Swift's alias plan makes cross-border payments feel domestic. They aren't

A phone number is not an account

A mobile number is a pointer. Its value depends entirely on who maintains the directory it points into, and who pays when it points somewhere wrong.

Swift — the cooperative whose messaging network most of the world’s banks use to instruct each other to move money across borders — says it will let people send money internationally using a mobile number or email address rather than account details, reaching the alias systems behind Spain’s Bizum, Australia’s PayID and Brazil’s Pix. Strip away the framing and the claim is narrow and genuinely useful: the addressing step gets easier. Nobody has ever enjoyed transcribing an international account number from a screenshot.

But addressing was never the expensive part of sending money abroad. The expensive parts are verification, recourse and price. Domestic alias schemes solved those inside a single perimeter: one currency, one rulebook, one regulator, one complaints scheme, one set of directory rules about who may claim which phone number. Cross a border and every one of those becomes two.

The name on the screen is doing most of the work

Ask an Australian why PayID feels safe and the answer is not the short address. It is that you type the number, a name comes back, and you check it. That returned name is the fraud control. It is why people press send.

So the first question for any international version is where that name comes from and whether anyone warrants it. If it is returned from a foreign directory that the sending bank cannot audit, has no contract over and does not stand behind, then a reassuring string on the screen is a warranty from nobody. Add transliteration between scripts, account nicknames, business trading names and the ordinary looseness of how people register aliases, and a sender in Melbourne is being invited to make a judgement on information whose provenance is invisible to them.

This is not hypothetical risk-mongering. It is the specific mechanism by which alias payments go wrong domestically, and the reason domestic schemes wrote rules about it.

Fast, final, and now easier to misdirect

The second question is loss allocation. An alias payment is quick, it feels casual, and once it lands in another jurisdiction, getting it back depends on the receiving institution’s law and goodwill rather than on the sender’s scheme rules.

Australia has spent years arguing about who should bear the loss when a customer is deceived into authorising a payment. Reasonable people disagree about the answer. What is not reasonable is answering it by omission — building a rail where the sender is the only party with no ability to verify anything and no forum to complain to that can compel the other end.

The price hides behind the convenience

Third, cost. A domestic alias transfer is effectively free. A cross-border one is not: there is a currency conversion, and a spread inside it, and fees that may be deducted along the way. Making an international payment look and feel like a domestic one invites the assumption that it costs like a domestic one.

If the experience is to be borrowed from the domestic system, the disclosure standard should be borrowed from the cross-border one: total amount debited, amount the recipient will receive, and the rate used, all shown before the sender confirms — not after, and not in a follow-up release.

The case against this argument

The strongest defence is that the alternative is worse. Today the realistic options for sending a few hundred dollars overseas are transcribing account details you may have been given incorrectly, or walking into a remittance shopfront and accepting whatever margin is on the board. An alias lookup sitting on top of existing bank rails is plainly better than both, and it is the incumbent banking system finally competing on convenience rather than complaining about those who do.

The second defence is that Swift is plumbing. It carries instructions; it does not hold the customer relationship. The protections come from the sending bank, the receiving scheme and each country’s law, and those are unchanged by making the address shorter.

Both points are fair, and I would rather this existed than not. But the second defence is exactly the claim that needs testing, because the protections do not travel. They stop at the border, while the interface tells the customer nothing has changed. When a payments network announces a consumer-facing experience, it takes on responsibility for what that experience implies — and this one implies domestic-grade certainty over an arrangement that does not have it.

What good would look like

None of this requires slowing the initiative down much. It requires writing three things down and publishing them: whether the returned name is warranted and by whom; how loss is split between sending and receiving institutions when an alias resolves wrongly or was registered by a fraudster; and what the sender must be shown about price before confirming.

Get those on paper and this is a straightforwardly good piece of financial plumbing. Leave them off and the first significant dispute will write the rule instead, in whichever jurisdiction happens to host it.