An operations clerk pastes an IBAN from the wrong row of a spreadsheet, and a client withdrawal for 40,000 euro leaves the account towards a stranger. Twenty minutes later somebody notices. The instinct is to call the bank and ask them to cancel it. That call will not cancel anything, because a credit transfer is not held in escrow while it travels. It is a chain of book entries, and by the time the mistake surfaces the entry at the far end has usually already been made.
What actually exists is a recall: a second message, sent along the same rails, asking the receiving bank to return funds it has already credited to its own customer. The distinction matters because it changes who holds the decision. Your bank does not decide. The beneficiary bank does, and in most cases it needs its account holder to consent before it can debit the account again.
How a recall message travels
On the SWIFT side, the sending bank issues a cancellation request referencing the original payment. If the payment has not yet been executed at the correspondent, it can sometimes be stopped in flight. That window is short and depends on cut-off times, weekends and the number of intermediaries in the route, which is exactly why a transfer through two correspondents is harder to claw back than a direct one. The mechanics of that routing, and what each hop costs you, are covered in our breakdown of SWIFT fees.
Inside the euro area the process is more standardised. SEPA defines a recall procedure with named reasons: duplicate sending, technical problem, and fraudulent origin, plus a separate request for return of funds used for wrong beneficiary or wrong amount. There are deadlines attached, typically counted in banking days from settlement, and the receiving bank is expected to answer rather than ignore the request. Standardised does not mean automatic. It means there is a defined form for saying no. The difference between the two rail families is worth understanding before you pick one for client payouts, which is the subject of SEPA versus SWIFT.
Why the beneficiary bank holds all the cards
Once money is credited to an account, it belongs to the account holder under local law. A bank that takes it back without permission is exposing itself to a claim from its own customer. So the sequence is: the bank freezes what it can, contacts the account holder, and asks them to authorise the return. An honest recipient who received a payment by mistake will normally agree, and the funds come back within days minus a handling fee.
A dishonest recipient does not agree, and a mule account has usually been emptied before anybody calls. This is the uncomfortable part of fraud recovery. If the account balance has gone, a recall achieves nothing and the matter becomes a police report and possibly a civil claim in a foreign jurisdiction. Speed is the only variable you control.
A recall is a request, not an instruction. Nobody in the chain can promise a return, and any provider who tells a client that a wire "will be reversed" is setting up a complaint that will land back on your desk.
The first hour, in order
Firms that recover money tend to do the same things quickly, and firms that do not tend to spend the first hour deciding who owns the problem. A practical order looks like this.
- Confirm the exact payment reference, value date and amount from the outgoing statement, not from the ticket the client raised.
- Call the bank on the phone and follow with the written recall request the same day, naming the reason code the rail expects.
- If fraud is suspected, say so explicitly. A fraud reason unlocks handling that a simple wrong-beneficiary claim does not.
- Freeze anything connected: the client account, the payout batch and any repeat instruction that would send a second copy of the same error.
- Write the incident up while it is fresh, because a regulator reviewing your controls later will read this file. What that review looks like is described in our piece on audit trails.
Recalls that go the other way
Brokers and prop firms also sit on the receiving end. A deposit lands, the client trades, and three weeks later the sending bank asks for the money back because the payer claims the transfer was unauthorised. Now you are the beneficiary bank's customer being asked to consent, and the balance may already have been traded away or withdrawn.
This is where deposit controls earn their keep. Matching the sender name to the verified account holder is the single most effective filter, because most recall claims on trading deposits involve a third party paying for someone else. Holding first deposits from new bank accounts for a short settlement period is unpopular with clients and quietly saves firms a lot of money. The same logic sits behind name checks in identity verification and behind holding rules on withdrawals, which we cover in AML holds.
What a recall costs
Even a successful recall is not free. The sending bank charges an investigation fee, the beneficiary bank usually deducts a handling fee before returning, and any currency conversion is done twice at the spread of the day. A wire sent in the wrong currency and returned a week later can come back meaningfully lighter, and the client will ask you who pays the difference. Decide that policy in advance and write it into your terms rather than negotiating it under pressure.
The wider lesson for a trading firm is that credit transfers are a poor fit for high-volume retail payouts precisely because errors are expensive and slow to fix. Rails with built-in validation, account name checks and instant confirmation reduce the number of recalls you ever need to file. That trade-off between cost, speed and reversibility is the whole argument in comparing payout rails.
"People assume a bank can pull a wire back the way you cancel a card payment. It cannot. Once the funds land, the only lever anyone has is a polite request to a bank that has no obligation to say yes."
— Roman Onta, Executive Director, SINGUARD
Key Takeaways
- A wire is not reversible by the sender. A recall is a request to the receiving bank, which normally needs its customer's consent before returning funds.
- SEPA has defined recall reasons and deadlines; SWIFT cancellations depend on cut-off times and how many correspondents sit in the route.
- In fraud cases the balance is usually gone within hours, so the recall must be filed the same day and labelled as fraud, not as a wrong beneficiary.
- Firms receiving deposits should match sender name to verified account holder, which removes most of the recall claims that hit trading businesses.
Frequently Asked Questions
Can a bank cancel a wire transfer after it is sent?
Only in the narrow window before the payment is executed at the receiving end. After the funds are credited, the sending bank can issue a recall request, but the receiving bank decides whether to return the money and normally needs its own customer to agree.
How long does a wire recall take?
A cooperative return inside SEPA is often settled within days of the request. Cross-border SWIFT recalls that pass through correspondents can take weeks, and a case where the recipient refuses or the account is empty may never resolve at all.
Who pays the fees on a recalled transfer?
Both banks typically charge for the investigation and handling, and any currency conversion happens twice. The returned amount can therefore be smaller than the amount sent, so a trading firm should state in its terms who bears that shortfall.
About the Author
Roman Onta is an Executive Director at SINGUARD. He builds the Prop Firm CRM, the Broker CRM, Scalegram and CopySignals side by side with his brother Alex Onta, and he helped on the design of eTrader, the division Alex built and leads. His ground is worldwide payment processing, AML compliance and the corporate structures brokers are built on, work the two of them carry together, shaped by executive roles in the UAE and international corporates. He lives and works in Dubai for most of the year. Meet the executive duo leading Singuard's five divisions.