Re Gulbenkian's Settlement Trusts (No 1) [1968] UKHL 5

House of Lords
Between
Re Gulbenkian's Settlement Trusts (No 1)Appellant
Before
Lord ReidLord HodsonLord GuestLord UpjohnLord Donovan

LORD DONOVAN

My Lords,Settlements were made by the late Mr. Calouste Gulbenkian in 1929 and 1938 under which the trustees " shall " during the life of his son Mr. Nubar Gulbenkian
" at their absolute discretion pay all or any part of the income of the " property hereby settled and the investments for the time being " representing the same (hereinafter called the Trust Fund) to or apply " the same for the maintenance and personal support or benefit of all " or any one or more to the exclusion of the other or others of the " following persons ..."

Among those persons were

" any person or persons in whose house or apartments or in whose " company or under whose care or control or by or with whom the " said Nubar Sarkis Gulbenkian may from time to time be employed " or residing ..."
The sole question in this appeal is whether this class of potential beneficiaries is so uncertain that these provisions cannot be operated by the trustees. It is not disputed that if the description of the class which I have quoted is too uncertain then the whole provision fails even although the other potential beneficiaries are easily ascertainable.This clause does not make sense as it stands. One of the permutations which the grammar requires is any person " by whom Mr. Gulbenkian is " residing ". But the client must not be penalised for his lawyer's slovenly drafting. Under modern conditions it may be necessary to relax older and stricter standards. If I adopt methods of construction appropriate for commercial documents and documents inter rusticos I must consider whether underlying the words used any reasonably clear intention can be discerned. I think that it is reasonably clear that this clause is the result of carelessly telescoping two separate clauses—(1) any person by whom Mr. Gulbenkian may from time to time be employed, and (2) any person in whose house or in whose company or under whose care or with whom he may from time to time be residing. Read literally the clause embraces any person in whose house or in whose company or under whose care or with whom Mr. Gulbenkian may from time to time be employed. That might well be held to be too uncertain but I think it reasonably clear that that cannot have been intended: no rational person would insert provisions lake that. I was surprised to learn that this botched clause has somehow found its way into a standard book of precedents, so I realise that this matter may be of some general importance.If the clause is read in the way in which I think it must be read then it is not suggested that there is any uncertainty about the first limb—any person by whom Mr. Gulbenkian may be employed—but it is argued that the other limb is bad for uncertainty. It would be bad if it purported to impose on the trustees duties which, even with the aid of the Court, they could not properly carry out. So one must first see what their duties are. One argument, as I understand it, is that because this is admitted to be a mere power, it really imposes no duties on them at all. I find that difficult to understand. It is a power given not to the individuals who happen also to be trustees but to the trustees as such so that new trustees duly assumed or appointed can exercise it. In my view it must follow that the trustees are to act in their fiduciary capacity. They are given an absolute discretion. So if they decide in good faith at appropriate times to give none of the income to any of the beneficiaries the Court cannot pronounce their reasons2to be bad. And similarly if they decide to give some or all of the income to a particular beneficiary the Court will not review their decision. That was decided by this House in Gisborne v. Gisborne, 2 App. Cas. 300, But their " absolute discretion " must I think be subject to two conditions. It may be true that when a mere power is given to an individual he is under no duty to exercise it or even to consider whether he should exercise it. But when a power is given to trustees as such, it appears to me that the situation must be different. A settlor or testator who entrusts a power to his trustees must be relying on them in their fiduciary capacity so they cannot simply push aside the power and refuse to consider whether it ought in their judgment to be exercised. And they cannot give money to a person who is not within the classes of persons designated by the settlor: the construction of the power is for the Court.If the classes of beneficiaries are not defined with sufficient particularity to enable the Court to determine whether a particular person is or is not, on the facts at a particular time, within one of the classes of beneficiaries then the power must be bad for uncertainty. If the donee of the power whether or not he has any duty) desires to exercise it in favour of a particular person it must be possible to determine whether that particular person is or is not within the class of objects of the power. And it must be possible to determine the validity of the power immediately it comes into operation. It cannot be valid if the person whom the donee happens to choose is clearly within the objects but void if it is doubtful whether that is so. So if one can reasonably envisage cases where the Court could not determine the question the power must be bad for uncertainty. But it is not bad merely because such determination may be difficult in a particular case. The Respondents have inserted in their Case at the request of the trustees a statement that in the view of the trustees " it must " be unlikely that they would in practice be able to exercise the said power " or discretion except after obtaining a decision of the Court whether any " particular suggested object thereof did or did not fall within the said description ". That in itself is not sufficient to warrant a decision that the power fails for uncertainty. It may be that there is a class of case where, although the description of a class of beneficiaries is clear enough, any attempt to apply it to the facts would lead to such administrative difficulties that it would for that reason be held to be invalid. But that is not this case.The class of persons to be considered in this case are those(a) in whose house or apartments Mr. Gulbenkian is residing,(b) in whose company or with whom he is residing and(c) under whose care or control he is residing. It is often difficult in a particular case to determine whether a temporary sojourn amounts to " residence ", but that is the kind of problem which Courts often have to solve. And it is not much more difficult to say whether a man is residing in another person's company or under another person's care or control than it is to say whether he is residing in a particular house. I therefore reject the Appellants' first argument.Then the Appellants submitted a further argument, that a power is bad for uncertainty unless it is possible to make a complete list of the possible beneficiaries at the time when it falls to be exercised. It is said that trustees cannot properly exercise their discretion unless they can survey the whole field: otherwise there might be in existence potential beneficiaries whom they might regard as more deserving than those who are known to them. In my view that cannot be right. Suppose that a testator or settlor empowers his trustees to give money to such of the descendants of X (or former domestic servants of X) as they may in their absolute discretion select: X may be the testator or settlor himself or anyone else. On the face of it that power is perfectly valid, and it surely cannot become invalid because before the power is exercised some descendant of X has emigrated with his family and cannot be traced.The Appellants found on Inland Revenue v. Broadway Cottages Trust [1955] Ch 20 . There the trustees had a duty to apply certain income for the benefit of all or any one or more of certain defined classes of bene- ficiaries in such shares proportions and manner as the trustees in their3discretion might from time to time think fit. Admittedly those classes were sufficiently precisely defined to make it possible to determine whether any particular individual was or was not eligible. But the classes were so wide that it was admitted to be impossible to make a complete list of those comprised in them. This was held to involve uncertainty so as to make the provisions void. That seems to me to be a very odd kind of uncertainty. Does it involve the proposition that if at first all potential beneficiaries are identifiable then the provision is valid, but that if for some reason, such as some of them disappearing, it later becomes impossible to make a list of all the survivors then the provision becomes invalid? I could understand it being held that if the classes of potential beneficiaries were so numerous that it would cost quite disproportionate enquiries and expense to find them all and discover their needs or deserts, then the provision would fail. But that would not be on the ground of uncertainty as that term is generally understood. I gravely doubt some of the arguments used by the Court of Appeal.I will not deal with the authorities in detail because I think that the present position of the law is far from satisfactory. I agree with criticism by Lord Evershed in Re Hain [1961] 1 W.L.R. 440. by Sachs L.J. in Darwen v. Leek [1968] 2 W.L.R. 1385 and by Lord Denning M.R. in the present case. This case can be decided on the grounds suggested by my noble and learned friend, Lord Upjohn, and it is not an appropriate case for a review of the law. But I trust that there may be an early opportunity for reconsideration of some of the narrow and technical distinctions which have grown up in this chapter of the law.I would dismiss this appeal.

Lord Hodson

My Lords,I have had the advantage of reading the opinion of my noble and learned friend, Lord Upjohn. I agree with it and would dismiss the appeal.

Lord Guest

my lords.I have had the advantage of reading the opinion of my noble and learned friend, Lord Upjohn. I agree with it and would dismiss the appeal.

Lord Upjohn

My Lords,On the 31st May, 1929, the late Mr. Calouste Sarkis Gulbenkian made a Settlement expressed to be in consideration of his natural love and affection for his son Nubar Sarkis Gulbenkian, one of the Respondents, who, how- ever, has taken no part in the argument, for he has assigned away his rights, if any, to income during his life ; I shall call him " the son ".

(i) in these terms

" 2. (i) The Trustees shall during the life of the said Nubar Sarkis " Gulbenkian at their absolute discretion pay all or any part of the " income of the property hereby settled and the investments for the " time being representing the same (hereinafter called the Trust Fund) " to or apply the same for the maintenance and personal support or " benefit of all or any one or more to the exclusion of the other or " others of the following persons namely the said Nubar Sarkis " Gulbenkian and any wife and his children or remoter issue for the " time being in existence whether minors or adults and any person " or persons in whose house or apartments or in whose company or " under whose care or control or by or with whom the said Nubar Sarkis " Gulbenkian may from time to time be employed or residing and the " other persons or person other than the Settlor for the time being " entitled or interested whether absolutely contingently or otherwise4" to or in the trust fund under the trusts herein contained to take effect " after the death of the said Nubar Sarkis Gulbenkian in such propor- " tions and manner as the Trustees shall in their absolute discretion " at any time or times think proper.
" (ii) Subject to the discretionary trust or power hereinbefore contained " the Trustees shall during the life of the said Nubar Sarkis Gulbenkian " hold the said income or so much thereof as shall not be paid or " applied under such discretionary trust or power upon the trusts and " for the purposes and for which the said income would for the time " being be held if the said Nubar Sarkis Gulbenkian were then dead."
After the death of the son the capital and future income of the Trust Fund was to be held upon terms with which your Lordships are not concerned.On the 18th July, 1938, the father made another Settlement by way of further provision for the son. During the lifetime of the son the income was to be held upon trusts similar to those set out above save that the words " or in whose company or under whose care or control" are omitted, but Counsel on neither side has sought to differentiate between the two Settle- ments on this ground and I shall therefore refer only to Clause 2 (i) of the 1929 Settlement.The whole question before your Lordships is whether that clause is void for uncertainty or whether the clause is valid so that the son is an object of the discretion. If so, certain subsidiary questions are raised by the Originating Summonses which have been issued, but these questions are not before your Lordships.My Lords,upon the main arguments presented to your Lordships the decision upon this matter lies in a very small compass though, in view of the secondary argument of the Respondents with which the majority of the Court of Appeal agreed, I shall have later to deal with the issues rather more generally.It is agreed between the parties that the discretion to the Trustees in Clause 2(i) to pay all or any part of the income of the Trust Fund at their absolute discretion to one or more of the persons therein mentioned to the exclusion of the other or others or to apply it for their maintenance support or benefit is a mere or bare power or a power collateral, as it is sometimes called. It is not a trust power: the Trustees have no duty to exercise it in the sense that the Court has any power to compel the Trustees to exercise it or to exercise it itself if the Trustees refuse or neglect to do so. In so far as the power is not exercised by the Trustees or if it is void for uncertainty, the income falls to be held upon the Trusts declared by Clause 2(ii) It is curious that there is no long line of decided cases as to what is the proper test to apply when considering the validity of a mere power when the class of possible appointees is or may be incapable of ascertainment, but there is a body of recent authority to the effect that the rule is that provided there is a valid gift over or trust in default of appointment (which was fundamental to the decision of Clauson J. in Re Park [1932] 1 Ch. 580) a mere or bare power of appointment among a class is valid if you can with certainty say whether any given individual is or is not a member of the class; you do not have to be able to ascertain every member of the class.This was stated by Harman J. in Gestetner [1953] Ch. 672, followed by Roxburgh J. in Re Coates [1955] Ch. 495 and by me in Re Sayer [1957] Ch. 423. Its reasoning was, I think, approved in the Court of Appeal in Inland Revenue Commissioners v. Broadway Cottages Trust [1955] Ch 20 ; I say that because it is a little difficult to know whether on pages 32 and 33 Jenkins L.J. was doing more than setting out the Crown's argument. I note that Danckwerts L.J. (probably rightly) treated it in the Court below as part of the judgment of the learned Lord Justice.And the rule was in general terms approved by Evershed M.R. in Re Hain [1961] 1 W.L.R. 440 at 445. Counsel for the Appellants submits the fore- going authorities correctly state the law and on his first line of argument Counsel for the Respondents agrees.5Let me assume, then, for the present that is the right test. Does Clause 2(i) satisfy that test or is it too uncertain? A very similar clause came before Harman J. in Re Gresham [1956] 1 W.L.R. 573 and he held it void for uncertainty even after construing it, as he described it, " benevolently ".Counsel for the Appellants argued that you must give the words used their literal meaning and then apply the test to see whether you can predicate with certainty whether a given individual is or is not within the class and no modification of the literal language is permissible to make sense of it. This argument is based on a fallacy.There is no doubt that the first task is to try to ascertain the Settlor's intention, so to speak, without regard to the consequences, and then, having construed the document, apply the test. The Court, whose task it is to discover that intention, starts by applying the usual canons of construction; words must be given their usual meaning, the clause should be read literally and in accordance with the ordinary rules of grammar. But very frequently, whether it be in wills, settlements or commercial agreements, the application of such fundamental canons leads nowhere, the draftsman has used words wrongly, his sentences border on the illiterate and his grammar may be appalling. It is then the duty of the Court by the exercise of its judicial knowledge and experience in the relevant matter, innate commonsense and desire to make sense of the settlor's or parties' expressed intentions, however obscure and ambiguous the language that may have been used, to give a reasonable meaning to that language if it can do so without doing complete violence to it. The fact that the Court has to see whether the clause is " certain " for a particular purpose does not disentitle the Court from doing otherwise than, in the first place, try to make sense of it.My Lords,I do not think the late Mr. Withers' language (if indeed he was responsible for the draftsmanship) is open to very serious criticism ; the clause it is true ran together too many possible situations and did so rather ungram- matically, but its general object was clear ; it was a " spendthrift " clause and must be read in the light of that general intention and construed with the object to giving effect to it if it is possible to do so.I adopt the construction propounded by Harman J. in Gresham with amend- ments. I do not regard his construction as benevolent in the least degree but as the proper construction of the Settlor's imperfectly expressed language; indeed, I do not regard my own interpretation of his language as perfect; it is merely sufficient for the decision of this case.

So I would read the relevant words in this way : After the words—

" the said Nubar Sarkis Gulbenkian and any wife and his children or " remoter issue for the time being in existence whether minors or " adults "
I would paraphrase the words of the Clause thus :— " and any person or persons by whom the son may from time to time " be employed and any person or persons with whom the son from time " to time is residing whether in the house or apartments of such person " or persons or whether in the company or under the care or control " of such person or persons ".Is such a clause too vague and indefinite to satisfy the test, namely, whether a given person is within or without the class? Even adopting this construction Counsel for the Appellants argued that it was uncertain and Harman J. thought so because of the difficulty of the interpretation of the word " residing ".In a very careful argument Counsel for the Appellants advanced a number of points which he submitted showed there were 14 cases where the Trustees would have an impossible task to execute, but these alleged impossibilities can be classified I think under four headings; uncertainty upon the meaning of

"Residing"

persons " with whom " the son is residing

persons " in whose company ' the son is residing

persons " under whose care or control" the son is residing.6My Lords,upon this matter I agree entirely with the Court of Appeal. Many difficult and borderline cases may occur in any one of these situations. But mere difficulty is nothing to the point. If the Trustees feel difficulty or even doubt upon the point the Court of Chancery is available to solve it for them. It solves many such problems every year. I cannot for myself see any insuperable difficulty arising in the solution of any given state of affairs which would make it necessary to hold that the relevant clause as I have construed it fails to comply with the test. Of course I have not overlooked Sifton v. Sifton [1938] AC 656 but that was the entirely different case of a divesting clause. In my opinion, this clause is not void for uncertainty, and the Court of Appeal were quite right to overrule the decision of Harman J. in Re Gresham, where he held a similar clause was void on that ground.My Lords,that is sufficient to dispose of the appeal, but, as I have mentioned earlier, the reasons of two members of the Court of Appeal went further and have been supported by Counsel for the Respondents with much force and so must be examined.The Master of the Rolls propounded a test in the case of powers collateral. namely, that if you can say of one particular person meaning thereby, apparently, any one person only that he is clearly within the category the whole power is good though it may be difficult to say in other cases whether a person is or is not within the category, and he supported that view by reference to authority. Winn L. J. said that where there was not a complete failure by reason of ambiguity and uncertainty the Court would give effect to the power as valid rather than hold it defeated since it will not have wholly failed, which put—though more broadly—the view expressed by the Master of the Rolls. Counsel for the Respondents in his second line of argument relied upon these observations as a matter of principle but he candidly admitted that he could not rely upon any authority. Moreover, the Master of Rolls expressed the view that the different doctrine with regard to trust powers should be brought into line with the rule with regard to conditions precedent and powers collateral [1967] 3 W.L.R. 1115 G.So I propose to make some general observations upon this matter.If a donor (be he a settlor or testator) directs trustees to make some specified provision for " John Smith ", then to give legal effect to that provision it must be possible to identify " John Smith ". If the donor knows three John Smiths then by the most elementary principles of law neither the Trustees nor the Court in their place can give effect to that provision ; neither the Trustees nor the Court can guess at it. It must fail for uncertainty unless of course admissible evidence is available to point to a particular John Smith as the object of the donor's bounty.Then, taking it one stage further, suppose the donor directs that a fund or the income of a fund should be equally divided between members of a class. That class must be as defined as the individual; the Court cannot guess at it. Suppose the donor directs that a fund be divided equally between " my old friends ", then unless there is some admissible evidence that the donor has given some special " dictionary " meaning to that phrase which enables the Trustees to identify the class with sufficient certainty, it is plainly bad as being too uncertain. Suppose that there appeared before the Trustees (or the Court) two or three individuals who plainly satisfied the test of being among " my old friends " the Trustees could not consistently with the donor's intentions accept them as claiming the whole or any defined part of the fund. They cannot claim the whole fund for they can show no title to it unless they prove they are the only members of the class, which of course they cannot do, and so, too, by parity of reasoning they cannot claim any defined part of the fund and there is no authority in the Trustees or the Court to make any distribution among a smaller class than that pointed out by the donor. The principle is, in my opinion, that the donor must make his inten- tions sufficiently plain as to the objects of his trust and the Court cannot give effect to it by misinterpreting his intentions by dividing the fund merely among those present. Secondly, and perhaps it is the more hallowed principle, the Court of Chancery, which acts in default of trustees, must know with sufficient certainty the objects of the beneficence of the donor so as to execute7So if the class is insufficiently defined the donor's intentions must in such cases fail for uncertainty. Perhaps I should mention here that it is clear that the question of certainty must be determined as of the date of the document declaring the donor's intention (in the case of a Will, his death). Normally the question of certainty will arise because of the ambiguity of definition of the class by reason of the language employed by the donor, but occasionally owing to some of the curious settlements executed in recent years it may be quite impossible to construct even with all the available evidence anything like a class capable of definition (Re Sayer (supra)), though difficulty in doing so will not defeat the donor's intentions (Re Hain (supra)). But I should add this: if the class is sufficiently defined by the donor the fact that it may be difficult to ascertain the whereabouts or continued existence of some of its members ait the relevant time matters not. The Trustees can apply to the Court for directions or pay a share into Court.But when mere or bare powers are conferred upon donees of the power (whether Trustees or others) the matter is quite different. As I have already pointed out, the Trustees have no duty to exercise it in the sense that they cannot be controlled in any way. If they fail to exercise it then those entitled in default of its exercise are entitled to the fund. Perhaps the contrast may be put forcibly in this way: in the first case it is a mere power to distribute with a gift over in default; in the second case it is a trust to distribute among the class defined by the donor with merely a power of selection within that class. The result is in the first case even if the class of appointee among whom the donees of the power may appoint is clear and ascertained and they are all of full age and sui juris, nevertheless they cannot compel the donees of the power to exercise it in their collective favour. If, however, it is a trust power, then those entitled are entitled (if they are all of full age and sui juris) to compel the Trustees to pay the fund over to them, unless the fund is income and the Trustees have power to accumulate for the future.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, 3rd edition, volume 30, paragraph 445:
" The Court will not " compel trustees to exercise a purely discretionary power given to them " but will restrain them from using it improperly and if coupled with a duty " will compel them 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, 22 Ch D 521 at 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 proposition8advanced by Denning, M.R. and Winn, L.J. mentioned earlier, and agree with the proposition as enunciated in Re Gestetner and the later cases.My Lords,I would dismiss this appeal.

Lord Donovan

My Lords,Where trustees are given power to make payments out of a trust fund to members of a designated class, then if the trustees have a complete discretion as to whom they shall pay the money, or as to whether they shall make any payment at all. I see no reason why the whole clause should be regarded by the law as void for uncertainty simply because some members of the designated class might be unascertainable. It is true that this means that such members will never even get considered as possible objects of the trustees' discretion, but I regard that result as less unfortunate than depriving all the ascertainable members of any chance of benefit. I find myself therefore in complete sympathy with the rule that provided one can say with certainty whether a given individual is or is not a member of the class, the power collateral (as it is called) does not fail altogether simply because a complete list of every member cannot be drawn up.The Appellants in the present case, however, contend that the Trustees could not say with certainty that any given individual was a member of the class. Suppose, for example, that a particular claimant presented himself and said that Mr. Nubar Gulbenkian was " residing " with him. What tests would the Trustees apply to determine such " residence "? Would a short stay of a week be enough, or must there be something more enduring about the stay? The Trustees might even have to go further and ask themselves whether Mr. Gulbenkian was residing or employed under the care and control of the claimant. How would they decide what " care and control " meant in this context? I would agree that if one attempted an exhaustive definition of the persons and circumstances envisaged by the settlor when putting his signature to a settlement containing these words, the task would be impossible. But that simply means, I think, that one cannot determine at the outset all those persons comprised in the class—which is not fatal to the clause. When, however, a particular claimant presents himself, I think it is going too far to assert that the question whether he is qualified or not would be unanswerable. Take the word " residing " for example. There are numerous cases where the Courts have had to decide whether " A" was residing within the United Kingdom, or even " ordinarily" residing there: and though some of these problems were certainly difficult none was insuperable. If the Trustees were in doubt upon the matter, they could seek the ruling of the Court, which in my opinion would be forthcoming, not only as regards the question of " residing ", but also on the problems raised by the words " care and control", " employed" and so on. I agree, therefore, that the clause is not void for uncertainty. I should, however, like to associate myself with the criticism of my noble and learned friend, Lord Reid, of the language of the provision in dispute. I have never understood why some conveyancers should regard it as beneath their dignity to employ sub-paragraphs in a clause, so as to make their meaning plain. Much trouble and expense would have been avoided had that been done here.Other matters were canvassed before your Lordships. First, suppose that one individual alone is certainly within the class ; does the power collateral remain good even though the identity of the remainder of the class is uncertain: in other words, even though it would be impossible to say that any one given claimant out of the remainder was or was not within the class? The Master of the Rolls and Lord Justice Winn appear to think that certainty as regards one particular person only would be good enough: whereas hitherto the decided cases seem to have laid down that one must be able to tell of any given claimant whether he is within the9class or not, even though it may be impossible to make a complete list of the persons within the class. My noble and learned friend, Lord Upjohn, deals with this matter in his opinion, and while I am inclined to share his view I would reserve my opinion upon it, since in this case it would be purely obiter.Second, it has been suggested that it is irrational that in the case of a power which imposes a trust on the donee it should be necessary to the validity of the power that the whole range of objects eligible for selection should be capable of ascertainment, although this is not required in the case of a power collateral. At the moment I do not share this view. Nor did the Court of Appeal in I.R.C. v. Broadway Cottages Trust ([1955] 1 Ch. at page 36) for they described it as based on sound reasoning. This reasoning is developed in the opinion of my noble and learned friend, Lord Upjohn, and I agree with it. I say no more upon the matter for the point does not presently arise for decision.I agree that the appeal should be dismissed.(318582) Dd.l96999 200 10/68 St.S.