Instant delivery abroad quietly deletes the right to change your mind
Speed is not what makes sending money overseas expensive — the exchange rate is. And a payment that lands in seconds strips out the half hour in which American law lets a sender cancel.
FedNow is a service run by America's central bank, the Federal Reserve, that lets banks move money to one another in seconds, at any hour, seven days a week. It opened in 2023 and works only inside the United States. The Fed now says it wants the banks that use it to be able to send payments abroad through it as well. Sending money overseas today is slower and dearer than sending it at home.
Speed is not what makes sending money abroad expensive — the exchange rate is. And an instant, irreversible payment quietly removes the half hour in which American law lets a sender cancel. The Fed should settle both before it builds.
The thirty minutes that disappear
Under American law, someone who sends money overseas generally gets half an hour to change their mind and get it all back. That right survives only as long as the money has not arrived. Deliver the funds in three seconds and the protection is still on the statute book, still described in the disclosure the customer was handed, and completely unreachable.
That is the first thing the Federal Reserve has to answer now that it is, according to reports this week, pushing ahead with plans to let participants in its FedNow instant payments system make cross-border transfers. FedNow is the central bank’s own rail, live since 2023, moving money between American banks in seconds at any hour of any day. A FedNow payment is final on acceptance. There is no recall button, no overnight window in which an error can be unwound before it leaves the correspondent chain. Domestically that finality is a feature, and the fraud risk it creates is at least contained inside one legal system with one set of regulators.
The remittance rules written after the 2008 crisis assumed the opposite world. They assumed money sat somewhere for a while — in a correspondent account, in a batch file, in a Friday-night queue — and that the sender could catch it. The cancellation window is the single most useful remedy an ordinary person has when they realise, ten minutes later, that the builder they just paid is not a builder. Instant cross-border delivery does not repeal it. It simply makes it inoperative, and nobody has to vote on that.
What actually costs money
The second decision is about price, and it is the one the framing of “faster cross-border payments” tends to bury. For a retail customer sending money abroad, the headline transfer fee is rarely the expensive part. The expensive part is the gap between the exchange rate the customer is charged and the rate the institution can obtain wholesale. That spread is a price, it is not itemised, and it does not fall merely because the message travels faster.
If anything, instant settlement can widen it in thin corridors. Batching allows a bank to net offsetting flows and buy currency at a convenient moment. A payment that must be delivered on demand at three in the morning on a Sunday has to be funded then, in that currency, by somebody who will charge for the privilege. Speed converts a liquidity management problem into a per-payment cost. Someone pays it, and the customer will not see where.
American law already anticipates this for consumer remittances: the provider must disclose the exchange rate, the fees, and the amount the recipient will get. But that regime has exemptions for institutions that send few such transfers, and a great many banks have relied on being outside it. If FedNow becomes the cheap route for banks to enter cross-border retail payments at scale, the disclosure perimeter matters far more than it did when the same banks were sending a handful of wires a week.
The plumber is also the rule-writer
The Fed is unusual here in being both. It operates the system and writes the circular that governs what a payment through it means. That is precisely why “the Fed does not regulate consumer disclosure” is not an adequate answer. Where finality attaches on a cross-border leg is an operator’s choice, not a consumer bureau’s. So is whether a consumer remittance leg can be held briefly before release. So is whether the foreign end is a linked instant system, a correspondent, or a participant admitted directly — each of which puts the liability for a screened, blocked or misdirected payment in a different place.
The case for going ahead anyway
The strongest argument on the other side is that the current arrangement is worse. Correspondent banking traps prefunded liquidity in accounts around the world, charges for the credit risk of holding it, and takes days to tell anyone where a payment went. Faster settlement removes the credit exposure that justifies part of the margin, and over time competition should pass that through. And the cancellation right is already largely theoretical in the corridors people actually use: nonbank money transmitters deliver to mobile wallets in seconds today. On this view FedNow simply lets regulated banks compete with them, which should be good for price and for supervision.
Most of that is true. But “the protection has already been engineered away elsewhere” is an argument for restoring it, not for replicating the gap with central bank money and calling it modernisation. And the liquidity saving only reaches the customer if the customer can see the price. A rail that is free and instant while the spread is undisclosed is not a cheaper product; it is the same product with the margin moved somewhere harder to shop.
The Fed has time. There is no pilot yet, and no statutory deadline forcing the sequence. What it does not have is the option of deciding these things by default, in the technical annexes, after the first payment has gone out and not come back.
Before any pilot goes live, the Board of Governors should publish the amended FedNow operating circular and put it out for comment, answering three questions in plain terms: at what moment a cross-border payment becomes final and who bears the loss if it should not have been sent; whether sending institutions must display the amount the recipient will receive, in the recipient's currency, before the customer confirms; and what recall mechanism, if any, exists once the payment has left the United States.
The Consumer Financial Protection Bureau should say, in writing and before launch, whether instant delivery extinguishes the thirty-minute cancellation right for consumer remittances — and if it does, require a short release delay on consumer legs so that the right means something. A protection that only works when the system is slow is not a protection; it is a historical accident.
What would change this view
I am wrong if the Federal Reserve publishes an amended FedNow operating circular, or a Federal Register notice, that either limits cross-border transfers to institution-to-institution and business flows rather than consumer remittances, or builds in a short hold on consumer legs so the statutory cancellation window survives — and if that document also requires the sending institution to display the amount the recipient will receive before the payment is sent.
If that appears before any pilot, the design objection collapses and the remaining question is only price. I would also revise my view on cost if the Fed publishes corridor-level data after launch showing all-in spreads on cross-border FedNow transfers below comparable correspondent pricing.
Worked from
- FedNow to enable cross-border payments
Finextra · 2026-09-25