The protector is the most under-drafted office in offshore trust practice, and this year two things happened to it. On 19 March the Privy Council allowed an appeal from Bermuda and held that where a deed requires a protector's consent but says nothing about how that consent is exercised, the protector is not a reviewer of the trustee's legality. It holds a discretion of its own. On 20 May 2026 the States of Guernsey approved in principle amendments to the Trusts (Guernsey) Law that would reverse the presumption that a protector or other reserved-power holder is not a fiduciary. As approved, reserved powers would be presumed fiduciary unless the deed says otherwise.
Both moves run the same way: reserved power now attracts duty, not immunity. That is awkward for a device sold as comfort with no exposure, and it means the drafting decides which one you bought.
What the Board actually decided
[A v C [2026] UKPC 11](https://caselaw.nationalarchives.gov.uk/ukpc/2026/11) was heard in private, and the parties are anonymised; the judgment of Lord Briggs and Lord Richards is public. Deeds commonly give a protector a veto over capital appointments and dealings in the family company's shares, without saying what standard the protector applies. Counsel framed the choice as a "narrow role" — veto only what is unlawful or irrational — against a "wider role" — decide the proposal on its merits, subject to fiduciary duties.
The Board declined the binary. At paragraph 83 it reframed the question as what constraints the instrument actually imposes, construed in context and against the general law. The answer for these trusts was: almost none beyond the ordinary duties. A power of veto, the Board observed, places the holder under no constraint at all save possibly good faith — the analogy was a landlord's consent to assignment.
Three features of the deeds did the work: the protectors could waive their own consent powers, an odd thing to give a watchdog; where joint protectors disagreed the trustees could proceed after consulting them, which makes little sense if the protector's job were to police legality; and consent was required only for a limited set of the most important decisions. Fiduciary status did not narrow the role: the no-profit, no-conflict and proper-purpose duties are formidable, but they operate inside the wider role. The absence of a specified role must be taken to have been a deliberate gap, and there is "simply no peg on which to hang the Narrow Role". A settlor who wants a watchdog has to say so.
Why this is an asset-protection question
A trust protects assets by taking ownership away from the settlor. Reserved powers put some of it back, and the English courts have been explicit about what happens when too much comes back. The leading authority is Pugachev: [JSC Mezhdunarodniy Promyshlenniy Bank v Pugachev [2017] EWHC 2426 (Ch)](https://caselaw.nationalarchives.gov.uk/ewhc/ch/2017/2426), where Birss J construed New Zealand trusts whose first protector was the settlor himself and a named discretionary beneficiary. Control ran through the protector's office: a veto over distributions, investment and variation, a power to appoint beneficiaries, a power to remove trustees with or without cause, and a clause switching the office away if he fell under a court order he did not want to obey.
On their true construction the judge found those powers to be "purely personal powers which may be exercised selfishly". The consequence was not bad administration: the deeds never moved the assets at all. They amounted to a bare trust, and Mr Pugachev remained beneficial owner of what he had settled. The switch clause drew the sharpest language — an attempt "to make the trust judgment proof", and one the court would not accept.
The fork is why reserved powers decide asset-protection cases. These trusts were not shams precisely because they left the settlor in control; had the protector's powers instead been fiduciary, the judge held, the same deeds would have been shams, the settlor having intended a false impression of ownership and the trustees having gone along with it recklessly. Either answer defeats the structure, by a different route.
Where the sham line sits
Sham itself remains a high bar. The test is Lord Diplock's in Snook v London and West Riding Investments (1967), read through Arden LJ in Hitch v Stone (2001) and Munby J's survey in A v A (2007): the parties must have shared a common intention that the documents not create the rights they appear to create, and the false impression must be intended for third parties. A settlor's unilateral intention is not enough. Reckless indifference by the other side will do. A trust that was not a sham at the start cannot become one later.
That leaves an awkward distribution of risk. Sham is hard to prove against an independent trustee, which is why the creditor's practical attack is usually ownership instead: the argument that the settlor never parted with the assets, run as illusory trust or, more prosaically, true effect. The new material makes that argument easier to run rather than harder, because a protector exercising a real discretion on the merits generates the paper trail showing who decided.
Guernsey moves the default
Guernsey's reforms were approved in principle on 20 May 2026 and are not law yet: legislation still has to be drafted and approved.
What the States approved is a reversal of section 15(2)(b) of the Trusts (Guernsey) Law, 2007, under which a power holder or protector does not by default owe fiduciary duties unless the terms say so. On the approved basis, reserved powers would be presumed fiduciary unless the instrument provides otherwise, with section 32 aligned so that consent and veto powers are treated consistently. Ogier's account of the existing regime shows how wide it is: Guernsey legislation neither defines nor uses the word protector, while section 15 lets power over investments, amendments and consent be reserved to a named person.
The Hong Kong drafting consequence
Hong Kong's statute is narrower than the offshore pitch implies, and the gap is where the drafting risk sits. Part 4D of the Trustee Ordinance (Cap. 29), added by the 2013 amending Ordinance (13 of 2013) and in force since 1 December 2013, contains one reserved-powers provision. Section 41X(1) provides that a trust is not invalid only because the settlor reserved to himself "any or all powers of investment or asset management functions under the trust", and subsection (2) shields a trustee who acts in accordance with such a power. Part 4A defines asset management functions as the trustees' functions over the investment, acquisition, management and disposal of trust property.
That is the whole harbour. It says nothing about a veto over distributions, a power to add or remove beneficiaries, a power to remove trustees, or a power to change the governing law — the powers that in practice carry the control. Nor is it conclusive: section 41X(4) provides that where a trust's validity is questioned, the court may take into account subsection (1) in determining validity. The word protector does not appear in the Ordinance at all.
So a Hong Kong-law deed that reserves protector powers is governed by its own words and the general law — and after the Privy Council's decision, its own silence has a settled meaning. The checklist writes itself. State the standard the protector applies, expressly, rather than leaving the role to be inferred; say whether the powers are fiduciary or personal, and price the Pugachev analysis if they are personal. Keep the protector out of the beneficial class unless there is a reason to include them — both the Board this year and Birss J drew the line at a protector who is also a beneficiary and the settlor. Avoid the automatic switch that moves the office away the moment the settlor is under compulsion. And minute the reasoning: a real discretion exercised for the beneficiaries is the evidence that the structure did what the deed said.
None of this makes a protector a bad idea. It makes an unconsidered protector a bad idea — and for a family office in Hong Kong holding a deed drafted before 2026, the more urgent one. Guernsey's legislation is still to be written, and Hong Kong's statute will not close the gap on its own. The document to read next is the deed in the drawer, starting with the clause that appoints the protector.
