Three Certainties and Creation of Express Trusts — SQE1 FLK2 Trusts Law

A woman's will leaves her farm to her brother, in the confident expectation that he will pass it to whichever of their nephews he considers best suited to run it. The brother has moved in. He says the farm is his, and he is putting it on the market. Words of expectation are not words of command, so is he right? No. And the reason is not in that sentence at all. It is in the next one, the sentence most candidates skim.

This is Three Certainties and Creation of Express Trusts, the first topic in Trusts Law, and everything else in the subject sits on it. Three questions, asked in order, and a trust that fails any one of them was never a trust. Keep the farm in mind. We are coming back to it.

Here is the route. What an express trust actually is, and what happens when one fails. Then the three certainties in their usual order: intention, subject matter, objects. Then the line between a trust and a power, and between a fixed and a discretionary trust, because that line decides which test applies. Then constitution, which is what happens when the settlor means it but never quite does it.

Start with the shape. An express trust has three parties. The settlor creates it, usually by deed or by will. The trustee holds the legal title. The beneficiary holds the equitable title, which is the ownership that matters. And the trustee is under an obligation, not merely given a choice, a point we come back to.

Now the stakes. If any one of the three certainties fails, the trust does not exist. Not defective. Not voidable. It never existed. The property then does one of three things. It passes as an outright gift, or is held on resulting trust for the settlor, or fails entirely and falls into the estate.

Certainty of intention first. The question is not whether the settlor used the right vocabulary. It is whether the words, read in context, impose a binding obligation to hold property for someone else, or merely express a wish. Nothing turns on the word trust. It never has to appear.

The leading case makes that concrete. A man kept his damages in a bank account in his own name. He told his partner, repeatedly, that this money is as much yours as mine, and that she could draw on it when she needed to. No document, no deed, no lawyer. The Court of Appeal held that those words declared a trust of the balance. Paul v Constance. Informality is not the enemy of intention.

Now the other side, and this is where the marks are. Precatory words are words of hope, wish or desire. On their own they create nothing. A man left his estate to his wife in full confidence that she would do what was right as to the disposal between his children. The Court of Appeal held there was no trust. She took it absolutely, and the children took nothing.

Hold that phrase, in full confidence, and now take a will that used the very same words. This testator gave his wife the property in full confidence that she would leave it to such of his nieces as she thought fit. But the will went on. It directed that in default of any disposition by her, the property should be divided among the surviving nieces at her death. Same words, opposite result. A trust. Comiskey v Bowring-Hanbury.

What made the difference is the gift over in default. A man who tells you what happens if you do not do the thing is not expressing a hope. He is imposing an obligation and providing for its breach. So intention is judged on the instrument as a whole, never on the precatory phrase in isolation.

Which takes us back to the farm. The confident expectation, standing alone, would have given the brother the farm absolutely. But the will goes on to direct that if he makes no such provision, the farm is divided equally between the nephews on his death. That is a gift over in default. He holds the farm on trust, and he cannot sell it as his own.

Two more points on intention. A moral obligation is not a trust. I hope you will look after my mother imposes nothing a court will enforce. And in commercial dealings a trust is harder to find, because those parties are taken to be creating contractual rights rather than equitable ones.

Harder is not impossible, and the way through is usually segregation. A mail order company facing insolvency paid customer prepayments into a separate account named as a customers' trust deposit account. Keeping the money apart showed the intention. Those customers were beneficiaries, not creditors in the queue.

Before the second certainty, the distinction that governs everything after it. A trust is mandatory. The trustee must distribute. A power is permissive. The holder may distribute, or may not, and nobody can compel the exercise.

Try one. A will leaves the residue to trustees to be divided equally between the testator's three nieces. One niece has married money and plainly does not need it. The other two would benefit far more, and they have agreed in writing. May the trustees pay the whole residue to those two? No. Not a discretion in sight. Equally between my three nieces fixes three shares of a third each, and paying two of them the lot is a breach of trust, however sensible it looks.

So read the verb. Shall distribute is the language of obligation. May distribute, or shall distribute to such persons as my wife shall appoint, is the language of choice. Get that wrong and you apply the wrong test to the objects, which is the commonest way to lose the mark.

Certainty of subject matter. Two limbs, and both must hold. First, the trust property itself must be identifiable: you must be able to say what is held on trust. Second, the beneficial interest must be identifiable: you must be able to say what share each beneficiary takes. Fail either and the trust is void.

Take the first limb. A woman keeps a cellar of about 300 bottles, many vintages, many values. She signs a note telling her nephew that from today she holds 50 bottles from the cellar on trust for him. Nothing is moved, marked or listed. She never says which 50, and she goes on drinking from the cellar. She dies, and the cellar passes to her sister. Does the nephew get 50 bottles?

He does not. The intention is plain and the object is plain, but no particular bottles were ever set aside, so there is nothing the trust can attach to. That is the rule from Re London Wine, where customers who had paid for wine still sitting in bulk in a warehouse took nothing and ranked as unsecured creditors. A later case applied the same reasoning to unallocated gold in a vault.

Now change one fact and watch the answer flip. A man owns 1,000 ordinary shares of one class in a company. He signs a note declaring that he holds 200 of them on trust for his goddaughter. He does not say which 200, nothing is separated, and all 1,000 stay in his name. Valid, or void for the same reason as the wine?

Valid. The shares are intangible and identical, so it makes no difference which 200 answer the trust. Any 200 will do the job, and that cannot be said of 50 bottles out of 300 that differ in vintage and value. Hunter v Moss, where the owner of 950 shares declared a trust of 50 of them without saying which, and the Court of Appeal upheld it. Segregate tangible goods. Identical intangibles look after themselves.

The second limb catches a different mistake. A share that cannot be quantified is no share at all. But do not assume a loose sounding formula is fatal. A direction to pay a beneficiary a reasonable income has been held certain, because reasonableness is an objective standard a court can apply. Hard to calculate is not the same as impossible to define, and that distinction is about to become the whole of the third certainty.

Certainty of objects. Who the beneficiaries are. Here the test depends on which kind of trust you have, which is why the trust and power point came earlier.

For a fixed interest trust, where each beneficiary takes a defined share, the test is the complete list test. It must be possible to list every single beneficiary, which follows from the nature of the thing. If the fund is divided into equal shares, you cannot size a share until you know how many people there are. A class described as my friends, with nothing more, fails.

For a discretionary trust the test is far lower, and it comes from McPhail v Doulton. Lord Wilberforce held that it is enough if it can be said with certainty that any given individual is or is not a member of the class. No list is required. The trustees need only answer the question about whoever turns up in front of them. It is called the any given postulant test, or the is or is not test.

The trust in that case was for the employees and former employees of a company and their relatives and dependants. Could the trustees list every relative of every former employee? Plainly not. Was the class conceptually certain? Yes. And that is the distinction the examiners build questions on.

So take it slowly. Conceptual uncertainty means the definition of the class is itself vague or subjective, because the words have no settled edge. Evidential uncertainty means the definition is clear but the facts are hard to gather. All my descendants living at my death is conceptually clear and evidentially a nightmare. Only conceptual uncertainty kills a trust.

And the courts work hard to save a trust. When that employees trust came back to the Court of Appeal, three judges took three different views on applying the is or is not test to the word relatives. One required only conceptual certainty, leaving a claimant to prove they were in the class. One asked whether a substantial number clearly qualified. One took the strictest line, wanting an answer either way for anyone at all. All three upheld the trust.

Now what turns on fixed against discretionary, beyond the test. Under a fixed trust each beneficiary has a vested proprietary interest in a defined share from the moment the trust is created. Under a discretionary trust nobody has a proprietary interest at all. Each object has a right to be considered, and nothing more, until the trustees select.

And one rule you should know by name. Where the beneficiaries are all of full age and sound mind, and between them absolutely entitled to the whole fund, they can require the trustees to hand it over. Whatever the settlor wanted. That is the rule in Saunders v Vautier.

See it work. A fund pays income to the testator's brother for life, and the capital then goes to his niece absolutely. Both are adults of full capacity. Between them they exhaust the whole beneficial interest, so between them they can collapse the trust today, however long the testator meant it to run. On a discretionary trust the same rule needs every possible object to agree, which is close to unheard of.

Last, constitution, because a trust that satisfies all three certainties can still fail. A trust is incompletely constituted where the settlor has declared it but never transferred legal title to the trustee, and equity will not assist a volunteer. The rule is that the settlor must have done everything necessary, given the nature of the property, to transfer it. A settlor who tried to move shares but never had the transfer registered had not done it, and the trust failed.

Two exceptions, and the first is narrow. Where the settlor has done everything within their own power, the trust is constituted from the moment the settlor became powerless to stop it. That holds even if an administrative act by a third party, such as the company's registrar, still remains. The Court of Appeal has gone further. It has held a gift effective where the donor had not quite done everything, but where it would have been unconscionable for her to take it back.

The second exception is proprietary estoppel: a promise that property will be held for the claimant, reliance on it to the claimant's detriment, and unconscionability in going back on it. There is also a deathbed exception. A gift made in contemplation of imminent death, with delivery of the property or the means of control, takes effect on the death.

And if nothing helps, be clear where the property lands. Where a trust was intended but the transfer never completed, it is held on resulting trust for the settlor. Where an outright gift was intended, it stays with the settlor.

A word on how SQE1 tests all this. You will not be asked to recall a case name, and you will not be asked for a citation. You get a scenario, five answers, and one instruction: pick the best. The names in this episode are memory pegs. The rules are the marks.

If you keep only three pegs, keep these. Paul v Constance, where a man who never used the word trust created one anyway. Hunter v Moss, where 50 shares out of 950 were a valid trust and 50 bottles out of a cellar were not. And McPhail v Doulton, which asks only whether any given individual is or is not in the class.

Five traps. One: informality cuts both ways. Casual words can create a trust, and formal sounding words of confidence may create nothing. Do not decide from the tone of the language. Decide from whether an obligation was imposed.

Two: never read a precatory phrase on its own. The examiners will put the gift over in default in the following sentence, and that sentence is the answer.

Three: a vague sounding measure is not automatically uncertain. A reasonable income is certain, because reasonableness is a standard a court can apply. Ask whether there is an objective yardstick, not whether the words sound woolly.

Four: only conceptual uncertainty voids a trust. Difficulty in tracing people is evidential and does not matter. Five: settle whether the trust is fixed or discretionary before you touch certainty of objects. Divided equally between points one way. Such of my family as the trustees select points the other, and the test is different.

Quick check. A settlor transfers £250,000 to trustees by deed. The deed directs them to distribute the fund among such of his relatives as are most deserving of help. The trustees hold a full family tree, so they can say of anyone whether that person is a relative. But several relatives emigrated decades ago, and tracing them would take months. A cousin, who takes the fund if the trust fails, says there is no valid trust.

Three candidate answers. One: valid, because the difficulty of tracing the emigrated relatives is only evidential uncertainty. Two: valid, because the trustees can say of any given person whether he or she is a relative. Three: void, because most deserving of help supplies no objective standard for the class. Pause here if you want a moment.

The answer is three. A discretionary trust needs conceptual certainty, and relatives is conceptually certain. That much the trustees have. But the settlor added a second requirement, and most deserving of help gives no yardstick by which the trustees or a court could decide who qualifies. The class cannot be defined, so the trust is void and the fund goes to the cousin.

Why the others fail. One states a correct principle about the wrong problem. The tracing difficulty is only evidential, but that is not where this trust breaks. Two answers only half the definition. Certainty about who is a relative does nothing to cure the vagueness of what the settlor added on top.

Five things to take away. One: three certainties, and a trust that fails any of them never existed, so ask what the property does instead. Two: intention is about obligation, not vocabulary, judged on the whole instrument. Three: segregate tangible goods, identical intangibles need no segregation, and the beneficial share needs an objective standard.

Four: fixed trusts need a complete list, discretionary trusts need only the is or is not test, and only conceptual uncertainty is fatal. Five: the certainties are not enough. Legal title has to reach the trustee, and equity will not assist a volunteer waiting for it.

And the farm. The confident expectation on its own would have handed the brother the farm outright. The direction about what happens if he makes no provision is what converts it into an obligation. He is a trustee. Next time, Formalities and Constitution of Express Trusts.

Practise this topic with exam-style questions at sqe1prep.co.uk. This episode is for education and exam revision only, not legal advice, and we are not affiliated with or endorsed by the SRA or Kaplan.