Group Litigation Orders: Supreme Court Clarifies When Courts May Order That A Judgment On A GLO Issue Is Not Binding On Other Parties Due To A Subsequent Change In The Law

The Supreme Court has dismissed an appeal against orders that decisions in test cases in the long-running CFC and Dividend GLO were not, or should not be, binding on other cases on the group register where there had been dramatic changes in the law since the test cases were decided: Axa Insurance UK Plc v Commissioners of Inland Revenue [2026] UKSC 24.

As the decision emphasises, the purpose of a GLO is to enable multi-party actions to be determined efficiently on a collective basis. To that end, under CPR 19.23(1)(a), a judgment or order in relation to a GLO issue automatically binds all other claims on the group register at the time, unless the court orders otherwise. As the court notes (citing the second edition of Class Actions in England and Wales, written by lawyers from HSF Kramer), this is seen as one of the key advantages of using a GLO to resolve multiple claims.

The Supreme Court identified two conflicting key considerations which it said are of great importance: the need to maintain the integrity of the GLO regime by treating a decision in a test case as binding; and a party's right to have their dispute decided according to a correct understanding of the law. 

Balancing those considerations means that the court should order that a test case decision will not be binding, based on a subsequent change in the law, only where it has been clearly established that the legal basis for the test case decision was wrong. It is not enough that there is some doubt about its correctness or that the parties to the follower cases wish to raise different, and potentially better, arguments than were raised in the test case. 

Even where the legal basis for the test case decision has been found to be wrong, the court will still need to consider whether, in the particular circumstances, it would be unjust to exercise the discretion to order that the decision will not be binding – for example because the result of doing so would be to impose on the parties further lengthy and costly litigation. All of this means that the discretion is likely to be exercised only in a rare case.

The decision also shows that, even where the court has made a decision on a GLO issue, the decision may not be binding on the follower claims because of the fact-specific nature of what was decided. 

Background

As set out in our blog post on the Court of Appeal's decision, here, this was an appeal in the context of the CFC and Dividend GLO, under which taxpayers brought restitutionary claims for the repayment of UK taxes levied in contravention of EU law, which were therefore paid under a mistake of law.

Claims by the Prudential group were tried as test cases and were determined by Henderson J in 2013, and ultimately by the Supreme Court in 2018. Preliminary issues were later tried as to the effect of those decisions on other claims on the group register, brought by the "AXA claimants", particularly in relation to limitation and set-off.

Set-off issue

The set-off issue concerned claims for the time value of unlawfully levied tax that had later been set off against lawfully levied tax. The GLO issues included how any compensation or relief should be assessed, and on what basis interest was payable. 

Before the trial of the Prudential test cases, the defendant conceded that compound interest was payable for part of the claim (category (a)), based on the House of Lords' decision in Sempra Metals Ltd v IRC [2007] UKHL 34, which had held that compound interest could be awarded in respect of money paid under a mistake. Henderson J noted the concession and held that compound interest was payable both for the category (a) claims and for other parts of the claim (categories (b) and (c)), which he saw "no rational basis" to distinguish. 

Henderson J's decision on categories (b) and (c) was, however, overturned by the Supreme Court (Prudential Assurance Co Ltd v HMRC [2018] UKSC 39), which departed from the reasoning in Sempra Metals. The Supreme Court noted that the category (a) claim would also have been rejected if not for the defendant's concession.

The Court of Appeal held (overturning the High Court's decision) that the Axa claimants were bound by Henderson J's decision on the category (a) claims – regardless of the fact that the Supreme Court would have reversed the decision if not for the concession – unless the court ordered otherwise. However, it considered it appropriate to order otherwise because there were special circumstances that would otherwise create injustice, including because when the defendant made its concession it could not have anticipated that Sempra Metals might be overruled. The Axa claimants appealed to the Supreme Court.

Limitation issue

The limitation issue related to when the claimants discovered the mistake of law or could with reasonable diligence have done so (the "discovery date"), so as to start the limitation period under s.32(1)(c) of the Limitation Act 1980. The GLO issue as originally defined asked from what date limitation started to run, but the list of GLO issues agreed in the run up to trial asked to what extent the claim was statute barred.

Henderson J held that the claims were in time. However, the decision underpinning that conclusion (Deutsche Morgan Grenfell Group plc v IRC [2006] UKHL 49, in which the House of Lords held that the discovery date was the date of the judicial ruling that established the unlawfulness of the relevant UK tax provisions) was later overruled by the Supreme Court.

The Court of Appeal held (in agreement with the High Court) that, although the question of whether the claims were statute barred was a GLO issue, Henderson J's findings on that issue were fact specific to the test claims and could not properly be interpreted as applying to all claimants in the GLO. The Axa claimants appealed to the Supreme Court.

Decision

The Supreme Court unanimously dismissed the appeal. Lady Rose and Lord Richards gave the leading judgment, with which Lord Hodge and Lord Hamblen agreed. Lord Leggatt gave a concurring judgment.

The court noted that, if the Axa claimants' claims had come before the court outside the context of the CFC and Dividend GLO, their submissions on the set-off and limitation issues would certainly be rejected, as they were contrary to Supreme Court authorities handed down since the Prudential test cases had been determined. The question was therefore whether the Axa claimants were entitled to orders based not on the current state of the law but on what the law was thought to be when the test cases were determined, since their claims were on the GLO register at that time.

Set-off issue

The Supreme Court noted that the power under CPR 19.23(1)(a) to order that the decision in a test case is not binding on follower claims is expressed in general terms. It rejected the Axa claimants' argument that it should be restricted to where the judgment or order in the test case itself provided that it was not to bind some or all of the follower claims. If that had been the drafter's intention, it would have been made clear by different wording.

Noting that the kinds of dispute in which the GLO procedure is used and the nature of the GLO issues identified are highly varied, the court said it would be "rash" to try to provide guidance for every case. It agreed with the Court of Appeal that the desire to avoid injustice should guide the court in exercising its power to order otherwise, as it guides the exercise of any discretion, but that did not mean that the court had a broad discretion. Having regard to the importance of not undermining the efficacy of the GLO regime, there would need to be some exceptional quality about the circumstances before it would be appropriate to consider whether to exercise the power. 

The Supreme Court said its focus was a case, such as the present, where it was argued that the court should order otherwise because the legal basis of the decision in the test case had been shown to be wrong. The court drew an analogy with issue estoppel, which may also preclude a party from putting forward strong arguments in support of its case because of the binding nature of an earlier judgment. It noted that issue estoppel presents a bar to relitigating issues except in special circumstances where this would cause injustice. 

The court considered that the cases on issue estoppel were helpful in indicating what kinds of exceptional circumstance should justify a departure from the binding force of a test case order. The leading case on when exceptional circumstances justify lifting an issue estoppel is Arnold v National Westminster Bank plc [1991] 2 AC 93. That case established that the relevant circumstances may include a subsequent change in the law as well as the discovery of new facts.

The Supreme Court pointed out, however, that issue estoppel does not generally raise the "additional weighty factor" that making an exception might have "wider and potentially undesirable repercussions for the operation of the overall GLO regime". In the GLO context, there were two conflicting key considerations: the need to maintain the integrity of the GLO by treating the test case as binding on follower claims; and the party's right to have their dispute decided by applying a correct understanding of the law at the time the order is made. The first key consideration meant the court should approach the exercise of its discretion to order otherwise with great caution, so as not to jeopardise a central purpose of the GLO regime – ie to enable multi-party actions to be determined efficiently on a collective basis. 

Balancing these two key considerations, the Supreme Court said it would be appropriate to order that a test case judgment is not binding only where it had been clearly established that the legal basis for that judgment was wrong – not merely where a party to the follower case wishes to raise different arguments or could point to dicta in subsequent cases that cast doubt on the correctness of the judgment. In that way, the exceptionality required was stronger than in cases of issue estoppel. 

Once it was established that the legal basis for the test case judgment was wrong, the court would still need to consider whether, in the particular circumstances, it would be unjust to exercise the discretion to order otherwise. Relevant factors might include:

• whether the result of agreeing to disapply the test case judgment would be to impose on the opposing party further lengthy and costly proceedings, in effect relitigating the issue that had been decided in their favour by the test case; 
• whether there was any particular unfairness to follower claimants arising from the fact that they were not chosen to be the test case in the GLO, bearing in mind that there may also be advantages to being a follower claimant (in particular in being spare significant investment of management time and resource); and
• whether a party had attempted to manipulate proceedings by making tactical decisions in the course of the test case with a view to later inviting the court not to treat the test case judgment as binding.

Applying these principles, the Supreme Court held that to enter judgment in a follower case on the basis of the Prudential judgment, in the very unusual circumstances that had arisen, would be unjust to such an extent that it was appropriate to exercise the discretion to order otherwise. This was because of the extraordinary nature of the developments in the law since the proceedings began, and in particular that it had been definitively established by Supreme Court decisions that the planks on which the Prudential judgment was based were wrong. 

The next question was whether there were factors which outweighed the injustice that would be caused to the defendant if the follower claims were decided on an incorrect legal basis. The Supreme Court held that there were no such factors. The court noted that the exercise of the power to order otherwise would not condemn the parties to yet more years of litigation to resolve factual or legal points, and there was no question of allowing points of law to be re-argued.

Limitation issue

The Supreme Court said that, in interpreting Henderson J's judgment and order, it was important to avoid hindsight and to recognise that, under Deutsche Morgan Grenfell, the discovery date was not a fact specific issue but the same date for everyone. 

The bald statement in Henderson J's order that the claims were "in time" mentioned no date and had to be interpreted by reference to what he said in his judgment, which appeared to be based on an underlying assumption that the claimants could not have been aware of their mistake before the relevant court decision. But that was not an issue that Henderson J decided. The Supreme Court agreed with the Court of Appeal that the discovery date for the purposes of s.32(1)(c) was simply not an issue.

The court emphasised that a decision on a GLO issue could be binding even if the answer to that issue was agreed or admitted, rather than determined after adversarial argument. It commented, "A party cannot attempt to keep their powder dry by conceding something which is then decided against them in the test case in the hope of arguing it out in full in a different case on the register." That was not, however, what happened in this case. 

Lord Leggatt's concurring judgment

Lord Leggatt delivered a concurring judgment in which he agreed with the court's overall conclusion that it should "order otherwise" under CPR 19.23(1)(a) in respect of the set-off issue, but rejected the analogy with cases of issue estoppel that had been endorsed by the Court of Appeal and in the judgment of Lady Rose and Lord Richards. In particular, Lord Leggatt did not accept that a change in the law since the decision in a test case may amount to "special circumstances" that would justify making an order under that rule.

Lord Leggatt noted that in a case of issue estoppel the claimant is bringing a new claim that has arisen since the previous judgment. Where in the meantime there has been a change in the law there are competing considerations of justice and the balance may sometimes favour allowing the issue to be reopened. In the GLO context, in contrast, the causes of action have already arisen and the claims are proceeding in parallel, with one or more test claims selected as a vehicle for deciding a common issue. In Lord Leggatt's view, if the law changes after a test claim has been decided, it would be unjust in principle to permit some but not all parties to reopen the common issue – particularly as it may be a matter of happenstance which claim is used as a test claim and how long the litigation continues after judgment in that claim is given. 

Accordingly, Lord Leggatt did not consider that the fact there had been a material change in the law was sufficient to justify making an order that the judgment in the Prudential test case was not binding in the Axa claimants' claims. However, he agreed it was appropriate to make such an order in this case because, on its unusual facts, doing so would not undermine the key purposes of the rule that decisions on GLO issues are binding on other claims on the register, namely avoiding inefficiency and the potential for inconsistent outcomes. It was significant, in particular, that an order otherwise would not result in a GLO issue decided in the test case being reargued, since the issue had been conclusively decided by subsequent Supreme Court decisions that the test case judgment was wrong in law. Further, the injustice of reaching inconsistent results in different claims under the GLO was, in Lord Leggatt's view, outweighed by the incoherence that would result from treating the test claim judgment as binding on category (a) claims when there was no rational basis for reaching different outcomes in category (a) claims as compared to category (b) and (c) claims.