The Three Certainties of a Trust

Introduction

A trust is a legal arrangement where a settlor causes property to be held by a trustee for beneficiaries. For the trust to take effect, three requirements must be satisfied, drawn from the classic case of Knight v Knight (1840) 3 Beav 148: certainty of intention, certainty of subject matter, and certainty of objects (beneficiaries). If any one is missing, the trust fails. The property may instead pass as an outright gift, be treated as a mere power, or result back to the settlor’s estate.

This guide sets out each certainty in plain terms, explains how courts assess them (including the different tests for fixed and discretionary trusts), highlights the leading authorities, and offers practical steps to avoid common drafting and exam pitfalls.

What You'll Learn

• What “certainty of intention” requires, beyond using the word “trust”
• How to identify a valid subject matter, including tricky bulk and fungible assets
• The different tests for fixed and discretionary trusts when defining beneficiaries
• Key cases: Knight v Knight, Re Adams, Comiskey, Re London Wine, Hunter v Moss, McPhail v Doulton, Re Baden’s (No 2), Re Gulbenkian, Re Hay’s
• How courts handle vague terms like “friends” or “relatives”
• What happens when a certainty fails: gift, power, or resulting trust
• Drafting and problem-question strategies to keep a disposition valid

Core Concepts

Certainty of Intention

The settlor must intend to impose a binding obligation on the recipient to hold property for others. There are no magic words: using “trust” helps, but it is not essential. The court asks, taking the words and context together, whether the language is imperative (binding) rather than merely hopeful.

Key points

• Words of wish or hope (precatory wording) usually do not create a trust.
• Imperative language, or structure that compels holding for others, supports intention.
• Conduct can support intention, especially in self-declarations of trust.

Leading illustrations

• Re Adams and the Kensington Vestry (1884) 27 Ch D 394: “in full confidence” that the widow would do the right thing for the children was held to be precatory, so no trust arose.
• Comiskey v Bowring-Hanbury [1905] AC 84: despite “in full confidence”, a mandatory gift over to nieces created a binding trust. The overall terms were imperative.
• Paul v Constance [1977] 1 WLR 527: “It’s as much yours as mine” and consistent conduct were enough to show intention to hold money on trust.

Practical takeaway: focus on substance, not labels. Ask, on an objective reading, whether the recipient is obliged to hold for others, or is free to treat the property as their own.

Certainty of Subject Matter

The trust property must be identifiable and the extent of each beneficial share must be clear. If the court cannot tell what is held on trust—or in what proportion—the trust fails.

Two recurring issues

• Segregation of tangible items: if the trust is over part of a bulk of goods, you generally need to separate the specific items to identify the trust property.
• Fungible intangible assets: identical units of the same class of shares can be treated differently.

Leading illustrations

• Re London Wine Co (Shippers) Ltd [1986] PCC 121 (and see Re Goldcorp Exchange Ltd [1995] 1 AC 74): no trust where bottles or bullion were not segregated from a larger indistinguishable mass.
• Hunter v Moss [1994] 1 WLR 452: valid trust of 50 out of 950 identical shares of the same class—no segregation required for fungible shares.
• Sprange v Barnard (1789) 2 Bro CC 585: “the remaining part of what is left” failed—subject matter uncertain because the quantum depended on the first taker’s spending.
• Re Golay’s Will Trusts [1965] 1 WLR 969: a direction to provide a “reasonable income” was sufficiently certain; “reasonable” could be assessed objectively.

Practical takeaway: name the assets, or define a clear method for identifying them. For bulk items, segregate. For shares of a single class, a specified number may suffice without segregation. Ensure any standard (for example “reasonable income”) is capable of objective assessment.

Certainty of Objects (Beneficiaries)

The court must be able to identify who can benefit. The applicable test depends on whether the trust is fixed (beneficial shares are predetermined) or discretionary (trustees choose who benefits, and by how much, within a class). The beneficiary principle also applies: non-charitable trusts generally require human beneficiaries who can enforce the trust.

Fixed trusts — the complete list test

• IRC v Broadway Cottages [1955] Ch 20: you must be able to draw up a complete list of beneficiaries. If you cannot identify everyone in the class, the fixed trust fails.
• Terms like “my children” are clear. “My friends” is usually too vague. Note: in Re Barlow’s Will Trusts [1979] 1 WLR 278, “friends” worked for a gift subject to a condition precedent, but that is different from a trust obligation.

Discretionary trusts — the “is or is not” test

• McPhail v Doulton [1971] AC 424: it is enough that, for any given person, you can say they are within or outside the class. You do not need a complete list.
• Re Baden’s Deed Trusts (No 2) [1973] Ch 9: “relatives” and “dependants” were acceptable classes. Conceptual clarity is needed; evidential uncertainty about whether a person falls within the class does not necessarily defeat the trust.
• Administrative workability: even if conceptually clear, a class can be too wide to administer sensibly (see guidance in Re Hay’s Settlement Trusts [1982] 1 WLR 202, and the approach later applied in cases like R v District Auditor ex p West Yorkshire MCC [1986]).

Powers of appointment

• Re Gulbenkian’s Settlements [1970] AC 508: for a mere power, the same “is or is not” test applies. However, very wide or capricious powers can fail.

Practical takeaway: define classes by objective criteria (“children”, “lineal descendants”, “employees with five years’ service”) rather than subjective terms like “friends”. For discretionary trusts, ensure the class is conceptually clear and workable in practice.

Key Examples or Case Studies

• Knight v Knight (1840) 3 Beav 148

  • Point: Sets out the three certainties (intention, subject matter, objects).
  • Application: Use it as your starting framework in any trust validity question.

• Re Adams and the Kensington Vestry (1884) 27 Ch D 394

  • Point: “In full confidence” was precatory; no trust arose over the gift to the widow.
  • Application: Be wary of soft language that expresses hope rather than obligation.

• Comiskey v Bowring-Hanbury [1905] AC 84

  • Point: Overall wording was imperative because of a mandatory gift over; a trust existed.
  • Application: Read the instrument as a whole—structure can turn vague phrases into an obligation.

• Re London Wine Co (Shippers) Ltd [1986] PCC 121 and Hunter v Moss [1994] 1 WLR 452

  • Point: Tangible goods in bulk require segregation; identical shares of one class may not.
  • Application: Ask whether the assets are fungible intangibles; if not, identify and separate them.

• Sprange v Barnard (1789) 2 Bro CC 585 and Re Golay’s Will Trusts [1965] 1 WLR 969

  • Point: “What remains” failed for uncertainty; “reasonable income” succeeded as an objective yardstick.
  • Application: Avoid subject matter that fluctuates at a beneficiary’s whim; do use objective measures.

• McPhail v Doulton [1971] AC 424 and Re Baden’s Deed Trusts (No 2) [1973] Ch 9

  • Point: Discretionary trusts use the “is or is not” test; “relatives” was acceptable with conceptual clarity.
  • Application: Define classes so you can say of any person if they are in or out, even if you cannot list everyone.

• Re Gulbenkian’s Settlements [1970] AC 508 and Re Hay’s Settlement Trusts [1982] 1 WLR 202

  • Point: The “is or is not” test extends to mere powers; very wide or unworkable classes may fail.
  • Application: If a trust looks too broad, consider whether it should be construed as a power—or whether it is administratively unworkable.

Practical Applications

Drafting pointers

• Use clear, imperative language: “I declare that X shall hold [property] on trust for…”
• Identify the property precisely. For bulk goods, segregate. For shares, specify class and number.
• State beneficial shares for fixed trusts, or set out a clear discretionary class for discretionary trusts.
• Avoid vague descriptors like “friends”. Prefer objective labels: “children”, “grandchildren”, “employees of [company] on [date]”, “lineal descendants”.
• If using “relatives”, consider defining the degree (for example “issue” or “descendants of my grandparents”).
• Consider a gift over if the trust cannot be operated (for example, no eligible beneficiaries).
• Add a non-binding letter of wishes for discretionary trusts to guide, but not fetter, trustee decisions.

Problem question workflow

1. Intention: Is the language imperative, taking the whole instrument and conduct into account? Cite Re Adams and Comiskey (and Paul v Constance for self-declarations).
2. Subject matter: Is the property and the quantum identifiable? Deal with Re London Wine / Hunter v Moss; consider Re Golay and Sprange.
3. Objects: Is this fixed or discretionary?
   • Fixed: can you compile a complete list? IRC v Broadway Cottages.
   • Discretionary: is-or-is-not test; conceptual certainty and workability; McPhail, Re Baden, Re Hay’s.
4. Salvage: If a trust fails, can it stand as a mere power (Re Gulbenkian)? If not, consider resulting trust.
5. Outcome: State the effect on the property (outright gift, trust valid, power, or resulting trust).

Trustee actions

• Do not distribute until the trust property is clearly identified and, where needed, segregated.
• Keep records of how you assessed class membership and exercised discretion.
• If the class appears unworkable or the terms unclear, seek directions from the court.

Summary Checklist

• Confirm intention: imperative wording or clear conduct? Avoid purely precatory phrases.
• Pin down the property: name the assets and quantify beneficial shares.
• For tangibles in bulk, segregate; for identical shares of one class, segregation may not be needed.
• Avoid “friends”; prefer objective class definitions and, for “relatives”, define the degree of kinship.
• Fixed trust → complete list test (IRC v Broadway Cottages).
• Discretionary trust → “is or is not” test (McPhail; Re Baden).
• Consider administrative workability for very wide classes (Re Hay’s; ex p West Yorkshire).
• If a trust fails, assess whether a power remains (Re Gulbenkian) or whether there is a resulting trust.

Quick Reference