Visa and Lloyds finished the stablecoin trial. Now publish the numbers

“Completed” is doing a great deal of work in this announcement.

Visa and Lloyds Banking Group say they have finished a live pilot of stablecoin-based settlement, and that the approach could support faster, more transparent and more flexible cross-border transactions. The first half of that is a fact: a thing was run with real money and it did not fall over. The second half is a claim, expressed as three adjectives, with no number attached to any of them. Cross-border payments have been the industry’s unfinished business for well over a decade, and those same three adjectives have introduced almost every attempt to fix it. They have stopped being informative.

Faster than what, exactly

A cross-border payment is not one movement. It is an instruction, a sanctions and fraud screen, a funding decision, a messaging leg, a settlement leg, and finally a credit the beneficiary can actually spend. Each has its own clock. Stablecoin settlement replaces one of those legs — the one where value moves between institutions.

In a correspondent chain that leg can indeed be slow, because of cut-off times, weekends and the number of intermediaries involved. But it is frequently not the binding constraint. Screening queues and the receiving bank’s own posting cycle routinely are. Swap a slow-ish leg for an instant one and leave the rest untouched and the corporate treasurer on the other end may see no change at all in when the money is usable.

So the only honest test is end to end: time from instruction to funds available to the beneficiary, measured against the correspondent baseline for the same corridor, the same currency pair and the same size of payment. Until somebody publishes that comparison, “faster” is not a finding. It is a hope with a press release.

The number that would actually be impressive

The real prize in wholesale cross-border payments is not speed. It is pre-funding. Banks park liquidity in accounts around the world so that payments can be made before the covering funds arrive. That money is trapped, and trapped money has a cost measurable in basis points on a large balance.

If a stablecoin settlement arrangement lets a bank hold materially less of that standing liquidity, that is a genuine, repeatable, auditable saving — and unlike speed it cannot be faked with a well-chosen corridor. But it is also not automatic. You cannot settle in a token you do not hold, so pre-funding may simply move from one place to another, out of foreign bank accounts and into token balances or standby conversion lines. Whether this pilot moved the liquidity or merely relocated it is the single most interesting thing either firm could tell us, and neither has.

Who ends up holding the backing

There is a third question sitting underneath all of this. A stablecoin is a claim on somebody, backed by reserves that somebody holds. Whoever holds those reserves earns the return on them.

When a bank settles across accounts, that balance is a deposit inside the banking system. When it settles in a third party’s token, the backing sits with the issuer and with whoever banks the issuer. Over a pilot, this is a rounding error. At scale it is a transfer of economics out of deposit-taking and into issuance. Banks have been vocal about deposit substitution when the subject is retail stablecoins. The wholesale version is the same arithmetic with larger numbers and far less attention, and a bank running these trials should be able to say where it thinks the balance ends up.

The case for saying nothing

The strongest defence of the silence is straightforward. A pilot is a learning exercise, not a benchmark. Volumes are small, corridors are hand-picked, and publishing timings and costs from a handful of transactions invites readers to generalise from statistics that cannot bear it. Operational detail is commercially sensitive. The proper audience for the findings is the supervisor, not the newspaper.

All of that would be persuasive if the firms had said nothing. They did not. They published a benefit claim, and publishing a benefit claim takes on an evidentiary burden. If the small sample is too weak to support a number, it is too weak to support the adjective. And “transparent” is a peculiar virtue to assert while declining to show the figures that would demonstrate it.

Why this is a supervisory question already

The Bank of England’s Financial Policy Committee spent its September meeting warning that vulnerabilities across the system are more likely to crystallise together, and singling out operational and cyber resilience as an intensifying concern. Separately, the Bank’s enforcement committee has a remit that expressly covers financial market infrastructures and critical third parties — which tells you the perimeter already contemplates settlement arrangements operated by firms that are not ordinary banks.

A pilot runs on a bilateral contract between two willing parties who can pick up the phone when something goes wrong. Infrastructure cannot. Infrastructure needs settled answers on finality, on what happens when one leg completes and the other does not, and on who carries the loss in the gap. Those questions are much easier to ask now, while the volumes are trivial, than after a corridor has quietly become load-bearing.