Fund tokenisation restructures who is accountable, not just the technology

The framework has three parts, and only the first is about technology.

The first part confirms that DLT-based registers are permitted. The FCA’s position, carried from CP25/28 into PS26/7, is that the COLL register rules are technology-neutral and outcomes-based, so recording a fund’s register of units on a distributed ledger, including a public, permissionless one, already sits inside the existing rules. The guidance clarifies the existing position rather than granting a new permission. Even this technical part carries a governance condition, though. The responsible firm has to keep the ability to make unilateral changes to the register, for court orders, deaths, divorces, fraud and mandatory redemptions, whether through private keys, a master-node function or contractual terms. To run the register lawfully, the manager must be able to override the ledger.

The second part is the one that reorganises the firm: the direct dealing, or direct-to-fund, model. In the conventional arrangement the authorised fund manager acts as principal in every investor deal, buying and selling units into and out of a manager’s box, standing between the investor and the fund. Direct dealing lets the fund itself, or its depositary, act as principal instead, with cash settling directly between the investor and the fund and the manager stepping out of the middle. The FCA has made this available to all authorised funds, tokenised or not, but it’s the natural pairing for a tokenised fund, and it’s where the accountability moves. This is a change in dealing counterparty, not systems.

The third part is the staged plan. The FCA has set out a path that runs from tokenised registers now, through tokenised assets and smart-contract-driven model portfolios, to fully on-chain products held directly in investor wallets. It has signalled that at the later stages the fund structure itself may start to give way to discretionary and model-portfolio arrangements, which is why it intends to review the portfolio-management conduct rules. The staged plan shows this is a structural programme.

When a manager adopts direct dealing, four distinct accountabilities move at once.

The principal moves. The manager stops acting as principal in unit deals and the fund or depositary takes the role. That single change removes the manager’s interim holding of units and the client-money consequences that came with it, which cuts the manager’s balance-sheet and capital burden while also removing a buffer the manager used to control. The deal now runs between the investor and the fund.

The client-money position changes. Because the manager is no longer the counterparty, the client-money rules apply differently, and a new dedicated account, the issues and cancellations account, comes in. That account is scheme property under the new rules, with its own protections, valuation treatment and a requirement that it not go overdrawn. A board that thinks it is approving a register migration is in fact approving a new way of holding cash, with client-asset implications of its own.

The anti-money-laundering obligation moves. This is the change most likely to be missed. Under direct dealing the relevant person for anti-money-laundering purposes can shift from the manager to the fund or depositary, and the scheme documents have to state who carries the obligation. A firm that adopts direct dealing without assigning AML responsibility in writing leaves an accountability gap.

The depositary’s role widens. The depositary picks up oversight of the new issues and cancellations account, custody questions for any digital assets, satisfaction on the custody of tokenised holdings, and access to a UK-legible reproduction of the on-chain register. Depositary boards are being told they need technology skills they didn’t previously carry.

None of these four is a technology decision. Each has to be decided and documented before the fund tokenises.

Two questions sit under the restructuring that the FCA’s framework raises without fully answering. A board is better off confronting both than waiting for them.

The first is a senior-managers question: who answers when a ledger-dependent or smart-contract-driven process fails. Say a fund uses an allow-list or eligibility check enforced in code to restrict who can hold its tokenised units, and the code fails, or admits a holder it should have kept out. Under the senior managers regime it rests with a named individual, and the firm has to decide, in advance, which senior manager owns the integrity of the tokenised processes and can evidence the reasonable steps taken to assure them. Liability for a tokenised fund on a public network has to be assigned in advance. And because using a permissionless network isn’t outsourcing, there being no provider to contract with, the firm can’t point at the network; it stays responsible for a rail it doesn’t own.

The second is a Consumer Duty question, and it ties tokenisation to fund economics. If tokenisation and direct dealing cut the manager’s costs in a real way, by taking out transfer-agency layers, capital and reconciliation, the price-and-value outcome at PRIN 2A.4 raises an awkward point: whether those savings should reach investors. A manager that keeps the whole efficiency benefit while holding fees flat is making the decision the assessment of value is designed to test, and it should expect to justify it. Tokenisation also affects the fair-value obligation under Consumer Duty.

The staged plan is where the change becomes strategic, not just operational. The FCA frames tokenisation as moving through three stages. First is the tokenisation of funds, recording the register on a ledger, which is the current UK position. Second is the tokenisation of assets, where investors come to hold tokenised assets more directly and managers run what amount to smart-contract-driven model portfolios, at which point some of what a fund does can be done without a fund wrapper at all. Third is the tokenisation of cash flows, where assets are broken into tokenised components and recombined, the world of composability and embedded, on-chain compliance.

The reason this matters to a board now, while the firm is only weighing stage one, is that the later stages point straight at the fund structure. If a manager can deliver exposure through tokenised assets and smart-contract model portfolios, the authorised fund, with its AFM, its depositary, its register and its dealing machinery, starts to look like one way of doing something that could be done other ways, and the FCA has already said it will review the portfolio-management conduct rules in anticipation. A manager should tokenise with a view on where the fund structure itself is heading, not just the first step. The strategic question is what business the firm is in once the ledger can do much of what the fund structure used to.

There’s a resilience problem a tokenised fund can’t treat as someone else’s. The FCA proposes to extend its operational resilience requirements to all crypto firms, and a tokenised fund running on a public, permissionless network sits squarely inside that, because the network is critical to the fund’s ability to deal and to keep its register, yet nobody owns it and there’s no provider to contract with. The FCA has been explicit that using a permissionless network isn’t outsourcing, because there’s no contractual counterparty, and that the firm stays fully responsible for the resilience and integrity of the service regardless. So the board that tokenises has to set an impact tolerance for a register and a dealing process that depend on infrastructure it doesn’t control, and has to hold a tested plan for a network outage, a congestion event or a settlement failure, as it would for a critical third party, except that here there’s no third party to call.

The FCA’s framing puts three decisions to fund boards.

Whether to adopt direct dealing is a business-model decision. It changes the firm’s role, its capital, its client-money exposure and its AML accountability, and the board should take it with those consequences understood, rather than arrive at it as a by-product of a register migration.

Where the firm intends to sit on the tokenisation value chain in three to five years is a strategy decision. The staged plan runs from tokenised registers toward a world in which the fund wrapper may give way to tokenised assets and model portfolios, and a manager that doesn’t decide where it wants to be on that path will have the position decided for it, by competitors and by the later stages of the FCA’s own programme.

Who carries liability for DLT-dependent processes is a governance decision. It has to be assigned to a named senior manager, evidenced, and built into the firm’s accountability framework before the fund goes live, because the alternative is an unowned risk on a public rail.

These decisions get skipped when tokenisation is framed as a technology project.

In each row below, what the technology team treats as a build is a shift in responsibility the board has to own. A tokenisation project belongs with the board before it reaches the engineers.

This insight is provided for general informational purposes only and doesn’t constitute legal, investment, or regulatory advice.