Funding a Trust with Digital Assets

The trust deed lists them. The asset schedule gives the amounts. The settlor mentioned them at every family meeting for six years, and everyone knows roughly what they were worth.

He died in March. The trustee has the deed, the schedule and the certificate of death, and it cannot reach a single unit of any of it.

This is the only essay in the sub-cluster in which the failure is final. A deed of assignment never registered can be registered late. Shares never entered in the register of members can be entered after a dispute. An assignment of a policy never notified can be notified, tardily, with priority lost but the asset intact. A key that nobody knows is not a contested asset. It is an absence, and no court in Nigeria or anywhere else can order it back into existence.

That question does more work in this asset class than every instrument counsel can draft, and it is the question the funding operation has to answer twice: who knows it now, and who will be able to learn it later.

Who else knows the secret?

Families use one phrase for holdings that behave in completely different ways, and the funding operation is different for each. Sort them before anything else.

The first is self-custodied. Cryptocurrency held in a wallet the settlor controls, where the only thing standing between the asset and anyone in the world is knowledge of a private key or a recovery phrase. There is a ledger, and it is public, and it records nothing about persons. It records that a certain address holds a certain balance. It does not know the settlor's name, has never heard of the trust, and cannot be told.

The second is platform-held. An account at an exchange, a payment platform, a broker, a marketplace, a monetised channel. Here a register does exist — the platform's own account records, which name a holder exactly as an insurer's register names a policy-holder. The difficulty is different: that register is closed. Most platform terms prohibit transfer of the account, prohibit sharing credentials, and treat the relationship as personal to the holder. The register names somebody and will not be persuaded to name anyone else.

The third is conventional property that happens to be digital. A domain name held through a registrar. Copyright in software, a catalogue, a course, a body of writing. A database. A registered trade mark used online. These are ordinary property rights in an unfamiliar wrapper, they are assignable in writing like any other chose in action, and the registrar or registry that records them behaves like the registers in the earlier essays. Families often overlook them entirely while worrying about the first category.

Only the third funds easily. The second frequently cannot be funded at all. The first can be funded, but only by an operation most families get precisely backwards.

Consider what it means to hand somebody a private key.

In every other asset class, transferring the thing that controls the asset divests the transferor. Hand over the original share certificate and you no longer have it. Execute the assignment of a policy and the insurer records somebody else. Pay money out of your account and the balance falls.

A key can be copied. Writing a recovery phrase on a card and giving the card to a trustee transfers nothing whatever, because the settlor still knows the phrase, or photographed it, or wrote it somewhere else in 2019 and has forgotten where. Two people now control the asset absolutely and independently, and neither can exclude the other. Nothing has moved.

This is the analytical heart of the essay. In self-custodied digital assets, control and ownership are the same fact, and control is knowledge. A gift is not complete until the giver can no longer take the thing back, and a settlor who still knows the phrase can take it back this afternoon from anywhere on earth. The instrument of gift that worked so well for cash in the essay on funding with cash does not complete anything here on its own, because there is no third party holding the asset who can be told the settlor no longer owns it.

The consequence is that the transfer must be made by movement, and to custody the settlor has never had and cannot reconstruct. Not by delivering a key. By generating a new one in the trustee's hands and sending the assets to it.

The order below assumes a settlor with holdings across all three categories, which is the ordinary case rather than the exotic one.

The determination for this asset class has two limbs, and the second is the one families resist.

A self-custodied digital asset cannot be settled by instrument. It is settled only by movement to custody the settlor has never held and cannot reconstruct. A deed describing the holding, a key handed over, a phrase written in a sealed envelope and a promise made in front of the family are, individually and together, not a transfer. Until the assets sit at an address whose keys the settlor never knew, the trust owns a description.

A platform-held asset usually cannot be settled at all. The account is personal, the terms forbid transfer, and the platform will deal with a personal representative rather than a trustee. What the trust can hold is the company that holds the account, or the proceeds once the holding is realised. Where the family wants the trust to hold the asset itself, the honest answer is frequently that it cannot, and the plan should be built around that rather than in denial of it.

One further consequence deserves stating to any settlor who holds meaningfully in this class. Digital assets are the only holdings in this sub-cluster where doing nothing produces total loss rather than an argument. An unregistered deed leaves a claim. An unassigned policy leaves a payment to the wrong person, which is at least a person who can be sued. An unrecoverable wallet leaves nothing at all — no defendant, no register to correct, no application to make. The asset does not pass to the wrong hands. It ceases to be reachable by any hands, permanently.

There is one instruction in this essay that admits of no exception, and it is the one most commonly got wrong by careful people trying to be thorough.

Never write a private key, a recovery phrase, a password or a set of credentials into a will. A will admitted to probate becomes a document of record, capable of being inspected. The registry is not a vault. A settlor who writes his recovery phrase into his will has arranged for his holdings to be published to anyone willing to make an application, at the precise moment when nobody is watching the wallet and everybody knows he is dead.

The same reasoning applies with almost equal force to a letter of wishes, which may be shown to beneficiaries and their advisers and is not written for confidentiality of that order. the essay on the letter of wishes treats what a letter of wishes is for, and this is not it.

What belongs in the will or the letter is the map, not the key: that holdings of this kind exist, who the custodian is, which professional holds the sealed instructions, and what the trigger for release is. The instruction should be sufficient for a trustee to know where to go and useless to anyone who reads it without authority.

The design problem underneath is real and should be named to the family rather than solved silently. Access wide enough to survive a death is access wide enough to be abused during life, and every arrangement in this field is a chosen point on that line. There is no configuration that eliminates both risks. There is only a decision, taken deliberately, written down, and reviewed as the value of the holding changes.

A key that has been shared has not been given. It has been duplicated, and the giver still owns everything.

Six asset classes have now been put to the same question, and the answers make a series.

Land answered with the register of a state. Shares answered with the register of members. Insurance answered with the register the insurer keeps, and with the date a notice reached it. The family business answered with six registers at once and the discovery that an enterprise is not an asset. Cash answered that there is no register of ownership at all, so the family must keep the book itself. And digital assets answer that there is a book, kept perfectly and publicly, which records addresses rather than people — and that the family must therefore keep two books, the ledger's and its own, and match them.

The sequence has been arguing one proposition in six costumes. A trust owns what the relevant record says it owns, and the record is never the deed. The deed states an intention. Something else — a registry, a register, a receipted notice, a signed acknowledgment, a contemporaneous instrument, a transaction to custody the settlor never held — converts that intention into ownership. Nigerian trusts fail at funding far more often than they fail at drafting, and they fail because the family stopped at the intention.

What follows this essay is the instrument the six were written toward: a funding and evidence checklist, built asset class by asset class, with one verification line for every step in these pages. It is the document a settlor should be handed at the first meeting and a trustee should work through before it accepts anything.

Every other asset in this sequence can be recovered by argument. This one cannot. When the key is gone, the asset is not disputed — it is simply gone.