Phillip Gwyn James Evans v Barclays Bank PLC & Ors

This is an appeal from the judgment of 31^st March 2022 of the Competition Appeal Tribunal (“CAT”) following a five day hearing in July 2021 which was then followed by a series of written submissions from the parties to the CAT in September and November 2021 (“the Judgment”). The appeal is brought with permission of the CAT upon the basis that the issues are novel, difficult and evolving and there was disagreement between the tribunal members on the central points. The majority included the President, Mr Justice Marcus Smith, and Professor Neuberger. The dissenting minority comprised Mr Paul Lomas. (Footnote: ¹)

The issues relate to the system of collective actions instituted by the Consumer Rights Act 2015 which led to amendments to the Competition Act 1998 (“CA 1998"). Under this system representatives (who generally combine a named individual as figurehead, lawyers and funders) apply to be certified to act collectively for a class of claimants. In this case there were two rival claims for certification before the CAT brought by Mr Evans (“Evans”) and Michael O’Higgins FX Class Representative Ltd (“O’Higgins”) as putative class representatives. If a claim is certified as suitable for a collective claim, the CAT then decides whether it proceeds upon an opt-in or opt-out basis. This system and related issues have been the subject of a series of judgments of the Supreme Court and the Court of Appeal. They have set out detailed expositions of the system. Rather than repeat what has been extensively set out therein, it suffices to refer to those judgments for the background. The judgments are as follows: Sainsbury's Supermarkets Ltd v Mastercard Inc [2020] UKSC 24 ("Sainsbury's"); Merricks v Mastercard Inc [2020] UKSC 51 ("Merricks"); Le Patourel v BT Group PLC and another [2022] EWCA Civ 593 ("Le Patourel"); LSER and others v Gutmann [2022] EWCA Civ 1077 ("Gutmann"); and, MOL (Europe Africa) Ltd and others v Mark McLaren Class Representatives Ltd [2022] EWCA Civ 1701 (“McLaren”).

In the Judgment the CAT gave permission for the Appellants to submit revised applications to be certified on an opt-in basis a claim for damages against certain banks which had engaged in an unlawful exchange of competitively sensitive pricing and other data with the object of reducing the risks normally attendant upon genuine competition. The claim is that this illegal cartel resulted in the participating banks earning artificially inflated returns at the expense of competitors and counterparties. The total claim with interest approaches £2.7 billion.

This judgment has been handed down at the same time as the judgment in UK Trucks Limited v Stellantis NV (Formerly Fiat Chrysler Automobiles NV) and others [2023] EWCA Civ [875] (“Trucks”). In that case an identically constituted Court of Appeal heard appeals in relation to a collective action against certain truck manufacturers for a price fixing cartel which it is alleged raised the price of trucks causing loss and damage to customers. There was overlap between some of the issues arising in the two appeals. Both cases are follow-on claims whereby, pursuant to section 60A CA 1998, the prior regulatory finding of the EU Commission is binding as to liability.

The present case was the first coming before the CAT for certification following the seminal judgment of the Supreme Court in Merricks (ibid). There are however now over 30 collective actions before the CAT with vast total claims. Many of these claims are highly complex legally and economically. In the intervening period the CAT has grappled with a range of novel procedural and legal issues and has instituted, and fine-tuned, many case management techniques, all with the object of bringing order and control to what otherwise risks the unleashing of litigation leviathans.

On 16^th May 2019, the Commission rendered two decisions finding infringements of Article 101(1) TFEU in relation to FX spot trading: (i) Case AT.40135 FOREX (“Three Way Banana Split”); and (ii) Case AT.40135 FOREX (“Essex Express”). The decisions are short form, settlement decisions, whereby the defendants made admissions to the Commission in return for a reduced fine. They have not therefore been made the subject of appeals. Both decisions are “object” decisions where liability is predicated upon the Commission proving, to the requisite standard, that the object of the cartel was to restrict and distort competition. This means that there is no detailed analysis of the “effects” of the cartel, which is the alternative predicate for liability. An important issue arising in this appeal is as to the probative value before the CAT of object based decisions. Taken at face value they provide limited information, evidence or guidance and offer only a relatively bare bones account of the impugned conduct and the law.

As it happens, on 5^th July 2022, following the hearing and judgment of the CAT and hence left out of account in the Judgment, the Commission published a fully reasoned decision addressed to Credit Suisse (which is not one of the defendant banks) finding an infringement of Article 101(1) TFEU in Case AT.40135-FOREX (“Sterling Lads”). This was not a settlement decision. As can be seen from a compare and contrast exercise with the two earlier decisions, the decision in Sterling Lads is a fleshed out version of the template used in Three Way Banana Split and Essex Express. There are numerous paragraphs of all these decisions which are identical. An issue in this appeal concerns the admissibility and probative value in these proceedings of the ordinary decision in Sterling Lads. The respondent banks argue that it is wholly inadmissible; but if admissible bears strictly limited evidential weight. The appellants argue that it is admissible, relevant and provides powerful support for their arguments on the appeal.

At the certification stage the putative class representatives adduced a substantial volume of factual and expert economic evidence. This was subjected to detailed scrutiny by the CAT majority which expressed concern that, whilst a plausible case at the level of economic theory had been advanced, they could not detect how, at the evidential level, causation was to be established between the breach (as per the Commission decisions) and the alleged loss, nor how disclosure would fill this gap. Significantly, the banks did not seek to strike out the claim or adduce evidence (expert or otherwise) or give disclosure to refute it. The CAT, of its own motion, however, considered whether it should nonetheless dismiss the claims.

The CAT majority ultimately decided not to strike out the claims. Instead, having set out in the Judgment why they considered a viable claim had not been formulated, and recognising that their concerns were described most fully and in some respects for the first time in the Judgment, they deferred a decision on strike out until the prospective class representatives had been given an opportunity to address the CAT’s comments in relation to the question of pleading. They did however reflect their negative views in the decision to be made on whether the claims should be opt-in or opt-out. The criteria for this are set out in CAT Rule 79(3) under which the considerations to be taken into account are at large but the rule expressly stipulates that (i) the strength of the claim and (ii) practicability, are relevant. In relation to strength of the claim, the CAT majority treated their view that the merits were weak as a reason leading it to the conclusion that the proceedings should be opt-in. In relation to practicability the CAT majority found, as a fact, that if the claims were certified upon an opt-in basis there would be insufficient take up for any claim to be viable. Nonetheless, they concluded that an opt-in claim was practicable because, applying an overarching principle of access to justice, the class to be represented were well resourced and sophisticated entities capable of bringing proceedings. If they decided not to, it was to be inferred this was a deliberate decision upon their part. This was not an access to justice problem. In his dissent, Mr Lomas disagreed with both conclusions. He was materially more sanguine than the majority as to the ability of the class representative to articulate a sustainable theory of harm in relation to causation. He concluded that a viable claim had been articulated. He also disagreed that the order of the CAT should be for opt-in. He considered that the order should be opt-out since otherwise there would be no action at all which was inconsistent with the statutory objectives behind the collective action regime.

The CAT also had to choose between the two competing applicant class representatives. There was unanimity in the decision to select Evans though the reasoning of majority and the minority differed. There was also unanimity that there was virtually nothing between the two candidates but since, perforce, a choice had to be made, the Evans team won, just. The majority judgment records that the Evans team had made a marginally better job of articulating a theory of harm. Though the majority added that the real answer to the question which of the two class representatives was suitable was, in view of their concerns about the strength of the claims, “neither”.

An agreed list of issues prepared by the parties identifies 20 issues, some incorporating multiple subsidiary questions. As a precursor, this Court must decide whether each is for the Court of Appeal by way of a statutory appeal or, alternatively, for the High Court by way of judicial review. The parties have, because of uncertainty as to the scope of the statutory appeal route, invoked both the statutory appeal route and commenced proceedings for judicial review. How each strand of the challenge is to be determined is a logically prior matter that must be determined for each of the issues arising.

I have grouped the various issues as follows:

Before the CAT the claim relied upon the decisions of the Commission in (i) “Three Way Banana Split” which operated between 18^th December 2007 and 31^st January 2013 and (ii) “Essex Express” which operated between 14^th December 2009 and 31^st July 2012. The parties to the Three Way Banana Split infringement included the defendants UBS, The Royal Bank of Scotland, Barclays, Citibank and JPMorgan. The Essex Express infringement included the defendants UBS, The Royal Bank of Scotland, Barclays, and MUFG. The Commission imposed cumulative fines exceeding Euros 1.1 billion. The ordinary decision in Sterling Lads involved only Credit Suisse, which is not a defendant to these proceedings (there is also a settlement decision in Sterling Lads, to which UBS, Barclays, RBS and HSBC are addressees). I take my account of the basic facts relating to the characteristics of the cartel and the market mainly, but not exclusively, from the ordinary decision in Sterling Lads.

The cartel covered at least the whole of the EEA and consisted of agreements and concerted practices that had the object of restricting and/or distorting competition in the sector of foreign exchange spot trading of G10 currencies (“FX”). The G10 currencies concerned comprised the US, Canadian, Australian and New Zealand Dollars, the Japanese Yen, the Swiss Franc, the Pound Sterling, the Euro, and the Swedish, Norwegian and Danish Crowns.

The affected market was that for FX spot trading of G10-currencies. This refers to an agreement between two parties to buy a certain amount (the “notional amount”) of one currency against selling the equivalent notional amount of another currency at the current value at the moment of the agreement, for settlement on the spot date (normally the transaction day plus 2 days).

The infringements involved the participation of the banks in conduct occurring within chatrooms in which traders engaged in recurrent and extensive exchanges of information through which they revealed to each other current or forward-looking information about confidential aspects of their market conduct. There were three types of orders relevant to the unlawful exchanges of information (Footnote: ²):

In Sterling Lads, Credit Suisse argued that the exchange was integral to the proper functioning of the FX market and, accordingly, it could not distort competition. The Commission disagreed holding that the market operated effectively absent the challenged exchanges of information:

“(14)

The Commission agrees that during the telephone era, FX trading was fairly opaque. As stated in recital (12), at that time, information about FX trades was proprietary to the two counterparties. However, in contrast, nowadays trades entered on electronic platforms are cleared and settled electronically. Therefore, contrary to what Credit Suisse claims, FX markets are now regarded as transparent because real-time prices (including the best available bid and ask prices) and corresponding trading volumes are available in the platforms to virtually all participants. This information on best bid and ask prices is constantly updated, without interruption. The information available on electronic platforms is therefore sufficient for traders to make their own judgment about the evolution of FX rates before entering into FX transactions.

(15)

There is no need for traders to engage in contacts, including participating in multilateral chatrooms, to gather non-public information not present on those platforms to make their trading decisions.”

(emphasis added)

FX transactions represent the largest share of the world’s financial transactions and are among the most liquid financial transactions. A 2013 survey conducted by the Bank of International Settlements indicated a total turnover of more than USD 5 trillion per day, of which USD 2 trillion resulted from FX spot transactions alone.

To execute an FX transaction, end-customers typically contact “dealers” (financial institutions who employ “traders”) acting towards end-customers as “market makers” via the latter’s sales desk or directly on their internet trading platform. Dealers typically quote two-way prices (a bid and an ask price). These vary depending upon the transaction size and on the traded currencies. An end-customer can therefore contact several dealers and choose among the most favourable quotes before trading. All currencies are quoted in pairs, because each FX trade involves buying and selling two underlying currencies and each currency is valued in relation to another. Since currencies are always traded in this way, in foreign exchange there is no such thing as a currency’s absolute value but instead there is a relative value compared with other currencies. The market price of one currency is set in a given currency pair, that is, a value if it is exchanged against another. The foreign exchange rate is the rate at which one currency will be exchanged for another.

Historically, FX transactions were executed directly by telephone on direct lines between parties. FX trading was limited to contacts between trading parties for their business needs and proprietary activities. In order to gather information, traders often called each other asking for quotes or to pass off open risk positions with the intention to trade. The early 1990s saw the development of electronic screen-based broking systems such as the Reuters and the Electronic Broking System (“EBS”) platforms. Electronic platforms now allow market participants to access streaming price information, which updates continuously. Trades entered on electronic platforms are cleared and settled electronically, thereby increasing operational efficiency, streamlining trade processing and settlement, reducing operational risks, and lowering trading costs.

There are three broad categories of market participants: (i) end-customers; (ii) dealers and traders; and (iii) brokers.

End customers: These typically include corporate customers and financial institutions, comprising asset managers, hedge funds, corporations, banks and central banks. They might use FX trading to support treasury operations associated with their core business activities. Corporate customers primarily use foreign currencies as a medium of exchange or as a means to hedge foreign cash flows. They pay little attention to future exchange rate movements and do not engage in speculative FX trading. Financial institutions cover a broad spectrum including: hedge funds, asset managers, banks and central banks. They usually trade larger amounts of currencies than corporate customers, hold FX positions for longer and use currencies primarily as a store of value. They have strong incentives to acquire information liable to influence the evolution of FX rates and will tend to be better informed than other end-users. Their trades may anticipate short-term FX movements and returns, hence, they are considered “informative”.

Dealers and Traders: Although the expression “dealers”’ refers to the financial institutions dealing with currency exchanges and the term “traders” to the agents employed by the financial institutions to conduct their currency exchange trades, these terms are used inter-changeably. Historically, large commercial and investment banks have been dealers. They set up specific trading desks where individual traders trade in specific currency groups. FX traders are employees of financial institutions who trade the currencies on a specific trading desk. They stand ready to trade with anyone needing foreign exchange at a moment’s notice. The sales desks are the interface between the traders and end-customers. They are responsible for maintaining good relationships with customers. FX traders will deal with large amounts of currencies, the transaction sizes being proportional to the size of the end-customers. Corporations or financial institutions trading with FX traders directly or via sales representatives are generally large companies. FX traders make money by selling a currency against another at a higher price than that at which they bought it. Trading revenue therefore depends on the amount of currency volume traded and on the difference between the purchase price and the sale price of the same currency (the “bid-ask spread”). Traders may also profit from holding a particular open risk position in their book (long or short) in anticipation of better trading conditions at a later stage. Traders typically receive bonuses tied to their individual profits and the profits of the entire trading floor while their individual risk-taking is constrained by position and loss limits.

Brokers: These are financial intermediaries who match counterparties to a transaction without being a party to the trade. They can operate electronically (electronic broker) or by telephone (voice broker). Brokers’ revenues are proportional to the amount of trades executed by traders on their platform.

An important issue in this case concerns the causal effect of the cartel upon the trading position of counterparties. In Sterling Lads, Credit Suisse argued that even in the context of an object case, the Commission had to undertake “a detailed assessment, in each individual case, of the market shares of the parties involved or the effects in the market” (paragraph [424]) citing Case C-67/13P Groupement des Cartes Bancaires (CB) v Commission EU:C:2014:2204 (11^th September 2014) at paragraph [53]. The Commission rejected this submission. The case law did not require that level of analysis of actual effects. Were the submission to be true it “would defy the very purpose of the category of restrictions by object” (paragraph [424]).

The Commission accepted however that it had to demonstrate, by reference to the appropriate legal and economic context, that there was a real likelihood of an actual effect on competition. It held that there was such a likelihood basing its conclusion upon (a) the innate or intrinsic likelihood that harm would occur from the type of illegal cartel in issue and (b) its analysis of the actual evidence. The impact upon “competition” was reflected in the financial position of the cartelists relative to competitors and customers. The Commission found that the cartelists enjoyed artificially beneficial trading terms which were, as a matter of probability or likelihood, at the expense of competitors and trading partners.

In relation to the intrinsically high likelihood of resultant harm, the Commission stated that a cartel of this type was “particularly likely to lead to a collusive outcome” and for this reason exchanges of this sort of information are “…by their very nature, harmful to the proper functioning of normal competition”. The Commission cited the Commission Horizontal Guidelines (paragraph [390]) which, applying well established case law, treated exchanges of sensitive forward looking information as especially harmful to competition: “The exchange of forward-looking information and price information is particularly likely to lead to a collusive outcome on the market. Therefore exchanges of information about such future intentions are, by their very nature, harmful to the proper functioning of normal competition.” At paragraph [445] the Commission held that there was a likelihood of adverse effects on competition:

“(445)

First, the role of experience and, therefore, foreseeability in that regard do not concern the specific category of an agreement in a particular sector, but the fact that it is established that certain forms of collusion, such as, in the case at hand, the recurrent and extensive exchanges of current or forward-looking commercially sensitive information between competitors about confidential aspects of their market conduct which allowed them to engage in coordination of their trading activities, are, in general and in view of the experience gained, so likely to have negative effects on competition that it is not necessary to demonstrate that they had such effects.”

(emphasis added)

In relation to likelihood of harm established by reference to the evidence, in paragraph [286] the Commission found on the facts that the frequent and recurrent exchanges of information could facilitate specific forms of coordination, which took place with a view to benefiting the participating traders or to avoiding trading against each other’s interest. The Commission took account of such matters as the nature of the products or the services provided, the size and number of the undertakings and the volume of the market (paragraph [387]). The cartel was able to confer an opportunity to obtain “additional benefits” on the defendants:

“(287)

By occasionally coordinating (or standing down), the participating undertakings sought to gain an advantage over competitors that did not participate in the STG Lads chatroom. This concerns certain instances where the participating traders who had disclosed that their open risk positions at the fix were of a certain type spotted the opportunity to potentially obtain additional benefits from it …. Specifically, this coordination consisted in a practice called “standing down”.

(288)

The occasional standing down practice concerned instances in which traders refrained from trading as they otherwise had planned to undertake during the time of the fix on account of another trader’s announced position or trading activity. The modification of some participating traders’ behaviour during this time reduced the risk that a transaction by the participating trader would not achieve the desired outcome and avoided simultaneous trading in opposite directions.”

“(342)

The extensive and recurrent exchanges of information facilitated occasional instances of coordination among the traders in the form of standing down with a view to securing commercial benefits for the undertakings concerned, as described in Section 4.1.3.2.”

(emphasis added)

In paragraph [394] the Commission refers to the ability of cartelists to “exploit” the confidential data “for their own benefit” in order to “boost” their profits which would be at the “expense” of counterparties. The Commission concluded that the benefits could be “large” and work to the “detriment” of competitors and counterparties:

“As regards the content of the conduct, the participating traders engaged in recurrent and extensive exchanges of information through which they revealed to each other certain current or forward-looking commercially sensitive information about confidential aspects of their market conduct. These exchanges included:

(a)

Information on outstanding customers’ orders …. These exchanges applied to: (i) customers’ conditional orders, (ii) WMR or ECB’s fix positions and (iii) customers’ immediate orders.

(i)

The exchange of information on customers’ conditional orders … increased market transparency for the chatroom members and thus enabled them to gain a better understanding of the direction towards which the market might move, when certain pricing levels were reached. In this respect, the recurrent update of knowledge of customers’ confidential conditional orders placed with the participating traders was capable of influencing the participating undertakings’ trading strategy and of increasing their ability to exploit that level of insider information in their trading activities for their own benefit.

(ii)

Regarding the information exchanges revealing orders for the fix … the participating traders had effective access to more specific and timely information about competitor positions than they would otherwise have had absent the exchanges. This conferred on them the ability to predict with a greater degree of confidence the direction in which the market may move at the time of the fix. The participating undertakings could take advantage of this improved degree of confidence about the market trends at or around the fix to adjust their trading strategy (for example, by trading at or in advance of the fix in order to hedge their net client orders, by choosing not to net off or by refraining from trading) and thereby attempt to boost their profits at the expense of other competitors and counterparties who did not have access to the information.

(iii)

The exchange of current or forward-looking commercially sensitive information related to customers’ immediate orders (such as the size or the direction of specific, non-aggregated orders or the type or name of customer) removed some of the uncertainties that are inherent to Forex trading and increased the level of transparency about the evolution of the involved exchange rates for the participating traders. Some of the exchanges of information … show that the participating traders disclosed information on their clients’ identities when discussing details of their customers’ immediate orders. The clients identified were mostly significant market participants, such as financial institutions whose trading activity is informative (see recital (40)) and, therefore, may anticipate short-term FX movements. Hence, the disclosure of details on such customers’ immediate orders increased the pool of information available to the participating undertakings concerning which way (up or down) currency prices were likely to move … .

(b)

… the recurrent knowledge update of the open risk positions of participating traders provided them with information which could be, for a window of minutes or until new information superseded it, relevant to their subsequent trading decisions and could enable the participating undertakings to identify opportunities for coordination. In particular, when one trader disclosed an open risk position, the other participating traders would refrain from trading by withholding bids or offers so that the price of the involved currency pair would not move in a direction adverse to the trader with the open risk position.

(c)

Information on bid-ask spreads quoted for specified currency pairs for certain trade sizes and for certain client types …. As a result of the exchanges on bid-ask spreads, the participating traders could reduce the risk inherent in trading currencies to their benefit, so that with the knowledge acquired from the exchanges with their competitors they could safely offer to their clients the upper range in the market price levels. Even a minor spread difference for large volume transactions, such as the ones the participating undertakings dealt with, could have resulted in large benefits for them to the detriment of their clients.

(d)

Information on current or planned trading activities …. The cumulative disclosure of other details of participating traders’ current or planned trading activities, or a combination of the topics described above (customer orders and/or open risk position) also removed some of the uncertainties that are inherent in Forex spot trading and increased the level of market transparency. The availability of this information provided the participating undertakings with valuable cumulative insights into current trading patterns of their competitors and comforted them in their risk assessment when developing their own trading strategies.”

Similarly broad findings about effects, and in particular detriment to counterparties, were repeated later:

“(399)

Price and risk management are parameters of competition … that the participating undertakings should have managed autonomously; however, the undertakings that were involved, via the participation of their traders in the chatroom, substituted competition by cooperation amongst them.

(a)

Regarding price (and bid-ask spreads), the exchanges of information on outstanding customer orders and the cumulative disclosure of other details of current or planned trading activities removed some of the uncertainties that are inherent in Forex spot trading. These exchanges increased the level of transparency about the participating traders’ trading strategies …. The participating traders disclosed in the chatroom information on their outstanding customer orders, such as the identity or type of client, which provided them with an insight into the subsequent movements in the involved exchange rate. As the undertakings’ pricing strategy depends on their expectations of FX rates tendencies, these exchanges informed their pricing behaviour about general pricing trends thereby reducing the risk inherent in pricing currencies (e.g. likely resulting in lower expected losses).

The exchanges of information on bid-ask spreads in the chatroom … also provided the participating traders with greater certainty on the prices they were quoting and informed their subsequent trading behaviour concerning spreads. Those exchanges were capable of enabling the participating traders to align their spreads for particular transactions and, thereby, their all-in price offered to a specific client for a particular transaction. Any potential counterparty who was not aware of such exchanges of non-publicly available information on spreads and who might have contacted more than one of the participating traders to get a price on a specific trade, might have received less competitive prices from them.

(b)

Regarding risk management, … traders are compensated not only for the immediacy of the service they provide but also for assuming the subsequent risk of holding a certain currency in their inventories (the open risk positions). The expertise of FX spot traders resides in their risk management skills and their capacity to reduce their risk of losses. For instance, in this particular case, traders adjusted their position in a currency depending on their expectations on its price evolution …; if a participating undertaking expected a decrease in the market price of a currency and held a long position in it, it reduced its position in order to reduce the risk of making losses linked to a price decrease of the involved currency. The exchanges of information on open risk positions … provided the participating traders with greater certainty on the trading intentions of each other and hence removed some of the uncertainty as to the potential evolution of a specific Forex rate, thereby helping the participating undertakings in the management of their own trading risk.

(400)

Therefore, the extensive exchanges of current or forward-looking commercially sensitive information … reduced uncertainty between the participating undertakings on their respective trading strategies and of the direction in which the market might move. These exchanges of information allowed the participating undertakings to make their daily risk management decisions comforted by the knowledge of their competitors’ trading behaviour, trading exposures and immediate plans. This could help them to better predict each other’s future conduct in the market and gave them the ability to inform their subsequent trading decisions. The participating undertakings would be in the position to either persevere in their intended course of action comforted in their risk assessment with the information they had, eventually adapting their price to that risk perception, or to change course and trade differently to how they would have traded absent that information. …

(401)

In conclusion, the continuous exchanges of commercially sensitive information provided the participating undertakings with the opportunity to subtract themselves from competition on the merits with regard to key parameters of competition (price and risk management). This constant flow of information exchanges within the chatroom also entailed an asymmetry of information between the participating undertakings and their non-participating competitors to the advantage of the former, since only the participating traders were continuously aware of their trading behaviours, trading exposures and immediate plans and this knowledge provided them more comfort when adopting their market behaviour.”

(emphasis added)

The evidence was enough to enable the Commission to conclude that there was a “sufficient degree of harm to competition” for the conduct to qualify as an object violation:

“(447)

… an overall assessment of the content of the conduct, in the light of the aims objectively pursued and the economic and legal context in which the conduct took place … reveals a sufficient degree of harm to competition to conclude that it can be qualified as a restriction of competition by object.”

(emphasis added)

In paragraph [473] the Commission reiterated that the intention behind the collusive exchange was to offer the most expensive spread to clients and thereby create “large” benefits for the cartelists and “detriment” to their clients:

“(473)

Moreover, the continuous exchanges within the STG Lads chatroom significantly reduced normal market uncertainties to the advantage of the participating undertakings compared to other market participants.

(a)

Regarding the exchanges revealing outstanding customers’ orders (conditional orders, orders for the fix and immediate orders) … and current or planned trading activities … the Commission only retains information exchanges where traders revealed pieces of confidential information on specific ongoing or immediately executable transactions that were not justified for the purpose of exploring trading or actually trading with each other as counterparties … . Through these information exchanges, the participating undertakings provided each other with an insight on their current or forthcoming behaviour on the market (timing, pricing, trade size, etc.) reducing the uncertainty that is inherent to a competitive scenario, where the parties must determine autonomously their pricing and risk strategy … .

(b)

Through the exchanges revealing their quoted or intended bid-ask spreads … the participating traders who were competitors in the market advised each other on strategies of pricing to quote to their clients. They disclosed the actual spread they quoted for specific currency pairs, trade sizes and client types, which may also affect the overall price paid by customers for trading currencies. In these exchanges, a participating trader consulted his competitors in the chatroom on the most convenient spread for a specific trade before offering a quote to his clients.

The exchanges of information on bid-ask spreads in the chatroom increased transparency and reduced market uncertainties for the participating traders regarding prices. These exchanges enabled the participating traders to obtain greater certainty on the spreads they were quoting and might have informed their subsequent trading behaviour concerning spreads. The information exchanges may also have allowed them to align their spreads for particular transactions and thereby their all-in price offered to a specific client for a particular transaction. A customer who is not aware of such exchanges of non-publicly available information on spreads may have contacted more than one of the parties’ sales desks to get a price on a specific trade and may have received less competitive prices from them due to the exchanges of information on bid-ask spreads between the participating traders…

Moreover, contrary to Credit Suisse’s claims, there is evidence showing that the intention of the participating undertakings with the exchanges on bid-ask spreads was to be able to offer the widest (most expensive for the client) spread possible to the clients, within the constraints they jointly perceived were imposed by the wider market. In an extract of 17 February 2012 not contested by Credit Suisse, [employee of Credit Suisse] asks what current spreads the other traders are offering for a volume of 50 million (presumably, EUR/USD). [Employee of non-addressee] answers “5 unfortunately”, and then adds “6 if you are lucky”, thereby showing that the intent of the exchanges was to enable the traders to identify, and offer, the widest spread possible given the market conditions at that time.

As a result of the exchanges of information …, the traders’ uncertainty was reduced and they could ‘safely’ offer to their clients the upper end of wider spreads. Even a minor spread difference for large volume transactions, as the ones the participating traders dealt with, might have resulted in large benefits for the participating traders to the detriment of their clients.”

(emphasis added)

There is substantial overlap between the formulation of the Evans and O’Higgins claims. There are however two significant differences which concern, first, the way in which they described the two principal categories of claim, and secondly, as to the types of transaction covered. I set these out now but they are relevant to the carriage selection issue addressed at section G below.

I take the following summary from the case as advanced by O’Higgins, because it is the broader of the two claims, but there is a comparable version set out by the Evans team, and nothing rests on the choice. Both take their core analytical format from the way in which the Commission analysed the position in Essex Express and Three Way Banana Split:

In both claims damages are said, broadly, to arise in two ways: (i) direct losses attributable to the agreements as between the defendants and counterparties; and (ii), “umbrella” damages which arise because of an alleged tainting effect of the cartel on the (un-cartelised) remainder of the market. The artificial increase in spreads brought about by the cartel enabled other traders, unilaterally, to follow suit and increase their spreads in a way that would not have been possible in a genuinely competitive market. Put another way the cartel caused market wide contagion with the consequence that counterparties paid more for FX trades across the entirety of the market, and not just when trading with the cartelists.

The O’Higgins claim treats both categories of loss as part and parcel of a single undifferentiated claim. The Evans case distinguishes claims into Category “A” and “B” with the former being direct claims and the latter being the umbrella claims.

Mr Jowell KC for O’Higgins argued that there was no need to differentiate between the two claims and that it was artificial to do so. He might turn out to be right. But the Evans team disagrees and Mr Kennelly KC, for the respondent banks, when addressing the merits of the claims, reserved his most serious criticisms for the inclusion of the Category B losses. Quite irrespective of ultimate merits there is at least a possibility that during preparation for trial and at trial, the evidence and analysis might begin to differentiate as between direct and umbrella losses. At the level of theory, the chain of causation also seems more complex in relation to Category B, umbrella, losses relative to Category A losses.

One expert instructed for the O’Higgins team sought to explain why an undifferentiated claim was unproblematic. He commented upon the difference in approach between the expert evidence of Mr Ramirez (for Evans) and Professor Breedon (for O’Higgins). I set this out to indicate only how the two claims could be viewed as different:

Cartel analysis and mechanisms

“(137)

Mr. Ramirez and Professor Breedon also adopt different class definitions. Professor Breedon defines a single group of class members in paragraph 1.8 of his report. Mr. Ramirez was instructed to consider two class definitions, designated Class A and Class B. Class A includes “all persons who entered into certain spot and/or FX outright forward transactions with a proposed defendant on any date when the proposed defendant was found by the EC to have been participating in either of the cartels,” whereas Class B includes “all persons who entered into certain spot and/or FX outright forward transactions with (i) a proposed defendant on dates during the overall infringement period other than those dates included in the Relevant Class A Period … and/or (ii) certain FX dealers that are not addressees of the EC Decisions … during the overall infringement period.” The difference between the two classes potentially matters because Mr. Ramirez hypothesizes that the cartel mechanisms he considers impacted the two classes differentially.

Class A. Citing Professor Rime’s report, Mr. Ramirez argues that harm to members of Class A arises because cartel dealers would have widened bid-ask spreads applicable to both Spot transactions and Forward transactions between members of Class A and the proposed defendants.

Class B. Again citing Professor Rime’s report, Mr. Ramirez argues that members of Class B were potentially harmed through two channels. First, they suffered from “less competitive conditions,” by which he means that the wider bid-ask spreads charged by the proposed defendants allowed their competitors to quote wider spreads as well. Second, they suffered from increased adverse selection costs, which caused non-cartel dealers to widen their bid-ask spreads.

(138)

Professor Breedon’s proposed use of models with multiple dummy variables allows him to accommodate concerns that the cartels may have had different effects on different sub-groups of trades, possibly through different mechanisms. His approach has the considerable advantage of exploiting the fact that all transactions are disciplined to a significant degree by the same market forces. It follows that analyzing all transactions together while allowing for appropriate dimensions of heterogeneity in the measured effects will enhance accuracy, precision, and reliability. Professor Breedon’s approach is also more flexible, in that it permits him to reach conclusions about the heterogeneity of the cartels’ effects based on the patterns in the data, rather than based on preconceptions concerning the important dimensions of heterogeneity, which are of necessity relatively uninformed at this stage of the analysis.”

(emphasis added)

The second difference lies in the choice of transactions which the two competing claims seek to encompass. As set out in the Commission decisions (see paragraph [16] above) there are three types of transaction. Different terminology is used to describe or label the different transactions. For present purposes it suffices to use the terminology used by the experts. The O’Higgins claim includes benchmark transactions, limit orders and resting orders. The Evans claim excludes these transactions.

In the Joint CPO response of the respondent banks, their views on the differences between the two claims are set out. They adopt a ‘plague on both your houses' approach, though in a footnote they express agreement with the narrower Evans claim (for obvious reasons), but still go on to explain how difficult the exclusion would be. Their view is interesting in the way in which the two claims are differentiated:

“63.

Mr Evans proposes to exclude benchmark transactions, limit orders and resting orders from the scope of the proposed action. However, contrary to Mr Knight’s evidence, the Respondents do not hold data which enables these transactions to be reliably identified across the claims period. As Judge Schofield held in the US certification ruling:

“… individualized inquiries would be required to identify and exclude certain types of trades. “Benchmark trades” are trades that were entered into at a benchmark price. Such trades are expressly excluded from both the OTC Class and Exchange Class. “Resting orders” are orders that are placed in advance, directing the bank to execute a trade if and when the market price for a particular currency pair hits a specified level. Resting orders would not be impacted by a conspiracy to widen spreads in the spot market, because clients do not “pay the spread” when they place resting orders. Because benchmark trades and resting orders cannot serve as a basis for liability in this case, the type of each transaction executed by class members is highly material to their claims. Identifying and excluding benchmark trades and resting orders cannot be accomplished through generalized proof. Rather, a fact-intensive individualized inquiry would be required -- for example, a review of the relevant communications between class members and Defendants. This would be an enormous undertaking; Plaintiffs have identified tens of thousands of OTC class members, and each class member, under the OTC Class definition, entered into at least ten FX transactions”

64.

Judge Schofield went on to record that it was the Plaintiffs’ own evidence that the Defendants did not “maintain records regarding certain trade characteristics” (p.434).

65.

Moreover, Mr. Ramirez acknowledges that there is little public data on either benchmark trades or resting orders. The determination of whether any putative class member has entered into any or all of the three excluded categories of FX transactions with the Respondents can therefore only realistically be carried out through disclosure from the class members themselves. The same applies in relation to trading with non-Respondents. In any event, Mr Ramirez’s proposed method for removing benchmark trades, limit, and resting orders on an aggregate basis, would not resolve the issue of class membership, i.e. the issue of whether any given entity only entered into transactions falling within these three categories, and no other categories, during the period of the Infringements. Such an entity would not fall within Mr Evans’ proposed class at all. The only solution to this issue is therefore disclosure by individual class members.”

In their evidence the Evans team explained that the reason for the exclusions was pragmatic. The excluded trades were of a type for which it could not be said with confidence that any loss would necessarily have been sustained by the class and, indeed, it was possible that some trades might even in fact have been profitable. The exclusion made the claim more focused and manageable.

I turn now to the preliminary question of how disputes about a judgment of the CAT should be addressed and whether this should be by way of an appeal under section 49(1A)(a) CA 1998 or by way of judicial review to the High Court, or even both. At a directions hearing held on 6^th December 2022 the Court ordered that all issues (whether procedural or substantive, interim or final) were to be addressed at a single rolled-up hearing with the Court sitting as both the Court of Appeal and a Divisional Court of the High Court. This was embodied in an order of the Court. By this device we have been able to consider all issues in the round. By clarifying the law it should be possible to avoid this unnecessary duplication of cost, resource and effort in the future. Nonetheless, should such a scenario arise again, this is a sensible way of dealing with these disputes.

The right of appeal is set out in section 49(1A)(a) CA 1998:

“An appeal lies to the appropriate court on a point of law arising from a decision of the Tribunal in proceedings under section 47A or in collective proceedings—(a) as to the award of damages or other sum (other than a decision on costs or expenses) …”

This contains two cumulative limits. First, an appeal is limited to “a point of law”; and secondly, that point of law must be “as to the award of damages or other sum”.

There is no difficulty in identifying a point of law when it concerns, for instance, the interpretation of a statute or other legal instrument. As set out below, the Supreme Court in Merricks identified many challenges which it treated as concerning “points of law”. The area of greatest difficulty lies in categorising challenges to findings of fact or to evaluations of combinations of fact. The classic formulation of the test here is found in Edwards (Inspector of Taxes) v Bairstow [1956] AC 14 (“Bairstow”) which held that an appellate court could reverse a finding of fact where it appeared that the decision maker acted without any evidence, or on a view of the facts which could not reasonably be entertained. Lord Radcliffe, at page 36, stated:

“…it may be that the facts found are such that no person acting judicially and properly instructed as to the relevant law could have come to the determination under appeal. In those circumstances, too, the court must intervene. It has no option but to assume that there has been some misconception of the law and that, this has been responsible for the determination. So there, too, there has been error in point of law.”

In Secretary of State for Education and Science v Tameside MBC [1977] AC 1014 the House of Lords confirmed that where a public body had misdirected itself, or made a mistake, as to a material fact, that may suffice to allow a challenge as law. In a typical case an error of law might, for example, arise if a decision maker has in a material respect: failed to take account of a relevant fact; taken into account an irrelevant fact; or drawn an incorrect inference or conclusion from an established fact.

Nonetheless, courts are careful to differentiate between challenges to evidence which amount properly to points of law and those which are, at base, disguised or camouflaged disagreements with the legitimate finding on the evidence of the decision maker. For example, in Napp Pharmaceutical Holdings Ltd v Director General of Fair Trading (No 5) [2002] EWCA Civ 796 the Court of Appeal refused to grant permission to appeal where the applicant sought to present as issues of law challenges to primary findings of fact about market behaviour.

In drawing the line, a court will take into account the nature of the jurisdiction exercised by the decision maker being appealed against and will be more reluctant to interfere where the lower court or tribunal applies a specialist expertise to the evaluation of facts. The CAT is one such body: See e.g. Merricks (ibid) paragraph [63]. In Le Patourel the Court of Appeal gave some guidance on when a dispute focusing upon the evaluation of facts could amount to a point of law:

“56.

Where the challenge is directed at a decision about facts prima facie this will not be a matter falling within the jurisdiction of the Court of Appeal because it will be the outcome of an exercise of judgment and not an error of law.  Nonetheless, the exercise of judgment over facts can on occasion amount to an error of law, for example where the decision in dispute is outside the (generous) bounds of that which the decision maker (here the CAT) could properly and reasonably make.  If for instance the CAT were wrongly to place the decimal point one place to the right in an equation relevant to the computation of aggregate damages (thereby magnifying the damages to be paid by 1000% - damages of £1m might become £10m), then the Court might, for example, treat this for instance as the CAT taking into account an irrelevant consideration and correct it.  We consider that issues such as this will very much be the exception and not the rule. 

57.

On the other hand when it comes to the weighing up of the various factors relevant to the choice of opt-out or opt-in this is essentially an exercise of judgment over facts and evidence by an expert, specialist, body, that will over time accrue an increasing well of experience in how to handle these complex cases.  The appellate courts recognise that the case management decisions of the CAT are exercises in pragmatism and that undue formalism and precision are not required:  See the summary of the case law in NTN v Stellantis NV and others [2022] EWCA Civ 16 at paragraphs [24] - [29].  These considerations broaden the Tribunal’s margin of discretion or judgment.  This Court should not interfere simply because it might, for the sake of argument, have drawn a different conclusion from the weighing exercise.  We would expect that most opt-out/opt-in decisions will involve a weighing exercise of this nature.”

The second limit is that the point of law must be “as to the award of damages” or other sum. The expression “as to” presupposes some degree of connection between the point of law and the damages or other sum, but there is no indication in the CA 1998 as to how close that nexus or connection must be. The judgment of the Supreme Court in Merricks(ibid) provides the clearest guidance as to what constitutes “a point of law … as to the award of damages or other sum”. The Court addressed seven grounds all of which were accepted as being admissible and it necessarily follows amounted to (a) points of law and (b) were “as to” damages. The Court held that there were errors in relation to five of the admissible grounds but not in relation to the other two.

Those where the Supreme Court found an error can be summarised as follows:

The following were admissible grounds of appeal where, on analysis, the Supreme Court found there to be no error:

Like the Supreme Court, the Court of Appeal in Merricks ([2018] EWCA Civ 2527) articulated a broad test for jurisdiction. Patten LJ considered that even though the decision of the CAT itself did not address the viability of the underlying claims for damages (paragraph [27]) it was nonetheless: “a decision in collective proceedings as to the award of damages”. This was because the refusal of a CPO was a determination of the Tribunal that the eligibility criteria were not met and the proposed representative was not therefore entitled to seek an aggregate award of damages under section 47C(2) which was a remedy “unique to collective proceedings”. The refusal of a CPO was likely to prevent individual members of the represented class who had suffered loss from obtaining any compensation. It was therefore the “end of the road for a class action of this kind and, as such, a decision as to the award of section 47C(2) damages”. The fact that class members were left with individual claims was “nothing to the point.” Coulson LJ, in a concurring judgment, agreed (paragraph [29]) that a decision not to grant a CPO was a decision as to the award of damages.

This court provided a summary of the above in Le Patourel(ibid) at paragraphs [50] – [55] and in McLaren(ibid) at paragraph [9]. The Court of Appeal has considered whether particular issues are “as to” damageson a few earlier occasions. In Enron Coal Services Ltd v English Welsh & Scottish Railway Ltd [2009] EWCA Civ 647 at paragraphs [23]-[24] (“Enron”) Patten LJ considered that decisions which amounted to the rejection of a claim (such as a strike out decision), and those refusing to strike out a claim were decisions “as to the award of damages” (within the meaning of section 49(1)(b) of the CA Act 1998, which was the relevant provision for the purposes of that case). The expression “as to” damages:

“24.

… was not… intended to limit the disappointed party's right of appeal to decisions of the tribunal either awarding or refusing an award of damages following a full hearing… it is difficult to believe that Parliament intended an unsuccessful claimant to be able to appeal against the dismissal of his claim after a full hearing but not to do so against its dismissal [at an interlocutory stage].”

A narrow interpretation of section 49(1A)(a) was to be avoided. It did not accord with Parliament’s intention in establishing the framework to bring collective proceedings. Interlocutory decisions could have a significant effect as to the damages awarded.

In Paccar Inc and others v Road Haulage Association Ltd and others [2021] EWCA Civ 299 (“Paccar”) the Court was addressing whether the proposed funding arrangements of a collective action amounted to an unenforceable damage-based agreement. It was not in dispute that if the issue was not within the scope of an appeal under section 49(1A)(a) it could nonetheless proceed by way of judicial review. The Court held that it was not appealable. The case was heard after the judgment of the Supreme Court in Merricks was handed down but does not treat that judgment as indicating the permissible bounds of an appeal. At paragraph [55] the Court treated section 49(A1)(a) as “… descriptive of the type of decision from which an appeal may be brought, and not a description as to the type of proceedings in which the decision is made.” This distinction was not arrived at by reference to the analysis of the Supreme Court in Merricks and was affected by the conclusion of the Court that “… there [was] … every reason to suppose that an acceptable way of dealing with the problem [of the proposed class representative] would have been found” (paragraph [59]). Hence the decision of the CAT was not an end of the road decision. Given the omission from the judgment of any reference to the Supreme Court in Merricks it should be seen as a decision on its own facts.

Finally on this point, the approach currently adopted by the CAT to interlocutory decisions was recently set out in the decision on permission to appeal in Merchant Interchange Fee Umbrella Proceedings [2022] CAT 50 (“MIF”) at paragraphs [4] - [22]. The question was whether Mastercard was entitled to appeal the substantive decision ([2022] CAT 31) on grounds concerning the extent of disclosure and witness evidence relevant to the issue of merchant pass-on. The CAT held that whilst there was jurisdiction to appeal (paragraph [22]) the appeal had no real prospect of success and permission should be declined. On the question of jurisdiction, the CAT asked “essentially whether the decision affects the amount of damages to be awarded in some causal way” (paragraph [14]). Further: “… a case where no damages will arise at all because of an interlocutory decision will be a decision as to the award of damages” (paragraph [15]). In pragmatic terms the CAT observed (paragraph [18]):

“Parties before the Tribunal can proceed on the basis that, assuming a point of law arises, contested interlocutory decisions, even of a contested case management nature, can be presumed, for the purposes of permission to appeal applications, to meet the requirement that they affect the final substantive outcome in terms of the level of damages awarded.” (emphasis in original)

I see force in the CAT’s analysis. The test: “whether the decision affects the amount of damages to be awarded in some causal way” highlightsthe need for there to be “some” (sufficient) causal link between the decision and damages. The guidance from Merricks is that the link or effect does not have to be very direct or close. The test is not one capable of being applied with mathematical exactitude. However case law indicates for example: that a decision which brings the possibility of a claim for damages to an end (such as a strike out) is “as to” damages; that a decision going to the amount of a possible claim (for instance a decision that part of a claim is unsustainable) is “as to” damages; that a decision that a claim should not be struck out is “as to”damages, not least because if the appeal prevails the effect is as if the CAT should have struck out the claim; and that decisions as to the procedure to be applied to determine damages claims are also “as to” damages because the procedure adopted could affect the ultimate quantum.

There will however be an outer limit. In argument, citing Paccar, it was suggested that the outer limit was whether the decision under challenge brought the claim or part of a claim to an end. But that analysis seems too narrow. It follows from Merricks that decisions which affect how claims are to be run or adjudicated upon are also “as to” damages even where the decision does not bring the claim to an end. So for instance the Supreme Court treated whether the CAT was right to hold a trial within a trial as a decision “as to” damages and it also held that a dispute about whether distribution should be taken into account at the certification stage was “as to” damages. Disputes as to how broad common law principles apply to the evaluation of evidence relating to damages and as to the judicial tools and techniques at the CAT’s disposal (such as the broad axe) have also been held to be proper subject of the statutory appeal process and are therefore “as to” damages. They are reasonably described as principles of law and procedure which govern how a damages claim is to be determined and they all could ultimately affect quantum.

I am loathe at this stage in the development of the case law to express a definitive view as to how bright the line is as between an appeal and judicial review. I am though clear that the statutory right of appeal should be construed broadly in order to minimise the scope of judicial review. One of the legislative purposes identified by the Supreme Court in Merricks as guiding the operation of the regime was judicial efficiency. Judged through this optic there is only judicial inefficiency flowing from forcing litigants seeking to challenge CAT decisions to go via judicial review or (as in this case), even worse, proceed simultaneously via judicial review and a statutory appeal.

There is no logic in a conclusion that Parliament wished to give an appeal route a narrow scope leaving judicial review with a concomitantly broader scope. To the contrary there are good reasons why an appeal should take precedence over judicial review. First, in terms of judicial hierarchy it makes sense for challenges to CAT decisions to flow, to the greatest degree possible and consistent with the legislative purpose, to the Court of Appeal. Institutionally the CAT is presided over by a specialist High Court Judge and in individual cases High Court judges with suitable experience are routinely appointed to sit as the presiding judge. Judges who sit in the CAT acquire specialist skills and receive specialist training. A CAT panel routinely includes an economist. If judicial review were a normal route of challenge this would entail a challenge from a three person specialist CAT, to a non-specialist High Court judge sitting in the Administrative Court which could then lead to an appeal to the Court of Appeal. Judicial review inserts an unnecessary non-specialist step in the progress of a CAT decision to an appeal. Secondly, it is relevant that in practical terms there is not a great deal of difference (if any) between an appeal on a point of law and judicial review. There is no clear benefit in permitting judicial review to have a broad scope where there is no inherent forensic value to the exercise. Both proceed upon the basis of facts as found by the lower court or tribunal and in both an appropriate margin of discretion or appreciation is accorded to the first level trier of fact, especially if it is a specialist body. The traditional grounds of an appeal on a point of law are closely related to the traditional grounds of judicial review. The authors of De Smith’s Judicial Review, 9^th Edition (2023) observe at paragraphs [16-018] and [16-019] that the powers of an appellate court will encompass all the grounds of judicial review within the rubric “points of law” and might “perhaps” even be greater.

It is also an elemental principle that parties seeking judicial review should exhaust other judicial remedies first. The rigour with which this is applied will always be dependent upon the nature of the “other” remedy: In Glencore Energy UK Ltd v Revenue and Customs Commissioners [2017] EWHC 1476 (Admin) (“Glencore”) at paragraphs [40] and [112] - [115] (upheld on appeal [2017] EWCA Civ 1716) the issue was whether the taxpayer could seek judicial review of a decision of HMRC instead of appealing to the Tax Tribunal. Where Parliament had created a statutory right of appeal it was important that the statutory route be taken advantage of for reasons of resources allocation and expertise. This is a principle which forms part of the Pre-Action Protocol for Judicial Review which, at paragraph [5], states that: “Judicial review should only be used where no adequate alternative remedy, such as a right of appeal, is available. Even then, judicial review may not be appropriate in every instance.” In my judgment wherever an appeal lies in cases such as these, judicial review should not.

Where there is any doubt about the route of challenge, the procedure adopted in this case whereby the Court sits as the Court of Appeal and High Court makes sense and avoids duplicated court time and expense. The occasions when the only issue is one of judicial review should be rare.

I turn now to the first substantive issue. Both applicant class representatives argue that the CAT erred because, on any analysis, their pleaded cases went way beyond the bare minimum needed to pass any threshold strike out hurdle. The Court was taken at length to the pleaded cases of the applicants including to their expert and trade witness evidence to show both the soundness of the econometric approaches proposed, the sufficiency of the data sources available to populate future economic regression modelling, and the way in which trade witness evidence could plug any gaps which might exist in the data. It was also argued that the CAT was wrong to take the issue of its own motion, there having been no dismissal application made by the banks.

It is important to be clear about the challenge. The CAT did not strike out the claims. The applicants live to fight another day. The challenge therefore is to the conclusion of the CAT that it would not, there and then, decide that the pleaded cases were sufficient as opposed to deferring the decision to a later day when the parties could replead having been placed on notice of the concerns of the CAT.

The applicants advance two main arguments: (i) the CAT applied the wrong test in law; and (ii), in any event the claims advanced to the CAT met any test the CAT was entitled in law to apply. The respondent banks retort that the CAT was correct to consider upon an own-motion basis whether the claims were viable and to conclude that the claims as presented were speculative, ill formed and, in the round, “hopeless”. The CAT was therefore justified in taking the weakness of the claim into account in relation to the opt-in v opt-out decision under CAT Rule 79(3). Mr Kennelly KC advanced a forensically spirited dissection of the applicants’ cases though he did so without his clients having adduced any expert or trade evidence or offered any pre-action disclosure.

In my judgment this issue proceeds by way of appeal only. It concerns the power of the CAT in law to consider strike out of its own motion; and the nature, scope and effect of the exercise of its case management powers to regulate how claims “as to” damages should be pleaded.

The CAT has the power, of its own motion, to determine whether a claim is viable. It can do this both at the certification stage and/or thereafter. This is expressly set out in the CAT Rules and was confirmed by the Supreme Court in Merricks at paragraphs [26] and [59]. This power is an important tool in the CAT’s gatekeeper armoury.

It is argued that the CAT applied too onerous a standard and as a result was excessively demanding as to the detail expected to be provided. The appellants rely in particular upon the test at the certification stage. In Merricks the Court emphasised that the regime did not contemplate a merits assessment at the certification stage and expressly said that the threshold was low (paragraphs [44] and [45]). It also held (paragraph [73]) that all an applicant had to establish in a follow-on case was a reasonable prospect of showing “some loss”:

“The fact that data is likely to turn out to be incomplete and difficult to interpret, and that its assembly may involve burdensome and expensive processes of disclosure are not good reasons for a court or tribunal refusing a trial to an individual or to a large class who have a reasonable prospect of showing they have suffered some loss from an already established breach of statutory duty. In the context of suitability for collective proceedings or aggregate damages, it is no answer to say that members of the class can bring individual claims. They would face the same forensic difficulties in establishing merchant pass-on, and insuperable funding obstacles on their own, litigating for small sums for which the cost of recovery would be disproportionately large.”

The case law on the threshold for certification is not the relevant test. In Merricks the Supreme Court recognised that there were exceptions to the rule that merits were irrelevant at the certification stage in that (i) putative defendants could seek to strike out the claim or obtain reverse summary judgment and (ii) the strength of the claim was relevant under CAT Rule 79(3) to the opt-in v opt-out decision.

The issue for this Court is not the same as it would have been had the CAT actually struck out the claims. If that had happened, the Court would have had to decide whether on the claim as presented (including the formal pleading but also the “methodology” proposed and other evidence presented) a viable claim had in law been made out. The issue on this appeal however is only whether as a matter of case management discretion the CAT majority erred in deferring the possibility of a dismissal decision on the merits. Applying traditional strike out law, the CAT did not dismiss the claims. It is not said that the CAT misunderstood the strike out test. I therefore do not agree that the CAT applied the wrong test in law.

The next strand of the argument goes to the heart of the issue. Did the CAT err in failing to accept as sufficient the cases as pleaded? If the CAT erred, then the logical consequence should have been that (i) there was no need to defer consideration of strike out or require any additional pleading; and (ii), the CAT should have taken a much more positive view of the strength of the claims when choosing between opt-in and opt-out (which is relevant to Issue III at section F below). It is important to set out the reasoning of the majority of the CAT.

The CAT majority criticised the applicants’ pleadings as based upon a plausible theory but lacking particulars of evidence. They could not see how the CAT could effectively case manage or ultimately try the claim (Judgment paragraphs [141] and [228ff]). A claim that was plausible in economic theory was insufficient in law to proceed. The CAT majority said under the heading “Market-wide harm”:

“230.

An important aspect of markets and competition is how undertakings in a market respond to an increase in the costs of doing business. The legal analysis to date has focused on the recovery of unlawfully caused costs. The four principal options are set out at paragraph 228 above, but these options are likely to operate not singly, but in parallel, and we would be surprised if these were the only options open to the undertaking: there will at least be variants on these themes. It must also be noted that the picture becomes even more complex when it is borne in mind that an undertaking is unlikely to react to an unavoidable increase in costs immediately. In the short term, an undertaking may well bear an unavoidable increase in costs by making less profit (or incurring a loss or a greater loss), but that is most unlikely to be the undertaking’s response in the medium or the long term. In the medium or long term, the undertaking will seek to maximise its profit and to cover its costs one way or the other.

231.

There is, thus, to the economist, a broad similarity between a cost that is passed from undertaking to undertaking (like the unlawfully excessive MIF) and a cost that represents an increase in the cost of doing business (like the cost of doing business in a market rendered less efficient by unlawful information exchanges). The way in which these costs arise is self-evidently different: but the way they are recovered by the undertaking may in essence be the same. We have no difficulty in economic theory postulating that an increase in costs may – one way or the other – result in an increase in prices. From this, it follows that we have no difficulty in economic theory postulating that a specific and unlawful cost (whether that be an excessive MIF or an unlawful information exchange) may be passed on or transmitted to the market in the form of increased prices. To be even more specific, we have no difficulty (as a matter of theory) in postulating or accepting that information asymmetries in the FX markets (including, but not limited to, unlawful information asymmetries) might generate increased costs to large numbers of participants in those markets, resulting in increased spreads charged to market counterparties.

232.

But economic theory does not, in and of itself, constitute an arguable legal claim. Put as we have put it, to the lawyer it amounts to no more than assertion, bereft of the particularity that is required in order to render the claim triable. Economic theory does not automatically or easily translate into a legal claim. A civil action requires, amongst other things:

(1)

Identified or identifiable claimants.

(2)

Identified or identifiable defendants.

(3)

Some kind of actionable and identifiable harm, caused by the defendants to the claimants.

233.

The economic theory of passed on or transmitted costs provides the answer to none of these questions. An essential problem in articulating market-wide claims of harm, which both the O’Higgins Application and the Evans Application need to have grappled with, lies in translating this possible or theoretical phenomenon into a series of averments capable of being tried in a court.”

In paragraph [234] the CAT majority stated that it had not overlooked the guidance given by the Supreme Court in Sainsbury’s which encouraged a highly pragmatic approach to quantification:

“234.

Before we turn to the specific issues that arise out of the O’Higgins and Evans Applications, it is important that we make clear that these courts are open to claims of market-wide harm, and have a number of tools to deploy in order to enable such a claim to be framed. As to this:

(1)

The courts in this jurisdiction are very much alive to the concept of “effectiveness”. Claimants cannot have imposed upon them insurmountable burdens in establishing their claims. If there are insurmountable burdens, then they should arise not from the rules of pleading, but from the inherent (de)merits of the case itself. The courts in this jurisdiction have shown remarkable flexibility in terms of what constitutes a properly pleaded case in order to ensure that proper cases are not denied access to the seat of judgment. Thus:

(i)

Articulation of the burden of proof on particular issues is of considerable importance, as the consideration in the Supreme Court in Sainsbury’s, shows. Pleadings play an important role in articulating which party, on a given issue, bears the burden of proof. In cases where proof of fact may be elusive, this is important.

(ii)

So far as the quantification of loss and damage is concerned, the courts can and do take a “pragmatic view” and Lord Blackburn’s dictum about compensation being accomplished “to a large extent by the exercise of a sound imagination and the practice of the broad axe” is rightly claimant friendly, particularly when read in light of the low hurdle of actionable loss.

(iii)

More generally, the courts are well able to take account of inferences pleaded out of anterior factual averments, and such inferences are valuable in demarcating the areas where disclosure and factual evidence will be required (from one side or the other) in due course. As we have noted, the pleadings are the critical source for identifying areas of disputed fact, so that the parties and the court can marshal the evidence that will be required to resolve them. Particularly in competition and markets cases – but more generally also – the courts are receptive to expert statistical evidence in support of a pleaded case provided it is not too abstract, theoretical, unrepresentative or uncertain. Furthermore, extrapolation based upon sampling that is not underpinned by statistical analysis may be a perfectly acceptable way of pleading a claim.

(iv)

It may be that in cases involving multiple transactions or claimants, a sample of transactions or claimants needs to be taken and set out in detail, so as to enable a properly extrapolated case to be articulated at the pleadings stage. As we have described, collective proceedings before this Tribunal have the very significant advantage of not obliging the claimant class to plead or prove individual loss. The loss can be articulated by reference to a class or classes of person.

(2)

We do not consider that market-wide harm cases can be pleaded at the level of economic theory only. The facts and matters set out in paragraph are unlikely to be provided with the specificity required to try a claim by theory alone. But we should stress that we are very much alive to the difficulties of pleading a market-wide harm case, and would be open to novel ways of articulating such claims provided they were sufficiently specific to enable the trial processes properly to go ahead.

(3)

We hesitate to be too specific as to how such a pleading might be framed, because this is a matter for the parties, and not the court. But it does seem to us that there are at least two ways in which a case of this sort could be pleaded. As to the two ways that we have identified:

(i)

First, a statistical correlation between infringement and effect on market spreads could be averred. The essence of such a plea would be that whilst the transmission mechanism of an additional and unlawful cost (i.e. the information imbalance) through the market would not be set out or averred, the statistical relation between the infringements found in the Decisions and the effects on the market was such as to amount to an arguable claim that the explanation for this correlation was that the widened spreads were caused by the infringements. In short, such a plea would involve looking at the start point and end point of a causal chain inferring those links out of the correlation between the data relating to the start and end points of the chain (the start point being the infringement and the end point being the spreads in the market).

(ii)

Secondly, the additional cost to the market of the unlawful infringements could be articulated and its transmission through the market described. This would involve the articulation of the links in the causal chain, which the first method side-steps. In such a case, the sort of statistical correlation described in sub-paragraph (3)(i) above would be unnecessary to plead, but there would have to be: (i) some particularisation of how the cost of the illegitimate information imbalance found in the Decisions manifested itself in the market; and (ii) how that additional cost was transmitted or passed on, so as to manifest itself in wider spreads. Again, we are under no illusions that this is a difficult case to make out. It might well involve extrapolation from specific examples. It would also likely require some sort of consideration of how price increases can be passed on in what is a competitive market.

(4)

We are, for obvious reasons, both reluctant and unable to spell out in any further detail how a market-wide harm case might be pleaded. There are, no doubt, other ways of articulating an arguable case. We turn, now, to the pleadings in the two Applications.”

In paragraphs [236] and [237] the CAT majority drew a distinction between a “theoretically plausible” pleading and one which particularised evidence sufficient to take the case beyond the theoretical and into the real:

“236.

We have therefore reviewed the cases articulated by the Applicants in the widest sense, so as to understand the cases they are making. Earlier drafts of this Judgment went to some length in considering all of the pleadings, including those submitted in response to our letter of 20 July 2021, as well as the expert evidence submitted by each Applicant. It is, however, unnecessary to set out the content of these materials in any greater detail than we have done. On the basis of all the materials we have considered, including in particular the expert evidence, we are prepared to proceed on the basis that the claims are theoretically plausible.

237.

However, we are satisfied that the facts and matters necessary to support a proper pleading have not been articulated in the pleadings as they stand in either the O’Higgins Application or the Evans Application. We are satisfied that this is not because of a failure to translate specific details that are contained in the expert reports into a legally framed document. Although the expert reports are detailed, these details amount to no more than a detailed expansion of a theoretical position. Our conclusion is that they do not contain material sufficient to support a proper plea of causation, loss and damage.”

The CAT did not consider that the identified deficiencies could be overcome by disclosure. At paragraph [238(5) and (6)] the majority said:

“(5)

At best, this plea is one where there is the hope – framed as an expectation – that something will “come out in the wash”, probably in the form of the regression analyses that can be conducted in relation to the data that would – if these actions were to proceed – be provided on disclosure, so as to enable a theoretical position to be fleshed out by some kind of freshly articulated case.

(6)

Allowing actions to proceed on a “wing and a prayer” is precisely what Nomura enjoins. Defective claims cannot be allowed to proceed in the expectation – even the confident expectation – that the deficiency will be made good by disclosure. The answer to this sort of problem is pre-action disclosure – and no application along these lines has even been suggested by the O’Higgins PCR.”

In paragraph [238(7)] the CAT majority said that to justify disclosure an arguable statistical case had to be pleaded first:

“(7)

Nor are we confident that the regression analysis would demonstrate the kind of correlation between the infringements found in the Decisions and the movements in the market (in particular, the widening and narrowing of spreads) so as, in and of itself, to make good the causative link between the infringements and the losses alleged. That is because of the multitude of other factors that may affect the level of spreads in the FX market, which will be hard to control for. We make this point about the utility of statistical analysis simply because it underlines the importance, and essential correctness, of cases like Nomura. If there is an arguable statistical case on causation, it should be pleaded first, with disclosure following.”

Under the heading “Conclusions” the CAT majority said:

“240.

The question is whether the level of generality or abstraction contained in the O’Higgins and Evans pleadings is sufficient to amount to “reasonable grounds for making the claim” within the meaning of rule 41(1)(b) of the Tribunal Rules. The short answer to this question is that we have no doubt that this test is not met and that both Applications could be struck out under this rule. We are acutely conscious that translating a phenomenon that may well commend itself to economic theory into an arguable claim is likely to be extraordinarily difficult. We are also well aware of the competing values of (i) the need for clarity and certainty in regard to a case being put forward and (ii) the principle of effectiveness. As the Supreme Court has noted in Sainsbury’s (SC),[i]t is the duty of the court to give full effect to the provisions of Article 101 by enabling the claimant to obtain damages for the loss which has been caused by anti-competitive conduct.” However, we are satisfied that the averments in both Applications lack the specificity to enable them to be tried, and that is both unfair to the Respondents and an impossible burden on this Tribunal.”

At footnote [152] the CAT majority added a critical qualification:

“We want to be very clear that we are not making any kind of determination on the merits. What we are articulating is a deficiency in how the case is pleaded. We are not saying that there is no causal link between infringement and loss. What we are saying is that we do not understand from the pleaded case how this link between infringement and causation arises. We could speculate as to how the case might be put. Or we could – using the economic expertise of the panel – try to “fill in the blanks”. But that is not our function.”

The final conclusion of the CAT was therefore that it should not exercise its strike out jurisdiction:

“241.

Equally, we are in no doubt that this is a jurisdiction that we should not – at this stage – exercise:

(1)

The reason we are in no doubt that these Applications could be struck out is because, as presently framed, we simply do not understand how they could properly be tried. The pleadings give no idea as to how the losses claimed have been suffered, and we do not consider that this Tribunal can effectively manage these cases to trial or at trial; nor do we consider that the Respondents can properly defend themselves in circumstances where – although the nature of the claims are understood at a theoretical level – there is, in reality, no pleaded case on causation.

(2)

However, as we have noted on a number of occasions, these Applications raise novel and difficult questions. In particular, “market harm” cases – where the class sought to be represented consists of participants in a market in which anti-competitive infringements took place – are novel. We accept that this is a new and (in pleading terms) untested area. It is right that the strike-out jurisdiction not be exercised in an area of law that is subject to some uncertainty and is in a state of on-going development, and not without the Applicants having the opportunity to address the concerns we have articulated much more clearly in this Judgment than we did during the hearing. For what we hope are understandable reasons, our attempts to understand the claims advanced by the Applicants have caused developments in our thinking, and it is entirely fair to say that there has not been an opportunity, on the part of the Applicants, to address our final thinking on the question of pleading.

(3)

Accordingly, we consider that, at this juncture, it would be inappropriate to strike out either Application. Rather, both Applicants need to be (and now are) on notice that absent significant amendment and revision a future strike-out application may very well be on the cards.”

I turn to my conclusions. The CAT majority examined causation in real depth. It addressed the facts it considered to be relevant and did not leave out of account facts that could be said to be germane to the analysis. It was cognisant of the approach it was required to take to the evidence. It did not for instance ignore that it had broad axe powers. There was a difficult and finely balanced judgment call to be made, as the existence of the minority judgment demonstrates, and all three judges were keen to emphasise the importance of a clearly formulated case on what was manifestly a complex and novel claim, even though they disagreed as to the intrinsic merits. Ultimately, the CAT has ordered further and better particularisation of the pleaded case on causation. The task of this court is not to decide whether we side more with the minority than the majority. We can ask only whether the CAT was within its broad case management discretion to defer the decision. In my judgment it was.

It was suggested by the respondent banks that in McLaren the Court had imposed an onerous burden on the CAT to ensure that cases going forward were viable and this justified the CAT’s very demanding approach to the merits in this case. On this I do not agree. To be clear, in McLaren the CAT had in its judgment identified “the” central issue in the trial but had then brought its analysis to an abrupt end. The view of this Court was no more than this created a lacuna in the exercise of the CAT’s post-certification gatekeeper role and that it needed to have some “blueprint” for managing the issue going forward. Having identified the central issue, the CAT had to case manage it in some appropriate manner. This Court did not however indicate how the CAT should go about this task nor indicate that the “blueprint” for the conduct of this central issue necessarily had to be detailed. What the CAT would require would be an exercise of its discretion and would be fact and context specific.

I should add one final observation concerning the applicants’ criticism of the respondent banks for the position they adopted. Having declined, no doubt for tactical reasons, to put forward an application to dismiss backed by expert and other evidence and even early disclosure, it is said by the applicants that the banks opportunistically stood on the side line throwing rocks, many predicated upon assumptions or assertions about facts they were unwilling to make good. It is also said that the banks adopted this position to avoid having to give any disclosure, including pre-action disclosure, which might have revealed why the Commission arrived at the conclusions it did in Three Way Banana Split and Essex Express. The CAT has a standalone power to strike out a non-viable case. I do not suggest that the CAT should never adopt this course of action. There is nonetheless a risk where it does so because it compels the CAT to do its own thinking without the assistance of a properly formulated, evidence based, objection from the putative defendants. The CAT does not obtain the same level of assistance from a respondent jumping upon a passing bandwagon whilst, at the same time, keeping its cards far distant from the table. The CAT has a continuing power to strike out non-viable claims which it is in principle entitled to exercise after a defendant has given, for instance, disclosure. If the CAT has concerns, it always has the option to adopt a wait and see approach.

In conclusion the CAT did not err in the application of the appropriate test to the facts.

The second issue concerns the decision of the CAT, by majority, to certify the claim upon an opt-in basis. The test is set out in CAT Rule 79(3) and is essentially at large. The CAT may take into account all matters it thinks fit, including the matters set out in CAT Rule 79(2) and (3):

“(3)

In determining whether collective proceedings should be opt-in or opt-out proceedings, the Tribunal may take into account all matters it thinks fit, including the following matters additional to those set out in paragraph (2) - (a) the strength of the claims; and (b) whether it is practicable for the proceedings to be brought as opt-in collective proceedings, having regard to all the circumstances, including the estimated amount of damages that individual class members may recover.”

The criteria of “strength” and “practicability” are just two of the matters that can be considered. “Strength” is a relative concept which (contrary to a submission advanced to the Court in oral argument) runs from the irredeemably weak to the compelling and includes everything betwixt.

The CAT unanimously held that it had the jurisdiction to choose as between opt-in or opt-out even where the applicants applied only for an opt-out CPO (Judgment [82] – [88]). It was plainly correct in this. Nothing in the CA Act 1998 compels the CAT to accept the choice made by class representatives. Its discretion, in public law terms, cannot be so fettered. Were it otherwise, class representatives would invariably select opt-out thereby making the statutory choice illusory.

The majority held that the proceedings should be certified upon an opt-in basis. In coming to this conclusion, they took account of the following two principal considerations. First, that the strength of the claim was very weak. Secondly, that by making the claim opt-in there was no access to justice deficit even though the CAT found as a fact that were it to make such an order the proceedings would not go ahead. The CAT also considered: (i) that the class representative did not amount to a pre-existing body such as a trade association; (ii) that the funding arrangements were such that there would be an incentive to settle at which point members of an opt-in class would be more likely to express an interest than in an opt-out case; and (iii), that the existence of settlements in the US showed that opt-in claims were practicable. These points were considered to be “weak” indicators of opt-in.

Mr Lomas dissented. He considered that it was illogical for the majority to conclude that absent an opt-out order there would be no claim at all, but then order opt-in proceedings nonetheless. This was antithetical to the principle of access to justice and to the public interest in ensuring that wrongdoers disgorge their ill-gotten gains. An opt-out CPO would provide access to justice for all class members, whereas an opt-in CPO would be neither practicable nor economically viable nor be in the interests of the class members: see Judgment paragraphs [415], [435] - [449] and [455].

The appellants argue that the CAT majority erred:

In my view this argument raises a point of law as to damages and is subject to the right of appeal. The issue concerns: (i) the correct interpretation and application of provisions of the CA 1998 and the CAT Rules; (ii) the proper inferences to be drawn from essentially common ground facts and their relevance to the exercise of the discretion under CAT Rule 79(3); and (iii), the bringing of the claim to an end since, at the practical level, it is common ground that this procedural decision will result in no claim for damages being advanced at all.

A consideration of statutory purposes is relevant to this dispute. The Supreme Court in Merricks held that in construing the rules it was important to interpret and apply them (ibid paragraph [45]) “… in their context as a special part of UK civil procedure and with due regard paid to their purpose”. Lord Briggs (ibid paragraph [37]) approved the threefold description of statutory purposes set out by Chief Justice McLachlin in Hollick v Toronto (City) 2001 SCC 68; [2001] 3 SCR 158 at paragraph [15] in relation to the Ontario Class Proceedings Act 1992:

“The Act reflects an increasing recognition of the important advantages that the class action offers as a procedural tool … class actions provide three important advantages over a multiplicity of individual suits. First, by aggregating similar individual actions, class actions serve judicial economy by avoiding unnecessary duplication in fact-finding and legal analysis. Second, by distributing fixed litigation costs amongst a large number of class members, class actions improve access to justice by making economical the prosecution of claims that any one class member would find too costly to prosecute on his or her own. Third, class actions serve efficiency and justice by ensuring that actual and potential wrongdoers modify their behaviour to take full account of the harm they are causing, or might cause, to the public. In proposing that Ontario adopt class action legislation, the Ontario Law Reform Commission identified each of these advantages … In my view, it is essential therefore that courts not take an overly restrictive approach to the legislation, but rather interpret the Act in a way that gives full effect to the benefits foreseen by the drafters.”

In Le Patourel the Court of Appeal summarised the legislative intention behind the creation of the collective damages regime:

“29.

Pulling the threads together, the principal object of the collective action regime is to facilitate access to justice for those (in particular consumers) who would otherwise not be able to access legal redress. Embraced within this broad description is the proposition that the scheme exists to facilitate the vindication but not the impeding of rights.  Also included is the proposition that a scheme which facilitates access to redress will increase ex ante incentives of those subject to the law to secure early compliance; prevention being better than cure.  Finally, emphasis is laid on the benefits to judicial efficiency brought about by the ability to aggregate claims.”

The respondent banks argue that there is a fourth guiding consideration which is that defendants should not be unfairly burdened or oppressed by this radical new regime which should therefore be strictly applied (in favour of defendants). Such an object is however not an identified legislative purpose and would run counter to the policy objective of ensuring that wrongdoing undertakings do not avoid the consequences of their illegal actions. On the other hand, in applying the broad principles governing the conduct of claims set out in CAT Rule 4, it is the responsibility of the CAT to ensure that defendants are treated fairly and proportionately but within the confines of the legislation properly construed. Taken as a whole the CA 1998 and the CAT Rules strike the proper balance between the various objectives and the need to protect defendants.

Ms Wakefield KC, for Evans, argued that it was illogical for the CAT majority to defer the decision on the merits accepting that the lacuna identified were capable of being filled yet, at the same time, treat what of necessity had to be a tentative and provisional view on the merits as definitive and fixed and attracting decisive weight in the opt-in/opt-out scales. Further, she argued that ifany weight was to be attributed to the merits it could only become relevant after the class representatives had exercised their right to resubmit a reformulated pleading. It was wrong in principle and procedurally unfair to pre-empt that right and treat a provisional view as, in effect, definitive without waiting to see if revised pleadings overcame the CAT’s concerns. I see the force in both these points. If, as was the case, the CAT was prepared to await a reformulated case before arriving at a conclusion on the merits then it could not logically treat its provisional view on the merits as legally definitive.

The second objection is that the CAT failed to explain how or why making an opt-in order would improve the conduct or fairness of the proceedings. Instead, the logic of the CAT was that by making an opt-in order the claim would collapse; in other words an opt-in order was the sanction for a non-viable claim. It is argued that this is an irrelevant consideration to take into account. The applicants add that the CAT did not have the benefit of the subsequent case law of the Court of Appeal in which this issue has been considered; had the CAT been aware of this case law it might have taken a different decision. It is pointed out that in Le Patourel the Court of Appeal, when considering the relevance of strength of the case under Rule 79(3) (ibid paragraphs [104] – [108], observed that in most cases the strength of a case might be neutral as a factor in the choice between opt-in and opt-out. The Court also observed that it might be hard for the CAT confidently to assess the merits of the claim “because of the complexity of the legal and economic issues arising and the absence at the certification stage of expert evidence and disclosure”.

Again I see the force in this. The strength of a claim (either way) is but one relevant factor that might (but need not) be taken into account. Generally, the strength of a claim will be neutral regardless of whether the proceedings are opt-in or opt-out. Though in follow-on cases liability will already be established so (as in the present case) the issue will be as to causation and loss. There might therefore be some relationship between the relative merits and the mechanics of a trial process. Even assuming the CAT was entitled to take its negative view on the merits into account it follows that it still needed to show how that assessment made opt-in preferable to opt-out. The factors the CAT will take into account should bear upon such questions as which option is better able to vindicate the claim, which affords better access to justice and which enables the case to be best case managed from the point of view of judicial efficiency, or by reference to some other relevant consideration (e.g. under CAT Rule 4). By way of example it is said that in the main opt-out aggregate damages claims are likely to be easier to fund than opt-in claims (cf Le Patourel (ibid) paragraph [107]) and therefore likely to result in better run litigation. Further, an opt-out damages claim is easier to prove in terms of causation because there is no need to establish a causal link with individual claimants, that being an issue for distribution which occurs later, and, as the Supreme Court observed in Merricks, was rarely likely to have an impact at the early threshold stages. As matters stand the basis for the CAT’s decision reflects a view that making the order will bring the claim to an end which is not, in my view, a consideration relevant to the choice to be made under Rule 79(2).

The applicants next argue that, contrary to its statement that it was not reviewing the merits, the CAT did in substance apply a strike out test and, having done so, engineered a dismissal of the claim by the back door. As such the Court of Appeal should decide whether the CAT was right to dismiss the claims as lacking merit. There is some considerable logic in this, but the position is not straightforward on the unusual facts of this case. On balance I think the best way to proceed is to accept at face value that the CAT left the merits to be decided in the future and analyse the case upon that basis. This means that the merits will have to be reconsidered when the matter returns to the CAT, now upon an opt-out aggregate damages basis.

However, since the matter was fully argued before us and raises some important points it is relevant to consider the main arguments advanced if only because they could be relevant when the issue returns to the CAT. Three central points were made by the appellants:

The Commission Decisions

The first issue was the subject of detailed oral argument. I have set out at paragraphs [25] – [32] above the findings made by the Commission in Sterling Lads about likely effects. In this case, the decisions relied upon in pleadings as the relevant follow-on decisions are those in Essex Express and Three Way Banana Split. These are short form settlement decisions, drafted in relatively skeletal form, and they do not refer to evidence nor set out any of the detailed reasoning whereby the Commission came to the conclusion that there was a sufficient likelihood of actual harm to competition to justify the making of object based violation decisions. The decisions do not feature in any material way in the Judgment.

The decision in Sterling Lads is of an altogether different nature. It is a disputed decision. Credit Suisse fought hard to refute the Commission’s claims against it. The decision is detailed and refers extensively to evidence. It explains why and how the Commission concluded that the cartel was, on the facts, likely, causatively, to have benefitted cartelists and harmed competitors and counterparties. The CAT did not have this full decision before it. The appellants argue that the decision is highly significant and provides the answer to the CAT’s concerns about theory not translating into reality. The respondent banks argue that the decision is inadmissible but alternatively of strictly limited, if any, probative value. It is contended that the rule in Hollington v Hewthorn [1943] KB 587 (“Hollington”) applies to the findings in the decision, rendering it inadmissible. I do not agree. I address below first whether the decision is admissible and, secondly, if so, the approach to be taken to an evaluation of its evidential weight.

I start with admissibility. The rule in Hollington is that absent the operation of estoppel, factual findings in civil cases in England and Wales are inadmissible in subsequent proceedings. The modern and most oft cited formulation of the rule is in the judgment of Christopher Clarke LJ in Rogers v Hoyle [2014] EWCA Civ 257 who noted that the rule had been extended to findings of facts of arbitrators (Land Securities v Westminster City Council [1993] 1 WLR 286), coroners (Bird v Keep [1918] 2 KB 692) and extra-statutory inquiries (Three Rivers District Council v Governor and Company of The Bank of England (No.3) [2001] UKHL 16, [2003] 2 AC 1). He stated:

“39.

… the foundation on which the rule must now rest is that findings of fact made by another decision maker are not to be admitted in a subsequent trial because the decision at that trial is to be made by the judge appointed to hear it (“the trial judge”), and not another. The trial judge must decide the case for himself on the evidence that he receives, and in the light of the submissions on that evidence made to him. To admit evidence of the findings of fact of another person, however distinguished, and however thorough and competent his examination of the issues may have been, risks the decision being made, at least in part, on evidence other than that which the trial judge has heard and in reliance on the opinion of someone who is neither the relevant decision maker nor an expert in any relevant discipline, of which decision making is not one. The opinion of someone who is not the trial judge is, therefore, as a matter of law, irrelevant and not one to which he ought to have regard.

40.

In essence, as the judge rightly said, the foundation of the rule must now be the preservation of the fairness of a trial in which the decision is entrusted to the trial judge alone.”

There are however a growing number of exceptions to the rule. Under section 11 of the Civil Evidence Act 1968, criminal convictions are admissible to evidence the fact that an offence has been committed. This reverses the position in Hollington which actually concerned a criminal conviction for careless driving. In Re W-A (Children: Foreign Conviction) [2022] EWCA Civ 1118, the issue concerned the admissibility of the previous conviction of a man for sexual offences against a child in Spain as evidence of relevant underlying facts in care proceedings before the Family Court. The Court of Appeal held that it was settled law in family proceedings that the Court could consider and attach weight to earlier findings. Any other approach would conflict with the overriding duty of the Court to discover the truth in the best interests of the child. It has also been held that a subsequent court can “have regard” to the evidence set out in an earlier case as part of the evidence in the later case leading the judge to arrive at the same conclusion: see e.g. Otkritie International Investment Management v Gersamia and Jemai [2015] EWHC 821 (Comm) per Eder J at paragraph [23].

Most importantly, it is well established that the rule does not apply to the CAT which has its own rules of procedure and evidence. CAT Rule 55(1)(b) makes clear that the CAT has a wide discretion as to the evidence to be admitted. This has been recognised on many occasions and is, in my view, correct: see e.g. Agents’ Mutual Limited v Gascoigne Halman Limited [2017] CAT 5 at paragraph [8]; Argos and Littlewoods v OFT [2003] CAT 16 at paragraph [105]; Aberdeen Journals v. OFT [2003] CAT 11 at paragraphs [126] and [134]), Consumer Association v Qualcomm [2023] CAT [9] (“Consumers Association”) at paragraph [18]. In Le Patourel the CAT had relied upon the findings in a prior settlement decision between the respondent, BT, and OFCOM. The Court of Appeal agreed with the CAT that the findings were relevant as showing a serious case to be advanced but made clear that they were not binding upon the CAT at trial (ibid paragraph [106]). And of course, there is already a statutory exception to the rule in section 60A CA 1998.

There is no need for the CAT to be hidebound by a common law rule on fairness. Whilst the CAT does not apply the strict rule in Hollington it does, of course, endeavour to secure fairness but it is a sophisticated tribunal well able to form its own view on the value, if any, of prior findings.

The CAT, if confronted with prior findings said to be relevant, will carefully decide what weight can be attached to those findings. Without intending to be exhaustive, it will examine such matters as: whether the decision is a follow-on decision and the limits of the binding effect under section 60A CA 1998; where not a follow-on decision, the extent of the overlap between the prior findings of facts and the present case; who the earlier decision maker was and whether it was a specialist fact finder or otherwise; what the standard of proof was which was applied to the findings; and the nature of the legal analysis in the prior decision and the extent to which this affects the findings of fact made. The CAT will also consider to what forensic use the earlier findings are sought to be deployed. There might be many relevant uses some of which fall short of reliance upon earlier conclusions about the ultimate merits. The earlier decision might for instance identify relevant evidence and thereby demonstrate lines of inquiry relevant only to disclosure. The CAT will be conscious of the risk that being invited to perform a detailed inquiry into how prior findings came about draws it into disproportionate, satellite, litigation: see Consumers Association (ibid) paragraph [30].

In the present case the CAT will consider the probative value of the Sterling Lads decision when the case returns to it. Various points were made to the Court which seem to me to have some force.

First, this is not a Hollington type decision because it concerned Credit Suisse, not a defendant (Footnote: ³). It is however a decision of the Commission in relation to more or less identical facts to those arising in the instant case. The findings also import a relatively high degree of probative value given the (quasi-criminal) standard of proof the Commission had to overcome before the findings could be made: See e.g. The Competition and Markets Authority v Flynn Pharma and others [2020] EWCA Civ 1847 at paragraph [136] and cases cited thereat. The analysis in the decision is based upon (i) the intrinsically high probability of conduct of this nature causing actual harmful effects and (ii) the evidence in the case justifying a conclusion that tangible harm was in fact likely to have been caused. The decision thus supports the case advanced by both class representatives that their cases on causation went beyond the merely theoretical.

Secondly, the decisions in Essex Express and Three Way Banana Split, which are the foundations for the follow-on, are from the same decisional template or mould as that in Sterling Lads. As Ms Wakefield KC for Evans submitted, a side by side comparison of the various decisions showed many paragraphs which were common across all the decisions. In her submission, by a process of logical reverse extrapolation and inference, the ordinary decision in Sterling Lads indicated what was hiding behind the short form decisions in Essex Express and in Three Way Banana Split. When confronted with short form decisions, care therefore had to be taken to avoid underestimating their potential relevance which might only become apparent after disclosure.

Thirdly, in an aggregate damages, opt-out, case generalised findings, such as might be found in an object violation decision (such as Sterling Lads), might have greater value than in an opt-in case. None of the decisions relied upon, including Sterling Lads, addressed how harm would be distributed as between the member of the class or categories thereof. However, in an opt-out case there may be less need for the CAT to delve into such details at the certification stage: see Le Patourel (ibid) paragraph [107]. It is only once the aggregate loss is determined that questions of distribution arise, and these can be resolved by ADR or by some flexible method of distribution (e.g. Le Patourel (ibid) paragraphs [87] – [99] and Gutmann (ibid) paragraph [87]). In other words, in an opt-out case questions of causation might be simplified. Given that the principle of aggregate damages involved a “radical” (see Merricksibid paragraph [58]) departure from normal tortious principles of compensation there remains scope for the CAT to apply the flexibility as to remedy which the Supreme Court emphasised existed in Merricks. There, in relation to distribution, the Court made clear that where any attempt to calculate individual losses was impossible the CAT was entitled to adopt “…some other method [which might] be more reasonable, fair and therefore more just” (ibid paragraph [77]). This is an indication of the flexibility with which the CAT might approach issues of causation and damage. For instance, in a case where there is evidence that defendants have gained from their unlawful conduct but it is difficult to determine to what extent this caused loss to the class, it might be possible for the Court to adapt less standard remedies, for instance by ordering disgorgement of profits. One purpose behind the collective damages regime is the public interest in wrongdoers not retaining their unlawful gains. The use of an account of profits, or some equivalent remedy, is by no means unknown in tort claims. It might even be considered that the gain made by the wrongdoers amounted to a proxy for the loss of the class.

Fourthly, whilst it was true that the Sterling Lads decision had little if anything to say about the existence of umbrella, Category B, damages (see paragraphs [35]-[38] above) the extent to which such damages are recoverable might depend upon findings by the CAT at trial about the manner in which Category A damages were sustained. One argument advanced by the respondent banks was that the traders concerned amounted to a few, random, individuals who would never, either individually or collectively as a cabal, have been able adversely to affect competition in either (a) the Category A segment of the market as a whole or (b), by extension of the same reasoning, the Category B segment of the market as a whole. The appellants in response refer to trade witness evidence from an FX trader who gave evidence on how FX market trading floors actually operated. It was explained that the FX traders concerned in the chat room cartel were relatively senior in their companies and operated in a trading room environment where they communicated real time information to other traders physically proximate to them who did not participate directly in the chat room but who were, thereby, able to benefit from the same artificially translucent trading conditions. Moreover, the defendants collectively accounted for up to 45% of the world FX trade so that the collusive contagion could have affected up to this volume of trade in the affected market. And if this was so, it was capable of causing a widening of spreads sufficient to have been visible to non-colluding banks who could, lawfully and unilaterally, have followed suit broadening their spreads accordingly leading to concomitant gains for them and losses for counterparties. The point goes only to show that a decision of relatively limited scope might act as a starting point or platform from which a broader case using statistical and witness evidence could be built to establish both Category A and B losses.

Fifthly, it was proper to infer – and should be inferred – that the respondent banks would not have engaged in highly risky illegal and collusive behaviour unless they were confident that they would in actual fact extract financial gain from their conduct. A court might take as a starting point an inference that there was an actual gain where it could be assumed that the wrongdoers understood the regulatory and/or reputational risk of participating in unlawful behaviour but persisted nonetheless. Similar points were made in McLaren (ibid) paragraph [42]; and Stelianos (ibid) paragraph [73].

Disclosure

The next point concerns disclosure. The CAT majority was sceptical that disclosure would prove to be valuable (see paragraph [72] above). However, the parties took the Court through the many different sources of data that might be available, including from third parties. This included material that would be relevant to a traditional regression analysis. They pointed to the decision in Sterling Lads as indicating how detailed the documentary evidence was that enabled the Commission to draw conclusions about the effects of the cartel in that case and they argued that this sort of evidence was likely to be available after disclosure from the respondent banks and would shed light on issues of cause and effect. Some of this could even be circumstantial. For example evidence of traders’ bonus and performance related pay might show how the pay of cartelists was affected (boosted) by the illicit exchanges of information.

One illustration referred to the Court showed the different periods in which different banks participated in the cartel (set out in Annex A to this judgment). This meant that it was possible to examine the trading record of the defendants during periods when they were in the cartel which could be compared with periods when they were not in the cartel but during which the cartel was still operating. They also had data which would enable the claimants to analyse the trading performance of the banks in the period before and after the cartel. All of this may be valuable in providing a test bed for regression analysis because modelling could apply to periods when the cartel did and did not operate and during when each defendant did and did not participate in the cartel.

As an indication of the scale of the data likely to be available from disclosure, the appellants relied upon evidence as to the disclosure exercise conducted in equivalent US proceedings. A witness statement summarised the data collection exercise in the US which gave an indication of the scale of the data available:

“16.

During the course of the US Proceedings, the US defendants have produced very large quantities of information and data, by way of discovery. Specifically, I understand that, as at 12 January 2018, the following had been disclosed by the US defendants:

a.

Documents: The US defendants produced approximately 1.6 million documents, amounting to more than 16.5 million printable pages.

b.

Transaction data: Additionally, the US defendants produced over 7,000 files of transaction data from over 30 different trading systems, amounting to approximately 10 billion rows, occupying 4 terabytes.

c.

Third party transaction data: Further, Lead Counsel also obtained – pursuant to subpoenas – an additional 2.5 terabytes of data from non-party sources, including Hotspot, Reuters Matching, EBS, and WM/Reuters.

17.

My colleagues at SSAAL believe the transaction database to be one of the largest ever assembled for use in a single piece of litigation. To produce this database, extensive work was conducted with Velador including data ‘cleaning’ (as the US defendants produced the data unfiltered) and data ‘normalization’ (as the data came from so many sources, it had to be put into uniform data extracts). In connection with the data cleaning and normalization processes, Velador had to develop over 1,000 scripts of code.

18.

A large proportion of these documents and databases obtained in the US and referred to above constitutes US Confidential Material. The Proposed Representative, the legal and expert teams and I have not seen any such US Confidential Material, but I can confirm from my review of the publicly available sources and from conversations with SSAAL partners that the US Confidential Material (together with other information) has been used to form the basis of the expert work in the US Proceedings and to implement the US Plan of Distribution for Option 1 Claims in the US Settlement.

19.

I have no reason to believe that, following disclosure in these proceedings, the legal and expert teams retained by the Proposed Representative will be unable to similarly process the disclosed materials for use in damages quantification and, ultimately, distribution of damages.”

It is quite impossible for this Court, exercising an appellate function, to immerse itself in the minutiae of the data so as to be in a position to form any definitive view of the availability and value of data in the context of complex econometric modelling. I endorse the view of the Supreme Court in Merricks that the test of necessity has to be relatively high level. From a reading of the detailed expert reports and the data sources they describe to be relevant to a regression analysis, and in the light of the findings in the Commission decisions, it does though seem to me that the disclosure exercise is intrinsically likely to generate relevant material, especially if gaps can be plugged with witness statement evidence and the judicially wielded broad axe.

The tools available to the CAT

The appellants say finally that this being the first case before the CAT following the Supreme Court in Merricks the CAT did not show the institutional confidence that it has shown in more recent cases where it has demonstrated that it is much more robust in addressing evidential problems such as those identified in the present case.

Almost all damages claims rest upon some species of regression analysis but virtually all such modelling suffers from a variety of reliability risks which are well recorded in the literature, accepted by economists, and commented upon by the courts in the past: See e.g. BAT and others v Secretary of State for Health [2016] EWHC 1169 at paragraphs [598ff]. The judgments of the Supreme Court and the Court of Appeal in this field have expressed confidence in the ability of the CAT to use the extensive powers at its disposal to overcome such evidential obstacles. More recently the CAT has evolved an array of case management and evaluative techniques to overcome what, at an earlier point, might have seemed to present almost insuperable forensic obstacles. The willingness of the CAT now to resolve difficulties is illustrated by the approach adopted in a number of recent decisions of the CAT. I do not however go into these cases because some are subject to pending appeals. I would though observe that in Royal Mail Ltd v DAF Trucks Ltd [2023] CAT 6 at paragraphs [475] – [481] the CAT, being critical of the accuracy of the regression analysis evidence before it, relied upon the wielding of the broad axe to resolve differences between the experts, rather than engaging in a detailed issue by issue evaluation of each expert’s conclusions (an exercise which the CAT decried as one of seeking “spurious accuracy” (paragraph [479])) (Footnote: ⁴). It does seem to me that such recent practice evinces a more confident and robust approach to evidence, which is generally to be welcomed.

The minority judgment on strength

I would, finally, refer to the analysis in the minority judgment of how to approach the strength of the case. I see the force of the points made. In the view of Mr Lomas, the strength of the case was of limited value given the absence of witness and expert evidence, and of disclosure. The experts had put together a detailed explanation of how they would conduct a regression analysis, but they had not yet done so given that they needed disclosure and witness statement evidence to populate and inform their modelling. An assessment of strength had to be “high level” and “clearly” could not be “detailed or constitute a mini-trial, not least in a case like this where the strength issues essentially relate to quality of evidence that will be obtained in the future to support a theory which, albeit not particularised in detail, is tenable intellectually, even if difficult to prove, and which must be considered at trial on all the evidence” (paragraph [453]). Mr Lomas thought that it was premature for the CAT to arrive at a conclusion on strength which could be factored into a decision on opt-in or opt-out, even if the concerns of the majority about causation were warranted:

“428.

… However, whilst recognising those very considerable concerns, although we have heard extensively from experts as to how they would approach that exercise, it has not yet been done, at least in evidence submitted to us (there are suggestions that it has been done in other related cases), and we have not yet heard other factual witness evidence that there might be relevant to the PCRs’ case. Accordingly, there are limitations as to the weight that can be given to the Strength criterion at this stage of the case.”

Mr Lomas also concluded that an approach indicating that stronger cases got opt-out and weaker cases got opt-in, was lacking in legal foundation. There was no basis in the legislation for the creation of a sliding scale of strength whether measured in percentage chances or otherwise (paragraph [431]). It would be an unworkable test to apply at the early stages of litigation of this complexity. To the contrary, the nub of the test to be applied was essentially practical: “…which of the two processes is the better one for having the merits of the PCMs’ case determined by the Tribunal” (paragraph [433]). The test was not:

“… whether this case is so weak that a CPO should only be granted on an opt-in basis because that will mean that it will not proceed (thereby saving costs and time) unless, contrary to the evidence before us, a sufficient number of PCMs now decide that they want it. That is not how the Tribunal Rules are set up.”

Pulling strands together, the argument is that the CAT erred in adopting an overly strict approach to the evidence needed to establish a viable case on causation. For the reasons set out above the criticisms advanced by the CAT are premature. New reformulated cases can be resubmitted to the CAT, this time in the context of an opt-out aggregate damages claim which will simplify at least some complications. The CAT will have the advantage of the fuller decision of the Commission in Sterling Lads to guide it. It will also have the accrued experience of over 30 cases where it has honed its ability to manage large scale collective actions. It will also have the advantage of a growing body of appellate case law which addresses how the relative strength of a case can have relevance under CAT Rule 79(3).

The second central component of the CAT’s analysis of opt-in v opt-out concerns practicability. The appellants argue that the CAT majority erred in the inferences drawn from the undisputed fact that absent an opt-out order there would be no claim. The CAT inferred that the class members were large and sophisticated entities that could afford to bring proceedings and that if they did not do so this was due to a deliberate decision on their part that they did not “want” to litigate. As such, the majority concluded, there was no access to justice deficit. In paragraph [381(9)] of the Judgment in this case the CAT majority said:

“Nor can it be said that the putative class members will be ignorant of these potential claims. To the contrary, the efforts of Hausfeld – contacting 321 firms – evidence that it appears not to be ignorance that is preventing a rush to join the proceedings. Rather, there appears to be a deliberate decision not to participate. We are conscious that we have not heard directly from any members of the putative classes. It may be that putative class members are so unimpressed with the claims that they do not wish to be associated with the actions; or it may be that those sufficiently interested have joined the Allianz proceedings; or it may be that the class members are so apprehensive about joining the proceedings because of the potential reaction of the Respondents that they are deterred from doing so; or it may be that decision-makers simply cannot be bothered to consider whether it is in their firms’ interests to opt in or not. We have no material on which to base so specific a conclusion. We can only say that we can see no reason why it is not practicable for the putative class to join on an opt-in basis, given all the circumstances and in particular given the general sophistication of the putative class, the class knowledge, and the potential size of claim. The inference (and we consider it a strong one) is that potential class members are not opting in because they do not want to, and not because opt-in proceedings are not practicable.”

In Merricks the Court explained that access to justice was a concept associated with practicality and proportionality. The statutory purpose was one “… of providing effective access to justice for claimants for whom the pursuit of individual claims would be impracticable or disproportionate…” (paragraph [37]). I start therefore by considering the evidence before the CAT which relates to practicability and proportionality. The CAT did not disagree with this evidence. It stated that it was “very helpful background” (paragraph [380]). The issue is as to the inferences to be drawn from the evidence.

First, there is the evidence describing the unsuccessful efforts made to enlist potential class members to proceed upon an opt-in basis. The main explanations for this were: the mismatch between the size of the claim and the costs and the inability to secure funding for such litigation; relationship harm; and organisational and logistical difficulties. A statement prepared by a lawyer acting for Mr Evans described the process undertaken to contact and persuade potential claimants to join a collective action opt-in group:

“Following this substantial exercise, and after having contacted approximately 321 organisations, Hausfeld were instructed by 14 clients to provide legal and strategic advice in connection with potential claims in respect of FX misconduct. However, after investigation, it transpired that their combined claims were not of sufficient size to bring a viable group claim bearing in mind the costs of litigating complex claims against several major banks. Many of the individual claims were estimated to be in the low (single digit) millions or less. This made it practically impossible to put in place funding, given that the likely budget required to pursue the claims (including the costs of acquiring adverse costs insurance) was likely to be similar or even exceed the total estimated damages for the group.

Unwillingness of potential claimants to commit to opt-in proceedings

As is clear from the above, despite the significant time and effort invested in identifying and approaching potential claimants, only a fraction of those we approached agreed to be retained for the purpose of pursuing claims. Based on my discussions with many of these potential clients, I believe there were four main reasons for their reluctance to participate:

a.

A key concern expressed by many of the organisations we contacted was that they did not want to embark on a legal fight with major banks for what seemed to them to be a modest to small level of potential damages (in most cases single digit millions or less). In particular, many of the organisations maintained ongoing relationships with those banks – including corporates and hedge funds and asset managers who worked in the finance sector – and did not want to take any steps that might cause conflict in those relationships.

b.

Several organisations, including sophisticated financial institutions, expressed a reluctance to invest time navigating their internal approval processes, which would require obtaining approval from senior managers, considering funding proposals and documents, identifying, retrieving and reviewing the relevant documents and determining whether the potential benefits of a possible damages payment outweighed the downside of allocating their internal resources for this purpose. I was also informed in many of these discussions that committing to litigation could also trigger reporting considerations within many organisations, and many of them were reluctant to draw attention to this issue internally.

c.

Many potential claimants expressed concern that participating in litigation would require them to share confidential business information with the other claimants, the defendant banks and/or more widely. Although it is possible to obtain protective court orders, such as a confidentiality ring, to protect against wider disclosure of confidential information, I was not able to give potential claimants a watertight guarantee that these measures would be put in place, because this would ultimately be a matter for the court to decide.

d.

The process of engaging a client itself often proved to be a practical deterrent for potential clients. Many of the 321 organisations and individuals we spoke to told us that gathering historic FX transaction records, along with any other relevant evidence to support their claim, would require a significant amount of internal resource. Several told us that they no longer retained the relevant records or only had partial records. For many, this was an unattractive prospect in circumstances where, at that stage, the outcome and quantum of the potential claims were uncertain.”

Witness statement evidence from the founder and managing director of a high frequency high volume FX trader was provided which elaborated upon relationship risk. He explained that participating in litigation could compromise the relationship between trader and bank and if this transpired other settling banks might refuse to provide prime brokerage services which would have the effect of shutting the trader out of the market. The witness stated: “The financial services industry is heavily dependant on relationships, trust and who you know. Customers are often reliant on the banks’ services such as access to credit or financing, or for prime broker services. This means there is not the same perceived bargaining power as there might be in other commercial relationships. I believe this leads to an inherent nervousness about suing banks, meaning any disputes will often be resolved informally”.

Data before the Court places the issue into context. A table (see Annex B to this judgment) drawn from expert evidence tendered on behalf of the Evans team, provides a breakdown of the average claim per class member both on an aggregate basis but also divided up into categories based upon type of financial institution concerned (as class member) and annual turnover. There are in total over 18,000 financial class members. The great majority are modest in size measured by turnover (approaching 16,000 with turnovers of £0-499,000). Whilst the average claim is £133,805 the typical claim for most class members is c. £16k. The CAT pointed to various uncertainties in the data, but it has not been suggested that it does not paint, in overall terms, a broadly accurate picture. Nothing in this data is, in my view, inconsistent with the explanations given in the evidence for the reluctance of potential class members to join an opt-in action.

With respect to the CAT, it is now clear from case law that where there would be no proceedings save on opt-out terms, that is a powerful factor in favour of a claim being certified as opt-out. Access to justice is not just about the size and sophistication of the class members, but encompasses also the size of the claim and whether it would be proportionate or practicable for the class members (whatever their size and degree of sophistication) to commence proceedings to recover that loss. In the present case even for the largest class members the sums at stake are relatively modest and on an opt-in basis could be dwarfed by the costs. There was argument before the Court as to whether access to justice was only for consumers and SMEs. There is reference in the documents leading to the adoption of the legislation that the new regime was to protect consumers and SMEs but there is nothing suggesting that it was limited in that way, and the admissible background documents refer to both SMEs and also to businesses more generally. Further, the legislation is drafted in broad and unlimited terms. At all events the evidence shows that a large portion of the class would be SMEs.

In Le Patourel the appellant (the putative defendant) argued that the only issue relevant to practicability was the identifiability and contactability of potential claimants. It was wrong in principle to draw a distinction between a class member’s willingness to opt-in at the outset and the willingness to opt-in after a favourable ruling on liability and the making of an aggregate damages award as part of the distribution process. The CAT disagreed and held that the real life convertibility of potential class members into actual litigants was a relevant consideration. The Court of Appeal agreed and stated:

“73.

In our judgment, and in line with the observations expressed in Lloyd and in Merricks, the CAT was entitled to conclude that if an opt-in was ordered the take-up could be very limited. Indeed, this seems to us to be a more or less obvious conclusion to arrive at on the facts. Both judgments demonstrate that the practicalities of collectively organised litigation might favour an opt-out solution where there are large numbers of potentially affected parties and relatively small sums at stake which might otherwise deter the take up of opt-in proceedings. 

74.

The ability of a claimant to convert identifiable contacts into litigants is hence an important factor which goes well beyond issues of identifiability and contactability.  The Tribunal examined relevant factors such as size of class, the scale of a possible award and the impact of these on funding as important considerations.  These might be sufficient, by themselves, to justify an opt-out decision.  The CAT also considered the more subjective characteristics of the class including age profile, social class and technical ability. These are case specific factors which can serve to reinforce an opt-out decision.  The CAT came to specialist conclusions which lay squarely within its broad margin of judgement.  There is in our judgment no basis in law upon which this court can properly interfere.”

In relation to the conclusion of the CAT that being large sophisticated commercial entities it was by that fact alone practicable for them to join an opt-in class, a similar argument about “doability” was rejected in Le Patourel:

“83.

… The concept of “practicability” is not defined in the CA 1998 or the Rules and it is not “the” test but simply “a” matter the Tribunal is entitled to take into account... . Ms Ford QC for BT did not demur from this analysis of legislative language.  She did argue that practicability meant “doability”; if it can be done then it is practicable and if it is therefore practicable then it pointed powerfully in favour of an opt-in process.  With respect we do not agree.  Practicability includes being “doable” but goes further; it requires the court to ask whether it is not only “doable” but also reasonable, proportionate, expedient, sensible, cost effective, efficient etc, to do it. There are many things that might be doable but where to do them would amount to a poor exercise of judgment.”

With regard to knowledge, which the CAT majority considered relevant, knowing about a claim does not thereby make it practicable to participate in it. Indeed, knowing about a claim might be the necessary precursor to forming a considered view that it was not practicable.

As to the scale of the claim the data relied upon (see Table at Annex B), even accepting that it is based upon a variety of assumptions, is not said to be wholly inaccurate or unreflective of reality and the respondent banks say that the quantum is wildly overblown so that average recoverable claims might in practice be considerably lower than the averages set out in the Table. The scale of the claims set out in the Table, assuming them therefore to be towards the upper end of recoverability, can therefore reliably serve as a benchmark for practicability. The table indicates relative to costs that the scale of typical claims is modest and the size of typical claimants is not large. In paragraph [381(6)] of the Judgment the CAT sets out that the average claim across the class is between £50-£60,000. The CAT says of this: “… it is clear that these are – on average – not insignificant individual claims, and that large institutions would have the potential to claim really quite large sums of money.” It might be that the total claim, as estimated, is between £2.1b and £2.7bn; but it is the claim of the individual that determines whether that individual is willing to become embroiled in litigation. I would respectfully disagree with the inferences the CAT has drawn from the Table.

Finally, contrary to the submission of the respondent banks, access to justice is not the only lodestar which guides this issue. Two other principles are important and relevant. In Merricks, at paragraph [54], the Supreme Court endorsed the proposition that the “…evident purpose of the statutory scheme was to facilitate rather than impede the vindication of those rights.”The collective action regime enabled: “… whole classes of consumers to vindicate their right to compensation and the large cost of the necessary litigation to be funded before an expert tribunal…”. The Supreme Court also highlighted the importance of the regime being applied in a manner which encouraged compliance with the law. The creation of strong enforcement powers “… serves as a disincentive to unlawful anti-competitive behaviour of the type likely to harm consumers generally” (paragraph [2]).  Anticompetitive conduct would not be “effectively restrained” if wrongdoers could not be “brought to book” by mass claims (paragraph [53]). As was pointed out in the dissenting judgment, both of these factors are relevant in this case and point in favour of opt-out proceedings.

In paragraph [383] the CAT identified three factors pointing “weakly” in favour of opt-in. These were: (i) the absence of a pre-existing body (paragraph [370(3)]); (ii) the nature and inadequacy of funding and the impact this could have on settlement decisions (paragraph [370(5)]); and (iii), the existence of the Allianz proceedings (paragraph [372(3)]). The issue is whether these are factors with a relevant impact upon or connection to the choice between opt-in and opt-out. The CAT addressed these issues shortly and they were not material to the outcome of the CAT’s decision. I address each briefly.

Absence of pre-existing body: In paragraph [370(3)] the CAT said:

“Pre-existing body. Neither the O’Higgins PCR nor the Evans PCR is a “pre-existing” body: … We consider that this is a factor pointing away from certifying on an opt-out basis. If we had before us a trade association, whose established purpose it was to represent a specific class that had suffered alleged harm, but (for good reason) found it difficult to corral members of the class into opting in, that would be a factor in favour of certifying on an opt-out basis. It seems to us that the fact that both PCRs in this case have come forward, not at the behest of the class, but at the behest of the lawyers they now instruct (who have themselves failed to “build a book”) is an indicator against certifying on an opt-out basis.”

Both appellants take a similar stance and say that the point is neutral and does not bear upon the ability of a representative efficiently to run the litigation. They point out that in many cases there will be no pre-existing trade association to assume the burden of carrying the litigation but even where there is there is no guarantee it will be capable of running the litigation efficiently. Further, many claims even run by existing trade associations will be viable only because of third party funders. In my view it is relevant that the approach of the CAT was drafted as a point of general application; claims pursued by lawyers and funders through special purpose vehicles carry an automatic negative weighting on the opt-out side of the scales. With respect, there is no such general principle or preference. Third party funders and legal representatives, who act as the motor force behind claims, for profit, are integral to the viability of many claims. Insofar as this creates a risk of abuse or misuse the CAT can exercise control through cost control and other case management measures. I cannot however see how it can be treated as a feature weighing generically against opt-out proceedings, even marginally.

Level of fundingand settlement incentives: In paragraph [370(5)] the CAT was concerned that the inadequate level of funding created a risk that representatives would be forced into the risk of early settlement in which case opt-in was preferable. The CAT said:

“(i)

Opt-in collective proceedings will have some – even if limited – involvement of the class, who might be expected to have views about any proposed settlement.”

(ii)

In the case of opt-out collective proceedings, the only safeguard is the scrutiny of the Tribunal. Of course, the Tribunal will discharge its obligations conscientiously and carefully, but the reality of this case is that quantum is hugely uncertain and – in any application for approval of the settlement – the Tribunal will be faced by both the class representative and the Respondents saying “this outcome is better than litigating to trial and judgment.”

(iii)

The level of funding does, therefore, slightly inclines us against opt-out collective proceedings for this reason, but the point is, we consider, a marginal but not immaterial one, of relatively little weight. We take it into account.”

The appellants argue that if anything the point indicates that opt-out proceedings are superior because they are more attractive to funders which means that funding will be larger than in opt-in proceedings and the risk identified by the CAT mitigated. There might be some force in this. On the other hand I can see that in an opt-in case the class members, being directly involved, might have more of a say over settlement discussions, which might be an advantage. And I can also see that in an opt-out case there might be particular pressures and incentives to settle which might not be wholly satisfactory, a point made more fully in paragraph [88(3)(v) – (vii)] of the Judgment. On balance, the CAT accorded this only marginal albeit some weight and the point makes no difference to the outcome of the appeal. I am inclined to accept, like the CAT, that in some cases it might be a factor which could attract some modest weight.

The Allianz proceedings: This concerned the existence of parallel proceedings. At paragraph [373(3)] the CAT said:

“… The risk of overlapping claims tells differently as between opt-in and opt-out collective actions. If there is a risk of overlap – and it is, in this case, very difficult to tell, which is part of the problem – then it is better to ensure that class members take the conscious decision to opt in, rather than being obliged to consider opting out. Furthermore, the Allianz proceedings are an indicator that there is an appetite to bring this sort of claim, albeit as an adjunct to instances where individuated or direct harm (through entering into a specific FX transaction at the wrong rate) has also been caused. This does, however, support (albeit marginally) the sense that the putative class members are choosing not to involve themselves in the proceedings the Applicants wish to bring on their behalf. We return to this point in the next sub-paragraph (but will avoid “double counting”).”

The appellants argue that the existence of the Allianz proceedings is irrelevant. Evans challenges the CAT’s conclusion that there is overlap between the claims, pointing out that its claim expressly excludes transactions covered by other court proceedings. Further, the parallel proceedings were brought by a group of well resourced companies that traded substantial volumes of FX. In contrast, in this case, the CAT accepted as a fact that companies would not suein the CAT on an individual or opt-in basis. In my view this point is part of the broader issue about the viability of opt-in proceedings, which I have dealt with above. I can see that, in principle, if there are overlapping individual or opt-in claims in a comparable domestic or foreign jurisdiction, that might indicate that opt-in is practicable and as such one factor to go into the balance. In the present case, however, evidence has been accepted indicating that absent an opt-out order no claim will be pursued in the CAT. The CAT says that this reflects the choice of the companies concerned, which is plainly correct but, with respect, does not indicate that the choice is irrational, contrived or false. In light of the statistical data (see paragraph [121] above and Annex B), which evidence has not been rejected, the fact that the situation might have been different in other proceedings does not, on the facts here, mean that the present claims in the CAT are practicable.

The CAT decision was taken at an early stage in the evolution of the jurisprudence and it did not have available to it as guidance the analysis in subsequent case law. The two factors of relevance to this decision are (i) strength of claim and (ii) practicability. The other (“weak”) factors are marginal and do not affect the final decision.

On the two relevant factors, in my judgment the CAT erred. First, in relation to “strength”, having concluded that it would form no final view on the merits pending the submission of reformulated cases by the applicants, it was wrong to treat that necessarily provisional view as definitive and accord it more or less decisive weight in the scales against opt-out, knowing and intending that this would bring the claim to an end. Further, it was necessary, in any event, to link any conclusion the CAT did have on strength to the choice it had to make. As this Court held in Le Patourel, the legislation creates no predisposition for or against any outcome and in most cases the merits will be a neutral factor. However, insofar as they are not neutral, there needs to be a relevant connection with the choice to be made. In this case, as the dissent by Mr Lomas suggests, it is wrong to treat strength as a sliding scale with a weaker case going to opt-in and a stronger case to opt-out. Secondly, in relation to practicability, the statistical evidence, which in broad sweep is unchallenged, explains why opt-in is impracticable. I respectfully disagree with the inferences drawn from this data and from the evidence by the CAT majority. That being so there was no reason why the proceedings should not proceed upon an opt-out basis.

For completeness, I refer to the analysis in the dissenting judgment. This largely chimes with the reasoning I have set out above. In respect of practicability Mr Lomas focused upon how to evaluate the evidence:

“435.

In creating an opt-in class, it would be necessary to establish a critical mass of core claimants to make such a claim viable as an action. The (formidable) costs of bringing this action are not materially dependent on the size of the class. However, the total size of the damages claim is critical because it supports the funding to pursue the claim. That is a function of the number of class members and the size of their claims. In essence, that total likely damages claim has to be large enough for the economics of bringing the claim, with its costs and risks, to be rational. Once sufficient (presumably larger) claimants opt in so that point is reached, and a claim is viable and proceeds, there is then a separate issue of the extent to which it is possible to contact other PCMs to give them a fair opportunity to join the class. In this sense, practicability has two elements: (i) would a claim happen at all; and (ii) if it did, would it be practicable to bring the claim to the attention of the remaining PCMs to give them a fair opportunity to consider whether they should opt-in.”

And later:

“447.

The extent to which an opt-in CPO is practicable is not a binary issue but a matter of degree and to be assessed in particularly uncertain circumstances (not least since there is limited current experience in the creation of an opt-in CPO and none, that we are aware of, in analogous circumstances). This is, not least, because (a) a degree of impracticability can be overcome by the application of greater effort and resources and (b) the concept of an opt-in CPO that is practicable must include some assessment of how widely it meets the interests of the PCMs as a whole (rather than, say, just a core element).

448.

Logically, however, the less practicable an opt-in CPO would be, the more the discretion should be exercised in favour of opt-out. In the limiting, perhaps theoretical, case where it is clear that an opt-in CPO is actually impracticable (in the sense that it cannot happen), it seems that ought to be an exclusionary test for the opt-in approach. It can make little sense for the Tribunal to order a CPO (which, by definition, has already passed the Authorisation Condition and the Eligibility Condition to be awarded a CPO) on a procedural basis which means that it will not happen.”

Having set out the basic facts and observed that the respondent banks had not adduced their own evidence or challenged that of the class representatives, Mr Lomas concluded:

“449.

In this case, the only evidence that we have is from the PCRs. It does not formally establish that an opt-in CPO is formally impracticable (impossible), it does establish (on a basis unchallenged by other evidence and certainly unrebutted) that, at the least, it is very unlikely that an opt-in CPO would proceed at all and certainly that a large percentage of the 40,000 UK based PCMs would opt in. In my opinion, this is a factor that weighs heavily in favour of an opt-out CPO. That reflects both the logic of the position and the overriding objective that we are balancing the respective interest of the PCMs, the Respondents and the administration of justice. I cannot see how we respect the interests of the PCMs and the administration of justice by adopting a process that we can only conclude is very likely not viable at all and where, even if it were to occur, many PCMs would never opt in, and, in many cases, would never have the opportunity to opt in.”

In the light of recent case law, in my view, the CAT erred in relation to both strength and practicability. The CAT should not have declined to certify the proceedings on the basis that they were proposed on an opt-out aggregate damages basis rather than on an opt-in basis.

Finally, I turn to the carriage issue. This concerns the choice of the CAT that Evans should have carriage of the claims of class members, as class representative, not O’Higgins. The CAT must apply a test of suitability. The discretion conferred is broad and multifaceted. The norm will be that the CAT will choose a single representative; it is unlikely to be sensible or feasible to appoint two representatives to represent the same class. But the CAT is not, in principle, precluded from choosing more than one representative, for example, if it became necessary in order to overcome an otherwise insoluble conflict between categories of class member. In the present case the CAT concluded that it should approve only one class representative and therefore, perforce, it had to make a choice.

The CAT discussed a wide range of factors in the body of the Judgment. In paragraph [389] it set out its overall conclusions:

“389.

We consider that if we were minded to certify on an opt-out basis, the carriage of the proceedings should be granted to the Evans PCR and not to the O’Higgins PCR. In other words, we would – on this basis – be minded to grant the Application of the Evans PCR and stay the Application of the O’Higgins PCR. We have reached this conclusion for the following reasons.

(1)

In many respects the Applications are (entirely unsurprisingly) very similar. In each case, we have the highest respect for each PCR and for the legal teams and experts they have instructed. Equally, the faults we have found (in particular an overspend on pre-certification costs and a shortfall in funding) apply similarly to both Applications. The question of carriage is a very marginal decision.

(2)

Although it is correct to say that the O’Higgins PCR was “first to file” in comparison with the Evans PCR, for the reasons we have given, we do not consider this to be a point in favour of the O’Higgins PCR.

(3)

The O’Higgins PCR undoubtedly has an advantage in terms of the extent of ATE insurance, which is a material point, but of limited weight given the costs that the Respondents are likely to incur. We consider that the ability of a successful defendant to recover taxed/assessed costs is important, but it is only one of many factors. Neither the O’Higgins PCR nor the Evans PCR was, in our judgment, providing security (in the form of ATE insurance) coming close to the taxed/assessed costs that the Respondents would be entitled to recover, assuming they were to succeed in their defence at trial. The £10 million-odd difference in ATE insurance cover between the rival PCRs is less than £2 million per Respondent, and does not amount to a particularly material difference.

(4)

We consider the claims of the Evans PCR to be better thought through. We stress, however, that we are drawing a distinction between two cases which have both only just survived strike-out on this occasion. In respect of each, we have very serious concerns about the manner in which the claims put forward have been articulated. With that very substantial proviso, we have concluded that, viewed side-by-side on a relative basis, the claims articulated by the Evans PCR have been better thought through and represent, to our mind, a marginally better attempt at capturing an elusive loss than that attempted by the O’Higgins PCR.

(5)

We stress that in reaching this conclusion, we are in no sense seeking to apply any kind of merits test. We are simply gauging the relative “strength” of the two claims in the sense described in paragraphs 98 to 118 above. To put the same point differently, we consider that the essential question to ask is which Applicant will better serve the interests of the victims that comprise the class(es) for whom the PCRs wish to act. Although we consider that the real answer to this question is “Neither”, if required to reach a conclusion, we conclude in favour of the Evans PCR.

390.

Accordingly, if we were required to do so, we would decide the Carriage Issue in favour of the Evans PCR, without taking into account the new material introduced by the Evans PCR after the oral hearings had concluded. It is to that new material that we now turn.”

The O’Higgins team challenge this decision. In written and oral submissions the following main points were made: (i) the CAT wrongly took into account the merits having stated, categorically, that they had not done so and indeed that to do so would be inappropriate; (ii) the CAT correctly identified as a differentiating factor that the Evans claim excluded certain types of transaction (see paragraphs [39] – [41] above) whereas O’Higgins had not and the CAT failed to infer from this that the O’Higgins claim better served the overriding interest of vindicating claims and serving consumers; (iii) the CAT erred in failing to give sufficient weight to the fact that the O’Higgins claim was better funded and financed and applied an unfair procedure in allowing the Evans team to supplement their financial offer after the hearing; (iv) the CAT erred into failing to take into account that the O’Higgins claim had been served first.

The first matter to consider is whether this proceeds by way of appeal or judicial review. Section 49(1A)(a) excludes from appeal points of law which relate to “…a decision on costs or expenses”.

The dispute is about the broad conclusion of the CAT about who can most efficiently conduct a damages claim. It is not a narrow dispute about, for instance, the exercise of the CAT’s discretion whether to order costs or about the quantum of costs. That sort of narrow dispute would appear to fall within the exception to the appellate jurisdiction. Does the fact that both Evans and O’Higgins seek to act as class representatives for commercial reasons and seek their return through the award of costs mean that this issue should likewise fall within the exception? If the test was simply, does the decision bring the claim to an end then the answer is “no”, since on the basis of the decision of the CAT the claim continues, carried by Evans. But for the reasons set out above, the end of the road test is not that which governed the analysis in Merricks (see paragraph [55] above) and in my view a broader optic is required through which to examine the issue.

Mr Jowell KC adopted this broader approach: the issue was as to “who” could most effectively carry the claim for damages. This was applying the language of the CA Act 1998 an issue which was “as to” damages. He further argued that because the O’Higgins claim was wider than the Evans claim it was “as to damages” in the sense that if it was not permitted to proceed that particular head of claim would be stymied, which was one test for a claim “as to damages”. Mr Jowell KC is correct in that if the CAT correctly chose Evans, a part of the overall claim pursued by O’Higgins will be precluded, not least because there is no argument that if the decision in favour of Evans is upheld anyone would then bring a claim (stand alone or opt-in) solely upon the basis of these excluded categories of transaction. As such actual victims would be denied a right to seek redress. Applying Mr Jowell’s logic, a point “as to damages” does arise. He is also correct that the answer to this question cannot ride upon whether the CAT and/or this Court thinks that Evans was correct to exclude these transactions from the claim; that would be to elide the merits with the broader but merits-neutral jurisdictional question of classification of whether the claim is “as to”damages.

I conclude that this part of the claim at least is “as to” damages. Adopting a purposive view of the appellate jurisdiction, I think that a choice as to who is best suited to advance a claim for damages is also “as to” damages. Judicial efficiency also dictates that the carriage dispute should not be split up. I would be very reluctant to arrive at a result whereby the issue of the scope of claim went to an appeal but the pure carriage issue went to judicial review. That would be the worst of all possible worlds. In my judgment this issue proceeds in its entirety by way of an appeal.

I turn to the merits. The choice made by the CAT majority was a quintessential multifactorial evaluation. The CAT considered in the round a variety of factors relevant to who could conduct the proceedings best. The challenge is as to the weight the CAT attached to the various considerations as to which the CAT, as the expert in how proceedings play out at the nuts and bolts level, is vastly better placed than the Court of Appeal to form a view. The threshold for persuading the Court of Appeal to reverse the CAT’s decision is commensurately high. I can see no basis upon which to interfere in the CAT’s decision. I address the four points raised by the O’Higgins team relatively briefly.

The CAT expressly stated that it had not taken merits into account. Mr Jowell KC argues that it can be seen from paragraph [389(5)] of the judgment (see paragraph [140] above) that the CAT did take account of the merits. I disagree. Fairly read all the CAT majority was saying was that its analysis proceeded from the starting point that the merits were the same as between the competing class representatives in that neither had advanced an arguable case. As such the merits were not a differentiating factor.

The CAT pointed out that the O’Higgins claim was more broadly framed. It did not however go on and analyse the relative pros and cons of the different claims. In an ideal world the CAT might have concisely set out what, if any, relevance it attributed to this fact. Nonetheless, it seems to me that since it does not feature in the summary section in paragraph [389] the most appropriate inference to draw is that it was not considered to be relevant and was attributed no weight. I have no difficulty with this conclusion. I have set out the relevant facts at paragraph [39] – [41] above. The mere fact that one putative class representative crafts a broader claim is not an indication that the claim is preferable. Were it otherwise all class representatives would be falsely incentivised to draft claims as widely as possible to obviate the risk that in a carriage competition having a narrower claim might tell against them. There may be many good reasons why a better articulated and thought-through claim will be narrower and not wider. There might be sensible trade-offs to be made between pursuing the more questionable outer limits of a claim (which might significantly add to costs) and focusing upon a narrower and stronger core claim (which might be more efficient to litigate). If, as seems to be the case, the CAT decided that this was not a factor of relevance then it was entitled so to do. If, equally, it considered that on the basis of the explanations given before it the exclusion of the additional transaction reflected a better thought through claim then again that was a legitimate conclusion for the CAT to form.

The CAT did not appear to view funding as a factor which differentiated as between O’Higgins and Evans: see Judgment paragraph [389(3)]. Mr Jowell KC argued that in this the CAT erred because, objectively speaking, the O’Higgins offer was superior to that of Evans. As at the end of the hearing below the O’Higgins team had put forward a more extensive funding package. To counter perceived deficits, but after the hearing, the Evans team submitted new evidence to the effect that (i) Evans had incepted an additional layer of ATE insurance providing additional cover in the amount of £10,500,000, thus bringing the total cover to £33,500,000; and (ii) the funder had agreed to increase the funding commitment by £12 million. The effect, if this was taken into account by the CAT, was to overtop the financial package hitherto offered by O’Higgins. There was an objection to the admissibility of this evidence from the O’Higgins team which the CAT described as entirely unsurprising (Judgment paragraph [394]). Having considered the objections, the CAT decided however to admit the evidence. They did so with some misgivings because of the risk of system “gaming”:

“406.

It follows that the objection on admissibility must be rejected, and that we are obliged not to disregard this material, but instead to take it into account. However, the extent to which we take it into account must be constrained so as to avoid a “gaming” of this jurisdiction. We do not want to encourage late changes to the basis upon which a PCR proposes to represent a class. Accordingly, it seems to us that where an applicant, as here, makes a late improved offering, which has the effect of improving that applicant’s offer as against that of a rival applicant, such that the outcome of the carriage dispute changes in the former’s favour, then the rival applicant who would otherwise have been awarded carriage of the litigation should be given the opportunity of matching or beating the improved offer. If that occurs, it is the rival who is awarded carriage of the litigation.”

The CAT then held that even without the new material the decision would still have gone in favour of Evans (Judgment paragraphs [407] – [409]). The CAT observed (Judgment paragraph [405]), to my mind significantly: “The fact is that both PCRs consider that they are best suited to represent the classes they wish to represent, and a responsible PCR will – we have no doubt – keep the quality of its offering under constant review.”

I agree with the CAT that in many cases there might be nothing to choose between funding packages even if one is, superficially, superior to another as at the date of the hearing. As the CAT was alluding to, there might be artificiality in making a snapshot evaluation at the certification stage when the true position is that the facility will be subject to change and adaptation. It might be better for the CAT to consider not just the snapshot of what was on offer, but in particular the ability and preparedness of each competing class representative and its funders to increase its facility over time; and the resilience of that facility, e.g. whether it is guaranteed and ringfenced. That might be a better indication of which carrier is best suited in the context of the efficient running of the ligation as a whole, and the important need to protect a successful defendant against costs.

A submission was made to the Court that the cost of finance for a prospective class representative was generally higher pre-certification, when the risks were greater than post-certification. If this be so then, if the CAT demanded to see financing in place before selection, this might increase the costs relative to an obligation to secure actual financing post-selection. We did not, though, see any evidence indicating quite how much the difference was in real terms and I would not express any view on the importance of such a point on the evidence before us.

Finally, Mr Jowell KC argued that the CAT was wrong (Judgment paragraph [389(2)]) to treat the fact that O’Higgins was the “first to file” as an essentially irrelevant consideration. We were taken to the position in various US and Canadian jurisdictions where the law and courts had addressed the relevance of a putative class representative being the first to file. Unlike in some jurisdictions there is no statutory embodiment of such a principle in the UK. I agree with the CAT that it is largely an irrelevant factor. If it were systemically accorded weight, it would risk encouraging premature and ill-thought out claims simply upon the basis that being first in time conferred a forensic advantage. It could penalise the more measured class representative that wished to road test a claim thoroughly before lodging or (as we were told is the case here) await publication or availability of a Commission decision. It is hard to see what, in policy terms, is to be gained from encouraging a race to file.

The final procedural point concerned the decision by the CAT to determine carriage at the same time as certification. It was said that the first to file argument would disappear as a relevant consideration if the CAT set up an early, even pre-certification, decision making process on carriage. It could, once an application has been made, set a deadline for any further prospective representative to apply with a guillotine coming down upon expiry of the deadline. It could then, applying a robust approach, make its decision. The timing of a carriage decision is for the CAT. I can see however that in some cases this might be suitable to avoid what risks becoming costly satellite litigation. There might be a degree of rough and readiness about the exercise but, equally, the CAT, armed with its rapidly growing expertise in the area will know what sorts of facts and matters are relevant. From the perspective of this Court if the CAT adopted such a procedure and gave concise reasons for its choice it would be most unlikely that any appeal would succeed or even that permission to appeal would be granted.

In conclusion the CAT did not err in selecting Evans to be the class representative.

There is one final matter to address. Mr Jowell KC for O’Higgins submitted that were the case to be remitted it should be to a differently constituted CAT given the strong and adverse view formed by the majority on the merits. It was suggested that remitting to the same panel unavoidably meant that there would be unconscious bias against any claim that proceeded. This submission is not pursued by the Evans team. I do not accept the point. This was, on any view, a complex and difficult case. as demonstrated by the divisions between the judges on key issues. The CAT itself granted permission to appeal upon the basis that there was a real prospect of success, demonstrating its impartiality and open minded approach. The composition of the tribunal going forward is a matter for the President but there is nothing to suggest that the same constitution would not act with complete objectivity upon a remittal.

As set out in footnote 1 above, following the hand down of the judgment on 25^th July 2023 two events occurred. First, the Supreme Court handed down its judgment in PACCAR. Secondly, the President of the CAT questioned the jurisdiction of the Court to amend a CPO as the Court had indicated it intended to do. For the reasons set out below, I have concluded that this Court does have jurisdiction to make or amend a CPO, that this point does not in fact arise on the facts of this case, and that the Court should make an order that provides for a wider remittal to the CAT to take into account (if appropriate) the implications of the decision in PACCAR.

I start by considering the issue raised by the President.

In paragraph [138] of my judgment delivered on 25 July 2023, I indicated that the CPO should be set aside to the extent that it made an order for opt-in proceedings and it should be amended so that the proceedings were made opt-out upon an aggregate damages basis. The President questioned whether it was open to this Court to make such an order. He identified the following points: First, the CAT was a statutory tribunal sitting outside of HMCTS and the Senior Courts Act 1981 with a UK-wide jurisdiction. Secondly, by primary legislation under section 47B(4) CA 1998 “Collective proceedings may be continued only if the [CAT] makes a collective proceedings order”, and any such order is required to specify whether the proceedings were opt-out or opt-in. Thirdly, in the present case the CAT certified the proceedings upon an opt-in basis, which decision was appealed. Fourthly, this Court decided that the CAT erred and that the order should be set aside to the extent that it made an order for opt-in proceedings and it should be amended so that the proceedings were made opt-out upon an aggregate damages basis. Fifthly, the basis for this jurisdiction was unexplained and it was not apparent that the Court had the power to amend the CPO. Had this been an appeal from the High Court the jurisdiction would exist under CPR part 52.20(1), which provides that the appeal court had all the powers of the “lower court”. However, this was secondary legislation which was irrelevant to statutory appeals from the CAT under the CA 1998. Finally, matters were exacerbated by the fact that although the CAT certified the jurisdiction as England and Wales, the class certified would likely include persons in Scotland and Northern Ireland.

In response the parties submitted that the power of the Court of Appeal in relation to an appeal from the CAT was broad and contained in primary legislation, namely section 15(3) Senior Courts Act 1981. That section makes clear that for all purposes of or incidental to an appeal the Court has all the authority and jurisdiction of the tribunal from which the appeal was brought:

“15 General jurisdiction of Court of Appeal.

(1)

The Court of Appeal shall be a superior court of record.

(2)

Subject to the provisions of this Act, there shall be exercisable by the Court of Appeal—

(a)

all such jurisdiction (whether civil or criminal) as is conferred on it by this or any other Act; and

(b)

all such other jurisdiction (whether civil or criminal) as was exercisable by it immediately before the commencement of this Act.

(3)

For all purposes of or incidental to —

(a)

the hearing and determination of any appeal to the civil division of the Court of Appeal; and

(b)

the amendment, execution and enforcement of any judgment or order made on such an appeal,

the Court of Appeal shall have all the authority and jurisdiction of the court or tribunal from which the appeal was brought.

(4)

It is hereby declared that any provision in this or any other Act which authorises or requires the taking of any steps for the execution or enforcement of a judgment or order of the High Court applies in relation to a judgment or order of the civil division of the Court of Appeal as it applies in relation to a judgment or order of the High Court.”

There is no indication in the 1981 Act that the word “tribunal” excludes a statutory tribunal such as the CAT.

CPR 52.20(1) is to similar effect. It provides that in relation to an appeal this Court has all the powers of “the lower court”, and CPR 52.1(3)(c) defines “lower court” to include a tribunal, which again I consider must include a statutory tribunal.

I therefore agree with the analysis of the parties. It is not correct to regard the CAT as operating wholly independent of any of the provisions of the Senior Courts Act 1981, nor to regard section 47B(4) CA 1998 as preventing the Court of Appeal from making a CPO in an appropriate case. Section 49 CA 1998 expressly provides for an appeal to this Court from the CAT. It does so on certain specified and limited grounds. Section 49 CA 1998 does not, however, purport to exclude or modify any of the powers conferred upon this Court by the Senior Courts Act 1981 when dealing with any such appeal. Parliament intended that this Court should have all the powers normally available to it when hearing appeals.

That being so, depending on the circumstances, it could be appropriate for the purposes of, or incidental to the determination of the appeal, for this Court either to exercise the authority and jurisdiction of the CAT to make a CPO if one had been refused; or, if a CPO had been made on an erroneous basis, to vary or amend any such order to give effect to the judgment of the Court. The CAT obviously has the power to make a CPO, and it also has the power to vary or amend a CPO under rule 85 of the CAT Rules. Although section 47B(4) CA 1998 envisages that collective proceedings may be continued only if the CAT makes a CPO, I see no reason why Parliament should have intended that this requirement should not be satisfied if such an order was made by the Court of Appeal exercising the “authority and jurisdiction” of the CAT.

In the instant case, the parties point out that no CPO of any description was actually made by the CAT. The CAT declined to make an order granting either of the applications for certification on an opt-out basis, and it stayed both applications in order to give the PCRs an opportunity to amend their pleaded cases but upon the conditional basis that if the cases went ahead they would proceed only upon an opt-in basis. Thereafter, the parties did not agree the terms of any such order, and the CAT did not make a CPO on an opt-in basis (or indeed any sealed order that we have seen). Instead, the parties pursued an appeal against the CAT’s refusal of certification on an opt-out basis.

The upshot is that this Court would have had the jurisdiction either to amend any CPO that had been made by the CAT, or simply to make such an order itself on the correct basis had it thought it appropriate to do so. It also remains open to the CAT on remittal to make a CPO on an opt-out basis if it considers it appropriate to do so in light of this judgment and the implications of the decision in PACCAR, to which I now turn.

The next question concerns the implications of PACCAR which concerned the enforceability of certain litigation funding agreements pursuant to which the funder was entitled to recover a percentage of any damages recovered. To be enforceable such agreements had to satisfy certain conditions. The agreements in question did not satisfy those conditions. The issue then was whether they fell within the category of agreements required to comply and if so as to their enforceability. The Supreme Court held, by a majority, that the agreements were subject to the relevant requirements and were unenforceable.

The Court of Appeal has not heard argument as to the relevance of this judgment to the viability of the present claims going forward. Unsurprisingly, the parties adopt different stances. The Evans team contemplate that there might have to be some modification of their funding arrangements but do not otherwise seem to contemplate any undue difficulties. The O’Higgins team, more optimistically, suggest that there is a “theoretical” possibility that the Evans PCR might not, in the light of this judgment, be able to proceed and that the O’Higgins team might be substituted. The respondent banks suggest that the funding arrangements of both Evans and O’Higgins appear to fall foul of the majority judgment in PACCAR and accordingly the CAT might have to revisit certification. In the judgment of this Court in Trucks (ibid) the Court contemplated the possibility that an unsuccessful applicant for authorisation as a PCR should not necessarily have its claim struck out if there was the possibility, in the future, that the CAT might direct either a substitution of the PCR or the ordering of an additional PCR to represent a subclass of the relevant consumers, for instance to take account of a conflict of interest at the level of the PCR.

Having reflected, I see the force of the suggestion that the CAT is better placed than we are to consider the implications, if any, of the decision in PACCAR. Accordingly,in addition to remitting this matter for consideration of the specific issues that were canvassed on the appeal, we also indicate that the CAT should consider what, if anything, follows for this litigation from that decision.

The Court, as part of the submissions following the hand down of the first judgment, received observations on the principles to be applied to costs. Given the novelty of many of the issues arising it is helpful to summarise, briefly, the principal considerations the Court has taken into account in determining costs and which will be reflected in the final Order of the Court. These considerations arise on the facts of this case. They do not reflect principles which would necessarily apply in other cases:

In all these circumstances the Court allows the appeal in relation to those areas where we differ from the CAT i.e., with respect to the opt-in v opt-out issue. Otherwise, we dismiss all of the appeals. We refuse permission to apply for judicial review to all of the applicants. We remit the matter to the CAT for further decision and case management in accordance with this judgment.

I agree.

I also agree.