Sainsbury's Supermarkets Ltd v Mastercard Incorporated & Ors (Rev 1)

The central question in these three appeals is whether the setting of default multilateral interchange fees (“MIFs”) within the MasterCard and Visa payment card systems contravenes article 101 of the Treaty on the Functioning of the European Union 2012/C326/01 (the “TFEU”). (Footnote: ¹) Article 101(1) provides that agreements between undertakings which may affect trade between Member States and which have as their object or effect the prevention, restriction or distortion of competition within the internal market are prohibited as incompatible with the internal market of the European Union. Section 2 of the Competition Act 1998 (the “1998 Act”) makes the same provision in relation to agreements which may affect trade within the United Kingdom, and which prevent, restrict or distort competition within the United Kingdom.

Two of the appeals are brought from the Commercial Court, and one from the Competition Appeal Tribunal (the “CAT”). In the broadest of outline, the CAT decided in Sainsbury’s v MasterCard (the “CAT case” or “Sainsbury’s v MasterCard”) that the MIFs charged within the MasterCard payment system were prohibited anti-competitive agreements, whilst the two Commercial Court judges decided, for different reasons, in each of AAM v MasterCard (the “AAM case” or “AAM v MasterCard”) and Sainsbury’s v Visa (the “Visa case” or “Sainsbury’s v Visa”) that the MIFs charged respectively within the MasterCard and Visa payment systems were not prohibited anti-competitive agreements. It falls to this court to resolve the considerable differences of approach between the three decisions under appeal.

Both the MasterCard and the Visa payment card schemes are known as four-party schemes, though there is in each case a fifth party, namely the scheme operator itself. The schemes operate in an essentially identical way, which can be represented by the following diagram:

The essentials of the four-party payment card scheme can be summarised as follows:

It is at this stage important to note that these appeals are not concerned with threeparty payment card schemes such as that operated in the UK by American Express and Diners Club. In that type of scheme, the scheme operator (American Express or Diners Club) itself deals directly with both cardholders and merchants, each of which pay it a fee. Payments are cleared through the scheme operator itself.

The relevant markets should also be noted. In this regard, we can gratefully adopt paragraph 47 of Phillips J’s first judgment as follows:

“It is common ground that four-party systems such as the [Visa] Scheme and the MasterCard system give rise to what is described as a “two-sided market”. One side is that in which Issuers compete with each other for the business of customers to whom they will issue cards (“the issuing market”), the other is that in which Acquirers compete for the business of Merchants (“the acquiring market”). The two sides are closely linked and dependent on each other: the value of a Visa card to a cardholder is dependent on the extent to which it is accepted by Merchants and, correspondingly, the benefit Merchants gain from accepting Visa cards is dependent on the extent that consumers have and use those cards. Precisely what benefits Merchants gain from card transactions is a matter of dispute, but it is common ground that they benefit at least to the extent that card transactions are cheaper for them than cash transactions, involving the time costs, increased staff costs and bank charges of handling and banking cash.”

Each of the appeals raises three primary issues which can be expressed shortly. (Footnote: ²)

The other significant issues that arise are as follows:

We have made every attempt in the judgment to use plain intelligible language rather than jargon. Some of what is written in this area of the law can be confusing, not because the concepts are particularly difficult, but because the premise for each proposition is not properly explained. We shall try to avoid that situation.

The detailed statutory foundation to the issues under consideration in these appeals is set out in annex 2 to this judgment. Annex 2 includes the relevant provisions of the TFEU, the 1998 Act, the European Commission’s Guidelines (the “Guidelines”) on which all parties placed considerable reliance, and the (Irish) Competition Act 2002 (the “Irish Act”).

The approach adopted in this judgment is to deal with matters in the following order:

On 24 July 2002, the European Commission (the “Commission”) decided that Visa’s intra-European Economic Area MIFs (“EEA MIFs”) were restrictive of competition for the purposes of what is now article 101(1), but would be exempt for a period of 5 years under what is now article 101(3), subject to undertakings from Visa (the “Visa Exemption Decision”). Visa undertook that it would: (i) reduce its EEA MIFs over the period so that (on a weighted average basis) debit and credit card transactions would respectively attract maximum MIFs of €0.28 and 0.7% of transaction value; (ii) further ensure that its MIFs did not exceed the sum of three defined categories of issuer costs, about which data was to be collected; (iii) make information available to merchants about MIF levels and the issuer cost percentages on which these were based; and (iv) differentiate MIFs for mail order and telephone transactions from those for face-to-face transactions (the “Visa Exemption”). MasterCard was not bound by this decision, and did not respond by reducing its EEA MIFs.

On 6 September 2005, the Office of Fair Trading (“OFT”) decided that MasterCard’s intra-UK MIFs (“UK MIFs”) restricted competition under article 101(1) and were not exempt under article 101(3). MasterCard appealed the OFT’s decision to the CAT, with Visa Europe intervening and making submissions. In response to the appeal, the OFT attempted to alter the factual basis on which it had arrived at its decision against MasterCard, and subsequently withdrew its decision on procedural grounds, as the CAT formally confirmed on 10 July 2006 ([2006] CAT 14). The OFT, and its successor the Competition and Markets Authority (“CMA”), continued to investigate MasterCard’s and Visa’s UK MIFs.

On 19 December 2007, the Commission decided that MasterCard’s EEA MIFs had, since 22 May 1992, been in breach of article 101(1), and MasterCard had not proved to the requisite standard that any of the first three article 101(3) exemption criteria were met (the “Commission’s decision”). MasterCard appealed the Commission’s decision to the General Court, and in the meantime reduced its EEA MIFs to zero. Visa did not respond by reducing its EEA MIFs.

The Visa Exemption expired on 31 December 2007 and, in 2008, the Commission recommenced its investigation into Visa’s EEA MIFs. In September 2008, it informed Visa that it favoured the Merchant Indifference Test (or MIT) for assessing whether its MIFs were lawful, rather than the issuer-based costs methodology previously used (including in the Visa Exemption Decision) (the “Issuer Costs Methodology”). On 3 April 2009, the Commission sent a Statement of Objections to Visa Europe concerning its EEA MIFs.

Meanwhile, MasterCard and Visa had been discussing their EEA MIFs with the Commission. In July 2009, MasterCard increased its EEA MIFs from zero to 0.3% for credit cards and 0.2% for debit cards. On 8 December 2010, the Commission adopted a decision accepting commitments offered by Visa regarding its EEA debit card MIFs. The decision (i) required Visa to reduce its weighted average EEA debit MIF to 0.2%, (ii) recorded the allegation that the MIFs had both the object and effect of restricting competition, and (iii) without making a finding on liability, and subject to compliance with the decision, held that the Commission would not take further action against Visa in relation to its EEA debit MIFs.

On 23 May 2012, Asda and Morrisons each issued claims against MasterCard in the Commercial Court. Both claims were for damages for infringements of article 101,

On 24 May 2012, the Commission’s decision was upheld on appeal by the General Court ([2012] 5 CMLR 5) (the “General Court’s decision”). MasterCard appealed the General Court’s decision to the CJEU.

On 30 July 2012, the Commission sent a Supplementary Statement of Objections to Visa Europe concerning its EEA credit MIFs and certain other of its cross-border acquiring (“CBA”) rules.

On 5 October 2012, Argos issued a claim against MasterCard in the Commercial Court. The claim was for damages for infringements of article 101, the 1998 Act, article 53 and section 4 of the Irish Act, in respect of MasterCard’s UK MIFs since 5 October 2006, EEA MIFs since 5 October 2007, and Irish domestic MIFs between 5 October 2006 and 5 January 2007 and since 20 January 2009 (see Popplewell J’s judgment at [26] for an explanation of the claim period relating to the Irish MIFs).

On 19 December 2012, Sainsbury’s issued a claim against MasterCard in the Chancery Division. The claim was for damages for infringements of article 101, the 1998 Act and article 53, in respect of MasterCard’s UK MIFs since 19 December 2006 (or 19 December 2007 in respect of transactions in Scotland).

In July 2013, the Commission published a proposal for a regulation capping interchange fees across Europe. To verify the levels of the proposed caps under the MIT, the Commission conducted (with assistance from Deloitte) a survey of EEA merchants’ costs data. 254 merchants from 10 EEA states responded, and the final survey report (the “Commission Survey”) was published on 18 March 2015.

On 18 December 2013, Sainsbury’s issued its claim against Visa in the Chancery Division (which was subsequently transferred to the Commercial Court by consent). The claim was for damages for infringements of article 101 and the 1998 Act, in respect of Visa’s UK MIFs since 18 December 2007.

On 26 February 2014, the Commission accepted an offer made by Visa that it would amend its CBA rules from 1 January 2015 to allow cross-border acquirers to elect between (i) the local domestic MIF rate or (ii) respective rates of 0.2% and 0.3% for debit and credit card transactions, and that it would cap its EEA credit MIFs at a weighted average of 0.3%.

In September 2014, the CMA announced that, due to the imminent European regulation capping interchange fees, its investigations into MasterCard’s and Visa’s UK MIFs were at an end.

On 11 November 2014, the CJEU dismissed MasterCard’s appeal from the General Court’s decision ([2014] 5 CMLR 23) (the “CJEU’s decision”).

On 1 December 2015, Barling J made an order transferring Sainsbury’s claim against MasterCard from the Chancery Division to the CAT pursuant to section 16 of the Enterprise Act 2002 (the “2002 Act”). His reasons for making the order are contained in his judgment of the previous day ([2015] EWHC 3472 (Ch)) (“Barling J’s transfer judgment”).

On 8 June 2015, the EU’s regulation on interchange fees for card-based payment transactions (Regulation (EU) 2015/751) (the “Interchange Fee Regulation”) came into force. Articles 3 and 4 respectively of the Interchange Fee Regulation set a maximum weighted average rate cap of 0.2% on domestic and cross-border debit MIFs, and a maximum ad valorem rate cap of 0.3% on domestic and cross-border credit MIFs, with effect from 9 December 2015. Member States may impose lower caps for domestic transactions, but the UK has not done so. Ireland has imposed a lower debit card interchange fee of 0.1%. Both MasterCard and Visa have had to reduce their debit and credit UK MIFs to comply with these caps. The CAT took the view at [17(4)(iii)] of its decision that “it was common ground, or at least not contested by Sainsbury’s” that Sainsbury’s could not claim in respect of transactions made after 9 December 2015.

On 14 July 2016, the CAT (Barling J, Professor John Beath OBE and Mr Marcus Smith QC) upheld Sainsbury’s claim against MasterCard, and awarded damages of £68,582,245 (subsequently adjusted to take account of the impact of corporation tax).

On 4 August 2016, MasterCard sought permission to appeal the decision of the CAT on 5 grounds, two of which related to liability and three of which related to quantum. Permission was refused by the CAT in respect of all 5 grounds on 22 November 2016.

On 30 January 2017, Popplewell J dismissed the AAM parties’ claims against MasterCard (viz those issued by Asda, Argos and Morrisons, which had by this time been combined) on the basis that MasterCard’s MIFs did not infringe article 101(1), and in any event would have been exempt under article 101(3). On 16 February 2017, he refused permission to appeal.

On 16 August 2017, Beatson LJ granted the AAM parties permission to appeal Popplewell J’s judgment on most, but not all, of their proposed grounds. On the same day, he granted MasterCard permission to appeal the CAT decision, in respect of all its grounds.

On 30 November 2017, Phillips J dismissed Sainsbury’s Commercial Court claim against Visa on the basis that Visa’s MIFs did not infringe article 101(1). The judge granted Sainsbury’s permission to appeal, and indicated that he would address the article 101(3) issues in a further judgment.

On 15 December 2017, Flaux LJ ordered that the Sainsbury’s appeal against the decision of Phillips J would be heard at the same time as the appeals against the decisions of the CAT and Popplewell J.

On 23 February 2018, Phillips J gave a further judgment in which he held that, had he reached a different view on the article 101(1) question, Visa’s MIFs would not have been exempt (at any level) under article 101(3).

On 8 March 2018, Flaux LJ granted the Commission permission to intervene in the appeals pursuant to article 15(3) of Regulation 1/2003 (the “Modernisation Regulation”). He also allowed the Commission to make oral submissions at the hearing.

After setting out the facts, issues and evidence in some detail, the CAT decided two issues which are no longer in dispute. At [85]-[95], it held that the setting of the UK MIF was a decision or series of decisions by the MasterCard entities as an association of undertakings. This is not appealed by MasterCard, and Visa did not seek to argue the contrary before Phillips J. At [97]-[102], the CAT held that the MIF did not amount to a restriction of competition by object. This is not appealed by Sainsbury’s, and the AAM parties did not pursue a similar line of argument before Popplewell J.

The CAT next turned to the question whether the MIFs amounted to a restriction of competition by effect within the meaning of article 101(1). As for the counterfactual against which their restrictive effects were to be tested, the starting point was a rule that MasterCard transactions would be settled at par, which was equivalent to a default MIF of zero ([141]-[143]). The CAT rejected a submission by Sainsbury’s that Visa should be assumed to have introduced a similar rule: the CAT reasoned that it was the effects of the MasterCard scheme that were being tested, and it would be wrong to make any presumptions regarding Visa that were not grounded in fact

As to what Visa would have “actually” done in the counterfactual world, the CAT thought that it would have maintained its MIFs as close to their then level as it felt it could achieve ([160]-[163]). The CAT considered that for the following reasons, this would not have resulted in issuing banks immediately leaving the MasterCard scheme due to the higher MIFs offered by Visa; rather, they would have sought bilaterally to agree interchange fees with acquirers ([174]-[178]). Although it would have been open to acquirers to refuse to agree anything (resulting in a zero MIF attractive to their merchant customers), they would not have taken this course for fear of (i) a Visa monopoly as issuers switched schemes and (ii) issuers withdrawing valuable benefits or features of the scheme in response to reduced MIFs ([182]-[197]). Instead, they would have secured new charging structures more favourable to particular types of merchants than the traditional ad valorem per transaction basis. Over the claim period these new structures would, on average, have equated to positive interchange fees of 0.50% of transaction value for credit cards and 0.27% for debit cards. Since Visa’s MIFs on debit cards were practically the same (0.26%), issuing banks would not have gradually abandoned MasterCard’s cards in favour of Visa’s cards [238]. Nor would they have drifted to Visa credit cards, despite the apparently higher MIFs on offer (0.80%), because (i) the new charging structures would render the actual difference in MIF levels between the schemes less stark, (ii) the MasterCard scheme is a wellfunctioning one with an established client base, and (iii) issuers do not choose whether or not to participate in schemes solely on the basis of MIF levels. Accordingly, the MasterCard scheme would not have collapsed in the counterfactual world, and acquirers would have been able properly to differentiate their services by competing on price for merchants’ business, resulting in lower prices overall. It followed that the MIFs as set amounted to a restriction of competition by effect ([267]-[269]).

Given its extensive reasoning on the first article 101(1) issue, the CAT was able to deal more briefly with the question of objective necessity. It considered that the question answered itself: the MIFs were “on no view inherently necessary” to the MasterCard scheme, which would “operate perfectly well – indeed, it would be more competitive and better – without the UK MIF” [279].

At [285]-[289], the CAT dealt with the article 101(3) exemption. It held that the MIFs as set did not satisfy any of the four conditions for the following reasons. They inhibited economic progress by frustrating bilateral negotiations between issuers and acquirers, creating upward pressure on merchants’ service charges and preventing the emergence of new charging structures. Accordingly, no benefits resulted and the second condition did not arise. As for the third and fourth conditions, the MIFs were not indispensable to the scheme (but only served to avoid the transaction costs of bilateral agreements), and did enable the parties to eliminate competition. In the light of its finding that interchange fees would be bilaterally agreed in the counterfactual, the CAT did not consider it necessary to decide whether the MIFs would have been exemptible under article 101(3) at any level lower than that at which they were actually set.

At [290]-[419], the CAT considered whether MasterCard could avail itself of the illegality defence as a result of Sainsbury’s Bank plc (“Sainsbury’s Bank”), a legal entity distinct from the claimant, having received MIFs as an issuer participating in the scheme. It held that it could not: there was no turpitude on the part of Sainsbury’s Bank and, even if there was, Sainsbury’s Bank was not part of the same undertaking (within the meaning of article 101(1)) as Sainsbury’s, its conduct was not attributable to Sainsbury’s, and Sainsbury’s Bank did not bear “significant responsibility” for MasterCard’s infringement. Those findings were not appealed by MasterCard and Visa does not advance any arguments on illegality.

Having upheld the claim, the CAT considered the amount of damages due to

Finally, the CAT decided what interest should be applied. Applying a “broad axe”, it held that, had there been no overcharge, 50% of an amount equivalent to the overcharge would have been passed on to Sainsbury’s customers (albeit not in a manner sufficient to reduce the damages due to it) and 50% would have been retained by Sainsbury’s ([509]-[526]). Sainsbury’s would have received interest on this second 50%. For 20% of it, the rate would be that which Sainsbury’s received on its cash balances and, for the remaining 30%, the rate would be that which the company paid on new debt ([527]-[547]).

With respect to the article 101(1) issues, Popplewell J agreed with the CAT that the starting point for the counterfactual was a rule that MasterCard transactions would be

Accordingly, Popplewell J would have found that the MasterCard MIFs infringed article 101(1), were it not for the ‘death spiral argument’. He expressed this argument in the following stages: (i) it is legally permissible for the counterfactual to take into account competition; (ii) the proper assumption in the present case is that Visa’s MIFs would have been the same in the counterfactual as they were in reality; and (iii) this would have led to the collapse of the MasterCard scheme as issuers abandoned it in pursuit of higher MIFs. With respect to the first stage, he held that it is permissible to consider competition, on the basis of CJEU jurisprudence, including [177]-[179] of the CJEU’s decision; the contrary principle stated by the Court of First Instance in Metropole Television (6) and others v Commission [2001] 5 CMLR 33 (“Metropole”) was out of line with that jurisprudence ([164]-[185]). Regarding the second stage, he held that Visa’s MIFs should be assumed to be the same in the counterfactual as they actually were, and not the same as MasterCard’s counterfactual MIFs, unless there was sufficient evidence that the two schemes were “materially identical”, which there was not ([186]-[219]). As for the third stage, he concluded, on the basis of the evidence of MasterCard’s witnesses and of both parties’ experts, that the MasterCard scheme would not have survived in such circumstances ([220]-[236]). Therefore, the MIFs as set did not restrict competition by effect, and were objectively necessary as an ancillary restraint, with the consequence that they did not infringe article 101(1).

Popplewell J then addressed the article 101(3) exemption, even though this was not strictly necessary in the light of his conclusions on article 101(1). Since it was common ground that the fourth condition was met, Popplewell J only had to consider the first three of the four article 101(3) conditions set out above ([262]). He then set out the law to be applied, as follows. The benefits claimed to satisfy the first condition must be causally linked to the MIFs, and such links must be sufficiently direct to be capable of proof ([264]-[265]). The second condition (whether merchants received a fair share of these benefits) would be met if the MIFs as set did not (i) exceed the benefits they produced for merchants, or (ii) generate “unduly high profits” for issuers ([287]). In the circumstances of the case, the third condition (whether the MIFs were indispensable to attainment of the merchant benefits) would be met unless the MIFs created “an unfair degree of profit for issuers”, because there was no realistic counterfactual in which something other than MIFs could confer the relevant benefits ([290]-[291]). Regarding burden of proof, it was for the defendants to prove whether the MIFs as set were exempt, but for the claimants to prove the maximum level of MIF that would have been exemptible under article 101(3), up to which they would not be entitled to damages, drawing an analogy with the requirement for claimants in sale of goods claims to establish as their measure of loss the difference between market value and price paid ([294]-[302]).

Applying these principles to the facts of the case, Popplewell J held that the first condition was met because the MIFs enabled issuers to offer incentives to cardholders, which increased card usage, in turn producing the following benefits for merchants: avoidance of the costs of other payment methods; competitive advantage over merchants who do not take cards (‘business stealing’); facilitation of online spending and e-commerce; guaranteed payment in the case of fraud or default; the avoidance of the cost of providing credit; and increased and earlier spending by customers ([308]-[335]). For his analysis of the second condition, Popplewell J took as a starting point the Commission Survey, which used the MIT to estimate the value to merchants of avoiding the costs of cash payments by accepting cards ([347]). He then adjusted the survey results, based on the parties’ expert evidence, such that they (i) applied to the average merchant, and not just large merchants ([363]-[368]), (ii) included the value of all of the benefits identified above, and not only the avoidance of the cost of cash payments ([369]-[397]), and (iii) reflected the extent to which issuers retained MIFs as profit rather than spending them on cardholder incentives, since the MIT assumed no such retention, and any MIF retained could not possibly contribute to merchant benefits ([398]-[410]). The resulting values exceeded the MIFs as set (except for EEA debit cards for part of the claim period), so that Popplewell J’s threshold requirement of merchant neutrality was passed. He did not consider the issuer profit margins on MIFs (estimated by one of MasterCard’s witnesses at 10% to 40%) to be excessive, and so the fair share condition was met. It followed that the third condition was also met ([409]). Accordingly, had it been necessary for MasterCard to rely on the article 101(3) exemption, Popplewell J would have held that the conditions were fulfilled.

Phillips J began his analysis of the article 101(1) issues in the same way as the CAT and Popplewell J: the starting point for the counterfactual was a rule that Visa transactions would be settled at par, and this was equivalent to a default MIF of zero ([98]-[100]). At [126]-[129] he agreed with Popplewell J that bilateral agreements would not be concluded in the counterfactual because:

“… despite the fact that MIFs have provided a default level of Interchange Fee for many years … bilateral agreements … are unknown in the UK market. That demonstrates the very considerable strength of the market forces which keep the Interchange Fees at the level of the default … In my judgment it would require clear evidence to support a finding that [bilateral agreements] would emerge in a default settlement counterfactual when they do not arise in the actual default Scheme … it is clear that there is no such evidence in these proceedings. On the contrary, the evidence was unanimous and unequivocal to the opposite effect”.

Phillips J then rejected an argument advanced by Sainsbury’s that settlement at par should be regarded as a “fixed and obvious starting point”, resulting in a “competitive process” which is absent where there is a MIF. His main reasons for doing so were expressed at [130]-[137] as follows:

“…there is simply no difference in the competitive process in the two scenarios under consideration in the absence of bilateral agreements. In either case, the market will not deviate from the default settlement rule set by the Scheme notwithstanding that the market participants are free to so …

… there is no a priori reason why settlement … should be at par rather than at a discount (or at a premium) and Interchange Fees are no more or less than another way of expressing such a discount (or a premium if they have a negative value) …

… the effect of the argument is that any level of MIF, on the infinite scale from infinitely positive to infinitely negative … is deemed to be a restriction of competition, all in comparison with an infinitesimally small point on that scale equating to there being no MIF (a figure of zero). But there is, in this context, no magic in the number zero and no reason why it represents an inherently more competitive situation than any other level.”

Next, Phillips J dealt with a submission by Sainsbury’s that he was bound by the CJEU’s decision to hold that Visa’s MIFs restricted competition within the meaning of article 101(1). He rejected that submission on the basis that the CJEU’s decision was based on a finding of fact by the Commission that bilateral agreements would emerge in the counterfactual; the CJEU did not decide that MIFs restricted competition as a matter of law ([138]-[148]).

Phillips J went on to address the question whether the MIFs as set restricted competition in the acquiring market by imposing a floor below which the merchants’ service charge could not fall. It is to be recalled that, had Popplewell J not accepted the ‘death spiral argument’, he would have held on this basis that the MIFs infringed article 101(1). Phillips J’s main reason for disagreeing with Popplewell J is to be found at [156] as follows:

“… the situation is exactly the same at any lower level of MIF, including a zero MIF or its equivalent, a no-MIF/default SAP counterfactual. At that lower level, the default settlement rule still provides a default level of Interchange Fee, and therefore (because of the lack of competitive pressure to depart from that default) both a floor and a ceiling for that fee. The only difference is the level. Popplewell J rejected that argument in the Asda Judgment, stating at §160 that “… in a no MIF counterfactual the alleged vice is not the same as the actual: there is no floor.” However, a zero MIF or no-MIF/default SAP counterfactual most certainly does give rise to a “floor”, both in economic terms and as a matter of logic, particularly in the context of a two-sided market: it prevents the possibility of market forces driving the MIF to a negative level (equivalent to a premium on settling the transaction price). As I have mentioned above, that is not merely a theoretical possibility, as all the expert economists recognised that negative MIFs could and do arise in the real world …”.

Phillips J concluded at [161] that the MIFs as set did not restrict competition within the meaning of article 101(1). Though his analysis and conclusions did not depend on the assumption to be made regarding MasterCard’s counterfactual MIFs, he disagreed with both the CAT and Popplewell J on that issue at [162]-[169]. He thought it difficult to conceive of circumstances in which one scheme would be unable to set any MIFs whilst the other continued to operate unconstrained. More importantly, such an assumption would mean that two unlawful schemes could each escape censure merely by virtue of the existence of the other, which could not be right.

Though not strictly necessary, Phillips J went on to consider the ancillary restraint exemption to article 101(1). In this respect, Visa had relied solely on the ‘death spiral argument’, which the judge had already rejected in the context of whether the MIFs restricted competition. He considered that his reasoning equally applied in the context of ancillary restraint ([179]-[180]). He disagreed with Popplewell J that the CJEU jurisprudence made it permissible to take into account competitors in either context ([181]-[190]). Accordingly, had Phillips J reached a different conclusion on whether the MIFs amounted to a restriction of competition, he would not have regarded the restriction as objectively necessary to the operation of the Visa scheme ([191]).

Phillips J’s second judgment addressed the question of what level of MIFs (if any) would or could have been exempt under article 101(3), had his first judgment reached a different conclusion on the article 101(1) issues. Like Popplewell J, he thought that only the first three of the four article 101(3) conditions applied in the circumstances of the case ([9]). He disagreed, however, with Popplewell J as to which party bore the burden of proving the maximum exemptible level of MIF for damages purposes. He considered that this burden lay on the defendant, and the correct analogy was with contributory negligence rather than sale of goods claims ([13]-[21]).

Turning to the question of whether the MIFs were exempt, Phillips J first considered the standard of proof to be applied. He concluded at [24] that “robust analysis and cogent evidence will be required to establish, on the balance of probabilities, that a restrictive agreement in fact and in the real world (as opposed to in theory) gives rise to pro-competitive effects”. He then summarised Visa’s case on article 101(3), which was, he said, fundamentally the same as that advanced by MasterCard to the Commission: the MIFs were used by issuers to incentivise card usage, the resulting increases in which produced benefits for merchants ([36]-[37]). Except for ‘business stealing’, which was omitted, the claimed benefits were the same as those put by MasterCard to Popplewell J. Phillips J, however, reached a different conclusion at [46]-[50], namely that:

“…there is in my judgment a complete absence of evidence of a real, observable and measurable link between MIFs and actions taken by Issuers to stimulate card usage …

… it is entirely impossible to discern, let alone demonstrate, the alleged increase in card usage arising from such increased stimulation (as opposed to the pre-existing stimulation). Visa has not attempted to prove an increase in usage from any particular increase in stimulation with empirical data … … For the above reasons I conclude that Visa has not established to the requisite standard (or anywhere close) that the UK MIFs contribute to net efficiencies …”.

Visa’s case on article 101(3) thus failed at the first hurdle, so that the MIFs would not have been exempt at any level. Given this conclusion, Phillips J did not attempt to examine the fair share condition, but he did say at [65]-[66] that, had Visa been able to prove the benefits mentioned above, he would have held that the MIFs were indispensable to achieving them.

Although it is not expressly stated in the wording of article 101(1), it is well established in EU law that a provision of an agreement which has the effect of restricting competition does not constitute an infringement if it is objectively necessary for the implementation of the “main operation” of the agreement, provided that the main operation does not itself infringe article 101(1).

A restrictive provision will only be objectively necessary if the main operation would be impossible to carry out in the absence of the restriction. This is clear from the judgment of the CJEU in MasterCard at [91] and [93]:

“91… Contrary to what the appellants claim, the fact that the operation is simply more difficult to implement or even less profitable without the restriction concerned cannot be deemed to give that restriction the 'objective necessity' required in order for it to be classified as ancillary …

93 … The objective necessity test… concerns the question of whether, in the absence of a given restriction of commercial autonomy, a main operation or activity which is not caught by the prohibition laid down in [article 101(1)] and to which that restriction is secondary, is likely not to be implemented or not to proceed.”

The merchants and the Commission submitted that the consideration of objective necessity is a relatively abstract exercise concerned with whether, without the restriction in question, a main operation of the type in question would be impossible to carry out. The test, they said, is not concerned with whether the restriction is necessary for the particular operation in question to compete successfully or be commercially successful. They also said that an analysis of the pro- and anticompetitive effects of the restriction is for article 101(3) and does not form any part of the article 101(1) exercise, including as to ancillary restraint. They submitted that this was clearly established by the decision of the Court of First Instance in Metropole at [107]-[109]:

“107 As regards the objective necessity of a restriction, it must be observed that inasmuch as, as has been shown in paragraph 72 et seq. above, the existence of a rule of reason in Community competition law cannot be upheld, it would be wrong, when classifying ancillary restrictions, to interpret the requirement for objective necessity as implying a need to weigh the pro and anti-competitive effects of an agreement. Such an analysis can take place only in the specific framework of [article 101(3)] of the Treaty.

108

That approach is justified not merely so as to preserve the effectiveness of [article 101(3)] of the Treaty, but also on grounds of consistency. As [article 101(1)] of the Treaty does not require an analysis of the positive and negative effects on competition of a principal restriction, the same finding is necessary with regard to the analysis of accompanying restrictions.

109

Consequently, as the Commission has correctly asserted, examination of the objective necessity of a restriction in relation to the main operation cannot but be relatively abstract. It is not a question of analysing whether, in the light of the competitive situation on the relevant market, the restriction is indispensable to the commercial success of the main operation but of determining whether, in the specific context of the main operation, the restriction is necessary to implement that operation. If, without the restriction, the main operation is difficult or even impossible to implement, the restriction may be regarded as objectively necessary for its implementation.”

The approach of the Court in Metropole was approved and applied by the General Court in MasterCard. At [89]-[90] the General Court said:

“89 As the case-law cited in paragraph 77 above [i.e. Metropole] shows, examination of the objective necessity of a restriction is a relatively abstract exercise. Only those restrictions which are necessary in order for the main operation to be able to function in any event may be regarded as falling within the scope of the theory of ancillary restrictions. Thus, considerations relating to the indispensable nature of the restriction in the light of the competitive situation on the relevant market are not part of an analysis of the ancillary nature of the restriction (see, to that effect, M6 and Others v Commission, cited in paragraph 77 above, paragraph 121).

90 Accordingly, the fact that the absence of the MIF may have adverse consequences for the functioning of the MasterCard system does not, in itself, mean that the MIF must be regarded as being objectively necessary, if it is apparent from an examination of the MasterCard system in its economic and legal context that it is still capable of functioning without it.”

It was submitted on behalf of the schemes, specifically by Mr Mark Hoskins QC, leading counsel for MasterCard, that (i) Metropole is inconsistent with earlier EU law and (ii) was not approved and thus effectively overruled by the CJEU in MasterCard. These were the arguments which were accepted by Popplewell J at [164]-[181] of his judgment, which was a critical aspect of his acceptance of the death spiral counterfactual in relation to the application of the ancillary restraint doctrine.

So far as the earlier EU law is concerned, Mr Hoskins relied upon the decisions of the CJEU in Remia BV & others v Commission (1985) Case 42/84; [1987] 1 CMLR 1 (“Remia”) and Gottrup-Klim Grovvaeforening v Dansk Landbrugs Grovvareselskab AmbA (DLG) (1992) C-250/92 (“Gottrup-Klim”).

The issue in Remia was whether restrictive covenants in agreements transferring businesses were objectively necessary to the main operation, the transfer. The CJEU took account of the risk that purchasers would face competition from vendors. Popplewell J considered that this was an analysis of the specific competitive effects on the purchaser. We agree, however, with the analysis of Phillips J in his first judgment at [187] that the CJEU was considering transfers of business in general and

That the CJEU in Remia was applying an objective and not a subjective test is also clear, as Mr Mark Brealey QC, leading counsel for Sainsbury’s, submitted, from its acceptance of the analysis of the Commission (in the sense of concluding that the Commission had not made any manifest error or reached incorrect findings of fact) referred to in [31] of the judgment: “An agreement which restricts competition cannot escape the prohibition in [article 101(1)] merely because it enables an undertaking to survive.”

The issue in Gottrup-Klim was whether a clause in a cooperative purchasing agreement which precluded members from joining a rival scheme was objectively necessary. We agree with the analysis of Phillips J at [188] of his first judgment that the CJEU considered that issue in relation to co-operative purchasing associations in general, as is clear from the paragraphs in the judgment which he cites, in particular [35]:

“Nevertheless, a provision in the statutes of a co-operative purchasing association, restricting the opportunity for members to join other types of competing co-operatives and thus discouraging them from obtaining supplies elsewhere, may have adverse effects on competition. So, in order to escape the prohibition laid down in [article 101(1)], the restrictions imposed on members by the statutes of co-operative purchasing associations must be limited to what is necessary to ensure that the cooperative functions properly and maintains its contractual power in relation to producers.”

We do not consider that there is anything in the judgment of the Court of First Instance in Metropole that is inconsistent with the decisions of the CJEU in Remia and Gottrup-Klim. So far as the decisions of the European Courts in MasterCard are concerned, we have already noted that the General Court approved the approach of the Court in Metropole. The only difference was that, whereas the Court in Metropole considered that it was sufficient for the main operation to be difficult to operate without the restriction, the General Court considered that, to be objectively necessary, the main operation had to be incapable of functioning without the restriction.

It was that narrow approach of the General Court to the objective necessity test which was the subject of one part of the appeal by MasterCard to the CJEU. As appears from [86] of the CJEU’s decision, MasterCard relied upon [109] of Metropole as a correct statement of the law, contending that in limiting objective necessity to where it

The CJEU said at [91] that the enquiry under the ancillary restraint doctrine is: “whether that operation would be impossible to carry out in the absence of the restriction in question”. The CJEU then rejected at [92] the suggestion that there had been an amalgamation by the General Court of the ancillary restraint exception under article 101(1) with the criterion of indispensability under article 101(3). It set out its reasoning at [93] and [94]:

“93.

In that regard, suffice it to note that those two provisions have different objectives and that the latter criterion relates to the issue whether coordination between undertakings that is liable to have an appreciable adverse impact on the parameters of competition, such as the price, the quantity and quality of the goods or services, which is therefore covered by the prohibition rule laid down in [article 101(1)], can nonetheless, in the context of [article 101(3)], be considered indispensable to the improvement of production or distribution or to the promotion of technical or economic progress, while allowing consumers a fair share of the resulting benefits. By contrast, as is apparent from paragraphs 89 and 90 of the present judgment, the objective necessity test referred to in those paragraphs concerns the question whether, in the absence of a given restriction of commercial autonomy, a main operation or activity which is not caught by the prohibition laid down in [article 101(1)] and to which that restriction is secondary, is likely not to be implemented or not to proceed.

94.

In ruling, in paragraph 89 of the judgment under appeal, that ‘[o]nly those restrictions which are necessary in order for the main operation to be able to function in any event may be regarded as falling within the scope of the theory of ancillary restrictions’, and in concluding, in paragraph 90 of the judgment under appeal, that ‘the fact that the absence of the MIF may have adverse consequences for the functioning of the MasterCard system does not, in itself, mean that the MIF must be regarded as being objectively necessary, if it is apparent from an examination of the MasterCard system in its economic and legal context that it is still capable of functioning without it’. The General Court did not, therefore, err in law.”

It was submitted on behalf of the schemes that, because the CJEU did not in [94] cite the previous sentence of [89] of the General Court’s decision which had stated, by reference to Metropole, that “examination of the objective necessity of a restriction is a relatively abstract exercise”, the CJEU was somehow disavowing or implicitly overruling Metropole. In our judgment that submission is unsustainable. In this part of its judgment the CJEU was citing only those passages of the General Court’s decision which it was necessary to approve in relation to the particular aspect of the appeal with which it was dealing, namely, as we have said, whether it was sufficient for the objective necessity test that without the restriction the main operation would be difficult to operate or whether without the restriction the main operation had to be impossible to operate.

Earlier in its judgment, at [81]-[82], the CJEU had quoted the entirety of paragraphs [89] and [90] of the General Court’s decision (albeit omitting the citation of authority) including the first sentence of [89], later omitted in the citation by the CJEU at [94], without any suggestion there or in any other part of its judgment that the statement of the applicable principles in [89] and [90] of the General Court’s decision, which followed Metropole, did not represent good law. It seems to us, therefore, that far from implicitly disapproving Metropole the CJEU was implicitly approving it, other than in relation to difficulty versus impossibility.

Furthermore, we note that the CJEU was not being invited by MasterCard in its appeal to conclude that Metropole was wrongly decided. On the contrary, as we have said, MasterCard was relying upon Metropole in support of its case that difficulty of operation was sufficient. In those circumstances, it would be surprising if the CJEU had expressly or implicitly disapproved the decision in Metropole. In our judgment, it did not do so. The principle established by Metropole, as approved and modified by the General Court’s decision in MasterCard, correctly states the law. It follows that the ancillary restriction must be essential to the survival of the type of main operation without regard to whether the particular operation in question needs the restriction to compete with other such operations. All questions of the effect of the absence of the restriction on the competitive position of the specific main operation and its commercial success fall outside the ancillary restraint doctrine, as [109] of Metropole makes clear.

Those questions of the competitive effect of the absence of the restriction are to be considered, if at all, under article 101(3). This was made clear by the decision of the General Court in Cartes Bancaires v Commission [2016] EU:T:2016:379 (“Cartes Bancaires”) at [126]-[127]:

“126.

The question of knowing whether the restrictive effects of the measures on the issuing market would be counterbalanced by the alleged restrictive effects on competition on the payment systems market that would occur in their absence stems from the analysis under [article 101(3)]. In this regard, in recital 368 of the contested decision, the Commission deemed that the Group’s argument relating to the indispensability of the measures for the survival of the CB system would be examined within the context of [article 101(3)].

127 Furthermore, it should be noted that, in its previous decision-making practice, i.e. in recital 59 of the Visa 2002 decision, the Commission had considered that Visa’s argument that in the absence of the MIF, the extent of Visa’s activities, and therefore, their competitive impact, would be greatly reduced, would be examined with regard to [article 101(3)] and not to [article 101(1)] for which the question that arose was to determine whether a clause was technically necessary for the functioning of the Visa payment system.”

It follows, in our judgment, that Popplewell J was wrong to conclude that the issue of whether, in the absence of the restriction in question, here the default MIF, the MasterCard scheme would survive in view of the competition from Visa, was one which could be considered under the ancillary restraint doctrine under article 101(1); and he was also wrong to hold that Metropole is contrary to other EU jurisprudence and had been implicitly disapproved by the CJEU’s decision in MasterCard.

There were few substantial differences between the parties in these appeals as to the legal principles applicable to exemption under article 101(3). The real differences were as to how that law had been applied by the two Commercial Court judges and by the CAT (to the extent that the article 101(3) issue was considered by the CAT at all) to the facts of the cases before them.

As we have already mentioned, it is common ground that in order to establish exemption under article 101(3) four cumulative conditions have to be satisfied as set out in the Guidelines, only the first three of which were engaged in this case:

Pursuant to article 2 of the Modernisation Regulation, the burden of proving that these cumulative conditions are satisfied is upon the schemes. Recital 5 to the Regulation makes it clear, however, that the standard of proof is for the national law, so that the usual civil standard of the balance of probabilities applies.

One of the issues which arose before both the Commercial Court judges below was the relationship of that standard of proof to the requirement of EU law, particularly in relation to the first condition, that a claim that a restrictive agreement creates efficiencies must be founded on detailed, robust and compelling analysis and that assumptions and deductions must be based on empirical data and facts and not economic theory alone.

As an analysis of how that requirement sits alongside the standard of proof under English law, we adopt what Phillips J said at [24] of his second judgment:

“In my judgment the distinction being drawn is between (a) real links to real efficiencies, capable of being observed and demonstrated on the facts by evidence (in other words, requiring empirical data), and (b) theoretical or logically assumed links and efficiencies based on broad economic or logical analysis, opinion or anecdotal evidence, perhaps sound in theory but possibly failing to take into account one or more of the many factors which arise in highly complex interactions in the real economy. I see no difficulty in this court determining whether the former has been proved on the balance of probabilities. That test is capable of accommodating varying requirements as to what is expected to meet the standard: contract terms must be “certain”, allegations of fraud must be “distinctly proved” and it is often said that “cogent” evidence is required to rebut certain presumptions. In the case of Article 101(3), it is recognised that robust analysis and cogent evidence will be required to establish, on the balance of probabilities, that a restrictive agreement in fact and in the real world (as opposed to in theory) gives rise to pro-competitive effects.”

We agree with Phillips J (at [25] of that judgment) that this analysis does not differ significantly from that of Popplewell J at [305] of his judgment, but to the extent that there are any differences, we prefer the analysis of Phillips J. In so far as Ms Dinah Rose QC, leading counsel for Visa, sought to argue that Phillips J adopted too prescriptive an approach and that any evidence should suffice provided it meets the civil standard of proof, we do not accept that argument. We consider that Phillips J was right that regard should be had to the requirement of the Commission and the CJEU for cogent and convincing arguments and evidence (see GlaxoSmithKline Services v Commission [2006] ECR II-2969; [2006] 5 CMLR 29 (“GlaxoSmithKline”) at [235], which was applied and followed in the General Court’s decision in MasterCard at [196]).

Although the standard of proof is a matter of English law, the nature of the evidence which will satisfy that standard must be informed by EU law and Commission decisional practice since, ultimately, whether a party is entitled to exemption involves the application of a European treaty. Furthermore, in that context, it is important to maintain a consistency of approach across Member States as to the requirements of article 101(3).

We also reject the suggestion that requiring cogent evidence based on facts and empirical data and analysis rather than economic theory would mean that the standard of proof required of the schemes would be impossible to meet, for reasons we consider in more detail later in the judgment in relation to the specific issues under article 101(3).

We emphasise various principles which emerge from the Guidelines and the European jurisprudence on article 101(3), which are relevant to the present appeals.

First, the relevant benefits for the purposes of the benefits requirement must be causally linked to the relevant restriction, here the default MIF. As Popplewell J correctly recognised in stating the principles at [264] of his judgment, it is not sufficient to identify benefits which result from the use of credit cards or debit cards generally or from the particular MasterCard or Visa scheme generally. This is because it is the restriction of the default MIF which, on this hypothesis, has been found to be a restriction of competition under article 101(1) and has not been shown to be objectively necessary under the ancillary restraint doctrine, and which therefore requires justification to be held exempt under article 101(3). It is any alleged procompetitive effect of the default MIF which falls to be weighed against the anticompetitive restrictive effect: see [207] of the General Court’s decision in MasterCard, which was upheld in the CJEU’s decision at [232].

Secondly, the causal link between the restriction and the relevant benefits must be established by facts and evidence supported by empirical analysis and data and not just economic theory. This is clear not just from the Guidelines but also from [689][690] and [695] of the Commission’s decision:

“689 … it cannot just be assumed, as MasterCard does, without detailed economic and empirical analysis, that a MIF maximises the overall benefits of a system to merchants and cardholders “by reducing costs, increasing services levels and contributing to overall economic welfare”. The mechanism may overburden one side of the scheme with (artificial) costs while not yielding any positive effects on scheme growth and overall efficiency.

690.

Hence, whether a MIF should be paid by acquirers to issuers or vice versa, and whether it should be set at a certain amount or at zero, cannot be determined in a general manner by economic theory alone. A claim that an interchange fee mechanism creates efficiencies within the meaning of article [101(3)] therefore must be founded on a detailed, robust and compelling analysis that relies in its assumptions and deductions on empirical data and facts. Apart from MasterCard’s general assertion that balancing of the demand of cardholders and merchants leads to a better performance of the MasterCard system, is inherent and indispensable to the operation of a four-party payment card system, contributes to overall economic welfare and therefore “undoubtedly” fulfils the first condition of article [101(3)] no such analysis and empirical evidence was provided to the Commission.

…

695… In the context of the first condition it has to be ascertained that the restrictive effects are offset by efficiencies. In this context the undertakings concerned must demonstrate whether a MIF generates the positive effects which the underlying model claims to achieve, here: an increase of system output and possible related efficiencies. To the extent that objective efficiencies cannot be established empirically, they cannot be balanced with the restrictive effects. Some form of convincing empirical evidence on the actual effect of a MIF on the market is therefore required.”

These passages emphasising the need for empirical evidence were expressly approved by Phillips J at [32]-[35] of his second judgment, and we cannot accept the schemes’ challenge to that part of his judgment.

Thirdly, as [54] of the Guidelines makes clear, the causal link must be sufficiently direct to be capable of proof and an indirect effect will not generally be sufficient, precisely because cogent evidence of the link based on empirical analysis and data and not merely economic theory is required. Ms Rose submitted that the requirement in the Guidelines of a direct causal link had not been followed by the General Court in GlaxoSmithKline. In that case GSK had entered a restrictive agreement to restrict parallel trading. In seeking exemption under what is now article 101(3), they argued that the additional profits from the restriction could be invested in research and development. The Commission contended that GSK had failed to show the necessary direct causal link. The relevant passage in the judgment is at [280]:

“…it must be observed that that argument [the need for a direct link], which was raised most recently at the hearing, cannot be accepted. That distinction is not to be found at recitals 155 to 161 to the Decision, to which recital 169 refers, since those recitals unreservedly conclude that there is no link between the General Sales Conditions and the contribution to the promotion of technical progress. Nor is that distinction provided for in [article 101(3)], which allows the exemption of agreements producing a gain in efficiency without distinction as to whether that effect is direct or indirect, and a distinction cannot in principle be drawn where the Treaty draws no distinction (Consten and Grundig v Commission, paragraph 110 above, p.

339). In accordance with the case-law cited at paragraphs 247 and 248 above [which included Consten and Grundig], any advantage in the form of a gain in efficiency must therefore be taken into account, provided that it is objective and appreciable and that its existence is proved convincingly.”

Ms Rose submitted that the CJEU had therefore rejected a submission based upon the Guidelines that a direct causal link was required. We do not read this part of the CJEU’s judgment in that way. The Guidelines at [54] do not say that there must always be a direct causal link, but that it must normally be direct, because indirect effects are normally too remote and uncertain. The Guidelines then give a specific example of an indirect effect in the form of increased profits enabling more investment in research and development, in effect the GlaxoSmithKline case. Whilst it is true that the Guidelines say that such an indirect link is generally not sufficiently direct to be taken into account, they do not exclude that possibility if there is convincing evidence of the link. All the CJEU in that case was saying was that, in effect, an indirect causal link will be sufficient if it is established by convincing evidence. We see no inconsistency between the Guidelines and that decision.

Fourthly, in the context of these specific cases, establishing the requisite causal link involves two critical stages: (i) that the default MIFs in each case incentivise the issuers to take steps they would not otherwise have taken; and (ii) that the steps taken did indeed increase card usage or increase the efficiencies of transactions which would have been card transactions anyway. It was not really in issue at trial that both these stages had to be established: see [310] of Popplewell J’s judgment and [37] of the second judgment of Phillips J, although the AAM parties submitted that Popplewell J had failed to keep in mind the need for both stages to be established by empirical evidence, a matter to which we will return later in this judgment.

In order to satisfy the benefits requirement, a balancing exercise is required, namely that the restriction under consideration “must in particular display appreciable objective advantages [for the relevant consumers] of such a character as to compensate for the disadvantages which [the restriction] entails for competition” (the CJEU’s decision in MasterCard at [234] citing its previous decision in Consten and Grundig v Commission [1966] ECR 299 at [348]). The CJEU rejected an argument by MasterCard that the wider system output of the scheme, in the sense of benefits to society as a whole, should be considered under the first condition. The CJEU held at [237] that in a two-sided system, such as the MasterCard scheme, regard must be had for the purposes of that first condition to the net advantages not only for the consumers on the acquiring market on which the restriction was established, but also for the consumers on the other side of the system, in the issuing market. In other words, for the purposes of the benefits requirement, the court is looking at net advantages to both cardholders and merchants - see the judgment of Popplewell J at [277]-[278]. This was not the subject of appeal and the contrary was not argued by Visa before Phillips J or on appeal.

It follows that, in order to establish the requisite causal link, the schemes have to satisfy the court that, when the balancing exercise is undertaken, the objective advantages of the default MIFs to both cardholders and merchants from increased card usage and efficiencies outweigh the disadvantages of the restriction. In the case of cardholders, the specific “disadvantage” would be the fact that not all MIF income is passed through to them, but rather some is retained by the issuers as profit. In the case of merchants, a specific disadvantage would be the cost to them of default MIFs, which they always have to bear, even on transactions where the cardholder would have used a scheme card anyway, irrespective of the MIF.

The fifth principle is related to the previous one. As [43] of the Guidelines makes clear, for the purposes of the first two conditions, where a restriction affects more than one market, its effect on all such markets must be considered. The effect of the default MIF must, therefore, be considered in both the issuing market (as regards cardholders) and the acquiring market (as regards merchants). It is also made clear in [43], however, that where overall there are negative effects on consumers in one market, those cannot be balanced against and compensated by positive effects on consumers in another market, unless the group of consumers in each market is substantially the same, which is not the case here.

This point was made clearly at [242] of the CJEU’s decision in MasterCard:

“However, as is recalled in paragraph 234 of the present judgment, examination of the first condition laid down in [article 101(3)] raises the question whether the advantages derived from the measure at issue are of such a character as to compensate for the disadvantages resulting therefrom. Thus, where, as in the present case, restrictive effects have been found on only one market of a two-sided system, the advantages flowing from the restrictive measure on a separate but connected market also associated with that system cannot, in themselves, be of such a character as to compensate for the disadvantages resulting from that measure in the absence of any proof of the existence of appreciable objective advantages attributable to that measure in the relevant market, in particular, as is apparent from paragraphs 21 and 168 to 180 of the judgment under appeal, where the consumers on those markets are not substantially the same.”

It seems to us that the same point was being made by the CAT at [289(2)] of its judgment, cited with approval by Popplewell J at [271] of his judgment.

The sixth principle also follows on from that point and concerns the correct interpretation of the second condition, the fair share requirement. It was determined by the European Courts in MasterCard that this involves consideration not just of whether there are net benefits to the consumers as a whole (merchants and cardholders) under the first condition, but also whether there were net benefits to the merchants, being the consumers who are affected by the restriction of default MIFs.

At [281] of his judgment, Popplewell J held that a fair share for the merchants must not leave them worse off as a result of the restriction in question, so that, unless they obtain greater benefits from the default MIF than the anti-competitive disadvantage it imposes upon them, the second condition will not be satisfied. That analysis was not the subject of any Respondent’s Notice from MasterCard and, during the course of argument, Mr Hoskins said that it was agreed.

In the Visa case, however, Phillips J accepted the argument of Visa, based upon its analysis of the relevant section of the CJEU’s decision in MasterCard at [240]-[243] and [247] that, for the purposes of the fair share requirement, the consumers as a whole (both cardholders and merchants) must be considered and: “benefits accruing to cardholders can therefore be taken into account in determining whether benefits at least equal the disadvantage of the MIF. There must, however, be at least some objective advantages for Merchants, even if less than the burden they suffer” ([62] of the second judgment). On that interpretation of the second condition, even if merchants are worse off overall, the second condition can be satisfied, provided they

Sainsbury’s appeals that conclusion. Ms Rose, for Visa, maintains that Phillips J’s interpretation of the CJEU’s decision is correct. Accordingly, it is necessary for us to consider in some detail what the CJEU decided. At [240] the Court is clearly considering the first condition, as it reiterates the point already made at [234] that it was necessary for the purposes of that condition “to take into account all the objective advantages flowing from the MIF, not only on the relevant market, namely the acquiring market, but also on the separate but connected issuing market”, in other words advantages to both merchants and cardholders.

Paragraphs [241]-[242] are also dealing with the first condition and the balancing exercise to which we have referred, as is clear from the opening sentence of [242] which we quoted above. At [243] the CJEU went on to approve the approach of the General Court at [226] of its judgment, which had essentially applied the principle enunciated in [242] (which in turn, as we have said, reflects [43] of the Guidelines), in concluding that since there was no proof of the existence of objective advantages flowing from the MIFs enjoyed by merchants, it was not necessary to examine the advantages flowing from the MIFs for cardholders, since they could not by themselves compensate for the disadvantages resulting from the MIFs (clearly a reference back to the disadvantages to the merchants).

At the end of [243] and in the next two paragraphs [244]-[245], the CJEU rejected the various criticisms levelled by MasterCard at [229] of the General Court’s decision. It is important to read that paragraph in its context, by reference to the preceding paragraphs [226]-[228], which the CJEU was clearly also approving and read as follows:

“226 It must be concluded therefore that, in the absence of proof of a sufficiently close link between the MIF and the objective advantages enjoyed by merchants, the fact that the MIF may contribute to the increase in MasterCard system output is not, in itself, capable of establishing that the first condition laid down under [article 101(3)] is satisfied.

227

The applicants also criticise the Commission for failing to take into account the advantages to cardholders that arise from the MIF and, moreover, for acting as a ‘price regulator’ in respect of the MIF.

228

With regard to the first criticism, it is indeed settled case-law that the appreciable objective advantages to which the first condition of [article 101(3)] relates may arise not only for the relevant market but also for every other market on which the agreement in question might have beneficial effects, and even, in a more general sense, for any service the quality or efficiency of which might be improved by the existence of that agreement (Case T-86/95 Compagnie générale maritime and Others v Commission

[2002] ECR II-1011, paragraph 343, and GlaxoSmithKline Services v Commission, cited in paragraph 196 above, paragraph 248). However, as merchants constitute one of the two groups of users affected by payment cards, the very existence of the second condition of [article 101(3)] necessarily means that the existence of appreciable objective advantages attributable to the MIF must also be established in regard to them.

229 Therefore, in the absence of such proof, the applicants’ criticism that insufficient account was taken of the advantages of the MIF for cardholders is, in all events, ineffective.”

At [247] of its judgment the CJEU then dealt with the criticism made by MasterCard (to which the CJEU had referred at [223]):

“As regards the appellants’ argument that the General Court did not explain why the first two conditions in [article 101(3)] could not be satisfied on the basis only of the advantages the MIF produce for cardholders, it is sufficient to refer to paragraphs 240 to 245 of the present judgment.”

In our judgment, although expressed in somewhat oblique terms, this can only sensibly be interpreted as the CJEU saying that the position as regards the second condition is the same as regards the first condition, namely that where the restriction affects two markets, if the restriction causes disadvantages overall to the consumers in the market under consideration (here the merchants in the acquiring market), those disadvantages cannot be compensated by advantages to consumers in the other market (here the cardholders in the issuing market), unless the two groups of consumers are substantially the same, which they are not in this case.

Ms Rose’s contrary argument was that, at [241] (and thus [247]), the CJEU was approving the proposition for both the first two conditions that, provided there were some advantages to the merchants (even if overall they were disadvantaged by the restriction), the advantages to the cardholders in the other market could be taken into account to arrive at an overall net benefit position. This seems to us to ignore [242][245] which follow and the context of this whole part of the CJEU’s decision, namely as we have said, approval of the analysis at [226]-[229] of the General Court’s decision.

There are not “appreciable objective advantages” to the merchants in the acquiring market unless the advantages caused to them by the MIF outweigh the disadvantages. Only then can it be said that they have received a “fair share” of the benefits. That is the sensible interpretation of both the General Court’s decision and the CJEU’s decision, and it accords with what is stated in [43] of the Guidelines (to which Phillips J does not refer):

“Moreover, the condition that consumers must receive a fair share of the benefits implies in general that efficiencies generated by the restrictive agreement within a relevant market must be sufficient to outweigh the anticompetitive effects produced by the agreement within that same relevant market.”

The footnote to that passage confirms that the test under the second condition is whether the consumers in the relevant market, here the merchants, have received advantages from the restriction which outweigh the disadvantages. It states:

“The test is market specific, see to that effect Case T-1 31/99, Shaw, [2002] ECR II-2023, paragraph 163, where the Court of First Instance held that the assessment under [article 101(3)] had to be made within the same analytical framework as that used for assessing the restrictive effects, and Case C360/92 P, Publishers Association, [1995] ECR I-23, paragraph 29, where in a case where the relevant market was wider than national the Court of Justice held that in the application of [article 101(3)] it was not correct only to consider the effects on the national territory.”

We consider, therefore, that nothing in the judgments of the European Courts in MasterCard alters the established position under the fair share requirement, that the consumers in the specific market, here the merchants in the acquiring market, will only receive a fair share of the benefits if the advantages to them caused by the restriction outweigh the disadvantages, so that, as Popplewell J said, they are no worse off. On this point we consider that his analysis was correct and that of Phillips J was wrong.

The seventh principle relates to the indispensability requirement in the third condition. The party seeking exemption has to prove that the restriction in question, here the default MIF, was indispensable to the attainment of the relevant benefits or efficiencies. This condition only arises if the first two conditions are satisfied. Both the Commercial Court judges below proceeded on the basis that, if the first two conditions were satisfied in relation to the MIFs, then the MIFs were indispensable by definition. It seems to us that that approach ignores that a restriction will only be indispensable if there are no other less restrictive means of achieving the same benefits or efficiencies (see [75] of the Guidelines). In the context of the MIFs it follows that the schemes have to prove that the particular level of MIF for which they contend was indispensable to achieving the relevant benefits or efficiencies.

Finally in relation to the Guidelines, although, as was submitted on behalf of the schemes, they are not legally binding and therefore some flexibility in whether they should be applied and followed is permissible, we consider that, as with the nature of evidence required to satisfy the first condition, consistency of approach across Member States is important. We note that it was not suggested by either Ms Rose or Mr Hoskins that national courts in other member states have departed from the Guidelines in considering the issue of exemption under article 101(3).

It was argued on behalf of the schemes that the approach taken in the Commission’s decision in MasterCard was out-of-date because in its commitment decisions the Commission and, in the Interchange Fee Regulation, the Commission and the European Parliament, have adopted the MIT as sufficiently accurate and robust to assess the level of exemption under article 101(3).

This overlooks, however, that, whilst it is correct that the Commission and the European Parliament have accepted the MIT for the purpose of setting an upper limit or cap under the Interchange Fee Regulation (which is how Rochet and Tirole themselves viewed the function of the MIT), the Commission’s consistent position has been that adoption of the MIT alone will not lead to automatic exemption. That is made clear by recital 14 to the Interchange Fee Regulation:

“The application of this Regulation should be without prejudice to the application of Union and national competition rules. It should not prevent Member States from maintaining or introducing lower caps or measures of equivalent object or effect through national legislation.”

As Ms Ronit Kreisberger, for the Commission, explained in her submissions to us, the Commission regards the MIT as a useful starting point but not as a substitute for the facts of the case. As she put it, the MIT is not a “silver bullet” for the schemes. In other words, to obtain exemption, a scheme still has to back up any reliance on the MIT as a benchmark with robust analysis and cogent evidence.

By the end of the oral argument, it was common ground between the parties that the test under article 101(1) was whether there is a “likelihood” that the agreement in question restricts competition. The debate that led to that common ground had relied on varying statements from European institutions referring to the need to show that an agreement has either the likelihood, capability or the potential of restricting competition (see, for example, John Deere v Commission [1998] 5 CMLR 311 at [72]-[79]). The Guidelines at [16] say that “[a]greements between undertakings are caught by the prohibition rule of [article 101(1)] when they are likely to have an appreciable adverse impact on the parameters of competition on the market”. It may be, as Ms Rose submitted, that the reason is that one needs to be able to establish whether an agreement is caught by article 101(1) at its inception and before its actual effects can be considered. We will proceed on the basis that what is required is that it be likely that the agreement in question restricts competition.

We shall deal with this issue by describing first the parties’ arguments, then dealing with those arguments. We explain in some detail the significance of each of the Commission’s decision, the General Court’s decision and the CJEU’s decision in MasterCard, because those decisions have formed the basis of many of the central disagreements between both the parties before us and the decisions under appeal. We end our analysis of this issue by dealing with each of the decisions under appeal and summarising our overall conclusions on whether the schemes’ rules restrict competition.

Ms Rose submitted that, properly characterised, MIFs are not a price for a service. They are a transfer of value from one side of a two-sided market to the other, in the form of either a discount on settlement at par, which is a positive MIF, or a premium over settlement at par, which is a negative MIF. Once this is understood, it becomes clear that a zero MIF is legally and economically equivalent to a settlement at par rule, and that a positive or negative MIF does not fix prices any more than does a rule requiring settlement at par. Since it is common ground that some form of default settlement rule is necessary for the schemes to function, the appropriate counterfactual must include some form of collusive agreement. This is what, submitted Ms Rose, distinguished this case from a typical cartel case, where the counterfactual necessarily involves the participants setting their prices independently. Accordingly, in this case, the competitive process will not differ in the counterfactual, because (i) it is common ground that a default settlement rule must be set, (ii) it is common ground that issuers and acquirers will not enter into bilateral agreements so as to depart from the default settlement rule, due to the ‘free rider problem’ (no acquirer will agree to pay a MIF unless it knows its competitors are paying the same), (iii) the MIF is a transparent common cost, which is passed on by acquirers to merchants, and does not figure in the negotiations between them, and (iv) acquirers nonetheless compete for merchants’ business in relation to the acquirer’s margin and the additional services they offer.

Ms Rose pointed to the unchallenged factual evidence of Mr Ashworth of WorldPay, the biggest UK acquirer, to the effect that no difference in competitive dynamics resulted from the merchants being able to negotiate 100% of the merchants’ service charge as they could in the counterfactual, as against their being able, in reality, to negotiate only some 10% of it. They are negotiating over the same sterling amount (the acquirer’s margin), and the nature of the negotiation is the same. According to Ms Rose, the point was also accepted by AAM’s expert economist, Mr Dryden, as Phillips J acknowledged at [159], and is determinative of this case on the evidence.

Ms Rose submitted that the claimants’ arguments prove too much. They would apply equally to scheme fees and allocation of the costs of fraud, both of which (i) are also decisions by associations of undertakings, (ii) also allocate costs in a two-sided market, and (iii) impact on what is charged to customers. Further, if the claimants were right, the MIF would be unlawful at any level, because it would at any level inflate or set a floor under the merchants’ service charge (including at zero, when compared with a negative MIF). This would be remarkable when the Interchange Fee Regulation allows MIFs up to a certain level.

Ms Rose relied on the death spiral argument in relation to this issue as well as the ancillary restraint issue. She submitted that the merchants’ service charge would go up if Visa were taken out of the market as a result of its competitors charging MIFs when it could not in the counterfactual (because Visa’s MIFs were generally lower than those of its competitors).

Finally, Ms Rose submitted that the court was not bound by the CJEU’s decision, because it depended on a finding of fact as to the likelihood of acquirers agreeing bilateral interchange fees. The Commission found at [460], based on the statements of retailers, that (i) the default settlement at par counterfactual would lead to a period of bilateral negotiation; (ii) these negotiations would create uncertainty amongst acquirers as to what other acquirers were paying; (iii) this uncertainty would enable merchants to exert greater pressure on acquirers when negotiating the merchants’ service charge; and (iv) in the long term this dynamic would drive MIFs down to zero. These findings of fact underpinned the decisions of both the General Court at [134] and [143] and the CJEU at [195] (which only sets out and approves the final sentence of [143] of the General Court’s judgment, and not the preceding sentence). By contrast, the unanimous expert evidence before Phillips J (and Popplewell J) was that there would be no bilateral agreements between issuers and acquirers in the counterfactual (see [111]-[129] of Phillips J’s first judgment and [141]-[150] of Popplewell J’s judgment). Phillips J, therefore, made a finding of fact that there would be no change in competitive dynamics in the counterfactual, which turned on the expert evidence about the impossibility of bilateral interchange fees, and which the court could not and should not disturb. To do so would be inconsistent with the House of Lords’ decision in Crehan v Inntrepreneur Pub Co (CPC) and another (Office of Fair Trading and others intervening) [2006] UKHL 38 (“Crehan”).

Mr Hoskins, for MasterCard, adopted Visa’s submissions on this issue in so far as they were relevant to MasterCard. He further submitted that the CAT’s findings about bilateral interchange fees should be quashed because there was no evidence to support them. There was clear authority in the O2 Germany case (Case T–328/03) [2006] 5 CMLR 5 at [68]-[71], and also in the CJEU’s decision in MasterCard at [169], that the correct counterfactual was a question of fact, because mere theory is not enough, and the court must consider whether the counterfactual would be likely and realistic in the actual context. This is also supported by the General Court’s decision in Cartes Bancaires at [108]. Moreover, even if the CJEU’s decision in MasterCard did not turn on a factual finding about bilateral agreements, the Commission plainly relied on evidence to reach its conclusion that there would be more competition without the MIF due to merchants being able to exert greater pressure on acquirers. The Commission referred to statements of retailers at the start of [460] and in footnote 516. The references to “merchant pressure” at [143] of the General Court’s decision and [195] of the CJEU’s decision can only be referring to [460] of the Commission’s decision, because merchant pressure is not mentioned anywhere else. The decisions were therefore based on evidence. The evidence before Phillips J was the opposite, namely that there would be no greater competition in the absence of the MIF. The court was therefore squarely in Crehan territory.

By contrast, Mr Jon Turner QC, leading counsel for the AAM parties, submitted that, even though charging higher prices alone because of the MIF did not engage article 101(1), charging higher prices to customers because of an agreement to impose uniformly agreed charges on them, certainly did (see [64] of the Advocate General’s opinion in MasterCard). A default rule providing for any MIF, whether positive or negative, was fundamentally different from a default settlement at par rule. The former is a collusive agreement to impose uniformly agreed charges on one side of a two-sided market, whereas the latter is not, because it imposes no charge on either side. This was the special significance of zero (see [453] of the Commission’s decision where a similar point is made).

There is greater competition in a counterfactual where there is a default settlement at par rule, even in the absence of any bilaterally agreed MIFs. Even if acquirers are competing for merchants’ business in relation to the same sterling amount of the merchants’ service charge as in the real world (i.e. the acquirer’s margin), they are competing on the entirety of the merchants’ service charge in the counterfactual as opposed to only 10% of it in the real world (see Krupp Thyssen Stainless GmbH

Mr Brealey for Sainsbury’s submitted that the whole purpose of the counterfactual exercise is to ask whether, in the absence of the measures in question, there would be greater competition (Cartes Bancaires at [111]). In the present case, the measures in question are the collective agreement to impose interchange fees. These measures cannot therefore be present in the counterfactual (see the CJEU’s decision at [161] and [172]).

The merchants submitted that the CJEU’s decision could not be distinguished on the basis of any findings of fact that there would have been bilateral negotiations in the counterfactual. This is not what the Commission was saying at [460]. It was merely saying that the realistic outcome in the counterfactual was no interchange fees, as demonstrated by footnote 517. It is irrelevant whether the “no interchange fee” outcome occurs immediately or after an interregnum. The conclusions of the General Court at [143] and the CJEU at [195] were conclusions of law on materially indistinguishable facts, by which this court is bound. Crehan has no application.

Mr Turner submitted that Popplewell J’s judgment at [161] explained that Mr Dryden had not accepted that, absent a MIF, there was no difference in competition in the acquiring market. Moreover, Ms Rose was wrong to say that, if the MIFs were unlawful so were scheme fees and fraud protection rules, because scheme fees might well pass the objective necessity or article 101(3) tests, and fraud protection costs are not passed on to cardholders, as a common cost, by all issuers. Issuers compete for cardholders’ business, including by consideration of how efficiently they manage fraud.

The Commission supported the merchants’ arguments submitting that the European Courts have consistently found that rules providing for default MIFs in such payment card schemes harm competition in the acquiring market by impeding the ability of merchants to negotiate the fees charged by acquirers below the threshold imposed by the MIF. The MIF is a collective device which exploits the merchant’s dependence on payment cards. Under EU law, article 101(1) is infringed in circumstances where the MIF gives rise to a price floor in the acquiring market below which the fees charged to merchants cannot be driven (see the CJEU’s rejection of MasterCard’s argument at [195], upholding [143] of the General Court’s decision).

Phillips J misconstrued the CJEU’s decision, which did not depend on a determination of fact by the Commission that, in the absence of MIFs, there would be a “highly competitive process” between issuing and acquiring banks in the form of bilateral negotiations which amounted to “actual competition”. The CJEU’s decision at [195] expressly referred to the effect of the MIF being to limit the commercial pressure which merchants were able to exert on acquiring banks. That was a restriction of the competitive process on the acquiring market. The restriction in question was the impediment to the merchants’ ability to drive down prices charged by acquirers, due to the setting of the price floor, not the absence of bilateral negotiations. Since that impediment does not arise in a payment card scheme providing for settlement at par, in which competitive forces can operate unfettered, a positive default MIF is necessarily restrictive compared to a zero MIF counterfactual.

The Commission’s approach to the nature of the anti-competitive restriction in the acquiring market in both the 2002 Visa and 2007 MasterCard decisions, should, as a matter of principle and logic, apply to the analysis of the effects of MIFs in cases where the relevant counterfactual is a zero MIF, irrespective of whether bilateral negotiations between banks are also considered likely or not. In the words of the General Court at [143], such harm “necessarily” follows where a positive MIF is compared with a zero MIF.

The General Court said at [111] in Cartes Bancaires that:

“… the analysis of the competitive situation in the absence of the measures in question aims to determine whether the measures restrict the competition that would have existed in their absence. This concerns, in particular, determining whether, in the absence of the measures in question, the competitive situation would have been different on the relevant market, that is to say whether the restrictions on competition would or would not have occurred on this market.”

In our judgment, the schemes’ arguments as to the correct counterfactual ignore these fundamental propositions. The “measures in question” in this case are the agreements between the issuers and the acquirers to be bound by the scheme rules set by the

It is true, as Ms Rose argued, that there has to be a rule as to settlement, but it is not true that such a rule has to include a MIF, negative or positive. The magic of zero is that the “measures in question”, namely the agreements to impose default interchange fees, are absent.

It is, therefore, necessary to ask whether, in a world without the scheme rules that set a MIF in default of bilateral interchange fees being agreed, there would or would not be more competition in the acquiring market.

The first question then is whether this court is in fact bound to follow the CJEU’s decision in MasterCard. The domestic court is obviously required to do so if the decision is one of law, since the EU competition rules have direct effect equally in all Member States, but is not necessarily required to do so if the decision was one made on the facts that were found in that case by the Commission. In this regard, when interpreting and applying our national competition law as opposed to EU competition law, we must have regard to the provisions of section 60(1) of the 1998 Act which provide that its purpose is to “ensure … so far as is possible [that] questions arising under this Part in relation to competition within the United Kingdom are dealt with in a manner which is consistent with the treatment of corresponding questions arising in [EU] law in relation to competition within the [EU]”. Moreover, section 60(2)(b) provides that this court must “act … with a view to securing that there is no inconsistency between … the principles applied, and decision reached, by [this] court … and … the principles laid down by the [TFEU] and the European Court, and any relevant decision of that Court, as applicable at that time in determining any corresponding question arising in Community law” and must “in addition, have regard to any relevant decision or statement of the Commission”.

Crehan does not alter what we have already said. The question in Crehan was whether the English court was bound by factual findings made by the Commission, not whether the English court would have been bound by legal questions decided by the CJEU, which it obviously was. That much is clear from Lord Bingham’s speech at [7], where he said that the question in that case was whether Park J, the trial judge, should “have treated the Commission’s factual assessment of the United Kingdom beer market in its Whitbread, Bass and Scottish and Newcastle decisions as effectively binding upon him”. At [11], Lord Bingham summarised the position as follows:

“[EU] law prohibits the making by national courts of decisions which contradict decisions of [EU] institutions on the same subject matter between the same parties, and strongly discourages the making by national courts of decisions which may be inconsistent with decisions which may yet be made by [EU] institutions on the same subject matter between the same parties. But it does not, as the analysis of the relevant authorities by … Lord Hoffmann, shows, go to the length of requiring national courts to accept the factual basis of a decision reached by [an EU] institution when considering an issue arising between different parties in respect of a different subject matter.”

It is, therefore, necessary to consider what precisely the Commission, the General Court and the CJEU decided in MasterCard.

Looking at the Commission’s decision as a whole, it can readily be seen that the Commission was dealing with the same factual situation as in these cases in relation to both Visa and MasterCard: a default MIF set by the scheme in the absence of any bilateral interchange fees being agreed between issuers and acquirers. The Commission’s conclusion was broadly the same as that agreed before both Popplewell and Phillips JJ, namely that, in the counterfactual situation in the absence of the challenged restriction, issuers and acquirers would ultimately not agree bilateral interchange fees so that the situation would revert to settlement at par, with negotiations between merchants and acquirers being undertaken as to the merchants’ service charge, absent the MIFs.

The Commission considered the article 101(1) issue at section 7 of its decision starting at [330]. It turned to deal with restriction of competition by effect at paragraph 7.2.2 of the decision starting at [408], having concluded at [407] that “given that it can be clearly established that the MasterCard MIF has the effect of appreciably restricting and distorting competition to the detriment of merchants in the acquiring markets it is not necessary to reach a definite conclusion as to whether the MasterCard MIF is a restriction by object”.

The Commission stated its conclusion at [410] that:

“MasterCard’s MIF constitutes a restriction of price competition in the acquiring markets. In the absence of a bilateral agreement, the multilateral “default” rule fixes the level of the interchange fee rate for all acquiring banks alike, thereby inflating the base on which acquiring banks set charges to merchants. Prices set by acquiring banks would be lower in the absence of this rule and in the presence of a rule that prohibits ex post pricing. The MasterCard MIF therefore creates an artificial cost base that is common for all acquirers and the merchant fee will typically reflect the costs of the MIF. This leads to a restriction of price competition between acquiring banks to the detriment of merchants (and subsequent purchasers).”

The reference to an absence of “a bilateral agreement” is to describe the nature of the rule which provides for a MIF to be the default, absent a bilateral agreement.

The Commission then described the two quantitative analyses it had undertaken at [425] to “see whether and to what extent the Intra-EEA [fall-back] interchange fees set a floor under the merchant fees”.

The Commission explained the “decisive question” at [448] of its decision by saying first that “the purpose of the second quantitative analysis was to assess the differential between merchant fees paid by larger and small merchants to assess the extent to which larger ones are in a position to negotiate [a merchants’ service charge] below the MIF”. It continued:

“[t]he decisive question is whether in the absence of the MIF the prices acquirers charge to merchants at large would be lower. This is the case, because the price each individual bank could charge to merchants would be fully determined by competition rather than to a large extent by a collective decision among (or on behalf of) the banks.”

Under the heading “7.3.2.1.5 Commission Assessment of MasterCard’s arguments why its MIF would not restrict competition between acquiring banks” starting at [439], the Commission dealt with MasterCard’s various arguments, turning at (f) to the argument that the MIF was not a restriction because its effect would be like “excise tax”. At [456], the Commission said that “factually, MasterCard neglect[ed] that a MIF does not have neutral effects on all acquirers but that it may well disadvantage certain acquirers to the benefit of others”. The Commission then disagreed with the argument by explaining at [458] that MasterCard’s approach would entirely deprive article 101(1) of its effet utile. The default MIF “not only creates an (artificial) common cost for acquirers and thereby sets a floor for the fees each acquirer charges to merchants”, but “[a]cquirers also know precisely that all of their competitors pay the very same fees”, which eliminated an element of uncertainty for all suppliers involved. The Commission said that, in the absence of the MIF, the merchants’ service charge would be set taking into account only “the acquirer’s individual marginal cost and his mark up”: see the last sentences of both [459] and [460].

It is in this context that [460] of the Commission’s decision needs to be understood. It was explaining why the MIFs were not like an excise tax, but actually restricted competition between acquirers and forced up prices for merchants. It referred to “statements of retailers demonstrat[ing] that they would be in a position to exert that pressure if acquirers were not able to refer to the interchange fee as the “starting point” (that is to say, as the floor) for negotiating the [merchants’ service charge]”. The Commission explained that “without a default that fixes an interchange fee rate in the absence of a bilateral agreement, merchants could shop around to contract with the acquirer who incurs the lowest interchange costs”. It then explained why, even if there were some bilateral agreements for a time, the process in the counterfactual would end up without a MIF at all and with settlement at par. Footnote 517 to [460] makes clear that it was the Commission’s view that “in the absence of a default MIF banks may or may not enter into bilateral agreements on interchange fees”. The Commission explained its view at [522] that “[i]n the presence of a MIF the marginal cost of acquirers are inflated, thereby setting a floor under the merchant fee”.

The General Court considered whether the default MIFs were a restriction on competition in the section of its decision generally entitled “Law” starting at [60]. Within that section, the relevant parts of the decision appear under the headings “b) The part of the plea relating to errors of assessment in the analysis of the effects of the MIF on competition” starting at [123] and “The complaints relating to the assessment of competition in the absence of the MIF” starting at [129].

It is clear from [129] that exactly the same points were put to the General Court as have been put to us by Ms Rose. The General Court referred to the applicants saying “that the MasterCard system could not function without a default transaction settlement procedure” and that “the Commission wrongly concluded that, in the absence of the MIF, bilateral negotiations would be held between issuing banks and acquiring banks and that such negotiations would in due course lead to the disappearance of interchange fees”.

These points were rejected at [131] on two grounds. First, at [132], the General Court said that, for the reasons it gave in [94]-[120], a MasterCard system operating without a MIF - solely on the basis of a rule prohibiting ex post pricing - was economically viable, and that was sufficient to justify it “being taken into consideration in the context of the analysis of the effects of the MIF on competition”. This was the aspect of the General Court’s decision that was later said to be wrong by the CJEU (see [169] and [198] of the CJEU’s decision), but it demonstrates that the General Court was approving the “no MIF” plus prohibition of ex post pricing counterfactual, which the CJEU also later approved for slightly different reasons (see [173]-[174] of the CJEU’s decision). The “no MIF” plus prohibition of ex post pricing counterfactual is not materially different from the no default MIF plus settlement at par counterfactual that the parties are agreed upon in this case. Both admit the possibility of bilateral interchange fees, but assume that in default there will be no imposed standard MIF and also settlement at par.

Secondly, in relation to what the Commission had said about negotiating bilateral interchange fees, the General Court held at [133]-[134] that the Commission had not been manifestly incorrect to refer to them at [460]:

“essentially in order to point out that in a MasterCard system operating without a MIF acquirers accepting interchange fees on a bilateral basis would risk failing to remain competitive in the acquiring market, and that, therefore, in the absence of a MIF, it was to be expected that interchange fees would in due course cease to be charged on the settlement of transactions”.

It is clear from [142] that the General Court was dealing with the same arguments as were addressed to us. It recorded that MasterCard had submitted that the fact that the MIF had an impact on the level of the merchants’ service charge did not affect competition between acquirers, because the MIF applied in the same way to all acquirers, and operated as a cost that was common to all of them. MasterCard argued that the prohibition of ex post pricing effectively imposed a MIF set at zero which “from a competitive aspect, would be equivalent to and just as transparent as the current MIF, the only difference being the level at which it is set”.

Again, the General Court rejected this line of argument at [143] where it effectively repeated that, since the Commission had been legitimately entitled to find that “a MasterCard system operating without a MIF would remain economically viable”, it necessarily followed that “the MIF has effects restrictive of competition”. The error that the CJEU found in this latter passage did not impinge on its legal determination of the appropriate counterfactual.

The General Court went on in [143] to explain that “[b]y comparison with an acquiring market operating without them, the MIF limits the pressure which merchants can exert on acquiring banks when negotiating the [merchants’ service charge] by reducing the possibility of prices dropping below a certain threshold”. This reasoning seems to us to be a conclusion of law based on the factual premises stated, which it may be noted are precisely the same factual premises as are agreed in this case.

In analysing the CJEU decision, it is first necessary to understand the error of law that it identified in the General Court’s decision. The CJEU’s main reasoning about the counterfactual used for the article 101(1) analysis began at [161]. There, the CJEU recorded the argument that, in assessing a restrictive effect on competition, the Commission should have considered what the actual counterfactual hypothesis would have been in the absence of the MIF. It referred to a line of established authority to the effect that counterfactual competition should be assessed within the actual context in which it would occur in the absence of the agreement in dispute (see also [164][166]).

The CJEU then explained at [162] that the General Court had relied on the same counterfactual as it had used for the “ancillary restraint” analysis (to which we shall come in due course), namely the premise of “a MasterCard system operating without a MIF - solely on the basis of a rule prohibiting ex post pricing”. This is important because the CJEU later expressly approved that counterfactual for article 101(1) purposes at [173]-[174]. It did, however, point out at [163] that it was not always appropriate to use the same counterfactual hypothesis for both purposes.

The CJEU’s criticism of the General Court at [167] was that it had not “in any way [addressed] the likelihood, or even plausibility, of the prohibition of ex post pricing if there were no MIF” in the context of its analysis of the restrictive effects of the MIFs. That was why the CJEU concluded at [169] that the General Court should not at [132] and [143] of its decision have relied on “the single criterion of economic viability” to justify the inclusion of a rule prohibiting ex post pricing as part of its counterfactual. The General Court, said the CJEU, had nowhere explained “whether it was likely that such a prohibition would occur in the absence of MIF”. That was an error of law, but one that the CJEU thought at [170] would not allow the decision to be quashed if “its operative part [was] shown to be well founded on other legal grounds”.

The CJEU held at [171]-[173] that the “ancillary restraint” counterfactual that the General Court had justified at [94]-[96] of its decision was appropriate for the primary article 101(1) analysis. The General Court and the Commission had been entitled to conclude that the possibility of issuers “holding up” acquirers who were bound by the Honour All Cards Rule could only, in effect, be solved by a scheme rule prohibiting ex post pricing. Such a rule was less restrictive of competition than MasterCard’s existing MIF solution. That led the CJEU to conclude at [173] that the ex post pricing prohibition could be regarded as a counterfactual hypothesis that was “not only economically viable in the context of the MasterCard system but also plausible or indeed likely, given that there is nothing in the [General Court’s] judgment … to suggest, and it is common ground … that MasterCard would have preferred to let its system collapse rather than adopt” that solution.

At [174], the CJEU concluded that, despite the General Court’s error, it had been entitled to rely on the same counterfactual it had used in the context of its objective necessity analysis “albeit for reasons other than those … in [132] and [143]” of the General Court’s decision. The error identified at [169] therefore had no bearing on the analysis of the restrictive effects carried out by the General Court by reference to the counterfactual it used. At [175] the CJEU made clear that “[l]ikewise, that error has no bearing on the operative part of the [General Court’s] judgment …, which [was] well founded on other legal grounds”. We emphasise that the CJEU thought the General Court had been deciding a legal issue in identifying the relevant counterfactual.

The CJEU again endorsed at [192] the counterfactual employed by the General Court, observing that the General Court had not regarded MIFs as, by their very nature, injurious to the proper functioning of normal competition, but that it had properly analysed the competitive effects of the MIFs at both [143] and [123]-[193] (see below).

It is true that at [193] the CJEU approved the last sentence of [143] of the General Court’s decision to the effect that the MIFs had restrictive effects, in contrast with an acquiring market operating without MIFs, in that they limited the pressure which merchants could exert on acquirers when negotiating the merchants’ service charge by reducing the possibility of prices dropping below a certain threshold. This passage makes it clear that the counterfactual approved by the CJEU was one that involved an absence of MIFs, with the abrogation of the default MIF rule and the imposition of an ex post pricing rule.

In [195], the CJEU dealt with the argument that the General Court’s decision was based on the premise that high prices in themselves constituted the infringement of article 101(1). The CJEU said expressly that it was apparent from [143] of the General Court’s decision that high prices, arising as the result of the MIFs, themselves limited the pressure which merchants could exert on acquiring banks, with a resulting reduction in competition between acquirers as regards the amount of the merchants’ service charge. The General Court had not merely assumed that the MIFs set a floor for the merchants’ service charge; it had undertaken a detailed analysis in [157]-[165] to determine that was the case.

Finally, the CJEU said at [196] that it was in a position to carry out its review of the analysis underlying the statements in [143] of the General Court’s decision. That is important because it was suggested by the schemes that [193] of the CJEU’s decision had not endorsed the first sentence of [143] of the General Court’s decision. In fact, however, the CJEU endorsed the analysis in [143] “taking into account the considerations in” [183]-[195], and held that “the General Court [had given] reasons to the requisite legal standard for its analysis relating to the effects of the MIF on competition”.

In our judgment, the proper analysis of the CJEU’s decision on these points is that it endorsed the counterfactual adopted by the General Court as a matter of law. It rejected the arguments (i) that the “no default MIF” and prohibition on ex post pricing counterfactual was inappropriate, (ii) that there was no basis for saying that the MIF set a floor on the merchants’ service charge (see also the CJEU at [197]), and (iii) that the imposition of the MIFs did not restrict competition between acquirers because the merchants could still compete in relation to the parts of the merchants’ service charge that were unaffected by the MIF.

It would be remarkable if the same scheme rule requiring the payment of MIFs in default of the agreement of bilateral interchange fees were held to be in breach of article 101(1) in one Member State, but not in breach of it in another Member State, whatever the factual or expert evidence might have been as to what might have happened in the postulated counterfactual. We say this because factual and expert evidence as to what will happen in a counterfactual position (i.e. in the absence of a particular agreement) is not hard-edged. It is, by its very nature, a kind of informed speculation, as we have seen very clearly from [180]-[181] of the CAT’s decision and

We turn to consider the reasoning of the lower courts on the main article 101(1) issue.

Popplewell J concluded at [154]-[155] that there was no distinction to be drawn in this case between a restriction counterfactual and an ancillary restraint counterfactual, and that one realistic counterfactual which would or might arise was “(i) a zero MIF (which is the same as no MIF with a prohibition on ex post pricing)”. He held that, subject to the death spiral argument, the MasterCard MIFs “did amount to a restriction of competition on the acquiring market by comparison with a counterfactual of no MIF”, because they imposed a floor below which the merchants’ service charge could not fall, because “acquirers had to pay at least that much to issuers and had to recoup it from the merchants, which in turn led to higher prices charged by acquirers to merchants through the [merchants’ service charge] than if the MIF were … zero”. Such a floor, said Popplewell J, restricted competition because it interfered with the ability of acquirers to compete for merchants’ business by offering merchants’ services charges below such floor. It was no different from a collective agreement by manufacturers to maintain inflated wholesale prices, which prevents wholesalers competing on the retail market below those prices.

Popplewell J then held at [156]-[158] that these conclusions had been “consistently the view of the Commission in relation to EEA MIFs” and applied equally to UK MIFs. We agree, although we would not have expressed our conclusions as to what the Commission decided, and what the General Court and the CJEU approved, in precisely the terms adopted by Popplewell J.

Popplewell J considered the “death spiral argument applied to the zero MIF counterfactual” at [163] onwards. In our judgment, Popplewell J fell into error (particularly at [182]-[185]) in considering the death spiral argument at all in relation to the question whether the measures were a restriction of competition under article 101(1). It is common ground that the correct approach to deciding the primary article 101(1) question was set out at [111] in Cartes Bancaires as follows: “determining whether, in the absence of the measures in question, the competitive situation would have been different on the relevant market, that is to say whether the restrictions on competition would or would not have occurred on this market”.

It is common ground that the relevant market for article 101(1) purposes is the acquiring market. That is stated in the first issue agreed between the parties under article 101(1). But the death spiral argument does not concern a comparison between the state of competition in the acquiring market with and without the “measures in question”. Instead, it concerns the effects on the inter-system market and the issuing market of issuers switching to a competing scheme in order to earn MIFs in the absence of MIFs being imposed in the MasterCard scheme. It is true that the putative decline of business in the inter-system market and the issuing market affects the level

The General Court made this point clear at [172]-[173] as follows: “the Commission took the view that four-party bank card systems operated in three separate markets: an inter-systems market, an issuing market and an acquiring market, and relied on the restrictive effects of the MIF on the acquiring market” and “[i]t must be held that such a definition is not manifestly erroneous”. This approach was approved at [178] and [180] of the CJEU’s decision.

It is no justification for the course Popplewell J adopted that the CJEU’s decision at [177]-[179] also mentioned the need to consider the restriction within its actual context and the possibility of taking into account the two-sided market at the article 101(1) stage. The CJEU had rejected at [180]-[182] the argument that the General Court ought to have taken into account the economic advantages of the two-sided nature of the system at the 101(1) stage. The CJEU approved the General Court’s concentration on the acquiring market at the 101(1) stage, and said that no contrary argument had been addressed to it.

We will explain in detail, when we come to deal with the death spiral argument at the ancillary restraint stage, the reasons why we think the argument ought not to have succeeded even at that stage.

In setting out the principles applicable to the article 101(1) analysis at [83]-[97], Phillips J identified at [89] an “heretical assumption about MIFs” to the effect that “MIFs are necessarily the result of restrictive agreements within the meaning of Article 101(1) because they result in higher [merchants’ service charges] than would be charged in their absence”. This was central to Phillips J’s reasoning, but we disagree with him.

Phillips J referred to Bookmaker’s Afternoon Greyhound Services Ltd v Amalgamated

Having considered the issues as to the correct counterfactual and the evidence before him in detail, Phillips J concluded at [137] that the effect of the merchants’ argument was that any level of MIF, on the scale from infinitely positive to infinitely negative (including an infinitesimally small level), was deemed to be a restriction of competition, all in comparison with an infinitesimally small point on that scale equating to there being no MIF (a figure of zero). He held that there was “in this context, no magic in the number zero and no reason why it represent[ed] an inherently more competitive situation than any other level [of MIF]”. We take the view that [137] of Phillips J’s judgment is beside the point. As we have already said, the exercise under article 101(1) is to consider whether there would be more competition in the absence of the measure in question. The measure in question here was the rule that, in the absence of bilateral agreements, a default MIF would be imposed. In the absence of such a rule, there would have been no bilateral agreements and no MIFs would have been charged, because there would have been either a settlement at par rule or an ex post pricing restriction, as the CJEU held. Accordingly, we think that Phillips J was wrong to think that there was “no magic in zero” just because of the possibility of negative MIFs. Moreover, the Commission saw any positive MIF as setting a floor under and inflating merchants’ service charges, as it said at [522].

In dealing with the CJEU’s decision, Phillips J wrongly held at [142] and [148] that [460] of the Commission’s decision demonstrated that its “conclusion that the MIF restricted competition in the acquiring market was based on its finding of fact that, in the absence of the MasterCard MIF, there would be bilateral negotiations and agreements in the intra-EEA market, with Acquirers negotiating different levels of Interchange Fees, those who agreed higher Fees becoming less competitive than those achieving lower levels”. In fact, as we have already said above, the Commission’s decision at [460] was explaining why the MIFs were not like an excise tax, but actually restricted competition between acquirers and forced up prices for merchants. The reference to statements of retailers demonstrating that they would be in a position to exert pressure on acquirers in the absence of a floor to the merchants’ service charge was there to explain that, without a default rule that fixed a MIF in the absence of such bilateral agreements, merchants could shop around to contract with the acquirer who incurred the lowest interchange costs.

Moreover, there is no inherent difference between the Commission’s conclusion that there might be bilateral agreements for a time ending up in no interchange fees, and the position where there are just no MIFs without any bilaterals being agreed (see the Commission’s footnote 517 making it clear that in the absence of a default MIF banks may or may not enter into bilateral agreements on interchange fees).

In these circumstances, and for reasons we have also already given, we take the view that Phillips J was mistaken at [148] to conclude that the CJEU had not decided that positive MIFs of the kind charged by MasterCard are, as a matter of law, a restriction on competition. That, in the circumstances of the MasterCard decisions and these cases, was precisely what the CJEU decided.

Finally in this connection, Phillips J dealt with Popplewell J’s conclusion at [156] of his judgment that the MasterCard MIFs did amount to a restriction of competition on the acquiring market by comparison with a counterfactual of no MIF, because they imposed a floor on the merchants’ service charge. We consider that Phillips J was misled by Ms Rose’s argument as to a sliding scale of MIFs including the zero MIF as one point on that scale. As we have said, that ignores the basic question one is required to ask under article 101(1), namely whether there would be more competition without the measure in question, that is to say the rule imposing a default positive MIF in the absence of bilateral agreement. The answer to that question was delivered by the Commission, and approved by the General Court and the CJEU. The correct counterfactual envisaged no default MIF and a prohibition on ex post pricing. The MIF did set a floor on the merchants’ service charge, and restricted competition between acquirers, because the higher prices resulting from it limited the pressure which merchants could exert on acquirers, reducing competition between acquirers as regards the amount of the merchants’ service charge.

The CAT concluded at [267] that the UK MIF was a restriction of competition by effect within the meaning of article 101(1) because, in its counterfactual world of bilateral agreements, there would be very significant and better competition in the acquiring market than existed in the real world over the claim period. The CAT considered whether the MasterCard scheme would enter a death spiral, but decided that it would survive. Importantly for present purposes, the CAT also indicated obiter at [271(2)] that it had not concluded that any MIF must, ipso facto, be a restriction on competition:

“The reason we consider the UK MIF as set in the real world to be a restriction by effect is because, although the UK MIF is an Interchange Fee ostensibly set as a default rate, the rate selected in fact precludes or inhibits the agreement of a true market price. That is the mischievous effect of the UK MIF on competition. As we have described in paragraph 266(3) above, given the dynamic between Acquiring Banks and Merchants on the one hand and Issuing Banks on the other, there is a danger that if the MIF is set too high, Issuing Banks will be disinclined to negotiate, and Acquiring Banks/Merchants will not have the market power to make them”.

The argument before us did not give much consideration to this approach, save that the schemes argued that it demonstrated that the decision was one of fact, and submitted that one of the vices in the merchants’ submissions was that any level of MIF would be unlawful, even one allowed by the Interchange Fee Regulation. We accept that, in theory, it could have been argued that the schemes’ actual MIF rates during the relevant periods were so low as to differentiate themselves from the legal position determined by the CJEU’s decision. Plainly, the reasoning of the CJEU to which we have referred does not mean that any very small default MIF would automatically be a restriction on competition. The factual premise, however, of the MasterCard scheme that the Commission was considering and of the schemes that we are considering was that the default MIFs made up a large percentage (some 90%) of the merchants’ service charge. (Footnote: ⁵) In these circumstances, the fact that the CAT may have been correct to say that not every default MIF, however small, would automatically be a restriction on competition violating article 101(1) does not deprive

The CAT approached the relevance of the death spiral argument to the initial article 101(1) analysis in a similar way to Popplewell J’s later judgment, by concluding at [135] that the CJEU’s decision at [177] had required a look beyond the acquiring market alone. The CJEU had said that it was necessary “to take into account any factor that is relevant, having regard, in particular, to the nature of the services concerned, as well as the real conditions of the functioning and the structure of the markets, in relation to the economic or legal context in which that coordination occurs, regardless of whether or not such a factor concerns the relevant market”, and “[w]here there are several markets that are inter-connected, that very inter-connection, in our view, is a matter that needs to be taken into account”. In our judgment, however, as we have already said, it is necessary to read [177]-[182] of the CJEU’s decision as a whole. The CJEU was not gainsaying the need to concentrate on the acquiring market in determining whether the imposition of default MIFs were restrictive of competition. It was saying that such an evaluation had to be in an appropriate context, that context including the structure within which the acquiring market was set. We do not consider that the CJEU was sanctioning a broad consideration of the competitive effects within the entire inter-scheme market at the first stage. The CAT’s determination to consider the inter-scheme and issuing markets as part of the article 101(1) question also raises questions about the way it approached the evidence about what would have occurred in the counterfactual world.

It is now appropriate to consider the bilateral interchange fees issue and to ask whether the CAT was right to employ a counterfactual that assumed that acquirers would agree bilateral interchange fees in the absence of MIFs. Although it is generally only possible to appeal from the CAT on a point of law, it was agreed exceptionally for the Sainsbury’s v MasterCard appeal that the normal approach to appeals from the High Court should be adopted, because that was the basis on which the case was originally transferred to the CAT.

Mr Hoskins accepted for MasterCard that the Court of Appeal would only interfere with a trial judge’s (here the CAT’s) findings of fact where it properly determined that the “finding of fact is unsupported by the evidence or where the decision is one which no reasonable judge could have reached” (Note 52.21.5 to the CPR at page 1858). Mr Hoskins made essentially three submissions concerning the CAT’s findings that bilateral interchange fees would be agreed between issuers and acquirers in the counterfactual world. First, the findings should be quashed because there was no evidence to support them. Secondly, in so far as there was evidence before the CAT, it was contrary to the findings that the CAT made, and thirdly, even if the findings were theoretically open to the CAT, the intended findings that were a creation of the CAT itself ought to have been put by the CAT to the witnesses before appearing as findings in the decision.

The CAT was considering four options for issuers in relation to the counterfactual world, which it stated at [153]: (i) to negotiate bilateral interchange fees with acquirers; (ii) to accept participation in the MasterCard Scheme without any interchange fees from acquirers; (iii) to participate in an alternative settlement system, other than the interchange operated by MasterCard; and (iv) to leave the MasterCard Scheme altogether in the UK. For reasons we have given, we do not see the latter two

The CAT considered the question of whether its option 1 was realistic at [179]-[197]. It concluded at [197] that the detail of how the acquiring market might evolve in the counterfactual world where no UK MIF existed was “precisely the sort of speculative question that arises on a counterfactual hypothesis”. The CAT repeated at [180] that the “question what would have happened in the counterfactual world is a necessarily hypothetical question, and not a factual one”, but said that it agreed it had to apply the law so that its “counterfactual hypothesis had to be in accordance with the requirements laid down by” the CJEU’s decision. The CAT then described the process as “one of evidentially based speculation” because “[n]o amount of factual enquiry can ever conclusively tell us what would have happened on the counterfactual hypothesis”.

We agree that the counterfactual adopted by the CAT had to be in accordance with the requirements laid down by the CJEU. In these circumstances, we cannot see any proper basis for the CAT’s conclusion on this issue. The CJEU’s decision plainly approved a counterfactual in the same factual circumstances as the MasterCard scheme of “no default MIF and a prohibition on ex post pricing”.

We emphasise that we are not holding that no amount of evidence could have made it appropriate to find that, even in a “no default MIF and a prohibition on ex post pricing” counterfactual, bilateral interchange fees would have been agreed. It might have been possible to show that the economic background to the MasterCard scheme in question was so different to that being considered by the Commission, the General Court and the CJEU that a different outcome would have occurred in a similar counterfactual world. The evidence relied upon by the CAT in [182]-[197], however, comes nowhere near to achieving that objective.

The CAT purported to place greater reliance on the factual than the expert evidence at [181], but even the factual evidence was exiguous. None of Sainsbury’s witnesses said it would have volunteered to pay an interchange fee to its acquirers. Mr Coupe, Sainsbury’s Chief Executive Officer, said that it “would use whatever negotiating leverage we could create in order to reduce the costs, to reduce the prices that we were charged”. The CAT acknowledged that no bilateral interchange fees were agreed in the UK ([183]), but took comfort from some inconclusive evidence from Mr Douglas, the Executive Vice President and General Manager of MasterCard in the USA ([185]), and from Mr Willaert, Head of MasterCard’s Interchange Fee Team from 2010-12 ([184]). The CAT itself acknowledged that the expert evidence was not supportive of bilateral interchange fees being agreed in the way it found to be likely. Dr Gunnar Niels, MasterCard’s expert, said that, if the default MIF was zero, there would be no negotiation of positive bilateral interchange fees because the merchants would insist on a short-term costs benefit ([194]-[195]). Mr Nils von Hinten-Reed, Sainsbury’s expert, thought that bilateral negotiations would only produce a very low interchange fee ([186]-[187]), so that the costs of negotiation would not be worth the effort.

It is not necessary for us to engage in a complete review of the evidence before the CAT in order to reach our conclusion that the CAT did not have a sufficient evidential foundation to come to the conclusion that significant bilateral interchange fees would have been agreed in the absence of the MIF. There was simply no substantive evidence supporting the proposition that “bilateral interchange fees would be likely to

We accept Mr Hoskins’s submission that the CAT’s findings as to the likelihood of bilateral interchange fees being negotiated between issuers and acquirers in the counterfactual world should be set aside on the ground that it was inadequately supported by the evidence. It follows that the CAT’s evaluation of the level of those bilateral interchange fees must also be set aside.

Our conclusions on the primary article 101(1) issue can be summarised quite shortly. The correct counterfactual for schemes like the MasterCard and Visa schemes before us was identified by the CJEU’s decision. It was “no default MIF” and a prohibition on ex post pricing (or a settlement at par rule). The relevant counterfactual has to be likely and realistic in the actual context (see the O2 Germany case at [68]-[71] and the CJEU’s decision at [169]), but for schemes of this kind, the CJEU has decided that that test is satisfied.

The CJEU’s decision also made clear at [195] that MasterCard’s MIFs, which resulted in higher prices, limited the pressure which merchants could exert on acquiring banks, resulting in a reduction in competition between acquirers as regards the amount of the merchants’ service charge. This is not a decision from which this court either can or should depart. It answers the schemes’ argument that, whether as a matter of evidence or not, the competitive process will not differ in the counterfactual. The default MIFs may be a transparent common cost, which is passed on by acquirers to merchants, and which does not figure in the negotiations between them, but it does not follow that acquirers nonetheless compete as strongly for merchants’ business in relation to the acquirer’s margin and the additional services they offer, as they would in the absence of the default MIFs.

Ms Rose is, in our judgment, wrong to submit that all MIFs will infringe article 101(1) as a result of our decision, even those permitted under the Interchange Fee Regulation. We do not discount the possibility that some evidence might conceivably enable other schemes to distinguish different MIFs from those upon which the CJEU was adjudicating. In the present case, however, the MIFs are materially indistinguishable from the MIFs that were the subject of the CJEU’s decision. In both cases, the MIFs represented the vast majority of the merchants’ service charge, and the appropriate counterfactual was a “no default MIF” plus a prohibition on ex post pricing.

The death spiral argument is not relevant at this stage of the debate because the article 101(1) question must be asked in relation to the acquiring market.

We have concluded that the CAT fell into error when it held that it was likely that bilateral interchange fees would be negotiated between issuers and acquirers in the counterfactual world. That decision, and its decision as to the level of the likely bilateral interchange fees, must be set aside.

We, therefore, agree with Popplewell J that the rules of the MasterCard scheme providing for a default MIF in the absence of bilateral interchange fees infringed article 101(1), and we disagree with Phillips J’s contrary conclusion in respect of the Visa scheme.

Mr Brealey submitted that, on the basis of the correct legal analysis to be applied to the issue of ancillary restraint as we have held it to be in Part IV of this judgment, the approach accepted by Popplewell J was wrong because he adopted a test which was essentially subjective rather than objective. The fallacy of his approach was demonstrated by the statement at [178] of his judgment, to which Mr Brealey took particular exception: “If competition would kill off the main operation without the restriction being in place, the restriction is necessary for the main operation”. Mr Brealey submitted that this proposition cuts across the fact that the whole purpose of competition was to compete on price, quality and service and beat one’s competitors. The judge’s approach would enable an inefficient enterprise which entered into a restrictive agreement to rely upon the ancillary restraint doctrine to argue that it needed the restrictive agreement to prevent its more efficient competitors putting it out of business.

The same point was made by Mr Turner, on behalf of the AAM parties, that, if subjective competitive considerations were relevant to the ancillary restraint issue, a group of substandard businesses, which could only stay afloat because of a collusive restrictive agreement, would be able to say that there was no unlawful restriction of competition. This was wrong in principle.

Mr Turner submitted that the correct question under the ancillary restraint doctrine, which provided a narrow exception to the restriction of competition under article 101(1), was whether the relevant restriction, here the default MIF, was necessary for the survival of the type of main operation, here a four-party card payment scheme. That question had to be answered in the negative since the schemes accepted that a four-party scheme would not collapse simply because the issuing banks did not receive MIFs.

The Commission supported the merchants on this issue, both on the correct legal test (which we have addressed in Part IV of our judgment) and in relation to the error in Popplewell J’s acceptance of the asymmetrical counterfactual. Ms Kreisberger submitted that, if the schemes could rely upon the death spiral argument as a valid counterfactual, this would undermine the full effectiveness of article 101(1), contrary to fundamental principles of EU law and underlying competition policy.

As we have already noted in Part IV of this judgment, the principal argument advanced by Mr Hoskins on behalf of MasterCard, in seeking to uphold the judgment of Popplewell J in relation to the death spiral issue, was that the judge had been

Mr Hoskins submitted that Popplewell J had been correct to say that material identity cannot be assumed and must be the subject of evidence, on which the burden was on the merchants, who had raised the point of material identity, and that there was insufficient evidence to conclude that Visa’s MIFs were materially identical and therefore unlawful. Indeed, we consider that Popplewell J was clearly imposing a requirement of showing material identity under both article 101(1) and article 101(3) in [204] of his judgment. Mr Hoskins submitted that Popplewell J had been correct to say that establishing material identity required an analysis of both article 101(1) and 101(3). In response to the reaction of the court that, if correct, this would have involved in the MasterCard trial many days of evidence and submissions about the Visa scheme and whether its MIFs would have been exempt under article 101(3), Mr Hoskins submitted that this was an inevitable consequence of the AAM parties seeking to raise an issue of material identity on which the burden was on them.

Ms Rose on behalf of Visa directed the preponderance of her argument on the death spiral to its impact on restriction of competition, with which we have already dealt in the main part of our conclusions on article 101(1). In her submissions in support of Visa’s Respondent’s Notice, Ms Rose also made various points supportive of Mr Hoskins on the death spiral in the context of the ancillary restraint doctrine. She submitted that Phillips J had been wrong to conclude in [168(ii)] of his first judgment that it was difficult to conceive of a scenario where one scheme was unable to set default MIFs, but the other remained unconstrained. This was exactly what had occurred in Hungary in 2010, as had been in evidence before Phillips J, when Visa had significantly cut its debit MIFs pursuant to a commitment given to the Commission, but MasterCard had not, and Visa had lost half its market share in that country. This was also what occurred in the United Kingdom when, in order to satisfy Commission decisional practice, Visa maintained lower credit card MIFs than MasterCard. Visa lost substantial market share.

On this issue, we will apply the legal principles applicable to the ancillary restraint doctrine as set out in Part IV of this judgment. On that basis, Popplewell J was wrong, as we have said, to conclude that the issue of whether, in the absence of the default MIF, the MasterCard scheme would survive in view of the competition from Visa was one which could be considered under the ancillary restraint doctrine under article 101(1). Such questions relating to the application of the so-called asymmetrical counterfactual are not for the ancillary restraint issue under article 101(1), but for the issue of exemption under article 101(3).

We agree with the merchants that, if questions of the subjective necessity of a restriction for the survival of the particular main operation were relevant for the purposes of the ancillary restraint doctrine, it would enable failing or inefficient businesses that could not survive without a restrictive agreement or provision to avoid the effects of article 101(1), which would undermine the effectiveness of that provision of EU law and the underlying competition policy.

The only question in relation to the potential application of the ancillary restraint doctrine in the present context is whether, without the restriction of a default MIF (which is the relevant counterfactual), this type of main operation, namely a fourparty card payment scheme, could survive. The short answer to that question is in the affirmative and the contrary was not suggested by MasterCard or Visa. There are a number of such schemes in other parts of the world which operate perfectly satisfactorily without any default MIF and only a settlement at par rule.

Even if Popplewell J had been correct in his conclusion that the decision of the Court of First Instance in Metropole was implicitly disapproved by the CJEU in MasterCard, so that it was appropriate to consider, in the context of the ancillary restraint doctrine, the competitive effects of the removal of the restriction in question on the specific main operation, we consider that his adoption of the asymmetrical counterfactual was incorrect for two related reasons.

First, as the CJEU’s decision makes clear at [108]-[109], the counterfactual must be a realistic one. The asymmetrical counterfactual which Popplewell J accepted assumes that MasterCard would be prevented from setting default MIFs but Visa would remain unconstrained. As Phillips J said at [168(ii)] of his first judgment, addressing the mirror argument made by Visa in that case, that situation is “not merely unrealistic but seems highly improbable”. As Phillips J said, the schemes are engaged in the same business, using the same model and are fierce competitors. We were not impressed in this context by the arguments on behalf of the schemes that there have been inconsistencies in approach on the part of the Commission and other competition authorities and regulators. Whilst there have been differences in the detail, as appears from the chronological background set out at Part II of this judgment, the competition authorities and regulators have sought to constrain both schemes in a broadly similar fashion. We consider that a realistic counterfactual would assume that, if one of the schemes was unable (whether for commercial or legal reasons) to set default MIFs, the other scheme would be similarly constrained.

The correctness of that conclusion was not undermined by the points made by Ms Rose about what had happened historically in Hungary or even in the United Kingdom. The critical point is that the hypothesis of the asymmetrical counterfactual is that one of the schemes would be prevented from setting any default MIF but the Commission and the UK competition authorities and regulators would allow the other scheme to carry on setting its default MIFs, without any constraints being imposed. That seems to us to be completely unrealistic and improbable. Realistically there would be similar constraints on both schemes.

Secondly, Popplewell J accepted at [189] of his judgment that, if the AAM parties were right that the two schemes were materially identical, he would have had to assume that, in the counterfactual world, Visa’s MIFs would be constrained to the same extent as MasterCard’s. His essential reasoning for that conclusion at [190][193] of his judgment was that it should not be open to one unlawful scheme to save itself by arguing that it otherwise would face elimination by reason of competition from the other scheme, which is itself unlawful.

On the evidence before him, however, Popplewell J considered that the AAM parties had not established that the Visa scheme was materially identical to the MasterCard scheme he was considering. He concluded at [204] that what was material was whether and to what extent Visa’s MIFs as set constituted an unlawful restriction of competition infringing article 101, which involved considering all the features of the

This conclusion suffers from the same fallacy as Popplewell J’s acceptance of the argument that, for the purposes of the ancillary restraint doctrine, it is permissible to look at the competitive or commercial effect of the removal of the restriction in question on the specific main operation. It brings into the article 101(1) analysis matters which are only to be considered under article 101(3). Once it is recognised that the relevant test is only satisfied if the restriction is objectively necessary for the survival of the type of main operation in question and the subjective necessity of the restriction for the survival of the specific main operation is irrelevant, it is clear that it is only material identity in respect of matters relevant to article 101(1) that would have to be established.

We consider that the two schemes are materially identical for the purposes of the article 101(1) analysis. They are both four-party card payment schemes with an Honour All Cards Rule for credit and debit cards, in which default MIFs are set which are paid to issuing banks and passed on to the merchants as part of the merchants’ service charge imposed by acquiring banks. In those circumstances, even if Popplewell J had been correct that it was appropriate to consider, in the context of the ancillary restraint doctrine, the competitive effects of the removal of the restriction in question on the specific main operation, he should have gone on to conclude that the schemes were materially identical, so that in the counterfactual world Visa’s MIFs would be constrained to the same extent as MasterCard’s.

For all these reasons, we consider that Popplewell J erred in accepting the death spiral argument and should have upheld his initial conclusion that MasterCard’s MIFs were a restriction on competition under article 101(1). By parity of reasoning, Phillips J was correct to reject the death spiral argument in his first judgment.

In view of our conclusion on the death spiral issue, it is not necessary to consider the AAM parties’ further ground of appeal that, even if Popplewell J was correct to accept the death spiral argument, he misapplied it to the facts because MasterCard’s actual MIFs were set at higher rates than the “survival” rates which the judge found were objectively necessary at 0.2% less than Visa’s MIF rates.

Although there is some overlap between the AAM parties’ appeal against Popplewell J’s judgment in favour of MasterCard on article 101(3) and Visa’s appeal against Phillips J’s judgment holding that Visa had not established exemption under article 101(3), the two appeals raise separate considerations and need to be considered separately. We will consider first the appeal against Popplewell J’s judgment.

In order to provide the proper context for the parties’ arguments and our decision on this aspect of the AAM parties’ appeal, it is necessary first to describe in a little more detail than we have done above the judge’s reasoning in this part of his judgment. Having set out the three conditions of article 101(3) which had to be satisfied and his conclusions on the burden and standard of proof (to which we have already referred), the judge turned to consider the first condition, the benefits requirement. At [308] he identified what Mr Hoskins described as “the virtuous circle”, namely the 6 benefits to merchants of accepting credit and debit cards. In terms of benefits derived from the MasterCard scheme generally, although four of those were challenged before the judge, only one remains challenged by the AAM parties on appeal, so-called “business stealing” to which we return later in this judgment.

At [310], the judge then set out the reasons given by MasterCard as to why all those benefits to merchants were “at least to some extent the result of charging a positive MIF”. He recorded at [311] the merchants’ complaint that this was a rehash of the “system output” argument which had failed before the Commission and the General Court and which suffered from the error of confusing benefits conferred by the scheme generally with those caused by the MIF.

He rejected that complaint at [312], as follows:

“... MasterCard’s argument that charging positive MIFs led to an increase in the use of cards and therefore an increase in the amount of the benefits enjoyed by merchants as a result of the use of cards is made good on the evidence before me. So too is its case that because cardholders received benefits from issuers which were funded by the MIF, the benefits to merchants of card use are to some extent directly caused by the MIF. That does not mean that all the benefits enjoyed by merchants are directly attributable to the level of MIFs charged by MasterCard. It does, however, mean that a MIF at some positive level is directly causative of some benefits to merchants. That is the starting point for the Article 101(3) process. There then remains to be addressed the difficult quantification exercise involved in valuing those merchant benefits which are directly attributable to the MIF. This raises difficult questions, which I address below, but none of them were issues which MasterCard sought to address before the Commission, or which the Commission, General Court or CJEU needed to address.”

The judge then noted at [313] the merchants’ argument that, based upon the expert evidence of Mr Dryden and the Rochet and Tirole 2008 study (Footnote: ⁶) (which developed the MIT to which we referred earlier in this judgment), maximising system output (i.e. maximising card usage) can be positively deleterious rather than beneficial. The judge continued:

“The theory, in a nutshell, is that payment card systems can exploit merchants’ fears of losing business, so as to cause them to be willing to pay higher interchange fees than are justified by the benefits received by taking a card payment in place of cash payment. This was characterised as the “must take cards” phenomenon. The view that this creates system “inefficiencies” assumes (a) that business stealing is an irrelevant merchant benefit, and (b) that merchant benefits are to be measured by reference only to the difference between cash and card transactions, neither of which are correct assumptions for the reasons which I explain below.”

It is these conclusions of the judge at [312]-[313] that increased card usage, which is to an extent caused by the MIF, is always beneficial to the merchants and never deleterious, which Mr Turner on behalf of the AAM parties submitted contain a fundamental fallacy that then infects the remainder of the judge’s analysis.

What the judge then did over some 20 paragraphs was to consider in detail the 6 areas of merchant benefits derived from the virtuous circle. He did so primarily by assessing the extent to which the particular benefit was conferred by the card scheme as a whole, as opposed to focusing specifically on the extent to which MasterCard had established that the benefit was caused by the default MIF. At places in this analysis, the judge did refer to the MIF being causative of benefits, but this was very much on the basis of the economic analysis propounded by the experts. One exception was the fraud guarantee, in respect of which he referred to the evidence of Mr Willaert of MasterCard, although, as the judge noted, Mr Willaert accepted that MasterCard would have deployed anti-fraud technology even if MIF revenue had not been available.

The judge then set out his conclusion at [335] that the MIF directly contributes to some extent to each of the 6 benefits to merchants. From [336] onwards, he dealt with the quantification of benefits and specifically with the MIT; and, from [369] onwards, he made adjustments to the MIT he considered necessary to reflect the benefits other than the avoided cost of cash sales he had found merchants receive from the use of MasterCard cards, of which the MIF was to some extent directly causative, and which were not reflected in the MIT MIF methodology used in the Commission Survey. He arrived on that basis at adjusted figures for the MIT MIF to reflect the value to merchants of the use of cards, which he set out at [389].

He then dealt at [390]-[397] with “business stealing” or the competitive advantage of accepting cards, to which he ascribed an increase in the MIT MIF value of 0.4% for credit cards and 0.2% for debit cards, to arrive at a value to merchants of accepting cards under his adjusted MIT MIF methodology.

It is only at this point of his analysis that the judge turned to adjustment of those figures to reflect the extent of issuer pass-through, in other words, the extent to which issuers actually pass the MIF revenue to cardholders (in the sense of using it to incentivise card usage, whether by reward schemes, marketing or innovation such as “Contactless”) as opposed to simply retaining the MIF revenues as profits for themselves and their shareholders; what the Commission at [730] of the Commission decision called “extracting rents”. At [398] the judge recognised that the level of issuer pass-through was relevant to all of the first three conditions under article 101(3).

At [399] he identified the ways in which issuers “potentially” pass through the MIF to cardholders, although at this stage he did not analyse the extent to which any of these could be said actually to incentivise card usage, a point which he dealt with at [403]. At [400] he said that quantification of pass-through was difficult to estimate because MIF revenue was not an isolated pot of money. He noted that neither expert “sought to quantify the extent of pass-through other than “[by] adjectival terms”: “very high” according to Dr Niels, MasterCard’s expert, and “a significant proportion” according to Mr Dryden, the AAM parties’ expert.

At [401] Popplewell J set out three sources of “potentially helpful data” referred to by Mr Dryden, one of which, an econometric analysis by the Commission in 2006, the judge then discounted as “not a secure indication of issuer pass-through generally for the purposes of the current litigation”. The judge described the other two sources as follows:

“(2)

An analysis of debit interchange fees in the US by Evans Chang and Joyce applied an indirect method of measuring how announcements of proposed caps on debit interchange fees in the US affected the price of issuers’ shares. The authors inferred that 80% of debit card interchange fee was passed through to buyers.

(3)

PwC analysed research conducted in 2001 by Research International which asked 648 respondents about their likely response to price change in credit cards, including the introduction of a transaction charge of 0.25% on all credit cards. The conclusion was that the introduction of such a charge would decrease the total transaction volume on credit cards by 26%. This suggests that in the UK the MIF subsidisation of costs which would otherwise have to be recouped in card fees has a very significant impact, implying a high level of pass through.”

At [402] the judge said that:

“Dr Niels based his opinion that pass through was very high on economic theory and the competitive nature of the issuing market: as a matter of economic theory, in a perfectly competitive market suppliers set prices equal to marginal costs; and a single cost change to an industry which is highly competitive results in all or almost all of the cost change being passed on to the consumer; there are a large number of issuers in the UK and in general the UK market is regarded as highly competitive, particularly for credit cards…”

He then elaborated on various examples Dr Niels had given, including publicly available information about issuers who reduced their reward offerings following the cap imposed in late 2015 by the Interchange Fee Regulation. The judge discounted these as follows:

“…in my view these examples provided no firm empirical basis to support or undermine his estimate. They are not a representative sample and do not reflect the end point in any reduction of benefits. His estimate of 80%-100% was not based on any identified data or calculations.”

At [403], the judge said that Mr Dryden had not attempted any estimate of the percentage. He then referred to the further point made by Mr Dryden that it was not enough for the MIF to be passed through in the abstract, but the pass-through must cause the card holders to respond, in other words incentivise card use. The judge accepted that was sound in principle:

“…because a MIF which does not translate into card use cannot have a direct causative effect on merchant benefits arising out of card use. Mr Dryden and Dr Niels accepted that pass through for credit cards, which were stand-alone products, met this hurdle but Mr Dryden opined that for debit cards it "may not".”

The judge set out Mr Dryden’s reasons for reaching that conclusion and noted at [404] that Dr Niels disagreed. At [405], the judge said there was some force in Mr Dryden’s points, which led him to conclude that the extent of effective pass-through was less for debit cards than for credit cards.

At [406] the judge dealt with Mr Dryden’s analysis that, if pass-through was less than 100%, it was necessary that, in effect, the number of switching transactions (i.e. increased card usage) was sufficiently large relative to the number of “always card” transactions (i.e. transactions which would always have been made using a card where the MIF would always be payable by the merchants). The judge rejected this analysis as having been made on the false premise that the only relevant benefit to merchants is that of not taking cash and having ignored both the various other benefits from the use of cards the judge had identified and the existence of Amex as a competitor, to which a significant proportion of MasterCard premium business would have been lost if MIFs were not set at competitive levels for issuers. Mr Turner was particularly critical of the judge’s dismissal of the always card/switching analysis, which he submitted was part of the balancing exercise required by the first condition of article 101(3).

At [407] the judge mentioned the question of issuer profits from card related income generally and said at [408] that: “I have not had substantial evidence of issuer profits or issuer finances which might assist on this question. Neither side called any witness with substantial contemporaneous experience as an issuer. The documentary evidence was exiguous.” He then summarised the evidence which he had.

The judge set out his conclusions on issuer pass-through and issuer profits at [409] as follows:

“Doing the best I can, my conclusions are that the degree of issuer pass through is likely to be at least 75% for credit cards and at least 40% for debit cards. These percentages are intended to reflect the causative effect of the MIF in incentivising card use. The level of profit for issuers is not such as to violate the fair share or indispensability requirements. A profit at levels which varied across the market between 10% and 40% is not indicative of an unreasonable return on capital for a bank and would not of itself represent such a degree of profit as to cause the fair share or indispensability requirement to be breached if the threshold test of merchant benefit neutrality were fulfilled.”

At [410] he then set out the “value of merchant benefits caused by the MIF using the adjusted MIT MIF methodology” rounded to the nearest basis point for each of the types of cards he was considering, for example for UK credit cards 1.01% being 75% of 1.35% (his earlier adjusted figure set out at [397]) and 0.38% for UK debit cards being 40% of his earlier adjusted figure of 0.95%. At [418] he went on to conclude that these levels of MIFs would be exempt under article 101(3).

Mr Turner emphasised the need for a causal link between the restriction, here the default MIFs, and the net benefits, as was made clear by the CJEU’s decision in MasterCard at [234], which required a balancing exercise to establish that the advantages to the relevant consumers caused by the restriction outweighed the disadvantages which the restriction entailed for competition. He submitted on the basis of CJEU jurisprudence and Commission decisional practice that this causal link cannot be based merely on economic theory but must be supported by robust analysis

Mr Turner submitted that Popplewell J fell into error because at [312]-[313] he started from the assumption that the default MIFs always bring benefits to merchants, and that therefore all he needed to do was to quantify that benefit. He held that MasterCard’s argument that positive MIFs benefited the merchants was made good on the evidence before him, but there is no indication as to what that evidence was. Phillips J was correct to say at [51] of his second judgment that the assumptions and estimates in Popplewell J’s judgment constituted “little more than putting a finger in the air”.

He submitted that the judge had paid insufficient regard to the concerns expressed by the Commission and the CJEU that MIFs can overburden merchants and that it cannot simply be assumed, even if MIFs increase card use to some extent, that this will produce benefits to consumers that outweigh the disadvantages, hence the requirement for robust analysis and cogent evidence showing a link between MIFs and benefits. By assuming that positive MIFs led to increased card usage, which in turn benefited the merchants, without considering properly the disadvantages in relation to transactions where the cardholder would always have used the scheme card anyway irrespective of the MIF, the judge had failed to conduct the balancing exercise required under the benefits requirement: see the CJEU’s decision at [234].

The judge erred in his approach to pass-through because he ignored the need for cogent factual evidence as to both the extent to which MIF income was passed through to cardholders and the extent to which such MIF revenue as was passed through did stimulate additional card usage by cardholders. There was no such evidence in this case. MasterCard could have produced evidence from issuers but did not. In the circumstances, the judge should have concluded that MasterCard could not establish the necessary causal link between MIFs and any net benefits.

The “always card” issue is a problem for merchants because, in a mature card market such as the UK, the great preponderance of card transactions are ones where the cardholder would always have paid with a scheme card anyway irrespective of the MIF, so the MIF is a burden on the merchant with no corresponding benefit. Phillips J correctly recognised this at [49] of his second judgment, but Popplewell J wrongly rejected the always card point at [406]. Contrary to his conclusion (and to Mr Hoskins’s intervention during Mr Turner’s submissions), the AAM parties’ case at trial was not linked to switching away from cash specifically, but applied whatever the alternative payment method from which there was switching to the scheme card.

Mr Turner submitted that the judge had been wrong to regard business stealing as a competitive advantage and thus a benefit to merchants. The Commission and the CJEU had regarded it as a predicament rather than a benefit because merchants fear losing business even with a high level of MIF. The judge’s approach led to whatever the banks could get away with, in terms of imposing high MIFs on merchants, being converted into a benefit leading to exemption under article 101(3).

On behalf of the Commission Ms Kreisberger supported the AAM parties’ submissions as to why Popplewell J’s approach was incorrect. The Commission’s approach to article 101(3) can be expressed in four propositions. First, there must be a causal link between the default MIFs and increased card usage demonstrated by empirical data and evidence and not just economic theory: see the Commission’s decision at [695] and the Commission Survey at [54]. A “ledger” approach, however, is not required. It was accepted that qualitative benefits, such as contribution to development of new products, may not be capable of quantification, but evidence must still be provided: see [57] of the Guidelines.

Secondly, it must be proved that increased card usage gives rise to net benefits or efficiencies. Maximising card usage cannot be assumed to be a good thing, particularly in a mature market: see [60] of the Commission Survey. Thirdly, passthrough of MIF revenue to card holders must be established. Fourthly, there must be a reasonable channel through which MIFs can promote the use of cards to consumers: see [72] of the Commission Survey where we note that the Commission makes the point that reward schemes for debit cards do not exist and cardholding is widespread. MasterCard’s arguments on article 101(3)

On behalf of MasterCard, Mr Hoskins submitted that Popplewell J had adopted the correct approach, taking the MIT as a benchmark or starting point and then considering the evidence which supported the case for exemption. He had made findings of fact which this court should not overturn unless no reasonable judge could have reached them.

Mr Hoskins said that, contrary to the submissions for the merchants, there was detailed evidence before Popplewell J on how MasterCard sets its rates, namely that of Mr Willaert (of MasterCard) and Mr Sidenius (of Edgar Dunn, a consultancy firm engaged by MasterCard to assist in determining the right level of MIFs). There was also detailed evidence on the extent of issuer pass-through, albeit mostly not from issuers themselves. He referred to the three pieces of data set out at [401] of the judgment, together with an FCA November 2015 report which concluded that there was strong competition in the credit card market (but without specific reference to MIFs) and evidence from four issuers that they had reduced their reward offerings to cardholders following the Interchange Fee Regulation (which the judge rejected).

In relation to the “always card” argument, Mr Hoskins submitted that it depended upon the market being mature, but the claim period dated back to 2006 and without evidence it could not be assumed the market was mature throughout this period. He submitted that the judge had been right to reject the argument for the reasons he gave in [406].

In relation to business stealing Mr Hoskins referred to [33] of the Guidelines which says that improving competition by offering better services to customers is a relevant benefit. There was evidence from merchants before Popplewell J (not disputed by the AAM parties) that accepting cards provides a better service to customers. It follows that any business stealing that results is a relevant benefit and the judge’s analysis was correct.

Despite the obvious care and detail devoted by the judge to the article 101(3) issue, there are a number of flaws in Popplewell J’s approach. First, in considering the first critical stage in the causation analysis which we identified at [88] above, namely whether the issuers were incentivised to increase card usage to a greater extent than they would have been anyway, the judge has noted towards the end of his analysis at [408] the absence of any factual evidence from the issuers but he paid insufficient

The significance of this point is clear from the Commission’s decision which referred at [686] to the fact that the other income from cards may have been a sufficient commercial incentive to issuers to invest in seeking to increase card income, emphasising in a footnote that, in the United Kingdom, issuing banks generated 90% of their revenues from income other than MIFs, such as interest charges, and only 10% from MIFs. The Commission made a similar point in relation to debit cards at [720].

This was a matter on which only factual evidence from the issuers could have remedied an obvious gap in MasterCard’s evidence. We were unimpressed by arguments advanced on behalf of both schemes that such evidence would have been difficult if not impossible to obtain, given that, as the judge found, MIF income was not an isolated pot of money. Issuing banks receive the MIF income and must know what they do with it in broad terms. Documentary evidence could have been obtained from issuers in the form of board minutes and internal memoranda and from studies as to how the banks spend income from credit and debit cards, their motives and intentions, how they view the stream of MIF income and whether and, if so, to what extent it causes them to provide promotional benefits to cardholders. Witnesses from issuers could have been called to speak to those matters. We accept that a “ledger” approach of giving precise figures for MIF pass-through might not be possible, but see no reason why issuers could not provide some estimation of the extent to which MIF income is passed through to cardholders. It was not suggested that MasterCard had attempted to obtain such evidence from issuers but been unable to do so. It simply did not attempt to obtain any such evidence.

Secondly, the judge hardly addressed the second critical stage of the causation analysis identified at [88] above, namely the extent to which card usage actually increased as a consequence of the steps taken by the issuers to incentivise it. The judge mentioned this point in passing at [403] and, as we have said, recognised that it was sound in principle. To be fair, the experts agreed, at least as regards credit cards, that pass-through would incentivise card use. There was, however, no empirical, factual evidence on this point, which seems to have been based on economic theory. If issuers were passing through MIF revenue to cardholders with a view to incentivising card usage, one would expect some factual or empirical evidence from those issuers as to the extent to which such pass-through was having the desired effect. It may be that the judge did not analyse this point further because he assumed (as is apparent from [312]-[313]) that increased card usage would always benefit merchants.

Thirdly, and perhaps of even more significance than the first two matters, the judge effectively failed to carry out the balancing exercise identified at [89] above to establish that overall the restriction, here the MIF, provided appreciable objective advantages for the relevant consumers of such a character as to compensate for the disadvantages which the restriction entailed for competition and, in this context, the burden it imposed on merchants. The disadvantages to which the default MIF gives rise are twofold: so far as the cardholders are concerned, the extent to which the MIF income is not passed through to them but simply retained by the issuers, and so far as

We agree with Mr Turner that the judge’s analysis overlooked or ignored these disadvantages and failed to carry out the relevant balancing exercise. Even on the judge’s assessment, issuers do not pass through 25% of the MIF income on credit card transactions and 60% of the MIF income on debit card transactions but retain it. The judge did not take sufficient account of this point. Likewise, the judge’s analysis does not take into account that a large percentage of overall transactions in a mature card market, which Phillips J at [49] of his second judgment considered the UK market to have been during the claim period, would always have been transactions using the relevant scheme’s cards anyway, so that the merchants bear the burden of the MIFs on those transactions without any corresponding advantage.

We consider that the judge has overlooked or discounted this critical aspect of the analysis because of two related errors in his reasoning. The first is that, as we have already noted, the judge simply accepted at [312]-[313] of the judgment MasterCard’s case that the default MIFs led to increased card usage, which in turn always conferred benefits on merchants. This assumption is inconsistent with the point made by the Commission at [695] of the Commission decision, which we cited at [85] above, reiterated at [730] and [732] (and at [60] of the Commission Survey) that it cannot be presumed that an increase in card usage will be beneficial, so that cogent evidence to that effect is required. This was emphasised forcefully by the following footnote 840 to [695], which the judge did not refer to:

“Again, it should be noted that an increase in system output does not constitute an objective efficiency if the benefits of increased card usage only accrue to banks, while customers and merchants are worse off due to higher retail prices and increased merchant fees. Hence, evoking the maximisation of system output also requires a convincing analysis that consumers benefit from this.”

Although the judge said at [312] that the increase in card usage and so in the benefits enjoyed by merchants from the charging of MIFs was “made good on the evidence before [him]”, it is entirely unclear to what evidence he was referring, let alone whether it was the cogent factual and empirical evidence which European law requires.

The second error in the judge’s reasoning was his rejection of the “always cards” point at [406] of his judgment. As Mr Turner pointed out, contrary to what the judge said, Mr Dryden’s “framework” did not assume that the only relevant benefit to merchants was not taking cash. It is clear from the expert evidence at trial that the socalled framework was on the basis of “switching” from other methods of payment to MasterCard cards, whether cash, store credit, PayPal or other card schemes such as Amex. Dr Niels accepted in cross-examination that this point would apply equally to comparators other than cash. In other words, there was no flaw in the framework as the judge thought.

Furthermore, even if there had been a flaw in the framework, the overall point remained a valid one, namely that the number of “switching” transactions (i.e. where cardholders are incentivised by issuers to use a scheme card rather than another payment method), on which (on this hypothesis) the merchant gained an advantage from the default MIF, is outweighed by the great preponderance of “always card” transactions (i.e. where the cardholder would always have used a scheme card, irrespective of incentives offered by issuers), on which the merchant bore the burden of the MIF without gaining an advantage. In rejecting Mr Dryden’s “framework” the judge lost sight of this overall point, which is a critical aspect of the balancing exercise required.

The fourth flaw in Popplewell J’s analysis is related to the criticism of his failure to carry out the balancing exercise required under article 101(3) because he rejected the “always cards” point. This concerns his treatment of pass-through. Mr Hoskins can legitimately point out that, in the trial before Popplewell J, unlike in the trial before Phillips J, it was not being contended by the AAM parties that the scheme had failed to establish pass-through at all. As the judge recorded at [400], Mr Dryden accepted that a significant proportion of MIF income was passed through to cardholders. His point and the AAM parties’ case was that, in circumstances where it was not 100% (or, as he put it in cross-examination, “materially incomplete”) and MasterCard could not establish by evidence the extent of pass-through, for the purposes of the first and second critical steps, the extent to which MIF revenue was used to incentivise card usage and did in fact stimulate additional card usage could not be established.

This inability to establish how much MIF revenue was passed through was said by the AAM parties to be fatal to MasterCard’s case for exemption. They submitted that MasterCard could not satisfy the requirement, pursuant to the balancing exercise, to show that the advantages to the relevant consumers caused by the default MIFs outweighed the disadvantages inherent in the restriction, specifically the fact that, on any view, a proportion of the MIF revenue was not passed through to cardholders but was retained by issuers and the fact that, on the preponderance of transactions where a scheme card would always have been used, the merchant bears the burden of the default MIF without it conferring any benefit.

We agree with this analysis. As we have already said, in dismissing the “always card” point on the basis of the flaws Popplewell J (incorrectly) identified in relation to alternative payment methods other than cash, the judge appears to have lost sight of the underlying validity of the argument that, in a mature card market such as the United Kingdom, from whatever payment method any “switching” took place, the relatively small incremental advantage to the merchant from such switched transactions is far outweighed by the preponderance of transactions on which a scheme card would always have been used anyway, where the merchant bears the burden of the MIF, with no corresponding advantage.

We consider that the judge should have concluded, by reference to this “always cards” point, that MasterCard could not establish, even on the basis of economic theory, that the extent of pass-through was such that the advantages thereby conferred outweighed the disadvantages to the relevant consumers. We do not consider that the various materials referred to by Mr Dryden, which the judge set out at [401], satisfy the requirement for cogent factual or empirical evidence of pass-through. We have already noted that the judge discounted the Commission’s econometric analysis on the basis that it was not a secure indication of pass-through. The Evans, Chang & Joyce analysis in the United States of inferring 80% pass-through from issuers’ share prices is merely another species of economic theory, inferences drawn by analysts and hardly empirical evidence.

The PwC research does contain empirical data but not about pass-through as such. It was a survey of the response of consumers to a proposed increase in card fees, concluding that a proposed 0.25% charge on all transactions using a credit card would lead to a 26% decrease in the use of such cards. The judge said of this survey: “This suggests that in the UK the MIF subsidisation of costs which would otherwise have to be recouped in card fees has a very significant impact, implying a high level of pass through”. That was the judge’s own inference. It does not seem to us that the inference is necessarily a valid one but, in any event, it is hardly evidence.

The judge should have concluded that, in the absence of any evidence as to the actual extent of the pass-through, MasterCard had failed to establish by robust analysis and cogent evidence, or otherwise, a sufficient causal link between the default MIFs and any net benefits, so that their claim for exemption under article 101(3) failed.

What the judge did instead was to seek to do the best he could on the exiguous evidence available, to arrive at what was no more than a “guesstimate” of the extent of issuer pass-through, which he then used to arrive at a further guesstimate of the extent to which the default MIFs were causative of a net benefit. He did so because, having started from the erroneous assumption that increased card usage always benefited the relevant consumers, he considered that he had to make some quantification of the extent of the pass-through and thus of the net benefits. On the contrary, the judge should have concluded, on the basis of the evidence before him, that the first condition of article 101(3), the benefits requirement, was not satisfied so that MasterCard had not established entitlement to an exemption under article 101(3).

In view of his conclusion that default MIFs led to increased card usage which was always beneficial to the relevant consumers, including the merchants, the judge did not need to return, at the end of his analysis in relation to the benefits requirement, to consider whether the second condition, the fair share requirement, was satisfied. In the AAM parties’ appeal, it is common ground that this requirement is looking at the position of the merchants and is only satisfied if the merchants were no worse off as a consequence of the restriction, so that unless they obtain greater benefits from the default MIF than the anti-competitive disadvantage it imposes upon them, the second condition will not be satisfied. In our judgment, had the judge carried out the necessary balancing exercise, he would inevitably have concluded that MasterCard could not satisfy the second condition either.

In the circumstances, the AAM parties’ appeal against the judgment of Popplewell J must be allowed.

It follows that it is not necessary to consider various detailed sub-grounds of appeal, such as that concerning the judge’s treatment of debit cards, as they are subsumed within the principal ground of appeal, on which the AAM parties succeed.

We will, however, deal briefly with one of those grounds, namely business stealing, since this was one of the matters which led the judge to his erroneous conclusion that increased card usage was always beneficial to merchants. The judge accepted the argument for MasterCard that merchants who accept MasterCard cards gain a competitive advantage over rivals because they gain sales from rivals who do not accept such cards and they avoid the loss of sales to MasterCard card-accepting rivals which would otherwise occur (see [316]-[317] of the judgment). He rejected the contrary argument for the AAM parties that the relevant pool of merchants to be considered was all merchants, not just those who take MasterCard cards, so that business stealing was a “zero sum game” (see [318]-[320] of the judgment).

We agree with Mr Turner that the judge’s analysis gives insufficient weight to the corollary of the supposed competitive advantage, which is that merchants who accept MasterCard cards take such cards and bear whatever level of MIF the scheme imposes for fear of otherwise losing business. The Commission and the General Court in MasterCard recognised this as a burden or predicament, not a benefit. The Commission said at [705]:

“Merchants will accept cards, in part, to attract customers from each other and this will increase the amount they will be prepared to pay to accept cards above a level determined solely from transactional benefits they obtain from accepting cards. Since some or most of the additional sales won by one will be taken from its competitors, the acceptance of cards may have little or no effect on total sales from the perspective of merchants as a whole.”

The same point was made at [222] of the General Court’s decision:

“As regards merchants, while an increase in the number of cards in circulation may increase the utility of the MasterCard system as far as they are concerned, it also has the effect of reducing their ability to constrain the level of the MIF and, therefore, of increasing the applicants' market power. It is reasonable to conclude that the risk of adverse effects on merchants' custom of a refusal to accept this method of payment, or of discrimination in that respect, is higher the greater the number of cards in circulation.”

We consider that this analysis is correct and that the acceptance of MasterCard cards vis-à-vis business rivals is a predicament rather than a benefit. It is also relevant in this context that the rationale for the development and adoption of the MIT was to eliminate from the MIF the so-called “business stealing” effect, namely what merchants would pay from fear of losing business to rivals if they did not accept cards: see the Commission Survey at [74]. Accordingly, it is not appropriate to give credit for business stealing as a benefit and add it to the level of the MIT MIF, as the judge has done.

Furthermore, the judge’s approach depends upon competition between merchants themselves but inter-merchant markets are not relevant markets for the purposes of article 101(3). The judge should have had regard to merchants as a whole and concluded that acceptance of MasterCard cards was not a competitive advantage but simply shifted value from one merchant to another.

As Phillips J succinctly put it at [7] of his second judgment:

“the fact that accepting a payment card enables Merchants to win business from competitors who do not accept that card (referred to as "Business Stealing") is a benefit for the accepting Merchants but not, in itself, for the economy as a whole: their competitors suffer an equal and opposite loss, achieving no more than transferring business from one to the other with no net gain.”

Visa did not seek to run the business stealing argument in that trial.

Even if there had been any force in Popplewell J’s analysis that “business stealing” was a competitive advantage, we do not see how the value to merchants which he attributed to it of 0.4% for credit cards and 0.2% for debit cards (a substantial proportion in each case of the overall MIF) was justified. At [393] the judge concluded that the differential between the MIF on credit cards and that on debit cards

We agree with Mr Turner that the effect of this approach is to attribute as the benefit to merchants of business stealing the premium issuers impose on merchants for credit cards as opposed to debit cards, once the value of credit functionality is taken out. On that basis, the greater the differential between the MIF which the issuers were able to impose in respect of credit cards than in respect of debit cards, the greater the value to be attributed to business stealing, which cannot possibly be an appropriate basis for assessing the value of a benefit.

In relation to debit cards, the judge adjusted his MIT MIF figures by 0.2% at [396] to reflect the value of the competitive advantage of taking such cards. This was, as the judge said, a rough and ready approach. The problem with the approach, quite apart from the fact that essentially it plucks a figure from the air, is that it is inconsistent with the evidence of Dr Niels who accepted that the percentage of merchants accepting MasterCard debit cards in the UK was very high, possibly approaching 100%, so that the number of sales which could be “stolen” from merchants who do not accept such cards is correspondingly very low. In those circumstances, a substantial uplift in the MIF of 0.2% seems to us not to be justifiable.

Finally, we make it clear that, even if we had thought that the judge was correct about business stealing, that would not have affected our overall conclusion that the judge’s approach was flawed for the reasons we have given.

We have summarised Phillips J’s essential reasoning in his second judgment in Part III of this judgment at [55]-[57] above. This second judgment dealt with the issue of exemption under article 101(3) on the hypothesis that the judge was wrong in his conclusion on article 101(1) in his first judgment. Accordingly, Visa’s challenge to the second judgment before this court is by way of its Amended Respondent’s Notice.

We have dealt elsewhere in this judgment with the criticisms levelled by Visa at the judge’s approach to the standard of proof under article 101(3). For the reasons we have given at [78]-[82] above in Part V, we consider that the judge’s approach to the standard of proof was unimpeachable and there is nothing in the criticisms. We deal below in the section on the quantum issues with the burden of proof in relation to loss. That leaves two aspects of the Respondent’s Notice:

Ms Rose submitted, in relation to the findings made by the judge at [43]-[50] of his judgment, that the judge’s statements on a number of occasions that Visa had produced no empirical or factual evidence or that Visa’s case was dependent upon economic theory alone, were wrong. Visa had produced such evidence, which the judge had overlooked.

She identified a number of areas of criticism in relation to the judge’s findings in [45] as to the 6 “channels” by which additional card usage might be stimulated by issuers, specifically in relation to rewards and innovation. Some of the material to which she referred the court was confidential, so that we will anonymise it for the purposes of this judgment. To the extent necessary, we summarise the material which Ms Rose showed us in the discussion and conclusions section below.

On the basis of that material, which she described as “the tip of an iceberg”, Ms Rose submitted that this was not a case where the judge had considered and evaluated all the evidence. He had overlooked or ignored a great deal of evidence.

In response, Mr Brealey referred to a number of other pieces of evidence which, he submitted, justified the judge’s overall conclusions on the evidence. For example, in relation to Contactless, there was evidence that the technology was already far advanced and that the merchants themselves invested significant sums, together with evidence that the MIF actually hindered Contactless, to which the judge was referring in the second half of [45(c)].

Mr Brealey referred, in particular, to the absence of any evidence from issuers, specifically any cost benefit analysis they had conducted. The judge had indicated in [43] the sort of evidence he would have expected and was critical of its absence in [46]. In that paragraph he accepted that there was some link between MIFs and cardholder rewards but said there was insufficient evidence as to its extent. In other words, he had insufficient evidence to conduct the balancing exercise. This was a point to which the judge returned in [49] and [51]. Mr Brealey submitted that the judge’s overall conclusion that Visa had not satisfied the first condition by robust analysis and cogent evidence was justified. It is well established that, for the purposes of article 101(3), the evidence must be of sufficient quality to enable the court to determine whether the alleged efficiency or benefit compensates for the harm resulting from the restriction. The judge was entitled to conclude Visa’s evidence was not.

Having considered the materials to which Ms Rose drew our attention, we are satisfied that (perhaps because of the length of time between the end of the trial and the second judgment) the judge has overlooked evidence which went beyond economic theory and that he was wrong to say that Visa had produced no empirical evidence or data or factual evidence to support its case on the benefits requirement and the causal link between the MIF and increased card usage.

We can state our reasons for this conclusion relatively briefly. First, in relation to the channel of rewards provided to cardholders by issuers with which the judge dealt at [45(a)] of his judgment, the judge referred to news reports collated by Dr Caffarra, Visa’s expert, demonstrating that, since the introduction of the Interchange Fee Regulation at the end of 2015 capped MIF rates for credit cards, 11 issuers had reduced the level of rewards they provided to cardholders. The judge said, however, in the first sentence: “Visa did not adduce direct empirical evidence that Issuers have reduced their reward programmes because of the lowering of MIF levels following the introduction of the [Interchange Fee Regulation]”.

In fact, as Ms Rose demonstrated, there were internal reports from three issuing banks in evidence before the judge. The first internal report was of a business unit committee meeting some two years before the Interchange Fee Regulation came into force, which referred to evidence from Australia and the United States where legislation reducing MIF rates had been introduced and the rewards to customers had reduced. The report made the point that with a lower capped MIF, customers who were previously profitable would become loss-making. One of the press reports produced by Dr Caffarra related to this bank and reported, three months before the Interchange Fee Regulation came into force, that it had halved rewards to customers in anticipation of the Interchange Fee Regulation.

The second internal report was from a consultant to another issuer in March 2015, articulating similar concerns about the impending regulation, and referred to what had happened when similar legislation had been introduced in Australia and the United States and suggested that the UK market response would be similar. The consultant made a bold proposal that the bank should not cut cardholder benefits in response to the Interchange Fee Regulation in order to seize market share, but that proposal was not accepted. Dr Caffarra produced a press report showing that that bank had cut its customer rewards by 60% and blamed the Interchange Fee Regulation.

The third report was a minute of a strategic management committee meeting of a third issuing bank in April 2014, making the same point that what was currently a profitable model would cease to be so if the Interchange Fee Regulation went ahead. The report expressed reluctance on the part of the bank to cut its rewards to customers, but the press report produced by Dr Caffarra showed that it nevertheless did so.

Visa also called evidence from senior Barclaycard personnel, Mr Gary Hoffman (former chief executive officer) and Mr Craig Evans (currently a senior executive), dealing with a number of the 6 channels. Mr Hoffman gave specific evidence about how, on past occasions when there were reductions in MIF rates, Barclaycard had cut its rewards to its customers in response. Mr Evans gave evidence from his experience about the correlation between the level of interchange fee income and the level of customer rewards.

It is clear that the judge has failed to deal with that factual evidence, which provided cogent support for what was said in the press reports. The first sentence of [45(a)] is, therefore, wrong and cannot stand. We consider that the judge was unduly dismissive of the press reports. Overall, the judge’s statement at the end of [45(a)] that this material (the press reports) is “the very best evidence Visa adduced to demonstrate a link between MIF levels and the extent to which issuers utilise the 6 “channels”” was incorrect.

The other channel on which Ms Rose focused was innovation, which included the development of Contactless. Phillips J had said the following at [45(c)]: “Gary Hoffman … and Mr Evans gave evidence that the revenue from Interchange Fees was a primary driver in the decisions of Barclaycard to pursue Contactless as a proposition for credit cards, but it is unclear why they were not equally motivated by the far larger revenue generated in the form of interest from revolvers and costs savings for Issuers”.

We agree with Ms Rose that that is an inadequate summary of their evidence, particularly that of Mr Hoffman who was at Barclaycard when it was a pioneer in the development of Contactless. In his witness statement and in cross-examination, he explained that the reason why it was initially developed for credit cards only was that credit cards generated revenue, by way of MIFs, to fund Contactless and MIF revenue was a crucial part of the business case for its innovation. Mr Evans gave similar evidence and, when it was put to him in cross-examination that the MIF was not needed to introduce contactless technology, he said that Contactless would not have been a profitable enterprise without the MIF. The judge did not deal with that factual evidence. It cannot be said that in [45(c)] he has considered it, weighed its significance and then rejected it.

As Ms Rose pointed out, there was also empirical evidence comparing the speed of roll-out of Contactless in the UK market as contrasted with other countries, specifically Australia, with lower or zero MIFs (and, in the case of the EFTPOS scheme in Australia, a negative MIF on debit cards) where contactless and certain online technology was not introduced until 2015. The judge did not refer to the evidence that, in other jurisdictions with lower MIFs, contactless cards and other technology took far longer to eventuate than in this country.

In relation to the development of e-commerce, the judge said this at [45(e)]:

“Visa contends that MIFs incentivise Issuers to strike the “right” balance between online security and the need for online purchases to be as “frictionless” as possible. However, Visa’s case in this regard is entirely theoretical: there was no concrete evidence whatsoever as to what steps Issuers would or would not have taken in the absence of the MIF or what the effect of those steps would be.”

The judge was wrong to characterise Visa’s case as entirely theoretical. As Ms Rose pointed out, there was evidence from Mr Hoffman about the ways in which MIFs had incentivised the development and promotion of e-commerce. There was empirical evidence from customer surveys about the extent to which customers would be deterred from making online purchases with cards by additional security online and similar evidence from retailers about customer behaviour. Both Mr Hoffman and Mr Evans gave evidence that Barclaycard would not have been prepared to invest in the online protection necessary for customers without MIF revenue. There was also evidence from Mr Sheedy, a senior Visa executive, as to what happened to ecommerce when MIFs were reduced: “[issuers] threw sand in the online process because they would suffer the cost of it, i.e. fraud, and get no benefits”.

Not all Ms Rose’s criticisms of the judge’s treatment of the evidence are valid and to that extent we agree with Mr Brealey. The core point she makes, however, does seem to us to be compelling, even on the limited material she was able to show us due to inevitable time constraints on what was a lengthy series of appeals. This is not a case where the judge has set out or summarised the factual, empirical evidence before him and said, for example, that he does not accept it or given reasons for discounting it. Rather, it is a case where, in a number of respects, he has overlooked empirical and factual evidence which was before the court, and made a number of general statements about the absence of any such evidence.

We have already referred to the judge’s description of the press reports produced by Dr Caffarra as “the very best evidence” of a causal link between the MIFs and stimulation of card usage. At [46], after he had dealt with the 6 channels, the judge said this:

“In summary, there is in my judgment a complete absence of evidence of a real, observable and measurable link between MIFs and actions taken by Issuers to stimulate card usage. The best material that has been adduced may support some relationship between decreases in credit card MIFs and decreased levels of rewards, but its existence is a matter of supposition, there being no attempt to rule out the possibility of other causes. Even if some link was sufficiently clear, its nature and extent is not.”

Mr Brealey is no doubt correct that, if all that this were referring to was the absence of direct evidence from the issuers of cost benefit analyses and the like, it would be hard to criticise the judge’s conclusion. The difficulty is that the judge appears to be making a sweeping statement about the complete absence of factual and empirical evidence as opposed to economic theory, which cannot be justified in the light of some of the evidence to which we have been referred by Ms Rose. In relation to issuer revenue, pass-through and always cards transactions, the judge was critical at [49] and [51] of the absence of empirical evidence and the quality of the expert and other evidence available, saying at [51]:

“Despite the volume of that evidence and the eminence of the experts, they have all ultimately engaged in (and invited me to undertake) an exercise which involves making sweeping assumptions and the broadest of estimates, many of them requiring, in the end little more than putting a finger in the air.”

Ms Rose referred us to a Note which she provided to the judge in closing submissions summarising Visa’s positive case on those issues and the expert and factual evidence which was adduced. It is not possible or appropriate for this court to evaluate the evidence referred to, but we are of the view that there is considerable force in her submission that the judge has failed to deal with that positive case and the evidence relied upon. It may be that, had he done so, he would still have reached the same conclusion, but it is impossible for this court to say that is inevitably so, and accordingly that is no answer to his failure to address the evidence in question.

In conclusion, we agree with Ms Rose that this is not a case, as in many appeals on issues of fact, where it is a sufficient answer for Mr Brealey to establish that there was evidence on which a reasonable judge could have made the findings the judge made or that there was evidence running in the opposite direction to the evidence on which Visa relied or that it was a matter for the judge to decide what weight to give to particular evidence. This was a case in which, as Ms Rose said, the judge did not weigh all the evidence. He ignored or mischaracterised it or said that there was no evidence on particular matters when there was.

In the circumstances, we consider that the only just course is to set aside the judge’s order and to remit the case for renewed consideration. The fact that, after such renewed consideration, the same conclusion might be reached is not relevant to the terms of the order we should make. We deal later in this judgment with the disposition issue, in relation to which we are anxious that, so far as possible, outstanding issues in the light of our determination of the appeals should be determined by one Court or tribunal, the obvious candidate being the CAT.

As we noted above, in its Amended Respondent’s Notice, Visa raised a further point, namely that the judge’s finding that, if the Visa default MIFs were a restriction of competition, they would not be exempt at any level was not open to him on the basis of the case run by Sainsbury’s. This is a reference to the fact, recorded at [9] of Phillips J’s first judgment, that in closing submissions at the trial Sainsbury’s accepted that a UK MIF of up to 0.2% for debit cards and 0.19% for credit cards would be lawful.

When this point was raised by Ms Rose orally on the last day of the hearing before us, Mr Brealey sought to answer it by saying that, in his closing submissions at trial, he had made clear to the judge that his primary case was that there was no exemptible level of MIF at all but his alternative case was that the 0.2% and 0.19% levels, to which we have referred, were lawful. Ms Rose pointed out that, as she had said to the judge in her closing submissions, Sainsbury’s had never advanced any pleaded case that only a zero MIF was lawful. Its pleading had accepted, after correction by its expert, that MIFs of 0.19% for debit cards and 0.17% for credit cards were lawful. In fact, after further correction in evidence by the expert, the figures were accepted in Sainsbury’s written closing submissions at trial to be 0.2% for debit cards and 0.19% for credit cards.

Those are the figures, which the judge recorded in [9] of his first judgment, that Sainsbury’s had accepted as lawful in its closing submissions. There was no appeal against that finding in the first judgment and we do not consider it to be open to Sainsbury’s to challenge it now. In his analysis in his second judgment, the judge seems to have forgotten what he had recorded in his first judgment as having been accepted by Sainsbury’s as lawful MIF levels.

After the draft judgment containing [297] to [299] above had been sent out to the parties for them to provide the Court with any typing corrections or suggested corrections of any obvious errors, the Court received detailed written submissions from Sainsbury’s which contended that it had not made the concession to which we have referred. It was contended by reference to passages in the written opening submissions and the written and oral closing submissions at trial of Sainsbury’s, together with other material, such as expert evidence, that Sainsbury’s had made it clear to the judge that its primary case was that no level of MIF was exemptible and that the case which accepted that a UK MIF of up to 0.2% for debit cards and 0.19% for credit cards would be lawful was an alternative case.

In those circumstances, we adjourned hand-down of the judgment, at least in part to deal with this issue and to afford Visa the opportunity to provide a substantive response. We have received written submissions on the issue from Visa and also heard oral submissions from both parties at a further hearing on 2 July 2018, which was held in private at the time to seek clarification on this issue, but not to allow new substantive points to be taken.

Ms Rose submitted that what we said in [298] and [299] above was entirely correct. As we have said at [298], she pointed out to the judge in her closing submissions at trial that Sainsbury’s had never pleaded a case that only a zero MIF was lawful, so that to the extent that Mr Brealey was submitting that only a zero MIF was lawful, that case was not open to him.

Having considered the parties’ submissions and the material to which our attention has been directed, it appears that there was a certain lack of clarity on the part of Sainsbury’s at trial as to the precise nature of its pleaded case. It is clear, however, that throughout Visa were taking the point that there was a concession in Sainsbury’s pleading. We consider that Phillips J clearly decided this point at [9] of his first judgment on the basis that Sainsbury’s pleadings made a concession and the pleadings were never amended. Sainsbury’s did not raise any issue with the judge as to what he had said in [9] of the first judgment between it being provided to the parties in draft and hand-down of the judgment. Furthermore, as we have said, there was no appeal by Sainsbury’s against that finding in the judgment.

In the circumstances, we do not consider that it should be open to Sainsbury’s to challenge that finding now or to seek to go behind it. On 18 June 2018, we received an application from Sainsbury’s to amend its Particulars of Claim to plead a positive case that only a zero MIF would have been lawful. We do not consider that Sainsbury’s should be entitled at this late stage to withdraw the concession made or to amend its pleadings.

Finally in relation to the article 101(3) issues, we note that the CAT did not deal with the issue of exemption under article 101(3) in any detail at all. Such limited analysis of whether the existing MasterCard MIFs were exempt by reference to the four conditions to be satisfied under article 101(3) as was undertaken at [288]-[289] of the CAT’s decision was tied to its finding that there would have been bilateral agreements as to interchange fees in the article 101(1) counterfactual. As we have concluded earlier in this judgment, that finding cannot stand. It follows that, although MasterCard did not pursue a ground of appeal on exemption, Sainsbury’s v MasterCard must be remitted to the CAT for reconsideration of the article 101(3) exemption issues in the light of our judgment.

There were only two quantum issues that remained at the hearing:

The issues raised by MasterCard on its appeal from the CAT in relation to the assessment of damages were not pursued by MasterCard at the hearing.

This issue is raised by the AAM parties’ appeal from Popplewell J and by Visa’s Respondent’s Notice in the appeal from the decision of Phillips J. The issue is, as we have said, whether, if the agreement is not exempt under article 101(3), the merchants nevertheless carry the burden of proving what MIF agreement, if any, would have

The merchants’ contention is that, once it has been established that the default MIF as set is illegal, the established loss is the full amount of the MIF element of the merchants’ service charge, and that, if MasterCard asserts that it could and would have charged a lawful lower MIF, then it is for MasterCard to prove such assertion.

Popplewell J rejected that contention. He distinguished between exemption under article 101(3) and exemptibility in connection with the assessment of damages. He held that the claimants do not establish that the extent of their loss is the full amount of the MIF merely by establishing that the default MIFs as set were unlawful. He said that, as the claim is for tortious damages, the principles of causation in tort apply. It is, he said, for the merchants to establish the extent of their loss by reference to the extent of the unlawfulness; and so it is for the merchants to establish as their measure of loss the difference between, on the one hand, what MasterCard could lawfully have charged by setting an exemptible MIF and, on the other hand, the amount of the MIFs actually charged.

Popplewell J acknowledged that, on this approach, the burden of proof would lie on MasterCard on each of the criteria when seeking to prove exemption but on the merchants when seeking to prove exemptibility for loss purposes. He said, however, that the principled allocation of the burden of proof should not be discarded merely because of apparent difficulties or anomalies in its application; and that it is not unheard of for one party to bear the burden of proof on a particular issue for one purpose and the other for another purpose.

Having regard to the burden of proof on the merchants to establish the extent of their recoverable loss, Popplewell J held that, when considering exemptibility for the assessment of damages, it was appropriate to increase by 10% the levels at which the default MIFs would be exempt under article 101(3). Accordingly, having concluded that below the following levels the default MIFs as set would be exempt under article 101(3) - UK Credit 1.01%, UK Debit 0.38%, Irish Credit 1%, Irish Debit 0.39%, EEA Credit 1.28%, EEA Debit 0.38% - rounding to the nearest basis point, the following levels of default MIF would be exemptible under article 101(3) for the purposes of calculating any damages claim - UK Credit 1.11%, UK Debit 0.42%, Irish Credit 1.10%, Irish Debit 0.43%, EEA Credit 1.41%, EEA Debit 0.42%.

It followed that, so far as Popplewell J was concerned, the default MIFs as set were below the exempt and exemptible levels, save for the EEA debit card MIF for the earliest part of the claim period prior to June 2008.

On these appeals MasterCard supports Popplewell J’s approach. MasterCard emphasises that, in relation to the exemptible level of the MIF, where the “broad axe” is relied upon by the merchants to estimate their loss, the court must err on the side of under compensation, and this is in contrast to the burden on the schemes under article 101(3). MasterCard says that it is always open to the merchants to show that no positive MIF would meet the criteria for exemption, for example, by showing that the MIF did not generate any relevant benefits or the benefits were insufficient to outweigh the costs of the MIF; and it will also be open to the merchants to put forward evidence and analysis indicating that an exemption could be justified within a range but only up to an identified maximum.

Phillips J disagreed with the approach of Popplewell J on this issue. He said that, having reached a decision, on the extensive evidence before him, as to what levels of MIFs could be shown by Visa to be exempt (if any), those levels were necessarily the same for both exemption and for the assessment of damages. He rejected the proposition that a percentage “discount” should be applied to the outcome based on an assessment of that evidence to reflect a theoretical difference in the burden of proof.

We agree with the conclusion of Phillips J on this issue. The correct analysis is to apply articles 101(1) and (3) in order to determine whether or not the default MIF, as charged, is in whole or in part unlawful, and then to assess damages on the unlawful amount or level as so determined.

We also agree with Phillips J that, in any event, as a matter of principle, the burden of proving any particular exemptible level of default MIF for the purpose of assessment of damages should lie on the scheme rather than the merchant. The burden of proving that some agreement, other than the actual agreement, would have been lawful should lie on the party putting forward that assertion. Otherwise a heavy burden would be placed on the merchants, incompatible with the enforcement of competition legislation through private claims in national courts: see Case C-295/04 Manfredi v Lloyd Adriatico SpA [2007] Bus LR at [89]-[91]. It would require the merchants to prove a complex negative, namely the highest level at which the MIF would be exempt. It would require the merchants to satisfy the court as to what the defendant could and would have done, that is to say something which on the face of it would be based on facts within the scheme’s knowledge.

That makes no sense in the context of the requirement that, in order to prove an exemption under article 101(3), the defendant is required to lead precise empirical evidence and cannot rely simply on economic theories, such evidence being within the defendant’s own possession. It would make no sense, in a case where the scheme cannot or in any event does not adduce evidence to establish any level of exemption under article 101(3), to impose the burden on the merchants, in order to recover damages for an undoubtedly unlawful default MIF charge, to have to establish the highest level at which the MIF would have been lawful. As Mr Turner submitted, requiring the merchants to prove the cut-off between lawful and unlawful MIFs operated by the schemes, based on theoretical defences which the merchants do not accept, is unrealistic and excessively difficult.

We agree with the merchants that the present cases are not ones where the so-called “broad axe” principle, as explained by Popplewell J at [307] of his judgment, applies: that is to say, where the court is compelled to use a broad brush in the absence of precision in the evidence as to the extent of the harm suffered by the claimant, and so the court should err on the side of under-compensation so as to reflect the uncertainty as to the loss actually suffered and to give the defendant the benefit of any doubts in the calculation. The cases cited by Popplewell J in support of that proposition, SPE International Limited v Professional Preparation Contractors (UK) [2002] EWHC 881(Ch) at [87], and Blayney (t/a) Aardvark Jewelry) v Clogau St David’s Gold Mines Ltd [2002] EWCA Civ 1007, [2003] FSR 19 at [31]-[34], were cases in which there was a lack of evidence on which the claimant could rely to prove loss. In cases such as those with which we are concerned, the analysis under articles 101(1) and (3) will show what unlawful amount has been charged by way of the default MIF.

In competition cases, pass-on arises where the direct purchaser passes on all or part of an unlawful overcharge to its own customers, the indirect purchasers. It is common ground that, under the jurisprudence of the CJEU and the common law, the cause of action for damages is then split between the direct purchaser and, to the extent that the unlawful charge has in fact been passed on, the indirect purchasers. The CAT regarded Sainsbury’s as an indirect purchaser, as regards the default MIF, but it is common ground that, for the purposes of the legal treatment of pass-on, Sainsbury’s is to be regarded as a direct purchaser. MasterCard claims that Sainsbury’s passed on to its customers at least part of the default MIF and, to that extent, any damages payable by MasterCard should be reduced.

The actual point in issue on this appeal is a narrow one, namely whether, as MasterCard contends, the CAT made inconsistent findings in holding, on the one hand, that MasterCard had failed to prove that any part of the MIF had been passed on to its customers and, on the other hand, that Sainsbury’s was entitled to compound interest on only 50% of the MIF because 50% had been passed on.

There was little dispute between counsel as to the applicable law.

The concept of pass-on is well established in EU law. The starting point is that persons harmed by breach of EU competition law have a right to compensation in the domestic courts of EU member states under domestic laws and rules which comply with EU principles of equivalence (not less favourable than those governing similar domestic actions) and effectiveness (do not render practically impossible or excessively difficult the exercise of rights conferred by EU law) but such domestic laws and rules may prevent unjust enrichment from over-compensation: Case C459/99 Courage Ltd v Crehan [2002] QB 507 at [25]-[26] and [29]-[30]. Where an unlawful charge has been borne not by the direct purchaser but by the customer of that purchaser, to whom the cost has been passed on, repayment of the full amount of the unlawful charge to the direct purchaser would amount to paying the direct purchaser twice over, and it is for the domestic courts to decide whether that would in all the circumstances amount to unjust enrichment: Case 199/82 Amministrazione Delle Finanze Dello Stato v San Giorgio Spa [1983] ECR 3595, [1985] 2 CMLR 658 (unlawful health inspection fees) at [12]-[14].

Whether or not the unlawful charge has been passed on is a question of fact, the burden of proving which lies on the defendant, who asserts it: Amministrazione Delle Finanze Dello Stato v San Giorgio Spa at [12]-[14]; Case C-192/95 Société Comateb v Directeur général des douanes et droits indirects [1997] ECR I-165, [1997] STC 1006 (unlawful dock dues) at [23] and [25]; Case C-147/01 Weber’s Wine World Handels GmbH v Abgabenberufungskommission Wien [2003] ECR I-11365, [2004] 1 CMLR 7 (beverages duty) at [93]-[96].

Damages which would reimburse the full amount of an unlawful charge will only amount to unjust enrichment of the claimant if there has been a direct passing on of the charge by the claimant to another person. The claimant will not be required to give credit for collateral advantages. The situation under consideration in Case C398/09) Lady & Kid v Skatteministeriet [2012] 1 CMLR 14 was that the Danish

Finally, the EU has legislated on pass-on in article 3 and chapter IV of Directive 2014/104/EU of the European Parliament and of the Council of 26 November 2014 (the “Damages Directive”). They reflect the above principles of EU law. The Damages Directive also makes clear that Member States must provide effective rules and procedures to enable indirect purchasers, as well as direct purchasers, to recover compensation for the harm caused by breaches of competition law.

Those principles of EU law are entirely consistent with common law principles of the assessment of damages and, in particular, mitigation.

In the well-known case of British Westinghouse Electric and Manufacturing Company Limited v Underground Electric Railways Company of London Ltd [1912]

Hodgson v Trapp [1989] 1 AC 807 was a case concerning damages in tort for personal injury. Lord Bridge, with whom all the other members of the appellate committee agreed, emphasised (at page 819) that damages are intended to be purely compensatory and that, where the damages claimed are essentially financial in character, the basic rule is that it is the net consequential loss and expense which the court must measure. He said that if, in consequence of the injuries sustained, the claimant had enjoyed receipts to which she would not otherwise have been entitled, prima facie, those receipts were to be set against the aggregate of the plaintiff’s losses and expenses in arriving at the measure of damages.

Sums received, which have diminished the loss, are only to be taken into account if there is a sufficiently close causative link between them and the wrong committed by the defendant. In Fulton Shipping Inc v Globalia Business Travel SAU [2017] UKSC

There are a few short points to add to those general principles. Firstly, the Damages Directive has been incorporated into our domestic legislation through schedule 8A of the 1998 Act, which came into force on 9 March 2017. Secondly, Parliament introduced a collective action procedure by the Consumer Rights Act 2015 (the “2015 Act”), which can facilitate a class action by indirect purchasers to whom an unlawful anti-competitive charge has been passed on. Those provisions, inserting new sections 47B-47D in the 1998 Act, came into force on 1 October 2015. Thirdly, the approach of the US Supreme Court in United Shoe Machinery Corporation v Hanover Shoe Inc

On the other hand, we accept Mr Hoskins’s submission that in each case it is a matter for the judge to decide whether, on the evidence before her or him, the defendant can show that there is a sufficiently close causal connection between an overcharge and an increase in the direct purchaser’s price. We see no reason why that increase should not be established by a combination of empirical fact and economic opinion evidence. It is not appropriate for us on these appeals to be more specific as to the nature and type of evidence capable of satisfying a trial judge that there is a sufficiently close causal connection.

The CAT stated (at [526]) that it had applied a broad axe in reaching its conclusion that Sainsbury’s was entitled, pursuant to Sempra Metals Ltd v IRC [2007] UKHL 34, [2008] 1 AC 561, to damages representing interest at a compounded rate on 50% of the overcharge. As we explain more fully below, that statement was in the context, not of establishing the recoverable amount of the overcharge as a matter of fact, but of an economic assessment as to the consequences for Sainsbury’s of the overcharge in the context of its claim for interest.

Fifthly and finally, we accept Mr Turner’s observation, made by reference to recital 40 of the Damages Directive, that passing-on in the form of an increase in the retail price might itself give rise to a loss of profit, which would need to be taken into account on the assessment of damages.

The actual ground of appeal from the CAT’s decision on the issue of pass-on is, as we have said, a very narrow one. MasterCard maintains that the CAT made two inconsistent findings on pass-on. On the one hand, the CAT concluded (at [484]) that MasterCard’s pass-on defence must fail as (1) no identifiable increase in retail price was established before the CAT, still less one that was causally connected with the UK MIF, and (2) MasterCard was not able to identify any purchaser or class of purchasers of Sainsbury’s to whom the overcharge had been passed who would be in a position to claim damages.

On the other hand, as we have already said, the CAT concluded (at [525]-[526]) that, for the purposes of Sainsbury’s claim to damages representing compound interest, such interest should be calculated on the basis that 50% of the UK MIF would have been passed on. MasterCard claims that, since that finding of the CAT has not been appealed by either party, it should be endorsed by this court as the only proper finding on pass-on binding on the parties.

Contrary to MasterCard’s contention, there is no inconsistency between those two findings. It is plain that, in reaching its conclusion at [484], the CAT applied the legal principles for establishing pass-on.

There was disagreement between Mr Hoskins and Mr Brealey as to whether the second point at [484] of the CAT’s decision - viz that MasterCard was not able to identify any purchaser or class of purchasers of Sainsbury’s to whom the overcharge had been passed – is a substantive point of law which must be satisfied in order to establish a pass-on and so distinct from the first point – viz that no identifiable increase in retail price was established. Although it is not necessary to resolve that issue on this appeal, we consider that it is not an essential condition for recovery: it would reflect the kind of policy decision which motivated the US Supreme Court in the Hanover Shoe case and is inconsistent with the principle that damages are compensatory rather than punitive. In any event, it is sufficient that MasterCard accepts on the appeal that the CAT was entitled to come to the conclusion that MasterCard failed to satisfy the CAT that there was no identifiable increase in the retail price attributable to the unlawful MIF.

It is equally plain that, in restricting compound interest on the basis that 50% of the UK MIF was passed on by Sainsbury’s, the CAT was making economic assumptions different from the legal principles. Indeed, that was expressly stated by the CAT at [525] when it said that:

“… We consider that a substantial amount of the UK MIF - 50% - would have been passed-on (albeit not in a manner which would have amounted to a ‘defence’ of pass-on, for the reasons given at paragraphs 484 to 485).”

The CAT had said at [484(4)] that, while the notion of passing on cost is a very familiar one to an economist, the legal definition of a passed-on cost differs from that of the economist in two respects. First, whereas an economist might well define passon more widely (i.e. to include cost savings and reduced expenditure), the pass-on defence is only concerned with identifiable increases in prices paid by the claimants’ customers. Secondly, the increase in price must be causally connected with the overcharge, and demonstrably so. Reflecting the view of an economist, the CAT was effectively saying at [525]-[526] that, when it comes to compound interest claimed as damages, the award of such interest should reflect the fact that a significant portion of the cost of the MIF was absorbed internally by Sainsbury’s by savings and the like.

Whether or not the CAT was entitled to limit compound interest by making those economic assumptions is not an issue in the appeal from the CAT. That would be a matter for Sainsbury’s to challenge and it has not done so.

We, therefore, reject this ground of MasterCard’s appeal from the decision of the CAT.

The correct counterfactual to test the restrictive effects of schemes like the MasterCard and Visa schemes before us was identified by the CJEU’s decision. It was “no default MIF” and a prohibition on ex post pricing (or a settlement at par rule).

Popplewell J was broadly right, therefore, to hold that the rules of the MasterCard scheme providing for a default MIF in the absence of bilateral interchange fees infringed article 101(1), and Phillips J was wrong to reach the contrary conclusion in relation to the Visa scheme. We do not discount the possibility that some evidence might conceivably enable other schemes to distinguish different MIFs from those upon which the CJEU was adjudicating. The death spiral argument is not relevant to this issue, because the article 101(1) question must be asked in relation to the acquiring market.

The CAT was wrong to decide that it was likely that bilateral interchange fees would be negotiated between issuers and acquirers in the counterfactual world. That decision, and its decision as to the level of the likely bilateral interchange fees, must be set aside.

Popplewell J erroneously concluded that the CJEU had disapproved the decision of the Court of First Instance in Metropole, which led him into the further error of concluding that the competitive effect of removal of the restriction on the particular scheme was relevant to the ancillary restraint doctrine. He should have concluded that all such issues of pro- or anti-competitive effect of the particular scheme were for article 101(3) and that the only question under the ancillary restraint doctrine was one of objective rather than subjective necessity. The right test was to ask whether a default MIF was essential to the survival of this type of main operation, namely a four-party card payment scheme, to which the clear answer was negative, so that the default MIF could not be justified under the ancillary restraint doctrine.

Popplewell J also adopted too stringent an approach to whether the two schemes were materially identical in holding that the AAM parties had to establish material identity in relation to matters relevant to article 101(3) and not just article 101(1). He should have concluded that the two schemes were materially identical so that the only realistic counterfactual was that, if one scheme could not impose a default MIF, the other scheme would have been similarly restrained.

Phillips J was correct to reject the death spiral argument for the reasons he gave in his first judgment.

Popplewell J’s approach to the exemption issue was flawed, primarily because he proceeded on the fallacious assumption that the default MIFs would lead to increased card usage, which was always beneficial to merchants. He failed to have regard to the “always card” point (to the effect that, in the mature UK market, most switching would be from one scheme card to another, and therefore of no overall benefit to merchants) and, accordingly, did not engage appropriately in the balancing exercise required by article 101(3). There was no cogent factual or empirical evidence of the extent of issuer pass-through, so that the judge should not have sought to estimate percentages of credit and debit card pass-through in order to assess exemptible levels of MIF. Rather, he should have concluded that MasterCard had failed to satisfy the first condition of article 101(3) so that its case for exemption failed.

Popplewell J’s analysis as to the meaning of “fair share” in the second condition was correct and in accordance with Commission’s decisional practice and EU law. The contrary interpretation of the condition by Phillips J was erroneous.

In reaching the conclusions which he did in his second judgment, Phillips J overlooked or ignored important factual and empirical evidence which was before him, so that the Visa case requires remission for renewed consideration of all the evidence.

In agreement with Phillips J and disagreement with Popplewell J, we hold that the merchants do not bear the burden of proving the lawful level of MIF. The correct analysis is to apply articles 101(1) and (3) in order to determine whether or not the default MIF, as charged, is in whole or in part unlawful, and then to assess damages on the unlawful amount or level as so determined.

The CAT was right not to have reduced Sainsbury’s damages for ‘pass-on’. There is no inconsistency between the CAT’s findings regarding pass-on at [484] and [525][526] of its decision which concerned loss and interest respectively.

Section 16(4) of the 2002 Act provides that “[t]he court [the High Court, a county court, the Court of Session or a sheriff court] may transfer to the Tribunal [the CAT], in accordance with rules of court, so much of any proceedings before it as relates to a claim to which section 47A of the 1998 Act applies”.

Section 47A of the 1998 Act, as amended by sub-paragraph 4(1) of Part 1 of Schedule 8 of the 2015 Act, provides as follows:

Sub-paragraph 4(2) of part 1 of schedule 8 of the 2015 Act provides that the amended section 47A applies with retrospective effect (i.e. to claims arising before the 2015 Act came into force on 1 October 2015). The current position, therefore, is that claims in respect of infringement decisions or alleged infringements of chapter I of the 1998 Act, chapter II of the 1998 Act, article 101 of the TFEU or article 102 of the TFEU (but not of articles 53 and 54 of the Agreement on the European Economic Area, since these fall outside the wording of section 47A) may be transferred to the CAT.

As it seems to us, such claims should in normal circumstances be transferred to the CAT. We say this because of the specialist nature and other advantages enjoyed by the CAT, which were appropriately summarised in Barling J’s transfer judgment, as follows:

“15 The 1998 Act recognised that competition law was an area which justified a specialist court to deal, not just with appeals in cases concerning public enforcement of the competition rules, but also with some private law claims for damages. One obvious feature of competition litigation is the almost ubiquitous presence of expert economic evidence, often of a complex and technical nature. Another common feature, related to the last one, is evidence as to the characteristics and dynamics of specific industries and markets. Mindful of these features, Parliament provided for the specialist competition tribunal to have a multi-disciplinary constitution. In this way panels have the potential to include not just lawyers but also, for example, distinguished economists, accountants or industry experts, selected for each case from the members appointed to the CAT by reason of their knowledge and experience in these areas. Expertise of this kind is of considerable assistance in understanding and resolving the difficult issues which are a common feature of competition litigation. This has long been recognised in the UK, the former Restrictive Practices Court having had a similar constitution. Although it is not impossible for a judge sitting on a case in the High Court to enlist the assistance of a court expert, this is relatively uncommon, and there are resource and other obstacles to the adoption of that course on more than very exceptional occasions.

16

Furthermore, CAT panels benefit from outstanding logistical and legal support provided by the CAT staff and legal assistants (“referendaires”). This is of particular value in lengthy and complex actions …

17

… the CAT has the best of both worlds, in that it is also able to tap into the expertise of the High Court in this field. For many years High Court judges of the Chancery Division have been appointed as CAT Chairmen, and have regularly sat in the CAT. In this way the CAT is in a position to draw on the assistance of experienced judges who have heard competition law cases in both the High Court and the CAT ….”

Where proceedings raise issues with which the CAT is permitted to deal under section 47A, but also raise other issues, it is possible under section 16(4) of the 2002 Act to transfer to the CAT only those issues with which it is permitted to deal. Whether or not this course is appropriate will depend on considerations specific to the particular proceedings, such as how important, and how easily separable from the other issues, the competition issues are. Where this course is not appropriate, the case should remain in the Competition List of the Business and Property Courts.

The appeals in AAM v MasterCard and in Sainsbury’s v Visa will be allowed on the article 101(1) issue, and MasterCard’s appeal on the bilateral interchange fees issue in the CAT case will also be allowed. The AAM parties’ appeal will be allowed in AAM v MasterCard on the ancillary restraint death spiral issue. We will make appropriate declarations in each of the three cases to the effect that the agreements are restrictive of competition infringing article 101(1).

The CAT case will be remitted to the CAT for reconsideration of the article 101(3) exemption issue and for the assessment of the quantum of the claim. Sainsbury’s v Visa will also need to be remitted for reconsideration of the article 101(3) issue and the assessment of the quantum of the claim. In relation to AAM v MasterCard, we consider below the question of what article 101(3) arguments can be addressed on remission.

On the final day of the hearing, we invited short written submissions from the parties on whether, if more than one of the present appeals were to be remitted for reconsideration, it would be sensible for them to be heard together by the CAT. It is not necessary to set out all the arguments made by the parties on that issue. We should, however, briefly address a submission by Visa that the appropriate course would be to remit the matter to a High Court judge, who could hear full argument by the parties as to appropriate forum, taking into account the content of this judgment, before making a decision. We disagree. It is, we think, entirely appropriate to deal with the matter in this judgment.

We need now, therefore, to consider two questions: (a) the tribunals to which each case should be remitted, and (b) the nature and extent of the necessary reconsideration in each case.

We have taken the clear view that each of the three cases which are the subject of the present appeals should be remitted to the CAT, in accordance with the principles set out above. The CAT will be able to use its specialist expertise to deal with all the remaining matters in accordance with the guidance contained in this judgment. In view of the inconsistencies that have resulted from the separate hearings that have taken place so far, we take the view that the three cases should, so far as possible, be heard by the same tribunal and at the same time. The tribunal in question should be chaired by a High Court Judge.

With regard to the nature and extent of the necessary reconsideration, both the CAT case and Sainsbury’s v Visa will go back to the CAT for reconsideration of article 101(3) exemption issues in accordance with this judgment. As part of this exercise, we consider that the CAT should give effect to the acceptance by Sainsbury’s at the trial in Sainsbury’s v Visa that MIF levels of 0.2% for debit cards and 0.19% for credit cards would be lawful.

The outstanding question is then whether, when AAM v MasterCard goes back to the CAT, it should be open to the CAT to reconsider the applicability of the article 101(3) exemption. We are conscious that we have held that Popplewell J ought to have concluded on the evidence that he heard that MasterCard’s claim for exemption under article 101(3) failed. Since, however, the schemes in the CAT case and in Sainsbury’s v Visa will have an opportunity to re-argue the article 101(3) exemption issue based on the principles set out in this judgment, we have considered whether MasterCard should have the same opportunity in AAM v MasterCard, even though it has not filed a Respondent’s Notice.

We take the view that, despite what we have said above, it is not certain that, had Popplewell J had the benefit of this judgment and thus been fully aware of the need for empirical data and facts in order to prove an exemption, MasterCard’s case on article 101(3) would have failed in its entirety. It is possible, bearing in mind the acceptance by Sainsbury’s and the CAT in the other two cases that there was a lawful level of MIF, that the judge would have found that there was some exemptible level of MIF, albeit a lower one than he in fact found. Altogether removing the article 101(3) issue from reconsideration could therefore result in an unjustified windfall for the AAM parties. It seems far more just to us that the issue should be reconsidered in all three cases, based on the same principles, by the same tribunal. There is no real injustice to the AAM parties in the course we propose, since the windfall to which we have referred would have arisen from the procedural mishap caused by the separation of three cases raising almost identical issues. If the CAT is now able to reach a consistent conclusion in all three cases on the exemption and quantum issues, that will produce a fair and just outcome for all the parties. It would be a triumph of form over substance if we were to hold that we were unable to reach a just solution simply as a result of a procedural accident.

We emphasise that the cases will be remitted for reconsideration and not for a retrial, so that it will not be open to any party to adduce fresh evidence before the CAT, save in respect of quantum in Sainsbury’s v Visa and AAM v MasterCard. We note in that context that there was no application to adduce fresh evidence before this court. We have, however, devoted some thought to the question whether the parties should be confined to the evidence that they adduced in each case, or entitled to rely on evidence that was adduced in the other two cases. We have ultimately preferred the latter course on the basis that it would be unsatisfactory, in view of our decision that the cases should all be heard by the same tribunal at the same time in order to ensure consistency, if arbitrary results were reached as a result of the parties in each case being unable to rely on generic evidence from the other two cases that was equally applicable to both the schemes and to all merchants. The CAT should be entitled to take such evidence into account in all three cases.

The Visa Europe Operating Regulations (the “Visa Scheme Rules”) apply to the Visa scheme as a whole. Regulation 7.1H provides that transactions must be settled at par, as follows:

“Reimbursement for Interchange Transactions

Each Issuer must pay the Acquirer the amount due for Transactions occurring with the use of a valid Card.”

Section 9.9 deals with MIFs, as follows:

“9.9

Interchange Reimbursement Fees

This Section 9.9 specifies the fees reimbursed by one Member to a Customer or vice versa to cover Interchange for International Transactions. These fees shall also apply to Visa Europe Transactions and Domestic Transactions where a Member's domestic operating regulations do not provide for an equivalent fee.

For the avoidance of doubt, no Interchange Reimbursement Fees applicable to International Transactions shall be applied, by default, to Visa Europe Transactions.

9.9A Merchant Transactions

For Transactions originating at a Merchant, an Acquirer reimburses the Issuer, or, where applicable, the issuer that is a Customer, an Interchange Reimbursement Fee for each Interchange Transaction. This fee is calculated as a percentage of net sales (Transaction Receipt totals less Credit Transaction Receipts).

9.9B Default Domestic Interchange Reimbursement Fee

For Visa Europe Transactions, the Interchange Reimbursement Fees as specified in this Section 9.9, serve as the default Interchange Reimbursement Fees for Domestic Transactions in Visa Europe countries where Multilateral Agreements and/or Private Agreements are not in place.

9.9C Domestic Interchange Reimbursement Fee Variances

The Visa Europe Board may, on request, establish country-specific default Interchange Reimbursement Fees for Domestic Transactions if the Members in that country are unable to reach agreement on appropriate default Interchange Reimbursement Fees for Domestic Transactions, or in other exceptional circumstances.”

Specific provisions for the UK domestic market are made in the Visa Operating Regulations for the UK and Gibraltar (the “Visa UK Rules”), Chapter 9 of which provides as follows:

“This Chapter 9 details Member-to-Member fees applicable to Domestic Transactions in the United Kingdom and Gibraltar … where these fees differ from the [Scheme Regulations] and in the absence of Private Agreements.”

The fees detailed in Chapter 9 are the UK MIFs. Different MIFs are set for different types of transactions, depending on factors such as whether the transaction is paid for with a credit or debit card, whether the cardholder is present in store, and whether the payment is “chip and pin” or contactless.

MasterCard’s rules are similar in nature. Although various versions have applied during the claim periods, the parties agree that the relevant rules have remained substantially the same throughout, and that we can refer to the 28 May 2015 version (the “MasterCard Scheme Rules”) for the purposes of the appeals.

Rule 8.2 concerns the settlement of transactions between issuers and acquirers, and provides as follows:

“8.2

Net Settlement

A Customer that uses the Interchange System for the authorization and clearing of Transactions is required to net settle in accordance with the Corporation’s settlement Standards. However, an Acquirer and an Issuer may, with respect to a particular Transaction, agree to settle directly between themselves pursuant to a bilateral agreement.

Standards describing net settlement and bilateral agreement rights and obligations are set forth in the Settlement Manual ….”

The effect of this rule is that, in the absence of bilateral agreements with acquirers, issuers are only allowed to make deductions from the settlement obligation which are permitted under the MasterCard Scheme Rules.

Interchange fees are one such permitted deduction. In this regard, the rules specific to the Europe region provide as follows:

“1.7.3.7 Interchange Fee Requirements

If a central Acquirer acquires an lntracountry Transaction, the following principles apply to the interchange fee:

Rules 8.3 and 8.4 provide as follows:

“8.3

Interchange and Service Fees

A Transaction settled between Customers gives rise to the payment of the appropriate interchange fee or service fee, as applicable. The Corporation has the right to establish default interchange fees and default service fees (hereafter referred to as “interchange fees,” “service fees,” or collectively, “fees”), it being understood that all such fees set by the Corporation apply only if there is no applicable bilateral interchange fee or service fee agreement between two Customers in place. The Corporation establishes all fees for Interregional Transactions and Intraregional Transactions, and may establish fees for Intracountry Transactions.

The Corporation will inform Customers, as applicable, of all fees it establishes and may periodically publish fee tables. Unless an applicable bilateral interchange fee or service fee agreement between two Customers is in place, any intraregional or interregional fees established by the Corporation are binding on all Customers.

8.4

Establishment of lntracountry Interchange and Service Fees

This rule is applicable only to lntracountry Transactions.

If intracountry interchange and service fees are not established by the Corporation, such fees may be established in one of two ways: by agreement of Customers in the country as set forth in Rule 8.4. 1, or by application of intraregional interchange and service fees to lntracountry Transactions as set forth in Rule 8.4.2. Such fees may also be established by bilateral agreement between two Customers as set forth in Rule 8.4.3.

For any Transaction that is subject to a bilateral agreement between two Customers, the interchange and service fees set forth in the bilateral agreement prevail.

For any Transaction that is not subject to a bilateral agreement between two Customers, the default intracountry fees established by the Corporation apply, or if none, the default intracountry fees established by Customers pursuant to these Rules apply, or if none, the intraregional fees apply, or if none, the interregional fees apply. Any multilateral Customer fee agreement must comply with all requirements set forth in Rule 8.4. 1. The Corporation reserves the right to determine if multiple bilateral agreements are deemed to be a multilateral agreement.”

The default MIF levels referred to above are published separately from the scheme rules. As is the case for Visa, they vary according to transaction type.

Articles 101(1) and 101(2) provide that:

“1.

The following shall be prohibited as incompatible with the internal market: all agreements between undertakings, decisions by associations of undertakings and concerted practices which may affect trade between Member States and which have as their object or effect the prevention, restriction or distortion of competition within the internal market, and in particular those which:

(a)

directly or indirectly fix purchase or selling prices or any other trading conditions;

(b)

limit or control production, markets, technical development or investment;

(c)

share markets or sources of supply;

(d)

apply dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage;

(e)

make the conclusion of contracts subject to acceptance by the other parties of supplementary obligations which, by their nature or according to commercial usage, have no connection with the subject of such contracts.

2.

Any agreements or decisions prohibited pursuant to this Article shall be automatically void.”

Although not evident from the above wording, it is well established in EU case law that a provision of an agreement which restricts competition does not infringe article 101(1) if it is objectively necessary for, and proportionate to, the implementation of the “main operation” of the agreement, provided that the main operation does not itself infringe article 101(1). This is known as the “objective necessity” or “ancillary restraint” doctrine. In the present case, the card schemes are the main operations, so the question is whether the MIFs are necessary for, and proportionate to, their implementation.

Article 101(3) provides another exemption to article 101(1), as follows:

“3.

The provisions of paragraph 1 may, however, be declared inapplicable in the case of:

-

any agreement or category of agreements between undertakings,

-

any decision or category of decisions by associations of undertakings,

-

any concerted practice or category of concerted practices,

which contributes to improving the production or distribution of goods or to promoting technical or economic progress, while allowing consumers a fair share of the resulting benefit, and which does not:

(a)

impose on the undertakings concerned restrictions which are not indispensible to the attainment of these objectives;

(b)

afford such undertakings the possibility of eliminating competition in respect of a substantial part of the products in question.”

The European Commission has published guidelines on the above provisions (referred to in the body of the judgment as the “Guidelines”), which provide insight into how it interprets them. Although not legally binding, all the parties relied on the Guidelines in argument. We therefore set out below the most important sections from them.

Most relevantly, [26]-[29] of these Guidelines provide as follows:

“26.

If a horizontal co-operation agreement does not restrict competition by object, it must be examined whether it has appreciable restrictive effects on competition. Account must be taken of both actual and potential effects. In other words, the agreement must at least be likely to have anti-competitive effects.

27.

For an agreement to have restrictive effects on competition within the meaning of Article 101(1) it must have, or be likely to have, an appreciable adverse impact on at least one of the parameters of competition on the market, such as price, output, product quality, product variety or innovation. Agreements can have such effects by appreciably reducing competition between the parties to the agreement or between any one of them and third parties. This means that the agreement must reduce the parties’ decisionmaking independence, either due to obligations contained in the agreement which regulate the market conduct of at least one of the parties or by influencing the market conduct of at least one of the parties by causing a change in its incentives.

28.

Restrictive effects on competition within the relevant market are likely to occur where it can be expected with a reasonable degree of probability that, due to the agreement, the parties would be able to profitably raise prices or reduce output, product quality, product variety or innovation. This will depend on several factors such as the nature and content of the agreement, the extent to which the parties individually or jointly have or obtain some degree of market power, and the extent to which the agreement contributes to the creation, maintenance or strengthening of that market power or allows the parties to exploit such market power.

29.

The assessment of whether a horizontal co-operation agreement has restrictive effects on competition within the meaning of Article 101(1) must be made in comparison to the actual legal and economic context in which competition would occur in the absence of the agreement with all of its alleged restrictions (that is to say, in the absence of the agreement as it stands (if already implemented) or as envisaged (if not yet implemented) at the time of assessment). Hence, in order to prove actual or potential restrictive effects on competition, it is necessary to take into account competition between the parties and competition from third parties, in particular actual or potential competition that would have existed in the absence of the agreement. This comparison does not take into account any potential efficiency gains generated by the agreement as these will only be assessed under Article 101(3).”

The intended purpose of these Guidelines is set out in their introductory paragraphs, as follows:

“4.

The present guidelines set out the Commission’s interpretation of the conditions for exception contained in [article 101(3)]. It thereby provides guidance on how it will apply [article 101] in individual cases. Although not binding on them, these guidelines also intend to give guidance to the courts and authorities of the Member States in their application of [article 101 (1) and (3)] of the Treaty.

5.

The guidelines establish an analytical framework for the application of [article 101(3)]. The purpose is to develop a methodology for the application of this Treaty provision. This methodology is based on the economic approach already introduced and developed in the guidelines on … horizontal co-operation agreements … The Commission will follow the present guidelines, which provide more detailed guidance on the application of the four conditions of [article 101(3)] than the guidelines on … horizontal co-operation agreements … also with regard to agreements covered by those guidelines.

6.

The standards set forth in the present guidelines must be applied in light of the circumstances specific to each case. This excludes a mechanical application. Each case must be assessed on its own facts and the guidelines must be applied reasonably and flexibly.”

[29] explains the objective necessity exemption to article 101(1) referred to above, as follows:

“… the concept of ancillary restraints covers any alleged restriction of competition which is directly related and necessary to the implementation of a main non-restrictive transaction and proportionate to it. If an agreement in its main parts, for instance a distribution agreement or a joint venture, does not have as its object or effect the restriction of competition, then restrictions, which are directly related to and necessary for the implementation of that transaction, also fall outside [article 101(1)]. These related restrictions are called ancillary restraints. A restriction is directly related to the main transaction if it is subordinate to the implementation of that transaction and is inseparably linked to it. The test of necessity implies that the restriction must be objectively necessary for the implementation of the main transaction and be proportionate to it …”.

[34] sets out the four conditions which must be satisfied in order to engage the article 101(3) exemption, as follows:

“The application of the exception rule of [article 101(3)] is subject to four cumulative conditions, two positive and two negative:

(a)

The agreement must contribute to improving the production or distribution of goods or contribute to promoting technical or economic progress,

(b)

Consumers must receive a fair share of the resulting benefits,

(c)

The restrictions must be indispensable to the attainment of these objectives, and finally

(d)

The agreement must not afford the parties the possibility of eliminating competition in respect of a substantial part of the products in question.

When these four conditions are fulfilled the agreement enhances competition within the relevant market, because it leads the undertakings concerned to offer cheaper or better products to consumers, compensating the latter for the adverse effects of the restrictions of competition.”

[43] addresses the question of which markets are relevant for each of the first two conditions, as follows:

“The assessment under [article 101(3)] of benefits flowing from restrictive agreements is in principle made within the confines of each relevant market to which the agreement relates. The Community competition rules have as their objective the protection of competition on the market and cannot be detached from this objective. Moreover, the condition that consumers must receive a fair share of the benefits implies in general that efficiencies generated by the restrictive agreement within a relevant market must be sufficient to outweigh the anti-competitive effects produced by the agreement within that same relevant market. Negative effects on consumers in one geographic market or product market cannot normally be balanced against and compensated by positive effects for consumers in another unrelated geographic market or product market. However, where two markets are related, efficiencies achieved on separate markets can be taken into account provided that the group of consumers affected by the restriction and benefiting from the efficiency gains are substantially the same. Indeed, in some cases only consumers in a downstream market are affected by the agreement in which case the impact of the agreement on such consumers must be assessed. This is for instance so in the case of purchasing agreements.”

Further guidance on the application of the first condition (the benefits requirement) is contained in the following paragraphs:

“49.

It follows from the case law of the Court of Justice that only objective benefits can be taken into account. This means that efficiencies are not assessed from the subjective point of view of the parties. Cost savings that arise from the mere exercise of market power by the parties cannot be taken into account. For instance, when companies agree to fix prices or share markets they reduce output and thereby production costs. Reduced competition may also lead to lower sales and marketing expenditures. Such cost reductions are a direct consequence of a reduction in output and value. The cost reductions in question do not produce any pro-competitive effects on the market. In particular, they do not lead to the creation of value through an integration of assets and activities. They merely allow the undertakings concerned to increase their profits and are therefore irrelevant from the point of view of [article 101(3)].

50.

The purpose of the first condition of [article 101(3)] is to define the types of efficiency gains that can be taken into account and be subject to the further tests of the second and third conditions of [article 101(3)]. The aim of the analysis is to ascertain what are the objective benefits created by the agreement and what is the economic importance of such efficiencies. Given that for [article 101(3)] to apply the pro-competitive effects flowing from the agreement must outweigh its anti-competitive effects, it is necessary to verify what is the link between the agreement and the claimed efficiencies and what is the value of these efficiencies.

51.

All efficiency claims must therefore be substantiated so that the following can be verified:

(a)

The nature of the claimed efficiencies;

(b)

The link between the agreement and the efficiencies; (c) The likelihood and magnitude of each claimed efficiency; and (d) How and when each claimed efficiency would be achieved.

52.

Letter (a) allows the decision-maker to verify whether the claimed efficiencies are objective in nature, cf. paragraph 49 above.

53.

Letter (b) allows the decision-maker to verify whether there is a sufficient causal link between the restrictive agreement and the claimed efficiencies. This condition normally requires that the efficiencies result from the economic activity that forms the object of the agreement. Such activities may, for example, take the form of distribution, licensing of technology, joint production or joint research and development. To the extent, however, that an agreement has wider efficiency enhancing effects within the relevant market, for example because it leads to a reduction in industry wide costs, these additional benefits are also taken into account.

54.

The causal link between the agreement and the claimed efficiencies must normally also be direct. Claims based on indirect effects are as a general rule too uncertain and too remote to be taken into account. A direct causal link exists for instance where a technology transfer agreement allows the licensees to produce new or improved products or a distribution agreement allows products to be distributed at lower cost or valuable services to be produced. An example of indirect effect would be a case where it is claimed that a restrictive agreement allows the undertakings concerned to increase their profits, enabling them to invest more in research and development to the ultimate benefit of consumers. While there may be a link between profit- ability and research and development, this link is generally not sufficiently direct to be taken into account in the context of [article 101(3)].

55.

Letters (c) and (d) allow the decision-maker to verify the value of the claimed efficiencies… Given that [article 101(1)] only applies in cases where the agreement has likely negative effects on competition and consumers (in the case of hardcore restrictions such effects are presumed) efficiency claims must be substantiated so that they can be verified. Unsubstantiated claims are rejected.

56.

In the case of claimed cost efficiencies the undertakings invoking the benefit of [article 101(3)] must as accurately as reasonably possible calculate or estimate the value of the efficiencies and describe in detail how the amount has been computed. They must also describe the method(s) by which the efficiencies have been or will be achieved. The data submitted must be verifiable so that there can be a sufficient degree of certainty that the efficiencies have materialised or are likely to materialise.

57.

In the case of claimed efficiencies in the form of new or improved products and other non-cost based efficiencies, the undertakings claiming the benefit of [article 101(3)] must describe and explain in detail what is the nature of the efficiencies and how and why they constitute an objective economic benefit.”

Further guidance on the application of the second condition (the fair share requirement) is contained in the following paragraphs:

“84.

The concept of ‘consumers’ encompasses all direct or indirect users of the products covered by the agreement, including producers that use the products as an input, wholesalers, retailers and final consumers, i.e. natural persons who are acting for purposes which can be regarded as outside their trade or profession. In other words, consumers within the meaning of [article 101(3)] are the customers of the parties to the agreement and subsequent purchasers. These customers can be undertakings as in the case of buyers of industrial machinery or an input for further processing or final consumers as for instance in the case of buyers of impulse ice-cream or bicycles.

85.

The concept of ‘fair share’ implies that the pass-on of benefits must at least compensate consumers for any actual or likely negative impact caused to them by the restriction of competition found under [article 101(1)]. In line with the overall objective of [article 101] to prevent anti-competitive agreements, the net effect of the agreement must at least be neutral from the point of view of those consumers directly or likely affected by the agreement. If such consumers are worse off following the agreement, the second condition of [article 101(3)] is not fulfilled. The positive effects of an agreement must be balanced against and compensate for its negative effects on consumers. When that is the case consumers are not harmed by the agreement. Moreover, society as a whole benefits where the efficiencies lead either to fewer resources being used to produce the output consumed or to the production of more valuable products and thus to a more efficient allocation of resources.

86.

It is not required that consumers receive a share of each and every efficiency gain identified under the first condition. It suffices that sufficient benefits are passed on to compensate for the negative effects of the restrictive agreement. In that case consumers obtain a fair share of the overall benefits. If a restrictive agreement is likely to lead to higher prices, consumers must be fully compensated through increased quality or other benefits. If not, the second condition of [article 101(3)] is not fulfilled.

87.

The decisive factor is the overall impact on consumers of the products within the relevant market and not the impact on individual members of this group of consumers. In some cases a certain period of time may be required before the efficiencies materialise. Until such time the agreement may have only negative effects. The fact that pass-on to the consumer occurs with a certain time lag does not in itself exclude the application of [article 101(3)]. However, the greater the time lag, the greater must be the efficiencies to compensate also for the loss to consumers during the period preceding the pass-on …

102.

Consumer pass-on can … take the form of qualitative efficiencies such as new and improved products, creating sufficient value for consumers to compensate for the anti- competitive effects of the agreement, including a price increase.

103.

Any such assessment necessarily requires value judgment. It is difficult to assign precise values to dynamic efficiencies of this nature. However, the fundamental objective of the assessment remains the same, namely to ascertain the overall impact of the agreement on the consumers within the relevant market. Undertakings claiming the benefit of [article 101(3)] must substantiate that consumers obtain countervailing benefits (see in this respect paragraphs 57 and 86 above).”

Guidance on the application of the third condition (the indispensability requirement) is contained in the following paragraphs:

“73.

According to the third condition of [article 101(3)] the restrictive agreement must not impose restrictions, which are not indispensable to the attainment of the efficiencies created by the agreement in question. This condition implies a two-fold test. First, the restrictive agreement as such must be reasonably necessary in order to achieve the efficiencies. Secondly, the individual restrictions of competition that flow from the agreement must also be reasonably necessary for the attainment of the efficiencies.

74.

In the context of the third condition of [article 101(3)] the decisive factor is whether or not the restrictive agreement and individual restrictions make it possible to perform the activity in question more efficiently than would likely have been the case in the absence of the agreement or the restriction concerned. The question is not whether in the absence of the restriction the agreement would not have been concluded, but whether more efficiencies are produced with the agreement or restriction than in the absence of the agreement or restriction.

75.

The first test contained in the third condition … requires that the efficiencies be specific to the agreement in question in the sense that there are no other economically practicable and less restrictive means of achieving the efficiencies. In making this latter assessment the market conditions and business realities facing the parties to the agreement must be taken into account. Undertakings invoking the benefit of [article 101(3)] are not required to consider hypothetical or theoretical alternatives. The Commission will not second-guess the business judgment of the parties. It will only intervene where it is reasonably clear that there are realistic and attainable alternatives. The parties must only explain and demonstrate why such seemingly realistic and significantly less restrictive alternatives to the agreement would be significantly less efficient.”

All the parties, and the judges below, agreed that the relevant provisions of the 1998 Act do not differ in substance from the corresponding provisions of article 101 TFEU.

Section 2 of the 1998 Act, which is the domestic counterpart to article 101(1) TFEU, relevantly provides as follows:

“Agreements … preventing, restricting or distorting competition.

(1)

Subject to section 3 [which is not relevant for present purposes], agreements between undertakings, decisions by associations of undertakings or concerted practices which—

(a)

may affect trade within the United Kingdom, and

(b)

have as their object or effect the prevention, restriction or distortion of competition within the United Kingdom,

are prohibited unless they are exempt in accordance with the provisions of this Part.

(2)

Subsection (1) applies, in particular, to agreements, decisions or practices

which—

(a)

directly or indirectly fix purchase or selling prices or any other trading conditions;

(b)

limit or control production, markets, technical development or investment;

(c)

share markets or sources of supply;

(d)

apply dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage;

(e)

make the conclusion of contracts subject to acceptance by the other parties of supplementary obligations which, by their nature or according to commercial usage, have no connection with the subject of such contracts.

(3)

Subsection (1) applies only if the agreement, decision or practice is, or is intended to be, implemented in the United Kingdom.

(4)

Any agreement or decision which is prohibited by subsection (1) is void.

(5)

A provision of this Part which is expressed to apply to, or in relation to, an agreement is to be read as applying equally to, or in relation to, a decision by an association of undertakings or a concerted practice (but with any necessary modifications).

…

(8)

The prohibition imposed by subsection (1) is referred to in this Act as “the Chapter I prohibition.”

Section 9, which is the domestic counterpart to article 101(3) TFEU, provides as follows:

“Exempt agreements

(1)

An agreement is exempt from the Chapter I prohibition if it—

(a)

contributes to—

(i)

improving production or distribution, or

(ii)

promoting technical or economic progress, while allowing consumers a fair share of the resulting benefit; and

(b)

does not—

(i)

impose on the undertakings concerned restrictions which are not indispensable to the attainment of those objectives; or

(ii)

afford the undertakings concerned the possibility of eliminating competition in respect of a substantial part of the products in question.

(2)

In any proceedings in which it is alleged that the Chapter I prohibition is being or has been infringed by an agreement, any undertaking or association of undertakings claiming the benefit of subsection (1) shall bear the burden of proving that the conditions of that subsection are satisfied.”

Section 60 sets out certain principles to be applied when determining questions under the above provisions, as follows:

“Principles to be applied in determining questions.

(1)

The purposes of this section is to ensure that so far as is possible (having regard to any relevant differences between the provisions concerned), questions arising under this Part in relation to competition within the United Kingdom are dealt with in a manner which is consistent with the treatment of corresponding questions arising in [EU] law in relation to competition within the [European Union].

(2)

At any time when the court determines a question arising under this Part, it must act (so far as is compatible with the provisions of this Part and whether or not it would otherwise be required to do so) with a view to securing that there is no inconsistency between—

(a)

the principles applied, and decision reached, by the court in determining that question; and

(b)

the principles laid down by the Treaty and the European Court, and any relevant decision of that Court, as applicable at that time in determining any corresponding question arising in [EU] law.

(3)

The court must, in addition, have regard to any relevant decision or statement of the Commission.

…

(5)

In subsections (2) and (3), “court” means any court or tribunal.

(6)

In subsections (2)(b) and (3), “decision” includes a decision as to—

(a)

the interpretation of any provision of [EU] law;

(b)

the civil liability of an undertaking for harm caused by its infringement of [EU] law.”

The Irish Act relates only to the AAM parties’ claims in respect of MasterCard’s Irish MIFs. The parties agree that its relevant provisions are the same in substance as article 101 TFEU.

The main provision is section 4, which relevantly provides as follows:

“(1)

Subject to the provisions of this section, all agreements between undertakings, decisions by associations of undertakings and concerted practices which have as their object or effect the prevention, restriction or distortion of competition in trade in any goods or services in the State or in any part of the State are prohibited and void, including in particular, without prejudice to the generality of this subsection, those which—

(a)

directly or indirectly fix purchase or selling prices or any other trading conditions,

(b)

limit or control production, markets, technical development or investment,

(c)

share markets or sources of supply,

(d)

apply dissimilar conditions to equivalent transactions with other trading parties thereby placing them at a competitive disadvantage,

(e)

make the conclusion of contracts subject to acceptance by the other parties of supplementary obligations which by their nature or according to commercial usage have no connection with the subject of such contracts.

(2)

An agreement, decision or concerted practice shall not be prohibited under subsection (1) if it complies with the conditions referred to in subsection (5) or falls within a category of agreements, decisions, or concerted practices the subject of a declaration for the time being in force under subsection (3).

…

(5)

The conditions mentioned in subsections (2) and (3) [subsection 3 not being relevant for present purposes] are that the agreement, decision or concerted practice or category of agreement, decision or concerted practice, having regard to all relevant market conditions, contributes to improving the production or distribution of goods or provision of services or to promoting technical or economic progress, while allowing consumers a fair share of the resulting benefit and does not—

(a)

impose on the undertakings concerned terms which are not indispensable to the attainment of those objectives,

(b)

afford undertakings the possibility of eliminating competition in respect of a substantial part of the products or services in question …”.

This decision concerned MasterCard’s MIFs for intra-EEA and Single Euro Payments Area (SEPA) transactions since 22 May 1992. It should be noted that MasterCard’s organisational structure changed during the relevant period: until the company’s Initial Public Offering (“IPO”) on 25 May 2006, its member banks had been grouped into 6 regions, with each regional board having the power to set MIFs for its region; after the IPO, responsibility for setting MIFs was transferred to MasterCard itself.

After setting out the procedural and factual background, the Commission considered the relevant market. It decided at [257]-[282] that there were three distinct product markets in play: the payment systems market, the issuing market and the acquiring market. The relevant one for the purposes of what is now article 101(1) was the acquiring market ([283]-[316]), which was national in scope rather than a single EU market ([317]-[329]).

The first issue in relation to article 101(1) was whether MasterCard remained an association of undertakings in the period following its IPO (the company having conceded that it was an association of undertakings prior to the IPO). The Commission held at [331]-[399] that it did, for reasons including that its global board still took decisions relating to the MIFs “virtually on behalf of” its member banks.

The next issue was whether the MIFs restricted competition within the meaning of article 101(1). At [401]-[407] the Commission considered it unnecessary to determine whether they were a restriction of competition by object, because it could be clearly established that they were a restriction by effect. When compared with a counterfactual of a settlement at par rule and prohibition on ex post pricing, the MIFs reduced price competition in the acquiring market by inflating the base on which acquiring banks set charges to merchants ([408]-[410]). Such a finding was consistent with the Commission’s previous case practice [412]. Further, it was supported by two separate quantitative analyses undertaken by the Commission, as well as a survey of more than 200 merchants concerning the impact of the MIFs on their negotiations with acquiring banks ([425]-[438]).

The Commission then rejected various arguments put by MasterCard as to why the MIFs did not restrict competition ([439]-[521]). One such argument was that the MIFs were similar to an excise tax because they were “a common identical cost … that does not influence price competition between acquirers in terms of determining the level of [merchants’ service charges]”. This argument was rejected for three reasons. First, it ignored the fact that the MIFs were decisions by an association of undertakings. Secondly, it would deprive article 101 of its effet utile. Finally, the MIFs not only created an artificial common cost for acquirers and thereby set a floor on the merchants’ service charge, but also eliminated an element of uncertainty for acquirers, who knew that their competitors all paid the same fees ([455]-[460]).

In relation to the ancillary restraint doctrine, the Commission rejected MasterCard’s argument that the MIFs were objectively necessary because they improved “system output”. Metropole and Gottrup-Klim made clear that the doctrine does not involve an assessment of whether the main operation would be less commercially successful in the absence of the restraint; such considerations fall under article 101(3) ([526][547]). Whilst some form of default settlement rule was necessary for open payment

In relation to the article 101(3) exemption, the Commission rejected MasterCard’s argument that the first condition was “undoubtedly” fulfilled because the MIFs were “a method of balancing the demand of cardholders and merchants”. The scheme had relied merely on economic theory, and had not produced empirical evidence sufficient to demonstrate that its theory held true in the real world. Such evidence was required in the light of European Central Bank statistics showing that schemes operating without MIFs still had relatively good system output. Requiring such evidence did not place an excessively high burden on MasterCard. Therefore, whilst the Commission accepted that the MasterCard scheme contributed to technical and economic progress, MasterCard had failed to demonstrate that such progress was causally linked to the MIFs, and the first condition of article 101(3) was not met ([670]-[733]).

In relation to the second article 101(3) condition, the Commission considered it necessary to show that all consumers received a fair share of any benefits generated by the MIFs. In a scheme where MIFs were paid by acquirers to issuers, this meant that the efficiencies must “in particular counterbalance the restrictive effects to the detriment of merchants”. MasterCard had failed to demonstrate that this was the case. Whilst merchants may have benefitted from enhanced network effects from the issuing side, this did not necessarily offset their losses sustained from paying inflated interchange fees. It followed that it was unnecessary to examine whether cardholders sufficiently benefitted from the MIFs, and the second condition was not satisfied ([739]-[747]).

The third article 101(3) condition was dealt with rather more briefly. MasterCard had not proven to the requisite standard that the MIFs were indispensable to the claimed efficiency benefits, particularly in view of the fact that several payment card schemes had successfully operated in Europe without MIFs ([748]-[752]).

Accordingly, the MasterCard MIFs were a restriction of competition under article 101(1), did not engage the ancillary restraint doctrine, and were not exempt under article 101(3). By way of remedy, the Commission required MasterCard to cease setting MIFs for intra-EEA and SEPA transactions (excluding on commercial credit and charge cards, in relation to which an investigation of possible efficiencies was ongoing) within 6 months of the decision. No fine was imposed, but the company would be subject to a daily periodic penalty payment of 3.5% of its daily revenue in the prior business year in the event of non-compliance ([753]-[776]).

After summarising the Commission’s decision and the key procedural steps which had followed it, the General Court addressed four pleas made by MasterCard in support of its application to annul the decision.

MasterCard’s first plea was that the Commission had erred in law in concluding that the MIFs restricted competition under article 101(1). The plea fell into two parts, the first of which related to restrictive effects and the second of which related to the ancillary restraint doctrine. The General Court dealt with the second part first. It rejected MasterCard’s complaint that the Commission should have held that the MIFs were an ancillary restraint because it would have been difficult to implement the main operation without them. Metropole had held that the only question was whether the restriction is necessary, and considerations relating to the competitive situation on the relevant market are not part of the analysis ([77]-[92]). MasterCard’s other arguments were also rejected, such as they fell to be considered, with the consequence that the second part of MasterCard’s first plea failed ([93]-[122]).

The General Court then turned to the first part of MasterCard’s first plea, which alleged various errors of assessment in the Commission’s analysis of the restrictive effects of the MIFs. It upheld the Commission’s counterfactual (i.e. no default MIF, with a prohibition on ex post pricing) on the basis that it did not render the scheme economically unviable. It explained that the Commission had referred in [460] to bilateral negotiations between issuers and acquirers merely in order to point out that interchange fees would cease to be charged in the counterfactual, and said that the Commission’s analysis was not manifestly incorrect ([129]-[134]). Further, the Commission was entitled to take into account the fact that competition between schemes resulted in upward pressure on MIFs ([135]-[141]). The General Court considered MasterCard’s argument that, because the MIFs operated as a cost common to all acquirers, they did not affect competition between acquirers, but merely the level of the merchants’ service charge. It held that the Commission had not erred in concluding that the MIFs restricted competition by setting a floor under the merchants’ service charge, and thereby limiting the pressure which merchants could exert on acquiring banks. The General Court rejected MasterCard’s challenges to the quantitative analyses and surveys which had underpinned the Commission’s findings ([142]-[147]). Nor had the Commission been wrong to find that there was a distinct acquiring market; any “system output” arguments concerning the inter-relationship of the issuing and acquiring markets fell to be addressed under article 101(3) ([148][182]). Finally, the General Court rejected various procedural complaints, with the result that MasterCard’s first plea was rejected in its entirety ([183]-[193]).

MasterCard’s second plea concerned article 101(3), and was again divided into two parts. The first part alleged that the Commission had imposed an excessively high evidential burden on it, and the second that the Commission had made various manifest errors of assessment. The General Court considered that it was not possible to examine the Commission’s approach in the abstract, and so dealt with both parts of the plea together ([193]-[198]). It began by upholding the Commission’s approach to the first article 101(3) condition. It was necessary to ask whether any benefits were specifically caused by the MIF, rather than the MasterCard system as a whole ([199][207]). The Commission was entitled to conclude, on the evidence before it, that MasterCard had failed to prove a sufficiently close link between the MIFs and the objective advantages to merchants of accepting MasterCard cards ([208]-[226]). In the absence of such proof, the Commission could not be criticised for failing to consider any benefits that the MIFs produced for cardholders ([227]-[229]). Finally, the Commission’s approach had not imposed an excessively high evidential burden on MasterCard. Indeed, it might be said that any such difficulty had resulted from the way MasterCard chose to put its case, which meant that it needed to prove a sufficiently close correlation between the MIF and the costs of providing issuing services, taking account of issuing banks’ card revenues from sources other than the

The General Court then considered and rejected MasterCard’s third plea, namely that the Commission had erred in characterising its payment scheme as an association of undertakings following the MasterCard IPO ([238]-[260]).

MasterCard’s fourth and final plea was that the Commission’s decision was vitiated by various procedural errors and errors of fact. This was rejected in its entirety. Accordingly, MasterCard’s appeal was dismissed ([261]-[302]).

In addition to MasterCard’s appeal, which requested that the General Court’s decision described above be set aside, there were cross-appeals before the CJEU from Royal Bank of Scotland and Lloyd’s TSB to the same effect ([1]-[2]).

After setting out the procedural background, and rejecting an objection by the Commission that the cross-appeals were inadmissible ([20]-[26]), the CJEU addressed the third plea in the main appeal, which was that the General Court had made various errors of law with regard to the admissibility of several annexes to the application at first instance. It held that this plea was founded on a misreading of the General Court’s judgment, and that the General Court had made no error of law ([27]-[47]).

The CJEU then considered the second plea in the main appeal, which alleged an error of law and/or inadequate reasoning by the General Court with regard to the question whether MasterCard constituted an association of undertakings following its IPO. This plea was also rejected ([48]-[77]).

The CJEU dealt with the first plea in the main appeal, which alleged an error of law and/or inadequate reasoning with respect to the General Court’s analysis of the ancillary restraint doctrine. The first part of this plea was an argument that the doctrine was engaged if the main operation would be difficult to run without the restraint. The CJEU rejected this argument on the basis that it was wrong in law (the requirement being that it would be impossible to run the main operation without the restraint) and would undermine the effectiveness of the article 101(1) prohibition ([86]-[95]). It then dealt with the second and third parts of the first plea, which sought to challenge the objective necessity counterfactual used by the Commission and upheld by the General Court. MasterCard’s central argument was that the counterfactual should not have included an ex post pricing rule, because such a rule would not have been adopted without regulatory intervention. The CJEU rejected this argument, considering that the objective necessity counterfactual is not limited to the situation that would arise in the absence of the restriction, and may extend to other realistic situations that might arise ([96]-[114]). The fourth and final part of the first plea, which contended that the General Court had failed to apply the required standard of judicial review, was also rejected, meaning that MasterCard’s appeal failed in its entirety ([115]-[121]).

The CJEU then turned to the first plea in the cross-appeals, which maintained that the General Court had erred in law in its assessment of restrictive effects under article 101(1). It was argued that the General Court had wrongly relied on the sole criterion of economic viability to justify including an ex post pricing rule in its counterfactual, and should have also asked whether such a rule would be likely to have occurred in the actual context. Although the CJEU accepted this contention as a matter of law,

The second plea in the cross-appeals related to article 101(3), and was expressed in three parts. The first part (relating to the evidential burden on MasterCard) was rejected as inadmissible, on the basis that it merely repeated the same arguments put to the General Court and did not seek to explain how that court had erred in evaluating those arguments ([200]-[219]). The second part alleged that the General Court had been wrong to focus solely on benefits to merchants, and to ignore benefits to cardholders, for the purposes of the first two conditions of article 101(3). In respect of the first condition, the CJEU said that, in the context of a two-sided scheme such as MasterCard’s, benefits to both merchants and cardholders should be taken into account. The General Court had not, however, ignored the latter, but had concluded that the evidence was insufficient to support MasterCard’s claimed efficiencies. Further, having properly concluded that there was no proof that the MIFs produced any appreciable benefits for merchants, it was unnecessary for the General Court to consider benefits to cardholders, because benefits to a market other than the one harmed by the restriction cannot “in themselves” compensate for the harm. The second part of the second plea was therefore rejected ([244]-[250]). The third and final part, to the extent it was admissible, made similar arguments to those summarised above, and was also rejected. Consequently, the cross-appeals were dismissed in their entirety ([251]-[259]).