(back to preceding text)

    96. Thus all the alternative solutions have their difficulties. But there are two problems which they fail to solve, but which are easily solved if the beneficial interest remains throughout in the lender. One arises from the fact, well established by the authorities, that the primary trust is enforceable by the lender. But on what basis can he enforce it? He cannot do so as the beneficiary under the secondary trust, for if the primary purpose is fulfilled there is no secondary trust: the pre-condition of his claim is destructive of his standing to make it. He cannot do so as settlor, for a settlor who retains no beneficial interest cannot enforce the trust which he has created.

    97. Dr Chambers insists that the lender has merely a right to prevent the misapplication of the money, and attributes this to his contractual right to specific performance of a condition of the contract of loan. As I have already pointed out, this provides no solution where the arrangement is non-contractual. But Lord Wilberforce clearly based the borrower's obligation on an equitable or fiduciary basis and not a contractual one. He was concerned to justify the co-existence of equity's exclusive jurisdiction with the common law action for debt. Basing equity's intervention on its auxiliary jurisdiction to restrain a breach of contract would not have enabled the lender to succeed against the bank, which was a third party to the contract. There is only one explanation of the lender's fiduciary right to enforce the primary trust which can be reconciled with basic principle: he can do so because he is the beneficiary.

    98. The other problem is concerned with the basis on which the primary trust is said to have failed in several of the cases, particularly Toovey v Milne 2 B & A 683 and the Quistclose case itself [1970] AC 567. Given that the money did not belong to the borrower in either case, the borrower's insolvency should not have prevented the money from being paid in the manner contemplated. A man cannot pay some only of his creditors once he has been adjudicated bankrupt, but a third party can. A company cannot pay a dividend once it has gone into liquidation, but there is nothing to stop a third party from paying the disappointed shareholders. The reason why the purpose failed in each case must be because the lender's object in making the money available was to save the borrower from bankruptcy in the one case and collapse in the other. But this in itself is not enough. A trust does not fail merely because the settlor's purpose in creating it has been frustrated: the trust must become illegal or impossible to perform. The settlor's motives must not be confused with the purpose of the trust; the frustration of the former does not by itself cause the failure of the latter. But if the borrower is treated as holding the money on a resulting trust for the lender but with power (or in some cases a duty) to carry out the lender's revocable mandate, and the lender's object in giving the mandate is frustrated, he is entitled to revoke the mandate and demand the return of money which never ceased to be his beneficially.

    99. There is a further point which is well brought out in the judgment of the Court of Appeal. On a purchase of land it is a commonplace for the purchaser's mortgagee to pay the mortgage money to the purchaser's solicitor against his undertaking to apply it in the payment of the purchase price in return for a properly executed conveyance from the vendor and mortgage to the mortgagee. There is no doubt that the solicitor would commit a breach of trust if he were to apply it for any other purpose, or to apply it for the stated purpose if the mortgagee countermanded his instructions: see Bristol and West Building Society v Mothew [1998] Ch 1, 22. It is universally acknowledged that the beneficiary of the trust, usually described as an express or implied trust, is the mortgagee. Until paid in accordance with the mortgagee's instructions or returned it is the property of the mortgagee in equity, and the mortgagee may trace the money and obtain proprietary relief against a third party: Boscawen v Bajwa [1996] 1 WLR 328. It is often assumed that the trust arises because the solicitor has become the mortgagee's solicitor for the purpose of completion. But that was not the case in Barclays Bank Plc v Weeks Legg and Dean [1999] QB 309, 324, where the solicitor's undertaking was the only communication passing between the mortgagee and the solicitor. I said:

"The function of the undertaking is to prescribe the terms upon which the solicitor receives the money remitted by the bank. Such money is trust money which belongs in equity to the bank but which the solicitor is authorised to disburse in accordance with the terms of the undertaking but not otherwise. Parting with the money otherwise than in accordance with the undertaking constitutes at one and the same time a breach of a contractual undertaking and a breach of the trust on which the money is held."

The case is, of course, even closer to the present than the traditional cases in which a Quistclose trust has been held to have been created. I do not think that subtle distinctions should be made between "true" Quistclose trusts and trusts which are merely analogous to them. It depends on how widely or narrowly you choose to define the Quistclose trust. There is clearly a wide range of situations in which the parties enter into a commercial arrangement which permits one party to have a limited use of the other's money for a stated purpose, is not free to apply it for any other purpose, and must return it if for any reason the purpose cannot be carried out. The arrangement between the purchaser's solicitor and the purchaser's mortgagee is an example of just such an arrangement. All such arrangements should if possible be susceptible to the same analysis.

    100. As Sherlock Holmes reminded Dr Watson, when you have eliminated the impossible, whatever remains, however improbable, must be the truth. I would reject all the alternative analyses, which I find unconvincing for the reasons I have endeavoured to explain, and hold the Quistclose trust to be an entirely orthodox example of the kind of default trust known as a resulting trust. The lender pays the money to the borrower by way of loan, but he does not part with the entire beneficial interest in the money, and insofar as he does not it is held on a resulting trust for the lender from the outset. Contrary to the opinion of the Court of Appeal, it is the borrower who has a very limited use of the money, being obliged to apply it for the stated purpose or return it. He has no beneficial interest in the money, which remains throughout in the lender subject only to the borrower's power or duty to apply the money in accordance with the lender's instructions. When the purpose fails, the money is returnable to the lender, not under some new trust in his favour which only comes into being on the failure of the purpose, but because the resulting trust in his favour is no longer subject to any power on the part of the borrower to make use of the money. Whether the borrower is obliged to apply the money for the stated purpose or merely at liberty to do so, and whether the lender can countermand the borrower's mandate while it is still capable of being carried out, must depend on the circumstances of the particular case.

    101. After this over-long exposition, it is possible to dispose of the remaining objections to the creation of a Quistclose trust very shortly. A trust must have certainty of objects. But the only trust is the resulting trust for the lender. The borrower is authorised (or directed) to apply the money for a stated purpose, but this is a mere power and does not constitute a purpose trust. Provided the power is stated with sufficient clarity for the court to be able to determine whether it is still capable of being carried out or whether the money has been misapplied, it is sufficiently certain to be enforced. If it is uncertain, however, then the borrower has no authority to make any use of the money at all and must return it to the lender under the resulting trust. Uncertainty works in favour of the lender, not the borrower; it cannot help a person in the position of Mr Leach.

    102. Like all resulting trusts, the trust in favour of the lender arises when the lender parts with the money on terms which do not exhaust the beneficial interest. It is not a contingent reversionary or future interest. It does not suddenly come into being like an eighteenth century use only when the stated purpose fails. It is a default trust which fills the gap when some part of the beneficial interest is undisposed of and prevents it from being "in suspense".

    103. In my opinion the Court of Appeal were correct to find that the terms of paragraphs 1 and 2 of the undertaking created a Quistclose trust. The money was never at Mr Yardley's free disposal. It was never held to his order by Mr Sims. The money belonged throughout to Twinsectra, subject only to Mr Yardley's right to apply it for the acquisition of property. Twinsectra parted with the money to Mr Sims, relying on him to ensure that the money was properly applied or returned to it. Mr Sims act in paying the money over to Mr Leach was a breach of trust, but it did not in itself render the money incapable of being applied for the stated purpose. Insofar as Mr Leach applied the money in the acquisition of property, the purpose was achieved.

(4)     Knowing (or dishonest) assistance

    104. Before turning to the critical questions concerning the extent of the knowledge required and whether a finding of dishonesty is a necessary condition of liability, I ought to say a word about the distinction between the "knowing receipt" of trust money and "knowing (or dishonest) assistance" in a breach of trust; and about the meaning of "assistance" in this context.

    105. Liability for "knowing receipt" is receipt-based. It does not depend on fault. The cause of action is restitutionary and is available only where the defendant received or applied the money in breach of trust for his own use and benefit: see Agip (Africa) Ltd v Jackson [1990] Ch 265, 291-2; Royal Brunei Airlines Sdn Bhd v Tan [1995] 2 AC 378, 386. There is no basis for requiring actual knowledge of the breach of trust, let alone dishonesty, as a condition of liability. Constructive notice is sufficient, and may not even be necessary. There is powerful academic support for the proposition that the liability of the recipient is the same as in other cases of restitution, that is to say strict but subject to a change of position defence.

    106. Mr Leach received sums totalling £22,000 in payment of his costs for his own use and benefit, and Twinsectra seek their repayment on the ground of knowing receipt. But he did not receive the rest of the money for his own benefit at all. He never regarded himself as beneficially entitled to the money. He held it to Mr Yardley's order and paid it out to Mr Yardley or his companies. Twinsectra cannot and does not base its claim in respect of these moneys in knowing receipt, not for want of knowledge, but for want of the necessary receipt. It sues in respect of knowing (or dishonest) assistance.

    107. The accessory's liability for having assisted in a breach of trust is quite different. It is fault-based, not receipt-based. The defendant is not charged with having received trust moneys for his own benefit, but with having acted as an accessory to a breach of trust. The action is not restitutionary; the claimant seeks compensation for wrongdoing. The cause of action is concerned with attributing liability for misdirected funds. Liability is not restricted to the person whose breach of trust or fiduciary duty caused their original diversion. His liability is strict. Nor is it limited to those who assist him in the original breach. It extends to everyone who consciously assists in the continuing diversion of the money. Most of the cases have been concerned, not with assisting in the original breach, but in covering it up afterwards by helping to launder the money. Mr Leach's wrongdoing is not confined to the assistance he gave Mr Sims to commit a breach of trust by receiving the money from him knowing that Mr Sims should not have paid it to him (though this is sufficient to render him liable for any resulting loss); it extends to the assistance he gave in the subsequent misdirection of the money by paying it out to Mr Yardley's order without seeing to its proper application.

    108. The classic formulation of this head of liability is that of Lord Selborne LC in Barnes v Addy (1874) LR 9 Ch App 244, 251. Third parties who were not themselves trustees were liable if they were found

"either making themselves trustees de son tort, or actually participating in any fraudulent conduct of the trustee to the injury of the cestui que trust".

In the next passage of his judgment, at p 252, he amplified this by referring to those who

"assist with knowledge in a dishonest and fraudulent design on the part of the trustees".

    109. There were thus two conditions of liability: the defendant must have assisted (i) with knowledge (ii) in a fraudulent breach of trust. The second condition was discarded in Royal Brunei Airlines Sdn Bhd v Tan [1995] 2 AC 378. Henceforth, it was sufficient that the defendant was accessory to any breach of trust whether fraudulent or not. The question for present decision is concerned with the first condition. Since that case it has been clear that actual knowledge is necessary; the question is whether it is sufficient, or whether there is an additional requirement of dishonesty in the subjective sense in which that term is used in criminal cases.

    110. Prior to the decision in Royal Brunei Airlines Sdn Bhd v Tan the equitable claim was described as "knowing assistance". It gave a remedy against third parties who knowingly assisted in the misdirection of funds. The accessory was liable if he knew all the relevant facts, in particular the fact that the principal was not entitled to deal with the funds entrusted to him as he had done or was proposing to do. Unfortunately, the distinction between this form of fault-based liability and the liability to make restitution for trust money received in breach of trust was not always observed, and it was even suggested from time to time that the requirements of liability should be the same in the two cases. Authorities on one head of liability were applied in cases which concerned the other, and judges embarked on sophisticated analyses of the kind of knowledge required to found liability.

    111. Behind the confusion there lay a critical issue: whether negligence alone was sufficient to impose liability on the accessory. If so, then it was unnecessary to show that he possessed actual knowledge of the relevant facts. Despite a divergence of judicial opinion, by 1995 the tide was flowing strongly in favour of rejecting negligence. It was widely thought that the accessory should be liable only if he actually knew the relevant facts. It should not be sufficient that he ought to have known them or had the means of knowledge if he did not in fact know them.

    112. There was a gloss on this. It is dishonest for a man deliberately to shut his eyes to facts which he would prefer not to know. If he does so, he is taken to have actual knowledge of the facts to which he shut his eyes. Such knowledge has been described as "Nelsonian knowledge", meaning knowledge which is attributed to a person as a consequence of his "wilful blindness" or (as American lawyers describe it) "contrived ignorance". But a person's failure through negligence to make inquiry is insufficient to enable knowledge to be attributed to him: see Agip (Africa) Ltd v Jackson [1990] Ch 265, 293.

    113. In his magisterial opinion in Royal Brunei Airlines Sdn Bhd v Tan [1995] 2 AC 378, every word of which merits close attention, Lord Nicholls firmly rejected negligence as a sufficient condition of accessory liability. The accessory must be guilty of intentional wrongdoing. But Lord Nicholls did not, in express terms at least, substitute intentional wrongdoing as the condition of liability. He substituted dishonesty. Dishonesty, he said, was a necessary and sufficient ingredient of accessory liability. "Knowingly" was better avoided as a defining ingredient of the principle, and the scale of knowledge accepted in Baden v Sociétié Générale pour Favoriser le Developpement du Commerce et de l'Industrie en France SA [1993] 1 WLR 509 was best forgotten. His purpose, as he made clear, was to get away from the refinements which had been introduced into the concept of knowledge in the context of accessory liability.

    114. In taking dishonesty to be the condition of liability, however, Lord Nicholls used the word in an objective sense. He did not employ the concept of dishonesty as it is understood in criminal cases. He explained the sense in which he was using the word at [1995] 2 AC 378, 389 as follows:

"Whatever may be the position in some criminal or other contexts (see, for instance, R v Ghosh [1982] QB 1053) in the context of the accessory liability principle acting dishonestly, or with a lack of probity, which is synonymous, means simply not acting as an honest person would in the circumstances. This is an objective standard. At first sight this may seem surprising. Honesty has a connotation of subjectivity, as distinct from the objectivity of negligence. Honesty, indeed, does have a strong subjective element in that it is a description of a type of conduct assessed in the light of what a person actually knew at the time, as distinct from what a reasonable person would have known or appreciated. Further, honesty and its counterpart dishonesty are mostly concerned with advertent conduct, not inadvertent conduct. Carelessness is not dishonesty. Thus for the most part dishonesty is to be equated with conscious impropriety. However, these subjective characteristics of honesty do not mean that individuals are free to set their own standards of honesty in particular circumstances. The standard of what constitutes honest conduct is not subjective. Honesty is not an optional scale, with higher or lower values according to the moral standards of each individual. If a person knowingly appropriates another's property, he will not escape a finding of dishonesty simply because he sees nothing wrong in such behaviour. In most situations there is little difficulty in identifying how an honest person would behave. Honest people do not intentionally deceive others to their detriment. Honest people do not knowingly take others' property. Unless there is a very good and compelling reason, an honest person does not participate in a transaction if he knows it involves a misapplication of trust assets to the detriment of the beneficiaries. Nor does an honest person in such a case deliberately close his eyes and ears, or deliberately not ask questions, lest he learn something he would rather not know, and then proceed regardless."

    115. In R v Ghosh [1982] QB 1053 Lord Lane CJ drew a distinction between dishonesty as a state of mind and dishonesty as a course of conduct, and held that dishonesty in section 1 of the Theft Act 1968 referred to dishonesty as a state of mind. The question was not whether the accused had in fact acted dishonestly but whether he was aware that he was acting dishonestly. The jury must first of all decide whether the conduct of the accused was dishonest according to the ordinary standards of reasonable and honest people. That was an objective test. If he was not dishonest by those standards, that was an end of the matter and the prosecution failed. If it was dishonest by those standards, the jury had secondly to consider whether the accused was aware that what he was doing was dishonest by those standards. That was a subjective test. Given his actual (subjective) knowledge the accused must have fallen below ordinary (objective) standards of honesty and (subjectively) have been aware that he was doing so.

    116. The same test of dishonesty is applicable in civil cases where, for example, liability depends upon intent to defraud, for this connotes a dishonest state of mind. Aktieselskabet Dansk Skibsfinansiering v Brothers [2001] 2 BCLC 324 was a case of this kind (trading with intent to defraud creditors). But it is not generally an appropriate condition of civil liability, which does not ordinarily require a guilty mind. Civil liability is usually predicated on the defendant's conduct rather than his state of mind; it results from his negligent or unreasonable behaviour or, where this is not sufficient, from intentional wrongdoing.

    117. A dishonest state of mind might logically have been required when it was thought that the accessory was liable only if the principal was guilty of a fraudulent breach of trust, for then the claim could have been regarded as the equitable counterpart of the common law conspiracy to defraud. But this requirement was discarded in Royal Brunei Airlines Sdn Bhd v Tan [1995] .2 AC 378

    118. It is, therefore, not surprising that Lord Nicholls rejected a dishonest state of mind as an appropriate condition of liability. This is evident from the opening sentence of the passage cited above, from his repeated references both in that passage and later in his judgment to the defendant's conduct in "acting dishonestly" and "advertent conduct", and from his statement that "for the most part" (ie not always) it involves "conscious impropriety". "Honesty", he said, "is a description of a type of conduct assessed in the light of what a person actually knew at the time." Usually ("for the most part"), no doubt, the defendant will have been guilty of "conscious impropriety"; but this is not a condition of liability. The defendant, Lord Nicholls said, at p 390E, was "required to act honestly"; and he indicated that Knox J had captured the flavour of dishonesty in Cowan de Groot Properties Ltd v Eagle Trust Plc [1992] 4 All ER 700, 761 when he referred to a person who is "guilty of commercially unacceptable conduct in the particular context involved." There is no trace in Lord Nicholls' opinion that the defendant should have been aware that he was acting contrary to objective standards of dishonesty. In my opinion, in rejecting the test of dishonesty adopted in R v Ghosh [1982] QB 1053, Lord Nicholls was using the word to characterise the defendant's conduct, not his state of mind.

    119. Lord Nicholls had earlier drawn an analogy with the tort of procuring a breach of contract. He observed, at p 387 B-C, that a person who knowingly procures a breach of contract, or who knowingly interferes with the due performance of a contract, is liable in damages to the innocent party. The rationale underlying the accessory's liability for a breach of trust, he said, was the same. It is scarcely necessary to observe that dishonesty is not a condition of liability for the common law cause of action. This is a point to which I must revert later; for the moment, it is sufficient to say that procuring a breach of contract is an intentional tort, but it does not depend on dishonesty. Lord Nicholls was not of course confusing knowledge with dishonesty. But his approach to dishonesty is premised on the belief that it is dishonest for a man consciously to participate in the misapplication of money.

    120. This is evident by the way in which Lord Nicholls dealt with the difficult case where the propriety of the transaction is doubtful. An honest man, he considered, would make appropriate enquiries before going ahead. This assumes that an honest man is one who would not knowingly participate in a transaction which caused the misapplication of funds. But it is most clearly evident in the way in which Lord Nicholls described the conduct of the defendant in the case under appeal. The question was whether he was personally liable for procuring or assisting in a breach of trust committed by his company. The trust was created by the terms of a contract entered into between the company, which carried on the business of a travel agency, and an airline. The contract required money obtained from the sale of the airline's tickets to be placed in a special trust account. The company failed to pay the money into a special account but used it to fund its own cash flow. Lord Nicholls described the defendant's conduct, at p 393:

"In other words, he caused or permitted his company to apply the money in a way he knew was not authorised by the trust of which the company was trustee. Set out in these bald terms, the defendant's conduct was dishonest."

There was no evidence and Lord Nicholls did not suggest that the defendant realised that honest people would regard his conduct as dishonest. Nor did the plaintiff put its case so high. It contended that the company was liable because it made unauthorised use of trust money, and that the defendant was liable because he caused or permitted his company to do so despite his knowledge that its use of the money was unauthorised. This was enough to make the defendant liable, and for Lord Nicholls to describe his conduct as dishonest.

    121. In my opinion Lord Nicholls was adopting an objective standard of dishonesty by which the defendant is expected to attain the standard which would be observed by an honest person placed in similar circumstances. Account must be taken of subjective considerations such as the defendant's experience and intelligence and his actual state of knowledge at the relevant time. But it is not necessary that he should actually have appreciated that he was acting dishonestly; it is sufficient that he was.

    122. This is the way in which Lord Nicholls' use of the term "dishonesty" was understood by Mance LJ in Grupo Torras SA v Al-Sabah [1999] CLC 1469. It is also the way in which it has been widely understood by practitioners: see William Blair QC "Secondary Liability of Financial Institutions for the Fraud of Third Parties" (2000) 30 Hong Kong Law Journal 74; Jeremy Chan "Dishonesty and Knowledge" (2001) 31 Hong Kong Law Journal 283; Andrew Stafford QC "Solicitors' liability for knowing receipt and dishonest assistance in breach of trust" (2001) 17 Professional Negligence 3. Mr Blair QC, at p 83, welcomed the "more pragmatic and workable test of objective dishonesty". Mr Stafford QC, at p 14, invited your Lordships to

"Reiterate that honesty is an objective standard and that individuals are not free to set their own standards of proper conduct;

This is almost entirely objective. The only subjective elements are those relating to the defendant's knowledge, experience and attributes. The objective elements include not only the standard of honesty (which is not controversial) but also the recognition of wrongdoing. The question is whether an honest person would appreciate that what he was doing was wrong or improper, not whether the defendant himself actually appreciated this. The third limb of the test established for criminal cases in R v Ghosh [1982] QB 1053 is conspicuously absent. But there is no trace of it in Lord Nicholls' opinion in Royal Brunei Airlines Sdn Bhd v Tan [1995] 2 AC 378 either.