LORD WALKER (with whom Lord Neuberger, Lady Hale, Lord Mance, Lord Clarke, Lord Sumption and Lord Carnwath agree)

Introduction

1. These appeals raise important and difficult issues in the field of equity and trust law. Both appeals raise issues about the so-called rule in Hastings-Bass. One appeal (Pitt) also raises issues as to the court's jurisdiction to set aside a voluntary disposition on the ground of mistake. It is now generally recognized that the label 'the rule in Hastings-Bass' is a misnomer. The decision of the Court of Appeal in In re Hastings-Bass, decd [1975] Ch 25 can be seen, on analysis, to be concerned with a different category of the techniques by which trust law controls the exercise of fiduciary powers. That decision is concerned with the scope of the power itself, rather than with the nature of the decision-making process which led to its being exercised in a particular way (see R C Nolan, Controlling Fiduciary Power [2009] CLJ 293, especially pp 294-295 and 306-309). The rule would be more aptly called 'the rule in Mettoy', from the decision of Warner J in Mettoy Pension Trustees Ltd v Evans [1990] 1 WLR 1587. But the misnomer is by now so familiar that it is best to continue to use it, inapposite though it is.

2. As Mettoy illustrates, the rule is concerned with trustees who make decisions without having given proper consideration to relevant matters which they ought to have taken into consideration. It has also been applied to other fiduciaries (in Pitt Mrs Pitt was acting as a receiver appointed by the Court of Protection). Mettoy was concerned with the rules of an occupational pension scheme, as are some other cases on the rule. But since the turn of the century there have been several cases concerned with family trusts, and in particular with tax“planning arrangements involving trusts, where the arrangements have for one reason or another proved unexpectedly disadvantageous, and the court has been asked to restore the status quo ante under the Hastings-Bass rule.

3. Futter is such a case, as Norris J pointed out in blunt terms at the beginning of his judgment, [2010] EWHC 449 Ch, [2010] STC 982, para 2:

'This is another application by trustees who wish to assert that they have acted in an un-trustee-like fashion and so have failed properly to exercise a power vested in them. The trustees wish to take advantage of this failure to perform their duties in order to enable the beneficiaries to avoid paying the tax liability consequent upon the trustees' decision. Put like that (and I am conscious that that is not the only way in which the situation may be described) the possibility is raised that the development of the rule may have been diverted from its true course.'

These appeals are the first cases on the Hastings-Bass rule in which the Commissioners of HM Revenue and Customs ('the Revenue', so as to include their predecessors, the Commissioners of Inland Revenue) have been joined as parties in the proceedings. It is the Revenue that has taken on the task of challenging, if not the existence, at least the limits of the Hastings-Bass rule. It is no coincidence that the judgment of the Court of Appeal in these two appeals (which were heard together in that court also) is the first fully considered judgment above first-instance level, and the first to come on further appeal to the Supreme Court (Mettoy was not cited to the Court of Appeal in Stannard v Fisons Pension Trust Ltd [1991] Pen LR 225, discussed in para 34 below).

4. Rescission of a voluntary disposition on the ground of mistake is, by contrast, a topic on which there is a good deal of authority, including a decision of the House of Lords, Ogilvie v Allen (1899) 15 TLR 294. But some of the authorities are quite old, and others are debatable. There has been much discussion of the distinction drawn by Millett J in Gibbon v Mitchell [1990] 1 WLR 1304, 1309, between a relevant mistake having to be 'as to the effect of the transaction itself and not merely as to its consequences or the advantages to be gained by entering into it.' So here too review by the Supreme Court is appropriate.

5. This court has therefore had to consider a large volume of case-law, culminating in the judgment of Lloyd LJ in the Court of Appeal in these appeals: [2011] EWCA Civ 197, [2012] Ch 132. That judgment, described by Longmore LJ, para 227, as 'remarkable', and by Mummery LJ, para 230, as a 'very fine comprehensive and clarifying judgment', runs to 226 paragraphs. I share their admiration, and I agree with Lloyd LJ's main conclusions as to the scope of the Hastings-Bass rule, and the outcome of the appeals on that issue. But I will say at once that I take a different view of the disposal of the appeal in Pitt on the mistake issue.

6. Before any detailed consideration of the case-law it may be helpful to identify, in general terms, some of the principal topics in the appeals. It has often been said (for instance, by Norris J in Futter, para 21) that the rule in Hastings-Bass is not founded in the law of mistake, and in his judgment Lloyd LJ dealt with them as almost completely separate topics. They do cover different areas, in that the Hastings-Bass rule is restricted to decisions by trustees and other fiduciaries, and does not necessarily require the decision-maker to be under a positive misapprehension: mere absence of thought may be sufficient. The court's wider jurisdiction to rescind a transaction on the ground of mistake is not limited to transactions entered into by fiduciaries, and does generally require there to have been something that can be identified as an operative mistake. The significance of fault in the error or inadvertence is a further point of distinction.

7. Nevertheless there is a degree of overlap between the two principles in their practical application. In some of the first-instance cases on the Hastings-Bass rule judges have drawn attention, with evident surprise, to the absence of any alternative claim for relief by way of rectification or rescission on the ground of mistake. In some of the cases (such as Abacus Trust Co (Isle of Man) v Barr [2003] EWHC 114 (Ch), [2003] Ch 409, the facts of which are summarized at paras 36 and 37 below) rescission on the ground of mistake would seem to have been the natural remedy for the trustees to seek. There must be some suspicion that reliance on the Hastings-Bass rule has come to be seen as something of a soft option, or at any rate as a safer option, at a time when it was supposed, wrongly, that the application of the rule did not require the granting of a remedy which was discretionary in the sense that it might be withheld because some equitable defence was established.

8. The way in which the law seemed to be developing, especially in cases concerned with unsuccessful tax-planning arrangements, led one legal scholar (Professor Charles Mitchell, Reining in the rule in In Re Hastings-Bass, (2006) 122 LQR 35, 41-42) to ask:

'Why should a beneficiary be placed in a stronger position than the outright legal owner of property if he wishes to unwind a transaction to which he has given his consent, but which turns out to have unforeseen tax disadvantages?'

Professor Mitchell went on to comment, presciently:

'The courts will have to look elsewhere for the means of reining in the rule in Re Hastings-Bass, most probably to the equitable bars to unwinding a transaction that would come into play if it were decisively recognised that the rule renders transactions voidable rather than void.'

This court now has the opportunity of confirming the Court of Appeal's recognition of that essential point.

THE HASTINGS-BASS RULE

The three strands of the problem

9. In the Court of Appeal [2012] Ch 132, para 227 Longmore LJ described the appeals as

'. . . examples of that comparatively rare instance of the law taking a seriously wrong turn, of that wrong turn being not infrequently acted on over a 20-year period but this court being able to reverse that error and put the law back on the right course.'

If the law did take a seriously wrong turning it was because a number of first-instance judges were persuaded that three separate strands of legal doctrine, all largely associated with practice in the Chancery Division, should be spun or plaited together so as to produce a new rule.

10. The first strand of legal doctrine starts with the entirely familiar proposition that trustees, in the exercise of their fiduciary discretions, are under constraints which do not apply to adult individuals disposing of their own property. I made some uncontroversial observations about this in Scott v National Trust for Places of Historic Interest or Natural Beauty [1998] 2 All ER 705, 717:

'Certain points are clear beyond argument. Trustees must act in good faith, responsibly and reasonably. They must inform themselves, before making a decision, of matters which are relevant to the decision. These matters may not be limited to simple matters of fact but will, on occasion (indeed, quite often) include taking advice from appropriate experts, whether the experts are lawyers, accountants, actuaries, surveyors, scientists or whomsoever. It is, however for advisers to advise and for trustees to decide: trustees may not (except in so far as they are authorised to do so) delegate the exercise of their discretions, even to experts. This sometimes creates real difficulties, especially when lay trustees have to digest and assess expert advice on a highly technical matter (to take merely one instance, the disposal of actuarial surplus in a superannuation fund).'

The same principles apply, at least in a modified manner, to other persons acting in a fiduciary capacity.

11. There are superficial similarities between what the law requires of trustees in their decision-making and what it requires of decision-makers in the field of public law. This was noted by the Court of Appeal in its judgment, delivered by Chadwick LJ, in Edge v Pensions Ombudsman [2000] Ch 602, 628-629. It was also noted by Lord Woolf MR in Equitable Life Assurance Society v Hyman [2002] 1 AC 408, para 20. The analogy cannot however be pressed too far. Indeed it was expressly disapproved by the Court of Appeal in these appeals (Lloyd LJ at para 77 and Mummery LJ at para 235). In Abacus Trust Co (Isle of Man) v Barr [2003] Ch 409, para 29 Lightman J identified three important differences as the discretionary nature of relief on judicial review, a different approach to nullity, and strict time limits.

12. The second strand is that a voluntary disposition (typically a gift, outright or in settlement) may be set aside on the ground of mistake. As already noted, this branch of equitable jurisdiction is distinct from the Hastings-Bass rule, but similar issues arise as to the nature and gravity of the relevant error or inadvertence, and in practice they sometimes overlap. The mistake jurisdiction was considered as a separate issue in paras 164 to 220 of Lloyd LJ's judgment. He identified the correct test as derived in part from the judgment of Lindley LJ in Ogilvie v Littleboy (1897) 13 TLR 399, 400 (approved by the House of Lords as Ogilvie v Allen (1899) 15 TLR 294), a case which 'emerged from the shadows to be cited to the court' after a century of obscurity. He also considered recent decisions including Gibbon v Mitchell [1990] 1 WLR 1304 and In re Griffiths decd [2008] EWHC 118 (Ch), [2009] Ch 162.

13. The third strand of legal doctrine, and the most abstruse one, is concerned with the partial validity of an instrument which cannot be entirely valid because it infringes some general rule of law. It is an issue which arises, often under the rubric of severance, in many different areas of law. One example is contract law, especially in the context of illegal restraints on trade (see the judgment of Jonathan Sumption QC in Marshall v NM Financial Management Ltd [1995] 1 WLR 1461, upheld by the Court of Appeal [1997] 1 WLR 1527). Another example is bye-laws held to be partly ultra vires (see the speech of Lord Bridge in Director of Public Prosecutions v Hutchinson [1990] 2 AC 783).

14. In the field of trust law the most common invalidating factor, until the Perpetuities and Accumulations Act 1964, was the unreformed rule against perpetuities, or remoteness of vesting. This applied relentlessly both to dispositions of property made by settlors or testators of property at their free disposal, and to dispositions made in the exercise of special (that is, restricted) powers of appointment over settled property. Special powers of appointment might be exercisable either by individual donees (for instance, by a parent with a life interest in favour of children with interests in expectancy) or by the trustees as a body. But in either case the power could be exercised only within the limits, and for the purposes, marked out by the donor of the power. And in either case the interests appointed had to conform to the rule against perpetuities as it applied to lives in being at the time of the creation of the power (that is, the date of the original settlement, or the date of the testator's death).

15. These matters were once familiar (indeed, elementary) to almost all chancery practitioners. Law and practice at the chancery bar have moved on. The rule against perpetuities has lost its terrors since the Perpetuities and Accumulations Act 1964 (which was almost completely non-retrospective) gradually came to apply to more and more trusts, followed by the Perpetuities and Accumulations Act 2009. Family trusts are now a shrinking enclave designated as 'private client' work, and pensions trusts, burdened by increasingly complex regulatory statutes, are another enclave reserved for pensions specialists. But in order to investigate the origins of the disputed rule in In re Hastings-Bass it is necessary to revisit, without much nostalgia, this area of trust law as it was about 50 years ago. There was a body of fairly arid case law, now almost entirely obsolete, about the validity of interests in settled property which were 'ulterior to' but not 'dependent on' antecedent interests which infringed the rule against perpetuities. In re Hubbard's Will Trusts [1963] Ch 275 and In re Buckton's Settlement Trusts [1964] Ch 497 are examples from just before the enactment of the reforming statute. In re Abrahams' Will Trusts [1969] 1 Ch 463 and In re Hastings-Bass, decd [1975] Ch 42, discussed below, can be seen as a final chapter in that case law.

16. There is one further background matter to be noted. Under traditional family settlements, when the modern type of discretionary settlement was still fairly rare, the most common dispositive power exercisable by trustees was the power of advancement. This is a power to accelerate the interest of a beneficiary interested in capital, exercisable with the consent of any beneficiary with a prior interest (typically a parent with a prior life interest). Such powers were so much common form that section 32 of the Trustee Act 1925 provided a default power, which could be excluded or (as often happened) extended by the trust instrument. The power was typically exercisable by a payment or transfer 'to or for the advancement or benefit' of the beneficiary.

17. In In re Pilkington's Will Trusts [1964] AC 612 the House of Lords, differing from the judge on one point and from the Court of Appeal on another, held that a power in those terms could (in principle, and apart from the rule against perpetuities) be exercised for the benefit of a minor beneficiary (the testator's nephew's daughter, who was only two years old when the proceedings started in 1959) by a transfer of up to half of her expectant share, with her father's consent, to the trustees of a new settlement under which she would attain a vested interest in capital at 30. This would lawfully avoid estate duty on her father's death if he lived for a further five years. But the House of Lords also held that the new settlement must, for the purposes of the rule against perpetuities, be treated as if it were an appointment made under a special power conferred by the testator's will. The trusts of the new settlement did not meet that requirement as the child was not a life in being at the testator's death in 1935. But valid trusts to much the same effect could have been achieved (and may eventually have been put in place) by referring to the alternative contingency of survival until 21 years after her father's death, as he was alive in 1935.

Vestey, Abrahams and Hastings-Bass

18. In the Court of Appeal Lloyd LJ correctly identified the decision of the Court of Appeal in In re Vestey's Settlement [1951] Ch 209 and that of Cross J in In re Abrahams' Will Trusts [1969] 1 Ch 463 as the most important precursors to the decision of the Court of Appeal in In re Hastings-Bass [1975] Ch 25. Lloyd LJ analysed these three cases very thoroughly at paras 33 to 67 of his judgment. Because his analysis is so full and accurate I can deal with the cases more briefly, especially as to the facts. It is worth noting that although all three cases had an important tax element, in each case the trustees' misunderstanding was not about tax law. It was about the general law: in the first case about the effect of section 31 of the Trustee Act 1925, and in the other two about the effect of the rule against perpetuities.

19. In Vestey the trustees of a large settlement made by Lord Vestey and his brother Sir Edmund Vestey exercised their discretion over the allocation of income with the apparent intention of income being accumulated during the minorities of a number of beneficiaries. They set out to do this by a sort of 'framework' resolution that income should 'belong' to the minor beneficiaries in specified shares, followed by further half-yearly resolutions to the effect that income was not required for the beneficiaries' maintenance, and should therefore be accumulated under section 31 of the Trustee Act 1925. The difficulty was that the language of section 31 did not really fit such a situation.

20. At first instance Harman J held that the resolutions were ineffective. That result would have avoided surtax but left the income in limbo (Evershed MR's suggestion in the Court of Appeal that the income would have been held on a resulting trust for the settlors seems, with respect, very doubtful). But the minor beneficiaries appealed, and the Court of Appeal gave effect to the framework resolution, treating the references to accumulation under section 31 as peripheral. Evershed MR stated ([1951] Ch 209, 220 to 221):

'I do not think it can or ought to be said that if, as I hold, the trustees wrongly thought that section 31 would operate, then a result is produced substantially or essentially different from that which was intended.'

The result was that for the period covered by the trustees' resolutions, the minor beneficiaries got their income, but the Revenue got their surtax on that income.

21. Abrahams and Hastings-Bass were both cases about plans to save estate duty by terminating a life interest and passing on settled property to the next generation. The plans (carried out in 1957 and 1958 respectively) were on the same general lines as that in Pilkington, the first-instance decision in which ([1959] Ch 699, Danckwerts J) had provided an encouraging precedent (the Revenue were joined in the proceedings and given leave to appeal in 1960). The Revenue were also parties to the Abrahams and Hastings-Bass cases, and in each case (ironically, in view of later developments, as Norris J pointed out) it was the Revenue which argued for the complete invalidation of the resettlement, partly through the direct operation of the rule against perpetuities, and partly (as an argument against severance) because 'the effect of the operation of the rule is wholly to alter the character of the settlement', as Cross J put it in Abrahams at p 485. Cross J rejected an argument approximating an advancement by way of resettlement to the exercise of a power of appointment. Although they were treated in the same way for perpetuity purposes, in his view the similarity ended there (p. 485 D-E):

'The interests given to separate objects of an ordinary special power are separate interests, but all the interests created in Carole's fund were intended as part and parcel of a single benefit to her.'

Cross J held, therefore, that there was no valid exercise of the power of advancement.

22. In Hastings-Bass the Court of Appeal, in a single judgment delivered by Buckley LJ, took a different view of a similar duty-saving transaction. The true ratio of the decision has been much debated, both in forensic argument and by legal scholars. It has been considered twice by Lloyd LJ, first in Sieff v Fox [2005] EWHC 1312 (Ch), [2005] 1 WLR 3811 paras 43 and 44 (his last first-instance case before his promotion to the Court of Appeal) and again, at much greater length, in his judgment in this case (paras 46 to 67).

23. It is perhaps simplest to start with what Hastings-Bass did not decide. It was not about mistake. Although one case on mistake (Wollaston v King (1869) LR 8 Eq 165) was cited, it was not referred to in the judgment. It would not have been enough for the Revenue to establish that the exercise of the trustees' power might have been voidable at the instance of a beneficiary. The Revenue could succeed only by establishing that there had been no valid advancement at all. Nor did the decision turn on any inquiry into what was actually in the minds of the trustees in exercising the power of advancement. There seems to have been no evidence of this, and in Buckley LJ's discussion at pp 39-41 (extensively quoted by Lloyd LJ at paras 53-56) the recurrent theme is what the trustees, as reasonable trustees, 'should' or 'would' have considered or intended. The third negative point to make is that Hastings-Bass did not overrule Abrahams. It was distinguished on the basis that in Abrahams the attenuated residue of the sub-settlement not struck down by the rule against perpetuities may not have been for the benefit of the beneficiary in question. But Buckley LJ did differ from Cross J's view that the benefit conferred by an advance by way of resettlement was of 'a monolithic character', preferring the view that it was 'a bundle of benefits of different characters.' If and so far as it is an issue of severability, it is obviously easier to sever part of a bundle than part of a monolith.

24. Buckley LJ's own statement of the principle of the decision in Hastings-Bass seems to be the passage at p 41 which has often been cited in later cases:

'To sum up the preceding observations, in our judgment, where by the terms of a trust (as under section 32) a trustee is given a discretion as to some matter under which he acts in good faith, the court should not interfere with his action notwithstanding that it does not have the full effect which he intended, unless (1) what he has achieved is unauthorised by the power conferred upon him, or (2) it is clear that he would not have acted as he did (a) had he not taken into account considerations which he should not have taken into account, or (b) had he not failed to take into account considerations which he ought to have taken into account.'

Lloyd LJ did not accept that as the true ratio. He thought that the Court of Appeal had already decided the case on the ground that the advancement, so far as not struck down by the rule against perpetuities, must stand unless it could not, in that attenuated form, reasonably be regarded as beneficial to the advancee. That is an objective test which does not call for an inquiry into the actual states of mind of the trustees.

25. Lloyd LJ expanded this line of thought in para 66:

'If the problem to be resolved is what is the effect on an operation such as an advancement of the failure of some of the intended provisions, because of external factors such as perpetuity, it is not useful to ask what the trustees would have thought and done if they had known about the problem. The answer to that question is almost certainly that they would have done something different, which would not have run into the perpetuity or other difficulty. It is for that reason that the test has to be objective, by reference to whether that which was done, with all its defects and consequent limitations, is capable of being regarded as beneficial to the intended object, or not. If it is so capable, then it satisfies the requirement of the power that it should be for that person's benefit. Otherwise it does not satisfy that requirement. In the latter case it would follow that it is outside the scope of the power, it is not an exercise of the power at all, and it cannot take effect under that power.'

On this analysis, limb (1) of Buckley LJ's statement of principle covers the whole ground, and limb (2) adds nothing. I respectfully agree with Lloyd LJ's criticism of the statement of principle. I think it is also open to criticism for the generality of its reference to unintended consequences ('notwithstanding that it does not have the full effect which he intended'). That is a far-reaching extrapolation from one case about section 31 of the Trustee Act 1925 and two cases about the rule against perpetuities. It set ajar a door that was pushed wide open in Mettoy Pension Trustees Ltd v Evans [1990] 1 WLR 1587 and other later cases.

Mettoy

26. In Mettoy Warner J applied the supposed new principle in the context of an occupational pension scheme, and in circumstances where the trustees' exercise of a discretionary power was within the scope of that power. There was no invalidating factor, such as the rule against perpetuities, applicable under the general law. In doing so Warner J dismissed two significant arguments for limiting the scope of the new principle.

27. The employer, Mettoy Co Plc, and the trustees of its pension scheme had in 1983 executed a deed to replace a 1980 deed (and some supplementary deeds) which were ineffective because of an error about the trusteeship. The rules scheduled to the 1980 deed included rule 13, providing for the winding-up of the scheme in certain circumstances, the priority of claims in the winding up and the disposal of any ultimate surplus. This rule differed from an earlier winding-up rule in several respects. Most materially, the discretion to use any ultimate surplus in augmenting benefits was to be exercisable by the employer (instead of by the trustees, as provided by the earlier rule). Moreover, in 1983 Mettoy's financial position was precarious (as a result of an ill-advised diversification from die-cast model vehicles into personal computers) so that winding-up of the scheme was much more than a remote possibility. In the event the scheme had to be wound up in 1984. The trustees issued an originating summons raising a number of questions, the most important being (in effect) whether the 1983 deed was wholly invalid, or valid except for rule 13, or valid except that the power of augmentation remained exercisable by the trustees. These questions arose because the trustees had admittedly not considered, or been advised about, the significance of rule 13.

28. In response to another question raised by the originating summons, Warner J held that the power of augmentation was, even when exercisable by the employer, a fiduciary power. On that basis it was not clear that the trustees, if they had fully considered the matter, would have objected to the change effected by rule 13 ([1990] 1 WLR 1587, 1628A-1630A). But by then Warner J had upheld (in a passage from pp1621G to 1626A) the existence of 'a principle which may be labelled 'the rule in Hastings-Bass''. He took Buckley LJ's statement of principle in that case (set out at para 24 above) and reformulated it in positive terms, and so far as relevant to the facts of the case, as follows (p 1621H):

'where a trustee acts under a discretion given to him by the terms of the trust, the court will interfere with his action if it is clear that he would not have acted as he did had he not failed to take into account considerations which he ought to have taken into account.'

29. Warner J rejected the submissions of Mr Edward Nugee QC, recorded at pp 1622G to 1623G, that the principle, although existent, was of very narrow scope, and that the cases of Vestey, Abrahams and Hastings-Bass (together with Pilkington, where there was a proposal for a resettlement rather than a completed transaction):

'. . .were about the consequences of what [Mr Nugee] referred to as an 'excessive execution' of a power, ie the purported exercise of a power in a way that the law rendered partially ineffective.'

Warner J dismissed this argument at p1624B-C:

'If, as I believe, the reason for the application of the principle is the failure by the trustees to take into account considerations that they ought to have taken into account, it cannot matter whether that failure is due to their having overlooked (or to their legal advisers having overlooked) some relevant rule of law or limit on their discretion, or is due to some other cause.'

30. Warner J also dismissed what he called Mr Nugee's 'all or nothing' argument (pp 1624H-1625A). In some cases the court would have to declare void the whole of some purported exercise of discretion by trustees. But in other cases (for instance where the trustees would have decided, had they thought about it properly, to omit some particular provision from a deed) the appropriate course would be to declare that provision alone to be void.

31. At p 1626D Warner J referred to 'the all important third question: what would the trustees have done if they had considered the matters that they failed to consider?' His meticulous review of the oral and documentary evidence, including the cross-examination of Mr Lillyman (who was at all material times closely involved as the employer's company secretary and a director of the corporate trustee) shows that he was concerned to establish, so far as he could, what these particular trustees (and not some hypothetical reasonable trustees) would have done. His approach was subjective, not objective.

32. I respectfully agree with Lloyd LJ's view that the basis on which Mettoy was decided cannot be found in the reasoning which led to the decision in Hastings-Bass. It can claim to be an application of Buckley LJ's summary statement of principle, but only if that statement is taken out of context and in isolation from the earlier part of the judgment. If the principle applied by Warner J merits a name at all, it should be called the rule in Mettoy. But the rule as formulated by Warner J has given rise to many difficulties, both in principle and in practice.

From Mettoy to Sieff

33. Mettoy was not much considered by the court during the 1990s. It was cited but not referred to in the judgment of the Court of Appeal in Edge v Pensions Ombudsman [2000] Ch 602. That decision, on an appeal by the Pensions Ombudsman from the judgment of Sir Richard Scott V-C [1998] Ch 512, was largely concerned with the jurisdiction of the Pensions Ombudsman under Part X of the Pension Schemes Act 1993. The general tenor of the Court of Appeal's judgment is that neither the Ombudsman nor the court has power to intervene in decisions made by trustees unless they have acted in breach of duty. That can be seen as putting down a marker that Lloyd LJ has since recognised.

34. In Stannard v Fisons Pension Trusts Ltd [1991] Pen LR 225, in which Hastings-Bass but not Mettoy was cited, the Court of Appeal modified Buckley LJ's formulation, without any full discussion of the point, by putting the test in terms of what the trustees might, rather than would, have done if fully informed. The facts were that trustees had taken a decision about transfer values on the basis of an out of date valuation of the pension fund. The Court of Appeal's modification of the test seems questionable since the legal significance of the error must have depended on the scale of the change in market value rather than on the precise nature of the trustees' hypothetical second thoughts.

35. It was not until about the year 2000 that Hastings-Bass and Mettoy began to be called in aid in cases where tax-planning arrangements involving trusts had gone wrong. The first case seems to have been Green v Cobham, decided by Jonathan Parker J in January 2000 but reported at [2002] STC 820, followed by Abacus Trust Co (Isle of Man) v National Society for the Prevention of Cruelty to Children [2001] STC 1344 (Patten J) and Breadner v Granville-Grossman [2001] Ch 523 (Park J). Breadner was an unsuccessful attempt to extend the principle so as to circumvent a missed time limit for the exercise of a power of appointment. Park J observed at para 61:

'There must surely be some limits. It cannot be right that whenever trustees do something which they later regret and think that they ought not to have done, they can say that they never did it in the first place.'

36. The most important decisions, prior to the present appeals, are the decisions of Lightman J in Abacus Trust Co (Isle of Man) v Barr [2003] Ch 409 and Lloyd LJ in Sieff v Fox [2005] 1 WLR 3811. In the former case Mr Barr had participated in a management buy-out of an engineering company and in 1992 he had settled his shares in the buy-out vehicle, held through an Isle of Man holding company, in a settlement of which Abacus Trust Co (Isle of Man) ('Abacus') was trustee. Abacus was administered by the Isle of Man firm of Coopers and Lybrand ('CandL (IoM)'). CandL (IoM) had close links with the English firm of Coopers and Lybrand, which advised on the buy-out. Mr Ward-Thompson of the English firm was Mr Barr's main contact.

37. Under the settlement Mr Barr had a life interest, but Abacus as trustee had an overriding power of appointment in favour of a wide class of beneficiaries. Very soon after the creation of the settlement Mr Barr told Mr Ward-Thompson that he wished 40% of the trust fund to be appointed on discretionary trusts in favour of his sons and their families, to the exclusion of himself and any wife of his. Through some misunderstanding this was conveyed to CandL (IoM) as a wish for 60% of the fund to be appointed, and on 22 April 1992 an appointment in that form was made. The mistake was discovered in August 1992 but nothing was done to try to remedy it until 2002. In the meantime, in 1994 the buy-out vehicle was floated on the London Stock Exchange and the holding company controlled by Abacus embarked on a programme of sales of its shares.

38. The judgment of Lightman J is impressively brief and incisive. He pointed out that Abacus was not seeking either rectification or rescission for mistake, and added in relation to the Hastings-Bass rule (para 13):

'But in considering the ambit of the rule it is necessary to bear in mind that it is only one of the protections afforded to beneficiaries in respect of the due administration of the trust by the trustees. It is also important to have in mind that equity does not afford a trustee or a beneficiary a free pass to rescind a decision which subsequently proves unpalatable or unfortunate and substitute another. Relief is only available if the necessary conditions for its grant are satisfied.'

He referred to the authorities already discussed and observed that he did not need to resolve the issue posed by Stannard, since (para 20) 'clearly the trustee would not have appointed 60% of the trust fund if it had known of the settlor's true wishes.' He then addressed four issues: (1) whether there had to be a fundamental mistake; (2) whether the rule applied if there was any relevant mistake or ignorance on the part of the trustee, regardless of how it arose (and in particular, regardless of any breach of duty on the part of the trustee); (3) following from the last point, whether the rule applied on the facts of the case before him; and (4) whether, if the rule applied, the appointment was void or voidable.

39. On the first issue Lightman J decided, correctly in my view, that a fundamental mistake was not necessary. A fundamental, or at least serious mistake may be necessary for rescission on the ground of mistake (that is relevant to the second ground of appeal in Pitt), but for the rule which Abacus was invoking (para 21):

'the rule does not require that the relevant consideration unconsidered by the trustee should make a fundamental difference between the facts as perceived by the trustee and the facts as they should have been perceived. All that is required in this regard is that the unconsidered relevant considerations would or might have affected the trustees' decision, and in a case such as the present that the trustee would or might have made a different appointment or no appointment at all.'

But as his decision on the second point shows, it must be sufficiently serious as to amount to a breach of duty.

40. On the second issue, Lightman J held that a breach of duty on the part of the trustee is essential to the application of the rule (para 23):

'What has to be established is that the trustee in making his decision has, in the language of Warner J in Mettoy Pension Trustees Ltd v Evans [1990] 1 WLR 1587, 1625, failed to consider what he was under a duty to consider. If the trustee has in accordance with his duty identified the relevant considerations and used all proper care and diligence in obtaining the relevant information and advice relating to those considerations, the trustee can be in no breach of duty and its decision cannot be impugned merely because in fact that information turns out to be partial or incorrect.'

41. That is in my view a correct statement of the law, and an important step towards correcting the tendency of some of the earlier first-instance decisions. If in exercising a fiduciary power trustees have been given, and have acted on, information or advice from an apparently trustworthy source, and what the trustees purport to do is within the scope of their power, the only direct remedy available (either to the trustees themselves, or to a disadvantaged beneficiary) must be based on mistake (there may be an indirect remedy in the form of a claim against one or more advisers for damages for breach of professional duties of care). This serves to emphasise that the so-called rule in Hastings-Bass was not in play in that case, or in Abrahams. In those two cases the trustees were not at fault in failing to foresee the House of Lords' decision in Pilkington several years later. But they purported to exercise their power of advancement in a way that was beyond the scope of that power, since it was contrary to the general law (that is the rule against perpetuities as clarified in Pilkington). The issue (resolved differently in Abrahams and Hastings-Bass) was whether the parts of the resettlement not void for perpetuity were sufficient to amount to a proper exercise of the power of advancement. In Mettoy and Barr, by contrast, it was never in doubt that the relevant deed fell within the scope of the trustees' power. This point is clearly made in paras 92 and 93 of Lloyd LJ's judgment in the Court of Appeal.

42. On the third issue Lightman J held that Abacus was in breach of duty, mainly because it had to take responsibility for Mr Ward-Thompson, who (para 27) 'has declined to give evidence and answer the case made or suggest a different scenario'. This part of the judgment turns on the particular facts of the case, but they are typical of many such cases, and I shall return to them in discussing the difficulties that still beset this area of the law.

43. On the fourth issue Lightman J held that in cases where the rule applies (as opposed to cases of equitable non est factum such as Turner v Turner [1984] Ch 100) it makes the trustees' disposition voidable, not void. The Court of Appeal agreed with his analysis, and so do I. The rule, properly understood, depends on breach of duty in the performance of something that is within the scope of the trustees' powers, not in the trustees doing something that they had no power to do at all. Beneficiaries may lose their right to complain of a breach of trust by complicity, by laches or acquiescence or in other ways. Lightman J adjourned the case, expressing the hope (para 34) that a compromise would be possible. The absence of any further reported decision suggests that his hope was realised.

44. In Sieff v Fox [2005] 1 WLR 3811 Lloyd LJ (as he had become by the time he handed down his lengthy reserved judgment) fully considered all the authorities mentioned above, and other authorities on mistake. I can take his judgment fairly briefly because he had occasion to reconsider it, and on one important point to depart from it, in his judgment in the Court of Appeal in these appeals. The case related to valuable land and chattels comprised in the Bedford settled estates, and the facts as to the trusts, and their tax implications, are very complicated.

45. It is sufficient to note two points. First, the critical appointment (made in 2001 by the trustees in favour of Lord Howland, and with a view to a resettlement by him) required the consent of Lord Howland himself. In deciding whether or not to give consent Lord Howland was not acting in a fiduciary capacity. His consent (given in ignorance of some of the implications, including adverse tax consequences) was challenged, successfully, on the ground of mistake (see paras 115 and 119 (vii) of the judgment of Lloyd LJ). In his discussion of mistake, Lloyd LJ relied (paras 98 to 101) on Ogilvie v Littleboy (1897) 13 TLR 399, upheld on appeal as Ogilvie v Allen (1899) 15 TLR 294. The trustees' exercise of their power of appointment was challenged, also successfully, under the Hastings-Bass rule (see para 114, and compare para 119 (vi)).

46. The second point to note is that Lloyd LJ was inclined to differ from Lightman J as to the need for the vitiating element in a fiduciary decision to amount to a breach of trust. Lloyd LJ referred to the trustees in Abrahams not being at fault in failing to foresee that the first-instance decision in Pilkington would be reversed on an appeal made out of time. But Abrahams was a case in which the purported exercise of the trustees' power was outside its proper scope, because it infringed the rule against perpetuities. This is the point on which Lloyd LJ has modified the provisional view which he expressed in Sieff v Fox.

Futter v Futter: The facts and the first-instance decision

47. The appeal in Futter is concerned with incorrect advice given by solicitors as to the effect of provisions, primarily in section 87 of the Taxation of Chargeable Gains Act 1992 ('TCGA'), charging capital gains tax in respect of gains realised by non-resident trustees. There were two settlements, the No 3 settlement and the No 5 settlement, made by Mr Mark Futter in 1985. Initially both settlements had non-resident trustees, but in 2004 Mr Futter and Mr Cutbill, both resident in the United Kingdom, were appointed as trustees of the two settlements. Mr Cutbill was a partner in the London solicitors which gave the tax advice. At that stage both settlements had 'stockpiled' gains “ that is, gains realised while the trust was not resident, and not yet distributed to the beneficiaries or brought in to charge for capital gains tax purposes.

48. On the advice of the solicitors, the new, resident trustees on 31 March 2008 distributed the whole capital of the No 3 settlement to Mr Futter, in exercise of a power of enlargement, and on 3 April 2008 distributed £36,000 from the No 5 settlement to Mr Futter's three children in equal shares, in exercise of a power of advancement. Each of these transactions was squarely within the scope of the relevant power. Mr Futter and Mr Cutbill understood (correctly) that the 'stockpiled' gains would in consequence be attributed to Mr Futter and his children as if they were gains realised by those beneficiaries themselves. They also believed (incorrectly) that these attributed gains would be absorbed by allowable losses which they had realised so that no eventual tax liability would arise. This overlooked the effect of section 2(4) of TCGA as amended (the relevant amendment, for those interested in the fine detail, was that made by Schedule 21, para 2 of the Finance Act 1998, and not the further amendment made by Schedule 2, para 24 of the Finance Act 2008, which applied only from 5 April 2008). The result was a large capital gains tax liability for Mr Futter and a modest one for his children.

49. Mr Futter and Mr Cutbill applied, as trustees of the two settlements, to have the deed of enlargement and the deeds of advancement declared void. The first four defendants, the beneficiaries, did not appear. The fifth defendant, the Revenue, resisted the application.

50. Norris J began his judgment in spirited fashion, as already noted (para 3 above). However he went on to state that it was not an occasion for a first-instance judge to reconsider a developed rule. He took the judgment of Lloyd LJ in Sieff v Fox as the leading authority on the rule, as had Sir Andrew Park in Smithson v Hamilton [2008] 1 WLR 1453, para 52, and as had Mr Robert Englehart QC in Pitt v Holt [2010] 1 WLR 1199, para 18.

51. The Revenue's submissions were similar to those advanced in Pitt (para 57 below), apart from the receivership point. As it happens the first-instance judgment in Pitt was given on the first day of the first-instance hearing in Futter, so that there was no real opportunity for revision of the Revenue's case. As recorded in the judgment of Norris J the Revenue had three main lines of argument. The first was that the decision of the trustees 'was not in any meaningful sense different from what they intended (apart from the tax consequences)'. This argument echoed the distinction drawn by Millett J in Gibbon v Mitchell [1990] 1 WLR 1304, 1309-1310, between 'effect' and 'consequences'. Norris J rejected this argument on the ground that mistake was a different ground for relief, and that under the Hastings-Bass rule tax consequences are rightly regarded as something that trustees must take into account in exercising their discretions. The Revenue's second line of argument focused on the significance of the trustees' error. It was to some extent a variation on the first argument, and it was rejected on similar grounds. The Revenue's third submission (not pressed) was that so far from considering capital gains tax, the trustees had it in the forefront of their minds: 'the problem was that the advice was wrong' (para 28). But wrong advice on tax consequences could, the judge said, lead to a perfectly orthodox application of the rule.

52. Norris J held that the deeds were void, not voidable. He referred briefly (para 32) to the judgment of Lightman J in Barr, but noted that his reasoning (based on the trustees being at fault) was not accepted by Lloyd LJ in Sieff v Fox. Nevertheless Norris J considered (para 33) that the rigours of the 'void' analysis could be mitigated by the application of equitable principles.

Pitt v Holt: The facts and the first-instance decision

53. The facts relevant to the Pitt appeal are set out at length in the judgment of Lloyd LJ at paras 147 to 159, to which reference may be made for further detail. The claim was made by the personal representatives of Mr Derek Pitt, who died in 2007 aged 74. In 1990 he had suffered very serious head injuries in a road traffic accident, resulting in his mental incapacity. His wife, Mrs Patricia Pitt, was appointed as his receiver under the Mental Health Act 1983, and on his death she became one of his personal representatives, and the only beneficiary interested in his estate. Mr Pitt's claim for damages for his injuries was compromised by a structured settlement, approved by the court, in the sum of £1.2m. Mrs Pitt's solicitors sought advice from Frenkel Topping, a firm of financial advisers said to have specialist experience of structured settlements. They advised that the damages should be settled in a discretionary settlement, and this was done, with the authority of the Court of Protection, in 1994. The trust was referred to as the Derek Pitt Special Needs Trust ('the SNT').

54. Frenkel Topping gave their advice in a written report to Mrs Pitt (as receiver) which was made available to the Official Solicitor, who represented her husband in the application to the Court of Protection. The report referred to various advantages which the SNT was expected to secure, and it mentioned income tax and capital gains tax in its illustrative forecasts. But the report made no reference whatsoever to inheritance tax. The SNT could have been established without any immediate inheritance tax liability if (i) it had been an interest in possession trust or (ii) it had been a discretionary trust complying with section 89 of the Inheritance Tax Act 1984. In order to comply with section 89 its terms should have provided that at least half of the settled property applied during Mr Pitt's lifetime was applied for his benefit. But the SNT as drafted and executed contained no such restriction. The consequence was an immediate liability to inheritance tax of the order of £100,000, with the prospect of a further tax charge on the tenth anniversary in 2004. The deputy judge (Mr Robert Englehart QC) observed that by 2010 the total tax, together with interest and penalties (if exacted) must have amounted to between £200,000 and £300,000.

55. Mrs Pitt and her advisers became aware of the inheritance tax liabilities in 2003. In 2006 Mr Pitt (by a litigation friend) and the trustees of the SNT commenced proceedings against Frenkel Topping claiming damages for professional negligence. Mr Pitt died in 2007. After taking further advice his personal representatives (who were also two of the trustees of the SNT) commenced proceedings seeking to have the SNT set aside either under the Hastings-Bass rule, or on the ground of mistake. The first defendant was the remaining trustee of the SNT (who took no part in the proceedings) and the second defendant was the Revenue (which actively opposed the application). Evidence was given in writing and there was no cross-examination.

56. In his judgment the deputy judge discussed the principal authorities on the Hastings-Bass rule and observed (para 22) that three matters were not in dispute. First, it was agreed that the rule could apply without the need to identify a breach of duty on the part of the trustees or their advisers (so following Sieff v Fox rather than Barr). Second, it was unnecessary on the facts of the case to decide whether the application of the rule rendered a transaction void or voidable. Third, the rule would apply only if it was established that Mrs Pitt, if properly advised, would not have set up the SNT (rather than merely might not have done so).

57. The principal arguments for the Revenue were that the rule did not in any case apply to a receiver (as opposed to a formally constituted trustee); that the rule applied only to a limited class of cases 'where the immediate purpose of the act in question was not achieved'; and that tax consequences were never a sufficient basis for the application of the rule. The deputy judge rejected these submissions, holding that a receiver, as a fiduciary, was in essentially the same position as a trustee, and that the weight of the first-instance authorities supported a wider version of the rule. He set aside the SNT on that ground. He indicated that he was not satisfied that there was any real mistake, as opposed to a failure to think about tax at all. Even if there was a mistake of any sort, it was only a mistake as to the consequences of the transaction, rather than its effect.

Lloyd LJ's judgment on the Hastings-Bass rule

58. I have already indicated my general agreement with Lloyd LJ's judgment on the Hastings-Bass issue. Paragraphs 1 to 28 contain an introduction and a summary of the facts of the two appeals. Paragraphs 29 to 67 consider Vestey, Abrahams and Hastings-Bass. They come to the conclusion, with which I fully agree, that Buckley LJ's statement of the supposed rule (para 24 above) was wider than the true principle of the actual decision in Hastings-Bass. Paragraphs 68 to 91 consider more recent authorities, including Mettoy and Barr. All this is in a sense preliminary. Lloyd LJ's essential reasoning and conclusions are at paragraphs 92 to 131. He then applied what he saw as the correct principle to the facts of Futter (paras 132 to 145) and Pitt (paras 146 to 163). He then dealt with the issue of mistake, raised by the respondent's notice in Pitt (paras 164 to 223). The outcome was that both appeals were allowed (paras 224 to 226).

59. Longmore LJ and Mummery LJ both gave short concurring judgments expressing full agreement. Mummery LJ added a clear summary of five salient points (paras 233 to 238).

60. In the core of his judgment Lloyd LJ correctly spelled out the very important distinction between an error by trustees in going beyond the scope of a power (for which I shall use the traditional term 'excessive execution') and an error in failing to give proper consideration to relevant matters in making a decision which is within the scope of the relevant power (which I shall term 'inadequate deliberation'). Hastings-Bass and Mettoy were, as he rightly observed, cases in quite different categories. The former was a case of excessive execution and the latter might have been, but in the end was not, a case of inadequate deliberation. Lloyd LJ therefore withdrew his doubts about the conclusions that Lightman J had reached in Barr.

61. Lloyd LJ then addressed the difficult question of how a fraudulent appointment (that is, an appointment ostensibly within the scope of a power, but made for an improper purpose) is to be fitted into the classification. The exercise of an equitable power may be fraudulent in this sense whether or not the person exercising it is a fiduciary. A well-known example of trustees exercising a power for an improper purpose is provided by In Re Pauling [1964] Ch 303, in which a power ostensibly exercisable for the benefit of young adult beneficiaries was used to distribute trust capital to be frittered away on their improvident parents' living expenses.

62. There is Court of Appeal authority that a fraudulent appointment is void rather than voidable: Cloutte v Storey [1911] 1 Ch 18. In that case the appointee under an improper appointment had charged his equitable interest as security for a loan (and in doing so made two false statutory declarations as to the genuineness of the appointment). It was held that the lender had no security, even though it had no notice of the equitable fraud. It is an authority which has bedevilled discussion of the true nature of the Hastings-Bass rule. Lightman J found the judgment of Farwell LJ problematic (Barr, para 31) and Lloyd LJ shared his reservations (para 98). So do I. It is hard to know what to make of Farwell LJ's observations [1911] 1 Ch 18, 31:

'If an appointment is void at law, no title at law can be founded on it; but this is not so in equity: the mere fact that the appointment is void does not prevent a Court of Equity from having regard to it: eg, an appointment under a limited power to a stranger is void, but equity may cause effect to be given to it by means of the doctrine of election.'

The decision in Cloutte v Storey may have to be revisited one day. For present purposes it is sufficient to note that a fraudulent appointment (that is, one shown to have been made for a positively improper purpose) may need a separate pigeon-hole somewhere between the categories of excessive execution and inadequate deliberation.

63. In paragraphs 102 to 118, Lloyd LJ considered the duties of trustees in exercising their discretion, and in particular the relevance of tax considerations. He referred to some well-known authorities including In re Baden's Deed Trusts [1971] AC 424. That case was directly concerned with the correct test for certainty of objects of a discretionary trust (or trust power) but the speech of Lord Wilberforce contains, at pp 448-457, a general discussion of fiduciary discretions which has been very influential in the development of the law. This includes a passage at pp 456-457 as to the Court's intervention if trustees fail to exercise a trust power (that is, a discretion which it is their duty to exercise in some way). After referring to Lord Upjohn's opinion Lord Wilberforce said:

'I would venture to amplify this by saying that the court, if called upon to execute the trust power, will do so in the manner best calculated to give effect to the settlor's or testator's intentions. It may do so by appointing new trustees, or by authorising or directing representative persons of the classes of beneficiaries to prepare a scheme of distribution, or even, should the proper basis of distribution appear by itself directing the trustees so to distribute'.

Lloyd LJ did not refer to that particular passage, but Warner J had done so in Mettoy [1990] 1 WLR 1587, 1617-1618, since in that case a decision as to the exercise of the power to augment benefits would have to be taken by someone. The passage serves as a reminder that where trustees have been in breach of duty by exercising a discretion with inadequate deliberation, setting aside their decision may not be the only course open to the court.

64. In discussing what trustees should take into account, Lloyd LJ observed that the older cases tended to focus, not on what should be taken into account, but on what should not be taken into account. He instanced two cases. One was Klug v Klug [1918] 2 Ch 67, where one of the trustees strongly disapproved of her daughter's choice of husband, and for that reason refused to concur with the Public Trustee in exercising a power of advancement in her favour. The court overrode her objection because she had not considered whether or not it would be for her daughter's welfare that the advance should be made. She had therefore made no proper exercise of her discretion. The other case was In re Lofthouse (An Infant) (1885) 29 Ch D 921, in which income of a fund was available for the maintenance of a five year old girl whose mother had died very shortly after her birth. Her father had remarried and had three children by his second marriage. Bacon V-C and the trustees (of the will of the girl's maternal grandmother) evidently took very different views of how the father would spend the income if it was all paid to him for his eldest child's maintenance. The case was resolved by agreement in the Court of Appeal. The old cases as to the maintenance of children are rather exceptional, especially where the position was complicated by the child in question being a ward of court, as in In re Hodges (1878) 7 Ch D 754. Some judicial pronouncements in these cases should not be taken out of context.

65. At para 115 Lloyd LJ reaffirmed the view that he had expressed in Sieff v Fox, para 86, that 'fiscal consequences may be relevant considerations which the trustees ought to take into account'. I agree. In the private client world trusts are mostly established by and for wealthy families for whom taxes (whether on capital, capital gains or income) are a constant preoccupation. It might be said, especially by those who still regard family trusts as potentially beneficial to society as a whole, that the greater danger is not of trustees thinking too little about tax, but of tax and tax avoidance driving out consideration of other relevant matters.

66. That is particularly true of offshore trusts. They are usually run by corporate trustees whose officers and staff (especially if they change with any frequency) may know relatively little about the settlor, and even less about the settlor's family. The settlor's wishes are always a material consideration in the exercise of fiduciary discretions. But if they were to displace all independent judgment on the part of the trustees themselves (or in the case of a corporate trustee, by its responsible officers and staff) the decision-making process would be open to serious question. The Barr case (2003) Ch 409 illustrates the potential difficulties of unquestioning acceptance of the settlor's supposed wishes.

67. It is interesting, in this context, to compare the facts of some of the offshore cases with those of Turner v Turner [1984] Ch 100. That was a case in which a farmer made a discretionary settlement which he did not understand, and appointed as trustees family friends who never realised that they had any responsibility at all except to do as the settlor asked. They thought that it would be intruding into the settlor's affairs if they were to read the documents that they were asked to sign (see at pp 106-108). Anyone familiar with the duties of trustees may find this hard to contemplate (as Mervyn Davies J did, at p 109). But it may be that some offshore trustees come close to seeing their essential duty as unquestioning obedience to the settlor's wishes.

68. The Barr case also illustrates another practical difficulty in the application of the Hastings-Bass rule as it has developed. Lightman J was in my view right to decide that when the vitiating error is inadequate deliberation on relevant matters (rather than mistake) the inadequacy must be sufficiently serious as to amount to a breach of duty; and Lloyd LJ was right to change the contrary view which he had expressed in Sieff v Fox. It would set the bar too high (or too low, depending on the spectator's point of view) to apply the Hastings-Bass rule whenever trustees fall short of the highest standards of mature deliberation and judgment. Where, as in Barr, the trustee is a body corporate acting as a sort of in-house facility provided by a firm of professional advisers, it may be hard to decide whether the separate juristic personality of the trustee insulates it from responsibility for the errors of individual professionals within the firm. A rather similar problem arose on the facts of Futter.

69. It is a striking feature of the development of the Hastings-Bass rule that it has led to trustees asserting and relying on their own failings, or those of their advisers, in seeking the assistance of the court. This was pointed out in no uncertain terms by Norris J in his first-instance judgment in Futter, quoted in para 3 above. There may be cases in which there is for practical purposes no other suitable person to bring the matter before the court, but I agree with Lloyd LJ's observation (para 130) that in general it would be inappropriate for trustees to take the initiative in commencing proceedings of this nature. They should not regard them as uncontroversial proceedings in which they can confidently expect to recover their costs out of the trust fund.

70. Lloyd LJ stated the correct principle, as he saw it, at para 127:

'It seems to me that the principled and correct approach to these cases is, first, that the trustees' act is not void, but that it may be voidable. It will be voidable if, and only if, it can be shown to have been done in breach of fiduciary duty on the part of the trustees. If it is voidable, then it may be capable of being set aside at the suit of a beneficiary, but this would be subject to equitable defences and to the court's discretion. The trustees' duty to take relevant matters into account is a fiduciary duty, so an act done as a result of a breach of that duty is voidable. Fiscal considerations will often be among the relevant matters which ought to be taken into account. However, if the trustees seek advice (in general or in specific terms) from apparently competent advisers as to the implications of the course they are taking, and follow the advice so obtained, then, in the absence of any other basis for a challenge, I would hold that the trustees are not in breach of their fiduciary duty for failure to have regard to relevant matters if the failure occurs because it turns out that the advice given to them was materially wrong. Accordingly, in such a case I would not regard the trustees' act, done in reliance on that advice, as being vitiated by the error and therefore voidable.'

The requirement for breach of duty

71. In this court Mr Robert Ham QC undertook the main burden of the argument for the appellants on the Hastings-Bass rule. Mr Christopher Nugee QC adopted Mr Ham's submissions, and added some of his own, but concentrated his argument on the issue of mistake. Mr Ham's submissions centred on whether the court's jurisdiction under the Hastings-Bass rule is exercisable only if there is a breach of fiduciary duty on the part of the trustees (or other relevant fiduciary). He argued that this is a novel requirement which leads to arbitrary and unfair distinctions, especially in cases where incorrect advice on tax has been given by professional advisers who may or may not themselves be trustees. Mr Ham also had subsidiary but important arguments about the attribution to trustees of fault on the part of their advisers, and about the identification of relevant considerations for the purposes of the rule.

72. Mr Ham contended that the supposed need for establishing a breach of fiduciary duty, before the Hastings-Bass rule can come into play, was a novel requirement introduced in 2003 by Lightman J in his judgment in Barr. Reference to paras 16 to 20 of his judgment shows that Lightman J was relying on a number of earlier authorities, including the decision of the Court of Appeal in Edge [2000] Ch 602, 627-628, and the decision of Warner J in Mettoy [1990] 1 WLR 1587, 1625:

'In a case such as this, where it is claimed that the rule in Hastings-Bass applies, three questions arise: (1) What were the trustees under a duty to consider? (2) Did they fail to consider it? (3) If so, what would they have done if they had considered it?'

73. In my view Lightman J was right to hold that for the rule to apply the inadequate deliberation on the part of the trustees must be sufficiently serious as to amount to a breach of fiduciary duty. Breach of duty is essential (in the full sense of that word) because it is only a breach of duty on the part of the trustees that entitles the court to intervene (apart from the special case of powers of maintenance of minor beneficiaries, where the court was in the past more interventionist: see para 64 above). It is not enough to show that the trustees' deliberations have fallen short of the highest possible standards, or that the court would, on a surrender of discretion by the trustees, have acted in a different way. Apart from exceptional circumstances (such as an impasse reached by honest and reasonable trustees) only breach of fiduciary duty justifies judicial intervention.

74. Mr Ham relied heavily on the decision of the Court of Appeal in Kerr v British Leyland (Staff) Trustees Ltd (1986) [2001] WTLR 1071. In that case Mr Kerr, a member of the British Leyland staff pension scheme, suffered from angina and claimed a disability benefit payable on permanent disability (defined as 'so that no further employment of any kind is possible'). The scheme had a group policy with Legal and General, which obtained separate medical advice and indicated that it would reject the claim. The corporate trustee of the pension scheme decided to follow Legal and General in rejecting the claim. Mr Kerr took proceedings challenging the trustees' decision, and was successful at first instance. The Court of Appeal dismissed the corporate trustee's appeal. In doing so Fox LJ (with whom Mustill LJ and Caulfield J agreed) made plain that the corporate trustee's board was not at fault. There had been a failure of communication. As to the judge's declaration that Mr Kerr was entitled to a pension, Fox LJ stated (p 1080):

'I do not think he was entitled to do that. The decision whether to accept the claim is one for the trustee and not for the court. It seems to me that, in the present case, the decision of the trustee was simply ineffective since the board did not carry out their duty to give a properly informed consideration to the claim. That however does not entitle the Court to substitute its own view of the claim for that of the trustee.

I would, therefore, discharge the order of the judge and substitute an order that the decision of the trustees on 28 June 1978 to reject Mr Kerr's claim was of no effect and that the trustee should reconsider the claim.'

75. The Kerr case is of interest since (though not reported for 15 years) it is an early example, antedating Mettoy, of the application of something like the Hastings-Bass rule. But I think it is important to note that under the British Leyland scheme the corporate trustee did not have any real discretion about disability benefit. It had to exercise a judgment on an issue of fact (permanent disability from any employment). That is an issue on which the court would be much more ready to intervene if the trustee had failed to grasp the real facts. It is an intermediate situation which is arguably closer to a mistaken judgment on an issue of fact than to the defective exercise of a discretion.

76. Kerr may be compared with Mihlenstedt v Barclays Bank International Ltd [1989] IRLR 522. That was a comparable case except that there was a preliminary issue of construction as to whether the relevant rule (which began 'Early retirement due to ill health will be permitted only when . . .') imported a wider discretion. The Court of Appeal decided that the language of the rule was that of obligation and entitlement, and that the judge had erred in supposing that there was a wider discretion. But on the facts the majority of the Court of Appeal held that the trustee had not formed its opinion on an erroneous basis.

77. Mr Ham's fallback position was that if a breach of duty was an essential requirement, there could be a breach without fault on the part of any individual trustee being established. This general argument was developed in several different directions. I would identify these (though there was some overlap) as (1) strict liability (2) agency (3) resulting absurdity and (4) a special meaning of 'relevant'. These points are considered below, in turn. Mr Nugee, in supporting Mr Ham's position, attached most weight to the argument on strict liability.

78. It is undoubtedly correct that trustees may be liable for breach of trust even though they have acted in accordance with skilled professional advice. Such advice cannot protect trustees from potential liability for a loss to the trust fund resulting from a decision that is, judged objectively, beyond the trustees' powers and detrimental to the trust (though professional advice may lead to their obtaining relief under section 61 of the Trustee Act 1925). An example mentioned in argument is Dunn v Flood (1885) 28 Ch D 586, in which trustees had sold by auction 73 plots of freehold land at Reading, subject to special conditions which the court held to be severely depreciatory (as Fry LJ put it at p594, 'eminently calculated to frighten away purchasers'). The Court of Appeal, upholding North J, refused to force a doubtful title on a reluctant purchaser. The fact that the trustees had consulted respectable solicitors was no excuse. It was not a reasonable exercise of discretion (Baggallay LJ and Bowen LJ at p592; Fry LJ at pp593-594). But the trustee's breach of duty was not in the manner of their decision-making (as to which we know nothing other than that they consulted respectable solicitors) but the loss to the trust property that their unreasonable decision appeared to have caused.

79. Further examples are provided by the decision of the Court of Appeal in Perrins v Bellamy [1899] 1 Ch 797 and that of the Privy Council, on appeal from the Supreme Court of Victoria, in National Trustees Co of Australasia Ltd v General Finance Co of Australasia Ltd [1905] AC 373. These cases, discussed by Lloyd LJ at para 124 of his judgment, were both examples of action taken by trustees on professional advice which was unequivocally incorrect: one a sale of leaseholds when the trustees had no power of sale; the other a distribution (resulting from 'some extraordinary slip' by solicitors of high standing) of a deceased beneficiary's vested share to persons who were not entitled to it under the intestacy law of Victoria in force at the beneficiary's death. As Lloyd LJ observed, the issue in these cases:

'is altogether different, as it seems to me, from the question whether, if trustees take advice properly, and act on that advice in a matter which is within their powers, the fact that the advice has misled them as to the true position in a relevant respect means that they acted in breach of fiduciary duty.'

80. I respectfully agree. Trustees may be liable, even if they have obtained apparently competent professional advice, if they act outside the scope of their powers (excessive execution), or contrary to the general law (for example, in the Australian case, the law regulating entitlement on intestacy). That can be seen as a form of strict liability in that it is imposed regardless of personal fault. Trustees may also be in breach of duty in failing to give proper consideration to the exercise of their discretionary powers, and a failure to take professional advice may amount to, or contribute to, a flawed decision-making process. But it would be contrary to principle and authority to impose a form of strict liability on trustees who conscientiously obtain and follow, in making a decision which is within the scope of their powers, apparently competent professional advice which turns out to be wrong.

81. Such a result cannot be achieved by the route of attributing any fault on the part of professional advisers to the trustees as their supposed principals. Solicitors can and do act as agents in some clearly-defined functions, usually of a ministerial nature, such as the receipt and transmission of clients' funds, and the giving and taking of undertakings on behalf of clients. But they do not and may not act as agents in the exercise of fiduciary discretions. As I said in Scott [1998] 2 All ER 705, 717:

'It is however for advisers to advise and for trustees to decide: trustees may not (except insofar as they are authorised to do so) delegate the exercise of their discretions, even to experts.'

82. Mr Ham relied on some observations of Warner J in Mettoy [1990] 1 WLR 1587, 1625-1626:

'But the question is not in my view to what extent trustees may in practice have to rely on professional advice. The duty to take into account all material considerations is that of the trustees. The extent of that duty is not affected by the amount or quality of the professional advice they may seek or obtain. In In Re Hastings-Bass [1975] Ch 25 it was not relevant to what extent the trustees themselves were able to form an opinion on the effect of the rule against perpetuities.'

This passage was noted by Lloyd LJ in his discussion of the cases (para 71) but receives only a passing mention in para 124, the part of his judgment which discusses the significance of professional advice. I have difficulty with these observations of Warner J. They occur in the part of his judgment dealing with the first of the three questions that he had posed (para 72 above) and probably they must be read in that context. Moreover the last sentence at p1626 A-B suggests that Warner J was not clearly distinguishing the category of excessive execution in Hastings-Bass itself from the category of inadequate deliberation relevant to the issue before him. If his remarks cannot be limited to their context then I would say that Warner J was wrong in disregarding the 'amount or quality' of professional advice obtained by trustees, when the question relates to a decision within the scope of the trustees' powers.

83. Mr Ham submitted that a refusal to attribute to trustees fault on the part of their advisers or agents leads to counter-intuitive and arbitrary distinctions. He instanced an error in a letter setting out the settlor's wishes (a variation of the facts in Barr). On that particular example, such an error might be a sufficient ground for a voluntary disposition to be set aside on the ground of mistake, regardless of where responsibility for the error lay. But I would accept that there have been, and no doubt will be in the future, cases in which small variations in the facts lead to surprisingly different outcomes. That is inevitable in an area where the law has to balance the need to protect beneficiaries against aberrant conduct by trustees (the policy behind the Hastings-Bass rule) with the competing interests of legal certainty, and of not imposing too stringent a test in judging trustees' decision-making.

84. There is indeed a striking contrast between the court's conclusions as to the position of Mr Ward-Thompson in Barr and that of Mr Cutbill in Futter. Mr Ward-Thompson's position was considered in detail by Lightman J [2003] Ch 409, para 27:

'He was the one point of contact between on the one side the settlor and on the other side C and L, C and L Isle of Man, the trustee and the protector. For all practical purposes he was the emanation and only representative of C and L, C and L Isle of Man, the trustee and the protector in all their dealings with the settlor. C and L was through itself and its associated firm, C and L Isle of Man and its vehicles, the trustee and the protector, providing the settlor with a total corporate and trust holding service. As is common ground the solicitors who drafted the appointment were acting on behalf of the trustee: Mr Ward-Thompson in giving instructions for its preparation in the circumstances can only have done so acting as agent for the trustee.'

These findings (based, it appears, on witness statements not tested by cross-examination) show that it was an unusual situation in which Mr Ward-Thompson had an exceptionally important role. The judge's conclusion was reinforced by another passage in para 27:

'I should add that my view is reinforced by the consideration that any ambiguity in the structure and arrangements ought to be resolved in favour of the settlor: (1) the 'C and L' side were responsible for the structure and arrangements; (2) Mr Ward-Thompson has declined to assist the court; and (3) the trustee perhaps surprisingly failed to seek from the settlor an expression of his wishes in documentary form or provide him with a copy of the proposed appointment before it was executed. In short, on the material before me, on the third issue I am satisfied that the trustee failed in its fiduciary duty to ascertain the true wishes of the settlor to which the appointment was intended to give effect and accordingly the rule is brought into play.'

85. Cases of this sort will call for detailed fact-finding by the judge, and sometimes no doubt for cross-examination. Barr may be contrasted with Abacus Trust Company (Isle of Man) Ltd v NSPCC [2001] STC 1344, in which an artificial tax-avoidance scheme failed because a deed of appointment was executed on 3 April 1998, contrary to the clear advice of leading counsel that it should not be executed until after the end of the 1997-98 financial year. On Wednesday, 1 April 1998 the appointment was faxed to the corporate trustee in the Isle of Man by an English solicitor with the suggestion that it should 'be executed on Friday' (naturally taken as 3 April). But a director of the corporate trustee had attended the consultation with leading counsel, and had received a note of it, which he did not refer to when he received the fax. Patten J applied the Hastings-Bass rule without finding it necessary to reach any clear conclusion about breach of duty, which was not then recognised as an essential requirement.

86. In Futter Mr Cutbill, a partner in a London firm of solicitors, was involved both as a trustee and as a solicitor advising the trustees. The facts as to his involvement were found at first instance by Norris J [2010] STC 982. It so happened, as already mentioned, that the judgment of Mr Engelhart QC in Pitt was given on 18 January 2010, the first day of the hearing in Futter. In Pitt it had been common ground ([2010] 1 WLR 1199, para 22) that there was no need to identify a breach of duty by the trustees. It is not clear from Norris J's judgment whether the same incorrect concession was made and accepted before him. But Norris J seems to have accepted Sieff v Fox as the leading authority from which to obtain guidance, and Barr received only a passing mention (on the void or voidable? issue) in his judgment. Norris J did not therefore make any clear finding about breach of fiduciary duty. He simply recorded and accepted Mr Cutbill's written evidence, which included the statement that '[Mr Futter] and I failed to pay any regard to the provisions of section 2(4) [TCGA] at the time, and therefore failed to consider the full tax implications.'

87. The Court of Appeal was therefore in as good a position as Norris J to make a finding about breach of duty on the part of the trustees. This Court has before it all the written evidence and exhibits that were before the judge, and in the absence of concurrent findings below it is also in a position to make findings, if necessary, on that issue. I will return to it below when dealing with the disposal of the Futter appeal.

88. Finally, on this part of the case, there is the submission that the trustees' duty to take account of relevant considerations is to be interpreted as a duty to act on advice only if it is correct “ in effect, a duty to come to the right conclusion in every case. I have left this submission until the end because it is to my mind truly a last-ditch argument. It involves taking the principle of strict liability for ultra vires acts (paras 81 to 84 above) out of context and applying it in a different area, so as to require trustees to show infallibility of judgment. Such a requirement is quite unrealistic. It would tip the balance much too far in making beneficiaries a special favoured class, at the expense of both legal certainty and fairness. It is contrary to the well-known saying of Lord Truro LC in In re Beloved Wilkes's Charity (1851) 3 Mac and G 440, 448:

'. . . that in such cases as I have mentioned it is to the discretion of the trustees that the execution of the trust is confided, that discretion being exercised with an entire absence of indirect motive, with honesty of intention, and with a fair consideration of the subject. The duty of supervision on the part of this court will thus be confined to the question of the honesty, integrity, and fairness with which the deliberation has been conducted, and will not be extended to the accuracy of the conclusion arrived at, except in particular cases.'

The trustees' duty does not extend to being right ('the accuracy of the conclusion arrived at') on every occasion. The 'particular cases' that the Lord Chancellor had in mind may have included cases concerned with the maintenance of minor beneficiaries. They may also have included cases (such as Kerr v British Leyland (Staff) Trustees Ltd) in which the trustees have to make a particular factual judgment, rather than exercise a wide discretion.

89. As a first footnote on the topic of fault, I would mention that in para 128 of his judgment, Lloyd LJ observed that a claim by beneficiaries against trustees may often be precluded by an exoneration clause in the trust instrument. Mr Philip Jones QC (for the Revenue) disputed this, pointing out that even if a trustee is exonerated from liability to pay equitable compensation, he is still liable to injunctive relief to prevent a threatened breach of trust, and personal and proprietary remedies may be available against persons who receive assets distributed in breach of trust. Moreover an exoneration clause does not protect a trustee against removal from office by order of the court. The Futter No 3 and No 5 settlements contain exoneration clauses in conventional terms, stating that 'in the professed execution of the trusts and powers hereof no trustee shall be liable for a breach of trust arising from a mistake or omission made by him in good faith.' I would not treat that clause as ousting the application of the Hastings-Bass rule, if it were otherwise applicable.

90. As a second footnote, there was some discussion in the course of argument as to the significance, in situations of this sort, of a possible claim for damages against professional advisers for financial loss caused by incorrect advice (or incorrect implementation of instructions). Mr Nugee referred to Walker v Geo H Medlicott and Son [1999] 1 WLR 727, in which a strong Court of Appeal dismissed on two grounds a claim for damages for professional negligence in preparing a will. The second ground was that the aggrieved claimant's proper course was to mitigate his damage by seeking rectification of the will. That decision may reflect the court's view of the particular facts of the case, and the feeling that if the drafting of the will had gone wrong other beneficiaries under it should not obtain adventitious benefits. In principle the possibility that trustees may have a claim for damages should have no effect on the operation of the Hastings-Bass rule. In practice it will be rare for trustees to have so strong a claim that they can be confident of obtaining a full indemnity for their beneficiaries' loss and their own costs. In the Pitt case this court was told that the claim against Frenkel Topping has been settled. Had it gone to trial the claim, even if successful in establishing duty and breach, might have faced difficulties over causation, since Mrs Pitt executed the SNT under the authority of an order of the Court of Protection, which had considered its terms. That court's apparent lack of awareness of the importance of section 89 of the Inheritance Act 1984 is one of the most remarkable features of the whole sorry story.

Would or Might?

91. In his statement of the correct principle (para 127 of the judgment, set out in para 70 above) Lloyd LJ did not provide an answer to the 'would or might?' debate. That was not, I think, an oversight. The Hastings-Bass rule is centred on the failure of trustees to perform their decision-making function. It is that which founds the court's jurisdiction to intervene if it thinks fit to do so. Whether the court will intervene is another matter. Buckley LJ's statement of principle in Hastings-Bass (para 24 above) cannot be regarded as clear and definitive guidance, since Buckley LJ was considering a different matter “ the validity of a severed part of a disposition, the other part of which was void for perpetuity. In Mettoy itself the trustees had wholly failed to consider (or even to be aware of) an important change in the new rules (affecting the destination of surplus in a winding-up of the scheme), at a time when winding-up was a real possibility. But Warner J (applying Buckley LJ's 'would not' formulation) declined to set aside the adoption of the new rules, because the power over surplus remained a fiduciary power.

92. It has been suggested (partly in order to accommodate the decision of the Court of Appeal in Stannard, para 34 above) that 'would not' is the appropriate test for family trusts, but that a different 'might not' test (stricter from the point of view of the trustees, less demanding for the beneficiaries) is appropriate for pensions trusts, since members of a pension scheme are not volunteers, but have contractual rights. That is an ingenious suggestion, and in practice the court may sometimes think it right to proceed in that way. But as a matter of principle there must be a high degree of flexibility in the range of the court's possible responses. It is common ground that relief can be granted on terms. In some cases the court may wish to know what further disposition the trustees would be minded to make, if relief is granted, and to require an undertaking to that effect (see In re Baden's Deed Trusts [1971] AC 424, referred to in para 63 above). To lay down a rigid rule of either 'would not' or 'might not' would inhibit the court in seeking the best practical solution in the application of the Hastings-Bass rule in a variety of different factual situations.

Void or Voidable?

93. Counsel on both sides readily admitted that they had hesitated over this point, but in the end they were all in agreement that Lloyd LJ was right in holding (para 99) that,

'if an exercise by trustees of a discretionary power is within the terms of the power, but the trustees have in some way breached their duties in respect of that exercise, then (unless it is a case of a fraud on the power) the trustees' act is not void but it may be voidable at the instance of a beneficiary who is adversely affected.'

In my judgment that is plainly right, and in the absence of further argument on the point it is unnecessary to add much to it. The issue has been clouded, in the past, by the difficult case of Cloutte v Storey, a case on appointments that are fraudulent in the equitable sense, that is made for a positively improper purpose. Here we are concerned not with equitable fraud, nor with dispositions which exceed the scope of the power, or infringe the general law (such as the rule against perpetuity). We are in an area in which the court has an equitable jurisdiction of a discretionary nature, although the discretion is not at large, but must be exercised in accordance with well-settled principles.

94. The working-out of these principles will raise problems which must be dealt with on a case by case basis. The mistake claim in Pitt involves a problem of that sort. But it is unnecessary and inappropriate to prolong what is already a very long judgment by further discussion of problems that are not now before this court.

Disposal of the Hastings-Bass issues

95. In Futter the essential issue was whether the trustees of the No 3 and No 5 settlements, in deciding to take the steps which they took in March and April 2008, failed in their duty to take relevant considerations into account. Capital gains tax was a relevant consideration. Indeed, it is fairly plain that it was the paramount consideration, and the trustees thought about it a great deal. But the tax advice which they received and acted on was wrong, because an amendment to section 2(4) of TCGA had been overlooked. As Lloyd LJ put it succinctly (para 138):

'They did not overlook the need to think about CGT. They were given advice on the right point. The problem was that the advice was wrong.'

96. The only complication was that Mr Cutbill (the solicitor-trustee) was a member of both teams: the solicitors giving the erroneous advice, and the trustees receiving and acting on it. I agree with the Court of Appeal that it would be artificial to distinguish between the two trustees, who acted together in making and effectuating their decisions. I would if necessary go further and hold that the documentary evidence indicates that most if not all of the technical tax advice given by his firm came not from Mr Cutbill but from the assistant solicitor who was working with him, from January 2008, in a review of a number of different Futter family settlements. Until 27 March 2008 it was supposed, wrongly, that the No 3 settlement's stockpiled gains were relatively small, and the fact that they amounted to about £188,000 led to a last-minute change of plan. On 28 March 2008 there was a telephone conversation between the assistant solicitor and Mr Bunce, Mr Futter's accountant, at which, without reference to Mr Cutbill, she definitely confirmed that Mr Futter's personal losses could be set off against the section 87 gains. Mr Cutbill seems to have been, very properly, reluctant to put the blame on a junior member of his firm, and of course his firm must take legal responsibility for any actionable mistake by any of its fee-earners. But the documents in exhibit 'CDC 1' to Mr Cutbill's witness statement tend to confirm that he should not, as a trustee of the No 3 and No 5 settlements, be treated as having been personally in breach of fiduciary duty.

97. In Pitt the position was even clearer. As her husband's receiver under the Mental Health Act 1983 Mrs Pitt was in a fiduciary position but there is no suggestion that she had any professional qualifications. She devoted herself, alternating with a carer, to looking after her disabled husband. As anyone in that position would, she took professional advice from solicitors and specialist consultants. After hearing from her legal advisers and the Official Solicitor the Court of Protection made an order on 1 September 1994 authorising (not directing) her to execute the SNT and she acted on that authority on 1 November 1994 (the date in para 161 of Lloyd LJ's judgment seems to be an error; compare para 151). She had taken supposedly expert advice and followed it. There is no reason to hold that she personally failed in the exercise of her fiduciary duty. Unfortunately the advice was unsound.

98. I would therefore dismiss the appeal in Futter, and the appeal in Pitt so far as it turns on the Hastings-Bass rule.

RESCISSION ON THE GROUND OF MISTAKE

Mrs Ogilvie's litigation

99. In this part of his judgment Lloyd LJ began with the litigation conducted by Mrs Ogilvie at the end of the 19th century. Mrs Ogilvie was a very rich widow who had in 1887 executed two deeds settling large funds for charitable purposes. She was described by Byrne J (in the transcript included in the appendix printed for the eventual appeal to the House of Lords, pp 862-863) as

'undoubtedly a good woman of business, shrewd, clever and intelligent, having been accustomed to assist her husband in business matters. She had a proper sense of the responsibilities of great wealth, she was charitable and munificent. She had strong views on certain subjects, was impatient of any attempt to thwart or control her, and though perhaps at times inclined to be somewhat changeable, she was firm and decided as to her course of action when she had made up her mind and laid down what she terms her law in respect to any matter.'

Seven years later she brought an action to have the deeds set aside. She relied on grounds summarized by Byrne J (p 862) as follows:

'(1) That she had not preserved to her the absolute power of disposing of the capital, including the land, as she thought fit during her life, and that notwithstanding express instructions to the contrary. (2) That she had not a similar absolute power in respect to income. (3) That she is liable to interference by the Charity Commissioners and by her own Trustees, and to be called upon by them to account for her administration of the income and that notwithstanding express instructions to the contrary. (4) That she has not the power to apply moneys originally dedicated to London institutions to Suffolk institutions. (5) That she was not fully and properly advised and that she did not fairly understand the nature and effect of the documents she executed.'

100. These grounds were fully explored in the pleadings, in interrogatories, and in cross-examination at the nine-day trial. Originally there was an alternative claim for rectification but her counsel did not rely on that claim, although it seems (pp 903-905) that the Attorney General (who appeared in person at every stage of the proceedings) made an open offer for the case to be disposed of uncontentiously on that basis. Byrne J gave a judgment, over 50 pages long in the transcript, in which he said (p 901),

'The case is entirely wanting in any of those elements of fraud, undue influence, concealments of facts from the donor, want of separate and independent advice, surprise or pressure, which, or some of which, are commonly to be met with in cases of attempts to set aside or rectify voluntary instruments.'

The judge rejected almost entirely the criticisms that Mrs Ogilvie directed towards her legal advisers:

'The utmost that could be suggested against Mr Smith is that he misunderstood his instructions, or that he was guilty of error of judgment in not having with more insistence determined to see his client personally, and against Mr Smith, Mr Harding, Mr Sutherland, and their counsel, that possibly they allowed their natural and perfectly unselfish wish to see the charitable scheme carried through to permit them to neglect informing the plaintiff of every trouble and difficulty of detail which arose in the matter.'

(Mr Smith was the London solicitor of Mrs Ogilvie, who lived in Suffolk; Mr Harding was a respected member of the Society of Friends, who gave her advice; and Mr Sutherland was her late husband's confidential clerk.) Her action was dismissed.

101. She appealed to the Court of Appeal, where in view of the trial judge's clear findings the argument seems to have been more closely focused as mistake. Giving the judgment of the Court of Appeal Lindley LJ said (Ogilvie v Littleboy (1897) 13 TLR 399, 400):

'Gifts cannot be revoked, nor can deeds of gift be set aside, simply because the donors wish that they had not made them and would like to have back the property given. Where there is no fraud, no undue influence, no fiduciary relation between donor and donee, no mistake induced by those who derive any benefit by it, a gift, whether by mere delivery or by deed, is binding on the donor . . . In the absence of all circumstances of suspicion a donor can only obtain back property which he has given away by showing that he was under some mistake of so serious a character as to render it unjust on the part of the donee to retain the property given to him.'

Mrs Ogilvie's grounds of complaint seem to have been revised a little. The alleged mistakes as to the application of capital or income for non-charitable purposes, and as to the jurisdiction of the Charity Commissioners, could not be sustained on the judge's findings. As to the fourth ground (relating to sales of land) the Court of Appeal held that 'the mistake, such as it was, cannot be regarded as so material as to affect the validity of the deeds.' The fifth ground had been reformulated as a failure by her advisers to warn her that members of the Society of Friends might be unwilling to become trustees. As to this Lindley LJ observed (p 401):

'But, assuming the danger to be real, assuming that it was an error of judgment not to call the plaintiff's attention to it, is such an omission enough to entitle her to have the deeds set aside? We are not aware of any legal principle which goes this length or anything like it. The complaint is not that her intentions have not been carried out; it is that a possible danger known to her advisers was not pointed out to her.'

102. So the appeal was dismissed, as was a further appeal to the House of Lords (Ogilvie v Allen (1899) 15 TLR 294). Lord Halsbury LC said (p 295):

'The appellant, a lady, was desirous of establishing certain charities, and she now contends that, though she did intend to devote her money to charity, certain specific intentions as to management, control, independence of control, and the like were such essential and important considerations to her mind that in these respects she was misled, and now seeks to get rid of the effect of her deeds upon that allegation. Such questions, doubtless, may arise under circumstances where misunderstanding on both sides may render it unjust to the giver that the gift should be retained. It appears to me that there are no such circumstances here. I entirely concur with the judgment delivered by the present Master of the Rolls . . .'

So did Lord Macnaghten, who said that Lindley LJ's judgment 'deals with the case so fully and so satisfactorily that there is nothing more to be added.' Lord Morris concurred.

103. Lloyd LJ reviewed and discussed other 19th century and modern authorities, including the first-instance decisions in Gibbon v Mitchell [1990] 1 WLR 1304 and In re Griffiths, decd [2009] Ch 162. He questioned the result in the latter case. The framework of his conclusion (paras 210 and 211) was that for the exercise of the equitable jurisdiction to set aside a voluntary disposition there must be (1) a mistake, which is (2) of the relevant type and (3) sufficiently serious to satisfy the Ogilvie v Littleboy test. That is a convenient framework against which to consider the authorities, although there is obviously some overlap between the three heads. In general a mistake as to the essential nature of a transaction is likely to be more serious than a mistake as to its consequences.

What is a mistake?

104. For present purposes a mistake must be distinguished from mere ignorance or inadvertence, and also from what scholars in the field of unjust enrichment refer to as misprediction (see Seah, Mispredictions, Mistakes and the Law of Unjust Enrichment [2007] RLR 93; the expression may have first received judicial currency in Dextra Bank and Trust Co Ltd v Bank of Jamaica [2002] 1 All ER (Comm) 193). These distinctions are reasonably clear in a general sort of way, but they tend to get blurred when it comes to facts of particular cases. The editors of Goff and Jones, The Law of Unjust Enrichment, 8th ed. (2011) para 9-11 comment that the distinction between mistake and misprediction can lead to 'some uncomfortably fine distinctions,' and the same is true of the distinction between mistake and ignorance.

105. Forgetfulness, inadvertence or ignorance is not, as such, a mistake, but it can lead to a false belief or assumption which the law will recognise as a mistake. The Court of Appeal of Victoria has held that mistake certainly comprehends 'a mistaken belief arising from inadvertence to or ignorance of a specific fact or legal requirement': Ormiston JA in Hookway v Racing Victoria Ltd [2005] VSCA 310, (2005) 13 VR 444, 450. That case was on the borderline between voluntary disposition and contract. It concerned prize money for a horse race which was paid to the wrong owner because the official in charge of prize money was ignorant of a recent change in the rules of racing (permitting an appeal against disqualification after a drugs test). He made a mistake as to the real winner.

106. The best-known English authority on this point is Lady Hood of Avalon v Mackinnon [1909] 1 Ch 476. Under a settlement Lord and Lady Hood had a joint power of appointment, and later Lady Hood as the survivor had a sole power of appointment, in favour of the children and remoter issue of their marriage. They had two daughters. In 1888 half the trust fund had been appointed (subject to the prior life interests of Lord and Lady Hood) to their elder daughter on her marriage, and had been resettled by her. In 1902 and 1904, after Lord Hood's death, Lady Hood appointed a total of £8,600 to her younger daughter. Then, wishing to achieve equality, as she thought, between her daughters, and entirely forgetting the 1888 appointment, she appointed a further £8,600 to her elder daughter (so inevitably producing inequality, unless the appointment were set aside). The elder daughter did not oppose Lady Hood's action for rescission of the last appointment, but the trustees of the resettlement (which contained an after-acquired property covenant) did oppose it. Eve J granted relief, stating (pp 483-484):

'Having regard to the facts which I have stated, I must assume that Lady Hood, intending only to bring about equality between her daughters, was labouring under a mistake when she thought that equality would be brought about by the execution of the deed appointing £8,600 to her elder daughter. It was obviously a mistake, because the effect of the execution of that deed was to bring about that which Lady Hood never intended and never contemplated.'

107. In his judgment Eve J referred at length to the decision of the Court of Appeal in Barrow v Isaacs and Son [1891] 1 QB 417. In that case the Court of Appeal declined to grant relief, on the ground of mistake, from forfeiture of a lease for breach of a covenant against underletting. The solicitors acting for a respectable tenant had overlooked the covenant and the premises had been sublet to a respectable sub-tenant. Both Lord Esher MR and Kay LJ commented that there was no legal definition of 'mistake'. Lord Esher MR said (at pp 420-421) that the head tenant had had 'a mere passive state of mind':

'I should say that mere forgetfulness is not mistake at all in ordinary language. I cannot find any decision in Courts of Equity which has ever stated that mere forgetfulness is mistake against which equity would relieve.'

But Kay LJ (with whom Lopes LJ agreed) seems to have taken the view that there was a mistake which equity had power to relieve, although in the circumstances of the case the court declined to grant relief. The power to relieve would, it seems, have been based on the conscious belief or tacit assumption that the underletting was lawful.

108. The fullest academic treatment of this topic is in Goff and Jones at paras 9-32 to 9-42. The editors distinguish between incorrect conscious beliefs, incorrect tacit assumptions, and true cases of mere causative ignorance ('causative' in the sense that but for his ignorance the person in question would not have acted as he did). The deputy judge's first-instance decision in Pitt [2010] 1 WLR 1199, para 50 is suggested as an example of mere causative ignorance: 'If someone does not apply his mind to a point at all, it is difficult to say that there has been some real mistake about it'. The Court of Appeal adopted a different view of the facts, treating the case (para 216) as one of an incorrect conscious belief on the part of Mrs Pitt that the SNT had no adverse tax consequences. The editors of Goff and Jones are, on balance, in favour of treating mere causative ignorance as sufficient. They comment (at para 9-41, in answering a 'floodgates' objection):

'. . . denying relief for mere causative ignorance produces a boundary line which may be difficult to draw in practice, and which is susceptible to judicial manipulation, according to whether it is felt that relief should be afforded “ with the court's finding or declining to find incorrect conscious beliefs or tacit assumptions according to the court's perception of the merits of the claim.'

It may indeed be difficult to draw the line between mere causative ignorance and a mistaken conscious belief or a mistaken tacit assumption. I would hold that mere ignorance, even if causative, is insufficient, but that the court, in carrying out its task of finding the facts, should not shrink from drawing the inference of conscious belief or tacit assumption when there is evidence to support such an inference. I shall return (paras 131 and 132 below) to the suggestion that this may involve 'judicial manipulation.'

109. A misprediction relates to some possible future event, whereas a legally significant mistake normally relates to some past or present matter of fact or law. But here too the distinction may not be clear on the facts of a particular case. The issue which divided the House of Lords in Kleinwort Benson Ltd v Lincoln City Council [1999] 2 AC 349 was whether (as Lord Hoffmann put it at p398) the correct view was that, 'a person who pays in accordance with what was then a settled view of the law has not made a mistake' and 'that his state of mind could be better described as a failure to predict the outcome of some future event (sc a decision of this House) than a mistake about the existing state of the law.' There is another interesting discussion of this point in the judgments given in the Court of Appeal in Brennan v Bolt Burdon [2005] QB 303.

110. A problem about the boundary between mistake and misprediction arose in In re Griffiths, decd [2009] Ch 162, a decision of Lewison J. Like Sieff v Fox and some other cases on the Hastings-Bass rule, it was a case in which the Revenue was invited to intervene but declined to do so, despite the large sum of inheritance tax at stake. The case was therefore heard without adversarial argument as to the law or the facts. Lloyd LJ commented (para 198) that he did not criticise the judge, given the limited argument before him, but that he did question his conclusion. I agree with both limbs of that comment.

111. It is important to note the sequence of events in In re Griffiths. Mr Griffiths had a valuable holding in Iota, a property company (whose shares did not attract business assets relief). He was aged 73 when, in January 2003, he and his wife took advice about tax planning. They received a lengthy report setting out various options. Most involved making potentially exempt transfers, which progressively reduce inheritance tax on qualifying gifts if the donor survives for three years, and avoid tax entirely if the donor survives for seven years after making the gift. The report recommended that seven-year term insurance cover should be obtained. Mr Griffiths decided to take various steps, the most important of which was a settlement of Iota shares worth over £2.6m. This was effected by a two-stage process which was completed in February 2004. He decided not to obtain term insurance. Unfortunately he was diagnosed with lung cancer in October 2004, and died in April 2005. Had he done nothing, the Iota shares would have formed part of his residuary estate, in which his wife took a life interest, and no inheritance tax would have been payable on his death.

112. In those circumstances his executor commenced proceedings asking that the dispositions should be set aside on the ground of mistake ([2009] Ch 162, para 6):

'The relevant mistake on which they rely is that Mr Griffiths mistakenly believed, at the time of the transfers, that there was a real chance that he would survive for seven years, whereas in fact at that time his state of health was such that he had no real chance of surviving that long.'

The medical evidence (in the form of letters from his general practitioner, from a consultant oncologist and from a consultant rheumatologist) was inconclusive, but the GP expressed the view that it was 'extremely unlikely' that the cancer was present in April 2003. On this evidence the judge found that in April 2003 Mr Griffiths had a life expectancy of between seven and nine years. He went on to observe (para 18):

'It is unfortunate that in a case involving £1m-worth of tax a proper medical report was not placed before the court and that the claimants are compelled to rely on a single sentence in a letter from [the oncologist]. Although I have hesitated about this finding, I am prepared to find, by a narrow margin that he was suffering from lung cancer on 3 February 2004; and that following the onset of lung cancer at that time his life expectancy did not exceed three years in February 2004. Had the facts been contested, I might not have felt able to make this finding.'

113. On the rather uncertain foundation of that finding the judge decided that the assignment of 3 February 2004 should be set aside (para 30):

'By that time Mr Griffiths was suffering from lung cancer about which he was unaware. He did therefore make a mistake about his state of health. Had he known in February 2004 that he was suffering from lung cancer he would also have known that his chance of surviving for three years, let alone for seven years, was remote. In those circumstances I am persuaded that he would not have acted as he did by transferring his reversionary interest in the shares to trustees.'

The judge did not say whether this was (in the Goff and Jones formulation) an incorrect conscious belief or an incorrect tacit assumption. The editors of that work (para 9-36) treat it as a tacit assumption but it seems close to the residual category of mere causative ignorance. Had the judge not made his hair's breadth finding about the presence of cancer in February 2004 it would have been a case of misprediction, not essentially different from a failure to predict a fatal road accident. Lloyd LJ observed (para 198) that it was strongly arguable that, having declined to follow the financial consultants' recommendation of term insurance, Mr Griffiths was taking the risk of deterioration of his health and failure to survive the statutory period.

What type of mistake?

114. Some uncontroversial points can be noted briefly. It does not matter if the mistake is due to carelessness on the part of the person making the voluntary disposition, unless the circumstances are such as to show that he deliberately ran the risk, or must be taken to have run the risk, of being wrong. (There is an illuminating discussion of this point in Lord Hoffmann's speech in Deutsche Morgan Grenfell Group plc v Inland Revenue Commissioners [2007] 1 AC 558, paras 24-30). Nor need the mistake be known to (still less induced by) the person or persons taking a benefit under the disposition. The fact that a unilateral mistake is sufficient (without the additional ingredient of misrepresentation or fraud) to make a gift voidable has been attributed to gifts being outside the law's special concern for the sanctity of contracts (O'Sullivan, Elliott and Zakrzewski, The Law of Rescission (2007) para 29.22):

'It is apparent from the foregoing survey that vitiated consent permits the rescission of gifts when unaccompanied by the additional factors that must be present in order to render a contract voidable. The reason is that the law's interest in protecting bargains, and in the security of contracts, is not engaged in the case of a gift, even if made by deed.'

Conversely, the fact that a purely unilateral mistake may be sufficient to found relief is arguably a good reason for the court to apply a more stringent test as to the seriousness of the mistake before granting relief.

115. The Revenue's printed case (paras 70 to 74) seeks to play down the distinction between mistake in the law of contract and its role in equitable rescission of voluntary dispositions. It seeks to build boldly on the decision of the Court of Appeal in Great Peace Shipping Ltd v Tsavliris Salvage (International) Ltd (The Great Peace) [2002] EWCA Civ 1407, [2003] QB 679, which did not follow (and has effectively overruled) Solle v Butcher [1950] 1 KB 671. The argument is that logic requires that a deed which transfers property for no consideration can be set aside only for a mistake of a fundamental nature that would render a contract void. Mr Jones did not cite any authority for this heterodox submission, and there is high authority (starting with Ogilvie v Allen) against it. Equity will grant specific performance of a covenant only if it is supported by valuable consideration. This includes the marriage consideration, but only if the covenant is being enforced by or on behalf of a person or persons within the scope of the marriage consideration. The traditional rules of equity were considered and explained by the Court of Appeal in Attorney-General v Jacobs [1895] 2 KB 341, an account duty case, and In re Cook's Settlement Trusts [1965] Ch 902, a decision of Buckley J, illustrates their application within living memory. They are necessary to the understanding of cases like Ellis v Ellis (1909) 26 TLR 166, where the after-acquired property covenant in Mrs Ellis's marriage settlement was enforceable in equity, because there were children of the marriage. But the notion that any voluntary disposition should be accorded the same protection as a commercial bargain, simply because it is made under seal, is insupportable.

116. Leaving aside for the present the degree of seriousness of the mistake, there is also controversy about its nature (or characteristics), especially as to the distinction between 'effect' and 'consequences' drawn by Millett J in Gibbon v Mitchell [1990] 1 WLR 1304. In that case two funds (May's fund and Henry's fund) were settled in 1946 by Mr Henry Gibbon's parents on the occasion of the marriage of his sister May. The funds were settled on May and Henry respectively for life, on the statutory protective trusts in section 33 of the Trustee Act 1925 (with a modification in the case of Henry's fund), but with power for May to surrender her protected life interest so as to accelerate the interests of her children once they had attained vested interests. For some unknown reason there was no corresponding power in respect of Henry's fund. The consequence was that a purported surrender by Mr Gibbon would cause a forfeiture of his fixed interest and bring into operation a discretionary trust affecting income during the rest of his life.

117. In 1987 Mr Gibbon was a prosperous farmer aged 69, with two adult children. He wished to take steps to save inheritance tax and was advised by his accountants and solicitors to surrender his life interest, and at the same time release two powers of appointment, so as to accelerate his children's interests. This advice was expressed in terms of enabling Henry's fund to pass immediately to the two children. His professional advisers failed to recognise, until after the deed of surrender had been executed, that the protective trusts provided a trap. Mr Gibbon applied to the court to have the deed of surrender set aside on the ground of mistake, and also for relief under the Variation of Trusts Act 1958. Millett J set the surrender aside, and varied the trust by lifting the protective trusts. In his judgment he referred to several of the older authorities, in most of which solicitors had misunderstood or gone beyond their instructions: Meadows v Meadows (1853) 16 Beav 401, Walker v Armstrong (1856) 8 De G M and G 531, Ellis v Ellis (1909) 26 TLR 166 and In re Walton's Settlement [1922] 2 Ch 509. Ogilvie v Littleboy was not cited. Millett J set out the principle which he drew from them at p1309:

'In my judgment, these cases show that, wherever there is a voluntary transaction by which one party intends to confer a bounty on another, the deed will be set aside if the court is satisfied that the disponor did not intend the transaction to have the effect which it did. It will be set aside for mistake whether the mistake is a mistake of law or a fact, so long as the mistake is as to the effect of the transaction itself and not merely as to its consequences or the advantages to be gained by entering into it.'

It will be observed that this formulation does not include the Ogilvie v Littleboy requirement of seriousness, except so far as it might be argued that any mistake as to the effect of a disposition is likely to be relatively serious.

118. Millett J's judgment has been very influential. It is a mark of the high respect in which he is held that an extempore first-instance judgment, not (so far as appears from the judgment) based on much adversarial argument, is cited as one of the key authorities in most of the standard works on equity and trusts, including Snell, 32nd ed. (2010) 11-008, 22-052; Lewin, 18th ed. (2008) 4-58, 29-231; Underhill and Hayton, 18th ed. (2010) 15-28 to 15-34; and Thomas and Hudson, 2nd ed. (2010) 20.37. But the source from which Millett J's statement of principle is derived is far from clear and it has been the subject of some criticism, both from legal scholars and in more recent decisions of the court.

119. It is generally agreed that 'effect' must mean legal effect (in the sense of the legal character or nature of a transaction). In Dent v Dent [1996] 1 WLR 683, 693 the deputy judge (David Young QC) understood it as 'the purpose or object' of a transaction. Several other first-instance judges have commented that the distinction between 'effect' and 'consequences' is not always clear, including Davis J in Anker-Petersen v Christensen [2002] WTLR 313, 330. Lawrence Collins J went further in AMP (UK) plc v Barker [2001] WTLR 1237, para 70, saying of the distinction:

'If anything, it is simply a formula designed to ensure that the policy involved in equitable relief is effectuated to keep it within reasonable bounds and to ensure that it is not used simply when parties are mistaken about the commercial effects of their transactions or have second thoughts about them.'

On that view it comes close to Lindley LJ's more general requirement for the mistake to be serious. In Wolff v Wolff [2004] STC 1633, Mann J considered (para 23) that the test was not a limiting factor, and (para 26) noted that Lawrence Collins J had referred to commercial consequences, not legal consequences.

120. Lloyd LJ has now reviewed Gibbon v Mitchell twice, first in Sieff v Fox and then in Pitt v Holt. In Sieff v Fox, Ogilvie v Littleboy was brought to light after a long period of obscurity (though it is mentioned in Peter Birks' Introduction to the Law of Restitution, first published in 1985). Lloyd LJ noted (para 106) that a test based on the legal effect of a transaction could not cover the tax consequences of a transaction, but that Lindley LJ's more general test in Ogilvie v Littleboy might do so. He expressed no final view because of the special circumstances of the case before him ([2005] 1 WLR 3811, para 116).

121. In Pitt v Holt Lloyd LJ went further. He expressed the view (para 208) that some recent cases about offshore trusts did not accord with English law: Clarkson v Barclay's Private Bank and Trust (Isle of Man) Ltd [2007] WTLR 1703; In re Betsam Trust [2009] WTLR 1489; and In re A Trust [2009] JLR 447. He accepted the distinction made by Millett J in Gibbon v Mitchell but extended it (para 210) by formulating it as a requirement

'. . . that, for the equitable jurisdiction to set aside a voluntary disposition for mistake to be invoked, there must be a mistake on the part of the donor either as to the legal effect of the disposition or as to an existing fact which is basic to the transaction.'

This extension seems to have been primarily to accommodate cases such as Lady Hood of Avalon, where there was (para 206) 'a fundamental error of fact, in relation to a point which lay at the heart of the transaction.' He also seems to have had in mind the New Zealand case of University of Canterbury v Attorney-General [1995] 1 NZLR 78, which is discussed at para 199 of his judgment. The special feature of that case was that the University had to some extent encouraged, or at least failed to correct, the donor's error; it wished to return the gift but the Attorney-General, representing the public interest in charity, opposed that course. In addition, the mistake must, Lloyd LJ said, meet the Ogilvie v Littleboy test of sufficient gravity.

122. This approach has been criticised by the editors of Goff and Jones, paras 9-101 to 9-106. I do not agree with all these criticisms of what the editors refer to as the Court of Appeal's 'stricter, hybrid approach.' But I can see no reason why a mistake of law which is basic to the transaction (but is not a mistake as to the transaction's legal character or nature) should not also be included, even though such cases would probably be rare. If the Gibbon v Mitchell test is further widened in that way it is questionable whether it adds anything significant to the Ogilvie v Littleboy test. I would provisionally conclude that the true requirement is simply for there to be a causative mistake of sufficient gravity; and, as additional guidance to judges in finding and evaluating the facts of any particular case, that the test will normally be satisfied only when there is a mistake either as to the legal character or nature of a transaction, or as to some matter of fact or law which is basic to the transaction.

123. To confirm the Gibbon v Mitchell test as formulated by Millett J would in my view leave the law in an uncertain state, as the first-instance decisions mentioned in para 119 above tend to demonstrate. It would also be contrary to the general disinclination of equity to insist on rigid classifications expressed in abstract terms. Equity, unlike many continental systems, has not adopted Roman law's classification of mistakes: error in negotio (the nature of the intended transaction), error in corpore (the subject-matter of the transaction), error in persona (the identity of the other party to the transaction) and error in substantia (the quality of the subject-matter). The Gibbon v Mitchell test, at any rate if applied narrowly, would cover only the first of these categories. But in some situations errors in other categories may be just as basic and just as serious in their consequences.

The conscience test

124. Lindley LJ's test in Ogilvie v Littleboy, quoted at para 101 above, requires the gravity of the causative mistake to be assessed in terms of injustice “ or, to use equity's cumbersome but familiar term, unconscionableness. Similarly Millett J said in Gibbon v Mitchell [1990] 1 WLR 1304, 1310:

'Equity acts on the conscience. The parties [in] whose interest it would be to oppose the setting aside of the deed are the unborn future children of Mr Gibbon and the objects of discretionary trusts to arise on forfeiture, that is to say his grandchildren, nephews and nieces. They are all volunteers. In my judgment they could not conscionably insist upon their legal rights under the deed once they had become aware of the circumstances in which they had acquired them.'

125. The evaluation of what is or would be unconscionable must be objective. Millett J identified precisely the class of beneficiaries in whose interest it would be for the forfeiture to stand (apart from tax considerations which made it disadvantageous for the whole family), but he did not do so in order to embark on the impossible task of establishing the state of the consciences of minor and unborn beneficiaries. Nor (apart from a defence of change of position) would the relative prosperity of the donor and the donees be relevant, except so far as it was part of the mistake (as in Lord Scott's example in Deutsche Morgan Grenfell Group Plc v Inland Revenue Courts [2007] 1 AC 558, para 87: 'A gift of £1,000 by A to B where B is believed to be impecunious but is in fact a person of substantial wealth').

126. The gravity of the mistake must be assessed by a close examination of the facts, whether or not they are tested by cross-examination, including the circumstances of the mistake and its consequences for the person who made the vitiated disposition. Other findings of fact may also have to be made in relation to change of position or other matters relevant to the exercise of the court's discretion. Justice Paul Finn wrote in a paper, Equitable Doctrine and Discretion in Remedies published in Restitution: Past, Present and Future (1998):

'The courts quite consciously now are propounding what are acceptable standards of conduct to be exhibited in our relationships and dealings with others . . . A clear consequence of this emphasis on standards (and not on rules) is a far more instance-specific evaluation of conduct.'

The injustice (or unfairness or unconscionableness) of leaving a mistaken disposition uncorrected must be evaluated objectively, but with an intense focus (in Lord Steyn's well-known phrase in In re S (A Child) [2005] 1 AC 593, para 17) on the facts of the particular case. That is why it is impossible, in my view, to give more than the most tentative answer to the problems posed by Professor Andrew Burrows in his Restatement of the English Law of Unjust Enrichment (2013) p 66: we simply do not know enough about the facts.

127. I add a postscript as to the criticism made by the editors of Goff and Jones (para 9-41), already quoted at para 108 above, of 'a boundary line which may be difficult to draw in practice, and which is susceptible to judicial manipulation, according to whether it is felt that relief should be afforded “ with the court's finding or declining to find incorrect conscious beliefs or tacit assumptions according to the court's perception of the merits of the claim.' There is some force in this, although the term 'manipulation' is a bit harsh. The fact that a unilateral mistake is sufficient means that the court may have to make findings as to the state of mind, at some time in the past, of a claimant with a lively personal interest in establishing that there was a serious causative mistake. This will often be a difficult task. But as a criticism of the Court of Appeal in Pitt I would reject it. The case was heard on affidavit evidence, without cross-examination, and the Court of Appeal was in as good a position as the deputy judge to draw inferences and make findings of fact.

128. More generally, the apparent suggestion that the court ought not to form a view about the merits of a claim seems to me to go wide of the mark. In a passage in Gillett v Holt [2001] Ch 210, 225, since approved by the House of Lords (see especially the speech of Lord Neuberger, with which the rest of the House agreed, in Fisher v Brooker [2009] 1 WLR 1764, para 63) I said in discussing proprietary estoppel that although its elements (assurance, reliance and detriment) may have to be considered separately they cannot be treated as watertight compartments:

'. . . the fundamental principle that equity is concerned to prevent unconscionable conduct permeates all the elements of the doctrine. In the end the court must look at the matter in the round.'

In my opinion the same is true of the equitable doctrine of mistake. The court cannot decide the issue of what is unconscionable by an elaborate set of rules. It must consider in the round the existence of a distinct mistake (as compared with total ignorance or disappointed expectations), its degree of centrality to the transaction in question and the seriousness of its consequences, and make an evaluative judgment whether it would be unconscionable, or unjust, to leave the mistake uncorrected. The court may and must form a judgment about the justice of the case.

Mistakes about tax

129. In this court Mr Jones applied for and obtained permission to raise two points which had not been raised below. The first (to be found in paras 80 to 95 of the Revenue's case) was that a mistake which relates exclusively to tax cannot in any circumstances be relieved. This submission, for which no direct authority was cited, was said to be based on Parliament's general intention, in enacting tax statutes, that tax should be paid on some transaction of a specified type, whether or not the taxpayer is aware of the tax liability. Mistake of law is not a defence, Mr Jones submitted, to tax lawfully due and payable.

130. In my opinion that submission begs the question, since if a transaction is set aside the Court is in effect deciding that a transaction of the specified description is not to be treated as having occurred. In the case of inheritance tax, this is expressly provided by section 150 of the Inheritance Tax Act 1984. That section is expressed in general terms as applying where a transfer 'has by virtue of any enactment or rule of law been set aside as voidable or otherwise defeasible', and the effect is that tax which would not have been paid or payable 'if the relevant transfer had been void ab initio' is to be repaid, or cease to be payable. There is no exception in section 150 for avoidance on the ground of a mistake about tax. More generally, Mr Jones's submission that tax is somehow in a different category is at odds with the approach of the House of Lords in Deutsche Morgan Grenfell [2007] 1 AC 558: see the speech of Lord Hope at para 44 and my own observations at paras 133 and 140.

131. So far as Mr Jones cites any authority for his submission, he has referred, but only as an aside, to the decision of the Court of Appeal in Racal Group Services Ltd v Ashmore [1995] STC 1151. That was a claim to rectification. Rectification is a closely guarded remedy, strictly limited to some clearly-established disparity between the words of a legal document, and the intentions of the parties to it. It is not concerned with consequences. So far as anything in Racal is relevant to the different equitable remedy of rescission on the ground of mistake, it is relevant, not to establishing the existence of a mistake, but to the court's discretion to withhold relief in cases where it would be inappropriate for the court to grant it. That is Mr Jones's second new point and it is considered below.

132. I would therefore reject the first new point as much too wide, and unsupported by principle or authority. But it is still necessary to consider whether there are some types of mistake about tax which should not attract relief. Tax mitigation or tax avoidance was the motive behind almost all of the Hastings-Bass cases that were concerned with family trusts (as opposed to pensions trusts). In Gibbon v Mitchell there was a mistake as to the legal effect of the transaction, which was to plunge the family into the trap of forfeiture under the protective trusts, rather than to achieve the immediate acceleration of the adult children's interests. But the seriousness of the consequences of the mistake was greatly enhanced by the inheritance tax implications. On the test proposed above, consequences (including tax consequences) are relevant to the gravity of a mistake, whether or not they are (in Lloyd LJ's phrase) basic to the transaction.

133. In Pitt the special tax advantage available under section 89 of the Inheritance Tax Act 1984 was a valuable one, and its loss was certainly a serious matter for Mrs Pitt, both as her husband's receiver and on her own account as his wife and carer and as the eventual beneficiary of his estate. Lloyd LJ accepted that (para 215). He was also prepared to accept (para 216) that Mrs Pitt had an incorrect conscious belief, or made an incorrect tacit assumption, that the proposed SNT (which had been the subject of advice from two professional firms, and approved by the Court of Protection) had no adverse tax effects.

134. It was on the issue of 'mistake as to effect or as to consequence?' (para 217) that Lloyd LJ felt obliged to withhold relief. He saw the tax liability, even though it was immediate and backed by a statutory charge (imposed by section 237 of the Inheritance Tax Act 1984) on the property of the SNT, as no more than a consequence (para 218):

'The legal effect [of the disposition] was the creation of the Special Needs Trust, on its particular terms, and the fact that the lump sum and the annuity were settled upon those terms.'

An irony of the situation is that if the SNT had been framed so as to comply with section 89 (requiring at least half of the property applied during Mr Pitt's lifetime to be applied for his benefit) it would most probably have made no difference to the distribution of capital or income during his lifetime (as the deputy judge noted in para 13 of his judgment, in dismissing a Revenue argument that Mrs Pitt might have decided not to take advantage of section 89). It has not been suggested that the primary purpose of the SNT was other than Mr Pitt's welfare and benefit, and the maintenance of his wife as his carer. The SNT could have complied with section 89 without any artificiality or abuse of the statutory relief. It was precisely the sort of trust to which Parliament intended to grant relief by section 89.

135. In Futter this court declined to permit the appellants to raise for the first time the issue of mistake, primarily because there was no sufficient evidential basis for considering that issue for the first time on a second appeal. Gibbon v Mitchell received a passing mention in the judgment of Norris J [2010] STC 982, para 20, but only for the purpose of rejecting the Revenue's argument that the distinction between 'effect' and 'consequences' was relevant to the Hastings-Bass rule. Had mistake been raised in Futter there would have been an issue of some importance as to whether the Court should assist in extricating claimants from a tax-avoidance scheme which had gone wrong. The scheme adopted by Mr Futter was by no means at the extreme of artificiality (compare for instance, that in Abacus Trust Co (Isle of Man) v NSPCC [2001] STC 1344) but it was hardly an exercise in good citizenship. In some cases of artificial tax avoidance the court might think it right to refuse relief, either on the ground that such claimants, acting on supposedly expert advice, must be taken to have accepted the risk that the scheme would prove ineffective, or on the ground that discretionary relief should be refused on grounds of public policy. Since the seminal decision of the House of Lords in WT Ramsay Ltd v IRC [1982] AC 300 there has been an increasingly strong and general recognition that artificial tax avoidance is a social evil which puts an unfair burden on the shoulders of those who do not adopt such measures. But it is unnecessary to consider that further on these appeals.

Equity does not act in vain

136. Mr Jones's second new point was that Mrs Pitt should be refused relief because the granting of relief would serve no practical purpose, other than saving inheritance tax. He cited Sir Nicolas Browne-Wilkinson V-C in the Spycatcher case, Attorney General v Guardian Newspapers Ltd [1987] 1 WLR 1248, 1270:

'It is an old maxim that equity does not act in vain. To my mind that is good law and the court should not make orders which would be ineffective to achieve what they set out to do.'

In the event the House of Lords took a different view, by a bare majority, as to whether the continuation of the interlocutory injunctions would serve any useful purpose. The maxim exists, but as Mason CJ and McHugh J said in Corin v Patton (1990) 169 CLR 540, 557,

'Like other maxims of equity, it is not a specific rule or principle of law. It is a summary statement of a broad theme which underlies equitable concepts and principles.'

137. The fund subject to the SNT had many calls on its resources, with heavy professional costs and expenses as well as making provision for the welfare and care of Mr Pitt and the maintenance of his wife. On his death on 25 September 2007 there was only £6,259 in the trust (the deputy judge added, para 15, that that was 'on Mrs Pitt's case' but he had earlier stated, para 4, that the material facts were not in dispute at all). On Mr Pitt's death this sum, subject to any outstanding liabilities, vested in his personal representatives under Clause 3 of the SNT. Any remaining value in the fund was therefore in the same beneficial ownership as if the SNT had been set aside by the court.

138. On 22 November 2011, after this court had granted permission for Mrs Pitt to appeal from the Court of Appeal's decision, her solicitors wrote to the Solicitor's Office of the Revenue drawing attention to a submission in the Revenue's skeleton argument before the Court of Appeal, para 105:

'But, in any event, the settlement should not be set aside after this period of time, especially when the Court does not know what proprietary claim would vest in the estate against third parties.'

Apparently with a view to avoiding any doubt on this point, Mrs Pitt's solicitors set out the factual position as it was at that time and stated in the last paragraph of their letter:

'Please note that Mrs Pitt and Mr Shores [her co-executor] have irrevocably instructed us to indicate, that if the Supreme Court orders that Mr Pitt's settlement is set aside, no further claim (to monies or other relief), will be made by them in their capacity as Mr Pitt's personal representatives, or by Mrs Pitt in her capacity as sole beneficiary of his estate, whether against the trustees (from time to time) of Mr Pitt's settlement or the recipients of distributions or other payments from the trustees. Our clients will be satisfied with the effect of section 150 IHTA 1984 (consequent on the order setting aside Mr Pitt's settlement).'

139. In these circumstances Mr Jones has submitted that it would be pointless, and so contrary to equity's practical approach, to grant relief that would achieve nothing, apart from a tax advantage to Mrs Pitt. He has relied on the approval by Peter Gibson LJ (with whom Sir Iain Glidewell and Kennedy LJ agreed) in Racal Group Services Ltd v Ashmore [1995] STC 1151, 1157 of what Vinelott J had said below [1994] STC 416, 425:

'In my judgment the principle established by these cases is that the court will make an order for the rectification of a document if satisfied that it does not give effect to the true agreement or arrangement between the parties, or to the true intention of a grantor or covenantor and if satisfied that there is an issue, capable of being contested, between the parties or between a covenantor or a grantor and the person he intended to benefit, it being irrelevant first that rectification of the document is sought or consented to by them all, and second that rectification is desired because it has beneficial fiscal consequences. On the other hand, the court will not order rectification of a document as between the parties or as between a grantor or covenantor and an intended beneficiary, if their rights will be unaffected and if the only effect of the order will be to secure a fiscal benefit.'

But Peter Gibson LJ went on to differ from Vinelott J in applying this principle. He held that there was an issue capable of being contested. The appeal was dismissed on another ground, that is because of the inadequacy of the evidence to satisfy the high standard of proof required for rectification.

140. What the Court of Appeal decided in Racal was that it is sufficient, even for the closely-guarded remedy of rectification, that there is a genuine issue capable of being contested, even if the parties decide that they will not in fact contest it. The test for rescission on the ground of mistake cannot be stricter than that.

141. Until the solicitor's letter of 22 November 2011 there was at least a possibility of third-party claims arising, and the Revenue placed reliance on that as a reason for refusing relief. But for the letter, the Court might, if minded to grant relief, have required an undertaking to the same effect as the one that Mrs Pitt and Mr Shores have volunteered. Moreover the Revenue's argument ignores the fact that unless and until the SNT is set aside, there are potentially contestable issues between the Revenue and any persons who, not being purchasers for value without notice, have received distributions from the SNT. The statutory charge under section 257 of the Inheritance Tax Act 1984 would prima facie give the Revenue a proprietary claim against such third parties. For these reasons I would reject the Revenue's second new point also.

The mistake claim in Pitt v Holt

142. In my opinion the test for setting aside a voluntary disposition on the ground of mistake is that set out in para 126 above, and it is satisfied in Pitt v Holt. There would have been nothing artificial or abusive about Mrs Pitt establishing the SNT so as to obtain protection under section 89 of the Inheritance Tax Act 1984. There was a considerable delay in the commencement of the proceedings, but the Revenue do not rely on the delay. They do rely on rescission being pointless and therefore inappropriate, but I would reject that submission for the reasons set out above. The deputy judge found ([2010] 1 WLR 1199, para 15) that the setting aside of the settlement would have no effect on any third party (plainly he was not here treating the Revenue as a third party). I would discharge the orders below and set aside the SNT on the ground of mistake.