De-banking decisions are commercial and should be data led

bank building corporate money assets grey tall city sky

Recent data from the Financial Ombudsman Service shows that the number of complaints about a bank account closure has risen in each of the three last financial years, hitting 2,708 complaints for 2022/23.

More significantly, it shows that the number of complaints for this year are already at 60% of that number, and are on track to hit a new high. Complaints about de-banking are seemingly going nowhere.

In part, this reflects the fact that de-banking is not new. Many customers have been de-banked or refused service for different reasons over the years, but the practice has been thrown into the spotlight recently given the high profile case of Nigel Farage and Coutts.

As the FCA investigation into bank account closures shows, it’s important to ask whether there are any prejudices at play. Recent work by the FCA is a good step to making this a data led debate rather than a knee jerk reaction to a small sample of cases.

However, outside true financial crime cases, the reality of firms de-banking or otherwise stopping service provision is often a commercial rather than a political decision.

Financial crime frameworks and PEPs

Taking a Politically Exposed Person (PEP) or other high-risk customer onto a bank’s books involves multiple layers of checks and due diligence as well as annual reviews and systems investment, all of which have a resource and capital cost.

In rare cases, this cost is passed on to the particular customer involved, but usually it is borne across all customers and increases the overall cost base of the firm. For this reason, it can create an unequal playing field between retail and established banks. For many challenger and digital banks PEPs and other high-risk clients are too expensive to onboard and maintain and they actively avoid providing services to them.

PEPs are not just rich politicians – there is also an issue at the other end of the scale, where a UK resident will be flagged as high risk because of the passport of their country of origin and could be denied service on that basis alone, including those with refugee status.

Some of these people would not be able to afford to pay for the due diligence required for a bank to take them on, and at the same time banks often do not want to deploy the capital to carry out the mandatory checks. Often the retail banks end up taking on these customers because it is a regulatory obligation to prevent harm to customers, but the cost means it is not a sound commercial decision.

A financial crime framework such as the PEP regime in its current form, can divert resources from real risks and therefore fail to deliver an effective control. Systems and processes that assess customer risk and the decisioning around them clearly need to be made more efficient and effective.

Efficient use of technology and AI

Technological advances such as AI mean that a more complex decision on risk can be made automatically.

But given AI is still in its relative infancy, a purely AI-based approach can result in prejudice being ‘programmed in’ or ‘big and simple’ decisions being made in an effort to streamline the process.

A more nuanced approach to Customer Risk Assessment involving human interaction powered by AI, has the potential to be both more efficient and effective as well as fairer.

This approach could save costs, reduce the anti-competitive angle between challenger and established banks, and create less of a disincentive to take on vulnerable customers who are being screened out because of their nationality.


First published by Investment Week on 16 January 2024.