Philipp v Barclays Bank UK PLC

In 2018 Mrs Fiona Philipp and her husband, Dr Robin Philipp, fell victim to a fraud. They were deceived by criminals into instructing Barclays Bank (“the Bank”) to transfer £700,000 in two payments from Mrs Philipp’s current account with the Bank to bank accounts in the United Arab Emirates (“UAE”). The payments were made and the money was lost. In these proceedings Mrs Philipp claims that the Bank is responsible for this loss. She contends that the Bank owed her a duty under its contract with her or at common law not to carry out her payment instructions if - as was allegedly the case here - the Bank had reasonable grounds for believing that she was being defrauded.

The Bank applied to have the claim summarily dismissed on the ground that as a matter of law it did not owe Mrs Philipp the alleged duty. Judge Russen QC, sitting as a judge of the High Court, agreed with this submission and granted summary judgment in favour of the Bank: [2021] EWHC 10 (Comm), [2021] Bus LR 451. But an appeal by Mrs Philipp to the Court of Appeal was allowed: [2022] EWCA Civ 318, [2022] QB 578. Birss LJ, in a judgment with which Sir Julian Flaux C and Coulson LJ agreed, accepted the claimant’s argument that, in principle, a bank owes a contractual duty to its customer of the kind alleged: whether such a duty arose on the facts in this case is a question which can only be decided at a trial.

For the reasons explained in this judgment, the Court of Appeal was, in my opinion, wrong to reach this conclusion, which is inconsistent with first principles of banking law. It is a basic duty of a bank under its contract with a customer who has a current account in credit to make payments from the account in compliance with the customer’s instructions. This duty is strict. Where the customer has authorised and instructed the bank to make a payment, the bank must carry out the instruction promptly. It is not for the bank to concern itself with the wisdom or risks of its customer’s payment decisions.

It would be possible for a bank to agree as an express term of the contract that it will not comply with a payment instruction given by the customer if the bank believes, or if the bank has reasonable grounds for believing, that the customer has been tricked by a third party into authorising the payment. But it is not suggested that the contract between Mrs Philipp and the Bank contained any such express term. In the absence of an express term, no obligation of this kind can be implied or said to be inherent in the relationship between a bank and its customer. To the contrary, such an obligation would be inconsistent with the normal contractual basis on which banking transactions are conducted.

The Court of Appeal derived the alleged duty by extrapolating from the reasoning in the case of Barclays Bank plc v Quincecare Ltd [1992] 4 All ER 363. In this and other similar cases, courts have held that a bank has a duty not to execute a payment instruction given by an agent of its customer without making inquiries if the bank has reasonable grounds for believing that the agent is attempting to defraud the customer. However, as I will discuss later in this judgment, the reason why the bank owes a duty to its customer to make inquiries in such cases is to ensure that it does not make a payment which the customer has not authorised. This reasoning does not apply to cases of the present kind where the customer has unequivocally authorised and instructed the bank to make a payment.

The type of fraud which occurred here is a growing social problem and can undoubtedly cause great hardship to its victims, as the sad facts of this case make all too clear. Whether victims of such frauds should be left to bear the loss themselves or whether losses should be redistributed by requiring banks which have made or received the payments on behalf of customers to reimburse victims of such crimes is a question of social policy for regulators, government and ultimately for Parliament to consider. It is in fact, as I will mention in more detail shortly, the subject of new legislation. But it is not a question for the courts. It is not the role of the courts to formulate such policy, still less to impose on the parties to a contract an obligation to which they have not consented and cannot reasonably be presumed to have consented since it is inconsistent with the normal and established allocation of risk and responsibility under contracts of the relevant type.

In the following sections of this judgment I will first describe in further detail the type of fraud involved in this case, the relevant facts and the questions of regulatory policy which this type of fraud raises. I will then consider the relevant contractual duties owed by a bank to a customer who has a current account with the bank. Finally, I will examine the reasoning of the Court of Appeal and other arguments made by the claimant in this case.

The generic name for the type of fraud committed in this case is “authorised push payment” (APP) fraud - so called because the victim is induced by fraudulent means to authorise their bank to send a payment to a bank account controlled by the fraudster. (Fraud of this kind is contrasted with “pull” payment fraud, where payments are extracted from the victim’s bank account or debited to a card by a criminal without the victim’s authority.)

APP fraud can take a variety of forms. Half yearly reports published by UK Finance Ltd, the official trade body representing the banking and finance industry in the United Kingdom, describe eight different forms of such fraud, which are not exhaustive. One of these involves the criminal claiming to be a police officer or employee of the victim’s bank. The scam often begins with a phone call or text message claiming that there has been fraud on the victim’s account and that the victim needs to transfer the money to a “safe account” to protect their funds.

The present case is a particularly egregious example of this type of scam. Dr Philipp was contacted in February 2018 by an individual (the fraudster) who claimed to be working for the Financial Conduct Authority in conjunction with the National Crime Agency and to be investigating a fraud within HSBC and an investment firm, Tilney, where Dr Philipp held substantial savings. In a series of telephone calls Dr and Mrs Philipp were led to believe that their money needed to be moved to “safe accounts”.

It is unnecessary to relate the full history of events, as the main question raised on this appeal is one of law and (save in one limited respect) the factual details are not material. They can be found in the judgment of Judge Russen QC at paras 27-71. A striking feature of the facts is that Dr and Mrs Philipp were even persuaded by the fraudster not to cooperate with the police when they received a visit from a police officer, DC Claridge, warning them that she believed that a fraud was being perpetrated on them. The sophistication of the means employed is shown by the fact that telephone calls were made to Dr Philipp which appeared, from the number displayed on his mobile phone, to be coming on one occasion from the telephone number of the National Crime Agency and on another occasion from the mobile number of DC Claridge.

On 5 March 2018 (the same day as this visit from DC Claridge) Dr Philipp, at the instigation of the fraudster, caused £950,000 to be transferred from his investment account with Tilney to his wife’s current account with the Bank. On 10 March, and again on 13 March, Mrs Philipp attended a branch of the Bank in person with her husband and gave instructions for an international payment to be made from her account to a bank account in the UAE. On the first occasion her instruction was to transfer £400,000 to an account in the name of Lambi Petroleum Ltd (a company with which Dr Philipp told the cashier, falsely, that he had had previous dealings). On the second occasion the instruction given was to transfer £300,000 to an account in the name of Bonito Systems Ltd. On each occasion Dr and Mrs Philipp were following directions given by the fraudster. On each occasion, before making the transfer, the Bank telephoned Mrs Philipp to seek her confirmation that she had made the transfer request and wished to proceed with it. On each occasion Mrs Philipp provided the required confirmation. The Bank therefore made the payments in accordance with her instructions.

On 15 March 2018 the Philipps received a second visit from DC Claridge, accompanied by another uniformed officer. She told them that more people had been affected by the suspected fraud. Again they said that they did not wish to have any involvement with the police. However, on the next day the police contacted the Bank’s Police Liaison Officer, explained that they were undertaking a large-scale fraud investigation and said that they had received credible information that Mrs Philipp’s current account had been compromised by fraudsters in the UAE. The Bank immediately froze the account.

On 19 March 2018 Mrs Philipp visited her local branch of the Bank once more and sought to make a third transfer. On this occasion her instructions were to pay £250,000 (the remaining balance of the funds received from her husband’s account with Tilney) to the account in the UAE in the name of Bonito Systems Ltd. However, Mrs Philipp was told that her account had been blocked pending a review, and the transfer did not take place. Still acting at the request of the fraudster, she then called the Bank’s fraud department. The transcript of the conversation records that in this call Mrs and Dr Philipp tried, unsuccessfully, to persuade the Bank’s staff to lift the block on her account by claiming (falsely) that they needed to make an urgent payment under a contract which they would lose if the payment was not made on that day.

Following a third visit from DC Claridge on 26 March 2018, Dr and Mrs Philipp finally came to realise that they had been the victims of a fraud. Mrs Philipp notified the Bank of this on 27 March 2018. Some two months later, on and after 31 May 2018, the Bank made attempts to recall the funds which had been transferred to the UAE but these attempts were unsuccessful. Dr and Mrs Philipp have therefore lost £700,000. This money represented the bulk of their life savings.

Although the relationship between a bank and an account holder is one of contract, such contracts are made and performed in a heavily regulated legal environment. For example, where the customer who holds the account is a consumer, any term of the contract which is unfair to the customer is not binding on the customer: see section 62 of the Consumer Rights Act 2015. Banks are also subject to various regulatory duties imposed by or under legislation. The most significant legislation applicable to payments is the Payment Services Regulations 2017 (SI 2017/752). These Regulations, originally introduced to transpose the second EU Payment Services Directive 2015/2366 into national law, have been retained (with minor modifications) following the withdrawal of the United Kingdom from the European Union. They apply to a wide range of payment transactions.

Part 7 of the Payment Services Regulations establishes rights and obligations relating to the provision of payment services by banks and other “payment service providers”. Among other things, these regulations define when a payment transaction is to be regarded as having been authorised by the payer (regulation 67); impose obligations on payment service providers to execute authorised payment transactions promptly (regulations 86 and 91); and require a payment service provider to refund to the payer the amount of an unauthorised payment (no later than the end of the business day following the day on which it becomes aware of the unauthorised transaction) (regulation 76(2)).

The relevant provisions of the Payment Services Regulations are largely concerned with establishing rights and obligations of payment service users and providers in relation to unauthorised payment transactions. They do not provide for reimbursement of any payments which the payer has authorised. They therefore do not offer any means of redress for a customer who is a victim of APP fraud. Indeed, they have been perceived as containing an obstacle to the imposition of a regulatory obligation on payment service providers to reimburse customers who are victims of APP fraud. Regulation 90(1) provides that, where a payment order is executed in accordance with the unique identifier (ie the essential information identifying the account of the payee), the payment order is deemed to have been correctly executed.

The government agencies responsible for regulating the provision of banking services are the Financial Conduct Authority and, in relation to payments, the Payment Systems Regulator, established under section 40 of the Financial Services (Banking Reform) Act 2013. Section 68(1) of that Act allows a designated representative body to make a complaint to the Payment Systems Regulator that “a feature, or a combination of features, of a market in the United Kingdom for services provided by payment systems is, or appears to be, significantly damaging the interests of those who use … those services.” In September 2016 the Consumers’ Association, which is a designated representative body, made such a complaint to the Payment Systems Regulator about lack of protection for consumers against harm caused by APP fraud. The complaint argued that banks could take steps to reduce the risks of consumers authorising push payments to bank accounts controlled by fraudsters and that placing liability on banks (including the bank where the fraudster’s account is held) to reimburse losses from such scams would incentivise banks to take such measures.

Since this complaint was made, there have been a number of reports, consultations and regulatory initiatives on this subject. The main initiative has been the introduction in 2019 of a voluntary code for payment service providers, called the “Contingent Reimbursement Model Code”. This code covers measures aimed at reducing the incidence of APP fraud and also provides for the reimbursement of customers who are victims of such scams in certain cases (which do not include international payments). To date, however, the code has been adopted by only 10 payment service providers (who include Barclays).

The Financial Services and Markets Act 2023, which received Royal Assent on 29 June 2023, provides for a mandatory reimbursement scheme. Section 72 of the Act amends regulation 90 of the Payment Services Regulations to enable liability to be imposed “where the payment order is executed subsequent to fraud or dishonesty”. Section 72 requires the Payment Systems Regulator to impose a requirement for reimbursement by payment service providers in such “qualifying cases” of payment orders executed subsequent to fraud or dishonesty as the Regulator considers should be eligible for reimbursement. Such cases are, however, limited to payment orders executed over the Faster Payments Scheme. The proposed new scheme is also confined to consumers, charities and “micro-enterprises”: larger businesses are not included. It provides (subject to potential adjustment through a dispute resolution process) for a 50-50 allocation of losses between the sending and receiving providers. It is not proposed that the regulatory obligations arising under the scheme will be directly enforceable by bank customers.

It should go without saying that it is not the role of courts to make rules of this kind. There are several reasons for this. A fundamental reason is that the nature of adjudication requires courts to identify legal rights and duties which it is fair to treat as applicable to the parties to the lawsuit when the events giving rise to the dispute between them occurred. That is a very different exercise, which requires different forms of reasoning, from formulating policy or fashioning rules designed to regulate future dealings between different classes of persons.

Legislators and regulators have the institutional competence to take an overall view of a perceived social problem and to consider the appropriate policy response as a whole and from a variety of angles. They also have the competence and capacities: (i) to bring together a variety of perspectives from individuals with experience and expertise in relevant fields of knowledge; (ii) to acquire and evaluate information about the relative costs and benefits of different possible measures, both for those directly affected and for society at large; (iii) to consult a range of different bodies (and the public more generally); (iv) to design a comprehensive regime containing qualifications, exceptions and safeguards; and (v) in designing such a regime to set temporal, financial or other limits on its scope or otherwise to draw distinctions which may not have a principled basis but are considered to promote the common good by achieving an appropriate trade-off or compromise between different policy goals.

Courts are in a very different constitutional and institutional position. In deciding the case before them, they are bound to apply the laws made by Parliament and to respect precedents created by past judicial decisions. While courts, particularly at appellate level, have a responsibility to seek to adapt and develop the common law to keep it up to date which may sometimes require overruling previous decisions, they proceed by reasoning from established principles and are under a duty to promote consistency and predictability in the law. Nor do courts have the institutional capacities of legislators and regulators that I have described above.

There is a further fundamental reason why questions about the appropriate policy response to APP fraud are outside the scope of these proceedings. This is that the claimant’s case is based on a duty allegedly owed to her by the Bank under the contract between them governing the operation of her current account. The extent of the Bank’s responsibilities under this contract does not depend on an evaluation of whether it would be a good or a bad thing if banks were required, either generally or in some circumstances, to reimburse customers who are victims of APP fraud. It depends on an analysis of what the parties to the contract have agreed.

This requires the court to consider any relevant terms expressly agreed between the bank and the customer. A bank will invariably have standard terms on which it agrees to provide its services, and it will be necessary in due course to refer to the terms which, on the Bank’s case, were incorporated into its contract with Mrs Philipp. As with other types of commonly recurring contractual relationship, however, the contract between a bank and a customer who holds a current account is one for which there is an established legal template consisting of certain basic terms and duties which have come to be recognised by the common law (and sometimes statute) as ordinary incidents of contracts of this type. These implied terms and duties apply automatically by default unless modified or excluded by express agreement. In the case of contracts between a bank and an account holder, they are the subject of a well developed body of case law.

The claimant’s case in these proceedings is based on this body of case law. The key legal allegation made in the particulars of claim is that the Bank was under a duty “to refrain from executing an order from Mrs Philipp if and for as long as it was put on inquiry, by having reasonable grounds for believing that the order was an attempt to misappropriate funds from Mrs Philipp”. This duty is said to be implied by the common law into the contract between Mrs Philipp and the Bank. The central issue in this appeal is whether such a duty is either already recognised by the common law, or can and should be recognised by a principled extension of the existing case law, as an ordinary incident of the contract between a bank and its customer. To decide this issue, it is necessary to start by identifying certain basic legal aspects of the contract between a bank and its customer which are well established.

The starting point in understanding the contract between a bank and a customer who holds a current account with the bank is the decision of the House of Lords in Foley v Hill (1848) 2 HL Cas 28. This decision has been described as a historical breakthrough in the development of banking law: see Sir Ross Cranston, Principles of Banking Law, 3rd ed (2018), p 190. It established conclusively that under ordinary circumstances a bank is not a trustee or fiduciary of money deposited by a customer, but simply a debtor. Money deposited with a bank becomes the bank’s money, to lend or otherwise deal with (so far as the customer is concerned) as it thinks fit. The principal obligation owed by the bank is to discharge its debt to the customer when called upon to do so. Thus, the bank is obliged to repay to the customer on demand an equivalent sum to that deposited (plus any agreed interest and less any agreed charges) and also, so long as the account is in credit, to make payments in accordance with the customer’s instructions in reduction of its debt to the customer. Another basic principle that has long been established is that, in making such payments, the bank acts as the customer’s agent: see eg Westminster Bank Ltd v Hilton (1926) 43 TLR 124, 126 (Lord Atkinson).

As in the case of every contractual agency, a bank is bound to act in accordance with the authority conferred upon it by its principal and to perform what it has agreed to do: see eg Bowstead & Reynolds on Agency, 22nd ed (2021), article 36. The terms on which a bank is authorised and undertakes to carry out its customer’s instructions to make payments are generally referred to as the bank’s mandate from the customer.

Unless otherwise agreed, the bank’s duty to comply with its mandate is strict. Where the bank acts outside the mandate by making a payment which the customer has not authorised, it cannot debit the customer's account. Conversely, where the bank receives an instruction to make a payment given in accordance with the mandate, the ordinary duty of the bank is simply to carry out the instruction and to do so promptly. In Bodenham v Hoskins (1852) 21 LJ Ch 864, 869, Kindersley V-C said that:

“… the banker looks only to the customer, in respect of the account opened in that customer’s name, and whatever cheques that customer chooses to draw, the banker is to honour. He is not to inquire for what purpose the customer opened the account; he is not to inquire what the monies are that are paid into that account, and he is not to inquire for what purpose monies are drawn out of that account: that is the plain general rule, as between banker and customer.”

The same point was made in Lipkin Gorman v Karpnale Ltd [1989] 1 WLR 1340, 1356, where May LJ said that there is nothing in the contract between a bank and its customer which could require a banker to consider the commercial wisdom or otherwise of the particular transaction.

The main implied limit on the bank’s duty to carry out its customer’s authorised payment instructions (where the customer’s account is in credit) is that the bank cannot be obliged to act unlawfully. Thus, in Gray v Johnston (1868) LR 3 HL 1 the House of Lords held that, where a bank is instructed by a customer who is an executor or other fiduciary to make a payment which would be a breach of trust, the bank has no right to refuse to make the payment unless to do so would make the bank “in privity” (as Lord Cairns LC put it at p 11) with the breach of trust which would thereby be committed. Under the modern law this exception is defined by the principle that a person who dishonestly assists in a breach of trust or fiduciary obligation is liable to the party to whom the obligation is owed: see Royal Brunei Airlines Sdn Bhd v Tan [1995] 2 AC 378; Twinsectra Ltd v Yardley [2002] UKHL 12; [2002] 2 AC 164.

Unless otherwise expressly agreed, a genuine or reasonable concern on the part of the bank that it might incur a legal liability by carrying out the instruction is not enough: the concern must actually be valid. In Westpac New Zealand Ltd v MAP & Associates Ltd [2011] NZSC 89, [2011] 3 NZLR 751, a bank declined to comply with its customer’s instructions to make a payment and, when sued by the customer for breach of mandate, asserted in its defence that it reasonably believed that, by making the payment, it would have been dishonestly assisting in the commission of a breach of trust. The Supreme Court of New Zealand held that such a belief, even if based on reasonable grounds, did not afford the bank a defence. To justify its refusal to execute its customer’s instructions, the bank would have had to show that, had it done so, it would actually have incurred liability for dishonestly assisting in a breach of trust (or other wrongful conduct). As there was no evidence that there was in fact any wrongdoing which making the payment would have assisted, this requirement was not satisfied and the bank had no defence to the claim.

Another form of illegality which limits the bank’s duty to carry out its customer’s authorised payment instructions derives from legislation aimed at preventing money laundering. Under section 328 of the Proceeds of Crime Act 2002 it is an offence for a bank to make a transfer of funds if it suspects that this would facilitate the use of criminal property without first obtaining authorisation to make the transfer under section 338 of the Act. It is an implied term of the contract between the bank and its customer that the bank may refuse to execute the payment instruction if to do so would involve the bank committing such an offence: Shah v HSBC Private Bank (UK) Ltd [2012] EWHC 1283 (QB), paras 39-45. Once again, the basis for this exception to the bank’s duty is that the bank would be acting unlawfully if it executed the customer’s instruction.

As with any contract for the supply of services in the course of a business, there is a term implied by law in the contract between a bank and its customer that the bank must carry out the services with reasonable care and skill: see section 13 of the Supply of Goods and Services Act 1982 and section 49 of the Consumer Rights Act 2015. A similar duty is owed in tort; but as the duty in tort arises out of the contractual relationship, it can be no more extensive than the contractual duty and adds nothing to the analysis for present purposes.

The requirement to exercise reasonable care and skill only applies, and is only capable of applying, insofar as the contract gives the supplier any latitude in how the relevant services are carried out. Insofar as the contract prescribes what the supplier must do or achieve in carrying out the services, failure to do or achieve what is required will be a breach of the contract and it is irrelevant whether the supplier has acted with skill or care. The bank’s obligation to carry out payment instructions in accordance with its mandate from the customer leaves the bank with very little latitude in performing the obligation. But where the contract does not completely specify what the bank must do, it must act in the way that a reasonably skilful and careful banker would.

So, for example, where a customer’s instruction to transfer funds leaves the bank freedom to choose the method of transfer, the bank must exercise reasonable skill and care in making its choice: see Royal Products Ltd v Midland Bank Ltd [1981] 2 Lloyd’s Rep 194. Another situation where the duty of care applies is where the customer’s instruction leaves it unclear what the bank is being instructed to do. It is a general principle of the law of agency, applicable to a bank when acting in that capacity, that where the agent honestly and reasonably understands an instruction to bear a particular meaning and acts accordingly, the agent will not be in breach of contract by so acting: see eg Bowstead & Reynolds on Agency, 22nd ed (2021), para 6-010; Ireland v Livingston (1872) LR 5 HL 395; Midland Bank Ltd v Seymour [1955] 2 Lloyd’s Rep 147, 153. This presupposes, however, that in relying on this understanding the agent has acted reasonably. As Robert Goff LJ observed in European Asian Bank AG v Punjab & Sind Bank (No 2) [1983] 1 WLR 642, 656:

“If instructions are given to an agent, it is understandable that he should expect to act on those instructions without more; but if, for example, the ambiguity is patent on the face of the document it may well be right (especially with the facilities of modern communications available to him) to have his instructions clarified by his principal, if time permits, before acting upon them.”

In Hilton v Westminster Bank Ltd (1926) 135 LT 358 a customer sent a telegram instructing his bank not to pay a cheque which he had drawn. The telegram accurately identified the payee and the amount of the cheque but misstated the number of the cheque which the customer wanted the bank to stop. The bank paid the cheque on presentation and the question was whether, in doing so, the bank had acted in breach of duty. The Court of Appeal (Bankes, Warrington and Atkin LJJ) held that the circumstances were such as to put the bank on inquiry as to whether the customer had made a mistake in specifying the number of the cheque so that the bank should have sought clarification before paying the cheque. Its failure to do so was a breach of the bank’s duty to exercise reasonable skill and care in carrying out the customer’s instructions. On a further appeal to the House of Lords, no doubt was expressed about the principle that the bank owed such a duty of care, although the House of Lords held that there had not been a breach of this duty on the facts: Westminster Bank Ltd v Hilton (1926) 43 TLR 124.

At the centre of the argument on this appeal is the legal significance of a series of cases which have held that the duty of care owed by a bank to its customer requires a bank not to execute a payment instruction given by an agent of the customer without making inquiries if the bank has reasonable grounds for believing that the instruction is an attempt by the agent to misappropriate the customer’s funds.

The case which has come to be seen as the leading authority is Barclays Bank plc v Quincecare Ltd [1992] 4 All ER 363. However, it was not the first in this line of cases. The modern authorities start with Selangor United Rubber Estates Ltd v Cradock (No 3) [1968] 1 WLR 1555. That case arose from the collapse of a rubber company carrying on business in the Malay States. The directors had engaged in a fraudulent scheme to use the company’s funds to finance the purchase of its shares. A claim was brought in the name of the company to recover sums which had been misapplied. The defendants included two banks which had paid cheques drawn on the company’s accounts with them. The claims against the banks were advanced on two bases, one being that they had dishonestly assisted the directors’ breaches of duty and the other that they were negligent and in breach of a contractual duty of care owed to their customer in paying cheques without inquiring into whether the directors who signed the cheques were misapplying the company’s money.

After a lengthy review of earlier authorities, Ungoed-Thomas J concluded, at p 1609, that the bank’s duty to use reasonable skill and care “applies to interpreting, ascertaining, and acting in accordance with the instructions of a customer; and that must mean his really intended instructions as contrasted with the instructions to act on signatures misused to defeat the customer’s real intentions.” This duty was not excluded by the mandate authorising specified directors to sign cheques on the company’s behalf.

Selangor was followed in Karak Rubber Co Ltd v Burden (No 2) [1972] 1 WLR 602, another case involving a rubber company whose directors misapplied the company’s funds. Again, a claim was brought by the company against (among others) its bank for the value of a cheque drawn on the company’s account which the bank had paid. The judge, Brightman J, adopted and endorsed the reasoning in Selangor, holding that the bank was in breach of a duty of care owed to its customer to make inquiries before paying the cheque in circumstances where there were reasonable grounds for believing that the authorised signatories were “misusing their authority for the purpose of defrauding their principal or otherwise defeating his true intentions”: see p 629H.

The next case in this sequence was Lipkin Gorman (a firm) v Karpnale Ltd [1987] 1 WLR 987, where a partner in a firm of solicitors with authority to sign cheques drawn on the firm’s client account used such cheques to withdraw cash in order to fund his own gambling activities. Alliott J found the bank liable for dishonestly assisting in the solicitor’s breach of trust, though this finding was later reversed on appeal. The judge did not separately consider the firm’s claim for breach of a contractual duty of care but emphasised the importance of consistency between the “twin approaches” (see p 997A). He accepted, at p 1006, the following propositions put forward by counsel for the bank, Mr Sumption, as reflecting the established contractual duties of a bank:

“(1)

the bank is entitled to treat the customer's mandate at its face value, save in extreme cases; (2) the bank is not obliged to question any transaction which is in accordance with the mandate, unless a reasonable banker would have grounds for believing that the authorised signatories are misusing their authority for the purpose of defrauding their principal or otherwise defeating his true intention; (3) it follows that if a bank does not have reasonable grounds for believing that there is fraud, it must pay; (4) mere suspicion or unease do not constitute reasonable grounds and are not enough to justify a bank in failing to act in accordance with a mandate; and (5) a bank is not required to act as an amateur detective.”

This was the state of the authorities when Barclays Bank plc v Quincecare Ltd was decided in 1988. The judgment of Steyn J in that case was little noticed at the time and was not reported until 1992. But it has had a prominent afterlife.

The facts were that a bank agreed to lend money to a company (Quincecare) to purchase four chemist shops. Acting on instructions given by the company’s chairman, Mr Harry Stiller, the bank transferred funds purportedly in drawdown of the loan to a firm of solicitors. Mr Stiller had instructed the solicitors to receive the money on his behalf and transfer it to an account in the United States, which they did. By this means he misappropriated the money. The bank sued Quincecare (and another company which had acted as a guarantor of the loan) for repayment of the loan. In response the companies argued that the bank had paid out the money in breach of its mandate or of a duty of care owed to its customer.

In his judgment Steyn J first considered the terms of the mandate provided by Quincecare to the bank, which authorised the bank to comply with orders given on behalf of the company if signed by two directors or by the chairman. The chairman, Mr Stiller, had given a signed written payment order to the bank, and Steyn J found that this payment order was given in accordance with the mandate (see pp 371-374).

The judge then turned to consider Quincecare’s contention that the bank owed a duty of care to its customer in and about executing an order to transfer money from its current account. Steyn J referred to the earlier cases of Selangor, Karak and Lipkin Gorman, but said that “the importance of the matter warrants an independent examination of the bank’s duties” (p 375g). He then undertook such an independent examination of his own. His analysis has been treated in later cases as a satisfactory explanation for the conclusion reached without apparently being questioned. Because it is central to the claimant’s submissions and to the judgment of the Court of Appeal in this case, it will be necessary to look closely at Steyn J’s reasoning. For the moment I will quote in full, given its importance, the key passage of the judgment (at p 376):

“Given that the bank owes a legal duty to exercise reasonable care in and about executing a customer’s order to transfer money, it is nevertheless a duty which must generally speaking be subordinate to the bank’s other conflicting contractual duties. Ex hypothesi one is considering a case where the bank received a valid and proper order which it is prima facie bound to execute promptly on pain of incurring liability for consequential loss to the customer. How are these conflicting duties to be reconciled in a case where the customer suffers loss because it is subsequently established that the order to transfer money was an act of misappropriation of money by the director or officer? If the bank executes the order knowing it to be dishonestly given, shutting its eyes to the obvious fact of the dishonesty, or acting recklessly in failing to make such inquiries as an honest and reasonable man would make, no problem arises: the bank will plainly be liable. But in real life such a stark situation seldom arises. The critical question is: what lesser state of knowledge on the part of the bank will oblige the bank to make inquiries as to the legitimacy of the order? In judging where the line is to be drawn there are counter-vailing policy considerations. The law should not impose too burdensome an obligation on bankers, which hampers the effective transacting of banking business unnecessarily. On the other hand, the law should guard against the facilitation of fraud, and exact a reasonable standard of care in order to combat fraud and to protect bank customers and innocent third parties. To hold that a bank is only liable when it has displayed a lack of probity would be much too restrictive an approach. On the other hand, to impose liability whenever speculation might suggest dishonesty would impose wholly impractical standards on bankers. In my judgment the sensible compromise, which strikes a fair balance between competing considerations, is simply to say that a banker must refrain from executing an order if and for as long as the banker is ‘put on inquiry’ in the sense that he has reasonable grounds (although not necessarily proof) for believing that the order is an attempt to misappropriate the funds of the company…”

On the facts of the Quincecare case Steyn J found that the duty to refrain from executing a payment order did not arise, as there was nothing in the history of the matter which should have put the bank on inquiry as to Mr Stiller's honesty (p 381).

Shortly after Quincecare was decided, the Court of Appeal heard an appeal by the bank in Lipkin Gorman v Karpnale Ltd [1989] 1 WLR 1340. The appeal was allowed on the ground that the evidence did not justify the judge’s finding that the bank manager had dishonestly turned a blind eye to the fact that the solicitor was drawing on his firm’s client account in fraud, nor even a finding that the bank had been negligent. The bank did not dispute, however, that it would have been liable for breach of a contractual duty of care if a reasonable banker would have had reasonable grounds for believing that the solicitor was operating the account in fraud of the claimant firm (see p 1377). Parker LJ, with whose judgment Nicholls LJ agreed on this issue, identified the relevant question (at p 1378B) as being:

“whether, if a reasonable and honest banker knew of the relevant facts, he would have considered that there was a serious or real possibility, albeit not amounting to a probability, that its customer might be being defrauded ...”

May LJ, at p 1356, emphasised that any duty requiring the banker to exercise care must be limited and that, “having in mind the vast numbers of cheques which are presented for payment every day in this country,” he would expect it to be only in rare circumstances that a cheque should not be paid immediately on presentation and inquiry made.

There was a further appeal to the House of Lords but by that stage the firm’s claim against the bank was no longer pursued: see Lipkin Gorman (a firm) v Karpnale Ltd [1991] 2 AC 548, 570.

After the appeal in Lipkin Gorman, this line of authority lay more or less dormant in England and Wales for many years. That changed with the decision in Singularis Holdings Ltd v Daiwa Capital Markets Europe Ltd [2017] EWHC 257 (Ch), [2017] Bus LR 1386. In that case Rose J held that a stockbroker and investment bank had acted in breach of duty when it paid out money held for a corporate customer on instructions given by the company’s sole shareholder, chairman, president and treasurer (Mr Al Sanea) who was the only authorised signatory for the account. The payments were made to other companies in a group which Mr Al Sanea controlled. The judge found that any reasonable banker would have realised that there were obvious signs that Mr Al Sanea was perpetrating a fraud on the company when he instructed the money to be paid to other parts of his business operations; and that in making the disputed payments without proper or any inquiry, the bank had been negligent.

There does not appear to have been any dispute in Singularis about the legal duty owed by the bank. The only two cases cited on that issue were Quincecare and the decision of the Court of Appeal in Lipkin Gorman. The main disputed issue of law was whether Mr Al Sanea’s fraud should be attributed to the company which he controlled so as to defeat the claim. Rose J held that it should not. That decision was affirmed by the Court of Appeal (see [2018] EWCA Civ 84, [2018] 1 WLR 2777) and ultimately by the Supreme Court (see [2019] UKSC 50, [2020] AC 1189). Although the existence and nature of the bank’s duty was not in issue, it is apparent that Baroness Hale of Richmond, who gave the only judgment in the Supreme Court with which the other Justices agreed, saw the duty as limited to a situation where a payment instruction is given to the bank by an agent of the customer. She said, at para 35:

“The context of this case is the breach by the company's investment bank and broker of its Quincecare duty of care towards the company. The purpose of that duty is to protect the company against just the sort of misappropriation of its funds as took place here. By definition, this is done by a trusted agent of the company who is authorised to withdraw its money from the account.”

I observe in passing that, so far as I have been able to find, Singularis is the first case in which the relevant duty was referred to as the “Quincecare duty”.

Since Singularis, there have been further cases in which the “Quincecare duty” has been raised, but they have also not required the courts to analyse its juridical basis. In Nigeria v JP Morgan Chase Bank, NA [2019] EWCA Civ 1641, [2019] 2 CLC 559, the Court of Appeal affirmed a decision of Andrew Burrows QC (sitting as a High Court judge) that particular express terms in an agreement between a bank and its customer did not, on their proper interpretation, have the effect of excluding a “Quincecare duty of care”. In JP SPC 4 v Royal Bank of Scotland International Ltd [2022] UKPC 18, [2023] AC 461, the Privy Council on an appeal from the Isle of Man held that no “Quincecare duty” could be owed to a person who was not a customer in a contractual relationship with the bank but was the beneficial owner of funds held in the customer’s account. In Stanford International Bank Ltd v HSBC Bank plc [2022] UKSC 34, [2023] 2 WLR 79, this court upheld a decision to strike out a claim based on an alleged breach of the “Quincecare duty”, which Lady Rose (who gave the majority judgment) described, at para 4, as:

“a duty on the bank to refuse to comply with a payment instruction given by the person mandated by the customer to give such an instruction when the bank is on notice that the instruction may be part of a fraud on the customer, unless and until the bank’s inquiries satisfy it that the instruction is validly authorised by the customer.”

It was assumed, however, for the purpose of the strike out application that the alleged duty was owed and that the bank was in breach of it. So, as Lady Rose made clear, the judgment was not concerned with the scope of the “Quincecare duty”.

Leaving aside JP SPC 4, which involved an unsuccessful attempt to extend the scope of the bank’s duty to someone who was not its customer, all the cases in this line of authority have involved the same basic factual situation. In each case a payment instruction was given to the bank by an agent who was an authorised signatory of the customer’s account but was acting in fraud of the customer. As just noted, in Singularis Baroness Hale regarded this common factor as a defining characteristic of the “Quincecare duty”.

Giving the judgment of the Court of Appeal in this case, Birss LJ acknowledged that this common factor is “undeniable”: [2022] QB 578, para 27. But he observed, quite rightly, that what matters is the reasoning in these cases. The claimant’s main submission, which was accepted by the Court of Appeal, is that the reasoning in the Quincecare line of cases does not depend on whether the instruction is given by an agent but is capable of being applied with equal force to the situation of an instruction given by a customer who is an unwitting victim of APP fraud, where the bank is on inquiry as to the fraud. Alternatively, the claimant submits that the “Quincecare duty” is simply one aspect of the bank’s duty, implied in the contract, of reasonable skill and care in and about executing the customer’s orders, and forms part of a wider genus which requires a bank to refrain from executing an order without making inquiries if it has reasonable grounds for believing that the customer is a victim of fraud.

I have quoted at para 46 above the key passage of the judgment in Quincecare containing Steyn J’s analysis of the bank’s duties. Notably, he saw the case as one “where the bank received a valid and proper order which it is prima facie bound to execute promptly” but “where the customer suffers loss because it is subsequently established that the order to transfer money was an act of misappropriation of money by the director or officer.” As I will explain later, it was in this initial characterisation of the case as one where the bank received a “valid and proper order” that the analysis, in my opinion, started to go wrong.

From this starting-point, the reasoning proceeded in two stages. The first step was to depict the bank as owing two conflicting contractual duties to its customer. On the one hand, the bank has a duty to execute a valid order to transfer money promptly. On the other hand, the bank owes a duty to exercise reasonable care in and about executing a customer’s order to transfer money. Steyn J posed the question: how are these conflicting duties to be reconciled?

The answer, as he saw it, lay in striking a balance between countervailing policy considerations, described as follows:

“The law should not impose too burdensome an obligation on bankers, which hampers the effective transacting of banking business unnecessarily. On the other hand, the law should guard against the facilitation of fraud, and exact a reasonable standard of care in order to combat fraud and to protect bank customers and innocent third parties.”

The second step in the reasoning was to balance these competing considerations to arrive at what Steyn J judged to be a “sensible compromise”, being that “a banker must refrain from executing an order if and for as long as the banker … has reasonable grounds … for believing that the order is an attempt to misappropriate the funds of the company.”

The claimant argues, and the Court of Appeal accepted, that the “Quincecare duty” as formulated by Steyn J is directly applicable in cases of APP fraud on the footing that in such cases the order to transfer money is “an attempt to misappropriate” the customer’s funds. This is how Mrs Philipp’s case is pleaded in her particulars of claim (see para 27 above). The Court of Appeal adopted this characterisation of the case. It is a persistent theme of the judgment of Birss LJ that, where the customer is a victim of APP fraud, the order given to the bank is “an attempt to misappropriate funds” (see paras 27, 30, 50, 71, 76).

As counsel for the Bank pointed out, however, this argument involves a distortion of the language used in Steyn J’s judgment. A customer who is tricked by a fraudster into instructing her bank to make a payment is not attempting to misappropriate funds. She is attempting to cause the bank to transfer funds which are hers to dispose of in accordance with her own wishes. The consequence of executing the customer’s instruction may be that the funds are misappropriated by a third party. But that is a different factual situation from the one considered in Quincecare and other cases in that line of authority where a person giving an instruction purportedly on the customer’s behalf is thereby attempting to misappropriate funds from the customer. The claimant’s case and the Court of Appeal’s judgment gloss over this distinction.

I do not, however, agree with the submission of counsel for the Bank that this point is fundamental to the Court of Appeal’s analysis. The critical part of the reasoning of Birss LJ does not depend on it. What is critical is the notion that a bank’s primary duty to execute a valid payment order potentially conflicts (or, as Birss LJ preferred to say, “operates in tension”) with its duty to exercise reasonable skill and care in and about executing such an order and that this conflict or tension may be resolved by requiring the bank in certain circumstances not to execute an order without making inquiries (see paras 27-28, 34). This reasoning is drawn directly from the judgment of Steyn J in Quincecare. If Steyn J’s reasoning is followed, then the question whether a duty of the kind alleged arises in cases of APP fraud depends upon an assessment of countervailing policy considerations. The relevant assessment was not made in Quincecare itself, as Steyn J was only concerned with the situation where the payment instruction is an attempt by a dishonest agent to misappropriate the principal’s funds. The characterisation of the present case as falling within what was decided in Quincecare therefore cannot be accepted. But if the appropriate method for determining the scope of the bank’s duty is by seeking to strike a “sensible compromise” or “fair balance” between competing policy considerations, then it is open to a litigant to seek to persuade the court that this balance favours extending the “Quincecare duty”, or recognising a similar duty, in cases of APP fraud.

Where in such cases the line is to be drawn would depend, if Steyn J’s approach is followed, on trying to judge the extent to which imposing liability on banks would, on the one hand, hamper the effective transacting of banking business and, on the other hand, be effective in protecting customers against fraud, and on deciding how much weight to attach to each of these policy goals. If this were a necessary or appropriate exercise for the court to undertake, then I do not think that the Court of Appeal could be faulted for holding that the right occasion on which to undertake it would be at a trial, with the assistance of evidence from experts on banking practice. All that the Court of Appeal ultimately decided as a matter of law is that “it is … at least possible in principle that a relevant duty of care could arise in the case of a customer instructing their bank to make a payment when that customer is the victim of APP fraud” (see para 78). They considered that whether such a duty in fact arises in this case is not a question suitable for summary determination. If the reasoning of Steyn J in Quincecare is correct, this is a logical approach to take.

In my view, however, the analytical approach adopted in Quincecare and followed by the Court of Appeal here does not withstand scrutiny. It is flawed at each stage. The flaws stem from trying to find a way to reach a sound conclusion (already reached in Selangor and Karak) starting from a mistaken premise. The arguments in this case have served to expose these flaws, which I will first point out before addressing the initial premise.

The first flaw is to regard the bank’s duty of care as potentially conflicting with its duty to execute its customer’s payment instruction. (To state that the duties “operate in tension” seems to me just another way of saying the same thing.) On a proper understanding of the bank’s duties, there cannot be such a conflict. As discussed earlier, the duty to exercise reasonable skill and care only arises where the validity or content of the customer’s instruction is unclear or leaves the bank with a choice about how to carry out the instruction (see paras 35-36 above). In such cases, as described in Selangor, the duty of skill and care “applies to interpreting, ascertaining, and acting in accordance with the instructions” of the customer (see para 40 above). Where the bank receives a valid payment order which is clear and leaves no room for interpretation or choice about what is required in order to carry out the order, the bank’s duty is simply to execute the order by making the requisite payment. The duty of care does not apply.

In his judgment in Quincecare Steyn J initially appeared to recognise this when he said (at p 376c) that the bank’s duty to exercise reasonable skill and care in and about executing a customer’s order to transfer money “must generally speaking be subordinate” to the bank’s other duties. This is clearly right. The duty to exercise skill and care is subordinate to the bank’s duty to carry out a customer’s order to transfer money because the exercise of skill and care, where there is any scope for it, is directed solely to the effective execution of the order. It cannot therefore provide an independent basis for failing or refusing to execute a valid payment order. It is impossible to derive from a duty to observe reasonable skill and care in and about executing a customer’s order a duty not to execute the customer’s order. Indeed, as Lord Sumption NPJ recently observed in PT Asuransi Tugu Pratama Indonesia TBK v Citibank NA [2023] HKCFA 3, para 14:

“The law cannot coherently treat compliance with an authorised instruction as a breach of duty...”

The point goes further. It is not only a mistake to suppose that the bank’s duty of care is capable of conflicting with and potentially displacing its duty to execute a valid payment order from its customer. It is also wrong to suppose that, in the absence of some expressly or impliedly agreed exception to that primary duty, there can be any contractual justification for not executing such a payment order which would afford a defence to a claim by the customer for breach of mandate.

The second flaw in the analysis in the Quincecare case follows on from the first. It lies in the method used to try to escape from the predicament of having to reconcile conflicting contractual duties. Having perceived a conflict which does not in reality exist, Steyn J had no principled way in which to resolve it. In these circumstances he resorted to reliance on policy considerations.

The problem with this approach is that it is not an appropriate method for identifying what duty is owed by a party pursuant to a contract. What rule would represent a “sensible compromise” or “fair balance” between broad policy goals is a matter for legislators and other policy-makers to consider. But such regulation must be seen for what it is - imposing rules on banks and their customers, whether they would choose them or not, in the interests of what is considered to be the overall public good. In deciding whether a party to a contract can be regarded as having undertaken an obligation to the other party without having done so expressly, the aim of the courts is the more modest one of seeking to give effect to the presumed common intention of the contracting parties. A duty to combat fraud or to protect customers (let alone innocent third parties) against fraud is not an ordinary incident of the contractual relationship between a bank and its customer. Nor is any wider public interest in promoting those goals a proper basis on which to identify an implied term of the contract.

These flaws in the reasoning in the Quincecare case have led scholars to doubt the existence of the legal duty which that reasoning was seeking to justify: see eg Peter Watts QC, “The Quincecare duty: misconceived and misdelivered” [2020] JBL 403; Chua Rui Yuan, “The Quincecare duty: an unnecessary gloss?” [2023] JBL 161. There is, however, a significant body of authority, apart from Quincecare, to which I have referred that recognises the existence of the duty; and rejecting the reasoning in the Quincecare case should not, in my view, lead to rejecting the conclusion. The attempt in Quincecare to justify the duty went down a wrong track because it proceeded from a false premise. This obscured the true basis for the duty. The false premise was a mistaken assumption about the effect of dishonesty on an agent’s authority.

This assumption is reflected in the statement of Steyn J in Quincecare, at p 376c, that:

“Ex hypothesi one is considering a case where the bank received a valid and proper order which it is prima facie bound to execute promptly on pain of incurring liability for consequential loss to the customer” (p 376c).

It appears to have been assumed without discussion that the validity of the payment order given by Quincecare’s chairman, Mr Stiller, to the bank was unaffected by the fact that it was given dishonestly and for the purpose of defrauding the company.

The same assumption can be seen in the trilogy of cases mentioned earlier which were decided before Quincecare. In each of those cases authorised signatories who signed cheques instructing the bank to make a payment for the purpose of defrauding their principal were described as “misusing” their authority: see Selangor at p 1609D (quoted at para 40 above); Karak at p 629G-H (quoted at para 41 above); and Lipkin Gorman at p 1006, proposition (2) (quoted at para 42 above). Misuse of authority is distinct from absence of authority. The underlying assumption is that an agent who is acting dishonestly and in fraud of their principal is not on that account acting without authority: the only consequence is that the agent is in breach of duty towards the principal and therefore liable to the principal for loss caused by the breach.

This view reflects how, when these cases were decided, the law was stated in the leading commentary on the English law of agency, Bowstead on Agency (now called Bowstead & Reynolds on Agency, reflecting the great contribution of Professor Francis Reynolds to this work). The 15th edition (which was cited, albeit on another point, by Steyn J in Quincecare at pp 375-376) stated at article 74:

“An act of an agent within the scope of his actual or apparent authority does not cease to bind his principal merely because the agent was acting fraudulently and in furtherance of his own interests.”

This statement is unexceptionable as regards apparent authority (which I will discuss shortly). The proposition, however, that an agent’s actual authority includes authority to act fraudulently and in furtherance of his own interests cannot be supported.

The text was changed in the 17th edition, published in 2001. The principle is now clearly, and in my opinion correctly, stated in Bowstead & Reynolds on Agency, 22nd ed (2021) at article 23 as follows:

“Authority to act as agent includes only authority to act honestly in pursuit of the interests of the principal.”

The basis for this proposition is elucidated both in the comment that follows it and in a valuable article by the current main editor, Peter Watts, “Actual authority: the requirement for an agent honestly to believe that an exercise of power is in the principal's interests” [2017] JBL 269. I would express it as follows.

In principle, the scope of an agent’s authority is a matter of agreement between the agent and the principal. Where that agreement is recorded in writing, the question is one of interpretation of the document. No doubt it would be possible in theory for a principal in appointing an agent to agree that the agent may bind the principal even if and when the agent is acting dishonestly with the aim of defrauding the principal. But it seems inconceivable that any sane person would ever agree, or could reasonably be presumed to have agreed, to confer such authority on an agent. As is generally the case in commerce, parties to an agency relationship naturally deal with each other on an unspoken common assumption that each will act honestly in relation to the other. It goes without saying that authority conferred on an agent does not encompass acting dishonestly to further the agent’s own interests in opposition to the interests of the principal.

A clear statement of the legal principle can be found in Lysaght Bros & Co Ltd v Falk (1905) 2 CLR 421, a decision of the High Court of Australia, where O’Connor J said, at p 439:

“Every authority conferred upon an agent, whether express or implied, must be taken to be subject to a condition that the authority is to be exercised honestly and on behalf of the principal. That is a condition precedent to the right of exercising it, and, if that condition is not fulfilled, then there is no authority, and any act purporting to have been done under it, unless in a dealing with innocent parties, is void.”

I agree with the comment in Bowstead & Reynolds on Agency, para 3-012, that the reference in this passage to dealing with innocent parties is to be understood as a reference to the possibility of apparent authority.

The main case which provided some support for a contrary view was Hambro v Burnand [1904] 2 KB 10. The facts were that Burnand, who had authority to underwrite guarantee insurance policies on behalf of the other defendants, underwrote a policy insuring the claimants against the risk that a company to which they had lent money would fail to repay the loan. That company was in fact, to Burnand’s knowledge, insolvent. Furthermore, Burnand had a financial interest in the company (which he did not disclose to the claimants or to the other defendants) and was personally interested in keeping it afloat. The trial judge found that, in issuing the insurance policy, Burnand was acting in pursuit of his own interests, and not the interests of his principals, and that in these circumstances Burnand was acting outside the scope of his underwriting authority. The Court of Appeal, however, reversed the judge’s decision. The essential reason was, as Lord Collins MR put it at p 20, that:

“It would be impossible, as it seems to me, for the business of a mercantile community to be carried on, if a person dealing with an agent was bound to go behind the authority of the agent in each case, and inquire whether his motives did or did not involve the application of the authority for his own private purposes.”

See also Mathew LJ at pp 25-26.

This reasoning is a clear expression of the rationale for the principle of apparent authority. That principle protects the expectations of a third party who has reasonably relied on a representation by the principal that the agent has authority to bind the principal to a transaction of the relevant kind. It does so by preventing the principal from denying, as against the third party, that the agent had such authority: see eg Freeman & Lockyer v Buckhurst Park Properties (Mangal) Ltd [1964] 2 QB 480; Bowstead & Reynolds on Agency, 22nd ed (2021), article 72. The principal is thus treated, vis-à-vis the third party, as if the agent had the authority which the principal represented the agent to have, even though the agent had no such actual authority.

It seems to have been assumed in Hambro v Burnand that the claimants could not rely on the principle of apparent authority because there was no direct communication between the defendants who had appointed Burnand as their underwriting agent and the claimants. But today it would be regarded as a sufficient representation of authority that those defendants had placed Burnand in a position (as their underwriting agent) which carried with it the usual authority to underwrite insurance policies of the relevant kind in their names.

At all events, the expectations of third parties are irrelevant to the existence or otherwise of actual authority, which depends solely on what has been agreed between the principal and the agent. It is precisely because the third party is not privy to that agreement that the doctrine of apparent authority is needed to protect the third party’s reasonable expectations. To determine whether Burnand had actual authority to issue the insurance policy in question, it was therefore necessary to focus solely on whether, as between Burnand and the other defendants, his authority to underwrite policies on their behalf would reasonably have been understood to authorise him to issue policies intended to further his own private interests and which he did not honestly believe to be in the interests of his principals. The judge had rightly recognised that the answer to that question was obviously not.

The confusion from a modern perspective in the reasoning of the Court of Appeal is illustrated by the observation of Romer LJ, at p 25, that the claimants would not have been entitled to recover on the insurance policy against the persons defrauded (ie the other defendants) if they had taken the policy with notice of the fraud of the agent (ie Burnand). If Burnand had indeed had actual authority to bind the defendants to insurance policies written dishonestly in his own interests, the claimants’ rights to sue on the policy would not have been affected by notice that Burnand was acting dishonestly. Such notice is relevant only to the question of apparent authority.

The decision in Hambro v Burnand can be justified only on the ground that, although Burnand was acting outside the scope of his actual authority in issuing the guarantee policy in the names of the other defendants, he nevertheless had apparent authority to do so in circumstances where there was nothing to put the claimants on inquiry that Burnand was acting dishonestly in pursuit of his own interests.

Although the decision of the Court of Appeal in Hambro v Burnand was not overruled, its reasoning was effectively rejected by the House of Lords in Reckitt v Barnett, Pembroke & Slater Ltd [1929] AC 176. The claimant, Sir Harold Reckitt, gave a power of attorney to his solicitor, Lord Terrington, to manage his affairs while he was abroad. The powers granted included authority to sign cheques on his behalf, which Sir Harold confirmed in a letter to his bank was intended to apply “without restriction”. Lord Terrington drew a cheque on Sir Harold’s account in favour of the defendants to pay for a Rolls-Royce car which he was buying for himself and to settle his garage bill for a car that he was selling to the defendants in part exchange. The defendants accepted the cheque without inquiry. Sir Harold later sued them to recover the proceeds of the cheque. In response, the defendants asserted that the cheque was drawn by Lord Terrington with Sir Harold’s authority.

The trial judge rejected this defence, but the Court of Appeal by a majority (Scrutton and Sankey LJJ, with Russell J dissenting) reversed that decision: [1928] 2 KB 244. The rival contentions were set out clearly by Scrutton LJ, at p 257:

“I think the difference between the two points of view is this. One side starts with the proposition: ‘It cannot be within the authority given by a principal to his agent to rob him.’ If, therefore, you get an act done by an agent with his principal’s property for the agent’s own benefit, it is outside his authority … I understand the other point of view to be: The principal may be able to say to his agent: ‘You have abused your authority; account to me for the abuse.’ But when he has put the agent in a position to do a certain class of acts, and the agent has done an act of that class, apparently on behalf of his principal but really for his own benefit, the principal cannot say to a third party, who without notice of the agent’s abuse of authority has accepted the act which the agent has been put there to do: ‘I am not liable, for my agent, though purporting to act for me, acted for his own benefit.’ The apparent authority is the real authority.”

It is clear on the face of this passage that “the other point of view”, which represented Scrutton LJ’s own view, is an expression of the principle of apparent authority. It was misleading to say that “the apparent authority is the real authority”. If this meant that in the case posed by Scrutton LJ apparent authority is the only form of authority which the agent has, then it is true. But apparent authority is not actual authority. The point of it is that, where apparent authority arises, the principal is bound with respect to a third party as if the agent had authority to do the act in question when in reality the agent did not have any such actual authority.

In his dissenting judgment Russell J reasoned that the power of attorney given to Lord Terrington, properly construed, was limited to acting in the management of his principal’s affairs and could not be taken to have authorised him to use the principal’s money for his private purposes. As he put it, at p 268:

“An attorney cannot, in the absence of a clear power so to do, make presents to himself or to others of his principal’s property.”

Russell J further considered that the circumstances were such as to put the defendants on inquiry as to whether Lord Terrington had authority to use Sir Harold’s funds to settle his own private debt and that, had they made inquiries, they would have found that he did not.

On an appeal to the House of Lords, the reasoning of Russell J was approved and the appeal allowed. Viscount Sumner spelt out, at p 189, with particular clarity why the power conferred on Lord Terrington to draw cheques on his principal’s account (said in a letter to the bank to be “without restriction”) could not be taken to authorise drawing cheques other than for the principal’s purposes:

“It is unreasonable so to interpret the words that … he can under the letter, by drawing enough cheques on the appellant’s account, transfer all the principal’s property into his own pocket or apply it to his own use. We know that in fact nothing of this sort could have been meant by the parties to the power, for such self-sacrificing trust would be beyond the limits even of romance and it would be wrong to attach to words, which at most are ambiguous, a meaning which bears no relation to the realities of life.”

The power of attorney granted by Sir Harold Reckitt to Lord Terrington led to more litigation. As well as buying himself a Rolls-Royce, Lord Terrington drew a series of cheques as Sir Harold’s attorney made payable to himself and paid them into his own bank account with Midland Bank. Sir Harold sued Midland Bank for conversion of the cheques and the bank relied on a statutory defence under section 82 of the Bills of Exchange Act 1882 that it had received payment of the cheques in good faith and without negligence. The defence failed. Once again, the case reached the House of Lords on appeal: Midland Bank Ltd v Reckitt [1933] AC 1. The House affirmed the decision of the Court of Appeal that, in circumstances where the agent could be seen to be paying his own debt out of his principal’s money, the bank was negligent in making no inquiry as to the agent’s authority to make the payments. The same facts prevented the bank from relying on any apparent authority conferred by the power of attorney. In the words of Lord Atkin, with whose judgment the other law lords agreed, at pp 17-18:

“the notice found to exist defeats reliance on ostensible equally with actual authority. Neither in the one case nor in the other can the agent be assumed to have authority to pay his own debts with his principal’s money.”

Midland Bank Ltd v Reckitt illustrates the limit on the principle of apparent authority that it cannot exist where the third party has notice that the agent is acting without actual authority: see Bowstead & Reynolds on Agency, 22nd ed (2021), article 73. As already stated, the principle of apparent authority protects the expectations of a third party which acts in reliance on a representation by a principal that an agent has authority to bind the principal. This protection is only justified, however, if reliance on the representation is reasonable. Thus, the doctrine will not protect a third party which has reason to believe that the agent is acting without authority and fails to make inquiries that a reasonable person would have made in the circumstances to verify that the agent had authority.

Although this limit on apparent authority is supported by a long line of cases, doubt was created by the judgment of Lord Neuberger of Abbotsbury NPJ in the Hong Kong Court of Final Appeal in Thanakharn Kasikorn Thai Chamkat (Mahachon) v Akai Holdings Ltd [2011] 1 HKC 357; (2010) 13 HKCFA 479, paras 49-60. An issue in that case was whether a company’s chief executive officer had apparent authority to bind the company to a transaction with a bank. Lord Neuberger accepted a submission by counsel for the bank, Mr Sumption QC, that the bank was entitled to rely on the apparent authority of this individual unless the bank had actual knowledge of his lack of actual authority or its belief that he had such authority was dishonest or irrational.

This decision was followed in several cases in England and Wales but was the subject of trenchant criticism by Professor Peter Watts: “Some Wear and Tear on Armagas v Mundogas - The Tension between Having and Wanting in the Law of Agency” (2015) 1 LMCLQ 36, 48-56; and see also Bowstead & Reynolds on Agency, 22nd ed (2021), paras 8-049 – 8-050. In East Asia Co Ltd v PT Satria Tirtatama Energindo [2019] UKPC 30, [2020] 2 All ER 294, paras 70-93, the Privy Council found force in that criticism, disagreed with the test expressed in Akai and reaffirmed the orthodox view.

In a recent judgment in the Hong Kong Court of Final Appeal, Lord Sumption NPJ has suggested that there is in fact no difference between the law as stated by the Privy Council in East Asia and as stated in Akai and that the supposed conflict stems from a misunderstanding of the reasoning of Lord Neuberger in the latter case: see PT Asuransi Tugu Pratama Indonesia TBK v Citibank NA [2023] HKCFA 3, para 20. I confess to some scepticism about this suggestion but, be this as it may, I am in no doubt that this court should adopt the statement of the law in East Asia and hold the correct test to be that a third party cannot rely on the apparent authority of an agent if it failed to make the inquiries that a reasonable person would have made in all the circumstances to verify that the agent had that authority.

When these principles of agency law are applied to the factual circumstances of the Quincecare line of cases, the justification for the legal conclusion reached in those cases becomes clear. The authority conferred on an agent by a customer of a bank to sign cheques or give other payment instructions on behalf of the customer does not include authority to act dishonestly in pursuit of the agent’s own interests and in fraud of the customer. An agent acting in this way will therefore lack actual authority to give the instruction on behalf of the customer. The agent will still in general have apparent authority to do so by virtue of the customer’s representation to the bank that the agent is authorised to give payment instructions on its behalf. But not if there are circumstances suggestive of dishonesty apparent to the bank which would cause a reasonable banker before executing an instruction to make inquiries to verify the agent’s authority. In such circumstances the bank’s duty to exercise reasonable skill and care in and about executing the customer’s instructions requires the bank to make inquiries to ascertain whether the instruction given is one actually authorised by the customer. If the bank executes the payment instruction without making such inquiries, the bank will therefore be acting in breach of duty. Furthermore, the instruction will not bind the customer, as the dishonest agent will lack apparent as well as actual authority to give it on behalf of the customer.

There is no conflict between the bank’s duty of care to verify the agent’s authority and its duty to execute a valid order to transfer money promptly. The duty of care requires the bank, if put on inquiry, not to act without checking that the order is indeed a valid order of the customer to transfer money. The situation is analogous to the one discussed earlier where the customer’s instruction is patently ambiguous. If in either case the bank without taking steps to clarify the customer’s intention executes the order, the bank will be acting in breach of its duty of care and will also be acting outside the scope of its mandate.

Professor Watts, who has done much to illuminate this area of the law, disputes that conclusion. In a case note on the decision of the Court of Appeal in this case, he submits that the bank derives its authority to follow the agent’s instructions directly from the customer; hence the bank’s authority does not stand or fall with that of the agent, and prima facie the bank retains actual authority even where the agent is acting dishonestly in fraud of the principal: see Peter Watts, “Playing the Quincecare card” (2022) 138 LQR 530, 533; and also “Quincecare in the Hong Kong Court of Final Appeal”, Oxford Business Law Blog, posted 28 February 2023. The real question, he argues, is the content of the implied condition attached to the bank’s authority. In his view, the optimal default rule for a bank is that a bank should always follow the instructions of the customer’s authorised agent unless it actually knows and understands (or is wilfully blind to the fact) that the agent is acting dishonestly in relation to the principal: see Peter Watts, “The Quincecare duty: misconceived and misdelivered” [2020] JBL 403, 406.

I agree with Professor Watts that a mandate giving an agent power to operate a bank account may, depending on its wording, not only represent to the bank that the agent is authorised to sign or give other instructions on the customer’s behalf but also confer authority directly on the bank to follow such instructions. I do not, however, agree that - without express contrary wording - such a mandate should be construed as authorising the bank to follow instructions which the agent has neither actual nor apparent authority to give. The principle of apparent authority reflects the accommodation reached by English commercial law between the values of not subjecting persons to transactions to which they have not agreed and protecting the expectations of parties who have relied on an expression of such agreement made by someone purporting to act for such a person. As discussed above, the principle requires a representation of authority (or holding out) of the agent by the principal and reasonable reliance by the third party on that representation. I see no reason why a different default rule should apply to banks. A bank that relies unreasonably on an agent’s authority despite notice of matters that would have caused a reasonable banker to make inquiries cannot legitimately expect (in the absence of express agreement) to be immune from liability.

A further reason for this conclusion is that the requirement of reasonable reliance incorporated in the doctrine of apparent authority aligns with the duty to exercise reasonable skill and care which a banker, like any other supplier of a service, impliedly owes to its customer unless the duty is excluded. Professor Watts contends that the applicable test should vary according to the type of agent and the particular circumstances. However, the standard of reasonableness provides the necessary flexibility to adapt the principle to different types of commercial transaction and accommodate practices and expectations particular to a field of commerce. Exigencies of banking practice, such as the vast number of transactions which banks process on a daily basis and the speed at which, in the interests of efficiency, most banking business is expected to be conducted, will naturally affect what reasonable skill and care requires. As Steyn J said in the Quincecare case, at p 377:

“Everything will no doubt depend on the particular facts of each case. Factors such as the standing of the corporate customer, the bank's knowledge of the signatory, the amount involved, the need for a prompt transfer, the presence of unusual features, and the scope and means for making reasonable inquiries may be relevant.”

The judgment which has best articulated the juridical basis of the bank’s duty is the judgment of Ungoed-Thomas J in Selangor. In that case the customer’s mandate to the bank authorised named individuals to sign cheques on the company’s behalf and provided that cheques so signed “shall be valid and binding on the company”. It was submitted for the bank that its only concern in these circumstances was that the cheques should be signed by the authorised persons. The judge did not accept that this could be the correct construction of the mandate. After referring to Midland Bank Ltd v Reckitt, he said, at pp 1608-1609:

“The mandate is to enable the signatories to sign ‘on behalf of the company’ and bind it vis-à-vis the bank. Clearly the signatories, because they were authorised to sign ‘on behalf of the company’, could not rely on the mandate to pay for their own purposes. In reality, they would be signing on their own behalf. Nor could the bank rely on the mandate as conclusive to make such payment for the signatories’ purposes out of the company’s account.”

This passage recognises that a customer’s mandate is not to be construed as conferring authority on named signatories to instruct the bank to make payments on behalf of the customer for their own purposes, nor as conferring authority on the bank to make such payments on instructions given without the customer’s actual or apparent authority. But the judge did not fully follow through the logic of his own reasoning. The logical consequence of it is that a payment made without inquiry when the bank is on notice that the payment is being made for the signatories’ own purposes is not only a breach of the bank’s duty of care but is also outside the scope of its mandate. Hence the bank is not entitled to debit the amount to the customer’s account. This has the practical significance that, unless the customer is claiming damages for consequential losses over and above the amount of the payment, it is unnecessary for the customer in order to succeed to prove that, if reasonable inquiries had been made, the agent’s dishonesty would have been revealed and the loss avoided.

In summary, the duty of a bank which has come to be referred to as the “Quincecare duty” is not, as that epithet might suggest, some special or idiosyncratic rule of law. Properly understood, it is simply an application of the general duty of care owed by a bank to interpret, ascertain and act in accordance with its customer’s instructions. Where a bank is “put on inquiry” in the sense of having reasonable grounds for believing that a payment instruction given by an agent purportedly on behalf of the customer is an attempt to defraud the customer, this duty requires the bank to refrain from executing the instruction without first making inquiries to verify that the instruction has actually been authorised by the customer. If the bank executes the instruction without making such inquiries and the instruction proves to have been given without the customer’s authority, the bank will be in breach of duty. It will also in making the payment be acting outside the scope of its own authority from the customer and will therefore not be entitled to debit the payment to the customer’s account.

These principles are not limited to corporate customers. They apply wherever one person is given authority to sign cheques or give other payment instructions to a bank on behalf of another. They apply, for example, where under the mandate for a joint account either account holder has power to bind the other. It is therefore a misplaced criticism to suggest that the law gives companies a protection which is denied to individuals.

Similar reasoning would also apply where a bank is on notice, in the sense of having reasonable grounds for believing, that the customer lacks mental capacity to operate a bank account or manage her financial affairs. As illustrated by the decision of the Singapore Court of Appeal in Hsu Ann Mei v Oversea-Chinese Banking Corp Ltd [2011] SGCA 3, the bank’s duty of care may require the bank not to execute its customer’s instructions in such circumstances until further inquiries can be made.

On the other hand, these principles have no application to a situation where, as in the present case, the customer is a victim of APP fraud. In this situation the validity of the instruction is not in doubt. Provided the instruction is clear and is given by the customer personally or by an agent acting with apparent authority, no inquiries are needed to clarify or verify what the bank must do. The bank’s duty is to execute the instruction and any refusal or failure to do so will prima facie be a breach of duty by the bank.

Counsel for Mrs Philipp have argued that instructions given by an agent acting in fraud of the customer and instructions which the customer has been induced to give by fraud form part of a wider genus on the ground that in both cases the payment instruction does not reflect the customer’s true, genuinely held intention. In his oral submissions Mr Hugh Sims KC referred to the observation of Ungoed-Thomas J in Selangor (quoted at para 40 above) that the bank’s duty to act in accordance with the instructions of a customer “must mean his really intended instructions” and argued that the instructions given by Mrs Philipp to transfer money to accounts that she believed to be “safe accounts” but were in fact controlled by criminals did not represent her “really intended” instructions.

This argument is not a good one. The fact that an intention or desire results from a mistaken belief does not make it any less real or genuinely held. That is so however the mistaken belief has come about, including where it has been induced by another person’s deceit. The law will sometimes relieve a person from the consequences of a mistake. But it is untrue to say that in such cases the person did not “really” intend to act as they did. It is essential to intellectual clarity to distinguish between a situation where an act is not really intended and one where the act really is intended but the person would not have acted or formed the intention to act in that way if, for example, a lie had not been told to that person. As Lord Nicholls of Birkenhead put it in Shogun Finance Ltd v Hudson [2004] 1 AC 919, para 6:

“Fraud does not negative intention. A person’s intention is a state of mind. Fraud does not negative a state of mind.”

There is a similar obfuscation in the submission, which the Court of Appeal appears to have accepted, that a payment instruction is “vitiated” by APP fraud: see [2022] QB 578, para 75. This is similar to the misleading saying that fraud vitiates consent. In English law fraud does not vitiate consent to a contract or other transaction in the sense of nullifying consent and requiring it to be treated as non-existent: see eg Whittaker v Campbell [1984] QB 318, 326-328 (Robert Goff LJ); Shogun Finance Ltd v Hudson [2004] 1 AC 919, paras 6-8 (Lord Nicholls). Rather, the effect of fraud is generally to give the person who was induced to consent to a transaction by fraud a right to set aside the transaction (which that person may choose whether to exercise or not).

Importantly, the right to set aside the transaction applies only in relation to the fraudster and not in relation to other, innocent parties involved. In the context of banking transactions, the legal position is precisely explained in Fox, Property Rights in Money (2008), paras 5-153 – 5-155, in the following passage which merits quotation in full:

“The originator [ie the customer who gave the payment instruction] could not require the bank to re-credit his or her account when he had made a voidable payment to the beneficiary. This would be true even if the beneficiary had induced the payment by fraud. The reason is the abstraction of the mandate between the originator and the bank from the transaction between the originator and the beneficiary. The fraud in the payment transaction would not invalidate the originator’s mandate to the bank. Nor even could the originator argue that his or her instruction conferred a voidable mandate on the bank, which the originator could then rescind. The bank would not have committed any fraud on the originator so he or she could not exercise any right of rescission against it. The practical consequence is that the originator would be the only party with any rights of action arising from the voidable payment. The bank would not need to have any recourse against the beneficiary as it would have if it had paid without any effective mandate at all.”

In short, the fact that the customer’s payment instruction was induced by fraud entitles the customer to claim repayment from the fraudster but does not invalidate the instruction or give rise to any claim against the bank.

The duty of a bank to carry out its customer’s valid payment instructions is not without limit. I have already mentioned the implied condition that the bank cannot be required to carry out an unlawful act (see paras 31-33 above). It was also accepted by Mr Patrick Goodall KC on behalf of the Bank that, as stated in Karak at p 624, there is an implied condition that a bank will act honestly towards its customer.

The possibility of a further implied limitation is suggested by the Australian case of Ryan v Bank of New South Wales [1978] VR 555, 579, where McGarvie J pointed out that there can be circumstances in which a person who has a duty to execute an order given by another person would not reasonably be expected to comply literally with the order. He gave some examples:

“A carrier carrying goods under contract and ordered to deposit them on a particular loading platform at the factory of destination would act unreasonably in complying with the order if, on arrival, the factory was on fire and the platform about to be consumed by the fire. A security officer ordered by his employer to deliver the employees’ weekly pay to the paymaster would act unreasonably in complying with the order, if he knew from what the paymaster had told him, that the paymaster intended to steal it. A paymaster ordered by his employer to take the employees’ weekly pay to the pay office would act unreasonably in complying with the order if, on the way, he learnt that the pay office was occupied by a gang of armed robbers.”

As McGarvie J observed, these examples show that even provisions in a contract expressed in unqualified language will be interpreted as subject to an implied qualification which prevents them from applying in particular circumstances covered by the literal language of the provision, if it goes without saying that the parties must have intended there to be that qualification. Applying this principle to the situation where a bank receives a payment instruction, McGarvie J held that the bank should not comply with the instruction “if a reasonable banker properly applying his mind to the situation would know that the [account holders] would not desire their orders to be carried out if they were aware of the circumstances known to the bank”: see Ryan at p 581.

It is not necessary for the purpose of deciding this appeal to express any concluded view on whether this is the correct test to apply. The question is not one on which the court heard full argument. What is, however, relevant to note is that this test presupposes that the circumstances known to the agent (here the bank) are (as the agent is aware) not known to the principal. Thus, if a bank receives reliable information from a source such as the police suggesting that a customer’s payment instruction has (unknown to the customer) been procured by fraud, it may be right for the bank to refrain from executing the instruction without first alerting the customer to this information and verifying whether the customer wishes to proceed with the transaction. On the other hand, circumstances such as the destination and amount of the payment which the customer has instructed the bank to make cannot support an inference that the customer, if aware of these circumstances, would not desire the instruction to be carried out when it is plain that the customer is already aware of these circumstances.

In the present case circumstances which the claimant contends should have put the Bank on inquiry that a fraud was being perpetrated on her include: the large and unprecedented sum of money received in her current account; the size of the payments which she instructed the Bank to make; the fact that the payments were to bank accounts in the UAE; and the fact that the payees were companies with which she had no previous history of dealing. All these circumstances, however, were facts of which Mrs Philipp was to the Bank’s knowledge aware, as she gave the instructions and confirmed the payment details to members of the Bank’s staff in person. The Bank therefore had no reason to doubt whether, if its customer was aware of these circumstances, she would desire the Bank to make the payments. Incontrovertibly, she knew the circumstances known to the Bank and was adamant that she wished the Bank to make the payments. There is no basis on the undisputed facts of this case for asserting that (at any rate before 16 March 2018, when the Bank received a “tip-off” from the police) the Bank had material information relating to the transactions which its customer did not have.

So far in this judgment I have been considering the ordinary incidents of the contract between a bank and its customer governing the operation of a current account which are implied by law unless something different is agreed. As I have mentioned, however, a bank will invariably have standard terms of business which are incorporated into the contract. In this case it was accepted for the purpose of argument on this appeal that the contract between the Bank and Mrs Philipp incorporated the Bank’s terms and conditions for personal customers, as updated in January 2018.

These terms and conditions included the following warnings:

“If you authorise a payment and none of your details have been compromised, we have to assume it was a payment you wanted to make.

If someone persuades you to pay them money and you feel you've been cheated or you should not have agreed to pay them, we are unlikely to be able to refund the money.”

It is accepted by counsel for Mrs Philipp that these are warnings which reflect the fact that ordinarily a bank will not attract liability for APP fraud.

The Bank’s terms and conditions also contain, under the heading “When we do not have to follow your instructions,” certain express exceptions to the Bank’s duty to carry out the customer’s instructions. These include circumstances where:

“• [the instruction] isn't complete or clear, or we don't think it came from you or someone authorised to give us the instruction on your behalf

• you don't have the money (including any available overdraft) to cover the payment

• by carrying out the instruction we might break a law, regulation, code or other duty that applies to us, or it might expose us to claims from third parties

• we reasonably think that a payment into or out of an account is connected to fraud or any other criminal activity, including where the funds are being obtained through deception”

The last of these circumstances would include a situation where the Bank reasonably believes that a payment instruction given by the customer is the result of APP fraud. In accordance with this express term of the contract, the Bank therefore had a right in that event to decline to carry out the instruction. Having such a right, however, is obviously not the same as being under a duty. Pursuant to this term, the Bank was entitled to decline to carry out the instruction given by Mrs Philipp on 19 March 2018 to make a third transfer of funds to the UAE, having been told by the police on 16 March 2018 that her current account had been compromised by fraudsters in the UAE (see paras 13-14 above). Were it not for that contractual term, Mrs Philipp could potentially have sued the Bank if the result of its refusal to carry out her instruction had been to cause her loss (instead of being, as was actually the case, to save her from losing even more money).

It remains to consider a fallback argument made by the claimant that the Bank was in breach of duty after the fraud had been discovered in not taking adequate steps to recover the money which had been transferred to the UAE. Having decided the main issue in Mrs Philipp’s favour, the Court of Appeal did not think it necessary to address this point separately. But on the view I take that the Bank cannot be faulted for carrying out Mrs Philipp’s instructions to make the transfers, it becomes relevant to do so.

Mrs Philipp’s particulars of claim allege that on a date no later than 16 March 2018 the Bank should have contacted the banks in the UAE to which the funds had been sent and requested them to freeze the funds. 16 March 2018 was the date on which the Bank received the “tip-off” from the police and, in response, immediately froze Mrs Philipp’s current account (see para 13 above). It is alleged that, if the Bank had taken steps at that time to recall the payments made pursuant to the instructions given on 10 and 13 March 2018, there is at least a substantial chance that the funds transferred would have been successfully reclaimed from the UAE banks. As it was, the Bank took no steps to attempt to recall the payments until on or about 31 May 2018, by which time it was too late.

Judge Russen QC evidently regarded as untenable any suggestion that the Bank should have taken steps to try to recall the payments before 27 March 2018: see [2021] Bus LR 451, para 122. That was when Mrs Philipp first notified the Bank that she believed she was the victim of a fraud, having at last come to realise this fact following a further visit from DC Claridge on the previous day (see para 15 above). The judge was undoubtedly correct to take that view. Up until 27 March 2018 Mrs Philipp had not countermanded her instructions to make the two payments even though she had been alerted by the police and the Bank that she might have been the target of a fraud (see paras 13 and 14 above). Indeed, on 19 March 2018 she had instructed the Bank to make a third payment to the UAE of £250,000. As discussed, the Bank was entitled under its standard terms of business not to carry out that instruction. But it had no authority, let alone obligation, to attempt to reverse earlier transactions when to do so would have been directly contrary to its customer’s payment orders. It is notable that Mrs Philipp’s response to the Bank’s refusal to make the third transfer was not to reconsider her instructions relating to the earlier payments but was to try to persuade the Bank to send yet more money to one of the UAE bank accounts, including by making a false claim that she and her husband would lose a contract if the payment was not made immediately (see para 14 above).

There is no pleaded allegation that on 27 March 2018 Mrs Philipp expressly instructed the Bank to attempt to recall the two earlier payments. But it is arguable that, when she reported that she had been induced to make the payments by fraud, the Bank’s staff should have sought her instructions on this point - which would surely have been given - as it was clear that Mrs Philipp would now wish any available steps to be taken to recover the money. The fact that the Bank made attempts on and after 31 May 2018 to recall the funds which had been transferred to the UAE (see para 15 above) indicates that there were steps that could be taken to try to do so and prompts the question of why the Bank did not take those steps sooner. These are not matters that can be resolved at this stage of the proceedings on an application for summary judgment.

The likelihood that, even if prompt action had been taken by the Bank on or immediately after 27 March 2018, any of the money transferred to the UAE would have been successfully reclaimed seems slim. The judge understandably expressed strong doubts about whether there is any realistic basis for Mrs Philipp’s claim that the Bank’s delay caused her to lose a substantial chance of getting any of the money back: see [2021] Bus LR 451, paras 125-126. Nevertheless, he ultimately concluded that, despite his doubts about the Bank’s ability to recover the payments after 27 March 2018, “there are too many imponderables in this counterfactual scenario for the matter to be decided against Mrs Philipp on paper”: para 182. That was a matter of judgment on which I do not think it would be right for this court to override the view taken by the judge. Although the judge did not himself draw this inference, it seems to me logically to follow from that conclusion that the claim for the loss of a chance of recovering money from the UAE should not have been summarily dismissed.

For these reasons, I would allow the Bank’s appeal and restore the order of the judge giving summary judgment in favour of the Bank. I would, however, vary that order by limiting the judgment to the dismissal of Mrs Philipp’s claim insofar as it is based on the allegation that the Bank owed her a duty not to execute her payment instructions. I would refuse summary judgment in relation to her alternative case that the Bank was in breach of duty in not taking adequate steps to attempt to recover the money transferred to the UAE insofar as that case is based on inaction after 27 March 2018.