James v HSBC: When “Enough to Plead Fraud” Starts the s.32 Clock, and When the AIC/Barrell Jurisdiction Will Not Re-open a Strike-Out

1. Introduction

James & Anor v HSBC UK Bank PLC (Rev1) [2026] EWCA Civ 88 is a Court of Appeal decision affirming a High Court strike-out of claims brought by former customers, Mr and Mrs James, against HSBC UK Bank plc. The claim alleged that HSBC (through unidentified employee(s)) fraudulently opened a loan account in Mrs James’s name in 2004, moved money from the claimants’ genuine joint accounts to service that loan, and later pursued recovery and adverse credit reporting.

The litigation ultimately turned not on whether the fraud occurred (no trial took place), but on whether any viable causes of action were pleaded and—critically—whether any arguably viable proprietary/constructive trust or restitutionary claims were time-barred, notwithstanding section 32 of the Limitation Act 1980 (“LA 1980”) (postponement for fraud/concealment/mistake). A further procedural focus was whether the first-instance judge should have re-opened his earlier limitation conclusions under the re Barrell Enterprises Ltd line of authority, now commonly called the AIC jurisdiction after AIC Ltd v Federal Airports Authority of Nigeria.

Key issues

  • Viability of pleaded fraud/deceit: whether the pleaded case in “fraud/deceit” disclosed a reasonable cause of action.
  • Limitation and s.32 LA 1980: when the claimants discovered (or could with reasonable diligence have discovered) enough to plead dishonesty/fraud for limitation purposes.
  • Re-opening an unperfected judgment: whether an alleged mistake about the meaning of a 31 July 2013 letter justified revisiting the limitation analysis under the AIC/Barrell jurisdiction.
  • Proprietary framing against banks: the Court expressed “real doubts” about fitting constructive trust/restitutionary proprietary analysis onto intra-bank transfers given the debtor–creditor nature of bank accounts, but proceeded on arguability due to concessions.

Parties

  • Appellants/Claimants: Mr and Mrs James
  • Respondent/Defendant: HSBC UK Bank plc

Condensed factual timeline (alleged)

  • Sept 2004: Loan account opened in Mrs James’s name for £49,000 (allegedly without authority).
  • Mar 2005: £25,000 transferred from a joint account to the loan account; monthly interest-like transfers 2005–2008.
  • May 2010: demand for ~£28,306; adverse credit reporting alleged.
  • June–July 2013: formal complaint; HSBC sends 31 July 2013 letter.
  • Sept 2013: first SAR response (limited).
  • Oct 2016: second SAR response includes “CARM report”.
  • Apr 2018: expert report concludes likely rogue employee fraud.
  • Oct 2022: claim issued.
  • Oct 2023 & July 2024: strike-out decisions at first instance.
  • Feb 2026: Court of Appeal dismisses appeal.

2. Summary of the Judgment

The Court of Appeal (Newey LJ giving the lead judgment; Singh LJ and Newey LJ agreeing) dismissed the appeal. The Court held:

  1. There was no viable appeal against the strike-out of the original pleaded “fraud/deceit” claims; the judge was “plainly right” that fraud is not itself a cause of action, and that deceit requires reliance, which the claimants did not plead (indeed could not, because they denied the loan once they learned of it).
  2. The only live appellate issue was whether the judge should have re-opened his earlier limitation conclusion under the AIC jurisdiction, based on an alleged misinterpretation of the 31 July 2013 letter.
  3. Even assuming a possible misreading, it would not have mattered: on the evidence the judge was entitled to conclude that by 2013 the claimants knew enough facts to plead dishonesty/fraud for s.32 purposes, so the limitation clock began at the latest in 2013.
  4. Consequently, the amended proprietary/constructive trust/restitutionary claims (treated as subject to a six-year period) were time-barred, and the strike-out stood.

3. Analysis

3.1 Precedents Cited

The decision is precedent-rich in two areas: (i) pleading and limitation where fraud is alleged, and (ii) the procedural power to re-open an unsealed judgment.

(A) Re-opening judgments: “Barrell” and the modern AIC formulation

  • re Barrell Enterprises Ltd [1973] 1 WLR 19
    Cited as the historical root of the jurisdiction allowing a judge to reconsider a judgment before it is perfected by a sealed order. In this case, the claimants invoked that jurisdiction to invite the first-instance judge to revisit his 2013 “knowledge” finding.
  • AIC Ltd v Federal Airports Authority of Nigeria [2022] UKSC 16, [2022] 1 WLR 3223
    Treated as the leading modern authority; the Court of Appeal accepted counsel’s shorthand “the AIC jurisdiction”. The first-instance judge refused to re-open; the Court of Appeal held that refusal was justified because (i) he did not accept he had made a mistake, and (ii) even if he reconsidered, he would reach the same conclusion—so no material injustice arose requiring re-opening. The case thus exemplifies that AIC is not a merits “second bite”, but a controlled power aimed at preventing material error/injustice before perfection of the order.

(B) s.32 LA 1980: what counts as an “action based upon fraud” and “facts relevant”

  • Beaman v ARTS [1949] 1 KB 550
    Used for the orthodox interpretation that an “action based upon the fraud of the defendant” is not any action where fraud is alleged, but one where fraud is a necessary ingredient of the cause of action. This provided the lens through which the Court evaluated which of the proposed amendments could even potentially invoke s.32(1)(a).
  • Seedo v El Gamal [2023] EWCA Civ 330, [2023] Ch 473
    Cited twice: first to confirm the Beaman approach to s.32(1)(a); second (at [77] in Seedo) to confirm that a “fact relevant” to the right of action (s.32(1)(b)) means a fact without which the cause of action is incomplete. This significantly narrowed the utility of “concealment” arguments in relation to contract claims that were complete without proving fraud.
  • Johnson v Chief Constable of Surrey (CA, 19 October 1992)
    Cited for the “cause of action incomplete without the fact” understanding of “fact relevant” (s.32(1)(b)).
  • AIC Ltd v ITS Testing Services (UK) Ltd, The Kriti Palm [2006] EWCA Civ 1601, [2007] 1 All ER (Comm) 667
    Cited (Rix LJ, Nourse LJ, Buxton LJ) as authority on s.32(1)(b)’s meaning of “fact relevant”. The Court used it to reinforce that concealed material which merely strengthens a case is not necessarily “relevant” in the s.32(1)(b) sense if the cause of action is already complete.

(C) Constructive trust limitation

  • Williams v Central Bank of Nigeria [2014] UKSC 10, [2014] AC 1189
    Deployed to explain that a six-year limitation period applies to remedial constructive trust claims. This was decisive in framing the amended constructive trust route as time-barred unless s.32 postponed time. The case also corrected the first-instance judge’s initial instinct that “constructive trust” might be “not subject to any limitation”.

(D) Pleading dishonesty: when facts permit an inference

  • Sofer v Swissindependent Trustees SA [2020] EWCA Civ 699
    Central to the judge’s and the Court of Appeal’s approach to whether, by 2013, the claimants could (with reasonable diligence) have pleaded dishonesty. Sofer’s summary of principles was treated as the relevant framework.
  • Three Rivers District Council v Governor and Company of the Bank of England (No.3) [2003] 2 AC 1
    Cited within the Sofer principles: dishonesty must be specifically pleaded and sufficiently particularised; it is not enough if facts are equally consistent with innocence.
  • JSC Bank of Moscow v Kekhman [2015] EWHC 3073 (Comm)
    Cited for the proposition that claimants do not need facts only consistent with dishonesty; the test is whether dishonesty is more likely than innocence/negligence on the pleaded primary facts.
  • Walker v Stones [2001] QB 902
    Cited for reading particulars of dishonesty as a whole and in context.
  • Persons Identified in Schedule 1 to the Re-Amended Particulars of Claim v Standard Chartered Bank plc [2024] EWCA Civ 674, [2024] 1 WLR 4589
    Noted as a later authority (post-dating the first judgment) that, if anything, slightly relaxes pleading requirements. The Court treated it as confirming that the Sofer approach remained usable and that the judge’s conclusion was not undermined.

(E) Banking law baseline (debtor–creditor)

While not a “precedent” case citation, the Court anchored its doubts about proprietary remedies in orthodox banking principle (citing Chitty on Contracts (36th edn, 2026), Vol II, §37-342): absent special features, bank/customer is debtor–creditor, and a credit balance is a chose in action, not a proprietary pot of money. That analysis underlies the Court’s expressed doubts about constructive trust/proprietary restitution attaching to an internal book transfer.

3.2 Legal Reasoning

(A) Why “fraud/deceit” was struck out and not appealed

The Court confirmed, as a matter of basic private law taxonomy, that “fraud” is not a standalone cause of action: it is a descriptor that may be an element of recognised causes of action. The pleaded route was “deceit”, but deceit requires that the claimant was deceived—typically that the claimant relied on a false representation to their detriment.

The pleaded narrative undermined reliance: the claimants’ case was that once they learned of the loan, they challenged it. Thus, while the pleadings alleged false representations (to the claimants, credit agencies, and FOS), the essential reliance element could not be made out on their own pleaded facts. This explains why the “fraud/deceit” framing could not survive strike-out.

(B) The attempted pivot: contract, constructive trust, and proprietary restitution

After the first hearing, the claimants sought to amend to plead: (i) breach of contract (acting without instructions; lack of reasonable skill and care; mishandling SAR response), (ii) constructive trust over the sums transferred, and (iii) an alternative restitutionary proprietary claim based on continued ownership of stolen money.

The Court of Appeal drew an important distinction for s.32:

  • Contract claims: generally not “based upon fraud” because fraud is not a necessary ingredient; they can be complete without proving dishonesty. That makes reliance on s.32(1)(a) problematic.
  • Concealment (s.32(1)(b)): limited to facts without which the cause of action is incomplete; facts that merely reveal fraud or make it easier to prove do not necessarily postpone limitation for contract causes of action.
  • Constructive trust/proprietary restitution: at least “arguably” actions “based upon” fraud, potentially engaging s.32(1)(a), but still subject to a six-year period (per Williams v Central Bank of Nigeria) unless postponed.

(C) The heart of the appeal: did a “mistake” about the 31 July 2013 letter matter?

The claimants argued that the judge’s 2013 knowledge finding rested on a mistaken view that the 31 July 2013 letter recorded a repayment plan concerning the loan account, when in fact it concerned a different account (the joint Premier account). The alleged mistake was linked to the letter pages being out of order in the bundle.

The Court of Appeal’s reasoning proceeded in layers:

  1. Uncertainty of mistake: it was not clear the judge did misread the letter; even in the wrong order, the letter indicated the relevant paragraph was about the joint account.
  2. No materiality: even if there was a misreading, the judge stated he would have reached the same conclusion on reconsideration. The Court held he did not need to provide extensive further reasons: the logic was already apparent.
  3. Reasonable diligence threshold met by 2013: by 2013, Mrs James knew: (a) a loan existed in her name that she did not authorise/benefit from, (b) substantial sums (£25k+) had been moved from their joint accounts to that loan account, and (c) there was bank-recorded “repayment arrangement” information attributed to her circumstances that she said was fictitious. Whether (c) related to the loan account or the joint account did not change its evidential character: it was still an apparently invented record of her agreeing and renegotiating repayment terms.
  4. Inference of dishonesty: applying the Sofer/Three Rivers framework, these were not readily explicable by mistake or innocence. The judge was therefore entitled to conclude that dishonesty could be pleaded by 2013 and that time began to run “at the latest” then.

(D) The Court’s “real doubts” about proprietary remedies against a bank (but no decision)

A notable feature is the Court’s candid expression of “real doubts” that either constructive trust or proprietary restitution neatly fits a case where: the claimant’s “money in the bank” is legally a debt owed by the bank, and a transfer between accounts is an internal accounting operation that changes the bank’s indebtedness rather than transferring a specific, identifiable asset to the bank.

The Court emphasised the orthodox position: a credit balance is a chose in action. When HSBC “transferred” £25,000 from one account to another, it did not “acquire” the claimant’s asset in the proprietary sense; it adjusted ledgers and the quantum of its obligations.

However, because arguability had been conceded before the judge and not revisited on appeal, the Court did not decide the point. The judgment nonetheless signals a warning: proprietary labels may not overcome the debtor–creditor structure of bank accounts.

3.3 Impact

(A) Limitation strategy in fraud-adjacent bank disputes

The case underscores that in s.32 disputes the key question is often not “when did the claimant obtain definitive proof of fraud?” (e.g., an expert report), but “when did the claimant have enough primary facts to plead dishonesty as more likely than innocence?”. Here, the Court treated 2013 as the latest point—years before the 2016 CARM report or 2018 expert analysis.

(B) AIC/Barrell is not an appeal in disguise

The decision illustrates the practical limit of the AIC jurisdiction: even if a judge’s earlier reasoning is said to have contained an error, a re-opening will not be compelled where (i) the judge does not accept a mistake, and (ii) the alleged error is not material to the outcome. Litigants must therefore treat the first hearing as the critical moment to marshal limitation and knowledge arguments.

(C) Pleading discipline: deceit, reliance, and alternative causes

The Court’s unchallenged confirmation that deceit requires reliance is a cautionary pleading point. In banking “unauthorised account/loan” narratives, claimants may instinctively plead “fraud/deceit”, but unless they can show they were induced to act, deceit may be structurally unavailable.

(D) Proprietary remedies against banks: an implied headwind

Although not decided, the Court’s analysis of the debtor–creditor model and the absence of a “transfer of a specific asset” is likely to be cited by defendants resisting constructive trust/proprietary restitution attempts in misposting/unauthorised transfer cases. Claimants may need to pivot to:

  • contractual unauthorised payment claims;
  • Quincecare-style frameworks where applicable (not in issue here);
  • statutory data/credit-reporting remedies (again not determined here);
  • or carefully articulated equitable claims identifying a proprietary base (if any) beyond the mere account balance.

4. Complex Concepts Simplified

Strike out / Summary judgment
A strike out removes a claim because it discloses no legally viable cause of action (even assuming the facts are true), or is otherwise improper. Summary judgment ends a claim without trial because it has no real prospect of success and there is no other compelling reason to have a trial.
Deceit
A tort requiring (among other elements) a false representation, knowledge of falsity, intent that the claimant rely on it, actual reliance by the claimant, and resulting loss. If the claimant was not “taken in”, deceit usually fails.
Section 32 Limitation Act 1980
A rule that can delay the start of the limitation clock where (a) the action is “based upon” fraud, or (b) a fact necessary to complete the cause of action was deliberately concealed. The clock starts when the claimant actually discovered the fraud/concealment or could with reasonable diligence have discovered it.
“Based upon the fraud”
Not every claim mentioning fraud qualifies. Fraud must be a necessary ingredient of the cause of action (per Beaman v ARTS and Seedo v El Gamal).
Reasonable diligence and pleading dishonesty
The standard is not certainty. It is whether the known primary facts allow dishonesty to be pleaded as a more likely inference than innocence/negligence, applying authorities summarised in Sofer v Swissindependent Trustees SA.
AIC/Barrell jurisdiction
A judge may reconsider a judgment before it is embodied in a sealed order, but it is an exceptional procedural safety-valve, not a routine opportunity to re-argue.
Constructive trust (institutional vs remedial)
An institutional constructive trust is treated as arising automatically from circumstances recognised by law (and may have different limitation implications). A remedial constructive trust is imposed by the court as a remedy for wrongdoing; per Williams v Central Bank of Nigeria, a six-year period applies.
Chose in action
An enforceable right (like a debt) rather than a physical asset. A bank account credit balance is the customer’s right to be paid by the bank, not ownership of specific money.

5. Conclusion

James & Anor v HSBC UK Bank PLC (Rev1) [2026] EWCA Civ 88 is a limitation-and-procedure driven decision with broader practical lessons. It reinforces that:

  • Deceit requires reliance; alleging dishonesty is not enough if the claimant did not act on the misrepresentation.
  • For s.32 LA 1980, the question is when the claimant had enough primary facts to plead dishonesty—not when they obtained definitive proof or expert confirmation.
  • The AIC/Barrell jurisdiction will not be used to re-open a conclusion where the alleged error is uncertain or immaterial.
  • The Court’s dicta highlights a significant conceptual difficulty for proprietary/constructive trust claims against banks based purely on movements in account balances, given the debtor–creditor model of banking.

In the wider legal context, the judgment strengthens defendant reliance on early “knowledge” dates in fraud-adjacent limitation disputes and serves as a reminder that careful cause-of-action selection—and early limitation analysis—can be outcome-determinative.