For years the UK debate over authorised push payment fraud turned on one unresolved question: who eats the loss when a customer is tricked into sending money to a criminal. The voluntary CRM Code left it to bank discretion, and discretion produced a lottery. On 7 October 2024 the Payment Systems Regulator ended the lottery. Mandatory reimbursement made the sending and receiving banks split the bill 50/50, up to £85,000 per claim, for in-scope Faster Payments and CHAPS scams.
The industry’s warning was loud and specific. Forcing banks to pay would blunt any incentive for customers to be careful, invite first-party fraud, and hand criminals a state-backed refund guarantee to farm. A year of data is now in. On 1 July 2026 the PSR published an independent evaluation by Frontier Economics. On the victim side, the verdict is not close.
What the year one numbers say
APP fraud losses over Faster Payments fell by roughly £73m a year, with nearly 35,000 fewer scam cases. Reimbursement across all reported claims rose from 54 to 65 percent; for claims actually in scope, firms now pay out 97 percent of the time. Frontier put the short-term net benefit at £17m to £29m after payment firms’ detection and reimbursement costs, and called that estimate conservative. The predicted collapse did not happen: the PSR found no evidence of market exits or of reckless consumer behaviour.
So the victim question is answered. The interesting question is the one the headline number cannot settle: is fraud actually being deterred, or is the cost simply moving?
Two mechanisms, one number
That reduction can come from two very different places, and they are not equivalent.
The first is genuine deterrence. Banks that now carry the loss have a hard financial reason to invest in detection, and the evaluation credits exactly that. The firms with the worst prior fraud volumes posted the largest gains. That is the incentive working as designed. Reimbursement was never mainly a consumer-protection tool. It was a mechanism to put the loss on the party best placed to stop it. On that logic it is doing its job.
The second mechanism is displacement, and this is where the number gets slippery. The £73m is a Faster Payments figure, in scope only. It excludes international payments, business transactions, and on-us transfers where the sending and receiving bank are the same firm. It says nothing about card fraud, crypto, or the purchase and investment scams that originate on social platforms and telecoms networks before any payment moves. The PSR itself flags this. Criminals are adapting, and the regulator is pressing tech firms and telcos to carry more of the prevention load. A fraud that never enters Faster Payments never shows up as a Faster Payments loss. Some share of that reduction is deterrence. Some share is water finding a lower channel.
The evaluation cannot cleanly separate the two, and it does not claim to. For a payments professional that distinction is the whole story. Durable deterrence compounds: better detection models, tighter onboarding at receiving banks, fewer mule accounts. Displacement does not compound. It relocates, and it relocates toward rails where no reimbursement backstop yet forces anyone to internalise the cost.
The cost is being reallocated in plain sight
There is a second reallocation happening openly. The scheme is a transfer from payment firms to victims, financed 50/50 across the payment chain. In year one the reduction in fraud outweighed that transfer, which is why the net benefit is positive. But the positive figure is fragile in a specific way. It holds only while detection investment keeps suppressing volume faster than criminals migrate. If displacement accelerates and in-scope volume plateaus, banks keep paying 97 percent on what remains while losing the falling-volume tailwind that made year one look cheap. The net benefit is a snapshot of a race, not a settled equilibrium.
The £85,000 cap matters here too. It was cut sharply from the regulator’s original proposal, which limits bank exposure on the largest, most sophisticated scams. That protects payment firm economics and preserves the incentive to keep serving customers, but it also means the highest-value victims, the ones losing life savings to investment fraud, sit partly outside the guarantee. The cap is the seam where consumer protection and PSP solvency were negotiated, and it is where the next fight will be.
What the first year actually proved
The UK has run the experiment other regulators were afraid to run, and the sky did not fall. That is the exportable result. Mandatory reimbursement can cut fraud and clear victims without breaking the banks, at least in year one. Expect this evaluation to be cited on both sides of the EU’s payments-fraud debate.
But the honest reading of the first year is narrower than the press release. The scheme has proven it can move loss onto the party that can prevent it, and that party responded. It has not yet proven it deters fraud in aggregate rather than steering it toward channels no one is yet required to pay for. The metric that will settle the deterrence question is not the Faster Payments line. It is total consumer fraud losses across every rail, and whether they fall or merely rearrange. Year two, and an evaluation that follows the money out of Faster Payments, will tell us which one we are looking at.
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.