This paper – which can be found here – provides an overview of the UK MasterCard litigation. Mr. Veljanowski is likely very well placed to discuss this:  he was one of the two economic experts involved in a case recently decided by the CAT on the matter. He also seems to publish a paper about every court decision concerning the MasterCard litigation (see my post of 24 March 2017, regarding the Arcadia v MasterCard case).

The paper begins with a quick overview of the MasterCard litigation. As a result of the European Commission’s MasterCard decision, there are currently about 25 separate standalone and follow-on retailer actions making their way through the English courts concerning MasterCard and Visa’s card systems’ interchange fees. The first decision in these cases was adopted by the CAT last year (Sainsbury v MasterCard). The second one was the Arcadia v MasterCard case I posted about on 10 February. There are also more recent decisions by the CAT (regarding a collective action against MasterCard) and the High Court (regarding whether VISA is liable for damages), which I will review shortly.

The author notes that these should have been simple cases: competition agencies across the world have concluded that credit and debit interchange fees were excessive (including in Australia, US, UK, Europe, etc.). Furthermore, the European Commission made MasterCard agree to lower interchange fees. In the light of this, the author argues that one could have been expected the resolution of the cases to be straightforward.

Instead, not only does he consider that the two existing decisions are long and convoluted, but that they actually contradict each other (and I would point out that the recent Visa decision adopted yet a different approach that goes against both prior judgments on the matter). The main thesis of this article is that this is explained by difficulties regarding the counterfactual used, the applicable theory of harm, and the adequate approaches to calculating counterfactual interchange fees in practice.

An analysis of the respective decisions of the CAT and the High Court is then quickly pursued. A preliminary point is that these are not follow on claims (i.e. the Commission decision is not binding) because they refer to the inter-UK multilateral interchange fee, while the Commission’s decisions concerned cross-border interchange fees.

In any event, the CAT followed the European Commission in considering that the appropriate counterfactual was one where the interchange fees were established in bilateral negotiations between card issuers and card acquirers. The Tribunal set the counterfactual default interchange fee at zero, and hypothesized that individual card issuers and acquirers would negotiate positive but lower bilateral interchange fees. The Tribunal calculated the counterfactual bilateral interchange fees using MasterCard’s bi-annual costing data (similarly to what the European Commission did in its decision regarding Visa’s interchange fee) to arrive at maximum bilateral interchange fees of 0.5% and 0.3% of card transaction value for credit and debit cards respectively.

The High Court, on the other hand, rejected the idea that interchange fees could be negotiated bilaterally. Without interchange fees, the MasterCard card scheme would not have survived, and therefore the multilateral interchange fee was a necessary ancillary restraint. As such, the counterfactual was the multilateral interchange fee value above which there would be an infringement of competition law. Since the court concluded that the multilateral interchange fee effectively charged by MasterCard as above this threshold, the High Court concluded that there MasterCard’s UK interchange fees did not infringe competition law.

In a third section, the author argues that the courts reached different conclusions because they chose different counterfactuals. At this point, he engages in a quick discussion of the theory of counterfactuals.

Counterfactual propositions comprise initial conditions (the counterfactual antecedent), hypothesized consequences (the counterfactual consequent), and a theory or causal path that explains the link between the antecedent and the consequent. The counterfactual antecedent should satisfy three conditions: “minimalism”, “realism” and “cotenabilty”.  A counterfactual is likely to be more relevant if it involves as few changes as possible to the real-world situation – the so-called ‘minimal-rewrite rule’. At the same time, counterfactuals need to be plausible and probable in order to support meaningful causal statements. This means that the ‘minimal rewrite rule’ needs to be tempered by the inclusion of other changes that render a counterfactual likely – the “cotenability” criterion.  To change only one thing might be artificial and unrealistic, and therefore other related changes are justified. Lastly, the theory or causal path that explains the link between the antecedent and the consequent is what in competition law is commonly known as a “theory of harm”– and this theory should apply not only to the counterfactual but to the actual world as well.  As such, the court needs to choose between a number of different possible counterfactuals; and each counterfactual can have many outcomes depending on the applicable theory of harm.

He then suggests that there are three features of a counterfactual that appear to have been accepted in the case law, namely:

  • The counterfactual can be defined as a hypothetical situation in which the otherwise factual scenario does not contain the impugned conduct i.e. the antitrust infringement.
  • Article 101 effects cases will often have to deal with more than a single relevant, ‘valid’ counterfactual.
  • The counterfactual must be ‘realistic’.

In the cases under analysis, the author identifies three types of possible counterfactuals:

  • No multilateral interchange fees are allowed. This, in turn, leads to five possible counterfactuals (extracted from existing payment systems that operate without such a fee): (i) issuer and merchant banks enter into ex post bilateral agreements;  (ii) ex post pricing is prohibited; (iii) issuer and merchant banks voluntarily enter into ex ante bilateral agreements; (iv) issuer and merchant banks mandatorily enter into ex ante bilateral agreements; (v) parties engage in mandatory arbitration if no agreement is reached.
  • Multilateral interchange fees are allowed, but they are set at zero.
  • Multilateral interchange fees are allowed, but only up to some value.

The choice between which counterfactual to use ultimately depends on the court’s theory of how competition would develop in in a market. As such, he argues that the existence of diverging judicial decisions is a consequence of the relevant courts having adopted different theories of competition.

  • The CAT took a conventional view on the rivalry between two card schemes (Visa and MasterCard) and the merchants’ reaction to bilateral interchange fees, and concluded that bilateral negotiations would lead to substantially lower fees. While this might involve some issuers changing from MasterCard to VISA, MasterCard’s interchange fees would then be negotiated to retain issuers who might be tempted to join Visa.
  • The High Court, on the other hand, concluded that if MasterCard’s bilateral interchange fees were lower by just 0.2% percentage points, issuers would switch to Visa en masse, and MasterCard would be caught in a “death spiral”, which meant that multilateral interchange fees were necessary.

The author then explains that the trial procedure has inherent and obvious limitations – particularly in that it requires the court to deal with parties which activities may not encompass the full market, and hence do not reflect all the relevant market’s interactions. In this case, both cases had to be decided without the benefit of evidence from VISA, which would have affected the plausibility of each counterfactual. The author takes this as a starting point to criticise the High Court’s decision, the lack of realism of the counterfactual relating to the “bilateral negotiation between issuer and merchant banks’ and the “methodology to determine what the legal bilateral exchange fee” would be.


My main interest in this paper is in how it explains the logic of, and difficulties with counterfactuals.

The application of wide standards in the context of complex markets and business interactions will always be plagued with doubts and uncertainties. Effects are hard to measure, there are often many possible counterfactuals that one can legitimately accept, and the theory of harm may rely on shaky empirical foundations or be impossible to demonstrate in practice.

All our administrative and judicial systems are geared to ensure that the conclusions these systems arrive at are plausible,  but we are rarely able to say that these conclusions are true. Acceptance of these (trite) observations helps us understand why different people can sit at the same table and have vociferous disagreements about the outcome of individual cases (cough cough Visa and MasterCard, cough cough Google) and not doubt that the intelligent, knowledgeable person disagreeing with us is in good faith.

Courts are constrained by the evidence to which they have access.  If the cases and evidence are presented by reference to an economic model that contains a counterfactual and applies a theory of harm, then the court will have to assess the plausibility of the counterfactual and theory of harm against the evidence. In such scenarios, it will not be altogether surprising if the chosen counterfactual is not 100% realistic, but is nonetheless the one that better fits the evidence or the relevant economic theory. It is also not inconceivable that courts will reason backwards – seeing what conclusions they deem most plausible, and reasoning from there to select the combination of counterfactual and theory of harm that may lead them to that conclusion. Problematic this may be – but, if I may advance my own cause a bit here, this is typical of legal reasoning in hard cases in various areas of the law. What courts do not do, however, is compare economic models to see which one is more robust according to economic methods.

As I heard someone say yesterday, it is not only competition lawyers who should learn economics; economists would do well to learn a bit of law as well.

Author Socials A weekly email with competition/antitrust updates. All opinions are mine

What do you think?

Note: Your email address will not be published