KNOWING ASSISTANCE AND KNOWING RECEIPT: TAKING STOCK

Simon Gardner

Address: Lincoln College, Oxford, OX1 3DR.

Email : simon.gardner@law.ox.ac.uk

(1996) 112 L.Q.R. 56

THE personal liabilities of non-trustees for losses to trusts include the forms of liability commonly referred to as "knowing assistance" and "knowing receipt". The design of these two forms of liability has been the subject of perennial difficulty and debate.

Writing in 1986, Harpum offered a profound and exhaustive review of matters thus far. He was faced with a collection of apparently irreconcilable decisions. As regards knowing assistance, most notably, Selangor United Rubber Estates Ltd v. Cradock (No. 3), Karak Rubber Co. v. Burden (No. 2) and Baden v. Societe Generale pour Favoriser le Developpement du Commerce et de l'Industrie en France S.A. were willing to see liability arise in a comparatively wide set of circumstances, whilst Belmont Finance Corp. Ltd v. Williams Furniture Ltd demanded something narrower. As regards knowing receipt, Nelson v. Larholt, Belmont Finance Corp. Ltd v. Williams Furniture Ltd (No. 2) the Baden case and International Sales and Agencies Ltd v. Marcus took a wide view, Carl Zeiss Stiftung v. Herbert Smith & Co. (No. 2) a narrow one.

Harpum's analysis made it possible to believe that, although the authorities were evidently not unanimous, the differences between them were not so confounding as might be thought. Partly he achieved this effect by dividing the overall topic into a greater number of subdivisions than is commonly found, and grouping these in a slightly unusual way; the authorities within each subdivision could then be presented as moderately consistent, whilst legitimately differing from those captured by other subdivisions. And partly he achieved it by maintaining clear sight of what he himself perceived to be the correct principles in the matter, and accepting or rejecting the authorities accordingly.

Harpum's efforts were, however, greeted by a further flurry of reported litigation and other discussion over the years to date. The present article seeks to take stock of these newer contributions. They may at first sight appear to consist of a wilderness of disjointed thoughts, but on reflection a fair amount of useful mapping can be done. The rules defining these liabilities have developed significantly in the cases, to some extent through a clarification of the concepts of knowledge and notice, upon which the rules often draw. The article begins by reviewing the work done on those concepts, and in the light of it moves on to take stock of the rules as they appear now to stand. More notably still, however, judges and commentators have been much more forthcoming than previously in venturing arguments as to the proper mission(s) of these two forms of liability, and the shape(s) which their rules ought therefore to take. The article therefore devotes most space to examining these arguments, and also to suggesting additional or alternative considerations. It concludes by asking whether a lesson of importance is not to be drawn from the very confusion traditionally associated with this area, and whether attempts to arrive at concluded positions regarding it may not be misconceived.

KNOWLEDGE AND NOTICE

In a moment, then, we shall try to sketch the law as it emerges from the newer cases. Although it may not be inevitable that it should do so, as a matter of fact much of the discussion in these cases involves the concepts of knowledge and notice. It will therefore be useful to develop an understanding of the meanings which are being given to these concepts.

We must be aware of a possible source of confusion. In the Baden case, Peter Gibson J. articulated five categories of cognisance:

"(i) actual knowledge; (ii) wilfully shutting one's eyes to the obvious; (iii) wilfully and recklessly failing to make such inquiries as an honest and reasonable man would make; (iv) knowledge of circumstances which would indicate the facts to an honest and reasonable man; (v) knowledge of circumstances which would put an honest and reasonable man on inquiry."

Knowledge and notice are often defined by reference to these categories. Knowledge is identified with the first three, notice with the remaining two. This may be supportable as regards knowledge, but it is incorrect as regards notice. Categories (iv) and (v) have the defendant knowing the evidence from which the reasonable person would have drawn the crucial inferences, but (scirelicet) himself failing to draw them, albeit that this failure is not "wilful" or "reckless." True notice, however, extends further, to the case where the reasonable person would have discovered or inferred the crucial facts, regardless of whether the defendant himself did or was aware of any evidence from which to do so. In fact, Peter Gibson J. himself made no reference to notice, but consistently referred to all five of his categories as involving knowledge: category (i) being "actual knowledge" and the rest "imputed knowledge," or "constructive knowledge." His classification does not, therefore, offer a definition of knowledge and notice, but a (wide) view as to the meaning of knowledge.

Knowledge

Millet J. in Agip (Africa) Ltd v. Jackson and Vinelott J. in Eagle Trust Plc v. S.B.C Securities Ltd have been principally responsible for the development of ideas regarding knowledge.

For Millett and Vinelott JJ., a requirement of knowledge in the present context is (or ought to be) really a requirement that the defendant be dishonest. To them, following this principle, Peter Gibson J.'s categories (i)-(iii) represent knowledge, but nothing else does. The distinction between categories (ii) and (iii) on the one hand, and notice on the other, is not simply a difference in the degree of obviousness of the facts which the defendant failed to infer or discover. The distinction makes reference especially to the defendant's own position vis-a-vis those facts. With equally obvious facts, some defendants might "wilfully shut their eyes" to those facts or "wilfully and recklessly fail to make reasonable inquiries" about them, while others might remain uncomplicatedly oblivious of them. The former merit grouping together with defendants who actually know the facts (i.e. those in category (i)), because all three such types of defendant are dishonest; in contrast to the latter, who are not dishonest. Millet J. explains:

"If a man does not draw the obvious inferences or make the obvious inquiries, the question is: why not? If it is because, however foolishly, he did not suspect wrongdoing or, having suspected it, had his suspicions allayed, however unreasonably, that is one thing. But if he did suspect wrongdoing yet failed to make inquiries because he 'did not want to know' (category (ii)) or because he regarded it as 'none of his business' (category (iii)), that is quite another. Such conduct is dishonest, and those who are guilty of it cannot complain if, for the purpose of civil liability, they are treated as if they had actual knowledge."

However, Millett and Vinelott JJ. go on to say that such knowledge on the defendant's part can be inferred. The defendants in Agip (Africa) Ltd v. Jackson declined to give evidence, and Millett J. found knowledge and dishonesty on their part by this process of inference. And in Eagle Trust Plc v. S.B.C Securities Ltd, Vinelott J. remarks:

"If the circumstances are such that an honest and reasonable man would have appreciated that he was assisting in a dishonest breach of trust, the court may infer from the defendant's silence that he either appreciated the fact or that he wilfully shut his eyes to the obvious or wilfully and recklessly failed to make inquiries for fear of what he might learn. Whether in such circumstances a defendant can escape liability on the ground that he acted honestly but failed to draw the obvious inference by reason of his inexperience or because he was unusually or unreasonably trusting, or for some other reason, is a question which I do not need to consider further."

Although Vinelott J. goes on to note that Millet J. would allow rebuttal, as appears from Millett J.'s reference to foolishness in the passage set out above, the second sentence here suggests that Vinelott J. contemplates that the inference might be irrebuttable. If it is, there is nothing subjective at all about the idea of knowledge being used here. Allowing rebuttal will, however, often be of scant comfort to defendants. In actions of this kind against non-trustees, the plaintiff goes beyond the relatively easy business of suing the trustee: a step which is likely to be worth taking only if the trustee is not worth suing but the non-trustee is. A non-trustee worth suing will very commonly be a commercial entity of some kind. And it will not be a happy experience for a commercial entity to seek to rebut an inference of knowledge by pleading that it was too foolish to perceive the reasonably perceptible.

It is tempting, then, to say that inferred knowledge is the same as notice. But Vinelott J. insists that there remains a difference between the two. He remarks that "'notice' is often used in a sense or in contexts where the facts do not support the inference of knowledge." His thinking can be gathered from his ensuing remarks. The basis of inferred knowledge is what the reasonable person would have known in the situation. According to Vinelott J., however, notice goes wider than this, and deems a reasonable person to know of facts which in truth no reasonable person would have known, and which it would therefore be improper to infer that the particular defendant knew. This reading of notice is a possible one, if perhaps a little exaggerated. But it describes only the meaning which notice is given in the context of land transfer. In other contexts, as we are about to see, notice has a less exigent meaning, such that it would not be inaccurate to equate it with the case in which it is possible to infer knowledge.

Notice

The concept of notice has been most used in questions regarding rights over land. The relevant caselaw has given the concept a rigorous significance. A person is said to have notice of some fact if the reasonable person would have known or discovered it, and the reasonable person has been treated as highly perceptive, businesslike to the point of suspicious, and prepared to go to very considerable lengths in his search for the truth.

It is a well established position, however, that these exigent standards are inappropriate outside the land transfer context. This view was famously articulated by Lindley L.J. in Manchester Trust v. Furniss:

"The equitable doctrines of constructive notice are common enough when dealing with land and estates, with which the court is familiar; but there have been repeated protests against the introduction into commercial transactions of anything like an extension of these doctrines, and the protest is founded on good sense. In dealing with estates in land title is everything, and it can be leisurely investigated; in commercial transactions possession is everything, and there is no time to investigate title; and if we were to extend the doctrine of constructive notice to commercial transactions we should be doing infinite mischief and paralysing the trade of the country."

There is in fact some difficulty about Lindley L.J.'s remarks. They appear to say that the demanding standards developed in the land transfer context should be confined to that context and not be permitted to apply to "commercial transactions." The problem is that the two categories clearly overlap: many land dealings are commercial. Millett J. has addressed this difficulty in Macmillan Inc. v. Bishopsgate Investment Trust Plc. According to Millett J., the right approach is, in effect, to disregard the question of whether a transaction is "commercial," and look purely to whether it involves land transfer or not - or more precisely, to whether it involves the kind of transaction (normally involving land) in which rigour is the norm, or the kind of transaction (which may involve land but which very often does not) in which rigour is not the norm. That approach is adopted in the article, albeit with the two classes of case identified by Millet J. usually referred to elliptically as involving land transfer or otherwise. It should be said, however, that Lindley L.J.'s mention of "commercial transactions" was not otiose. Lindley L.J.'s concern was indeed to minimise the damage which, he perceived, a rigorous doctrine of notice would do to commerce, presumably in the form of high transaction costs. This concern led him to exclude notice in its land transfer guise from other contexts, as noted by Millet J., but it also led him in other decisions to soften the rigour of notice within the land transfer context itself

Lindley L.J.'s position of confining exigent standards to the land transfer context has been adopted many times since. Discussions of notice in cases concerning knowing receipt and knowing assistance have been no exception. These discussions accept that in the land transfer context, a reference to notice means what the land transfer cases say it means. But they propose that in other contexts, a reference to notice should have an alternative, less exigent, meaning. As to what this alternative meaning might be, however, they divide into two schools of thought.

One school is led by Vinelott J. in Eagle Trust Plc v. S.B.C Securities Ltd. Vinelott J. takes it that the reasonable person always behaves as the land transfer cases hold him to, so that a less exigent reading of notice cannot be achieved while continuing to invoke the reasonable person. Vinelott J. accordingly proposes that outside the land transfer context a reference to notice should be read not as a reference to what the reasonable person would have discovered, but as a reference merely, to what the defendant did discover: in other words, to actual knowledge. Vinelott J. found support for his view in Thomson v. Clydesdale Bank Ltd. There, the House of Lords was dealing with a defence of bona fide purchase without notice in a context which did not involve land transfer. For the most part the judgments render the requirement of (lack of) notice as (lack of) knowledge, or at any rate something close to it. Vinelott J.'s view was adopted by Knox J. in Cowan de Groot Properties Ltd v. Eagle Trust Plc, and by the Court of Appeal in Polly Peck International Plc v. Nadir (No. 2).

The other school of thought is to be seen above all in the judgments of Millett J. in El Ajou v. Dollar Land Holdings Plc and, most recently, Macmillan Inc. v. Bishopsgate Investment Trust Plc. Whereas Vinelott J. assumes that the reasonable person must always behave as rigorously as he does in the land transfer context (and so proceeds not to refer to the reasonable person outside that context), the key to Millett J.'s approach is the idea that the reasonable person behaves differently in different contexts. To be sure, the reasonable person takes great pains over land transfer. Millett J. terms the standard deduced from this "constructive notice in the strict conveyancing sense." But the reasonable person may take less trouble over other kinds of transaction. One can, therefore, continue looking to the behaviour of the reasonable person, but deduce a less exigent standard. This reasoning allows Millett J. to achieve the desired aim, that a reference to notice outside the land transfer context should be less demanding than such a reference within that context, but to do so without ceasing (as Vinelott J. does) to cast the test in the traditional terms of what the reasonable person would have discovered, as opposed merely to what the defendant did discover. Millett J. encapsulates his notion thus:

" I agree that ... there is no room for the doctrine of constructive notice in the strict conveyancing sense in a factual situation where it is not the custom and practice to make inquiry. But it does not follow that there is no room for an analogous doctrine in a situation in which any honest and reasonable man would have made inquiry."

Millett J. goes on to state just how much rigour the concept of notice does comport in, at any rate, high-level commercial dealings with companies' assets. He says:

"A recipient is not expected to be unduly suspicious and is not to be held [to have notice] unless he went ahead without further inquiry in circumstances in which an honest and reasonable man would have realised that the money was probably trust money and was being misapplied."

Although, as noted above, Vinelott J.'s approach is not unsupported by authority, Millett J.'s approach is probably the better-founded of the two. Its roots lie in the perception that the exigent profile given to notice by the land transfer cases is merely a deduction from a more abstract formula, to the effect that notice is a function of how the reasonable person would have behaved in the relevant circumstances; and that it is this formula, rather than the standards set in the land transfer cases, that truly defines notice. Articulations of such an abstract formula may indeed be found; a recent example is a dictum by Lord Browne-Wilkinson in Barclays Bank Plc v. O'Brien.

As the case law stands at the time of writing, however, both Millett J.'s and Vinelott J.'s views are extant. Both agree that a rule making reference to notice (such as that providing the defence of bona fide purchase without notice, or, if this be the case, that establishing liability for knowing receipt or knowing assistance) will apply a standard more lenient than "constructive notice in the strict conveyancing sense" outside the land transfer context. According to Millett J., the rule's reference to notice will direct attention to such facts as the reasonable person would have discovered as a result of the less strenuous efforts which are to be expected of him when land transfer is not involved, which in an appropriate context may mean only such facts as are patently obvious. According to Vinelott J., on the other hand, the reference to notice will in this case be read as a reference to the defendant's knowledge.

That said, the two positions are brought closer again by the fact that the required knowledge can once again be inferred. In Eagle Trust Plc v. S. B. C Securities Ltd, Vinelott J. states, in the context of a discussion of knowing receipt, that the inference will be made " . . . if the circumstances are such that an honest and reasonable man would have inferred that the moneys were probably trust moneys and were being misapplied...". In El Ajou v. Dollar Land Holdings Plc, Millett J. notes this position as tending towards an assimilation with his own view. The closeness of that assimilation depends on the question, left open by Vinelott J., whether the inference is rebuttable, and also on whether, if it is, many defendants will in practice wish to rebut it.

We are now in a position to give an account of the shape which the new decisions have given to knowing assistance and knowing receipt.

Knowing assistance

The new decisions seem generally agreed that notice will not suffice for knowing assistance. At first, the requirement was said to be one of knowledge, but the emphasis has more recently shifted to dishonesty.

The simplest position was taken by Megarry V-C. in Re Montagu's Settlement Trusts and by Alliott J. and May L.J. in Lipkin Gorman v. Karpnale Ltd. All three required knowledge rather than notice, and it seems clear that by knowledge they meant Baden categories (i)-(iii). This was also the basis on which counsel proceeded in Cowan de Groot Properties Ltd v. Eagle Trust Plc.

In Agip (Africa) Ltd v. Jackson, Millett J. asserted that the requirement is fundamentally one of dishonesty on the defendant's part. He went on, however, to seek the presence of knowledge, as the indicator of dishonesty. Vinelott J. in Eagle Trust Plc v. S.B.C Securities Ltd continued this analysis, as in turn did Scott L.J. in Polly Peck International Plc- v. Nadir (No. 2). But, as noted above, both Millett J. and (especially) Vinelott J. went on to say that, when the truth of the situation would have been obvious to a reasonable person, it can be inferred that the defendant knew it too.

Some confusion might have been caused by the Court of Appeal in Agip (Africa) Ltd v. Jackson. There, Fox L.J. cited, apparently with approval, older authority which allowed liability to arise on a wider range of facts. But it was not absolutely clear what Fox L.J. was driving at in the passage in question; Millett J. subsequently took Fox L.J.'s remarks to be directed rather to knowing receipt, though that seems improbable. And certainly, in other parts of his judgment Fox L.J. seemed clear that liability for knowing assistance depends upon dishonesty, which the discussions noted above took as their cue for requiring knowledge rather than notice. Drawing upon this latter point, Vinelott J. in Eagle Trust Plc v. S.B.C Securities Ltd was able to treat Fox L.J. as supporting the view that dishonesty must be present.

Most recently, the Privy Council in Royal Brunei Airlines Sdn. Bhd. v. Tan has also stated that the requirement is one of dishonesty. But in a departure from the approach until now, the judgment, by Lord Nicholls of Birkenhead, goes on to point attention directly to the concept of dishonesty itself, adding that "'knowingly' is better avoided as a defining ingredient of the principle." Accordingly, Lord Nicholls offers an analysis of the concept of dishonesty, something not to be found in the preceding authorities.

Lord Nicholls' idea of dishonesty is fundamentally straightforward enough. Dishonesty, he says, "means simply not acting as an honest person would in the circumstances." The invocation of the "honest person" serves to indicate that the defendant's behaviour is to be judged against an objective standard:

"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."

Some reflection is, however, needed on certain later passages in Lord Nicholls' judgment, which at first sight may seem hard to reconcile with the simplicity of this position. Thus, at one point, Lord Nicholls states that account should be taken of the reason why the defendant acted as he did. If the defendant acted with a good motive, it may be thought, his (otherwise objectionable) behaviour may turn out not to be dishonest; a person stealing food so as to feed his starving children might not be regarded as dishonest. At another point, his Lordship remarks that "acting in reckless disregard of others' rights or possible rights can be a telltale sign of dishonesty. The word "reckless" appears to connote that the defendant not only created a risk of harm to the plaintiff , but also was contemptuous of the plaintiff s right not to be harmed: the suggestion is thus that the defendant's attitude might make his behaviour dishonest. it might be asked whether these positions represent a departure from Lord Nicholls' basic objective standard. It seems not, in fact. Rather, they signify that the test is as follows: does the honest person regard it as honest to behave as the defendant did - that behaviour being defined by reference not merely to the external effects of the defendant's actions, but also to the motives and attitudes with which the defendant performed those actions? There is no contradiction of the underlying message that the defendant's standards of honesty are irrelevant. So if the honest person regarded it as dishonest to steal food even to feed one's starving children, it would avail the defendant nothing that he personally did not.

As regards Lord Nicholls' statement that reference to knowledge should be avoided, however, there could be some difficulty. One can sympathise with his Lordship's desire to go straight to the core notion of (in his view) dishonesty, and not to be tripped up on the way by unnecessary mediating concepts. And it must be said that even the previous decisions which adverted to dishonesty appeared to use knowledge in such a mediating role. To that extent, the Privy Council's view seems commendable. But Lord Nicholls appears to go further, and say that there should be no further analysis of the concept of knowledge: "in the context of this principle [of dishonesty] the Baden scale of knowledge is best forgotten." This cannot be right. Indeed, his Lordship himself states that an assessment of whether a certain action is dishonest requires reference to what the defendant knew as he performed it, both in the sense that his performance must have been advertent and (presumably) in the sense that he must have been alive to the potential effects. That seems correct, especially in view of the point elaborated in the previous paragraph. If a person's behaviour must be defined by a process which includes reference to the motives and attitudes with which he acted, reference must inevitably be made to that person's cognisance (to use a neutral term) of the potential harm in question. But one could not reasonably imagine that it should be self-evident what "cognisance" means here, any more than it is in any other context. To be sure, the criminal law sometimes remits conundrums to the jury, rather than essaying an institutional position. But even that does not amount to an assertion that there is no difficulty of epistemology to be in fact grappled with.

It is doubtful, then, that the new law of direct reference to a concept of dishonesty obviates any need for an exegesis upon cognisance. It is nevertheless possible that for a while the difficulties will be ignored, and positions taken sub silentio. In that sense, the analysis of knowledge in the cases immediately preceding Royal Brunei Airlines Sdn. Bhd. v. Tan may indeed therefore drop out of the picture. But those difficulties will not have evaporated, and such jurists who do continue to ponder them will presumably wish to include the work done in those cases in their ruminations.

At the same time as Lord Nicholls has insisted that liability for knowing assistance requires dishonesty on the part of the defendant, he has ruled that liability for knowing receipt can arise regardless of whether the trustees' breach was itself fraudulent. This ruling runs counter to Lord Selborne L.C.'s famous statement, requiring fraud on the part of the trustees, in Barnes v. Addy. Lord Nicholls observes that Lord Selborne L.C.'s remarks were per incuriam of some earlier decisions which extended the liability to any form of breach. He presents liability for knowing assistance as being directed against loss caused by interference with the jural rights which the plaintiff enjoys against some obligee. Here, the rights are those enjoyed under trusts or other fiduciary arrangements; an analogy may be drawn with the tort of interference with contractual relations. So conceived, it is nothing to the point of such liability whether the loss takes the form of a fraudulent or, at the other extreme, a quite innocent breach by the obligee. One might, in fact, query whether a breach of trust should be necessary at all. It seems that no breach may be necessary for the analogous liability for interference with contractual relations. Certainly, it would make sense for breach not to be required in the trusts context. If, in the events complained of, the trustees relied on the probity and competence of the defendant, and acted as ordinary prudent businessmen in doing so, they will often not be liable for breach. Yet if the defendant acted dishonestly, as the law now requires, he has just as much harmed the jural rights comprising the trust as if there were a breach. As yet, however, the cases appear not to have considered this issue.

Knowing receipt

The law on knowing receipt has in recent years veered markedly. In Re Montagu's Settlement Trusts, Megarry V.-C. held that liability for knowing receipt could arise only if knowledge was present. To Megarry V.-C., this apparently meant the kinds of cognisance identified in Baden categories (i)-(iii), though presumably they would be inferable in the manner more recently proposed by Vinelott J. in Eagle Trust Plc v. S.B.C Securities Ltd, discussed above. Re Montagu's Settlement Trusts was accepted as correct by Alliott J. in Lipkin Gorman v. Karpnale Ltd, whose thinking was in turn accepted as correct by Steyn J. in Barclays Bank Plc v. Quincecare Ltd. But a quite different approach prevailed in the decision of Millet J. in Agip (Africa) Ltd v. Jackson, which in turn heralded a series of further analyses whose precise significance requires some thought.

In Agip (Africa) Ltd v Jackson, Millett J. stated his view as follows:

" . . . the person who receives for his own benefit trust property transferred to him in breach of trust . . . is liable as a constructive trustee if he received it with notice, actual or constructive, that it was trust property and that the transfer to him was a breach of trust, or if he received it without such notice but subsequently discovered the facts. In either case he is liable to account for the property, in the first case as from the time he received the property and in the second as from the time he acquired notice."

So it seems clear enough that Millett J. was accepting notice as sufficing for liability. It is true that in the variant where the defendant originally receives the property without notice, Millett J. looks to his having subsequently "discovered" the facts, but in the next sentence he refers to this event as the recipient having "acquired notice," which effects a reconciliation with his original mode of expression. The only other reason for uncertainty is that he later refers to Megarry V.-C.'s "doubt [in Re Montagu's Settlement Trusts] whether constructive notice is sufficient," and continues: "whether the doubt is well-founded or not. ... " But taken with Millett J.'s other remarks, this passage seems a reflection rather of courtesy towards a brother judge than of uncertainty on Millett J.'s part over his own view.

Subsequent decisions have provided further discussion, principally by Vinelott J. in Eagle Trust Plc v. S.B.C Securities Ltd, by Knox J. in Cowan de Groot Properties Ltd v. Eagle Trust Plc, and by Millett J. again in El Ajou v. Dollar Land Holdings Plc. None of these rejects Millett J.'s original ruling in Agip (Africa) Ltd v. Jackson that a proper description of liability for knowing receipt will make reference to notice. But, as discussed above, they have demanded a more subtle understanding of what such a reference to notice entails. To recapitulate. All agree that "constructive notice in the strict conveyancing sense" applies only to land transfer cases. So far as other types of case are concerned, there is a difference of approach. According to Vinelott and Knox JJ., the reference to notice should be read as requiring knowledge, though such knowledge is inferable, in the manner discussed earlier. According to Millett J., on the other hand, the reference to notice remains an invocation of the standards of the reasonable person, but with the caveat that those standards vary according to the circumstances, and outside the context of land transfer are likely to be undemanding. In practice, however, the two approaches may largely coincide.

THE PRINCIPLES

The most significant feature of the period begun by Harpum's 1968 contributions has been the discussion, on a much larger and more impressive scale than previously, of the principles underlying knowing assistance and knowing receipt. We now turn to these.

Knowing assistance

Various arguments have been advanced in order to show that the law is right to require dishonesty, or perhaps knowledge per se, as a condition of liability for knowing assistance.

One argument posits that the liability is "necessarily fault-based." It may be observed that that in itself might not necessarily rule out a requirement of notice, as a species of fault may be involved in that concept. But there is a prior objection. Taken literally, the argument is fundamentally unsound. Even criminal liability is not "based" on the defendant's fault. Rather, it is "based" on the fact the defendant has done some harm or wrong. Consideration of fault enters in, if at all, in the ensuing effort to ensure that liability for doing the harm or wrong arises only where it is just, or otherwise appropriate. Civil liability is no different in this respect. Specifically, liability for knowing assistance is "based" on the defendant's having helped to cause loss to the trust. There then arises the question under discussion, of whether the liability should attach only if the defendant was at fault, and if so whether the required degree of fault should be dishonesty or something less.

A second argument used to start out by drawing attention to the requirement that the trustees' breach must be dishonest. This requirement was reaffirmed by the Court of Appeal in Belmont Finance Corp. Ltd v. Williams Furniture Ltd, following Lord Selbome L.C.'s formulation in Barnes v. Addy. The argument from that point on is put representatively by Sir Peter Millett:

"Given that the breach of trust in question must be a fraudulent and dishonest one, it logically follows that constructive notice of the fraud is not enough to make the accessory liable. There is no sense in requiring dishonesty on the part of the principal while accepting negligence as sufficient for his assistant. Dishonest furtherance of the dishonest scheme of another is an intelligible basis for liability; negligent but honest failure to appreciate that someone else's scheme is dishonest is not."

With respect, this was never a good argument. The invocation of logic was misplaced. It is entirely intelligible to impose liability for negligently assisting in the dishonest causing of loss. For example, that is presumably an exact definition of the liability of store detectives and security personnel. Their job is to guard against the loss caused to their hirers by thieves (who by definition are dishonest), and they incur liability if, negligently, they fail to do so. In reality, then, the argument rested not on logic, but on a presupposition that liability for knowing assistance is liability not merely in connection with a fraud (the required dishonest breach), but for participation in a fraud, in an especially pure form wherein the participant is appropriately to be regarded as a fraudster himself. That form of the analysis was intelligible, but stood or fell with its presupposition that the liability was indeed aimed at participation in a fraud. That presupposition has now been displaced by the decision of the Privy Council in Royal Brunei Airlines Sdn. Bhd. v. Tan. As we saw earlier, Lord Nicholls' judgment establishes that liability for knowing receipt can arise even where the trustees' breach was itself innocent. Arguments that knowing assistance is about (necessarily fraudulent) participation in a fraud have accordingly evaporated.

Harpum, in a new contribution, has essayed a third argument to the effect that, even without a requirement of dishonest breach, this liability should necessarily require knowledge. It should be noted at the outset that this argument appears to be for knowledge in its own right, rather than for dishonesty, or for knowledge as an indicator of dishonesty: so it is not quite an apologia for the prevailing cases, especially not for Royal Brunei Airlines Sdn. Bhd. v. Tan. Essentially, Harpum holds onto the idea of participation, albeit now in any breach of trust, and proposes that participation necessarily demands knowledge. He draws on the earlier work of Sales. Sales had identified liability for participation in a wrong, such as a breach of trust, as "secondary liability". Secondary liability, according to Sales, is the phenomenon whereby the law treats as jointly responsible for the wrong, alongside its prime mover, anyone who was sufficiently connected with its occurrence - exactly analogously with the criminal law's liability for complicity in another's offence, and as exemplified also by such forms of liability as that for inducing a breach of contract. Then, although Sales himself was silent on the point, Harpum maintains that secondary liability cannot arise in the absence of knowledge on the defendant's part of the wrong in question. The syllogism is thus complete. Knowing assistance is a species of secondary liability; secondary liability requires knowledge; so knowing assistance requires knowledge.

This argument is clearly of some attraction. Nevertheless, there are difficulties with it. First, it is not correct to say that secondary liability observably arises only where knowledge is present. Although there is classic authority to that effect, it is not unchallenged. This is especially so in the criminal context, where the authority requiring knowledge seemingly relates particularly to defendants who become embroiled in a crime as they pursue their otherwise legitimate activities; a significant corpus takes a tougher line as regards defendants who in any event are engaged in nefarious activities. Secondly, there is a query over the very idea of secondary liability as a discrete concept. As regards criminal complicity, it seems that a person who advances or contributes to the occurrence of a proscribed harm, and is at fault in so doing, attracts liability on those very grounds, rather than parasitically upon the liability of the person whose hand actually does the deed. Equally, in tort, a person who (by act or, in appropriate cases, omission) helps in a second person's causing of loss to a third person can incur a primary liability. It seems hard, then, to be sure that we should follow Harpum in his deductions from secondary liability. If anything, the material relied upon - especially in criminal law - shows rather that there is actually a range of possible conditions for liability, the choice between them being influenced by the situation in the round.

Finally, however, there is the argument of Lord Nicholls in Royal Brunei Airlines Sdn. Bhd. v. Tan. His Lordship ultimately favours a requirement of dishonesty. On examination, in fact, he offers little positive support for this choice. He asserts that a requirement of dishonesty is more meaningful than one of knowledge, but that is not to say that either is appropriate. Instead, his argument largely proceeds by debunking the case for the main rival to dishonesty (or knowledge), namely negligence. (Note his Lordship's reference to negligence rather than to notice. As discussed earlier, in contexts outside land transfer, a requirement of notice has sometimes been read as one of knowledge. On that view, a reference to notice would for the most part not have provided a contrast with dishonesty, as a reference to negligence does. Remember, however, the approach to notice which invariably makes reference to the standards of the reasonable person, but acknowledges that these are less exigent outside the land transfer context. Like his Lordship's reference to negligence, that approach does retain a contrast with dishonesty.)

Lord Nicholls' view appears to be that a liability for knowing assistance involving negligence is either unnecessary or inappropriate. He notes that many potential defendants in knowing assistance will be persons and entities such as banks - they may be loosely termed agents - who voluntarily enter into an engagement with the trustees, and consequently owe a duty of care, which will normally be in contract but may perhaps alternatively be in tort. His Lordship believes that a defendant who has undertaken such a duty of care should be liable under it alone, and not be liable in knowing assistance as well. A defendant who has not undertaken such a duty of care, however, should according to his Lordship not find himself placed under such a duty by knowing assistance being allowed to operate in cases of negligence: only in cases of dishonesty.

Lord Nicholls offers an analysis of when a defendant will and will not be under such a negligence liability to the trust in contract or tort. His Lordship's view appears to be that the liability will start out in life as a duty which the agent owes to the trustees, to save them from breaching their trust. It will crystallise, if, failing in this duty, the agent allows the trustees to breach their trust, and thus to incur a loss by way of their own liability for breach. The agent's liability is to make good this loss. Then, says his Lordship, this liability of the agent to the trustees is held by the trustees for the trust. And that is how the agent comes to owe a liability for want of care to the trust. This analysis is in fact a little curious. On the one hand, whilst some agents will owe trustees a duty to save the latter from breach, and must indemnify them if they fail in this, any such duty will by its nature be owed to the trustees personally, and not held by them on their trust. This liability thus has no common ground with knowing assistance, which is indeed concerned with the defendant's liability to the trust. On the other hand, such an agent will normally - much more normally - owe the trust itself a duty of care in contract or tort. This is the main effect of an agent being engaged by trustees to provide services for the trust, the trustees in this case certainly holding the benefit of the engagement on the trust. It is this material, therefore, which is comparable with liability for knowing assistance. With the benefit of a duty of care arising thus, however, it should be noted that a defendant who breaks it will be liable to the trust not only where the trustees are innocent, as Lord Nicholls thought, but also where they are fraudulent (as indeed, at the other extreme, where they are not in breach at all).

But this is ultimately a detail, concerning the exact point at which we traverse from one to the other of Lord Nicholls' two reasons why there is no role for a negligence liability in knowing assistance. The important thing is the reasons themselves. According to his Lordship there is no role for such a liability because it would either, first, replicate a negligence liability which the defendant had voluntarily undertaken, and so be unnecessary, or, secondly, impose such a liability where he had not voluntarily undertaken one, and so be unwarrantable.

His Lordship's first reason is reminiscent of the courts' reluctance in recent years to allow a claim in negligence in cases where the parties had a contract. That reluctance has been carefully explicated and justified by Stapleton. But it appears to be losing its attraction to the courts, who are warming once more to concurrent liability. So it is probably wrong to rule out a negligence liability in knowing assistance on the basis that it would overlap with a similar liability in contract or tort, rather than on more substantive grounds. His Lordship's second reason is on the face of it dubious, for it would rule out any negligence liability outside, contract (for example, in case of traffic accidents), and that obviously cannot be right. But perhaps he meant to say that there should be no negligence liability because an undertaking might have been got from the defendant but had not been. Couched thus, his second reason would have been essentially similar to his first, and of the same strength. But it is hard to think that his Lordship did mean this, for he appears to see his second reason as covering virtually every case not covered by his first, and plaintiffs in knowing assistance certainly could not secure undertakings from defendants as frequently as that. Or perhaps his Lordship was reaching for a rather different argument: that in this context, those defendants who give no undertaking of care will nearly always be so peripherally involved in the plaintiff's misfortune that it would be wrong to make them liable for it. This argument too is explicated and justified by Stapleton. She reveals its strength as the manner in which it prevents too great a divorce between a defendant's legal liability and his, individualistically conceived responsibility. She is clear, however, that the argument loses force if one moves away from an individualistic conception of responsibility, or from the view that such responsibility should determine the incidence of liability. In particular, liability against peripheral defendants gains attraction if one takes a more communitarian view of responsibility, whereby society functions better if burdens are shared, or a more welfarist view of the function of liability, whereby plaintiffs' losses are shifted to such defendants as can spread them. Even if the most sympathetic view is taken of Lord Nicholls' two reasons, therefore, they do not constitute an incontrovertible case for his Lordship's view that knowing assistance should be confined narrowly, to cases of dishonesty alone.

These, then, are the salient arguments that have been advanced in recent years about the design of knowing assistance. All of them contend for a requirement of dishonesty, or knowledge. None of them, however, seems ultimately to provide a wholly successful justification for that position. But that is not to say that it cannot be supported. It is, in fact, very intelligible, and can be grounded in some familiar and quite valid concrete concerns.

Consider Barnes v. Addy itself, with its original ruling that liability for knowing assistance requires knowledge on the part of the defendant, and that the trustees' breach should itself have been fraudulent. It is clear from Lord Selborne L.C.'s words that he gave this ruling so as to protect trusts' agents - solicitors, banks, and the like - from excessive exposure to the liability. The role of the knowledge requirement in this project is self-evident, but the requirement of a fraudulent breach by the trustees may also be part of it. At that time, the trustees could have been liable even on the basis of innocent incompetence. Within ten years, Speight v. Gaunt was to change this, by requiring a failure to show the skill of a prudent businessman. The idea, as Jessel M.R. made clear, was that trustees were expected to exhibit only normal behaviour. The exclusion of innocent breach by Barnes v. Addy can be seen as cognate with that, saying that trusts' agents were to be concerned only to avoid embroilment in abnormal behaviour by the trustees, perhaps in that it was only abnormal behaviour which they could be expected to identify. This kindliness of Lord Selborne L.C. towards agents seems to have sprung at least partly from a sense of simple justice. But he also seems to have been concerned not to frighten them away from acting for trusts. There is a connection here too with Speight v. Gaunt, where Jessel M.R. was visibly concerned not only to avoid unfairness to trustees, but also to secure that over-draconian liability did not frighten potential trustees away from that role, and furthermore to facilitate their employment of professional agents. Underlying both decisions is the point that ensuring the supply of good-quality trustees and professional agents is the most important means of securing a good performance of trusts, which is know to have been an important preoccupation in the late nineteenth century.

Perhaps it is less a preoccupation in modern times, in the sense that no worry appears to be felt that the supply is in any real doubt. Nevertheless, the bulk of the authority in knowing assistance remains sympathetic to defendants. In this, it strikes a chord with two other concentrations of doctine, and the broad sweep of the experience may be instructive. First, this treatment of knowing assistance may echo the more sympathetic of the decisions on criminal complicity. As noted above, these seem calculated to spare from frequent liability the ordinary trader, who becomes embroiled in a criminal enterprise as the principal offenders avail themselves of his services. These decisions stand in contrast to others, less sympathetic to the defendant, which generally concern persons who are more specifically and actively involved in criminal activities in concert with the principal offenders. Although there are doubtless exceptions, in the usual scenario of knowing assistance the trust's solicitors, banks and so on should probably be seen as ordinary traders, conducting their ordinary business, whose services happen to be taken advantage of in the commission of a breach by the trustees.

And second, the general modem leniency in knowing assistance may be an aspect of the wide tendency, visible in the last ten or fifteen years, to curb the liabilities of banks and other commercial undertakings. This tendency may arguably be seen too in material such as Caparo Industries Plc v. Dickman, Lloyd's Bank Plc v. Rosset, and Abbey National Building Society v. Cann. R. J. Smith, a commentator hardly given to steep claims, was moved to remark of the reasoning in the last of these that it "comes perilously close to a simple assertion that a mortgagee wins because he is a mortgagee." Such curbing of liabilities evidently cannot be a goal in itself, but must be a facilitator of one or more further goals. The latter presumably include considerations such as minimising the costs of such undertakings, perhaps as a means of encouraging their economic performance, thereby in turn benefiting the economic performance, and so the wealth, of the nation; that, at any rate, was the message of the political rhetoric with which such legal developments were coeval.

Such considerations as these are never negligible, and under the right circumstances may be felt compelling. On the other hand, however, the considerations which conduce towards leniency in knowing assistance in this way can by no means claim a monopoly of wisdom. Sound reasons can certainly be identified why liability for knowing assistance should arise more widely. Even if one is ultimately unpersuaded by these reasons, one cannot achieve a proper appreciation of what is at stake in this area of the law without an understanding of them.

One such consideration is that visible in Bartlett v. Barclays Bank Trust Co. Ltd. This decision established that banks acting as trustees must exhibit not merely the skill of a prudent businessman, but the greater skill of a prudent bank. The reasoning explicitly proceeds on the basis that it is right to hold banks to the high standards of expertise which they themselves profess. One might inquire why, in turn, that should be so. Possible answers might include the point that banks ought morally to deliver value for the money which they are paid; and the point that, since the market works very imperfectly in this context, broadly because it is so difficult for consumers to judge the rival products on offer, its laws require supplementation by legal norms calculated to supply the missing discipline. Although they bear no outward sign of it, it is not inconceivable that the more stringent decisions on knowing assistance may have been influenced by thinking of this kind.

A second such consideration involves seeing it as wholesome that trusts' agents should, via broad liability for knowing assistance, become insurers against trustees' defaults. This vision depends on the reflection that such agents will in their turn simply insure against the liability, passing their costs in doing so back to the trust in the form of their charges, in the usual way. These charges will in effect be borne by the beneficiaries, by way of subtraction from their receipts under the trust. The overall effect of broad liability is thus to compel beneficiaries to insure themselves against depredation by their trustees. It is the same effect as occurs when the law imposes liability in contract by disallowing an exemption clause, whereupon the supplier will insure and pass the cost back to the customer as part of the price paid for the good. Compelling self-insurance in this manner may be regarded as welfarist, in contrast to the individualist approach of minimising liability and leaving beneficiaries and consumers to decide for themselves whether to insure or to bear the risk. Such welfarism is far from absurd, and, as the reference to the law of contract reveals, is indeed quite routine in the law. In the contract context, it may be justified especially by reference to consumers' likely improvidence, so that they are prone not consciously to take that decision, but rather to find themselves bearing the risk by default. The trusts context features substantial numbers of situations where that point could well apply too: in particular, those of pension funds (where the matter is heightened by the fact that the beneficiaries' very livelihood in their old age is at stake) and family trusts.

A third reason why broad liability in knowing assistance might be advantageous is that it would tend to place solicitors, banks and so on in the role of policemen against breach by trustees. It is not difficult to see how the effect would be achieved. When trustees commit depredations on their trusts, their professional agents routinely assist in them, by way of the use made of the latter's services: as when a bank puts through a payment, or a solicitor draws a sale, or an auditor certifies the accounts as correct. if liability were imposed in case of failure to take care in considering the quality of the trustees' activities, such agents would have a duty and more especially a practical incentive to take such care. And given the close sight which such agents' work gives them of the trustees' activities, their taking such care could significantly reduce the danger of depredation by trustees. The advantage in turning trusts' professional agents into policemen in this way, if it can appropriately be done, is that such agents are virtually uniquely placed to raise the alarm; there is no-one else so placed, so if trustees are to be effectively policed at all, it must be by their agents. The question remains, however, whether this course is blocked, in that such agents cannot in the nature of things be cast as policemen. We must examine the possibility that that is so.

The argument cannot of course be that such a role would require such agents to see the invisible. The liability would be one of taking care, rather than achieving the impossible. The argument must be, rather, that such agents are unequipped even to take care in monitoring their trustee clients' affairs. It might be observed, for example, that the mechanised haste of a bank, or the client-orientated role of a solicitor, tends heavily against the taking of such care. That is undoubtedly a weighty point. But it is not the end of the line. The procedures of banks and so on are not pre-ordained, but are the product of the various demands upon them. If one such demand were that they take care against fraud by their trustee clients, there is no reason why their procedures should not reflect it too. Doubtless, the result would involve a degree of compromise, or even of anguish, as the various demands pulled in different directions; that is of course troubling, but nonetheless entirely usual.

And it is clear that policing - moreover, policing of their clients - has indeed become part of professional agents' work, especially in recent years. Consider for example accountants, in regard to their audit function. Their professional standards body, the Auditing Practices Board, articulates its view of proper conduct in the form of published Statements of Auditing Standards ("S.A.S."). S.A.S. 110 and S.A.S. 120, both of which were published in 1995, are relevant. They provide instructions about the appropriate course of action for auditors who become aware of a (possible) fraud or other illegality. But they do not treat it as a matter of happenstance whether auditors do become aware of such things. They stipulate for a culture under which auditors can be expected to do so. In these documents, for example, the Board characterises the correct overall approach for auditors as "an attitude of professional scepticism." To be sure, it records that "it is not the auditors' function to prevent fraud and error," and that "an audit cannot be expected to detect all possible noncompliance with law and regulations." But it proceeds to require auditors to "assess the risk that fraud or error may cause the financial statements to contain material misstatements," and continues: "based on their risk assessment, the auditors should design audit procedures so as to have a reasonable expectation of detecting misstatements arising from fraud or error which are material to the financial statements;" "auditors plan, perform and evaluate their audit work in order to have a reasonable expectation of detecting material misstatements in the financial statements." In the same vein, the Board states that "the auditors should perform procedures to help identify possible or actual instances of [illegality]," which procedures it then goes on to enumerate. It also lists the kinds of situation in which there is thought to be an increased risk of fraud, and to which it therefore expects auditors to be sensitive.

Consider too the position of those persons who assist in, or conceal, or fail to report money laundering operations. Such activities may amount to criminal offences, principally under the Drug Trafficking Offences Act 1986 and the Criminal Justice Act 1988, as amended by the Criminal Justice Act 1993. One of these offences, that of concealing or transferring tainted money, expressly extends to the case where the defendant did not know or suspect the facts, but had reasonable grounds to suspect them. And a number of others, while requiring knowledge or suspicion, place the onus at least partly on the defendant to show that he lacked such knowledge or suspicion. Here too, moreover, the law is not content to leave it to chance whether the duties implicit in these offences are actually discharged. For it goes on to require the undertakings likeliest, to encounter money laundering operations - banks, solicitors etc. - to maintain procedures designed to make them aware of any such operations being carried on within their orbit. In addition, various groups of undertakings within this sector have issued guidance documents which recognise their responsibility, if not to detect all nefarious operations in their vicinity, at any rate to be vigilant against these and to maintain procedures (which are articulated) calculated to assist in that regard.

These matters should not be treated as merely isolated issues. On the contrary, they figure significantly in the conduct of professional agents' daily work. Nor should they be regarded as peripheral to the present study. On the contrary, again, the factual situations to which they relate clearly include breaches of trust, above all breaches on a large scale, such as the depredation of pension funds and felonious activities by company directors. It may be seen, then, that professional agents' responsibilities in regard to such phenomena already demand a large measure of the proactive wariness which would be required of them if liability for knowing assistance arose in cases of failure to take care as well as of knowledge or dishonesty. It appears to follow that such an approach to knowing assistance is entirely contemplable. And that allows us to return to the argument, above, that it is desirable to formulate knowing assistance so as to require trusts' professional agents to police the trustees' activities, because they are uniquely well-placed to do so.

None of these considerations goes to show, of course, that liability for knowing assistance must be formulated in an onerous way, rather than in the more lenient way approved in most of the modem authorities, for it was seen earlier that there are palpable considerations supporting the latter approach too. It undoubtedly does show, however, that the option is very much there. With respect, therefore, it is not satisfactory to assert, as Millett J. has recently done, that "account officers are not detectives", and to continue that "unless and until they are alerted to the possibility of wrongdoing, they proceed, and are entitled to proceed, on the assumption that they are dealing with honest men". That may once have been a sustainable point of view (though it was never more than that). But today, as we have seen, the money laundering legislation has rendered such statements simply erroneous, and any bank or similar entity which did not train its staff on the contrary basis would ipso facto be in breach of the law.

One's choice between the more lenient and the more demanding position will depend on the weight which one attaches to the group of considerations underlying each of them. As to that, it may be commented, perhaps that in the times in which we live, it cannot be denounced as folly to conclude that those favouring broad liability are the more impressive. Major depredation by fiduciaries appears to be a disease of our age, and one might quite rationally think it more important to attend to that than, say, to minimise commercial costs, important though one could simultaneously concede the latter to be. Interestingly, however, when the Goode Committee, set up in the aftermath of the Maxwell affair, recommended a new regime for pension funds which nodded (some would say far from excessively) in the direction of securing such funds' safety, the ensuing Government proposals and legislation moved back towards laxity, expressly so as to save companies from the costs which would have been entailed.

The arguments for exigence in knowing assistance are considerably less strong as regards defendants other than professional agents. Such defendants, for example, have no advertised standards to live up to, or easy means of spreading losses so as to be a focus of welfarist liability, or obvious suitability for a policing role. The difference is not polar, however. Lay persons are in fact subject to the same liability as professionals in respect of money laundering, and, as noted earlier, that can extend to the case where there are reasonable grounds for suspecting what is going on. Perhaps the point is that money laundering is such a grave social evil that everyone must help guard against it. As just noted, it is not unarguable that depredation of at least some kinds of trusts should be regarded similarly. But a difference between lay and professional surely does remain. There is no reason why the rules governing liability for knowing assistance should not reflect this difference, just as there are different standards for lay and professional defendants in other areas.

If an exigent liability for professional agents were favoured, how would it be expressed? One's immediate response might be to use the term notice, or negligence. But remember that the cases interpret notice very leniently in contexts other than land transfer, taking it to mean that the defendant is liable only if he knows the facts or if they would have been patently apparent to the reasonable person in the circumstances. Negligence has been somewhat similarly attenuated, by the cases which, to identify the standard required, look to the behaviour of other practitioners in the same field: a test which is often undemanding, certainly as regards financial entities. It would be regrettable if this approach meant that the concepts of notice and negligence could not be of assistance here. In fact, they can probably be salvaged. Their fundamental reference is to the standard which is reasonably to be expected in the circumstances. They are sometimes taken leniently because it is felt that, in the circumstances in question, not a great deal is reasonably to be expected. But the reasonable expectation may change: notably, because the social culture surrounding the particular activity becomes more exigent, as it was suggested above has happened to the culture of the financial sector. In that event, the content of the concepts of notice and negligence ought to be able to follow suit.

If it were felt right to maintain a rule imposing relatively exigent notice or negligence liability upon professional defendants, how might the law signal that lay defendants should be treated more leniently? The most obvious way would be to require knowledge or dishonesty for them. This approach might be less than ideal, however. It would require each and every defendant to be categorised one way or the other - lay or professional - when the natural place of some might be near the line. And, more important, its message would be that lay defendants need not even take as much care as can reasonably be expected of them, when we might well feel (as indeed the money laundering legislation provides) that lay persons too should play their part, albeit that it can only be a small part, in guarding against depredation of trusts. In fact, these concerns could be met, compatibly with requiring less for lay than for professional defendants, by requiring notice or negligence across the board after all. Negligence is routinely given different significances for different kinds of defendants in this way. There is no reason why the variability of notice could not similarly be used to refer to the standards of a reasonable person or entity of the same kind as the defendant.

Knowing receipt

The most recent juridical material on knowing receipt tends to the view that notice, rather than knowledge, is required. (Though, as we have seen, the significance of "notice" varies with the context, and in some judges' eyes may sometimes equate with knowledge.) One searches the literature somewhat in vain, however, for a principled apologia for that position. The two dominant arguments lie to either side of it: to the effect that knowledge alone should suffice, on the one hand, or that neither knowledge nor notice should be required, on the other.

The argument that knowledge alone should suffice is based on the idea that knowing receipt is about wrongfully causing loss to trusts, and the further idea that wrongfulness requires knowledge. This argument is not strong, however. It is incorrect to say that knowledge alone will show a defendant to have acted wrongfully: he can in principle equally be said to act wrongfully if he is negligent, or has notice. Moreover, it is questionable whether knowing receipt is about wrongfully causing loss at all. There may be more than one other thing that it could be about, but most modem opinion takes it to be a restitutionary liability, based on the fact that the defendant has acquired the plaintiff's property.

It is the latter view that underlies the other dominant argument, that neither knowledge nor notice should be required for knowing receipt. The leading proponent of this view is Birks. Birks contends that the point of knowing receipt is to restore to the plaintiff the value of property lost by him to the defendant, and by which the defendant was unjustly enriched. And that can be the case, he says, even if the defendant had neither knowledge nor means of knowledge of the plaintiff's rights: it is enough that the defendant has indeed had the benefit of the plaintiff's property. Birks allows two palliatives against the strictness of this position, however. First, he recognises that a defence against liability may be available to a bona fide purchaser of the property for value without notice of the true position, and that reference to notice (though not to knowledge) may be condoned to that extent. (So he argues that the cases supporting a requirement of notice may be acceptable as discussions of the defence of bona fide purchase for value without notice, the defendants in them happening to be purchasers for value.) Secondly, he argues that the defence of bona fide purchase should be either replaced or joined by a defence of change of position. Under this, a defendant who acts to his detriment on the faith of the acquisition, innocently of the plaintiff's right to reclaim it, is to that extent freed from liability. Although the defence of change of position has only recently been recognised in English law, and is as yet somewhat uncharted territory, it seems clear that the presence or absence of something in the way of notice or knowledge appears relevant to the issue of innocence. These two palliatives do not cover the whole ground, however, and so according to Birks at any rate some volunteer recipients would be liable even with no knowledge or notice.

This argument seems strong. At the very least, it can claim to handle cases of unjust enrichment at the expense of a trust in the same fashion as cases of unjust enrichment at the expense of a straightforward owner of property, which seems commendable. One can see the same point in another way. A general principle of restitution for unjust enrichment is now recognised to exist in English law, by virtue of the decision of the House of Lords in Lipkin Gorman v. Karpnale Ltd. It is of strict liability, subject to defences such as bona fide purchase and change of position. There is no obvious reason why it should not apply to cases in which the loss is to a trust, as much as to any others. So if knowing receipt is not taken thus, it will be undercut and made irrelevant anyway.

Sir Peter Millett, writing extrajudically, has taken a similar position to Birks. In El Ajou v. Dollar Land Holdings Plc, moreover, he made statements which appear to lend his judicial authority to this view. Yet both there, and in his earlier decision in Agip (Africa) Ltd v. Jackson, he appeared ultimately, whilst rejecting a requirement of knowledge, still to require notice. The following explanation of this apparent oddity may be ventured. Whilst Sir Peter's preferred position was one of strict liability, this position was unattainable by him in his judicial capacity as a result of the precedents binding upon him and of the argument conducted before him. But requiring only notice took him closer to his preferred position than requiring knowledge would have done. It would go too far to term this insight a principled case for a requirement of notice, but it does afford an understandable explanation for it.

Should we conclude, then, that the best that can be said of a requirement of notice in knowing receipt is that it is a staging-post on the road towards strict liability and a component of a possible, but eclectic, defence, and that a requirement of knowledge for the most part lacks even these redeeming features? It must be said immediately that that conclusion is eminently possible. Perhaps one countervailing point may be considered, however. A regime of strict liability with a defence of change of position may be conceptually very pure. An additional defence of bona fide purchase for value may muddy the details somewhat, but it does not contaminate the essence of the regime, in the way in which a wholesale requirement of notice does. But is this purity too refined for the realities of life, and does the traditional requirement of notice serve to reconcile the law with the latter?

Consider some examples. Say I acquire trust money, without giving value and with no notice that it is such, and, still without notice, spend it on taking a round-the-world cruise, a luxury which I had always assumed far beyond my pocket. Here I appear to have a defence of change of position. Say, on the other hand, that I instead spend the money on paying various routine bills. Here I appear not to have the defence. The difference between the two cases is, or at any rate is intended to be, this. In the first case I am acting to my detriment on the faith of my resources having been increased, in that I would not have bought the cruise otherwise. But in the second I am not so acting, in that I would have paid the bills anyway. To put the point another way, in the first case I have incurred additional expense, whilst in the second case I have merely saved my own money. So in the second case it is, whereas in the first case it is not, fair to require me now to reimburse the trust.

Now the facts in these examples are very polarised, and their proper consequences accordingly very easy to discern. In real life, however, it may very often be extremely difficult to characterise the facts so confidently. For many people, much if not all expenditure is surely made neither quite on the faith of having exactly such-and-such a level of resources, nor quite regardless of their level of resources. For one thing, in most of their expenditure they have some room for manoeuvre, but they are also under broad pressures. This is actually true of most holidays and most bill-paying, at any rate when it is remembered that the question is not simply whether to make the expenditure at all but rather whether to make it now or at some later date. And for another, they may rather rarely, in deciding whether to make some item of expenditure, consciously match it against their perception of their resources, though no doubt they do try to ensure, or perhaps just as realistically hope, that income and expenditure will balance out over time, and they are no doubt aided in their decisions about individual transactions by a subconscious sense of their broad position. For that matter, their saving money is often to be regarded not simply as not spending it, but as part of an effort to reconcile their income and expenditure in the longer term. The reflection that an anticipated expense will be catered for by their savings may figure in their deciding to go ahead with some other outlay. But, again, the "reflection" may be subconscious, and neither-piece of expenditure either quite a luxury or quite unavoidable. It must of course be recognised that such a vague approach to spending money would be an impracticable luxury for very many other people, who have to think hard about every penny. It may be suggested, however, that these people would dearly like to be in a position to adopt such an approach too. It therefore represents an interest to be taken seriously even for them.

So real life tends to feature fact situations much cloudier than those often posited as examples of the workings of change of position. The defence can no doubt be made to deal with them by letting the onus of proof take the strain. If the onus is on the defendant to prove the defence, as would be the usual position, the effect will be that the defendant who proceeds in the rather unrigorous manner just sketched above will not have the defence, because he will not have shown that he acted to his detriment on the faith of the acquisition. We may, however, consider whether such a defendant deserves not to have a defence, and so find ourselves wondering whether change of position should represent his only relevant chance of one. It can still be said that in a broad sense, one too broad for change of position, this defendant ran his affairs on the basis of being able to assume a certain level of wealth with reasonable safety. And it may be felt that the ability securely to act in this way is an interest well worth protecting. If that is accepted, the implication appears not to be the creation of a defence structured, as change of position is, in terms of the relationship between an individual action and a particularised belief. It is that very kind of structure which is inappropriate to the underlying issue. Rather, the implication seems to be more or less to assume the importance of security in wealth to people's activities (unless, conceivably, in an aberrant individual case this can be disproved), and to protect that security as such. In other words, to rule that once an incoming asset has been bona fide assimilated into the recipient's coffers, it becomes unrecoverable. And it so happens that this is, in effect, the approach traditionally taken in knowing receipt, where so long as the defendant lacks notice (or indeed, according to some, knowledge) he is not liable.

On the other hand, it is certainly hard to see any good reason why the particular sphere to which knowing receipt applies should command a radically different treatment from that which obtains in the surrounding area of property and restitution. And the approach just outlined does seem to be so radically different. Its implication is to negate very many of the plaintiff's claims which are currently protected by the law of property and of restitution, where the statement "that asset is (or was) mine" is allowed great power to ordain that the defendant be deprived, however uncomfortably for him. So the argument cannot be that the traditional approach to knowing receipt should be preserved, if this counter-approach is also to remain. It is rather that knowing receipt provides, albeit at the price of some incoherence with the surrounding area, a reminder that a recipient's interest in broad security of wealth is not inconsequential; and that it is not inconceivable for the law to lean towards protecting that interest. Interestingly, in fact, that leaning may be better established in the law than has generally been recognised. Above all, it may be reflected in the defence of bona fide purchase for value without notice, which lies very widely indeed in this area. This defence is not the same in its impact as change of position, not least because the purchase price paid may be lower than the value of the asset which the purchaser thereupon becomes entitled to retain; and more fundamentally, its rules articulate no stress upon the relationship between the two. On the contrary, the defence of bona fide purchase seems very much to protect the purchaser's interest in being able to assume the security of his acquisition. Seen from this point of view, indeed, the fact that the bona fide purchase defence is confined to cases of purchase for value, whilst the knowing receipt rules also protect the recipients of gifts, may appear more a difference of degree than one of root philosophy.

Let us take it, however, that the more rigorous system envisaged by Birks will be accepted, so that knowing receipt is recast (or absorbed into restitution generally) as a strict liability to reimburse subject to a defence of change of position, either alongside or in place of a defence of bona fide purchase. In fact, this system does not remove all recourse to tests of notice or knowledge. These clearly arise in the defence of bona fide purchase, if it is to remain, but they also arise in change of position, in the question whether the defendant acted to his detriment on the faith of the acquisition. The question therefore remains of the approach to be taken over them: whether they should be demanding or lenient towards defendants. The argument for treating knowing receipt as a strict liability, with defences, seems to contain no message as to this question. It is apparently aimed purely at the organisation of the subject, and as such, in particular, contains no implication to the effect that the "bottom line" position - the chance of liability actually being incurred, after both the nature of the action and the defences are considered - should itself necessarily tend towards strictness. What considerations, then, should matter here?

In favour of liability being relatively infrequent, i.e. in favour of stipulating for knowledge or at any rate taking a lenient approach to notice, there is clearly the argument that purchasers should be protected. This is an immensely familiar consideration in the law. Its nub is perhaps not so much a tenderness to purchasers per se, as a wish to obviate the damage to trade which would be done if purchasers were less well protected. Broadly, this policy operates in two ways. One is by limiting the number of obligations to which purchasers may become subject even if they know of them. This manifests itself in such institutions as overreaching of trust interests upon sale; the privity rule by which a contractual duty binds only the promisor; and the limited canon of property rights. It may be seen in action in decisions such as Lloyds Bank Plc v. Rosset. It is not directly in point here. The other way in which the policy operates, which is directly in point here, is by minimising the difficulty - and so the cost - which a purchaser would encounter in discovering any obligations which might bind him. This manifests itself in the rules for the registration of encumbrances over land. So far as knowing receipt is concerned, its logic is to call for a rule whereby a purchaser need trouble only with those facts which he knows about or which are patently obvious, and is safe in inquiring no further. It seems clear that this consideration was present to the minds of the judges who laid down that, outside the land transfer context, the notice element of knowing receipt was to be taken narrowly, whether as a requirement of knowledge or as a recognition that the reasonable person would take but little care. Indeed, many of the authorities taking this approach base themselves upon the famous dictum of Lindley L.J. in Manchester Trust v. Furness, where the needs of commerce were Lindley L.J.'s express concern.

This argument evidently has no application where the receipt is by a volunteer. But an economic argument, to the same effect, can also be made as regards that case. It draws on the idea of the important interest in broad security of wealth, described above. That interest is, by definition in this context, pitched against the interest of the plaintiff trust in recouping its lost property. So it cannot expect absolute protection. In favour of its being granted substantial protection, however, is the point that a person who enjoys, and perceives himself to enjoy, broad security in his wealth will be more ready to spend it (other things being equal) than one who feels it necessary to look over his shoulder before doing so, so as to check that it is really his to spend - in our context, that it cannot be reclaimed from him by a trust. The former state of affairs thus allows the market to approach full efficiency more closely than the latter, and from an economic point of view is therefore to be preferred. The implication for our context is that change of position should protect even voluntary recipients by not being negated by anything less than knowledge, or the very easy means of knowledge, of the trust's claim to the property.

A contrary consideration, pointing to more exigent standards, is the thought that recipients of property, especially purchasers, might advantageously be recruited to help police the possible nefarious activities of trustees. This shares common roots with the similar thought that there should be an exigent standard in knowing assistance, denoting that defendants there should help to police these activities. As regards knowing receipt, the idea would be that if recipients could incur liability in this wide manner, they, presumably via their professional advisers, would take the greater pains to establish the rectitude of the transaction, and in the process would be the more likely to bring to light any transgression by the trustees.

Whilst this consideration has its attractions, it is perhaps less compelling than its counterpart in knowing assistance. In knowing assistance, we were largely focusing on persons and entities - banks, solicitors, accountants, etc. - who, first, will almost inevitably know that they are dealing with trustees, and, secondly, will also be accustomed to a policing role from other aspects of their work, as we saw. In the context of knowing receipt, however, the target group would simply be purchasers, or even more broadly recipients of property. This group is much more diffuse, and lacks the features, just identified, which make it relatively easy to imagine trusts' agents as policemen. It more closely resembles the class of nonprofessional defendants in knowing assistance, who were acknowledged to have a much stronger claim to a lenient standard. Moreover, even if recipients could successfully be recruited as policemen in this way, the required changes in market practice would probably raise transaction costs quite substantially, and so do perceptible economic damage. The avoidance of the latter was one of the considerations favouring narrow liability in knowing receipt, noted above.

To be sure, the economic arguments have never swept all before them. The major reminder of this is the experience in the context of land transfer. Here, recipients have traditionally been expected to put considerable effort into verifying that all is well with their acquisition, and have not been accorded broad security in it. Certainly, the overall tendency of the last century or so has been to reflect economic considerations, above all by the development of registration. But even within this, a stickier attitude has at times re-emerged, especially in contexts where the interest at stake is a particularly favoured one, such as a person's home: see for example Williams & Glyn's Bank Ltd v. Boland. The emphasis returned to economic efficiency, however, in Abbey National Building Society v. Cann.

The land transfer experience valuably serves to remind us that the chosen rule must necessarily be a compromise between the opposing considerations, and also that it cannot seriously be expected ever to be fixed once for all. That said, however, it would be a little surprising in our contemporary culture if the economic arguments did not in general get the upper hand. So far as knowing receipt is concerned, the implication of this is that liability should be imposed sparingly (or, on the other way of looking at the matter, that its defences should arise generously). And that in turn should mean that the liability should arise (or not be negatived) only where the defendant knew of the impropriety, or where the reasonable person would have known of it from the very face of the matter. Which is, substantially, the position in fact arrived at, albeit by slightly different analytical routes, in all the recent authorities.

CONCLUSION

In reviewing developments regarding knowing assistance and knowing receipt since Harpum's article of 1986, we may conclude, despite perhaps unpromising first impressions, that much has been achieved. The concepts from which the rules may be assembled have been more thoroughly explored, and seem better understood. The rules themselves, especially regarding knowing assistance, are more certain than they once were. Above all, a number of important attempts have been made to express the principles upon which these liabilities are or should be founded, and to design the rules accordingly.

These attempts generally give the impression of confidence in their reasoning and conclusions. Even the arguments which explicitly look to policy often have a zealous quality about them. For example, Birks' remarks, about knowing assistance, that "it is undesirable that ... another shadowy liability should haunt commercial life" and that "it is unlikely that the ... view ... which requires only negligence ... will ever prevail again". Even Lord Nicholls, whose judgment in Royal Brunei Airlines Sdn. Bhd v. Tan acknowledges a range of points of view, and whose reasoning is persuasive rather than compulsive, ultimately seems to find himself positing a particular rule as ineluctably right: "their Lordships consider that dishonesty is an essential ingredient here ... Beneficiaries cannot reasonably expect that all the world dealing with their trustees should owe them a duty to take care lest the trustees are behaving dishonestly."

The present article should prompt the thought that such confidence is misplaced. It has reviewed the different kinds of impact which different designs of these liabilities would give them, and argued that a good case can be made for more than one of the possibilities. This should be quite unsurprising. Our society is a complex and compromised one on any front. When dealing with the protection of interests as sophisticated as those under trusts (as opposed, for example, to life and limb), we are least likely to find answers in the shape of transcendent verities: any position will be close to the line. And we are concerned here largely with financial dealings, a phenomenon as regards which orthodoxies are particularly short-lived. A pluralistic appreciation is therefore indispensable.

This should not cause pessimism, however, even from the point of view of trying to achieve stability in the law. On the contrary, there are worse problems in the excessively confident espousal of a favoured position. Even if the position in question can be made to stick in the short term, it is over time likely to prove an example of hubris. For it will find it hard to co-exist with those concerns which it does not reflect, but which remain operative, and which will continue to press for legal recognition. Greater stability is in fact likely to be had from an approach which recognises that there are competing arguments, that they all have their strengths and weaknesses, and that no rule which we may adopt can capture all this more than partially, and temporarily. Perhaps, indeed, an under-recognition of this has contributed to the notoriously dissolute condition of the law regarding knowing assistance and knowing receipt over the years. Perhaps, ultimately, we should stop trying to embed our thoughts about this material into firm rules, and contemplate other ways of giving them legal expression. Millett J. may have made a profound point when, in a discussion of knowing assistance in Agip (Africa) Ltd v. Jackson, he said that the incidence or otherwise of liability "is essentially a jury question". Not that one would readily look to jury trial; but the jury is a powerful emblem of the law's accommodation with the relativity and mutability of social values, and of its acceptance of contingency in its import.

Here is information on how to order Drafting Trusts and Will Trusts and other books by James Kessler QC.

Home