SEC proposes custody amendments, acknowledging adviser self-custody for crypto

The Securities and Exchange Commission has proposed amendments to its custody rules, and the statement released with them on October 1 makes the case for the rewrite on grounds of practicality. Safeguarding client assets from misuse, misappropriation or loss remains the purpose of custody regulation, and the tenets the rules rest on—segregation of assets and proper controls—have not aged; what has aged, the statement says, is the machinery for delivering them, since a rule that firms cannot follow in practice protects nobody and supplies what it calls an illusion of protection.

The Commission has been here recently: its 2023 proposal would have required advisers to keep crypto assets with a qualified custodian while raising doubts in the same document about whether any qualified custodian could demonstrate the exclusive control the rule demanded. Staff Accounting Bulletin No. 121 tightened the bind by forcing on-balance-sheet recognition of crypto assets at the companies safeguarding them, and the net effect, in the statement's account, was an instruction to advisers to use custodians that were largely unavailable for that work or unwilling to take it. The commissioner who wrote the statement supported issuing the 2023 proposal as a way to force public discussion and criticized its construction at the same time, on the ground that it arranged for advisers to fail even when they tried to comply in good faith.

The proposal now on the table is presented as a workable path to compliance, and for novel crypto assets the statement concedes that self-custody by an adviser or a fund may be the only option available when no qualified custodian is willing or able to hold the position. It acknowledges the corollary as well: an adviser holding client assets itself carries an inherent conflict of interest, and the controls and disclosures that would attach to that route are the first thing a compliance officer will read in the full text.

The argument is partly about timing: custody rules were written for a world in which holding assets meant holding paper, and they have been stretched since — modestly for modern securities markets, more recently for digital assets. Two developments, in the statement's account, make a fresh look worthwhile: the emergence of new asset classes and the market infrastructure being built to support them. The claim is narrower than a change in philosophy, because segregation and control stay where they were while their application is treated as contingent on how assets are actually held, so that a paper certificate in a vault and a position recorded on a distributed ledger are not asked to satisfy the same physical mechanics.

For an advisory practice, nothing in the compliance manual changes on the strength of a proposal, and the statement does not map the route from proposal to effective rule. The reason the crypto custody question reaches practices at all is supply: where the only compliant route runs through a custodian that will not take the account, advice on digital assets stops at the edge of the asset class — the outcome the 2023 design tended toward, and the one the new proposal appears intended to undo. A firm that has kept digital assets off its shelf for custody reasons has a specific reason to track this rulemaking.

The binding constraint for most firms has rarely been the willingness to advise on digital assets; it has been the custodian list, the short set of institutions a firm can name when a client asks where the assets would sit. The 2023 proposal pulled the list and the rule in different directions: the regulation wanted a qualified custodian exercising exclusive control, and the accounting treatment of crypto made being that custodian an unattractive business. Read one way, the combination held an asset class many clients own outside the accounts that hold the rest of their wealth.

Custody has also become a commercial contest in the RIA channel, on terms that have nothing to do with the SEC: sweep revenue, product-shelf access and account minimums have turned platforms into competitors for client cash, and as this publication has argued, Fidelity's hard $100 million minimum turns custody for a small firm into an acquisition question. Both senses of the word land on the same test, which is whether a firm of modest size can hold what its clients own. A platform can decide the firm is not worth serving; a custody rule that requires infrastructure the firm cannot supply leaves it holding assets on terms the statement itself calls conflicted.

The 2023 version failed on supply as much as design: too few qualified custodians were willing to hold crypto assets, and the balance-sheet treatment of those assets gave them one more reason to stay out of the business. Whether that supply improves is the real test of the proposal, and the custodians who would have to stand behind it will settle the question.