England and Wales Court of Appeal (Civil Division) Decisions

IN THE COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE COMPETITION TRIBUNAL
Mr Justice Roth (President), Professor Colin Mayer CBE and Ms Clare Potter
[2017] CAT 16

		Royal Courts of Justice Strand, London, WC2A 2LL
		16 April 2019

Lord Justice Patten :

Introduction

1. This is the judgment of the Court.
2. On 19 December 2007 the EU Commission published a decision ("the EC Decision") concerning the setting of the multilateral interchange fee ("MIF") which is charged between banks in relation to transactions involving the use of Mastercard card. The EC Decision related only to the EEA MIF which applies when a card issued under the Mastercard scheme in one EEA Member State is used to purchase goods or services from a merchant in a different EEA Member State. But the determination by the Commission that the setting of the EEA MIF involved a breach of EU competition law extended beyond cross-border transactions to include some purely domestic transactions. In particular, the Commission found that the level of the EEA MIF had a direct effect on the amount of the service charge ("MSC") which merchants were required to pay to their own banks ("the Acquiring Bank") as a fee for operating the Mastercard system. It is common ground in these proceedings that the entire cost of the MIF (including UK MIFs in purely domestic transactions) was passed by the Acquiring Banks to the merchants as part of the MSC.
3. The Commission also considered that some part of the MIFs which had been passed on to merchants through the MSC was likely to have been passed on to consumers in the form of increased prices. The Commission did not carry out any detailed investigation as to the level of pass-on and recognised that it will have been affected by other market factors such as competition between merchants which may have led them (as they put it) to internalise the costs. The level of pass-on may also vary between different sectors of the market. But if and insofar as the cost of the MIFs was passed down to consumers in the form of increased prices (rather than some form of specific charge for the use of payment cards) then it will have affected all consumers who purchased goods or services during the relevant period from merchants who participated in the Mastercard scheme regardless of whether they used a payment card or cash to purchase the goods and services concerned.
4. On this basis the appellant, Mr Merricks as the proposed representative, has issued a collective proceedings claim form against three companies in the Mastercard group seeking an aggregate award of damages and interest totalling some £14.098bn. The claim is made on behalf of all individuals over the age of 16 who had been resident in the UK for a continuous period of at least 3 months and who between 22 May 1992 and 21 June 2008 purchased goods or services from businesses in the UK which accepted Mastercard. As explained earlier, the proposed class includes all purchasers from those businesses during the relevant period regardless of whether they used a Mastercard payment card to make the purchase. The class is estimated to include some 46.2m people.
5. The power to bring collective proceedings was introduced into the Competition Act 1998 ("the CA") by the provisions of the Consumer Rights Act 2015 ("CRA"). Sections 47B and 47C of the CA now provide:

"47B Collective proceedings before the Tribunal

(1)  Subject to the provisions of this Act and Tribunal rules, proceedings may be brought before the Tribunal combining two or more claims to which section 47A applies ("collective proceedings").

(2)  Collective proceedings must be commenced by a person who proposes to be the representative in those proceedings.

(3)  The following points apply in relation to claims in collective proceedings—

(a)  it is not a requirement that all of the claims should be against all of the defendants to the proceedings,

(b)  the proceedings may combine claims which have been made in proceedings under section 47A and claims which have not, and

(c)  a claim which has been made in proceedings under section 47A may be continued in collective proceedings only with the consent of the person who made that claim.

(4)  Collective proceedings may be continued only if the Tribunal makes a collective proceedings order.

(5)  The Tribunal may make a collective proceedings order only—

(a)  if it considers that the person who brought the proceedings is a person who, if the order were made, the Tribunal could authorise to act as the representative in those proceedings in accordance with subsection (8), and

(b)  in respect of claims which are eligible for inclusion in collective proceedings.

(6)  Claims are eligible for inclusion in collective proceedings only if the Tribunal considers that they raise the same, similar or related issues of fact or law and are suitable to be brought in collective proceedings.

(7)  A collective proceedings order must include the following matters—

(a)  authorisation of the person who brought the proceedings to act as the representative in those proceedings,

(b)  description of a class of persons whose claims are eligible for inclusion in the proceedings, and

(c)  specification of the proceedings as opt-in collective proceedings or opt-out collective proceedings (see subsections (10) and (11)).

(8)  The Tribunal may authorise a person to act as the representative in collective proceedings—

(a)  whether or not that person is a person falling within the class of persons described in the collective proceedings order for those proceedings (a "class member"), but

(b)  only if the Tribunal considers that it is just and reasonable for that person to act as a representative in those proceedings.

(9)  The Tribunal may vary or revoke a collective proceedings order at any time.

(10)  "Opt-in collective proceedings" are collective proceedings which are brought on behalf of each class member who opts in by notifying the representative, in a manner and by a time specified, that the claim should be included in the collective proceedings.

(11)  "Opt-out collective proceedings" are collective proceedings which are brought on behalf of each class member except—

(a)  any class member who opts out by notifying the representative, in a manner and by a time specified, that the claim should not be included in the collective proceedings, and

(b)  any class member who—

(i) is not domiciled in the United Kingdom at a time specified, and

(ii) does not, in a manner and by a time specified, opt in by notifying the representative that the claim should be included in the collective proceedings.

(12)  Where the Tribunal gives a judgment or makes an order in collective proceedings, the judgment or order is binding on all represented persons, except as otherwise specified.

(13)  The right to make a claim in collective proceedings does not affect the right to bring any other proceedings in respect of the claim.

(14)  In this section and in section 47C, "specified" means specified in a direction made by the Tribunal.

47C Collective proceedings: damages and costs

…

(2)  The Tribunal may make an award of damages in collective proceedings without undertaking an assessment of the amount of damages recoverable in respect of the claim of each represented person.

(3)  Where the Tribunal makes an award of damages in opt-out collective proceedings, the Tribunal must make an order providing for the damages to be paid on behalf of the represented persons to—

(a)   the representative, or

(b)   such person other than a represented person as the Tribunal thinks fit.

…

(5)  Subject to subsection (6), where the Tribunal makes an award of damages in opt-out collective proceedings, any damages not claimed by the represented persons within a specified period must be paid to the charity for the time being prescribed by order made by the Lord Chancellor under section 194(8) of the Legal Services Act 2007.

(6)  In a case within subsection (5) the Tribunal may order that all or part of any damages not claimed by the represented persons within a specified period is instead to be paid to the representative in respect of all or part of the costs or expenses incurred by the representative in connection with the proceedings.

(7)  The Secretary of State may by order amend subsection (5) so as to substitute a different charity for the one for the time being specified in that subsection.

…

(9)  In this section—

…

(b)  "damages" (except in the term "exemplary damages") includes any sum of money which may be awarded by the Tribunal in collective proceedings (other than costs or expenses);

…"

6. The type of claims which may be brought before the Competition Appeal Tribunal ("the CAT") (by combining two or more claims to which s.47A of the CA applies) include claims for damages arising out of an infringement decision or an alleged infringement of the prohibitions contained in ss.2 and 18 of the CA (anti-competitive agreements affecting trade within the UK and the abuse of a dominant position) and those imposed by Articles 101 and 102 of the Treaty on the Functioning of the European Union ("TFEU"). In the present case we are concerned with follow-on claims consequent on the EC Decision.
7. The full background to the collective claim was summarised by the CAT (Roth J; Professor Colin Mayer CBE and Clare Potter) in its decision of 21 July 2017 ([2017] CAT 16) as follows:

"8. Mastercard operates what is commonly known as a four party payment card scheme, since payments made under the scheme generally involve four parties: (1) a cardholder; (2) the cardholder's bank (known as the "Issuing Bank"); (3) a merchant; and (4) the merchant's bank (known as the "Acquiring Bank"). Issuing and Acquiring Banks are licensed by Mastercard. They must pay fees to Mastercard to participate in the scheme and comply with the Mastercard Scheme Rules.

9. The scheme operates on a contractual basis as between all four parties, and in addition Mastercard as the scheme operator, which may be represented diagrammatically as follows:



10. In order to pay for goods or services using Mastercard, the cardholder presents his or her card to the merchant. Details of the transaction are passed by the merchant to its Acquiring Bank, and then by the Acquiring Bank to the cardholder's Issuing Bank. In the case of credit cards, the Issuing Bank sends an invoice to the cardholder, typically on a monthly basis, and the cardholder either pays the whole of that invoice or takes advantage of further credit under the terms of his or her arrangement with the Issuing Bank. In the case of debit cards, the Issuing Bank deducts the amount chargeable to the cardholder for the transaction from the balance in the cardholder's account. In the meantime, the Issuing Bank transmits payment to the Acquiring Bank, less a transaction fee known as the interchange fee ("IF"). The Acquiring Bank in turn generally deducts the amount of the IF, along with a fee for its acquiring services, from the payment it makes to the merchant. The total deduction made by the Acquiring Bank from the amount paid to the merchant is called the merchant service charge ("MSC"). However, the IF accounts for the vast majority of the MSC.

11. The Issuing Bank and the Acquiring Bank may have bilaterally agreed the level of IF that will apply to transactions between them, or in some cases they may be the same bank. But except for those situations, the level of the fee defaults to one set by Mastercard. This default fee is known as the multilateral interchange fee: the MIF.

12. Different MIFs apply for different territories and card types. As to the territorial aspect, it is important for present purposes to note that:

(i) where a card issued in one EEA Member State is used at a merchant based in a different EEA Member State, a cross-border MIF applies. This is the EEA MIF referred to above which was the subject of the EC Decision;

(ii) where a card issued in the UK is used to pay a merchant based in the UK, the domestic UK MIF applies. We were told that around 95% of the value of the present claim is based on the UK MIF; and

(iii) outside of the EEA, where a card is used at a merchant based in a different global region from the Issuing Bank, for example if a US tourist uses a card issued by a US bank to make purchases in London, a different cross-border MIF applies.

13. As already mentioned, the EC Decision held that the setting of the EEA MIF by Mastercard constituted a decision of an association of undertakings. The EEA MIF was found, in effect, to set a minimum price which merchants had to pay to their Acquiring Bank for accepting Mastercard branded consumer credit and charge cards and Mastercard or Maestro branded debit cards. On that basis it had the effect of inflating the base on which Acquiring Banks set their MSC charged to merchants, thereby restricting competition between Acquiring Banks to the detriment of merchants (and subsequent purchasers). It was held that in the absence of the EEA MIF, the MSC set by Acquiring Banks would be lower both for cross-border transactions and for domestic transactions in those Member States where no separate domestic MIF had been agreed or where local banks had specifically agreed to adopt the EEA MIF. Further, some banks viewed the EEA MIF as a benchmark for setting domestic IFs. The EEA MIF was not objectively necessary, since a payment system such as Mastercard's could operate without a MIF. The EC Decision stated, at recital para 411:

"A further consequence of this restriction of price competition is that customers making purchases at merchants who accept payment cards are likely to have to bear some part of the cost of Mastercard's MIF irrespective of the form of payment the customers use. This is because depending on the competitive situation merchants may increase the price for all goods sold by a small margin rather than internalising the cost imposed on them by a MIF."

14. The infringement was found to last from 22 May 1992 until the date of the EC Decision (i.e., 19 December 2007), and Mastercard was directed to bring it to an end within six months.

15. Since the appeals before the European Courts against the EC Decision have been dismissed, that decision is binding on the Tribunal: sect 58A CA."

8. The collective proceedings claim form issued on 6 September 2016 sets out the relevant parts of the EC Decision and then pleads a claim for damages for breach of statutory duty based on the effect of the infringements on purchasers from merchants in the Mastercard system during the infringement period. In terms of causation, the appellant contends that in a lawful counterfactual world there would either have been no MIFs payable by the Acquiring Banks or the level of the interchange fees ("IFs") would have been set bilaterally at a lower level based on one of a number of alternative methods of calculation that would have been compliant with EU competition law. One suggested basis of calculation is the merchant indifference test ("MIT") which assumes a level of fee that the merchant would have been willing to pay on the basis of comparing the cost of use of a payment card with the cost of a non-card payment, most obviously cash.
9. The overcharge caused by the setting of the EEA MIF and its effect on domestic MIFs is alleged (in [106] of the claim form) to have been passed on in total via the MSC to businesses that accepted Mastercard. This, as we have said, is now common ground. In relation to pass-on to consumers, it is alleged that the overcharge was passed on in total (or near total) to all consumers in the proposed class regardless of whether they themselves engaged in cross-border or domestic transactions using Mastercard: see [108] of the claim form. Reliance is placed on findings made by the General Court of the European Union in relation to Mastercard's unsuccessful application to annul the EC Decision and to what is said to be Mastercard's position in a claim brought against it by Sainsbury's in relation to the MSC where the defendant's expert evidence was to the effect that the level of pass-on by Sainsbury's to consumers was closer to 100% than to 50%. In [110] of the claim form it is, however, pleaded that:

"The determination of the extent of pass-on to the represented class will be complex and will be a matter for expert evidence in light of any appropriate disclosure and factual evidence (as to which, the proposed class representative relies upon the relevant proposals set out in the collective proceedings litigation plan)."

10. The claim form seeks an aggregate award of damages. Any other basis of awarding damages is described in [120] as unworkable. In [46] the claim form addresses the issue of whether the claims are suitable for an aggregate award of damages in these terms:

"… An individual assessment of damages suffered by each member of the proposed class would be impracticable. For example, such assessment would require (i) the determination of the actual purchases of goods and/or services made by each member of the proposed class during the infringement period (see paragraph 94 below), by whatever means of payment (including cash), and (ii) the assessment of the extent to which each of the businesses from which those purchases were made passed on the higher charges resulting from the proposed Defendants' infringement of Article 101 TFEU. In fact, the only practicable way of proceeding is by way of an aggregate award of damages. The expert report on common issues explains at Section 5 how it is proposed that such an award will be calculated."

11. The claim form was supported by a witness statement from Mr Merricks (which set out his qualifications for being authorised to act as the representative of the class in the collective proceedings) and by an expert report prepared by Dr Cento Veljanovski and Mr David Dearman which addresses the calculation of an aggregate award of damages and, in particular, the issue of MSC pass-on. Subject to one point which is relevant only to the order which this Court may make on this appeal, no issue exists between the parties as to Mr Merricks' suitability to act as the class representative. But there are significant disputes about the level of pass-on of the MSC to consumers during the infringement period; the method and feasibility of being able to calculate that level even for the purpose of an aggregate award of damages in collective proceedings; and how any aggregate award of damages would be distributed between the members of the class in a way which adequately (or at all) reflects the loss which they have incurred as individuals as a result of the infringement.
12. The application for a collective proceedings order ("CPO") under s.47B(4) was made on the basis of this material. As appears from the provisions of 47B of the CA set out earlier, the jurisdiction of the CAT to make a CPO depends on it being satisfied that the proposed representative is someone whom it could authorise to act as the representative in collective proceedings and that the individual claims of the members of the proposed class are eligible for inclusion in such proceedings: see s.47B(5). Eligibility is further defined in s.47B(6) and requires the CAT to consider that the individual claims raise the same, similar or related issues of fact or law and are suitable to be brought in collective proceedings. These are matters for the CAT to assess and its decision as to whether the criteria are satisfied would only be open to challenge on legal grounds if it could be shown that it had misapplied the relevant legal test or had made a decision on the facts before it which no reasonable tribunal, properly directed as to the law, could have reached.
13. The procedure for commencing collective proceedings is also dealt with in the Competition Appeal Tribunal Rules 2015 (2015 No. 1648) made under s.15 of the Enterprise Act 2002 ("the Rules"). Rule 75 sets out the requirement for the proposed representative to file a collective proceedings claim form containing, inter alia, a description of the proposed class and a concise statement of the relevant facts and law. The relief sought must also be specified including (see Rule 75(3)(i)(i)):

"where applicable, an estimate of the amount claimed in damages, including whether an aggregate award of damages is sought, supported by an explanation of how that amount has been calculated;"

14. Once the collective proceedings claim form has been served an acknowledgement of service must be filed by the defendant but there is no requirement for the defendant to file a defence before the hearing of the application for a CPO: see Rule 76(11).
15. As provided in s.47B(5) the CAT may only make a CPO if the claims are eligible for inclusion in collective proceedings. Eligibility is defined in s.47B(6). These provisions are incorporated into Rule 77(1)(b) which directs eligibility to be determined in accordance with Rule 79. Rule 79(1)-(2) states:

"(1) The Tribunal may certify claims as eligible for inclusion in collective proceedings where, having regard to all the circumstances, it is satisfied by the proposed class representative that the claims sought to be included in the collective proceedings—

(a) are brought on behalf of an identifiable class of persons;

(b) raise common issues; and

(c) are suitable to be brought in collective proceedings.

(2) In determining whether the claims are suitable to be brought in collective proceedings for the purposes of paragraph (1)(c), the Tribunal shall take into account all matters it thinks fit, including—

(a) whether collective proceedings are an appropriate means for the fair and efficient resolution of the common issues;

(b) the costs and the benefits of continuing the collective proceedings;

(c) whether any separate proceedings making claims of the same or a similar nature have already been commenced by members of the class;

(d) the size and the nature of the class;

(e) whether it is possible to determine in respect of any person whether that person is or is not a member of the class;

(f) whether the claims are suitable for an aggregate award of damages; and

(g) the availability of alternative dispute resolution and any other means of resolving the dispute, including the availability of redress through voluntary schemes whether approved by the CMA under section 49C of the 1998 Act or otherwise."

16. Mr Hoskins QC for Mastercard placed considerable emphasis on the scope of the power conferred by Rule 79 and the need for the CAT to take into account all matters it thinks fit in deciding whether the claims are suitable for inclusion in the collective proceedings. He accepts that the reference to "common issues" in Rule 79(1)(b) must be construed as a reference to the type of issues described in s.47B(6) of the CA but it is for the CAT, he says, to be satisfied that the claims do raise common issues as defined and that they are suitable for inclusion in collective proceedings. This includes the assessment of whether the claims are suitable for an aggregate award of damages.
17. On 30 November 2016 Mastercard filed a response to the application setting out their grounds of opposition to the making of the order. This identified two main areas of objection to certification. They contended that an award of aggregate damages would be inimical to the compensatory nature of damages and impossible to assess on any reliable basis. In relation to the proposed method of distribution (which was that any award of damages should be distributed to all class members on an equal, per capita basis in respect of each year of the claim period in which they were members of the class), it was said that this would also fail to reflect the compensatory nature of damages and result in individuals receiving amounts which bore no reasonable relationship to their actual loss.
18. Mastercard's pleaded position in relation to the assessment of damages even in the form of an aggregate award was that it was necessary to show that the infringement had caused loss to each member of the class. Difficulties about individual assessment could be overcome by the adoption of a "sound imagination" and "the practice of a broad axe" which are references to a passage in the speech of Lord Shaw in Watson, Laidlaw & Co Ltd v Pott, Cassels and Williamson (1914) 31 RPC 104 at pages 117-118. But this does not mean that the assessments should be conducted on the basis of pure guesswork. The proposed representative, they say, must provide as much certainty and particularity as possible. Damages are not at large and in the court's discretion.
19. The argument that the assessment of damages should be conducted on a principled and compensatory basis is not in dispute as a general proposition. But the appellant relies strongly on s.47C(2) of the CA which enables the CAT to make an aggregate award of damages in collective proceedings "without undertaking an assessment of the amount of damages recoverable in respect of the claim of each represented person". Damages, he says, may therefore be assessed on a global or class-wide basis in terms of the amount of the overcharge passed on to consumers generally without the need to calculate how each member of the represented class was affected. It is this top-down method of calculation which formed the basis of the experts' report that was relied on in the application for certification.
20. The proposed representative's experts proposed a 3-step method of quantifying the loss from the overcharge which had been passed on to the represented class of claimants as a whole. In [5.1.2] of their report they acknowledged that this exercise will require the disclosure by Mastercard of key data and the exchange of witness evidence. The assessment of quantum which they had made was therefore a preliminary exercise. In summary, their proposed methodology was to:

(1) Quantify the total volume and value of all relevant Mastercard transactions accepted by businesses selling in the UK during the infringement period (referred to as the volume of commerce ("VOC"));

(2) Quantify the extent to which the VOC was subject to the overcharge in respect of Mastercard domestic or EEA MIFs; and

(3) Quantify the proportion of the overcharge that was passed on to the proposed class. This requires a determination of the level of MIF pass-on from Acquiring Banks to merchants as part of the MSC and the extent to which the MIF overcharge was passed on through retail prices to consumers.

21. In [5.4] the experts' report assumed a full pass-on of the overcharge to merchants as part of the MSC which is now common ground. But the more difficult and controversial issue is the level of pass-on to consumers. In Part 6 of the report the experts accept that the calculation of the scale of pass-on to consumers is complex:

"… and will require further investigation and analysis. We accept that MSC Pass-On is a matter for further evidence and analysis and that different views have been expressed in recent legal proceedings. However, we see no credible basis for the assertion that the MSC was not passed-on to consumers within the proposed class across the United Kingdom during what is a long Full Infringement Period.

6. 2 Our intended approach to assessing the MSC Pass-On

6.2.1 In our opinion, in the absence of evidence to the contrary, it is appropriate to assume a single, but not necessarily constant over time, weighted average MSC Pass-on rate across the United Kingdom economy. This approach is consistent with the approach adopted in the Mastercard and Visa Undertakings. The averaging of the MSC Pass-on rate takes account of any data limitations and the computational complexity of determining MSC Pass-On across the United Kingdom economy for over one and a half decades.

6.2.2 For the reasons set out in Section 6.3 below, the MSC Pass-On is likely to be high (50%-100%) and could have been fully passed-on.

6.2.3 We will undertake independent research to determine the most likely pass-on rates based on past research, studies and data. This will include, for example:

(a) market studies, competition authority decisions and other research as described in Section 6.3 below; and/or

(b) evidence and analysis filed by the different Businesses that are bringing similar damages claims against Mastercard. Assuming the MSC Pass-On rate is consistent across Businesses operating in the same sector, which we consider is a reasonable economic assumption at this preliminary stage, then, based on the evidence from those claims, it may be possible to estimate the MSC Pass-On across key sectors such as food and drink, clothing, household goods, motoring, entertainment, travel and other retailers. This covers approximately 70% of all payments processed with a card in the United Kingdom (see Appendix 3);

6.2.4 If MSC Pass-On rates are ultimately found to be significantly different for different sectors of the United Kingdom economy, then we may be able to calculate a weighted average MSC Pass-On rate (weighted by reference to the VOC and pass-on rate associated with each sector during each year of the Infringement Period). This approach will depend on the availability of evidence and whether that evidence relates to the same period as the Full Infringement Period.

6.2.5 We note that, whether we are quantifying the loss suffered by the proposed class as defined, or sub-groups of the proposed class, or even an individual consumer, the approach we will adopt would be the same. In other words, MSC Pass-On is a common issue amongst the proposed class.

6.3 Our current assumption

6.3.1 At this preliminary stage prior to disclosure and the exchange of witness evidence, our starting position is that during the Full infringement Period, it is likely that there was full pass-on of the MIF (including the Overcharge) to members of the proposed class."

22. The experts then refer to various matters to support their preliminary conclusions including: (1) the position taken by Mastercard and their experts in the litigation with Sainsbury's; (2) statements and conclusions by competition authorities and regulators in relation to Mastercard's MIF and the likely level of pass-on; and (3) statements by the British Retail Consortium and other studies about likely pass-on-rates. But the report acknowledges that none of these statements amount to conclusive evidence of MSC pass-on (6.3.4) and that the empirical evidence shows a variety of pass-on rates (6.3.6).
23. Mastercard in its Response sets out a number of reasons why it contends that this approach to assessing the quantum of an aggregate award will not bear any realistic relationship even to the loss suffered by the class as a whole. It disputes the likelihood of the MSC pass-on rate being uniform across all seven of the market sectors identified in the proposed representative's report and makes the point that pass-on rates may even vary between merchants operating within the same sectors having regard to differences in their size, level of profitability and method of trading. More generally (in [61d]) it contends that:

"There is no realistic prospect of obtaining sufficient evidence from a sufficient number of different merchants for the Full Infringement Period, which runs from 1992 until 2008 (i.e. between 8 and 24 years ago)."

24. On the proposed method of distribution, it takes issue with the suggested method which, as I mentioned earlier, is to distribute the damages awarded between the class on a per capita basis for each separate year of the infringement period. This (as the proposed representative concedes) will bear no relationship to the compensatory principle at least on an individual basis. This is said to be impermissible. "Distribution of money on such a mechanical basis has nothing to do with compensation": see [68] of the Response.
25. Some further guidance for applicants about what is required in relation to an application for a CPO is contained in the CAT's own Guide to Proceedings ("the Guide") published in October 2015 which has the status of a practice direction. In relation to applications for a CPO, the Guide states:

"6.13 The proposed class representative should send with the collective proceedings claim form any evidence relied on in support of the application for a CPO. That may include, for example, a witness statement by or on behalf of the proposed class representative addressing the considerations raised by Rules 78 and 79; and an expert's report regarding the way in which the common issues identified in the claim form may suitably be determined on a collective basis.

…

6.28 The Tribunal does not encourage requests for disclosure as part of the application for a CPO. However, where it appears that specific and limited disclosure or the supply of information (cf Rule 53(2)(d)) is necessary in order to determine whether the claims are suitable to be brought in collective proceedings (see Rule 79(1)), the Tribunal may direct that such disclosure or information be supplied prior to the approval hearing."

26. The certification hearing took place over three days in January 2017. No disclosure was ordered and Mastercard did not rely on any expert evidence of its own. But the proposed representative's experts were cross-examined at some length by counsel for Mastercard and were also questioned in some detail by members of the CAT. In a reserved judgment, the CPO was refused.
27. Having set out the relevant provisions of s.47B of the CA and Rule 79, the CAT in its judgment summarises the expert evidence in relation to the calculation of the VOC and the amount of the MIF overcharge based on a non-infringing counterfactual IF (including the use of different counterfactual IFs for credit cards and for debit cards). The CAT explains (at [35]) that the proposed representative's experts' favoured approach is to take an average MIF and an average counterfactual IF in order to calculate the overcharge percentage rate which was the approach applied by the CAT in Sainsbury's. But it also notes the point made in (6.2.4) of the report about the use of a weighted average pass-on rate should it appear (as Mastercard contends) that the pass-on rates differed during the infringement period between different sectors of the retail market. The report itself acknowledges that this methodology is dependent on the availability of evidence and in its judgment (see [42]-[44]) the CAT refer to some of the oral evidence given by Dr Veljanovski at the hearing when he commented on the potential sources of the data that might be available. Card expenditure for the 1998-2008 infringement period is broken down in a schedule to the report into 11 sectors and Dr Veljanovski accepted that the calculation of a weighted average rate of pass-on would be dependent on calculating the proportion of total expenditure attributable to different market sectors on the assumption that the rate of pass-on differed as between them, A detailed breakdown of card expenditure is available for 2005-2008 but not as yet for the earlier years. Reference was made by the experts to the possibility of extrapolating backwards and in some cases of having to make what were described as broad-brush estimates of the rate of pass-on.
28. On distribution the CAT noted the submission made by Mr Harris QC, for the proposed representative, that the issues about the method of distribution of any aggregate award were not relevant to the issue of certification. Distribution is dealt with in Rules 92-93 which apply once an aggregate award has been made. The CAT must then give directions for the assessment of "the amount that may be claimed by individual represented persons out of that award" and those directions may include a method or formula by which such amounts are to be quantified: see Rule 92(1) and (2)(a). But Mr Harris is recorded as having conceded that if no methodology could be devised which would produce a fair distribution of the award and therefore proper compensation then that was something which the CAT could take into account as a relevant matter under Rule 79(2). The transcript of the hearing confirms that Mr Harris did assent to that proposition when it was put to him by the President of the CAT. But the real issue is whether the distribution of an aggregate award must conform at least in broad terms to the loss which individual members of the represented class in fact suffered or whether it is permissible to distribute the award on a fair but not strictly compensatory basis having regard to the near impossibility of being able to calculate individual loss at this remove in time from the infringement period. The appellant's case is that the power under s.47(2) to make an aggregate award of damages without undertaking an assessment of the loss suffered and recoverable by each individual claimant would be rendered nugatory if distribution could only be conducted by reference to such individual loss. The vindication of the claims made in the collective proceedings is, he contends, achieved by the aggregate award and a defendant such as Mastercard has no real locus to challenge either the aggregate award or the application for a CPO on the basis of the proposed method of distribution. What will constitute a fair method of distribution will depend on the circumstances including the difficulty of proving individual loss and need not conform to a strict loss-based measure of compensation which would in this case be practically impossible.
29. The CAT refused the CPO for two principal reasons: a perceived lack of data to operate the proposed methodology for determining the level of pass-on of the overcharges to consumers and the absence of any plausible means of calculating the loss of individual claimants so as to devise an appropriate method of distributing any aggregate award of damages.
30. On the first issue of eligibility under Rule 79, it accepted that there were either common issues between individual claimants (or at least it was common ground) as to the impact of the EEA MIFs set by Mastercard on UK MIFs for both credit and debit cards; as to the amount by which those MIFs were higher than a counterfactual, non-infringing MIF; and as to the level of pass-on of the MIF overcharges to merchants in the form of the MSC. But the degree to which the overcharges were passed to consumers in the form of price increases and the amount which each individual claimant spent at those merchants were not in the view of the CAT common issues in the sense of being the same, similar or related issues of fact common to all individual claims in the collective proceedings as required by s.47B(6).
31. The experts had accepted that there would have been variations in the level of pass-on not only as between different classes of goods and services but also as between different retailers. But the CAT rejected the claimant's submission based on the defences in the Sainsbury's litigation that there had been an almost complete pass-on of the MSC by way of increased prices. The Sainsbury's action covered a different period (2006-2015) and in its judgment in that case the CAT held that Mastercard had failed to establish what level of pass-on had occurred in the form of higher prices and how otherwise the MIF costs had been absorbed by the retailer. Other current litigation by other supermarket chains was also regarded as at too preliminary a stage to be of any real assistance in analysing the prospect of the proposed representative establishing pass-on in the present case. Disclosure from third parties was considered as a possible source of information about the rate of pass-on but the CAT observed that it would be a very burdensome and highly expensive exercise.
32. The CAT then turned to consider the scope of published data as a relevant source of information. They said:

"75. The experts emphasised that there is a lot of published data and studies on the passing on of input costs, and on credit and debit card usage. However, that is precisely the material surveyed in the RBB Report quoted above (see para 51), which found it incomplete and difficult to interpret. We have no doubt that some sectors have been the subject of detailed study, but there is nothing before us to contradict the overall finding of the RBB Report. We note that no real attempt appears to have been made to consider what data are available for each of the broad sectors over the relevant period.

76. Overall, we recognise that the methodology put forward by the experts in their oral evidence, in response to the Tribunal's questioning, is considerably more sophisticated and nuanced than that set out, rather briefly, in their Experts' Report. We have to say that it is unfortunate that the written report did not set out and explain their approach in this way. Indeed, the statement in their Report that at this preliminary stage the experts' starting position was that "it is likely that there was full pass-on of the MIF (including the Overcharge) to members of the proposed class" is unsustainable and was not adhered to in their oral evidence.

77. We accept that in theory calculation of global loss through a weighted average pass-through, as explained in the evidence and as summarised above, is methodologically sound. But making every allowance for the need to estimate, extrapolate and adopt reasonable assumptions, to apply that method across virtually the entire UK retail sector over a period of 16 years is a hugely complex exercise requiring access to a wide range of data. We certainly would not expect that analysis to be carried out for the purpose of a CPO application, but a proper effort would have had to be made to determine whether it is practicable by ascertaining what data is reasonably available. Given the massive size of the claim, a difference of even 10% in the average pass-through rate makes a very substantial difference in financial terms.

78. Accordingly, applying the Microsoft test (para 58 above), we are unpersuaded on the material before us that there is sufficient data available for this methodology to be applied on a sufficiently sound basis. It follows that we are not satisfied, and indeed very much doubt, that the claims are suitable for an aggregate award of damages: see rule 79(2)(f)."

33. On distribution the CAT indicated that in its view it was only permissible to proceed to an assessment of loss on a top-down, class-wide basis if it was then possible to formulate a method of distribution which did involve a calculation of loss on an individual basis:

"79. If the total loss could be calculated in the aggregate manner discussed above, it is nonetheless necessary to consider how that would translate into determination of the level of individual loss. That is particularly important since, as we have pointed out, the proposed methodology does not really go to determination of a common issue to the individual claims, but in a sense circumvents the problem of an issue which is not common by seeking to go directly to determination of a total sum for all claims. Such an approach can only be permissible, in our view, if there is then a reasonable and practicable means of getting back to the calculation of individual compensation."

34. There is a discussion in the judgment of the obvious difficulties that would be involved in any such calculation (which are, as I have explained, conceded) and of the possibility of adopting some more broad-brush approach to that assessment. But the CAT records that the experts were agreed that the proposed method of per capita distribution bore no relationship to individual loss. They said:

"87. This cannot be dismissed as a "mere" question of distribution, to be addressed only after an aggregate award has been determined. First, it is largely because of the methodology of seeking to calculate the loss on a top-down, aggregate basis, and not on the basis of a common issue concerning loss suffered by each member (or most members) of the class, that the fundamental problem arises. As a result, if, hypothetically, a million people opted out of the proceedings, there would be no proper way of reducing the quantum of damages accordingly (and, conversely, of increasing it if a large number of people now domiciled outside the UK sought to opt in): it would simply lead to everyone in the class getting more (or less) money out of the total pot.

88. Secondly, even if it were possible to determine with some broad degree of accuracy the weighted average for pass-through and thus to estimate the aggregate loss for the class each year, it is the significance of the individual issues remaining which mean that it is impossible in this case to see how the payments to individuals could be determined on any reasonable basis. As we have explained above, there are three sets of issues which are relevant: individuals' levels of expenditure; the merchants from whom they purchased; and the mix of products which they purchased. There is no attempt to approximate for any of those in the way damages would be paid out. The governing principle of damages for breach of competition law is restoration of the claimants to the position they would have been in but for the breach. The restoration will often be imprecise and may have to be based on broad estimates. But this application for over 46 million claims to be pursued by collective proceedings would not result in damages being paid to those claimants in accordance with that governing principle at all.

89.  Accordingly, in our judgment, these claims are not suitable to be brought in collective proceedings as required by sect 47B(6) CA. It follows that the Tribunal cannot make a CPO in this case: sect 47B(5)(b) CA."

35. The appellant appeals against the decision of the CAT on three main grounds. He contends that they erred in law and adopted the wrong approach to the assessment of the evidence and the strength of the case on pass-on; that the CAT adopted the wrong test in relation to distribution, in particular in directing themselves that a method of distribution which did not in some way seek to link the distribution of an aggregate award of damages to individual loss was impermissible; and (partly for the same reasons) that the CAT adopted the wrong legal approach to its assessment of the degree to which the individual claims raise common issues. There is also an application by the appellant to adduce fresh evidence for the purposes of this appeal in the form of a supplemental expert report from Dr Veljanovski and Mr Dearman which updates their earlier report about the sources of relevant data and seeks to address the concerns expressed by the CAT in its judgment about there not being sufficient data available to operate the methodology proposed for determining the level of pass-on of the MIFs to consumers. The admission of this evidence is opposed by Mastercard on the basis that it does not satisfy the Ladd v Marshall conditions: in particular because the material could reasonably have been made available at the certification hearing. But it is also conceded by Mastercard that there was further relevant material relating to the availability of data which could have been placed before the CAT.

Pass-on

36. The first ground of appeal is in substance a challenge to the CAT's treatment of the claim for an aggregate award of damages and the proposed method of calculating it. Such an award can be made without requiring the CAT to undertake an assessment of the damages which each individual member of the class would be able to recover in separate proceedings: see s.47C(2). In principle, therefore, a top-down calculation of the level of pass-on of the MIFs to consumers as a whole is a permissible basis for the calculation and making of an aggregate award.
37. The CAT seems to have accepted this at least implicitly in its judgment at [77] whilst adhering to an analysis of individual loss both in relation to whether each individual claim raised a common issue or issues about pass-on and more generally in relation to distribution. The argument about the existence of common issues about loss may not, however, matter given the recognition (in [67]) that not all of the significant issues in the claim need be common issues.
38. The appellant's basic criticism of the CAT's approach to the evidence about pass-on is that they demanded too much and set too high a hurdle for the appellant at the certification stage. The criticism made at [77] of the experts' proposed methodology in terms of the data available is said to be inconsistent with the test in the Canadian Supreme Court decision in Pro-Sys Consultants Ltd v Microsoft Corp. [2013] SCC 57 ("Microsoft"), which the CAT purported to apply but instead applied a much stricter test which is hinted at in [57] of the judgment where the CAT refers to the need for the proposed representative:

"to do more than simply show that he has an arguable case on the pleadings, as if, for example, he was facing an application to strike out. Collective proceedings on an opt-out basis can bring great benefits, if successful, for the class members which those individuals (or small businesses) otherwise could never achieve; but like almost all substantial competition damages claims they can be very burdensome and expensive for defendants. The eligibility conditions set out in sect 47B(6) CA, and adumbrated in the CAT Rules, require the Tribunal to scrutinise an application for a CPO with particular care, to ensure that only appropriate cases go forward."

39. It is therefore convenient at this stage to look at some of the Canadian authorities that are relied on. In [58] of its judgment the CAT directed itself in accordance with the decision of the Supreme Court of Canada in Microsoft quoting what was said by Rothstein J at [118] of his judgment:

"… the expert methodology must be sufficiently credible or plausible to establish some basis in fact for the commonality requirement. This means that the methodology must offer a realistic prospect of establishing loss on a class-wide basis so that, if the overcharge is eventually established at the trial of the common issues, there is a means by which to demonstrate that it is common to the class (i.e. that passing on has occurred). The methodology cannot be purely theoretical or hypothetical, but must be grounded in the facts of the particular case in question. There must be some evidence of the availability of the data to which the methodology is to be applied."

40. Mr Hoskins submitted that the CAT was applying this passage in particular rather than adopting the Canadian jurisprudence on certification more generally. We doubt that but, even if that be so, it is evident both from [59] of its judgment and the way that the argument proceeded at the certification hearing that the CAT considered that the Canadian cases, as exemplified by the judgment in Microsoft, provide guidance as to the appropriate approach to be adopted in applications for certification under Rule 79 at least in relation to claims for aggregate damages. Clearly neither the CAT nor this Court is bound to apply those decisions but they were considered to provide perhaps the most useful and proximate model for the UK regime and the contrary has not been argued on this appeal. Our view is that the CAT was right to treat the Canadian jurisprudence on certification as informing the correct approach. Most of the provinces in Canada have enacted class proceedings legislation (not limited to competition cases) under which the Court may certify the proceedings as suitable for a class action where the claims or defences to the claims raise common issues and a class action would be the preferable procedure for the resolution of those issues: see e.g. the Ontario Class Proceedings Act 1992 s.5(1). The Court may also make an aggregate award of damages in such cases where "the aggregate or a part of the defendant's liability to some or all class members can reasonably be determined without proof by individual class members": see s.24(1)(c). The similarities between the Canadian and the UK regimes are obvious.
41. In Microsoft the Supreme Court of Canada was concerned with an application by the Microsoft defendants to set aside a certification order made under the equivalent provisions of the British Columbia Class Proceedings Act and to dismiss the action as disclosing no cause of action. The claims against Microsoft for overcharging were based on s.36 of the Canadian Competition Act and also on other alleged unlawful conduct including the torts of conspiracy and intentional interference with economic interests. The proposed class consisted of ultimate consumers of the Microsoft PC operating systems who had purchased the software from retailers or other re-sellers further up the supply chain. An order for certification had been made by the British Columbia Supreme Court but then set aside by the Court of Appeal. The Supreme Court of Canada allowed the appeal and re-instated the order. Much of the judgment is concerned with whether indirect purchasers from Microsoft who suffer the consequences of the overcharging only insofar as it is passed on down the retail chain have a cause of action against Microsoft either under the Canadian Competition Act or on some other basis. But the Supreme Court also considered the standard of proof required at the certification stage and the role of expert evidence in determining whether loss to the class members can be established on a class-wide basis. In giving the judgment of the Court, Rothstein J referred to what McLachlin CJ had said in the Court's decision in Hollick v. Toronto (City) 2001 SC 68 about the need for the class representative to show some basis in fact for each of the certification requirements rather than for the claim itself:

"[102]   … The "some basis in fact" standard does not require that the court resolve conflicting facts and evidence at the certification stage. Rather, it reflects the fact that at the certification stage "the court is ill-equipped to resolve conflicts in the evidence or to engage in the finely calibrated assessments of evidentiary weight" (Cloud, at para. 50; Irving Paper Ltd. v. Atofina Chemicals Inc. (2009), 99 O.R. (3d) 358 (S.C.J.), at para. 119, citing Hague v. Liberty Mutual Insurance Co. (2004), 13 C.P.C. (6th) 1 (Ont. S.C.J.)). The certification stage does not involve an assessment of the merits of the claim and is not intended to be a pronouncement on the viability or strength of the action; "rather, it focuses on the form of the action in order to determine whether the action can appropriately go forward as a class proceeding" (Infineon, at para. 65).

[103]   Nevertheless, it has been well over a decade since Hollick was decided, and it is worth reaffirming the importance of certification as a meaningful screening device. The standard for assessing evidence at certification does not give rise to "a determination of the merits of the proceeding" (CPA, s. 5(7)); nor does it involve such a superficial level of analysis into the sufficiency of the evidence that it would amount to nothing more than symbolic scrutiny.

[104]   In any event, in my respectful opinion, there is limited utility in attempting to define "some basis in fact" in the abstract. Each case must be decided on its own facts. There must be sufficient facts to satisfy the applications judge that the conditions for certification have been met to a degree that should allow the matter to proceed on a class basis without foundering at the merits stage by reason of the requirements of s. 4(1) of the CPA not having been met."

42. Section 4(1)(c) of the Canadian Competition Act requires the claims of the class members to raise common issues which are defined as meaning "common but not necessarily identical issues of fact". Part of Microsoft's argument in the Supreme Court was that the alleged instances of overcharging occurred over too long a period (24 years) and involved 19 different products so that it was impossible to prove loss on the part of each member of the proposed class. But Rothstein J rejected this as an obstacle to certification:

"[110]   The multitude of variables involved in indirect purchaser actions may well present a significant challenge at the merits stage. However, there would appear to be a number of common issues that are identifiable. In order to establish commonality, evidence that the acts alleged actually occurred is not required. Rather, the factual evidence required at this stage goes only to establishing whether these questions are common to all the class members.

[111]    Myers J. concluded that the claims raised common issues.  I agree that their resolution is indeed necessary to the resolution of the claims of each class member. Their resolution would appear to advance the claims of the entire class and to answer them commonly will avoid duplication in legal and factual analysis. Those findings are entitled to deference from an appellate court."

43. We have already quoted what the Court said (at [118]) about the quality of the expert evidence needed to establish loss on a class-wide basis as an issue common to the class. The Court recognised that in indirect purchaser actions the ability to treat the loss caused to consumers as a class as a common issue was dependent on the availability of an economic model and methodology that was capable of making that global (and therefore common) assessment. But the Court added:

"115. … The requirement at the certification stage is not that the methodology quantify the damages in question; rather, the critical element that the methodology must establish is the ability to prove "common impact", as described in the U.S. antitrust case of In Re: Linerboard Antitrust Litigation, 305 F.3d 145 (3rd Cir. 2002). That is, plaintiffs must demonstrate that "sufficient proof [is] available, for use at trial, to prove antitrust impact common to all the members of the class" (ibid., at p. 155). It is not necessary at the certification stage that the methodology establish the actual loss to the class, as long as the plaintiff has demonstrated that there is a methodology capable of doing so. In indirect purchaser actions, this means that the methodology must be able to establish that the overcharges have been passed on to the indirect-purchaser level in the distribution chain."

44. The question of what constitutes "some evidence of the availability of the data" necessary to operate the methodology is not taken further in the Canadian authorities we were referred to. In Fairhurst v. Anglo American PLC, 2014 BCSC 2270 (a first instance decision of the British Columbia Supreme Court) Brown J (at [64]) said that the Court was not in the position at the certification stage to engage in a detailed analysis of expert opinion and that the evidentiary threshold for certification is not onerous. Mr Harris also emphasised that in Microsoft the Supreme Court was prepared to infer the existence of the necessary data from the statement of the claimant's expert that he could see no theoretical reason why the methodology he proposed could not be applied to sales of Microsoft software in Canada. It seems to us that at the certification stage the proposed representative must be able to demonstrate that the claim has a real prospect of success. To do so in this case he had to satisfy the CAT that the expert methodology was capable of assessing the level of pass-on to the represented class and that there was, or was likely to be, data available to operate that methodology. But it was not necessary at that stage for the proposed representative to be able to produce all of that evidence, still less to enter into a detailed debate about its probative value. To that extent a certification hearing is no different from any other interlocutory assessment of the prospects of success in litigation made before the completion of disclosure and the filing of evidence. Its purpose is to enable the CAT to be satisfied that (with the necessary evidence) the claims are suitable to proceed on a collective basis and that they raise the same, similar or related issues of fact or law: not that the claims are certain to succeed. The specific considerations relevant to suitability which are set out in Rule 79(2) do not call for a different approach. None of them requires the CAT to be satisfied that the collective claim has more than a real prospect of success.
45. A critical issue in deciding whether the proposed methodology is a suitable and effective means of calculating loss to the class is to determine whether it is necessary to prove at trial that each member of the proposed class has in fact suffered some loss due to the alleged infringement. Although the expert evidence must obviously provide a means of calculating the level of pass-on of the MIFs from merchants to consumers via price, there is some controversy as to whether that is sufficient to make the global loss suffered by consumers a common issue absent being able to show that each member of the class was in some way adversely affected in their own purchases during the infringement period.
46. The CAT has, as I have said, avoided that controversy in this case by accepting that not all of the issues need to be common issues in order for the collective claim to be certified. But, as indicated above, there is no requirement under s.47C(2) to approach the assessment of an aggregate award through the medium of a calculation of individual loss and the appellant's experts have not attempted to do so. In that they have the support of the Canadian authorities which in cases like Microsoft have approved a top-down method of calculation on the basis that the level of pass-on to the class as a whole will be a common issue for all individual claimants. It seems to us that the same approach should be adopted in relation to collective proceedings under s.47B of the CA. Although the purpose of such proceedings is to combine into a single action what would otherwise be individual follow-on claims and that to be eligible for inclusion in collective proceedings they must raise common issues of fact or law (see s.47B(6)), the issue of whether the MIF overcharge was passed-on to consumers generally and in what amounts is an issue common to all such individual claims as a necessary step in establishing loss by the class as a whole.
47. To require each individual claimant to establish loss in relation to his or her own spending and therefore to base eligibility under Rule 79 on a comparison of each individual claim would, as I have said, run counter to the provisions of s.47C(2) and require an analysis of the pass-on to individual consumers at a detailed individual level which is unnecessary when what is claimed is an aggregate award. Pass-on to consumers generally satisfies the test of commonality of issue necessary for certification.
48. Leaving aside for the moment whether this controversy should be allowed to re-emerge in some form in relation to the issue of distribution, the first ground of appeal therefore really centres on whether the CAT demanded too much of the proposed representative at the certification stage. In our view it did.
49. The CAT refused to certify because it was "unpersuaded on the material before us that there is sufficient data available for this methodology to be applied on a sufficiently sound basis": see [78]. The experts had concluded that the data was presently incomplete but the material contained in their report (together with their evidence at the hearing) was sufficient to persuade the CAT that the calculation of global loss based on a weighted average pass-on was methodologically sound and not (to use the language of the judgment in Microsoft) purely theoretical or hypothetical.
50. If, as the Supreme Court explained in Microsoft, the function of the Tribunal at the certification stage is to be satisfied that the proposed methodology is capable of or offers a realistic prospect of establishing loss to the class as a whole then that requirement was satisfied. The availability of data sufficient to allow the methodology to be operated on what the CAT described as a sufficiently sound basis ought at the certification stage to be looked at in terms of what information can be made available for use at the trial.
51. The appellant's experts in their report and in their evidence at the hearing acknowledged that the evidence gathering process was still at an early stage but sought to identify the likely sources of the relevant material. Although the CAT was entitled to satisfy itself that the experts' proposed methodology was credible, it was not appropriate at the certification stage to require the proposed representative and his experts to specify in detail what data would be available for each of the relevant retail sectors in respect of the infringement period. The experts had identified expected sources of the data and other material such as the pleadings in the Sainsbury's litigation. The CAT pointed out the limitations on the usefulness and probative value of this material for the purpose of the present proceedings including that in many of the actions by retailers against Mastercard the issue of pass-on will not be explored or at least will not be determined until trial. But the preliminary nature of those proceedings is not a bar to their being relevant as a possible source of evidence in the present action. At the certification stage the question for the CAT was not what they were capable of proving had this been the trial of the action.
52. Although in [57] the CAT rejected the idea that they should carry out some form of mini trial, that is in our opinion more or less what occurred. They also required the proposed representative to establish more than a reasonably arguable case which would have been the test had Mastercard applied to strike the claim out. What this in practice involved was the appellant's experts being cross-examined at a pre-disclosure stage in the proceedings about their ability to prove the claim at trial by reference to sources of evidence which they had identified but had not yet been able fully to analyse or assess.
53. The certification hearing therefore exposed the claim to a more vigorous process of examination than would have taken place at a strike-out application and gave minimal weight to the experts' own (albeit preliminary) assessment that there had been pass-on of the overcharges to consumers as found in the EC Decision itself. The CAT's self-direction at the end of [57] to the effect that it was required to scrutinise the application for a CPO with particular care in order to ensure that only appropriate cases go through also ignores the facet of this process highlighted in a number of the Canadian cases which is that certification is a continuing process under which a CPO may be varied or revoked at any time: see s.47B(9). The making of a CPO does not therefore prevent the CAT from terminating the collective proceedings if it subsequently transpires, for example, that the proposed representative is unable to access sufficient data to enable the experts' method of calculating the rate of pass-on to be performed. But a decision of that kind is much more appropriate to be taken once the pleadings, disclosure and expert evidence are complete and the Court is dealing with reality rather than conjecture.
54. The assessment of these matters is, of course, a question for the CAT to decide using its substantial expertise and on this appeal we can only interfere with the decision to refuse a CPO if there has been a demonstrable error of law. But, in our view, the approach taken to the expert evidence in this case was based on a misdirection as to the correct test to be applied in relation to whether the proposed representative had demonstrated that the claims were suitable for inclusion in collective proceedings (and, in particular, for an aggregate award of damages) so far as they were based on an allegation of pass-on of the MIFs to consumers. Although, as Mr Hoskins pointed out, Rule 79(2) enables the CAT to take into account a wide variety of matters which may be relevant to whether the included claims are suitable for inclusion in collective proceedings, in this case the CAT was primarily concerned with the issue of availability of data as evidence to support and establish a claim for aggregate damages: see [78]. At the certification stage the proposed representative should not, in our view, be required to demonstrate more than that he has a real prospect of success. This is not the test which the CAT applied.
55. In these circumstances, we do not propose to consider the supplementary report prepared by the appellant's experts which, as we have said, is directed to other potential sources of data which they say could have been placed before the CAT had they realised the need to do so. The Mastercard defendants have accepted that further material could have been provided to the CAT but dispute the application which has been made to admit the new evidence for the purposes of this appeal. Our view is that the new evidence is not required in order fairly to dispose of the appeal but it will doubtless become relevant should this matter return to the CAT for a re-hearing of the application for a CPO.

Distribution

56. The argument between the parties on this ground of appeal narrowed considerably during the hearing. It is evident from [88] of the CAT's judgment that in relation to the issue of distribution certification was refused because the proposed method of distributing any aggregate award of damages would bear no relation at all to the loss suffered by individual members of the presented class. Some would receive considerably more than they had lost during the infringement period: others (who were high spenders) considerably less. The proposed representative does not pretend otherwise. But Mr Hoskins now accepts that there is nothing in s.47B of the CA or in the Rules which requires aggregate damages to be distributed on a compensatory basis of the kind envisaged by the CAT and that there was an error of law if the proper reading of [87]-[89] of the judgment is that the CAT considered that distribution must in all cases be carried out on some kind of compensatory basis however approximate.
57. The distribution of an aggregate award is included in Rules 92 and 93. As Mr Hoskins accepted, they do not require that the award should be distributed according to what each individual claimant has lost although, where that is readily calculable, it will probably be the most obvious and suitable method of distribution. But we reject the suggestion that a loss-based method of distribution is mandated by the statutory provisions or therefore that the proposed method makes it unsuitable for a CPO to be made under Rule 79. As Mr Harris submitted, the power to make an aggregate award under s.47C(2) with reference to individual loss would be largely negated in large-scale opt-out proceedings of this kind if a calculation of individual loss was a pre-requisite for any authorised method of distribution and therefore for certification.
58. In our view the CAT clearly did consider that the aggregate award had to be distributed in a way which paid sufficient respect to the basic principle that damages should aim to restore the proposed representative to the position he would have occupied but for the breach and, on this basis, the respondents concede that there was a misdirection. But I wish to add a few comments as to why we consider the approach taken by the CAT was too narrow.
59. The result of the refusal to grant the CPO in this case is that no follow-on proceedings for infringement based on the EC Decision are likely to be taken by consumers against Mastercard. The likely scale of loss caused to any individual consumer, coupled with the costs of the proceedings, makes litigation by way of individual claims a practical impossibility. This was recognised by the CAT in [91] of their judgment where they say that this is effectively the position in most cases of widespread consumer loss resulting from competition law infringements.
60. Be that as it may, the power to bring collective proceedings introduced into the CA by the Consumer Rights Act 2015 was obviously intended to facilitate a means of redress which could attract and be facilitated by litigation funding and had Parliament considered it necessary to limit this new type of procedure by what would be required for the assessment of damages in an individual claim then it would have said so. As it is the provisions for distribution of an aggregate award are open-ended. The CAT is expressly required under Rule 79(2) to take into account whether the claims are suitable for an aggregate award of damages when considering whether to make a CPO but not whether such an award can be distributed in any particular manner. The making of an aggregate award does not, as already mentioned, require the Court to calculate individual loss or importantly to assess the damages included in that award on an individual basis. Why, then, should they be distributed in that way?
61. The CAT adopted Mastercard's position which essentially is that damages must be awarded (and therefore it is said) distributed on a compensatory basis as a matter of established principle and that neither the CA nor the Rules mandated a different approach. But that is difficult to reconcile in our view with either s.47C(2) or the purpose of enabling individual claims to be brought in the form of collective proceedings. Once it is accepted that aggregate damages can be awarded and therefore assessed by reference to the loss suffered by the represented class as a whole, it becomes difficult to justify a reversion to an individual calculation of loss for the purposes of distribution. Claimants in this class will know that they stand to be compensated (if at all) on a per capita basis and can choose (if they wish) not to be included in the action. But once the assessment of the aggregate award is complete then, on one view, Mastercard has no locus to be heard on the question of distribution and no right to object if the class members who have suffered loss by the then proven infringement stand to be compensated in a way which does not exactly or perhaps even approximately replicate the individual losses which have been suffered. The vindication of the rights of individual claimants is achieved by the aggregate award itself.
62. More importantly, for present purposes, the CAT is not required under Rule 79(2)(f) for certification purposes to consider more than whether the claims are suitable for an aggregate award of damages which, by definition, does not include the assessment of individual loss. Distribution is a matter for the trial judge to consider following the making of an aggregate award: see Rules 92 and 93. We therefore consider that it was both premature and wrong for the CAT to have refused certification by reference to the proposed method of distribution: an error compounded by their view that distribution must be capable of being carried out by some means which corresponds to individual loss.

Conclusion

63. We therefore allow the appeal and set aside the order of the CAT refusing certification. Due to issues about funding which it is not necessary to go into, it is now common ground that if the appeal is allowed then the application for certification must be remitted to the CAT for re-hearing. We will therefore make an order in those terms.
© Crown copyright