect. Moreover, the Commission had a look at the redeeming features (the need to have an adequate network), but rejected these arguments.

This decision was then appealed by the various parties leading to the General Court judgments that were appealed, leading to the two judgments discussed here.

  1. The Object-Effect Dichotomy in Article 101(1) TFEU

This issue is dealt with primarily in the CB case, as that one turned to a significant extent on the short-cut analysis of such object restrictions. Ever since Consten and Grundig it is known that effects on competition need not be proven if a case can be brought under the object heading. This makes the life of anyone trying to prove that Article 101(1) TFEU was infringed a lot easier (see para. 35 of A-G Wahl’s Opinion). Making life for the Commission or a National Competition Authority easier invariably makes life for the accused undertaking(s) more difficult. As a result, the boundary between the object and effect category must be clear and the reasons for bringing a specific restriction in one or the other must be sound. The question is thus whether the judgment in CB provides a clear dichotomy and good reasons for the qualification of a restriction under the object or effects heading? Readers of this blog may recall that I have been critical for the Court’s conflation of the analysis required to classify a restriction under the object-effect dichotomy. In Allianz, to my mind, the Court ended up prescribing something that looks an awful lot like an effects analysis to determine whether or not something qualifies as an object restriction (see also paras. 48 – 52 of A-G Wahl’s opinion and Chillin’Competition and the Kartellblog). To make matters worse: the Court found a restriction by object in that case, where the actions of the companies were essentially aimed at preserving their market shares.

Interestingly, the Court relies extensively on the judgment in Allianz in paragraphs 49 – 53 of CB.

49 In that regard, it is apparent from the Court’s case-law that certain types of coordination between undertakings reveal a sufficient degree of harm to competition that it may be found that there is no need to examine their effects (see, to that effect, judgment in LTM, 56/65, EU:C:1966:38, paragraphs 359 and 360; judgment in BIDS, paragraph 15, and judgment in Allianz Hungária Biztosító and Others, C‑32/11, EU:C:2013:160, paragraph 34 and the case-law cited).

50 That case-law arises from 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 (see, to that effect, in particular, judgment in Allianz Hungária Biztosító and Others (EU:C:2013:160) paragraph 35 and the case-law cited).

51 Consequently, it is established that certain collusive behaviour, such as that leading to horizontal price-fixing by cartels, may be considered so likely to have negative effects, in particular on the price, quantity or quality of the goods and services, that it may be considered redundant, for the purposes of applying Article 81(1) EC, to prove that they have actual effects on the market (see, to that effect, in particular, judgment in Clair, 123/83, EU:C:1985:33, paragraph 22). 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.

52 Where the analysis of a type of coordination between undertakings does not reveal a sufficient degree of harm to competition, the effects of the coordination should, on the other hand, be considered and, for it to be caught by the prohibition, it is necessary to find that factors are present which show that competition has in fact been prevented, restricted or distorted to an appreciable extent (judgment in Allianz Hungária Biztosító and Others (EU:C:2013:160), paragraph 34 and the case-law cited).

53 According to the case-law of the Court, in order to determine whether an agreement between undertakings or a decision by an association of undertakings reveals a sufficient degree of harm to competition that it may be considered a restriction of competition ‘by object’ within the meaning of Article 81(1) EC, regard must be had to the content of its provisions, its objectives and the economic and legal context of which it forms a part. When determining that context, it is also necessary to take into consideration the nature of the goods or services affected, as well as the real conditions of the functioning and structure of the market or markets in question (see, to that effect, judgment in Allianz Hungária Biztosító and Others (EU:C:2013:160), paragraph 36 and the case-law cited).

So Allianz is here to stay, but the surprise is in paragraph 51, where the Court finally gives the reasons that underlie the choice for bringing a restriction under the object or effect heading: the overwhelming likeliness of negative effects that follows from experience. To the historically inclined competition law scholars out there: the reference to Clair as a precedent makes very little sense to substantiate the statement in CB as it does nothing more than reiterate the object effect dichotomy that we already know from Consten and Grundig. Take out that superfluous reference and we end up with reasoning that echoes that proposed by A-G Wahl in paragraph 56 of his Opinion:

Only conduct whose harmful nature is proven and easily identifiable, in the light of experience and economics, should therefore be regarded as a restriction of competition by object, and not agreements which, having regard to their context, have ambivalent effects on the market or which produce ancillary restrictive effects necessary for the pursuit of a main objective which does not restrict competition.

So, there is the reasoning underlying and test for the object-effect dichotomy combined in one neat package. We only need to ask ourselves if the harmful nature is proven in economics and by experience and is obvious. If there are any redeeming features, the object category should not apply. A-G Wahl tops this up with a soundly reasoned opinion that the object category should be approached with a ‘relatively cautious attitude’ (para. 58).

Whilst the Court does not explicitly reiterate this cautious attitude, the way it structures the application of the test to the facts of the case follows that proposed by the Advocate-General which represents a welcome departure from the way the Court applied the same test in Allianz. The judgment in CB confirms the extensive context test handed down in Allianz. Fortunately, it results in a completely different outcome. Whereas the agreements in Allianz were brought in the object box, the Court reaches a different conclusion in CB. This can be seen as caution on the side of the Court. This cautious attitude follows from the analysis of the CB scheme as a two-sided market and the corresponding need to take into account the ways in which the network effects and need to address free riding are being dealt with by CB (paras. 75 – 79). In the context analysis prescribed by the Court to determine whether the object or effects category applies, both sides of this market and the relation between these sides must be taken into account (paras. 78, 79). In the theory of harm put forward by the Court in paragraph 51 of CB, what matters is that experience shows that the behaviour under investigation leads to falls in production and price increases, resulting in reductions of consumer welfare, and this was simply not proven by the Commission in the decision.

  1. The Context Analysis, Object Box and the Exclusion of Competitors

The Court then goes on to make the interesting statement that the fees charged as part of the CB system may deter or even rule out market entry for potential competitors. This is put squarely in the effects box by the Court in paragraph 81. Moreover, this finding corresponds to the theory of harm in paragraph 51 that attaches primarily to the behaviour that has negative results for consumer welfare, not so much behaviour that impacts the market structure. In this regard, the Court also distinguishes the scheme in CB from that in BIDS case (paras. 83 – 85). The latter involved the decision by the members of the meat and veal processing industry to collectively phase out excess capacity in that industry, that was put in the object box. According to the Court putting BIDS in the object category was warranted because it would phase out almost 75% of the overcapacity in the industry and reduce overall capacity with 25% which amounts to ‘an appreciable change’ of the market structure.

The fact that the Court not only looks at effects on prices and output, but also on the market structure, shows that the Court’s theory of harm is broader than could be inferred from paragraph 51 of CB and definitely includes the protection of the competitive market structure (see to this point T-Mobile, para. 38). Confusingly, the Court appears to hold that when the effects on the market structure are appreciable – a term previously used only in connection with an effects analysis – the coordination can be brought in the object box. This raises two questions: Firstly: how deep should the context analysis be and, secondly, what does this do to Article 102 TFEU?

As to the depth or level of detail with which the context analysis should be undertaken, the judgment is unclear. Part of the context analysis appears to be defensive and deals with the redeeming features or justification of the allegedly restrictive conduct, likely to be put forward as defences by the accused undertakings, such as the prevention of free riding in CB. Another part of the context analysis appears to be more offensive, and could see the competition authority trying to prove that the behaviour may have appreciable exclusionary effects. This of course begs the question when exclusion or deterrence of market entry is appreciable? If the reduction of overcapacity with 25% suffices for the qualification as a restriction by object, how does this measure up against a reduction of potential competition in the French market for bank cards?

As to Article 102 TFEU, I would be inclined to say that this ruling puts exclusionary effects outside the Article 102 TFEU per se box, if they are appreciable. To my mind the treatment of an exclusionary effect should not be different whether it results from pre-existing market power (an Article 102 TFEU case) or from created market power (an Article 101 TFEU case). If such effects are brought under the object heading of Article 101 only when they are appreciable, the same should apply to the functional equivalent thereof in Article 102 TFEU, which is the per se box.

  1. Redeeming Features and Article 101 TFEU

Finally, some words on the possibility to take into account redeeming features or justifications. We have already seen that redeeming features or justifications play a part in the context analysis. The presence of such redeeming features refutes the obvious harmfulness of the behaviour, bringing it in the effects box. The question then becomes how such justifications are taken into account. Is there room for these in the first paragraph or should such balancing take place in the third paragraph of Article 101 TFEU?

MasterCard clarifies this with regard to the so-called ancillary restraints to the effect that such restraints must be objectively necessary and proportionate to the main operation in that this operation must be impossible to carry out without the restraint in question (para. 89 – 91). All other restraints,  those that reduce the costs of the main operation or make it easier to implement, can only be saved within the framework of Article 101(3) (para. 93). The essential question is therefore whether the main operation, in MasterCard a functioning credit card payment scheme, would work without the interchange fee scheme called MIF?  This opens up the Pandora’s Box of Counterfactuals and What Ifs. The Counterfactual in MasterCard was a prohibition of ex post pricing. This would ban any issuing bank from imposing costs on any merchant/acquiring bank. MasterCard regarded this as a rule that it would never impose without regulatory intervention. So does that rule it out as an appropriate counterfactual? The Court holds that the counterfactual hypothesis should be ‘appropriate to the issue it is supposed to clarify and that the assumption on which it is based is not unrealistic’ (para. 108). This means that the counterfactual was correct and that the General Court was right in saving it. This also ties in nicely with the fact that the MasterCard scheme apparently functions in some Member States where there are no MIFs at all (see Commission press release). This results in a finding of restrictive effects, meaning that the redeeming features could only be taken into account within the framework of Article 101(3) TFEU (para. 180).


This is already an excessively long post (not unlike the legal procedures that have resulted in MasterCard and CB) and there are still many more interesting things to say about these two judgments. They are certainly to be welcomed as clarifications of the law applicable to the object-effect dichotomy, but also as some of the first cases that deal with two-sided markets.

Of course, it takes two to tango. This is nice when it comes to that dance or the two-sided markets that were under investigation. It is not so nice when it takes an appeal from a General Court judgment to get a better understanding of something as fundamental as the object-effect dichotomy. It is probably even worse when it takes a second chamber within the Court to take the tango back on course after a faux pas in an earlier judgment.