Moore Stephens (A Firm) v Stone & Rolls Ltd (In Liquidation)

On 22 December 2006 Stone & Rolls Limited (“the company”) commenced a claim in the Commercial Court for damages against Moore Stephens, chartered accountants (“the firm”). The firm had been the company’s auditors. The essence of the claim is that it is said that the firm negligently failed in the course of various audits to detect the dishonest behaviour of Mr Zvonko Stojevic, who at the times relevant to the claim was the sole directing mind and will of the company. The nature of his dishonesty was to procure the company to engage in frauds on banks which enabled substantial sums of money to be channelled through the company and applied elsewhere for the benefit of Mr Stojevic and others also party to the fraud. The frauds gave rise to liabilities by the company to the banks, in particular to a Czech bank. That bank sued the company and Mr Stojevic in deceit and was awarded substantial damages against both. The company could not pay and, upon the bank’s petition, went into liquidation. It is insolvent. Its claim against the firm, brought by its liquidators, is for damages to compensate it for losses it is said to have suffered in consequence of the firm’s alleged negligence in the conduct of its audits. The claim is for just under US $174 million.

The firm denies negligence and there has as yet been no trial on that question. Its position is that there is no need for one because the claim is doomed to failure and should be struck out. Its point is that the company is asking the firm to indemnify it against the liabilities it has incurred by its own fraud. It says that such a claim is barred by the principle of public policy expressed in the maxim ex turpi causa non oritur actio. In February 2007 the firm issued an application notice for summary judgment on the claim (CPR Part 24.2), or its striking out (Part 3.4(2)).

The application came before Langley J who gave his judgment on 27 July 2007. He declined to strike the claim out. This is the firm’s appeal against his order. We had excellent arguments from both sides. There was no issue as to the facts that have been found or must be assumed. As for the law, the debate came down to two questions. Mr Brindle QC, for the company, does not question the general operation of the ex turpi causa maxim but submitted (i) that it cannot prevent the company from suing for recovery in respect of its own losses caused by the individual who was its directing mind and will in relation to the frauds. That is because the company is itself said to have been a victim of the frauds and should not have any knowledge of them attributed to it; and (ii) that the maxim can anyway provide no defence to the firm when the detection of dishonesty in the operation of the company’s affairs was “the very thing” that the firm, as auditors, was retained to do. Mr Sumption QC, for the firm, submitted that both propositions are wrong.

The judge found against the company on the first issue. It is not easy to identify his conclusion in relation to the second. He ultimately decided the case on the basis that the conscience of the ordinary citizen would not find the pursuit of the claim so repugnant that it ought to be prevented “by use of the unforgiving and uncompromising operation of the ex turpi maxim”. It was agreed before us that that approach involved a misapplication of the principles relating to the maxim that the judge had earlier correctly identified, and Mr Brindle did not defend that part of his reasoning. His position was, however, that the judge had nevertheless arrived at the right result. The company’s case in respect of both issues was raised in a respondent’s notice.

Langley J summarised the facts and none of what he said about them gave rise to any difference before us. My summary is taken from his judgment. The damage suffered by the company arises from a letter of credit fraud it committed in concert with an Austrian company called BCL Trading GmbH (“BCL”). The fraud consisted in the company’s presentation to banks of false documents, the receipt of funds by the company and the payment of them away to BCL and others party to the frauds. The documents were shams purporting to reflect commercial transactions which had not occurred. The individual behind the company who procured, and benefited from, the frauds was Mr Stojevic, who owned, controlled and managed the company. The firm were the company’s auditors for the years 1996, 1997 and 1998, when the frauds were carried out. The claim is that the firm negligently failed in the course of the audits to detect and stop the frauds.

The major loser as a result of the frauds was Komercni Banka AS (“KB”). KB’s claim against the company and Mr Stojevic for damages for fraud was tried by Toulson J, whose judgment is reported as Komercni Banka AS v. Stone & Rolls Ltd and Another [2002] EWHC 2263 (Comm); [2003] 1 Lloyd’s Rep. 383. Langley J summarised Toulson J’s decision as follows (S&R is the company):

“9.

The claimant (KB) was a Czech Bank. It was the major victim of the fraud. The letters of credit were issued by the Bank at the request of an Austrian company called BCL. The Letters of Credit related to purported sales of agricultural products by S&R to BCL. S&R was the beneficiary. The claims related to 30 letters of credit which were honoured upon presentation by S&R of the relevant documents but for which KB received no payment from BCL. They were issued between July 1998 and July 1999. The first letters of credit were issued at the end of December 1997 but they were paid on maturity. KB’s case was that the documents presented to it were false in particular by representing that the issuing warehouse as S&R’s agent was holding the invoiced goods in favour of KB when in fact there were no goods at all. KB alleged that Mr Stojevic was responsible for procuring the fraudulent presentation of documents to it and that he and S&R were jointly liable to KB in deceit. The defendants denied dishonesty.

10.

Toulson J found that the documents presented to KB by S&R were false to the knowledge of both S&R and Mr Stojevic; that their presentation had caused KB to pay S&R or its assignee (where discounted by S&R) the face value of the letters of credit; and so had caused the loss to KB arising from those payments. Toulson J also held that while KB had itself been negligent that was no answer for a fraudster. He gave judgment for KB against both defendants for US $94.5 million. That was the figure paid out by KB under the 30 unreimbursed letters of credit. The evidence before Toulson J showed that of the total amount received by S&R from the fraud of approximately US $90 million, some 80 million was paid to BCL or companies connected with BCL. The evidence before this court shows that the inwards receipt of monies was promptly followed by the outward payments.

11.

The judgment (no doubt because it was not necessary) contains no analysis of whether the liability of S&R was founded on vicarious liability for the fraud of Mr Stojevic, attribution of knowledge or otherwise.”

The judgment led, on the petition of KB, to the entry of the company into provisional and then compulsory liquidation. If the company’s claim against the firm were to succeed, KB would be the major beneficiary. There is no dispute that KB has no direct claim against the firm, which owed it no duty of care.

Langley J summarised the lengthy particulars of claim. They asserted that Mr Stojevic was a Croatian national, who was a shadow director of the company for which he held a power of attorney. Mr Stojevic controlled the company, which was owned by an Isle of Man company which was in turn owned by his family trust. The particulars asserted that Mr Stojevic was “a highly intelligent and secretive Croatian businessman who controlled numerous companies in various jurisdictions and used trustees and nominee directors in order to conceal his association with them.” The company’s fraud was conveniently summarised as follows in the particulars of claim:

“16.

Mr Stojevic’s intention throughout was to use S&R as a vehicle of fraud, i.e. it was intended to be, and became, a vehicle through which funds were extracted from banks which believed that they were financing bona fide commodity trades and then paid away to third parties who were under the influence or control of Mr Stojevic. The fraudulent transactions which he planned and executed through S&R became both larger and more obviously fraudulent as he realised that Moore Stephens had failed to detect his earlier frauds and would probably not detect his frauds in the future.”

It is the essence of the company’s claim that Mr Stojevic was its controlling mind and will. Nobody else was in a like position. In a real sense the company was his company. It was, for practical purposes, a “one man company”. It is a further part of the claim that the company was throughout used by Mr Stojevic as a vehicle for fraud, by extracting money from KB so that it could then be paid away to the fraudsters. The application before Langley J proceeded on the assumed basis that, had the firm discharged the duty of care it owed to the company in the performance of its audits, it would have detected the frauds and brought them to an end for the future.

This, Mr Brindle submitted to us, was “the very thing” (or at least one of the very things) that the firm, as the company’s auditors, should have been astute to expose; and there can be no rule of public policy that ought to afford the firm a defence to such a claim. He referred us to Caparo Industries PLC v. Dickman and Others [1990] 2 AC 605, and in particular to the speech of Lord Oliver of Aylmerton, at 630, where he explained the purpose of the statutory requirement for an annual audit of a company’s accounts. In that passage, Lord Oliver referred to the auditor’s function:

“… to ensure, so far as possible, that the financial information as to the company’s affairs prepared by the directors accurately reflects the company’s position in order, first, to protect the company itself from the consequences of undetected errors or, possibly, wrongdoing (by, for instance, declaring dividends out of capital) ….”

In what follows I will (i) explain the ex turpi causa principle, as to which (subject to (iii) below) there is no dispute; (ii) consider whether or not (as the firm contends but the company denies) knowledge of the fraud should be attributed to the company, as to which there is a dispute; and (iii) consider the “very thing” argument, as to which there is also a dispute. If the appeal is to succeed, the firm must succeed on both (ii) and (iii).

The principle of public policy expressed in this maxim goes back a long way. Lord Mansfield CJ explained it in Holman v. Johnson (1775) 1 Cowp. 341, at 343:

“No court will lend its aid to a man who founds his cause of action on an immoral or an illegal act. If, from the plaintiff’s own stating or otherwise, the cause of action appears to arise ex turpi causa … there the court says that he has no right to be assisted. It is on this ground the court goes; not for the sake of the defendant, but because they will not lend their aid to such a plaintiff.”

Although there is no dispute as to how the principle works, it is necessary to provide an explanation of it – and also to delve into a little bit of its history -- in part to explain the basis on which the judge appears ultimately to have decided the application and in part to explain the main authority upon which Mr Brindle relied in support of his “very thing” submission.

In this court’s decision in Tinsley v. Milligan [1992] Ch. 310, the majority (Nicholls and Lloyd LJJ) held that the application or otherwise of the ex turpi causa principle in any case in which a claim is tainted with illegality or immorality required a flexible approach. By reference to recent authorities, Nicholls LJ explained that the applicable test was a so-called “public conscience” test, namely whether it would be an affront to public conscience to accede to the claim despite the fact that it was so tainted. This required the court to “weigh, or balance, the adverse consequences of granting relief against the adverse consequences of refusing relief. The ultimate decision calls for a value judgment” (Nicholls LJ, at 319H). Lloyd LJ, whilst agreeing with Nicholls LJ in the result, decided the appeal on different grounds but (at 339C to G) also subscribed to the view that the “public conscience” test was part of the civil law relating to illegality.

Tinsley went to the House of Lords ([1994] AC 340), which, by a majority, upheld this court’s decision on different grounds. The House unanimously rejected the “public conscience” test, which was no more than the assumption of a judicial discretion as to whether or not in any particular case, despite the illegality or immorality, to grant or refuse relief. Although that test found support in the recent authorities upon which Nicholls LJ had drawn, they were out of line with prior binding authority pointing the other way. The essence of the decision of the House in Tinsley was (i) (with which all members of the Committee agreed) that the “public conscience” test was abolished; and (ii) (with which the majority agreed) that the correct test was as summarised by Lord Browne-Wilkinson, at 376E:

“In my judgment the time has come to decide clearly that the rule is the same whether a plaintiff founds himself on a legal or equitable title: he is entitled to recover if he is not forced to plead or rely on the illegality, even if it emerges that the title on which he relied was acquired in the course of carrying through an illegal transaction.”

That statement of principle was expressed in the context of the facts of Tinsley, a property dispute. But it is one I regard as applying generally and which Langley J conveniently described as a “reliance” test. The relevant question it identifies is whether, to advance the claim, it is necessary for the claimant to plead or rely on the illegality. If it is, Tinsley decided that the axe falls indiscriminately and the claim is barred, however good it might otherwise be. There is no discretion to permit it to succeed. The absence of any such discretion emerges from all their Lordships’ speeches. Lord Goff of Chieveley, who was in the minority with Lord Keith of Kinkel, gave the leading speech for the rejection of the “public conscience” test, with which the majority agreed. The essential difference between the minority and the majority views was whether the touchstone for the application of the ex turpi causa maxim was the reliance test favoured by the majority or the wider test favoured by the minority and regarded as applicable to the particular facts before the court. But once the maxim is engaged, it applies indiscriminately. After referring to Holman v. Johnson and to the subsequent application of Lord Mansfield’s principle, Lord Goff said ([1994] AC 340, at 355B):

“It is important to observe that, as Lord Mansfield made clear, the principle is not a principle of justice; it is a principle of policy, whose application is indiscriminate and so can lead to unfair consequences as between the parties to litigation. Moreover the principle allows no room for the exercise of any discretion by the court in favour of one party or the other.”

The reliance test has since been applied. In Clunis v. Camden and Islington Health Authority [1998] QB 978, this court dealt with a case in which the plaintiff had killed a man in an unprovoked attack, was convicted of manslaughter and detained in a secure hospital. He had a history of mental disorder and sued the health authority because of the alleged negligent failure of a hospital in which he had earlier been detained to treat him with reasonable professional care and skill. His case was that, had it done so, he would not have committed manslaughter and suffered the prolonged detention he faced. This court struck the claim out as barred by the ex turpi causa principle. That principle applied to all causes of action, including claims in tort, “[b]ut whether a claim brought is founded in contract or in tort, public policy only requires the court to deny its assistance to a plaintiff seeking to enforce a cause of action if he was implicated in the illegality and in putting forward his case he seeks to rely upon the illegal acts” (Beldam LJ, delivering the judgment of the court, at [1998] QB 978, at 987C, emphasis added). The plaintiff’s counsel sought to invoke the “public conscience” test in support of the proposition that, in the circumstances of the case, the court ought not to regard the ex turpi causa maxim as applicable. After referring to the rejection of that test by the House of Lords in Tinsley, the court declined to accept the submission.

Cross v. Kirby, Court of Appeal, 18 February 2000, unreported, can perhaps be read as broadening the reliance test. The claimant had trespassed on the defendant’s land and assaulted the defendant with a baseball bat, which the defendant wrested from him and struck him with it, causing serious injury. This court allowed the defendant’s appeal against the judgment against him. Beldam LJ, with whom Otton LJ agreed, found, as an alternative ground for doing so, that the plaintiff was barred from bringing his claim because it arose from his own criminal and unlawful acts. The court referred to Tinsley, but Beldam LJ did not accept that strict satisfaction of any “reliance” test was necessary before the ex turpi causa principle could be invoked, saying:

“I do not believe that there is any general principle that the claimant must either plead, give evidence of or rely on his own illegality for the principle to apply. Such a technical approach is entirely absent from Lord Mansfield’s exposition of the principle. I would, however, accept that for the principle to operate the claim made by the claimant must arise out of criminal or illegal conduct on his part. In this context ‘arise out of’ clearly denotes a causal connection with the conduct.”

Judge LJ expressed as follows the test in relation to the circumstances before the court:

“In my judgment, where the claimant is behaving unlawfully, or criminally, on the occasion when his cause of action in tort arises, his claim is not liable to be defeated ex turpi causa unless it is also established that the facts which give rise to it are inextricably linked with his criminal conduct. I have deliberately expressed myself in language which goes well beyond questions of causation in the general sense.”

The principles were further considered by this court in Hewison v. Meridian Shipping Pte [2002] EWCA Civ 1821; [2003] PIQR 252. That was a claim for damages brought by a crane operator for negligence by his employer causing him personal injury, the damages claimed including loss of future earnings he would otherwise have earned in his employment. Liability was admitted but the employer’s case was that to award damages for loss of future earnings would infringe the ex turpi causa principle. This was because the claimant’s continued employment would have required him to continue fraudulently to deceive his employers that he did not suffer from epilepsy. Morland J agreed, refused to award such damages and this court, by a majority, also agreed. Clarke LJ, after referring to Clunis, said:

“29.

To my mind the authorities support that approach. They seem to me to support the proposition that where a claimant has to rely upon his or her own unlawful act in order to establish the whole or part of his or her claim the claim will fail either wholly or in part. …”

Tuckey LJ, also in the majority, said much the same:

“51.

For this reason I favour a broad test of the kind proposed by Clarke LJ, namely: is the claim or the relevant part of it based substantially (and not therefore collaterally or insignificantly) on an unlawful act? …”

It is unnecessary to consider further the precise limits of the circumstances in which the ex turpi causa maxim will bar a claim. There is no dispute that, in the present case, the company’s claim relies upon, is based substantially on, arises out of and is inextricably linked with the fraud that was perpetrated on the banks. That fraud was actually perpetrated by Mr Stojevic, the company’s sole directing mind and will.

The next logical question is whether Mr Stojevic’s fraud can properly be attributed to the company, which was the vehicle for the fraud. The general principles by reference to which the acts of individuals carried out for or in relation to a company’s affairs can or will be attributed to the company itself are not in dispute. Mr Brindle’s position is, however, that those principles preclude any such attribution in the present case because the company was itself a victim of Mr Stojevic’s fraud. He accepts, however, that if he is wrong on that, the ex turpi causa defence is in principle available to the firm, but that in the particular circumstances of this case its application is trumped by his “very thing” submission.

There is no dispute that a company may not merely be vicariously liable for the acts of its agents, but can be directly liable for a dishonest act by reason of the attribution to it of its dishonest directing mind and will. There is also some degree of agreement that there will or may not be such an attribution to the company where the dishonesty of the company’s agents is directed against the company itself. The firm’s position is, however, that that is not this case. Mr Stojevic was not engaged in a fraud on, or directed at, the company. He was using it as the vehicle for a fraud of its own directed at the banks. It was the company which tendered the fraudulent documents to the banks and it was the company that was dealing dishonestly with them. It was the banks that were the victims of the fraud, not the company.

There was a time in the development of English law when the belief was held by some that a company, a persona ficta, could have no dishonest motive attributed to it. It could at most, in certain cases, be vicariously liable for the dishonest wrongs committed by its human agents. In Abrath v. The North Eastern Railway Company (1886) 11 App. Cas. 247, in which a claim for malicious prosecution was brought against a company, the jury acquitted the company of the necessary malice and this court upheld the judge’s direction as having been correct. The House of Lords dismissed the appeal but, at 251, Lord Bramwell took the opportunity to advance the trenchant view – on a point not argued before the House -- that the claim was anyway bound to fail “even if you assume the strongest case, namely, that of the very shareholders directing it or the very directors ordering it, because it is impossible that a company can have malice or motive.” Lord Fitzgerald, at 255, had no doubt that Lord Bramwell’s weighty observations would be instructive in the future and would “always carry with them that force before any tribunal which they so eminently deserve.” The Earl of Selborne, at 256, said much the same, but he would have wanted the point argued before expressing his own opinion on it.

In the event Lord Bramwell’s view did not carry the prophesied force. Citizens’ Life Assurance Company, Limited v. Brown [1904] AC 423, a decision of the Privy Council, involved a claim for malicious libel against a company. The libel was published, in the course of his employment, by an officer of the company against whom there was evidence of express malice. It was argued before the Privy Council, in reliance on Lord Bramwell’s speech in Abrath, that the officer’s malice could not be attributed to the company. Lord Lindley, giving the advice of the Privy Council, referred to Lord Bramwell’s speech, and to one to like effect by Lord Cranworth in Addie v. Western Bank of Scotland (1867) LR 1 HL, Sc., 145, and said, at 426:

“But these opinions have not prevailed, and their Lordships are not prepared to give effect to them. If it is once granted that corporations are for civil purposes to be regarded as persons, i.e., as principals acting by agents and servants, it is difficult to see why the ordinary doctrines of agency and of master and servant are not to be applied to corporations as well as to ordinary individuals. The doctrines have been so applied in a great variety of cases, in questions arising out of contract, and in questions arising out of torts and frauds; and to apply them to one class of libels and to deny their application to another class of libels on the ground that malice cannot be imputed to a body corporate appears to their Lordships to be contrary to sound legal principles.”

The Citizens’ Life case was not in fact one in which the company had the relevant malice directly attributed to it, it was one in which the case made and proved was that it was vicariously liable for the acts of its officer. But the importance of Lord Lindley’s statement is that it rejected Lord Bramwell’s sweeping assertion that a company could not have attributed to it the malice or motive influencing the mind of its human agents.

Over a hundred years on, it is clear that, in an appropriate context, the mind of such agents can and will be attributed to the company itself. Lord Hoffmann’s judgment in the Privy Council decision of Meridian Global Funds Management Asia Ltd. v Securities Commission [1995] 2 AC 500 provides an illuminating exposition of the principles, and I will devote a little space to it since its guidance assists in understanding some of the cases to which I will later have to refer. Lord Hoffmann explained that a company’s primary rules of attribution will be found in its constitution (for example, a provision that a decision of its board shall be a decision of the company) and the acts of its duly appointed agents and servants will, by a combination of the general principles of agency and the company’s primary rules of attribution, also count as the acts of the company. “It is therefore a necessary part of corporate personality that there should be rules by which acts are attributed to the company” (506C). Lord Hoffmann continued, at 507:

“There is in fact no such thing as the company as such, no ding an sich, only the applicable rules. To say that a company cannot do something means only that there is no one whose doing of that act would, under the applicable rules of attribution, count as an act of the company.

The company’s primary rules of attribution together with the general principles of agency, vicarious liability and so forth are usually sufficient to enable one to determine its rights and obligations. In exceptional cases, however, they will not provide an answer. This will be the case when a rule of law, either expressly or by implication, excludes attribution on the basis of the general principles of agency or vicarious liability. For example, a rule may be stated in language primarily applicable to a natural person and require some act or state of mind on the part of that person ‘himself,’ as opposed to his servants or agents. This is generally true of rules of the criminal law, which ordinarily impose liability only for the actus reus and mens rea of the defendant himself. How is such a rule to be applied to a company?

One possibility is that the court may come to the conclusion that the rule was not intended to apply to companies at all; for example, a law which created an offence for which the only penalty was community service. Another possibility is that the court might interpret the law as meaning that it could apply to a company only on the basis of its primary rules of attribution, i.e. if the act giving rise to liability was specifically authorised by a resolution of the board or an unanimous agreement of the shareholders. But there will be many cases in which neither of these solutions is satisfactory; in which the court considers that the law was intended to apply to companies and that, although it excludes ordinary vicarious liability, insistence on the primary rules of attribution would in practice defeat that intention. In such a case, the court must fashion a special rule of attribution for the particular substantive rule. This is always a matter of interpretation: given that it was intended to apply to a company, how was it intended to apply? Whose act (or knowledge, or state of mind) was for this purpose intended to count as the act etc. of the company? One finds the answer to this question by applying the usual canons of interpretation, taking into account the language of the rule (if it is a statute) and its content and policy”.

The present case, Mr Sumption submitted, is not one in which the company could merely be said to be vicariously liable for Mr Stojevic’s actions in relation to the frauds. It is one in which the company was itself their perpetrator. The essential difference between the two concepts lies in the degree to which the relevant agent is identified with the company. If a junior employee commits a fraud in the course of his employment, the company will be vicariously liable for the fraud, but the fraud will not automatically be imputed to the company itself (although it may be if the context requires it). But if the board of directors resolves on the commission of a fraud by the company, their dishonesty will be attributed to it, and it will itself be directly, and not just vicariously, liable for the fraud. That, said Mr Sumption, is this case, in which Mr Stojevic’s state of mind was that of someone who for the relevant purpose was to be regarded as the company. He was not just its agent. He was someone to whom the relevant part of the running of the company’s affairs had been delegated. He had a power of attorney to run it at his discretion. His acts on behalf of the company were the company’s acts. He was the company’s directing mind and will.

Attribution on such a basis is illustrated by Lennard’s Carrying Company, Limited v. Asiatic Petroleum Company, Limited [1915] AC 705. A cargo on board ship was lost by a fire caused by the ship’s unseaworthiness. The shipowner was a company that was managed by another company, of which the managing director was designated as the person to whom the ship’s management was entrusted. He was, as Viscount Haldane, Lord Chancellor, said at 712 “the active spirit” in the managing company. The question was whether the shipowner could be relieved of the liability levelled against it. That depended on whether, as it claimed, what had happened had happened without “his actual fault or privity”, those being the exempting words in section 502 of the Merchant Shipping Act 1894, the “his” being a reference to the shipowner. The shipowner was held liable. Viscount Haldane said, at 713:

“My Lords, a corporation is an abstraction. It has no mind of its own any more than it has a body of its own; its active and directing will must consequently be sought in the person of somebody who for some purposes may be called an agent, but who is really the directing mind and will of the corporation, the very ego and centre of the personality of the corporation. That person may be under the direction of the shareholders in general meeting; that person may be the board of directors itself, or it may be, and in some companies it is so, that that person has an authority co-ordinate with the board of directors given to him under the articles of association, and is appointed by the general meeting of the company, and can only be removed by the general meeting of the company. My Lords, whatever is not known about Mr Lennard’s position, this is known for certain, Mr Lennard took the active part in the management of this ship on behalf of the owners, and Mr Lennard, as I have said, was registered as the person designated for this purpose in the ship’s register. Mr Lennard therefore was the natural person to come on behalf of the owners and give full evidence not only about the events of which I have spoken, and which related to the seaworthiness of the ship, but about his own position and as to whether or not he was the life and soul of the company. For if Mr Lennard was the directing mind of the company, then his action must, unless a corporation is not to be liable at all, have been an action which was the action of the company itself within the meaning of s.502. It has not been contended at the Bar, and it could not have been successfully contended, that s.502 is so worded as to exempt a corporation altogether which happens to be the owner of a ship, merely because it happens to be a corporation. It must be upon the true construction of that section in such a case as the present one that the fault or privity of somebody who is not merely a servant or agent for whom the company is liable upon the footing respondeat superior, but somebody for whom the company is liable because his action is the very action of the company itself. It is not enough that the fault should be the fault of a servant in order to exonerate the owner, the fault must also be one which is not the fault of the owner, or a fault to which the owner is privy; and I take the view that when anybody sets up that section to excuse himself from the normal consequences of the maxim respondeat superior, the burden lies on him to do so.”

That is, for present purposes, an instructive case. The shipowner was liable because its managing agent, who had the relevant knowledge, was its directing mind and will for the relevant purpose. It could only escape liability under the statute if his knowledge could not properly be attributed to it but it was held that it should be so attributed. The concept of the “directing mind and will” was considered by this court in El Ajou v. Dollar Land Holdings plc and another [1994] 2 All ER 685, in which, at 695j, Nourse LJ said the doctrine “attributes to the company the mind and will of the natural person or persons who manage and control its actions”, explaining further, at 696a, that “[i]t is necessary to identify the natural person or persons having management and control in relation to the act or omission in point.” Hoffmann LJ, at 705f, after referring to Lennard’s case, said that “A person held out by the company as having plenary authority or in whose exercise of such authority the company acquiesces, may be treated as its directing mind.” Later, at 706d, in line with Nourse LJ, he said that “[t]he authorities show clearly that different persons may for different purposes satisfy the requirements of being the company’s directing mind and will.”

In Lennard’s case the question of the attribution to the company of its directing mind and will arose in the context of the company’s claim to enjoy the exemption from liability afforded by a statutory provision. In El Ajou the question was whether the knowledge of a particular director could be attributed to the company so as to fix it with the liability of a constructive trustee for knowing receipt. Questions of attribution will also be relevant in bids to fix a company with criminal liability for an offence involving the requirement of mens rea. Where there is such a requirement, “a corporation can be guilty of it if the offence is committed in the course of the corporation’s business by a person in control of its affairs to such a degree that it may fairly be said to think and act through him so that his actions and intent are the actions and intent of the corporation; in such a case his acts and state of mind are regarded in law as the acts and state of mind of the company” (Halsbury’s Laws of England, 4^th Edition, Volume 11(1), paragraph 38). The key question was identified by Lord Reid in Tesco Supermarkets Ltd v. Nattrass [1972] AC 153, in two passages at 170E to G and 171F to H respectively:

“I must start by considering the nature of the personality which by a fiction the law attributes to a corporation. A living person has a mind which can have knowledge or intention or be negligent and he has hands to carry out his intentions. A corporation has none of these: it must act through living persons, though not always one or the same person. Then the person who acts is not speaking or acting for the company. He is acting as the company and his mind which directs his acts is the mind of the company. There is no question of the company being vicariously liable. He is not acting as a servant, representative, agent or delegate. He is an embodiment of the company or, one could say, he hears and speaks through the persona of the company, within its appropriate sphere, and his mind is the mind of the company. If it is a guilty mind then that guilt is the guilt of the company. It must be a question of law whether, once the facts have been ascertained, a person in doing particular things is to be regarded as the company or merely as the company’s servant or agent. In that case any liability of the company can only be a statutory or vicarious liability. …

… Normally the board of directors, the managing director and perhaps other superior officers of a company carry out the functions of management and speak and act as the company. Their subordinates do not. They carry out orders from above and it can make no difference that they are given some measure of discretion. But the board of directors may delegate some part of their functions of management giving to their delegate full discretion to act independently of instructions from them. I see no difficulty in holding that they have thereby put such a delegate in their place so that within the scope of the delegation he can act as the company. It may not always be easy to draw the line but there are cases in which the line must be drawn. Lennard’s case [1915] AC 705 was one of them.”

Applying those principles to the facts of this case, Mr Sumption submitted that a criminal prosecution for fraud (or whatever the appropriate charge for the particular criminal acts the company committed) could have been brought against the company. Mr Stojevic’s acts were the acts not just of a servant or agent of the company. He was acting as its very embodiment. His dishonest intent would, for the purposes of such a prosecution, be regarded as the dishonest intent of the company itself. The reason the company could be so prosecuted is because the law recognises the concept of the guilty corporate mind. It recognises that a company can be made directly responsible, civilly and criminally, in consequence of the attribution to it of the dishonest actions of those who, in the context of the particular matter in question, can fairly be regarded as its directing mind and will. Mr Sumption said there could be no clearer example of that than the present case, in which Mr Stojevic owned the company and controlled it in its every relevant act. There was no one else.

It follows, submitted Mr Sumption, that a company that brings a civil claim in reliance upon the dishonest or illegal actions of its own in circumstances in which a guilty knowledge of such actions is properly attributed to it can and will be met by the plea of ex turpi causa, which will afford a complete defence to the claim. For the purposes of that principle, a company can be as much a wrongdoer as can an individual.

Mr Brindle did not advance any submissions in support of a reasoned disagreement with these propositions, although he did not concede that the company could have been held liable to others for the fraud on the basis that it was treated as having itself committed the fraud as opposed to being liable vicariously. Toulson J had not made a specific finding either way as to that in his judgment in KB’s claim and Mr Brindle did not accept that the authorities cited by Mr Sumption made good the point that the company could be personally liable for the fraud. But Mr Brindle nevertheless accepted that, for the purposes of the present debate, the company could be treated as having itself committed the fraud. His submission was that, in the circumstances of this case, the attribution of Mr Stojevic’s fraud to the company is anyway excluded by what was referred to in the argument as the Hampshire Land principle, just as, he said, it would be so excluded if the company’s liability to others were merely a vicarious liability. I turn to that principle, which is in the nature of a principle of non-attribution.

The first main difference between counsel arises in relation to the application or otherwise of this principle, which derives from the decision of Vaughan Williams J in In re Hampshire Land Company [1896] 2 Ch 743. Under the articles of Hampshire Land Company, Limited (“Hampshire”), its directors could not borrow more than the amount of the paid-up capital unless the larger borrowing had been authorised by a general meeting. Hampshire was closely connected with Portsea Island Building Society (“Portsea”), the companies having four directors in common and Mr Wills being the secretary of both. At a general meeting of Hampshire a resolution was passed authorising the borrowing of £30,000, which exceeded the paid-up capital by about £20,000. The money was then borrowed from Portsea. The notice convening the meeting had irregularly given no notice to the Hampshire shareholders of the intention to propose the resolution, although under Hampshire’s articles it should have done. Hampshire and Portsea went into liquidation and Portsea sought to prove in Hampshire’s liquidation for over £30,000, the money lent. There was no evidence that Wills had informed Portsea of the irregularity in the calling of the Hampshire general meeting.

Vaughan Williams J commenced his judgment by identifying the question as being which set of shareholders – both sets being innocent -- ought to bear the loss resulting from the irregular resolution. The Hampshire directors had no authority to borrow in the absence of a properly passed resolution but the internal management rule explained in Royal British Bank v. Turquand 6 E & B 327 entitled Portsea to assume that all the essentials of Hampshire’s internal management had been carried out. Portsea was therefore entitled to prove for the full amount of its loan unless knowledge of the irregularities could properly be attributed to it. Should such knowledge be so attributed?

The argument that it should turned on Mr Wills’s dual role as a common officer of the two companies. The judge accepted that he must be taken to have been aware of the irregularities. It was submitted that his knowledge obtained as an officer of Hampshire must therefore be attributed to Portsea, of which he was also an officer. It was, however, also accepted on both sides that there was no general proposition to the effect that the knowledge of an officer common to two companies is always the knowledge of both companies. The judge’s view, expressed at 748, was that “the knowledge which has been acquired by the officer of one company will not be imputed to the other company, unless the common officer had some duty imposed upon him to communicate that knowledge to the other company, and had some duty imposed on him by the company which is alleged to be affected by the notice to receive the notice.” The gravamen of his decision is at 749, where he said:

“The case is very much more like the one which [counsel for Hampshire’s liquidator] had to admit was an exception to the general rule that they sought to lay down, for they admitted that if Wills had been guilty of a fraud, the personal knowledge of Wills of the fraud that he had committed upon the company would not have been knowledge of [Portsea] of the facts constituting that fraud; because common sense at once leads one to the conclusion that it would be impossible to infer that the duty, either of giving or receiving notice, will be fulfilled where the common agent is himself guilty of fraud. It seems to me that if you assume here that Mr Wills was guilty of irregularity – a breach of duty in respect of these transactions – the same inference is to be drawn as if he had been guilty of fraud. I do not know, I am sure, whether he was guilty of actual fraud; but whether his conduct amounted to fraud or to breach of duty, I decline to hold that his knowledge of his own fraud or of his own breach of duty is, under the circumstances, the knowledge of the company [which I take to mean Portsea]. I must, therefore, admit the proof.”

The essence of the principle is therefore that a company will not have attributed to it knowledge of a fraud when that fraud is being practised on the company itself. The law does not attribute knowledge of a deception to the person who is being deceived. Why should it make such an attribution when its agent engaged on deceiving it would not himself disclose his perfidy?

The Hampshire Land principle was approved and applied by the House of Lords in J.C. Houghton and Company v. Nothard, Lowe and Wills, Limited [1928] AC 1. I refer, without citation, to the speeches of Viscount Dunedin at 14/15, Viscount Sumner at 19 and Lord Carson at 33. Lord Atkinson agreed with Viscount Dunedin and Lord Shaw of Dunfermline dissented. A more recent example of its application, from the latter half of the last century, is Belmont Finance Corporation Ltd v. Williams Furniture Ltd and Others [1979] Ch 250. There the plaintiff company’s claim was that it was the victim of a dishonest conspiracy involving its acquisition of the shares in another company for £500,000 (which it said was over eight times their worth), under an arrangement under which the plaintiff was itself then to be sold to one of the defendants for £489,000. The claim included the assertion that the transaction involved the giving by the plaintiff of illegal financial assistance for the purchase of its own shares contrary to section 54 of the Companies Act 1948. The plaintiff’s resolution to effect the conspiratorial purchase had been passed at a board meeting attended by three directors, two of whom were defendants who were alleged to be party to the conspiracy. The third director was not alleged to be involved in the conspiracy.

Foster J dismissed the claim at the close of the plaintiff’s case on the basis that it was itself a party to the conspiracy upon which it sued and so was disabled from suing. The only basis on which the plaintiff could be regarded as party to the conspiracy was if the guilty knowledge of the two defendant directors was attributed to it. This court held that their knowledge could not be so attributed, so that the plaintiff was not precluded from bringing its claim. Buckley LJ applied the Hampshire Land principle, albeit without reference to the case itself, which was not cited. He said, at 261F to 262B:

“On the footing that the directors of the plaintiff company who were present at the board meeting on October 11, 1963, knew that the sale of the Maximum shares was at an inflated value, and that such value was inflated for the purpose of enabling the third, fourth, fifth and sixth defendants to buy the share capital of the plaintiff company, those directors must be taken to have known that the transaction was illegal under section 54.

It may emerge at a trial that the facts are not as alleged in the statement of claim, but if the allegations in the statement of claim are made good, the directors of the plaintiff company must then have known that the transaction was an illegal transaction.

But in my view such knowledge should not be imputed to the company, for the essence of the arrangement was to deprive the company of a large part of its assets. As I have said, the company was a victim of the conspiracy. I think it would be irrational to treat the directors, who were allegedly parties to the conspiracy, notionally as having transmitted this knowledge to the company; and indeed it is a well-recognised exception from the general rule that a principal is affected by notice received by his agent that, if the agent is acting in fraud of his principal and the matter of which he has notice is relevant to the fraud, that knowledge is not to be imputed to the principal.

So in my opinion the plaintiff company should not be regarded as a party to the conspiracy, on the ground of lack of the necessary guilty knowledge.”

A case upon which Mr Sumption and Mr Brindle both placed reliance is Attorney-General’s Reference (No 2 of 1982) [1984] 1 QB 624. Two defendants were charged, either individually or jointly, with the theft of money from companies they wholly owned and controlled. At the conclusion of the prosecution case, the judge withdrew the case from the jury and directed the acquittal of the defendants. That was on the ground that they “were” the company and so could not steal from it. The Attorney-General referred the case to the court for its opinion as to whether a person in sole control of a company was, or two such persons acting in concert were, capable of stealing the company’s property. The judgment of the court (Watkins and Kerr L.JJ) was delivered by Kerr LJ. The question before the court arose in relation to the definition of theft in the Theft Act 1968, the relevant provision being part of the definition of “dishonestly” in section 2(1)(b), which provides:

“2(1) A person’s appropriation of property belonging to another is not to be regarded as dishonest –

…

(b)

if he appropriates the property in the belief that he would have the other’s consent if the other knew of the appropriation and the circumstances of it; …”

Subject to that issue of dishonesty, the defendants conceded that the other ingredients of theft were prima facie satisfied. Their case was that they were the sole directing mind and will of the company, which was therefore bound to consent to that which they consented: consent by the defendants to the appropriations necessarily involved consent by the company. Turning to the authorities, the court referred to Tesco Supermarkets Ltd as authority for the proposition that, in a case in which two individuals such as the defendants “are” the company, any offences committed by them in relation to the affairs of the company would be capable of being treated as offences committed by the company itself. But that case had no bearing on offences committed against the company. A case in which the courts considered the position when the controlling individuals acted illegally against the company was Belmont Finance. The court cited the passage from Buckley LJ’s judgment that I have cited, held that it established a principle directly contrary to the defendants’ argument and that there was no reason why it should not apply equally in the criminal law: in short, the court held that the dishonest intentions directed against the company by those in control of it will not be attributed to the company. The court went on to hold that, in the circumstances, section 2(1)(b) could afford no defence to the defendants. The court said, at 642C:

“The essence of the defendants’ argument is the alleged identity, in all respects, and for every purpose, between the defendants and the company. It is said, in effect, that their acts are necessarily the company’s acts; that their will, knowledge, and belief are those of the company, and that their consent necessarily implies consent by the company. But how then can the company be regarded as ‘the other’ for the purposes of this provision? One merely has to read its wording to see that it cannot be given any sensible meaning in a context such as the present, where the mind and will of the defendants are also treated in law as the mind and will of ‘the other.’ It is for this reason that in such cases there can be no conspiracy between the directors and shareholders on the one hand and the company on the other: Reg v. McDonnell [[1966] 1 QB 233.”

Mr Sumption said that there was a tension between the holding that the Belmont Finance principle applied and the holding that the company was unable to consent to the defendants’ acts because the company had no mind separate from those of the defendants. The tension lay, I understand, in the fact that, having held that the defendants’ intention to steal from the company could not, because of the Belmont Finance principle, be attributed to the company, the court could not reach a point at which the question could arise as to whether the company was capable of consenting to the defendants’ acts: ex hypothesi it knew nothing of them. But Mr Sumption submitted that the passage just quoted was part of the ratio of the court’s decision and that it was correct in principle. Its essence was that there was no one in the company capable of being deceived by defendants’ acts, just as in the present case there was no one in the company capable of being deceived by Mr Stojevic’s acts.

Mr Sumption relied on The Attorney-General’s Reference case in support of an argument that the Hampshire Land principle cannot apply to a one-man company such as the present claimant -- or to a two-man company such as that in the Attorney-General’s Reference case. The basis of the argument was that that principle can only apply in cases in which there is at least one other human being in the company to whom the relevant dishonest knowledge might be capable of being communicated. He said that, given that the rationale of the Hampshire Land principle is that the dishonest agent is not to be treated as implicitly transmitting to the company that which he would anyway not have disclosed to it -- because it was the target of his dishonesty -- it follows that there can be no scope for the operation of the principle save in circumstances in which there is some other human agent engaged in the affairs of the company whom he is intending to deceive. The submission was that the principles of justice and common sense that underlie the Hampshire Land principle cannot be regarded as going the distance of requiring the fiction of separate legal personality to be carried to the extreme point at which the company is either mindless or else does not know what is in the only human mind that it in fact has. It can, therefore, have no application to a one-man company such as the company in the present case, in which Mr Stojevic was its sole directing mind and will.

Whilst I follow the logic of this submission, I do not accept it. The court in the Attorney-General’s Reference case held it to be unsound. It held in terms that the Hampshire Land (or Belmont Finance) principle can apply to the case in which the fraud or dishonesty of a company’s sole (or, in that case, joint) directing mind and will is targeted against the company itself; and in principle that appears to me to be right. As for the suggested tension in the court’s reasoning, Mr Brindle’s submission was that there was none and that the decision simply had two ratios. Whether or not that is correct, I would anyway accept and follow what the court said about the application of the Hampshire Land principle. In the circumstances I will therefore do no more than list, without discussion, the further authorities that Mr Sumption invoked in support of his argument in relation to one-man companies. They were Regina v. McDonnell [1966] 1 QB 233; Royal Brunei Airlines Sdn. Bhd. v. Philip Tan Kok Ming [1995] 2 AC 378, at 393; and Berg, Sons & Co. Ltd v. Mervyn Hampton Adams and Others [2002] 1 Lloyd’s Law Reports Professional Negligence 41, at 54.

The real issue between counsel is whether the Hampshire Land principle applies to the facts of this case. I now turn to that.

The critical question is this: should the court regard the company as villain or victim? If it is to be regarded as the victim of Mr Stojevic’s dishonesty, the effect of the Hampshire Land principle is that his dishonesty will not be attributed to it, it will not be tainted with it and, in principle, it will be entitled to bring its claim. If, however, the correct analysis is that it was itself carrying out the frauds and that its claim to be a victim is (as it is) based on no more than the liabilities it incurred as an inevitable consequence of those frauds – and it is not usually possible to defraud others without incurring liabilities to them – Mr Sumption’s submission was that, for the purposes of the Hampshire Land principle, the company must be regarded as the villain. If it is a victim at all, it is only one in a subsidiary and immaterial sense.

Mr Sumption developed this by saying that it is quite unreal to regard the company as a victim. It was engaged, in concert with BCL, in dishonestly extracting money from KB for non-existent goods that BCL was pretending to buy from it. The scheme was one under which BCL procured the opening of the letters of credit with KB in favour of the company, paying a modest proportion of the price as a security deposit. The company joined in the pretence of a sale of the non-existent goods to BCL, tendering to KB false documents purporting to be those required by the letter of credit, including a purported warehouse warrant made out to the order of KB. Those documents represented KB’s main security, which it would hold until BCL paid it in full. The payments from KB went into the company’s account, whence they were promptly paid out to BCL. BCL used it to pay off the debt to KB on the letter of credit, the object of that being to establish a credit record and so enable the obtaining of further and bigger letters of credit and so build up the purported scale of commodity dealing month by month. For most of the time the fraud was being committed, the KB money just went round in circles. Ultimately, when the prize was big enough, BCL made off with the money and KB was left with its worthless security. The warehouse warrants were forgeries and there were no goods in the warehouse. This scenario was one in which the intended, and actual, victim of the fraud was KB. It was not the company. Whilst the company incurred a liability to KB, that was a necessary part of its fraud; and whilst it paid out to BCL the payments of which it had defrauded KB, that was simply part of the same scam. A villain such as the company, which defrauds a bank such as KB of millions of dollars and then makes off with the spoils, cannot claim to be a victim for the purpose of the Hampshire Land principle on the ground that it has, in carrying out part of its own deliberate scheme, left itself unable to repay its victim.

That approach to the limits of the Hampshire Land principle is said by Mr Sumption to be supported by two decisions. McNicholas Construction Co Ltd v. Customs and Excise Commissioners [2000] STC 553 is a decision of Dyson J. In that case the Commissioners suspected that bogus sub-contractors had issued VAT invoices to McNicholas, an engineering contractor, for services that had not been provided. The company had paid VAT to the sub-contractors under their invoices and had claimed by way of VAT relief as input tax the VAT so paid. The effect of the suspect transactions on the company was therefore neutral; but if the invoices proved to be bogus, the company would not have been obliged to pay the VAT to the sub-contractors, nor entitled to claim the relief it did in its VAT returns, and it would be liable to make good the loss to the Commissioners. In that turn of events, the transactions would cease to be neutral and would damage the company.

The Commissioners raised assessments on the basis that the company had wrongly claimed relief in respect of payments of VAT to sub-contractors. To make good their claim, the Commissioners had to prove fraud; and they claimed that the company, through its site managers, was party to three frauds directed at the Commissioners. The VAT Tribunal found the allegations proved in relation to most of the alleged sub-contractors. By the appeal to Dyson J, the company sought to set aside that decision.

The tribunal had attributed the relevant acts and knowledge of the site agents to the company. Those agents were not the company’s directing mind and will, but (applying the Meridian guidance) the tribunal regarded such an attribution as required by the policy of the Value Added Tax Act 1994. The judge agreed with the tribunal on this, subject to the question of whether the making of such an attribution was, as the company contended, excluded by the Hampshire Land principle. The basis of that contention was, as the judge summarised in paragraph 51 of his judgment:

“… that the company was in a very real sense a victim of the fraud in that it paid the VAT shown on each of the invoices, and if the commissioners’ argument is accepted, they were not entitled to claim input relief. On any view, the company suffered a cash flow detriment in paying the VAT and only subsequently being credited with the input relief.”

Thus the damage to the company was said to be at least the cash flow loss – and worse if the VAT relief had been wrongly claimed. After referring to the decisions in Hampshire Land, Houghton and Belmont Finance, the judge expressed as follows his reasons for rejecting the submission and upholding the attribution:

“55.

In my judgment, the tribunal correctly concluded that there should be attribution in the present case, since the company could not sensibly be regarded as a victim of the fraud. They were right to hold that the fraud was ‘neutral’ from the company’s point of view. The circumstances in which the exception to the general rule of attribution will apply are where the person whose acts it is sought to impute to the company knows or believes that his acts are detrimental to the interests of the company in a material respect. This explains, for example, the reference by Viscount Sumner in J.C.Houghton & Co v. Nothard Lowe and Wills Ltd [1928] AC 1 at 19 to making ‘a clean breast of their delinquency’. It follows that, in judging whether a company is to be regarded as the victim of the acts of a person, one should consider the effects of the acts themselves, and not what the position would be if those acts eventually prove to be ineffective. As the tribunal pointed out, in Director General of Fair Trading v. Pioneer Concrete (UK) Ltd [1995] 1 AC 456 the company suffered a large fine for contempt of court on account of the wrongful acts of its managers. The fact that their wrongful acts caused the company to suffer a financial penalty in this way did not prevent the acts and knowledge of the managers from being attributed to it.

56.

The Hampshire Land principle or exception is founded in common sense and justice. It is obvious good sense and justice that the act of an employee should not be attributed to the employer company if in truth, the act is directed at, and harmful to, the interests of the company. In the present case, the fraud was not aimed at the company. It was not intended by the participants in the fraud that the interests of the company should be harmed by their conduct. In judging whether the fraud was in fact harmful to the interests of the company, one should not be too ready to find such harm. In my view, the cash flow point made by Mr Purle [leading counsel for the company] comes nowhere near being serious enough to trigger the principle. Looking at the facts of this case from a common sense point of view, there was no VAT fraud or harm to the interests of the company. The tribunal were entitled to reach this conclusion. It was the correct conclusion to reach.”

That reasoning reflects that, in considering the range of the Hampshire Land principle, a common sense approach is required. The task is to consider at whom the fraud is directed. If it is directed at the company, the exception will apply. In considering that matter, it is necessary to consider the effects of the frauds themselves, not their consequential effect upon the company if they are eventually uncovered. So approaching the matter, the frauds were not directed at, nor intended to harm, the company. They were intended to be neutral as regards the company, with the real victim being the Commissioners. The cash flow point represented inevitable harm to the company, but was not serious enough to trigger the Hampshire Land exception. In assessing whether the acts do harm the company, the court should not be too ready to find such harm.

Applying that approach to this case, Mr Sumption submitted that the fraud of which Mr Stojevic was the architect was not directed at, nor intended to harm, the company. It was directed at, and intended to harm, the banks. It was directed at stealing from them. McNicholas shows that, in assessing whether the Hampshire Land principle applies, it is not appropriate to factor into the consideration the adverse consequences to the company when and if the fraud is found out.

That analysis is said to be endorsed by this court’s decision in Bank of India v. Morris [2005] BCC 739. The liquidators of BCCI had brought a claim against the Bank of India (“BoI”) under section 213 of the Insolvency Act 1986 for being knowingly a party to fraudulent trading by BCCI. The application concerned six transactions, the issue being whether those at BoI responsible for entering into five of them involving a Mr Samant knew that they were thereby assisting BCCI to perpetrate a fraud on its creditors. Patten J found that Mr Samant had the relevant knowledge in relation to the last four of the five transactions and held that his knowledge should be attributed to BoI, whose board had been content to take his assurances in relation to them. The judge made a contribution order against BoI.

BoI appealed. This court (Mummery, Neuberger L.JJ and Munby J) dismissed the appeal, the court’s judgment being delivered by Mummery LJ. The key passages are in a section headed “Attribution of fraud: points on the authorities”. They read:

“114.

Clearly there are some circumstances in which an individual’s knowledge of fraud cannot and should not be attributed to a company. The classic case is where the company is itself the target of an agent or employee’s dishonesty. In general, if would not be sensible or realistic to attribute knowledge to the company concerned, if attribution had the effect of defeating the right of the company to recover from a dishonest agent or employee or from a third party. Mr Moss [leading counsel for BoI] argued that there should be no attribution of knowledge as this was a case in which BoI was the ‘secondary victim’ of Mr Samant. His actions were harmful to the interests of BoI, as he had exposed it to the risk of potential liability for fraudulent trading. We have no hesitation in rejecting that submission. If it were correct, it would never be possible to attribute the knowledge of the individual to a company under section 213. That is contrary to the agreed position that a company is capable of being liable under s.213. Knowledge of fraud may be attributed to a company even though such attribution may expose it to the risk of liability under s.213.

115.

BoI criticised the judge for not applying to this case the principle that the knowledge of an agent should not be attributed to the principal when the agent is acting fraudulently or otherwise in breach of duty and in circumstances where it was contrary to common sense to consider that the agent would have passed on his knowledge of the fraud to his principal. It was argued by BoI that it was contrary to common sense to conclude that Mr Samant, who was found to have been dishonest and to have acted in breach of duty in causing BoI to take part in the transactions with BCCI, would have passed on to the directors of BoI his knowledge that BCCI was conducting its business with an intention to defraud its creditors.

116.

As appears from the principles laid down in Meridian (see above) the terms of the legislation and the circumstances of the case may make it appropriate to attribute knowledge of fraud to the company, even though a person with knowledge of the fraud has acted dishonestly, in breach of his duty to his principal or employer and in circumstances in which he would not have passed on his knowledge to his agent or employer. For example in McNicholas Construction Co Ltd v. Customs & Excise Commissioners [2000] STC 553 a site manager was dishonest in relation to the liability of his employer for VAT. It was held that the policy and content of the legislation overrode any principle that knowledge of fraud was not attributable to a company that was itself a victim of fraud. The company was held liable for VAT evasion.

117.

It was submitted that Patten J was wrong to follow McNicholas. It was distinguishable from the present case as it was necessary to have attribution of knowledge to the company in that case in order to make the VAT legislation work. That was not, it was contended, the case here. There was no need to discern or fashion a special rule of attribution in order to prevent the policy of the legislation from being frustrated. Section 213 can work perfectly well on the application of the primary rules of corporate responsibility. On the judge’s findings the acts of the employee, Mr Samant, were dishonest. His knowledge should not be attributed to BoI as the acts of Mr Samant were aimed at and harmful to its interests and it cannot be inferred that he would have passed on his knowledge of the fraud of BCCI to the board of BoI.

118.

We agree with Patten J on this point. As in McNicholas, the acts of Mr Samant were not in fact targeted at BoI. He was acting for, and in what he apparently believed to be the interests of, BoI in seeking to gross up the balance sheet for the purposes of the year end accounts. The potential liability of BoI under s.213 is irrelevant in deciding whether BoI was a victim of Mr Samant and whether his knowledge should be attributed to it for the purposes of s.213.”

Part of that reasoning is, I consider, to the effect that the policy underlying particular statutory provisions may require the attribution of the dishonest knowledge of a particular individual to a company, whether or not the company could be regarded as a victim (whether secondary or otherwise) of the dishonest individual. That view is supported by the latter part of paragraph 114, in relation to the policy underlying section 213; and a similar point is made in the latter part of paragraph 116 in relation to the policy underlying the VAT legislation with which McNicholas was concerned.

If that were the limit of what can be derived from this passage, it might not be of direct help in the present case, in which the question of attribution is not conditioned by legislative policy. But Mr Sumption submitted that paragraph 118 goes beyond what had preceded it. There the court in terms endorsed Dyson J’s approach in McNicholas that, in considering whether or not the Hampshire Land principle is applicable, it is necessary to consider against whom the dishonest acts were targeted. In that context, as Dyson J had held, it had been irrelevant to take into account McNicholas’s potential liabilities if the site managers’ frauds were exposed, just as in Bank of India it was irrelevant to consider BoI’s potential liability under section 213 of the Insolvency Act 1986 if a fraudulent trading claim were brought. By parity of reasoning, Mr Sumption submitted, the fact that Mr Stojevic’s scheme might, when and if the frauds were discovered, rebound upon the company was irrelevant in identifying the victim of the frauds. As to that, there was only one answer, it was the banks. The company might ultimately be made to answer for its liabilities incurred in consequence of the frauds. But that did not make it a target so as to enable it to invoke the Hampshire Land principle. There is, moreover, no good reason in principle for not attributing to a company the decision of its board (or of its directing mind and will) to procure it to defraud a third party.

The only authority to which we were referred that was said to point in a different direction, and upon which Mr Brindle placed reliance, was the decision of Rix J in Arab Bank PLC v. Zurich Insurance Co [1999] 1 Lloyd’s Law Reports 262. It was cited to the court in McNicholas, but not referred to in Dyson J’s judgment. It was also cited to, and discussed by this court, in Bank of India, and I shall return to that. The assumed facts in Arab Bank were complicated but it is unnecessary to detail them. Mr Browne, the managing director of JDW, was alleged to have made some fraudulent valuations in JDW’s name. The claimants obtained judgments against JDW and sought to enforce them against JDW’s insurers, Zurich, under the Third Parties (Rights against Insurers) Act 1930. Zurich sought to avoid the policy on the basis that JDW had dishonestly failed to disclose Mr Browne’s alleged frauds. The argument required the court to attribute knowledge of the frauds to the company; but it failed on the basis that such attribution was implicitly excluded by the terms of the policy. The importance of the decision is that Rix J also dealt with an alternative attribution argument, as he explained in this passage, at 282:

“There remains the question, raised by Mr Tomlinson’s submission [he was leading counsel for Zurich], whether the Hampshire Land doctrine is confined to cases of fraud where the principal is himself the victim of the fraud, or whether, as Mr Justice Vaughan Williams put it in Hampshire Land itself, the doctrine extends to other breaches of duty where common sense would destroy the inference of transfer of knowledge. In the typical case in which the doctrine has been applied, Houghton, Belmont, PCW and Group Josi Re, fraud has been found or assumed. In the present case, fraud is also assumed, but the primary victim of the fraud has been the lending institution which has relied on the valuation. I would accept, however, the plaintiffs’ submission that JDW was also a victim, even if only a secondary victim, of the assumed fraud. One consequence of that assumed fraud has been JDW’s liability to the plaintiffs, albeit in negligence. Moreover, even if it could be said that JDW, unlike the plaintiffs, was not the victim of Mr Browne’s fraud, Mr Browne has, on the assumed facts, been guilty of dishonesty, and one can hardly visualize a graver dereliction of duty to his company. Although the cases often involve fraud, Hampshire Land itself did not necessarily do so, and I note that in Group Jose Re Lord Justice Saville was prepared to accept as a working definition of the scope of the principle the cases of ‘the agent’s or director’s fraud or other breach of duty to the company’ (at p. 367). In my judgment, Mr Browne’s fault comes within the concept of an agent’s fraud on his principal, but, even if it does not, his fault is such a breach of duty to JDW as in justice and common sense must entail that it is impossible to infer that his knowledge of his own dishonesty was transferred to JDW. That conclusion is consistent with the view of Mr Justice Colman in Kingscroft v. Nissan Fire and Marine Insurance Co. Ltd., (unreported, Mar.4, 1996, at p. 12).

It follows that the Hampshire Land doctrine would in any event prevent Mr Browne’s or Mr Pitts’ knowledge being attributed to JDW.”

Mr Sumption submitted that that part of Rix J’s reasoning was obiter and that anyway, if that part of it commencing with the words “Although the cases …” was intended to be one of general principle, it was incorrect. Rix J did not ask himself the question that in McNicholas Dyson J was later to regard as the key one, namely what was the object of the fraud? Moreover, Rix J’s analysis cannot be reconciled with this court’s approach in Bank of India as to the circumstances in which a director’s knowledge will be attributed to the company. Putting it generally, the essence of the Hampshire Land principle is that the only circumstance in which the director’s knowledge will not be attributed to the company is the one in which the director would not himself disclose it because it is the company itself that is the target of his dishonest scheme or his breach of duty. Rix J can, however, be read as expressing the principle in significantly wider terms.

Arab Bank was cited to this court in the Bank of India case. The argument of Mr Moss QC, for BoI, was that McNicholas was inconsistent with Arab Bank. At paragraph 124 of its judgment, this court rejected the proposition that Arab Bank was authority for the proposition that there can be no attribution of knowledge where the company is a “secondary victim” of the individual’s wrongdoing or breach of duty.

Mr Brindle’s submission was that Rix J’s observations in the passage I have quoted from Arab Bank formed a second ratio of his decision; and he made that good by referring to Rix J’s introduction (at 280) to his second line of reasoning, where he said “Even if I am wrong about that construction [i.e. the first reason for his decision], however, I think that Mr Boswood and Mr Goldsmith are right in their reliance on the Hampshire Land exception.” Mr Brindle said that Arab Bank was therefore authority for the proposition that it is not a condition of the application of the Hampshire Land principle that the company should be the primary victim. It may be that KB was a victim in the present case. But the question, as in all cases, is whether it can realistically and sensibly be said that the company is also a victim. If it can, the Hampshire Land principle can apply. That is supported by Rix J’s approach. Why, asked Mr Brindle, should the company have to be the primary victim? He accepted that the company’s victimhood had to be significant and not merely theoretical and unconvincing. But, subject to this, there was no reason in principle why there should not be more than one victim of an agent’s dishonest stratagem. As to whether the company was a victim in this case, the fraud may well have been directed primarily against KB, but its consequential effect was to strip the company bare and leave it exposed to liabilities to KB for the amounts of which KB was defrauded. On any footing, the company was worse off after the frauds than before. The present case, he said, was quite different from McNicholas, in which the essential loss suffered by the company itself was a cash flow loss, and in which the key to Dyson J’s reasoning is in the first two sentences of paragraph 55 (quoted above). The damage to McNicholas is to be contrasted with the potential damage to the company in the present case. Mr Brindle said that Dyson J was not holding that the company could not be a secondary victim. He was saying no more than that, on the facts, McNicholas was not a real and substantial victim so as to engage the Hampshire Land principle.

As for the Bank of India case, Mr Brindle said that the only detriment that BoI relied upon as entitling it to invoke the Hampshire Land principle was the very liability that the liquidators of BCCI were seeking to impose on it, so that it was not surprising that the court rejected the argument – on the basis that, if it were accepted, it would deprive section 213 of, in many cases, all practical utility. Mr Brindle submitted further that, in paragraph 118, the court was simply making the point that, as in McNicholas, BoI was not a victim at all. Mr Samant was, the court there acknowledged, acting in what he believed to be BoI’s best interests. As for the court’s comment on Arab Bank in paragraph 124 of its judgment (summarised above), Mr Brindle said that if it was there saying that Arab Bank was not authority for the proposition that “secondary victimhood” could entitle a company to claim the benefit of the Hampshire Land principle, the statement was not justified. Rix J’s judgment was plainly authority for that proposition. Mr Brindle submitted that all that the court was really saying in paragraph 124, taking that paragraph in context, was that it is not enough for a company simply to put up a flag saying “secondary victim” and thereby claim that there should be no attribution. What instead has to be done – as McNicholas shows – is to consider whether the company was in a realistic sense a victim. Considerations of whether it was a primary or secondary victim do not come into it. If the answer is yes, Arab Bank is authority for the proposition that the Hampshire Land principle can apply.

The difference between counsel turns, therefore, on whether or not this is a case to which the Hampshire Land principle applies. Mr Sumption submitted that it does not, and Mr Brindle that it does. Mr Brindle recognised that the principle is of limited scope and will not ordinarily be available to protect a company from liability (civil or criminal) as a result of the dishonest or other wrongful acts of its agents directed at third parties. In particular, it will not prevent such agents’ dishonesty or other wrong from being imputed to the company itself in such a context. That appears to me to be in principle correct.

The area in which the Hampshire Land principle does come into play is one in which, were the agent’s dishonesty or wrongdoing to be attributed to the company, it would have the consequence of prejudicing the company in the pursuit of its own claims or rights against others. That will be because so to attribute the knowledge would bar the claim or extinguish the right. In the Hampshire Land case itself, the case proceeded on the basis that had Mr Wills’s knowledge of the irregularity of the meeting been attributed to Portsea, Portsea could not have proved for its loan. In Belmont Finance, the attribution to the plaintiff company of the conspiratorial intentions of the two defendant directors would have made the company a knowing party to the conspiracy directed against it in respect of which it sued and would have disabled it from suing. Mr Sumption’s submission was that the principle can only apply in circumstances in which the company was the intended victim of the dishonest agent and that this is not such a case. The frauds in which the company engaged admittedly resulted in its incurring liabilities of its own to those it defrauded. But that was a necessary consequence of the frauds, it was not the object of the enterprise and it does not make it a victim for the purposes of the relevant principle. The reason why the company was exposed to liability to KB was because it was the perpetrator of the fraud, not the victim of it.

Mr Brindle’s general point was that the principle cannot be such as to preclude the company, in the circumstances of a case such as this, from bringing a claim against Mr Stojevic himself (for breach of his various duties to the company resulting in its incurring of liabilities to third parties) or against an outsider such as the auditors for negligently failing to identify the frauds, blow the whistle and so stop it from stealing yet more money. Mr Brindle submitted that the frauds, even if directed against the banks, also had materially damaging consequences upon the company sufficient to make it a real victim as well. It must, therefore, be the case that such a claim could be brought; and it must in turn follow that the Hampshire Land principle must enable it to be brought. His proposition was, in effect, that even accepting (as he did) that the company can, as regards the third parties who were defrauded, be regarded as itself having committed the frauds, it must be notionally acquitted of any imputation of dishonesty so as to leave it free to sue those who could be said to have caused (by not stopping) the consequential damage that its own frauds caused it to suffer.

The logic of Mr Brindle’s submission led him into territory that he appeared to recognise as posing difficulties. The claim is brought by the company at the suit of its liquidators. If the company is entitled to the chose in action against the firm that it now asserts, it must have been entitled to the same chose in action immediately before the liquidation, at a time when Mr Stojevic was in control: the liquidation did not give it the benefit of a cause of action it did not have before. Let it be assumed that at some point before the company went into liquidation, Mr Stojevic had procured it to bring a like claim against the firm as the present one, for damages to compensate it for the incurring of its liability to KB. Mr Brindle’s position was that in a hypothetical scenario such as that, in which Mr Stojevic was pulling the strings behind the litigation, there could be no question of the company being allowed to bring such a claim. He adopted the company’s position as acknowledged before Langley J, namely that “we would be laughed out of court”. His submission to us on the point was that:

“The court is going to find some way of making sure that claim is knocked out. Hampshire Land is flexible enough, we submit, to cater for such a situation. It is indeed a flexible doctrine. It is there to protect, where they exist, the rights of innocent interests in the company. That is the whole point behind Belmont and the other cases that we have looked at. In a case where it is simply an elaborate and rather cheeky claim brought by Mr Stojevic effectively to get back on behalf of the company moneys which he has himself pocketed, qua crook, the court is simply not going to apply any [such] principle. There are no innocent parties to be represented and there is no reason why the court should [not] say, ‘In that case we will not apply the special rule of [non-] attribution which we would otherwise normally, namely Hampshire Land.’ The beauty of Lord Hoffmann’s elegant analysis in Meridian of the rules of attribution is they are flexible. It says it all depends on what the situation is.”

I respectfully disagree with that submission, which appears to me to be unprincipled. As I have said, if the company, now in liquidation, has a claim in negligence against the firm, it must have had the same claim before it was in liquidation. Mr Brindle’s proposition that the Hampshire Land principle is in the nature of a discretionary judicial switch that the court can turn on or off according to the identity of the assumed pockets into which the fruits of a particular corporate claim will or may ultimately go appears to me to be wrong in principle, as well as unsupported by authority. The Hampshire Land principle is a principle of non-attribution that, in circumstances in which it is properly applicable, will enable a company to bring a claim or pursue a right that it could not otherwise have brought. Its application cannot depend on a consideration of who will or might ultimately benefit from the bringing of such a claim or the pursuit of such a right, which would often be an impossible inquiry.

If Mr Brindle were right that the company could not have brought a pre-liquidation claim, it must follow that it cannot bring one now. But since I have rejected his submission in relation to that, there still remains the question of whether or not the Hampshire Land principle applies. The essential issue is whether (a) as Mr Sumption submitted, the principle applies only in the case in which the company is the intended victim of its agents, it being irrelevant that it may also suffer incidental or consequential damage; or (b) as Mr Brindle submitted, it applies also in the case where a third party is the intended victim but the company also suffers material consequential damage. Mr Sumption’s submission can be said to derive support from McNicholas and Bank of India. Mr Brindle’s relies on Arab Bank.

I find it a little surprising that McNicholas and Bank of India emerge as authorities contributing to the jurisprudence on the application of the Hampshire Land principle. They were both concerned with fixing liability on a company at the suit of a third party and a central question in each was whether the relevant statutory policy (respectively the VAT legislation and the insolvency legislation) required the attribution to the company of the acts of its agents, being agents who were not its directing mind and will. Once, as in each case it did, the court held that the applicable policy did require such attribution, I find it difficult to see on what basis it was considered that such attribution could or might be trumped by the Hampshire Land principle, which is primarily concerned not with a company’s liabilities to others but rather with its claims against others.

But, surprising or not, there is no escaping that both in McNicholas and in Bank of India the court discussed the scope of the Hampshire Land principle. In my judgment both cases support Mr Sumption’s submission that the principle will ordinarily only apply in circumstances in which the agents intend to harm the company (McNicholas, paragraph 56), or it is the target of their acts (Bank of India, paragraph 118), and that it is not enough to engage the principle that agent’s acts may result in harm to the company. In the former case it made no difference that the agents’ frauds were found out and resulted in material harm to McNicholas in the shape of assessments to tax of more than £1 million (see [2000] STC 553, at 775, paragraph 1); and in the latter case it made no difference that Mr Samant’s actions resulted in BoI being made liable under section 213 to a judgment of over US $80 million (see [2005] BCC 739, at 743, paragraph 1). In both cases the companies were, in the phrase used in argument, left “holding the baby”, just as the company is said to have been here. Both authorities support the view that being a “secondary” victim of this nature is not enough to engage the principle; what counts is the identification of the victim against whom the fraudulent acts are directed. The logic underlying this approach is that it is irrelevant in the present context to take account of the adverse consequences to the fraudster of being a fraudster: those are simply the consequences that the law visits on fraudsters, but they do not, in the present context, make the fraudster a victim. Whilst I recognise Arab Bank as pointing in a different direction, I take the view that this court in Bank of India preferred and approved the reasoning in McNicholas, and in my judgment we should take our lead from Bank of India.

In these circumstances I am of the opinion that this is not a case in which the Hampshire Land principle has any application. The essence of the case is that it is one in which the sole directing mind and will of the company procured it to enter into fraudulent transactions with banks. It was the company that dealt with the banks and, so it seems to me, clear that, as between the company and the banks, the principles of attribution require the dishonesty of the company’s sole human agent to be imputed to the company. Mr Sumption’s submissions satisfied me that this is a case in which such an imputation should be made and that the company should therefore itself be liable for the frauds. Whilst, as I have said, Mr Brindle did not accept that this is the correct analysis, he did not argue against it and he was prepared to accept it for the purposes of the present debate. It is not therefore a case in which the company was the target, or the victim, of its agent’s dishonesty. It was itself the fraudster, it was not the target of the fraud, and in my view it can make no difference that its frauds were likely, when and if found out, to result in the incurring of liabilities by the company itself. It now seeks to be notionally absolved of the dishonesty that must be attributed to it in order to bring claims for compensation for the losses that its own frauds have brought upon itself. Unless it is so absolved, its claim is admittedly (but subject to its separate “very thing” argument) barred by the principle of ex turpi causa.

In my judgment, subject as aforesaid, the company’s claim is so barred. I would uphold this part of Langley J’s decision.

The essence of the company’s “very thing” argument is this. If, as I have held, Mr Stojevic’s fraud is to be attributed to the company, then it is recognised that the making of the claim by the company against the firm does require the company to allege and rely on its fraud committed upon the banks. On the face of it, that means that it runs head on into the ex turpi causa principle of public policy barring it from bringing such a claim.

In response to that consequence, the company confesses and seeks to avoid. By way of avoidance, it is said that the public policy point simply does not apply in the particular circumstances of the case. That is because “the very thing” – or at least one of the very things – that the firm’s engagement as the company’s auditors required it to do was to detect whether the company was engaged in frauds of the nature that it was in fact engaged in. The assumed facts require the court to assume that the firm negligently failed to detect the frauds and that, had it done so, they would have blown the whistle and stopped them. It is said that the ex turpi causa principle cannot bar a claim based on the commission of a fraud when the prevention of that fraud was “the very thing” that the defendants had been charged to do.

The firm’s answer to the point is that it is wrong and that “the very thing” argument is no answer to the public policy point represented by the ex turpi causa principle. Mr Sumption submitted that the principle the company seeks to invoke is, on analysis, nothing to do with the application or otherwise of the ex turpi causa principle -- which sweeps indiscriminately across the board and bars all turpitudinous cases, however good they might otherwise be -- but is all to do (and only to do) with principles of causation. In particular, it is one that is capable of being deployed in answer to a variety of causation defences, for example volenti non fit injuria and novus actus interveniens. It is basic that in every claim in contract or tort the claimant must show that the loss he has suffered is within the scope of the duty he asserts. Put generally, if the particular misfortune that the claimant has suffered is “the very thing” that the defendant had a duty to prevent, it will not be open to the defendant to say that the occurrence of the misfortune broke the chain of causation. But, Mr Sumption said, accepting that the nature of the claim is a “very thing” claim, it still cannot succeed if it is reliant on illegality. In that event the ex turpi causa principle will bar it.

Mr Sumption illustrated his point by concrete examples. In discussing the illegality cases, I referred to Clunis. The central element of Mr Clunis’s claim was that the psychiatrist ought to have realised that he was in urgent need of treatment and was dangerous, that he should have been assessed before he was released and that, had he been, he would have been detained or consented to become a patient and would not have committed manslaughter. Consistently with the principle that the nature of the claimed loss must be within the scope of the claimed duty, his case was essentially that the doctor owed him a duty of care to take steps that would have avoided his commission of manslaughter. At [1998] QB 978, at 987B, Beldam LJ recorded that “the claim against the defendant is founded on the assertion that the manslaughter of Mr Zito was the kind of act which Dr Sergeant ought reasonably to have foreseen and that breaches of duty by the defendant caused the plaintiff to kill Mr Zito.” Prevention of the manslaughter was, therefore, one of the very things that the psychiatrist’s duty required of her. But Mr Clunis’s claim was still defeated by the application of the ex turpi causa maxim.

Similarly, Mr Sumption said that Cross v. Kirby was a “very thing” case. The defendant owed the claimant a duty not to hit him with a baseball bat and injure him. He did the very thing (or one of the very things) he should not have done to the claimant, who sued him. The claim was defeated on ex turpi causa grounds. In Hewison too the employer owed the claimant a duty of care, the scope of which included a duty (one of the “very things” comprised in it) not to cause a loss of future earnings. The bid to recover such damages was defeated by the ex turpi causa principle. The firm’s position is that, logically, almost every claim in contract or tort in which the claim has been defeated by that principle is, on analysis, a “very thing” case.

It appears to me that that probably rather overstates the width of the submission that Mr Brindle makes, which is essentially that “the very thing” principle applies only to cases in which the commission of the illegality is itself the thing that the due performance of the duty was intended to avoid. On that basis, it seems to me, neither Cross v. Kirby nor Hewison is within the principle, although I should have thought that, if the argument is correct, Clunis ought to have been within it, although Mr Brindle disagreed.

The main authority upon which Mr Brindle founded the submission is this court’s decision in Reeves v. Commissioner of Police of the Metropolis [1999] QB 169, a case with certain similarities to Clunis. Both cases were argued before different constitutions of this court in October 1997, with judgment being given in Reeves on 10 November 1997 and in Clunis on 5 December 1997. Neither case makes reference to the other, of which presumably each constitution was respectively unaware. I have said all that I need to say about how Clunis dealt with the ex turpi causa issue. In order to understand Reeves insofar as it deals with the like issue, it is necessary to look first at the earlier, also similar, case of Kirkham v. Chief Constable of the Greater Manchester Police [1990] 2 QB 283.

Mr Kirkham committed suicide whilst on remand at a remand centre. The police had known of his suicidal tendencies but failed to pass this information on to the prison authorities. Had they done so, the judge found, the probability is that he would have been placed on the hospital wing at the centre rather than in an ordinary cell. The claim was by his widow against the Chief Constable for damages for negligence. The judge upheld her claim and rejected a defence of ex turpi causa. The Chief Constable appealed to this court, re-opening his defences of volenti non fit injuria (which was rejected) and ex turpi causa.

Lloyd LJ explained (at 290G) that the judge had rejected the ex turpi causa defence on the simple basis that as, since the Suicide Act 1961, suicide was no longer a crime, it was not available: the widow’s claim was not based on any illegality. He said that the judge had not, however, been referred to three recent authorities on the topic of illegality, namely Thackwell v. Barclays Bank Plc [1986] 1 All ER 676; Saunders v. Edwards [1987] 1 WLR 1116; and Euro-Diam Ltd. v. Bathurst [1990] 1 QB 1. They showed that the ex turpi causa defence rested on a principle of public policy that the courts will not assist a plaintiff who has been guilty of illegal or immoral conduct. As Kerr LJ had put it in Euro-Diam, at 35, “[i]t applies if in all the circumstances it would be an affront to the public conscience to grant the plaintiff the relief which he seeks because the court would thereby appear to assist or encourage the plaintiff in his illegal conduct or to encourage others in similar acts ….”

The three cases just listed were those which later (in 1991) influenced Nicholls LJ in Tinsley (in this court) to recognise the “public conscience” test as the touchstone of how the courts should deal with assertions of illegality. That test was of course later (in 1993) outlawed by the decision of the House of Lords in Tinsley. It follows that, although it did not realise it at the time, the court in Kirkham approached the ex turpi causa principle on an erroneous basis.

Lloyd LJ asked himself whether to afford relief to the widow would affront the public conscience and held that it would not. The Suicide Act 1961 had abolished the crime of suicide and suicide was anyway no longer regarded with the same abhorrence as formerly. Importantly, the evidence was also that, whilst Mr Kirkham was sane in the legal sense, he was not of sound mind: he suffered from clinical depression and his judgment was impaired. In holding that the ex turpi causa defence did not apply, Lloyd LJ qualified his decision by saying it did not apply “at any rate where, as here, there is medical evidence that the suicide is not in full possession of his mind” (291H). Farquharson LJ said much the same, at 296B to D, with the same qualification. Sir Denys Buckley, at 297D, agreed with both lords justices.

I now come to Reeves. The facts were these. Martin Lynch was in custody at a police station. The officers responsible for his custody knew he was a suicide risk. Three months before he had attempted to strangle himself with his belt in a cell at a magistrates’ court. After a morning’s visit to court, when he again tried to strangle himself, he was returned to custody where he saw the police surgeon who gave instructions that, as a suicide risk, he should be kept under observation. Shortly after that examination, Mr Lynch hanged himself. He could do so because the hatch in the cell door had been left down, which enabled him to tie his shirt through the spyhole on the outside of the door and hang himself with it. Sheila Reeves, his administratrix, sued the police commissioner for alleged negligence. The judge found that the police owed Mr Lynch a duty of care because they knew he was a suicide risk. The content of that duty was to take reasonable care to prevent such a person being held by them from committing suicide. It was negligent of the police officers not to shut the hatch because it was reasonably foreseeable that it gave Mr Lynch the opportunity to strangle himself in the way he did. Subject to two matters, the judge found a causative link between the negligent act and Mr Lynch’s death.

The judge, however, held that the commissioner could avoid liability in reliance on the defences of volenti non fit injuria and contributory fault, which latter he fixed at 100%. He was attracted to the view that the ex turpi causa principle also afforded a defence but did not decide the case on that ground. On Mrs Reeves’s appeal to this court, the commissioner ran all three points plus the defence of novus actus interveniens, to which the judge had given no separate consideration, treating it as closely bound up with the volenti defence. It is necessary to consider each of the three judgments delivered by the court.

Buxton LJ, delivering the leading judgment, dealt first with volenti, which he held afforded no defence. In so deciding he referred to Kirkham. He gave several reasons for rejecting this defence, but I need refer only to the first. For that he drew in particular on Farquharson LJ’s opinion at [1990] 2 QB 283, at 295C, that “the [volenti] defence is inappropriate when the act of the deceased relied on is the very act which the duty cast upon the defendant required him to prevent” (emphasis added). As Buxton LJ put it, at [1999] QB 169, 178B:

“If the police’s obligation was to guard against suicide, that is, to protect Mr Lynch from a deliberate act against his own life, I do not see how they can be or should be exempted from liability because that deliberate act in fact occurred.”

Buxton LJ then turned to novus actus interveniens. The point made was that Mr Lynch’s death was caused by his own act, not by the acts or omissions of the police and so the necessary link between their negligence and any damage suffered by Mr Lynch was broken. Buxton LJ rejected that too, drawing this time on the observations of Tucker LJ in Stansbie v. Troman [1948] 2 KB 48. As Buxton LJ explained, at 180G:

“Tucker LJ said, at pp. 51-52, that the general rule with regard to responsibility for acts done by a third party was ‘even though A is in fault, he is not responsible for injury to C which B, a stranger to him, deliberately chooses to do’ (see Weld-Blundell v. Stephens [1920] AC 956, 986, per Lord Sumner), but that principle did not apply where ‘the act of negligence itself consisted in the failure to take reasonable care to guard against the very thing that in fact happened’” (emphasis added).

Buxton LJ then dealt with the defence of contributory fault. He rejected that as well. After referring to the need for the commissioner to show that the claimed damage was suffered as a result “partly of [Mr Lynch’s] own fault” (section 1 of the Law Reform (Contributory Negligence) Act 1945), he said, at 181H:

“The basic problem in applying that test to this case is the same as with the other defence relied on by the commissioner: that it is simply artificial to contend that a defence to liability can rest upon the performance by the deceased of the very act that the defendant was under a duty to take reasonable steps to prevent” (emphasis added).

Buxton LJ explained further difficulties inherent in that defence, which I need not relate. That brought him to the final defence, ex turpi causa. He summarised the approach of Lloyd and Farquharson L.JJ to the like question in Kirkham, referring to the “public conscience” test they had applied. The point made in Reeves was that, unlike Mr Kirkham, Mr Lynch was of sound mind and so his case was had been specifically reserved by Kirkham. The case made on behalf of the Commissioner in Reeves was, as Buxton LJ said at 185B, that “we should hold that the defence of ex turpi causa does indeed hold in this case.” Buxton LJ then said this, at 185B to C:

“When a judge is asked to hold that a particular outcome would affront the public conscience or shock the ordinary citizen it behoves him to proceed with caution, as did this court in the Kirkham case. No evidence will be available on which to base such conclusions, and therefore the exercise must be one of speculation, albeit one would hope intelligent speculation. In the present case, however, I feel able to address the issue without descending into the arena in that way, because there are in my view clear reasons why the defence of ex turpi causa cannot, as a matter of logic and of legal principle, be available in this case.”

The first two sentences of that passage reflect, with respect, an error of approach. They show that the court was being asked to decide the ex turpi causa issue by reference to the “public conscience” test. Whilst it is understandable why the court in Kirkham went down that route, by the time of Reeves it had been closed off by the House of Lords in Tinsley. The inference from Buxton LJ’s judgment – indeed from all three judgments in Reeves -- is that the case was nevertheless argued on the basis that the “public conscience” test was still good law and that this court in Kirkham had correctly directed itself on it. The arguments of counsel as summarised in the law reports do not reflect that they made any reference to Tinsley. But Buxton LJ, as one might expect, was himself aware of it, because at 185F he said that “the actual application of Kerr LJ’s exposition of ex turpi causa in the Euro-Diam case itself has been disapproved: see Tinsley v. Milligan [1994] 1 AC 340, 363.” Having said that, he then said “[n]evertheless, the exposition in my view remains a valuable guide to the basis of the defence, and was accepted as such by Lloyd LJ in the Kirkham case [1990] 2 QB 283.”

If I may respectfully say so, at that point Buxton LJ took a wrong turn. Tinsley shows that the discretionary “public conscience” test does not have – nor strictly ever had – any part in the consideration by the courts of an ex turpi causa defence. Buxton LJ could not correctly regard it as providing any guide, let alone a valuable one, as to the basis of that defence. The question for the court in Reeves under this head was simply whether or not the claimant was relying on an illegal or immoral act committed by Mr Lynch. If yes, her claim must fail. If no, the ex turpi causa defence was not available. Buxton LJ did not direct himself in those terms.

He did, however, anyway reject the ex turpi causa defence and I must cite the paragraph of his reasoning that is the inspiration for the company’s “very thing” argument. He said, at 185D to E (immediately after the passage at 185B to C that I have quoted):

“First, the defence fails on a logical ground similar to that which is fatal to the defence of volenti. If it shocks the conscience of the ordinary citizen that a suicide could recover, why is it the duty of the police, not merely as public officers but in the private law of negligence, to take reasonable steps to prevent suicide? The case is, again, quite different from the usual application of ex turpi causa, where the plaintiff suffers injuries in the course of a criminal enterprise such as an affray or burglary. Here, the alleged turpitudinous act is the very thing that the defendant had a duty to try to prevent, imposed by a law of negligence which itself appeals to public conscience or at least to public notions of reasonableness” (emphasis added).

That is the high point of the company’s argument, in particular the last sentence. With respect, however, it appears to me to provide no support for it. It is right to note that in the passage immediately before, Buxton LJ disclaimed that he was going to descend into the “public conscience” arena, although (i) if he was not going to do so, he did not indicate what different test he was applying; and (ii) as the paragraph twice, and materially, refers to “conscience” and “public conscience”, it appears to me that in fact Buxton LJ was applying the “public conscience” test.

The paragraph is premised on the basis that suicide is an illegal or immoral act such that a damages claim reliant upon it will attract the public policy defence, although it is not clear that Buxton LJ had in fact decided that suicide is such an act. The passage is also premised on the incorrect basis that the operation of the ex turpi causa defence is a discretionary one geared to considerations of possible affronts to the public conscience. The first point made is, in effect, that if it is just to impose a duty of care upon the police to take reasonable steps to prevent a suicide, why should it shock the conscience of the ordinary citizen that the estate of the suicide should recover compensation for a breach of that duty? Buxton LJ’s answer is that it should not cause any such shock. Another answer might perhaps be that, whilst the police may be subject to a duty to take care to prevent the commission of the (assumed) illegal or immoral act, it does not necessarily follow that the public conscience would not be affronted by the recovery of damages in reliance on that act. But there is no need to consider further the rival answers, or their respective merits, to the hypothetical question because it is irrelevant for present purposes. The point of the whole paragraph is that it was informed by the “public conscience” test, and it was in that context that Buxton LJ made his “very thing” reference, in effect a repetition of his first point and going to whether, as a matter of discretion, the ex turpi causa defence should be allowed to prevail.

The reason therefore why none of this helps the company’s submission is because Buxton LJ was applying the wrong test. He should have applied the indiscriminating, uncompromising reliance test explained in Tinsley: i.e. if the claim relies on illegality or immorality, it is automatically barred. Had he applied that principle, and on his apparent assumption that the claim in Reeves was founded on an act attracting the ex turpi causa defence, he could not have said that that defence was unavailable because of “the very thing” point; or, perhaps more accurately, there is nothing in the context of what he actually said to suggest that he would have said that. He ought logically to have held that the claim was defeated. If he was disposed not so to hold, he could not have done so without explaining why Tinsley did not compel that result.

The outcome of Buxton LJ’s judgment was, therefore, that he was in favour of allowing the appeal. Morritt LJ, who would have dismissed it, focused first on the novus actus defence, referring to Mrs Reeves’s argument that “the occurrence of the very thing which it was the duty of the defendant to use reasonable care to prevent cannot be a new intervening force so as to break the chain of causation (189C)” (emphasis added). He referred to Stansbie and said that counsel were unable to refer to any case in which “the very thing” was the voluntary and intentional action of a plaintiff of sound mind (189G). In his view the fact that Mr Lynch was of sound mind made all the difference. The police’s negligence may have increased the risk that he would take his own life, but it did not increase the risk in a causative sense. Morritt LJ would have dismissed the appeal on that ground. He also accepted that the claim was barred by volenti. As for ex turpi causa, he appears also to have regarded the “public conscience” test as the applicable one, but took the view that it was not appropriate for the court to brand as contrary to public policy an act of suicide when that was no longer a crime (195F to G). Mr Brindle did not suggest that anything Morritt LJ said assisted his “very thing” argument.

Lord Bingham of Cornhill CJ agreed with Buxton LJ that the appeal should be allowed. He did not regard the suicide as a novus actus, because to do so would deprive the duty on the police of meaningful content. “This was, after all, the very thing against which the defendant was duty-bound to take precautions” (196G) (emphasis added). He too appears to have regarded the ex turpi causa defence as turning on considerations of public conscience, but did not regard it as available. Since Mr Brindle regarded part of Lord Bingham’s observations as supporting his submission, I will quote the part he relied upon, at 197G:

“It cannot, however, be said, in my judgment, that by permitting recovery in a case such as this the law is covertly conniving at or countenancing suicide: it is indeed imposing a civil penalty on those who, having a duty to try to prevent suicide, fail to do so. I do not, either, think that the conscience of the ordinary citizen would be affronted by the awarding of damages to the estate of a deceased in a case such as this.”

That can, I consider, be said to reflect a similar approach to that adopted by Buxton LJ. It can be said to assume that a claim based on a suicide is ex turpi causa but it reflects that Lord Bingham did not regard such a claim as likely to affront the conscience of the ordinary citizen. The remarks do not, however, support the view that the ex turpi causa defence in the present case is overridden by “the very thing” argument. The reason is because, again, Lord Bingham was applying the discretionary public conscience test, whereas – if he regarded the claim as arising ex turpi causa – he should have held it to be automatically barred.

Reeves went to the House of Lords ([2000] AC 360), but the public policy defence was not pursued there. The argument was all about causation, and it was in that context that the House referred to “the very thing” point (see per Lord Hoffmann, at 367H, and Lord Hope of Craighhead, at 381B). There is no need to refer further to the decision of the House.

Mr Brindle also relied on a decision of the Singapore Court of Appeal in United Project Consultants Pte Ltd v. Leong Kwok Onn [2005] 4 Singapore Law Reports 214. The respondent was the auditor of the appellant company. The appellant was required to pay a tax penalty arising from its failure to make a proper return in respect of fees payable to its directors. The appellant sued the respondent for negligence. The case was based on the proposition that the respondent ought to have discovered that the appellant had been providing incorrect information to the tax authorities. The trial judge found that the respondent was not in a position to know that the appellant had been providing incorrect information and so it was unreasonable to impose on the respondent a duty to discover the making of incorrect tax returns. The judge also found that the ex turpi causa defence barred the appellant’s claim. An appeal against his decision was allowed. As regards the ex turpi causa defence, the court held that the appellant’s relevant conduct was not criminal in nature nor could it be classified as reprehensible or grossly immoral. “In short, we were of the view that the appellant had not engaged in an act that was so culpable as to attract the application of the illegality defence” (paragraph 57).

Having so found, the court also went on to hold that, even it were wrong in that conclusion, the illegality defence “must fail for the simple reason that the loss suffered by the appellant was precisely the loss that the respondent was engaged to avoid” (paragraph 58). In the following paragraphs it explained why, its explanation showing that the factual situation was rather different from that in the present case. The real nature of the appellant’s case was that the respondent had negligently failed so to advise it as to avoid the subsequent incurring of the penalties. I will let the court speak for itself:

“60.

The respondent, however, alleged that the commission by the appellant of a statutory offence constituted an illegal act that disentitled the latter from pursuing its claim in tort. This argument placed the proverbial cart before the horse. On a proper appreciation of the facts, the appellant’s running afoul of the Act could be attributed solely to the fact that the respondent had failed in his duty to warn. To allow the respondent to rely upon a consequence that was directly caused by his own failings and to absolve him from liability, would be to reward the wrongdoer and punish the innocent party.

61.

An example would, perhaps, be apposite. Suppose a layperson were to employ the services of a solicitor to execute a trust for the benefit of his dependants. The solicitor reviews the manner in which the client has structured his affairs and, whilst realising that there may be a potential statutory breach for evading stamp duties, chooses to remain quiet. IRAS subsequently imposes a penalty on the client for failing to pay the requisite amount of stamp duty. In these circumstances, it cannot be countenanced that the solicitor was not answerable to the client for having negligently failed to warn the latter of the potential liability for failure to pay stamp duty merely because the client had committed a statutory offence.”

That description of the facts shows why the court was not sympathetic to the illegality defence. The case is, however, I consider ultimately unhelpful because it appears that, although the court had the benefit of the citation of a good deal of authority, it did not include Tinsley, and the court’s reasoning was not therefore directed to whether or not, had it found illegality and been required to apply the strict Tinsley principle, the ex turpi causa defence would have succeeded. The most help that Mr Brindle can derive from the case is that the authorities cited included Reeves; and in paragraph 62 the court apparently regarded Reeves as decided in part on the basis that “the very thing” argument is an answer to an ex turpi causa defence. The court said:

“62 … The same court [the Court of Appeal] had also decided, in the earlier case of Reeves v. Commissioner of Police for the Metropolis [1999] QB 169 ([53] supra), that the illegality defence could not be afforded to the police, in a situation concerning a claim by the dependants of a deceased who had committed suicide in police custody, where the police were aware that the deceased was a suicide risk. The reason was that the claimant’s conduct was the very act that the police were under a duty of care to prevent.”

In my judgment that is not, with respect, an accurate summary of the decision in Reeves. It may be a fair reflection of the particular basis on which Buxton LJ explained his rejection of the illegality defence, and I have quoted and commented on the material part of his judgment. It was not the basis on which Morritt LJ or Lord Bingham of Cornhill CJ dealt with matter. Morritt LJ decided that the public policy question did not come into play because suicide was no longer a crime; and Lord Bingham appears to me simply to have applied the “public conscience” test. The holding the Singapore court attributes to this court in Reeves was not part of the ratio of the decision in Reeves.

In the court below, Langley J, having found against the company on the attribution issues, said in his conclusions that he regarded the Reeves point – “the very thing” argument -- as the key question. He did not, however, expressly answer it. He expressed his reason for not striking the claim out as follows:

“But trying to stand back, and at the risk of fudging the strictures to be found in Tinsley v. Milligan (albeit with some authority to encourage me), I do not think that the ‘conscience of the ordinary citizen’ would find anything so repugnant in S&R pursuing this claim which would justify ruling it impermissible by the use of the unforgiving and uncompromising operation of the ex turpi maxim”

As I said earlier, it is agreed that that reflected an adoption by the judge of the wrong approach to the ex turpi causa principle, which – in a case in which it applies – is an “unforgiving and uncompromising” one. It is surprising that the judge so expressed himself because, as I have also said, he had earlier correctly identified the applicable principles, including that the “public conscience” test no longer formed part of them.

We do not, however, have the benefit of the judge’s own thoughts on “the very thing” submission. Despite Mr Brindle’s sustained submission in support of the proposition that the ex turpi causa principle can afford no defence to a claim based on the illegality that it was the very duty of the defendant to prevent, I have come to the conclusion that the submission is mistaken. I agree with Mr Sumption that “the very thing” concept is all about causation. Reeves is a good example of that. The notion that the administratrix of a man of sound mind who chooses to kill himself should have a cause of action in tort against others for not taking reasonable care to stop him doing so is not one that is obviously well founded; in particular, it is likely to give rise to contentious causation issues. But if, on facts of the sort that arose in Reeves, the defendant does owe duties of care that include a duty to take reasonable steps to prevent an occurrence of an unusual type, being “the very thing” that in fact happened, then the claimant may be able to overcome those causation difficulties.

There is, however, in my judgment no support in the authorities that we were shown for the proposition that if “the very thing” from which the defendant owed a duty to save the claimant harmless is, or includes, the commission of a criminal offence, the public policy defence based on the ex turpi causa principle will be overridden so as to enable the bringing of the claim that relies on the claimant’s illegality. For reasons I have explained, there is no support for it in Reeves. That is because there the court was applying the wrong, discretionary “public conscience” test with regard to the ex turpi causa principle. Whilst at [1999] QB 169, at 185E, Buxton LJ used “the very thing” phrase in the context of a discussion of whether the claim should be barred as ex turpi causa, he was there saying no more than that, if the duty was to prevent the suicide, the public conscience would not be shocked by allowing the claimant to sue in respect of it.

It is accepted in the present case that the company’s claim relies upon its own illegality. The question for us is whether the claim must necessarily fail on ex turpi causa principles. In my judgment, the teaching of Tinsley is that it must and that there is no discretion in the matter. No member of the court gave consideration to that in Reeves. In my judgment Reeves provides no authority binding on us for the proposition that the ex turpi causa principle is trumped by “the very thing” argument. If the point were good, Clunis ought not to have been decided on the public policy point in the way it was. Clunis applied the ex turpi causa principle correctly, whereas Reeves did not. I would reject the company’s “very thing” argument. I have no doubt that the public policy principle encapsulated in the ex turpi causa maxim is intended to apply to a claim such as this one; and, if I may be forgiven just a little more Latin, to stop it in limine.

It follows that, whilst I agree with Langley J’s conclusion on the attribution issue, I consider that he ought to have held the company’s claim to be barred by the ex turpi causa public policy principle. I would allow the firm’s appeal and strike out the company’s claim against the firm.

I agree, as I do with the judgment of Mummery LJ, which I have had the benefit of reading in draft.

I agree with my Lord’s admirable analysis of the authorities and his conclusions on the overall legal position. I add a few comments on this astounding claim.

Who was the victim of the fraud? The bank KB. Mr Brindle QC said not so: the victim was Stone & Rolls, the one man company owned and controlled by Mr Stojevic. Acting entirely under his direction the company became party to his fraud. In my view, Mr Stojevic and his company together defrauded the bank KB. To suggest that the corporate creature used by Mr Stojevic as the vehicle for the fraud was the victim of the fraud committed by him with it turns the world upside down. There is no prospect of establishing at trial that the company was the victim of the fraud.

Is the firm potentially liable in negligence to this fraudulent company? No. The company’s case is that the firm failed to use reasonable care to detect Mr Stojevic’s fraud. In my view, the firm did not owe a duty of care to the company, which was a fraudster in the total grip of another fraudster. As a general rule a fraudster, individual or corporate, is legally liable for the losses that flow directly from the fraud and cannot blame such losses on the negligence of another person, such as the victim or whoever.

What difference does the liquidation of the company make? None. The fact that the company is in insolvent liquidation is irrelevant to the company’s cause of action against the firm. It is also irrelevant that, if successful, the claim could benefit the victim bank KB. The company in liquidation is asserting the same cause of action that was vested in it prior to liquidation. Liquidation of the company did not alter the existing cause of action or create a new cause of action. Prior to liquidation the cause of action against the firm could, if successful, benefit the fraudulent company. The same is true post-liquidation and is a reason for striking the claim out.

How about “the very thing” - the claim against the firm being for “the very thing” that they were retained and paid to do (i.e. to take reasonable care to detect fraud in the affairs of the company)? Does this affect the negligence claim against the firm? No. No duty of care is owed by the firm to the fraudster company, which was party to, and not itself a victim of, the fraud, to take reasonable care to detect its fraud.

What about the primary and secondary rules on attribution to the company of knowledge and of a guilty mind? They are relevant to whose acts count as acts of the company for the purposes of the substantive rule in question and to fixing it with responsibility for/liability for the fraud of its director or vicarious liability. In my judgment, in this case the knowledge of the fraudulent mastermind and the knowledge of his creature company are identical in targeting the victim bank. It is not a case of a company itself being an innocent victim of deception by one of its officers. This company was party to the fraud, not an innocent victim of it.

Does common sense matter? Yes. It is contrary to all common sense to uphold a claim that would confer direct or indirect benefits on the corporate vehicle, which was used to commit the fraud and was not the victim of it, and the fraudulent driver of the fraudulent vehicle.