On 18 February 2026, the Competition Appeal Tribunal (“CAT”) handed down its second landmark ruling in the ongoing Merchant Interchange Fee Umbrella Proceedings in less than a year.
After first finding that all multilateral interchange fees infringe competition law in the UK and Ireland in its 27 June 2025 judgment from the so-called ‘Trial 1’, the CAT has now held that, barring three narrow exceptions in the sectors analysed, the multilateral interchange fee (“MIF”) is not passed on by merchants to consumers in the form of higher prices.
Unlike in its prior judgments, including in the similarly high-profile Royal Mail Group Limited v DAF Trucks Ltd and Ors where the question of pass-on was assessed alongside other issues including liability, causation and overcharge, it was the CAT’s sole focus in this ‘Trial 2’ judgment.1 Indeed, to-date, ‘Trial 2’ was the first to be dedicated solely to the resolution of the question of pass-on. The Trial 2 judgment is accordingly of wider substantive and procedural significance. More immediately for the Umbrella Proceedings, it also means that, barring a finding at the upcoming ‘Trial 3’ that a certain level of MIF is ‘exemptible’ under Article 101(3) TFEU and s.9 Competition Act 1998, the charging of any level of MIF by Visa and Mastercard amounts to anticompetitive conduct that merchants bear the brunt of.
The CAT’s decision marked the culmination of a seven-week split trial held between November 2024 and April 2025 in which three different forms of pass-on were assessed. Although this article will focus predominantly on ‘merchant pass-on’, these three forms of pass-on can be defined as follows:
After an initial five weeks which addressed the latter two forms of pass-on (“Trial 2A”), and a subsequent one week focused on acquirer pass-on (“Trial 2B”), the CAT concluded as follows:
In reaching its conclusion, the CAT considered a variety of quantitative and qualitative evidence from merchants and economic experts. It also conducted a detailed analysis of how pass-on has been assessed to date across numerous cases – not just those relating to MIFs. The judgment therefore provides long-awaited clarification and detailed guidance on precisely what defendants must prove to successfully rely on a pass-on defence.
Ten years ago, in its judgment in Sainsbury’s Supermarkets v Mastercard, the CAT asserted that pass-on ‘is in reality not a defence at all’.4 Rather, the CAT concluded, it was concerned with ensuring that ‘a claimant is sufficiently compensated, and not over-compensated’ and defendants being required to prove whether another class of claimants exists downstream5. Since then, the test for proving pass-on has evolved. First, the Supreme Court’s judgment in that same case seemed to balance the scales by dispelling any suggestion that greater precision was required in assessing pass-on rates. In doing so, the Supreme Court appeared to depart from the prior judgment handed down by the Court of Appeal which had rejected the ability of defendants to rely on the CAT’s ‘broad axe’ to establish pass-on, and in so doing appeared to make the burden of proving pass-on a more difficult one to discharge.6 Moving forward to 2021, the CAT’s judgment in Royal Mail arguably re-raised the hurdles facing defendants by emphasising the importance of a clear factual basis for reliance on pass-on. By requiring defendants to prove specific and identifiable losses and limiting their ability to instead rely on economic theory and assertions that costs are always passed-on, the CAT re-iterated the high evidentiary bar facing defendants that seek to rely on pass-on.7 Most recently, however, seemingly in direct contrast to the Court of Appeal’s position in Sainsbury’s, the Tribunal’s willingness to wield its ‘broad axe’ when estimating pass-on rates has become apparent, as seen in Dr Rachael Kent v Apple where the Tribunal acknowledged the need to apply ‘informed guesswork’.8
Clearly, the legal battleground of pass-on is and will remain a slightly unpredictable one. It was against this judicial backdrop that Trial 2 took place, with the CAT’s judgment from it re-iterating that pass-on remains a valuable, if unreliable tool in the arsenal of defendants and that it remains rooted in the common law doctrine of mitigation of loss.
Ever since the European Commission’s decision in December 2007 that MIFs charged by Mastercard on cross-border card transactions within the European Economic Area (‘Intra-EEA’ transactions) breached Article 81 of the EC Treaty (now Article 101 TFEU), the continued charging of MIFs by both Visa and Mastercard have been subject to ongoing regulatory scrutiny and private damages actions.
Charged on every Visa and Mastercard card transaction as part of the wider MSC, MIFs are a small fee usually calculated as a percentage of the transaction’s value. Since 2015, the Interchange Fee Regulation has introduced caps on the level of MIFs payable on consumer card transactions. Initially charged to acquiring banks – who facilitate the acceptance of card payments for merchants – by issuing banks – who provide consumers with payment cards – Trial 2 was focused on the question of who ultimately bore the cost of the MIFs.
That individual MIFs are small posed a particular challenge from the outset in seeking to establish whether and to what extent acquirer and merchant prices were affected by MIFs.9 Before addressing how this ultimately affected the Tribunal’s findings, we first briefly set out the parties’ competing positions on the three forms of pass-on.
Given the common ground between the parties over the treatment of the MIF under IC+ and IC++ contracts, the principal question facing the Tribunal during Trial 2B concerned the degree to which acquirers absorbed a portion of the MIF for customers on blended contracts.10 While the parties’ experts agreed there was a ‘material degree of pass-on’ to merchants, they differed in their views of what that percentage of pass-on was. The estimate put forward by the expert for Mastercard, Rachel Webster, was the lowest of the three ranging from 60-80%. Derek Holt, Visa’s expert, concluded that pass-on was different for smaller merchants (classed as those whose annual turnover is £50 million or less) – at 75% - compared to larger merchants, for whom he estimated pass-on to be at 100%. Last, the expert for the merchant claimants, Dr Stefano Trento, estimated complete or 100% pass-on of the MIF by acquirers across all merchants, irrespective of size.
In general, the parties’ experts agreed in Trial 2A that MSCs (and therefore the MIF) were too small and varied too infrequently for any statistical correlations to be identified between them and the prices charged by merchants. This led the experts to identify proxy costs in lieu of MSCs for analysis in the hope of observing a correlation between changes in those costs and downstream prices. Whether an appropriate proxy had been selected by each parties’ experts for this purpose was therefore a key consideration for the CAT.
The Defendants, who faced the burden of proving that merchants had avoided loss by passing on MIFs through higher prices, argued for high merchant pass-on rates and placed heavy reliance on economic theory in the process. They also argued that it was impossible (but unnecessary) to establish the precise extent to which each claimant passed on the cost of the MIF. Mr Holt selected Costs of Goods Sold (“COGS”) as his proxy, on the basis MIFs were variable industry-wide costs incurred throughout the claim period. Ms Webster also treated MIFs in this way, with her analysis, which drew more from the documentary evidence provided by the merchants themselves, generating merchant high pass-on rates as well.
With Mastercard also defending the consumer claims however, Ms Webster sought to differentiate her findings of high pass-on to the merchants’ claim period, which started in 2007, from the period covered by the claims brought by Mr Walter Merricks. To do this, she cited the lower levels of card usage in the years dating back to 1992 to suggest that pass-on levels in that earlier period were significantly lower.
Justin Coombs, the expert for Walter Merricks, who was the Class Representative for the claim brought on behalf of UK consumers against Mastercard, opted for total costs as his proxy, which included COGS as well as all overhead and fixed costs. This was on the basis that all costs become variable costs in the long-run and are thus recovered in pricing decisions at some point in the future. His analysis also produced high rates of pass on.
Last, having considered the available data inadequate to model other, more characteristically comparable costs identified by the claimants’ pricing expert, Dr Trento ultimately opted to use total overhead costs as a proxy. This was despite conceding that the use of any large cost as proxy for such a small cost like MIFs was problematic. He selected this proxy on the basis that most merchants did not treat MSCs or MIFs as COGS but rather as part of their overhead costs and treated them accordingly in their accounts. The Defendants rejected the utility of this approach and suggested the evidence provided by the small number of Claimants was not representative of and could not be extrapolated to apply to the different economic sectors being assessed.
Aside from the economic question of selecting between the different proxies, a key legal battleground in Trial 2A concerned the relevant test for causation. Specifically, did this test include a requirement for the Defendants to prove a ‘direct and proximate causative link’ between the MSCs and downstream prices charged to consumers – or could the Defendants validly argue that all costs including MIFs were passed on through pricing in the long run.11 This was clearly a pivotal question, the answer to which would determine whether the Defendants had to show a plausible mechanism by which MSCs would have been consciously and actively taken into account by merchants when setting their prices.
Citing the conclusions of the Supreme Court in Sainsbury’s, where it asserted that ‘the question of legal causation is straightforward’, as well as the prior decision-making of the CAT itself in these proceedings which relied on Sainsbury’s, the Defendants argued that ‘proximity’ was neither a necessary nor relevant consideration when assessing pass-on.12 This meant that the only live issue to be determined was one of factual causation, the test for which the Defendants contended was whether, but for MIFs, merchants’ prices would have been lower. Given the apparent need for sufficient merchant evidence on costs recovery to answer this question, the Defendants, and particularly Mastercard, also argued for adverse inferences to be drawn from the limited volume of disclosure provided.
The notion of ‘supplier pass-on’ was not an issue that the parties’ economics experts had been instructed to consider. It was also a theory that all economic experts, including that instructed by Mastercard, conceded was not something economists would consider affects well-run corporations. This article will accordingly not dwell on the arguments made in respect of supplier pass-on – of which there were relatively few. Most, if not all arguments made relating to supplier pass-on were voiced by Mr Greg Harman – a Chartered Accountant instructed by Mastercard; Visa made no submissions in respect of supplier pass-on during trial. Mr Harman’s arguments in this regard largely related to indications in the merchant evidence that changes in MSC costs ‘may lead to action to control other costs’. Mastercard also submitted that adverse inferences should be drawn from an alleged failure on the part of the merchant claimants to provide sufficient documentary evidence – a point the Tribunal ultimately did not agree with.13
The Tribunal concluded that pass-on of MIFs under blended contracts occurred at a rate of 85%. This finding was based on ‘a broad axe assessment’, relying in part on three primary studies of data from the PSR and one of the three acquirers who provided documentary evidence which the Tribunal suggested had ‘evidential value’ of an ‘indicative rather than determinative’ nature.14 The Tribunal also rejected the specific theory of Mr Holt that merchants of differing annual turnovers experience differing levels of acquirer pass-on. Despite acknowledging the theoretical potential for acquirers to treat merchants differently based on their scale, the Tribunal ruled that the evidence before it was insufficient to ‘determine whether there are indeed different [acquirer pass-on] rates for smaller and larger merchants and, if so, at what level of annual card turnover that distinction arises’.15 The Tribunal’s finding of 85% was accordingly applied to all merchants on blended contracts, irrespective of their size.
The Tribunal’s findings on merchant pass-on were predicated on the resolution of two key questions – the test for causation, and the appropriate choice of proxy.
On the former, the CAT ultimately found that the Defendants had wrongly interpreted its prior determinations to mean that proximity did not feature in assessment of causation at all. In fact, those observations were made solely in relation to legal – and not factual – causation. In its July 2022 Judgment and March 2024 Ruling, the CAT, then under the Chairmanship of Marcus Smith J, had reiterated the distinction drawn by the Supreme Court in Sainsbury’s between factual and legal causation. It had also explained that ‘questions of legal causation… were not before the Tribunal because, as propositions, they were not arguable as a matter of law.’16
In reaching this position, the CAT relied on the conclusions of Green LJ in NTN Corp v Stellantis NV, as well as the Court of Appeal’s subsequent dismissal of the defendants’ appeal in Royal Mail Group Ltd v DAF Trucks Ltd.17 In particular, the CAT’s focus on the factual position detailed in the merchant claimants’ evidence seemed to reflect the approach adopted in Stellantis, where the Court of Appeal struck out the defendants’ attempts to rely on a hypothetical pass-on defence based on theory. The Trial 2 judgment echoes the views of both Green LJ – that legal causation and the need for a ‘proximate, causal, connection between the overcharge and the act of mitigation’ were separate concepts – and the Chancellor, Sir Julian Flaux, who also identified the size of the overcharge relative to the businesses’ revenues as a relevant consideration when assessing factual causation.18
In this regard, the CAT’s ruling in this case builds on and aligns with the Chancellor’s reasoning in both Royal Mail and Stellantis. The Tribunal’s findings that legal causation was about ‘legal policy’ and proving factual causation required a ‘direct and causative link’ were reflective of those of the Chancellor who, like the Tribunal, preferred to avoid any reference to proximity in its conclusions.19 This was despite the fact it had featured earlier on in the Court of Appeal’s judgment. ‘We think that the test is adequately encapsulated in the formulation of a “direct causative link” and that is what we propose to apply’, Green J explained.20
Accordingly, the Tribunal ruled that the test for factual causation was ‘more than a “but for” test of causation’ as it was also ‘necessary to show that the overcharge directly caused or led to an increase in prices’.21 A mere awareness of the MIF when budgeting or setting prices would therefore be insufficient to satisfy this requirement, as would pricing based on competitor behaviour or EBITDA targets which would only indirectly impact prices. On this point, the Tribunal was clear that the Defendants, except for three economic sectors, had not met this burden.
The factual evidence presented confirmed that, rather than COGS, most merchants treat MIFs as overhead costs, which are not specifically taken into account when setting retail prices. Without a direct causative link between the payment of MIFs and the prices charged, the Tribunal found that there was no pass-on to consumers. Although this determination somewhat relegated the question of the selection proxies to one of secondary importance – with the Tribunal itself conceding as much in stating it did ‘not actually rely on total overheads as a proxy’ – it was also on this basis that the Tribunal confirmed it preferred the approach taken by Dr Trento.22 His selection of total overheads was endorsed by the Tribunal as the proxy that was ‘treated similarly to the way that the Claimants treated the MSC in fact and would be passed on, if at all, through a similar mechanism’.23 It was this factual reality which meant the Tribunal did not accept the defendants’ criticism of Dr Trento’s approach as being one that was overly dependent on how the MSC was treated in the merchants’ accounts.
In this regard, the Tribunal’s judgment provides helpful guidance on what criteria a proxy should satisfy to be considered appropriate or viable. In order to pass both the ‘theoretical and a practical/evidential test’, a chosen proxy must: i) have ‘similar economic characteristics’; ii) be treated ‘in a similar way in practice’; iii) be ‘affected by the same or similar mechanism [of pass-on]’.24 It must also ‘be large enough and [show] sufficient variance over the observations to have an effect on the price variable’ so as to ‘avoid the “signal to noise ratio” problem’ that can accompany the assessment of particularly small cost items – like the MSC.25
Of the 193 pages in the Tribunal’s judgment, only three were devoted to the question of supplier pass-on. The Tribunal ultimately held that the Defendants had not proved, on the balance of probabilities, that merchants had mitigated their losses by passing them on to suppliers. This was perhaps a result of the ‘odd’ manner in which the Defendants argued for supplier pass-on, and the Tribunal’s general uncertainty as to ‘whether it has any real credibility’.26 However, more fundamentally, it reflected both the suggestion in economic theory that ‘supplier pass-on is implausible’, as it relies on the notion that merchants do not negotiate as effectively as possible with their suppliers, and the absence of any evidence of a direct link between changes in MSCs and attempted re-negotiations with suppliers.27
The Tribunal’s Trial 2 judgment is one that could have significant and long-lasting implications for defendants’ abilities to rely on the pass-on defence. Prior to Trial 2, the argument that merchant claimants passed on their losses to consumers through the raising of prices had been a key tool in the arsenal of defendants. Although it of course remains available, in cases such as this one, where the alleged overcharge is both small and ancillary as a cost component of the merchant’s downstream sales, the chances of it being deployed successfully certainly seem to have shifted in the favour of claimants.
While there are numerous key takeaways from the judgment, perhaps the most important is the certainty it has provided about the requirements for proving both factual and legal causation. Building on the prior decisions in Stellantis and Royal Mail, and as since re-iterated by the Court of Appeal in its decision in Granville Technology Group Limited v LG, the CAT found that ‘directness’ is necessary component of the test for proving factual causation.28 This means that moving forward (and subject to the outcome of the Card Schemes’ appeals), only pricing decisions made as a conscious and direct reaction to an overcharge in the short-term will result in a finding of pass-on. Conversely, arguments like those made by Visa and Walter Merricks that all costs become variable and are passed-on in the long-run will not be sufficient. Neither will arguments that ‘point in a generalised way to possible implicit internal mechanisms such as [EBITDA] margins or competitor monitoring, because this would only indirectly translate into prices.’29
What is particularly clear from the Tribunal’s judgment is the importance of rooting a claim, a defence or, as was seen in the recent Supreme Court in Evans v Barclays Bank Plc, a certification application in a strong foundation based on factual reality and evidence.30 Despite not being overly voluminous, the factual evidence presented by the merchant claimants was clear and consistent, and sufficiently compelling for the Tribunal to rely heavily on its content in reaching its decision. As such, the Tribunal seemingly attributed less weight to the economic experts’ evidence presented to it. This contrasted with the Court of Appeal in Granville which, when presented with ‘limited’ factual evidence, including no witness evidence (owing in no small part to the long-term liquidation of the claimants), ruled the judge at first instance was ‘entitled… to give such weight as he thought appropriate to the expert evidence’.31 It remains to be seen whether the greater prominence afforded to the expert evidence in Granville is to be attributed to the more unique circumstances of that case, rather than a prompt rebuttal of the CAT’s focus on the factual over the expert evidence presented in Trial 2.
Undoubtedly, the CAT’s findings will make the hurdle of proving factual causation that little bit higher for defendants to clear. This is particularly the case for cases where the alleged overcharge is small relative to the overall costs stack of which it is said to be a part. This was of course the case here and in both Royal Mail and the more recent Stellantis Auto SAS v Autoliv AB.32 As the CAT identified, the likelihood of ‘price adjustment costs’ and the simple fact that such small costs are unlikely to be monitored in the short-term makes it less likely that defendants will be able to show the level of directness the test for factual causation necessitates.33
This can of course be contrasted with the CAT’s conclusions on legal causation, which as a question of legal policy, seems increasingly unlikely to pose an obstacle to defendants seeking to rely on a pass-on defence. The question focusses on whether, even though factual causation has been proven (i.e. that claimants are found to have recovered the cost through higher downstream prices), legal policy dictates that claimants should nonetheless be entitled to claim the full amount of loss. The Supreme Court in Sainsbury’s cited personal injury claims, or instances where claimants receive an indemnity under an insurance policy as two examples where legal policy could intervene. By contrast, the recovery of losses by merchants through regular budgeting processes will clearly not give rise to the same concerns. Accordingly, no such policy reasons arose in this case, but it is also difficult to see how any challenges linked to legal causation could arise more widely in competition damages actions. In all likelihood therefore, it is the question of factual causation that will prove to be key in determining pass-on debates moving forward.
As a trial dominated by economic evidence and analysis, the Trial 2 judgment naturally contains some helpful guidance about the Tribunal’s expectations of experts and the evidence they present. Above all, what becomes clear is the importance of reflecting what is shown in the factual evidence in the economic analysis experts conduct. Expert evidence must be ‘grounded in reality’ and should seek to explain the actual, real-world behaviour of merchants set out in the factual evidence they present, instead of drawing solely or even predominantly on economic theory.34 This extends to the creation of economic models, the development of final theses and, particularly for this case, the selection of proxies.
In this regard, The Tribunal noted that Mr Holt had adopted a different approach, having ‘selected his favoured proxy before considering any data or evidence provided by the Claimants… [and] even before the Analysed Claimants had been selected’. The result was that ‘[his] persistence on using COGS ignore[d] the reality that the MSC was treated by most Claimants as an overhead cost’.35 The fact that a more precise, less-varied pass-on estimate could be found using COGS was insignificant in the eyes of the Tribunal given how far removed merchants’ treatment of COGS was from how they accounted for MIFs. This was in sharp contrast to overhead costs. Despite giving ‘rise to practical issues’ which are ‘problematic’ when used as a proxy, they were ‘treated similarly to the way that the Claimants treated the MSC in fact and would be passed on, if at all, through a similar mechanism.’36 As a result, Mr Holt’s insistence on using COGS as a proxy despite this difference – and his failure to acknowledge the weaknesses in doing so – meant his analysis was ‘fatally flawed’.37‘The fact that a more accurate result can be obtained by using COGS does not justify using it if it is the wrong proxy’, the Tribunal concluded.38
In reaching this position, the Tribunal also reiterated the views previously expressed in Royal Mail that experts should accept when other experts make valid observations and be prepared to acknowledge weakness in their own positions. The overriding duty of economic experts is and remains to the CAT and not the parties instructing them. In fulfilling this primary duty, experts should also present their evidence, both written and oral, in a concise, focussed manner that avoids the potential for disproportionately long experts reports and responses during cross-examination. On this point in particular, there can be little doubt that the CAT was seeking to highlight much of what is detailed in its December 2025 Practice Direction on expert evidence.
It is plain from these specific extracts, and the judgment more widely, that the Tribunal attributed more weight to the factual evidence presented by the merchant claimants than to the analysis and modelling of the experts. Indeed, the Tribunal itself conceded that, on the basis of its ‘factual conclusions… it may be thought that [it does] not need to examine closely the expert evidence.’ The Tribunal does of course do so but only, it suggests, ‘in case [it is] wrong on the law’ and ‘because the expert evidence helps to inform whether [it was] right to adopt the approach [it did].’ This apparent relegation of expert evidence to a position of secondary importance below factual evidence, rather than one of equal or even higher standing, when assessing questions of pass-on should be noted by those who have recently issued or are contemplating new claims. Moving forward, it will be interesting to see whether the Tribunal reverts to placing heavy reliance on economic evidence in proceedings where there is an absence of upstream data readily available – such as where claims have only been brought on behalf of consumers. Therefore, although the Trial 2 Judgment has seemingly raised the bar for defendants seeking to rely on the pass-on defence, it remains to be seen whether the Tribunal’s approach and reasoning will be confined to the facts of this case, or whether it will have far wider implications outside the world of MIFs.
1 Royal Mail Group Limited v DAF Trucks Ltd and Ors [2023] CAT 6.
2 It was accepted by the parties and upheld by the CAT in its judgment that the acquirer pass-on rate for ‘interchange plus’ (“IC+”) and ‘interchange plus plus’ (“IC++”) contracts (where the MIF component is separately identified) was 100%.
3 For the remaining three sectors, the following pass-on rates were found: Cash services – 100%; Insurance Underwriting – 46.7%; and Travel Agents and Online Intermediaries – 47.5%
4 Sainsbury’s Supermarkets Ltd v Mastercard Incorporated & Ors [2016] CAT 11, para 484(3).
5 Ibid. para 484(4).
6 Sainsbury’s Supermarkets Limited v Mastercard Incorporated & Ors [2018] EWCA Civ 1536 para 331.
7 [2023] CAT 6 para 228.
8 Dr Rachael Kent v Apple Inc & Ors [2025] CAT 67 para 1057.
9 In the case of domestic consumer card transactions, MIFs represent no more than 0.3% of a given transaction’s value.
10 Blended contracts do not specify or set out the value attributed to the MIF component of the MSC. Rather, the MSC is presented as one overall fee. They differ in this regard to IC+ and IC++ contracts which do identify the MIF amount within the overall MSC figure.
11 [2026] CAT 11, para 20. The judgment can be found here.
12 Sainsbury’s Supermarkets Ltd v Visa Europe Services LLC & Ors, Mastercard Incorporated & Ors [2020] UKSC 24, para 215.
13 [2026] CAT 11, para 493.
14 Ibid., paras 624, 618, 619.
15 Ibid., para 595.
16 [2022] CAT 31, para 50(2)(ii).
17 NTN Corp v Stellantis NV [2022] EWCA Civ 16 and Royal Mail Group Ltd v DAF Trucks Ltd & Ors [2024] EWCA Civ 181.
18 [2022] EWCA Civ 16, para 33.
19 [2026] CAT 11, paras 83-84.
20 Ibid. para 98.
21 Ibid.
22 Ibid. para 354.
23 Ibid. para 353.
24 [2026] CAT 11, para 334.
25 Ibid., para, 335.
26 Ibid., para 487.
27 Ibid., para 489.
28 Granville Technology Group Limited (In Liquidation) & Ors v LG Display Co. Limited & Anor [2026] EWCA Civ 409.
29 [2026] CAT 11, para 100.
30 Evans (Respondent) v Barclays Bank Plc & Ors (Appellants) [2025] UKSC 48.
31 Granville Technology Group Limited & Ors v LG Display Co. Limited & Ors [2026] EWCA Civ 409 paras 94-95.
32 Stellantis Auto SAS v Autoliv AB [2025] CAT 9.
33 [2026] CAT 11, para 325.
34 [2026] CAT 11, para 214
35 Ibid., para 344
36 Ibid., paras 350-351
37 Ibid., para 352
38 Ibid.
39 Ibid., para 212.
