McPhail and Ors v Doulton and Ors

[1971] A.C. 424

---

(Judgment by: Lord Wilberforce)

Between: McPhail and Ors - Appellants
And: Doulton and Ors - Respondents

Court:

Judges: Lord Reid
Lord Hodson
Lord Guest
Viscount Dilhorne

Lord Wilberforce

Subject References:
TRUSTS
DISCRETIONARY
Power of selection
Fund for employees
Trustees 'shall' apply income
Trustees not bound to exhaust income of any year
Power to apply accumulations as though income
Whether trust or power
Whether same or different test applicable in determining validity

Hearing date: 13-15, 19-21 January 1970
Judgment date: 6 May 1970

---

Judgment by:
Lord Wilberforce

My Lords this appeal is concerned with the validity of a trust deed dated July 17, 1941, by which Mr. Bertram Baden established a fund for the benefit, broadly, of the staff of the respondent company, Matthew Hall & Co. Ltd. Mr. Baden died in 1960 and the appellants are the executors of his will. They claim that the trust deed is invalid and that the assets transferred to the trustees by their testator revert to his estate. The trusts established by the deed are of a general type which has recently become common, the beneficiaries including a wide class of persons among whom the trustees are given discretionary powers or duties of distribution. It is the width of the class which in this and in other cases before the courts has given rise to difficulty and to the contention that the trusts are too indefinite to be upheld.

The trust deed begins with a recital that the settlor desired to establish a fund for providing benefits for the staff of the company and their relatives or dependants. The critical clauses are as follows:

"9.(a)

The trustees shall apply the net income of the fund in making at their absolute discretion grants to or for the benefit of any of the officers and employees or ex-officers or ex-employees of the company or to any relatives or dependants of any such persons in such amounts at such times and on such conditions (if any) as they think fit and any such grant may at their discretion be made by payment to the beneficiary or to any institution or person to be applied for his or her benefit and in the latter case the trustees shall be under no obligation to see to the application of the money.

(b)

The trustees shall not be bound to exhaust the income of any year or other period in making such grants as aforesaid and any income not so applied shall be dealt with as provided by clause 6 (a) hereof.

[Clause 6. (a) All moneys in the hands of the trustees and not required for the immediate service of the fund may be placed in a deposit or current account with any bank or banking house in the name of the trustees or may be invested as hereinafter provided.]

(c)

The trustees may realise any investments representing accumulations of income and apply the proceeds as though the same were income of the fund and may also (but only with the consent of all the trustees) at any time prior to the liquidation of the fund realise any other part of the capital of the fund which in the opinion of the trustees it is desirable to realise in order to provide benefits for which the current income of the fund is insufficient.

"10.

All benefits being at the absolute discretion of the trustees, no person shall have any right title or interest in the fund otherwise than pursuant to the exercise of such discretion, and nothing herein contained shall prejudice the right of the company to determine the employment of any officer or employee."

Clause 11 defines a perpetuity period within which the trusts are, in any event, to come to an end and clause 12 provides for the termination of the fund. On this event the trustees are directed to apply the fund in their discretion in one or more of certain specified ways of which one is in making grants as if they were grants under clause 9 (a). There are certain other provisions in the deed upon which arguments have been based, but these are of a subsidiary character and citation of them is unnecessary.

The present proceedings were started in 1963 by an originating summons taken out in the Chancery Division by the trustees of the deed seeking the decision of the court upon various questions, including that of the validity or otherwise of the trusts of the deed. It came before Goff J. in 1967. He first decided that the references in clauses 9 and 12 to employees of the company were not limited to the "staff" but comprised all the officers and employees of the company. There was no appeal against this.

On the main question of validity, the learned judge was, it seems, invited first to decide whether the provisions of clause 9 (a) constitute a trust or a power. This was on the basis that certain decided cases (which I shall examine) established a different test of invalidity for trusts on the one hand and powers on the other. He decided in favour of a power, and further that on this footing clause 9 (a) was valid. On appeal, the Court of Appeal by a majority upheld the decision in favour of a power, but held also that the learned judge had applied the wrong test for the validity of powers, the correct test being that stated (subsequent to the hearing before Goff J.) by this House in In re Gulbenkian's Settlement (Whishaw v. Stephens) [1970] A.C. 508 . The Court of Appeal therefore remitted the case to the Chancery Division to reconsider the validity of clause 9 (a) as a power.

In this House, the appellants contend, and this is the first question for consideration, that the provisions of clause 9 (a) constitute a trust and not a power. If that is held to be the correct result, both sides agree that the case must return to the Chancery Division for consideration, on this footing, whether this trust is valid. But here comes a complication. In the present state of authority, the decision as to validity would turn on the question whether a complete list (or on another view a list complete for practical purposes) can be drawn up of all possible beneficiaries. This follows from the Court of Appeal's decision in Inland Revenue Commissioners v. Broadway Cottages Trust [1955] CH. 20 as applied in later cases by which, unless this House decides otherwise, the Court of Chancery would be bound. The respondents invite your Lordships to review this decision and challenge its correctness. So the second issue which arises, if clause 9 (a) amounts to a trust, is whether the existing test for its validity is right in law and, if not, what the test ought to be.

Before dealing with these two questions some general observations, or reflections, may be permissible. It is striking how narrow and in a sense artificial is the distinction, in cases such as the present, between trusts or as the particular type of trust is called, trust powers, and powers. It is only necessary to read the learned judgments in the Court of Appeal to see that what to one mind may appear as a power of distribution coupled with a trust to dispose of the undistributed surplus, by accumulation or otherwise, may to another appear as a trust for distribution coupled with a power to withhold a portion and accumulate or otherwise dispose of it. A layman and, I suspect, also a logician would find it hard to understand what difference there is.

It does not seem satisfactory that the entire validity of a disposition should depend on such delicate shading. And if one considers how in practice reasonable and competent trustees would act, and ought to act, in the two cases, surely a matter very relevant to the question of validity, the distinction appears even less significant. To say that there is no obligation to exercise a mere power and that no court will intervene to compel it, whereas a trust is mandatory and its execution may be compelled, may be legally correct enough but the proposition does not contain an exhaustive comparison of the duties of persons who are trustees in the two cases.

A trustee of an employees' benefit fund, whether given a power or a trust power, is still a trustee and he would surely consider in either case that he has a fiduciary duty: he is most likely to have been selected as a suitable person to administer it from his knowledge and experience, and would consider he has a responsibility to do so according to its purpose. It would be a complete misdescription of his position to say that, if what he has is a power unaccompanied by an imperative trust to distribute, he cannot be controlled by the court unless he exercised it capriciously, or outside the field permitted by the trust (cf. Farwell on Powers, 3rd ed., p. 524). Any trustee would surely make it his duty to know what is the permissible area of selection and then consider responsibly, in individual cases, whether a contemplated beneficiary was within the power and whether, in relation to other possible claimants, a particular grant was appropriate.

Correspondingly a trustee with a duty to distribute, particularly among a potentially very large class, would surely never require the preparation of a complete list of names, which anyhow would tell him little that he needs to know. He would examine the field, by class and category; might indeed make diligent and careful inquiries, depending on how much money he had to give away and the means at his disposal, as to the composition and needs of particular categories and of individuals within them; decide upon certain priorities or proportions, and then select individuals according to their needs or qualifications. If he acts in this manner, can it really be said that he is not carrying out the trust?

Differences there certainly are between trust (trust powers) and powers, but as regards validity, should they be so great as that in one case complete, or practically complete, ascertainment is needed, but not in the other? Such distinction as there is would seem to lie in the extent of the survey which the trustee is required to carry out: if he has to distribute the whole of a fund's income, he must necessarily make a wider and more systematic survey than if his duty is expressed in terms of a power to make grants. But just as, in the case of a power, it is possible to underestimate the fiduciary obligation of the trustee to whom it is given, so, in the case of a trust (trust power), the danger lies in overstating what the trustee requires to know or to inquire into before he can properly execute his trust. The difference may be one of degree rather than of principle: in the well-known words of Sir George Farwell, Farwell on Powers, 3rd ed. (1916), p. 10, trusts and powers are often blended, and the mixture may vary in its ingredients.

With this background I now consider whether the provisions of clause 9 (a) constitute a trust or a power. I do so briefly because this is not a matter on which I or, I understand, any of your Lordships have any doubt. Indeed, a reading of the judgments of Goff J. and of the majority in the Court of Appeal leave the strong impression that, if it had not been for their leaning in favour of possible validity and the state of the authorities, these learned judges would have found in favour of a trust. Naturally read, the intention of the deed seems to me clear: clause 9 (a), whose language is mandatory ("shall"), creates, together with a power of selection, a trust for distribution of the income, the strictness of which is qualified by clause 9 (b), which allows the income of any one year to be held up and (under clause 6 (a)) either placed, for the time, with a bank, or, if thought fit, invested.

Whether there is, in any technical sense, an accumulation seems to me in the present context a jejune inquiry: what is relevant is that clause 9 (c) marks the difference between "accumulations" of income and the capital of the fund: the former can be distributed by a majority of the trustees, the latter cannot. As to clause 10, I do not find in it any decisive indication. If anything, it seems to point in favour of a trust, but both this and other points of detail are insignificant in the face of the clearly expressed scheme of clause 9. I therefore agree with Russell L.J. and would to that extent allow the appeal, declare that the provisions of clause 9 (a) constitute a trust and remit the case to the Chancery Division for determination whether on this basis clause 9 is (subject to the effects of section 164 of the Law of Property Act, 1925) valid or void for uncertainty.

This makes it necessary to consider whether, in so doing, the court should proceed on the basis that the relevant test is that laid down in Inland Revenue Commissioners v. Broadway Cottages Trust [1955] CH. 20 or some other test.

That decision gave the authority of the Court of Appeal to the distinction between cases where trustees are given a power of selection and those where they are bound by a trust for selection. In the former case the position, as decided by this House, is that the power is valid if it can be said with certainty whether any given individual is or is not a member of the class and does not fail simply because it is impossible to ascertain every member of the class (In re Gulbenkian's Settlements [1970] A.C. 508 ). But in the latter case it is said to be necessary, for the trust to be valid, that the whole range of objects (I use the language of the Court of Appeal) should be ascertained or capable of ascertainment.

The respondents invited your Lordships to assimilate the validity test for trusts to that which applies to powers. Alternatively they contended that in any event the test laid down in the Broadway Cottages case [1955] CH. 20 was too rigid, and that a trust should be upheld if there is sufficient practical certainty in its definition for it to be carried out, if necessary with the administrative assistance of the court, according to the expressed intention of the settlor. I would agree with this, but this does not dispense from examination of the wider argument. The basis for the Broadway Cottages principle is stated to be that a trust cannot be valid unless, if need be, it can be executed by the court, and (though it is not quite clear from the judgment where argument ends and decision begins) that the court can only execute it by ordering an equal distribution in which every beneficiary shares. So it is necessary to examine the authority and reason for this supposed rule as to the execution of trusts by the court.

Assuming, as I am prepared to do for present purposes, that the test of validity is whether the trust can be executed by the court, it does not follow that execution is impossible unless there can be equal division.

As a matter of reason, to hold that a principle of equal division applies to trusts such as the present is certainly paradoxical. Equal division is surely the last thing the settlor ever intended: equal division among all may, probably would, produce a result beneficial to none. Why suppose that the court would lend itself to a whimsical execution? And as regards authority, I do not find that the nature of the trust, and of the court's powers over trusts, calls for any such rigid rule. Equal division may be sensible and has been decreed, in cases of family trusts, for a limited class, here there is life in the maxim "equality is equity," but the cases provide numerous examples where this has not been so, and a different type of execution has been ordered, appropriate to the circumstances.

Mosely v. Moseley, Fin. 53 is an early example, from the time of equity's architect, where the court assumed power (if the executors did not act) to nominate from the sons of a named person as it should think fit and most worthy and hopeful, the testator's intention being that the estate should not be divided. In Clarke v. Turner, Free.Ch. 198, on a discretionary trust for relations, the court decreed conveyance to the heir-at-law judging it "most reputable for the family that the heir-at-law should have it." In Warburton v. Warburton, 4 Bro.P.C 1, on a discretionary trust to distribute between a number of the testator's children, the House of Lords affirmed a decree of Lord Keeper Wright that the eldest son and heir, regarded as necessitous, should have a double share, the court exercising its own discretionary judgment against equal division.

These are examples of family trusts but in Richardson v. Chapman, 7 Bro.P.C. 318 the same principle is shown working in a different field. There was a discretionary trust of the testator's "options" (namely, rights of presentation to benefices or dignities in the Church) between a number of named or specified persons, including present and former chaplains and other domestics, also "my worthy friends and acquaintance, particularly the Reverend Dr. Richardson of Cambridge."

The House of Lords (reversing Lord Keeper Henley) set aside a "corrupt" presentation and ordered the trustees to present Dr. Richardson as the most suitable person. The grounds of decision in this House, in accordance with the prevailing practice, were not reported, but it may be supposed that the reported argument was accepted that where the court sets aside the act of the trustee, it can at the same time decree the proper act to be done, not by referring the matter to the trustee's discretion, but by directing him to perform as a mere instrument the thing decreed (ibid., 326, 327). This shows that the court can in a suitable case execute a discretionary trust according to the perceived intention of the truster. It is interesting also to see that it does not seem to have been contended that the trust was void because of the uncertainty of the words "my worthy friends and acquaintance." There was no doubt that Dr. Richardson came within the designation.

In the time of Lord Eldon, the Court of Chancery adopted a less flexible practice: in Kemp v. Kemp, 5 Ves.Jr. 849 Sir Richard Arden M.R., commenting on Warburton v. Warburton, 4 Bro.P.C. 1 ("a very extraordinary" case), said that the court now disclaims the right to execute a power (i.e., a trust power) and gives the fund equally. But I do not think that this change of attitude, or practice, affects the principle that a discretionary trust can, in a suitable case, be executed according to its merits and otherwise than by equal division. I prefer not to suppose that the great masters of equity, if faced with the modern trust for employees, would have failed to adapt their creation to its practical and commercial character.

Lord Eldon himself, in Morice v. Bishop of Durham, 10 Ves.Jr. 522, laid down clearly enough that a trust fails if the object is insufficiently described or if it cannot be carried out, but these principles may be fully applied to trust powers without requiring a complete ascertainment of all possible objects. His earlier judgment in the leading, and much litigated, case of Brown v. Higgs, 8 Ves.Jr. 561 shows that he was far from fastening any rigid test of validity upon trust powers. After stating the distinction, which has ever since been followed, between powers, which the court will not require the donee to execute, and powers in the nature of a trust, or trust powers, he says of the latter that if the trustee does not discharge it, the court will, to a certain extent, discharge the duty in his room and place. To support this, he cites Harding v. Glyn (1739) 1 Atk. 469, an early case where the court executed a discretionary trust for "relations" by distributing to the next-of-kin.

I dwell for a moment upon this point because, not only was Harding v. Glyn described by Lord Eldon (8 Ves.Jr. 561, 570) as having been treated as a clear authority in his experience for a long period, but the principle of it was adopted in several nineteenth-century authorities. When the Broadway Cottages Trust case came to be decided in 1955, these cases were put aside as anomalous (see [1955] Ch. at pp. 33, 35), but I think they illustrate the flexible manner in which the court, if called on, executes trust powers for a class. At least they seem to prove that the supposed rule as to equal division does not rest on any principle inherent in the nature of a trust. They prompt me to ask why a practice, or rule, which has been long followed and found useful in "relations" cases should not also serve in regard to "employees," or "employees and their relatives," and whether a decision which says the contrary is acceptable.

I now consider the modern English authorities, particularly those relied on to show that complete ascertainment of the class must be possible before it can be said that a discretionary trust is valid.

In re H. J. Ogden [1933] CH. 678 is not a case which I find of great assistance. The argument seems to have turned mainly on the question whether the trust was a purpose trust or a trust for ascertained objects. The latter was held to be the case and the court then held that all the objects of the discretionary gift could be ascertained. It is weak authority for the requirement of complete ascertainment.

The modern shape of the rule derives from In re Gestetner Settlement [1953] CH. 672 , where the judgment of Harman J., to his later regret established the distinction between discretionary powers and discretionary trusts. The focus of this case was upon powers. The judgment first establishes a distinction between, on the one hand, a power collateral, or appurtenant, or other powers "which do not impose a trust on the conscience of the donee" (at p. 684), and on the other hand a trust imposing a duty to distribute. As to the first, the learned judge said (ibid.):

"I do not think it can be the law that it is necessary to know of all the objects in order to appoint to one of them."

As to the latter he uses these words (at p. 685):

"It seems to me there is much to be said for the view that he must be able to review the whole field in order to exercise his judgment properly."

He then considers authority on the validity of powers, the main stumbling-block in the way of his own view being some words used by Fry J. in Blight v. Hartnoll (1881) 19 Ch.D. 294, 301, which had been adversely commented on in Farwell on Powers (3rd ed., at pp. 168, 169), and I think it worth while quoting the words of his conclusion. He says ( [1953] CH. 672 , 688, 689):

"The settlor had good reason, I have no doubt, to trust the persons whom he appointed trustees; but I cannot see here that there is such a duty as makes it essential for these trustees, before parting with any income or capital, to survey the whole field, and to consider whether A is more deserving of bounty than B. That is a task which was and which must have been known to the settlor to be impossible, having regard to the ramifications of the persons who might become members of this class.

If, therefore, there be no duty to distribute, but only a duty to consider, it does not seem to me that there is any authority binding on me to say that this whole trust is bad. In fact, there is no difficulty, as has been admitted, in ascertaining whether any given postulant is a member of the specified class. Of course, if that could not be ascertained the matter would be quite different, but of John Doe or Richard Roe it can be postulated easily enough whether he is or is not eligible to receive the settlor's bounty. There being no uncertainty in that sense, I am reluctant to introduce a notion of uncertainty in the other sense, by saying that the trustees must worry their heads to survey the world from China to Peru, when there are perfectly good objects of the class in England."

Subject to one point which was cleared up in this House in In re Gulbenkian's Settlements [1970] A.C. 508 , all of this, if I may say so, seems impeccably good sense, and I do not understand the learned judge to have later repented of it. If the judgment was in any way the cause of future difficulties, it was in the indication given - not by way of decision, for the point did not arise - that there was a distinction between the kind of certainty required for powers and that required for trusts. There is a difference perhaps but the difference is a narrow one, and if one is looking to reality one could hardly find better words than those I have just quoted to describe what trustees, in either case, ought to know. A second look at this case, while fully justifying the decision, suggests to me that it does not discourage the application of a similar test for the validity of trusts.

So I come to Inland Revenue Commissioners v. Broadway Cottages Trust [1955] CH. 20 . This was certainly a case of trust, and it proceeded on the basis of an admission, in the words of the judgment, "that the class of 'beneficiaries' is incapable of ascertainment." In addition to the discretionary trust of income, there was a trust of capital for all the beneficiaries living or existing at the terminal date. This necessarily involved equal division and it seems to have been accepted that it was void for uncertainty since there cannot be equal division among a class unless all the members of the class are known. The Court of Appeal applied this proposition to the discretionary trust of income, on the basis that execution by the court was only possible on the same basis of equal division. They rejected the argument that the trust could be executed by changing the trusteeship, and found the relations cases of no assistance as being in a class by themselves. The court could not create an arbitrarily restricted trust to take effect in default of distribution by the trustees. Finally they rejected the submission that the trust could take effect as a power: a valid power could not be spelt out of an invalid trust.

My Lords, it will have become apparent that there is much in this which I find out of line with principle and authority but before I come to a conclusion on it, I must examine the decision of this House in In re Gulbenkian's Settlements [1970] A.C. 508 on which the appellants placed much reliance as amounting to an endorsement of the Broadway Cottages case [1955] CH. 20 . But is this really so? That case was concerned with a power of appointment coupled with a gift over in default of appointment. The possible objects of the power were numerous and were defined in such wide terms that it could certainly be said that the class was unascertainable. The decision of this House was that the power was valid if it could be said with certainty whether any given individual was or was not a member of the class, and did not fail simply because it was impossible to ascertain every member of the class. In so deciding, their Lordships rejected an alternative submission, to which countenance had been given in the Court of Appeal, that it was enough that one person should certainly be within the class. So, as a matter of decision, the question now before us did not arise or nearly arise. However, the opinions given were relied on, and strongly, as amounting to an endorsement of the "complete ascertainment" test as laid down in the Broadway Cottages case.

My Lords, I comment on this submission with diffidence, because three of those who were party to the decision are present here today, and will express their own views. But with their assistance, and with respect for their views, I must endeavour to appraise the appellants' argument. My noble and learned friend Lord Reid's opinion can hardly be read as an endorsement of the Broadway Cottages case. It is really the opinion of my noble and learned friend Lord Upjohn which has to be considered. Undoubtedly the main part of that opinion, as one would expect, was concerned to deal with the clause in question, which required careful construction, and with the law as to powers of appointment among a numerous and widely defined class.

But having dealt with these matters the opinion continues with some general observations. I have considered these with great care and interest: I have also had the advantage of considering a detailed report of the argument of counsel on both sides who were eminent in this field. I do not find that it was contended on either side that the Broadway Cottages Trust case was open to criticism - neither had any need to do so. The only direct reliance upon it appears to have been to the extent of the fifth proposition appearing on p. 31 of the report, which was relevant as referring to powers, but does not touch this case. It is consequently not surprising that my noble and learned friend Lord Upjohn nowhere expresses his approval of this decision and indeed only cites it, in the earlier portion, in so far as it supports a proposition as to powers.

Whatever dicta therefore the opinion was found to contain, I could not, in a case where a direct and fully argued attack has been made on the Broadway Cottages case, regard them as an endorsement of it and I am sure that my noble and learned friend, had he been present here, would have regarded the case as at any rate open to review. In fact I doubt very much whether anything his Lordship said was really directed to the present problem. I read his remarks as dealing with the suggestion that trust powers ought to be entirely assimilated to conditions precedent and powers collateral. The key passage is where he says [1970] A.C. 508 , 525:

"Again the basic difference between a mere power and a trust power is that in the first case trustees owe no duty to exercise it and the relevant fund or income falls to be dealt with in accordance with the trusts in default of its exercise, whereas in the second case the trustees must exercise the power and in default the court will. It is briefly summarised in Halsbury's Laws of England, 3rd ed., Vol. 30 (1959), p. 241, para. 445:

'... the court will not exercise or compel trustees to exercise a purely discretionary power given to them; but the court will restrain the trustees from exercising the power improperly, and, if it is coupled with a duty, the court can compel the trustees to perform their duty.'

It is a matter of construction whether the power is a mere power or a trust power and the use of inappropriate language is not decisive (Wilson v. Turner (1883) 22 Ch.D. 521, 525).

So, with all respect to the contrary view, I cannot myself see how, consistently with principle, it is possible to apply to the execution of a trust power the principles applicable to the permissible exercise by the donees (even if trustees) of mere powers; that would defeat the intention of donors completely.

But with respect to mere powers, while the court cannot compel the trustees to exercise their powers, yet those entitled to the fund in default must clearly be entitled to restrain the trustees from exercising it save among those within the power. So the trustees or the court must be able to say with certainty who is within and who is without the power. It is for this reason that I find myself unable to accept the broader proposition advanced by Lord Denning M.R. and Winn L.J., mentioned earlier, and agree with the proposition as enunciated in In re Gestetner Settlement [1953] CH. 672 and the later cases."

The reference to "defeating the intention of donors completely" shows that what he is concerned with is to point to the contrast between powers and trusts which lies in the facultative nature of the one and the mandatory nature of the other, the conclusion being the rejection of the "broader" proposition as to powers accepted by two members of the Court of Appeal. With this in mind it becomes clear that the sentence so much relied on by the appellants will not sustain the weight they put on it. This is:

"The trustees have a duty to select the donees of the donor's bounty from among the class designated by the donor; he has not entrusted them with any power to select the donees merely from among known claimants who are within the class, for that is constituting a narrower class and the donor has given them no power to do this" ( [1970] A.C. 508 , 524).

What this does say, and I respectfully agree, is that, in the case of a trust, the trustees must select from the class. What it does not say, as I read it, or imply, is that in order to carry out their duty of selection they must have before them, or be able to get, a complete list of all possible objects.

So I think that we are free to review the Broadway Cottages case [1955] CH. 20 . The conclusion which I would reach, implicit in the previous discussion, is that the wide distinction between the validity test for powers and that for trust powers is unfortunate and wrong, that the rule recently fastened upon the courts by Inland Revenue Commissioners v. Broadway Cottages Trust ought to be discarded, and that the test for the validity of trust powers ought to be similar to that accepted by this House in In re Gulbenkian's Settlements [1970] A.C. 508 for powers, namely, that the trust is valid if it can be said with certainty that any given individual is or is not a member of the class.

I am interested, and encouraged, to find that the conclusion I had reached by the end of the argument is supported by distinguished American authority. Professor Scott in his well-known book on trusts (Scott on Trusts (1939)) discusses the suggested distinction as regards validity between trusts and powers and expresses the opinion that this would be "highly technical" (s. 122, p. 613). Later in the second Restatement of Trusts (1959), s. 122 (which Restatement aims at stating the better modern view and which annotates the Broadway Cottages case), a common test of invalidity is taken, whether trustees are "authorised" or "directed": this is that the class must not be so indefinite that it cannot be ascertained whether any person falls within it. The reporter is Professor Austin Scott. In his abridgment, published in 1960 (Scott's Abridgment of The Law of Trusts, s. 122, p. 239), Professor Scott maintains the same position:

"It would seem that if a power of appointment among the members of an indefinite class is valid, the mere fact that the testator intended not merely to confer a power but to impose a duty to make such an appointment should not preclude the making of such an appointment. It would seem to be the height of technicality that if a testator authorises a legatee to divide the property among such of the testator's friends as he might select, he can properly do so, but that if he directs him to make such a selection, he will not be permitted to do so."

Assimilation of the validity test does not involve the complete assimilation of trust powers with powers. As to powers, I agree with my noble and learned friend Lord Upjohn in In re Gulbenkian's Settlements that although the trustees may, and normally will, be under a fiduciary duty to consider whether or in what way they should exercise their power, the court will not normally compel its exercise. It will intervene if the trustees exceed their powers, and possibly if they are proved to have exercised it capriciously.

But in the case of a trust power, if the trustees do not exercise it, the court will: I respectfully adopt as to this the statement in Lord Upjohn's opinion (p. 525). 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 for distribution appear by itself directing the trustees so to distribute. The books give many instances where this has been done, and I see no reason in principle why they should not do so in the modern field of discretionary trusts (see Brunsden v. Woolredge (1765) 1 Amb. 507, Supple v. Lowson (1773) 2 Amb. 729, Liley v. Hey (1842) 1 Hare 580 and Lewin on Trusts, 16th ed. (1964), p. 630). Then, as to the trustees' duty of inquiry or ascertainment, in each case the trustees ought to make such a survey of the range of objects or possible beneficiaries as will enable them to carry out their fiduciary duty (cf. Liley v. Hey). A wider and more comprehensive range of inquiry is called for in the case of trust powers than in the case of powers.

Two final points: first, as to the question of certainty. I desire to emphasise the distinction clearly made and explained by Lord Upjohn ( [1970] A.C. 508 , 524) between linguistic or semantic uncertainty which, if unresolved by the court, renders the gift void, and the difficulty of ascertaining the existence or whereabouts of members of the class, a matter with which the court can appropriately deal on an application for directions. There may be a third case where the meaning of the words used is clear but the definition of beneficiaries is so hopelessly wide as not to form "anything like a class" so that the trust is administratively unworkable or in Lord Eldon's words one that cannot be executed (Morice v. Bishop of Durham, 10 Ves.Jr. 522, 527). I hesitate to give examples for they may prejudice future cases, but perhaps "all the residents of Greater London" will serve. I do not think that a discretionary trust for "relatives" even of a living person falls within this category.

I would allow the appeal and make the order suggested earlier in this opinion. The costs of the appellants and of the respondents of this appeal taxed on a common fund basis should be paid out of so much of the trust fund subject to the trust deed of July 17, 1941, as was derived from Bertram Baden deceased.

Appeal allowed.

Solicitors: Slaughter & May; Gregory, Rowcliffe & Co.