The Comfortable Estimate: Europe's Wealth Threshold Rewards the Incurious

On 18 August 2026 the Financial Conduct Authority (FCA) published its wealth management survey report, drawn from around 400 firms in a sector managing close to £1 trillion for more than 5.5 million retail clients. Around one in ten did not verify the source of a client's wealth at all, which is the whole of the obligation left undone. That will be read as a failure of diligence, and it is one. Yet it draws attention away from something the other nine firms in ten certainly do hold, and almost certainly do not test: a figure for what each client is worth, supplied by the client, entered at onboarding and refreshed by nobody in particular.

The figure has always been context rather than control. Industry guidance put it there; the Wolfsberg Group's 2020 questions and answers advise that an estimate of net worth should form part of a source of wealth assessment, while warning that the exercise "should not be seen as a documentary exercise". Context is a forgiving thing to hold. From 10 July 2027, in the European Union, that same figure decides how hard a firm must look at the customer who supplied it.

The number already exists

Nothing about the estimate is new. Private banks have carried a net worth figure for as long as they have carried a client file, because a relationship manager needs one to price a service and judge whether an instruction looks ordinary. It arrives with the client, is occasionally supported by a valuation or a public source, and is rarely challenged, because until now nothing turned on it. Singapore's regulator penalised three banks and an insurer in 2023 for that habit: establishing the wealth of higher-risk customers from the customers' own representations, without adequate corroboration.

The surrounding data suggests how lightly it is held. Twenty-six per cent of the firms the FCA surveyed did not collect expected transaction frequency and thirteen per cent did not record expected investment amounts, which are two of the inputs that would let a firm notice that a client's activity had outgrown the wealth on file. Positions reflect data up to 31 December 2024, so the picture may have improved. But a number that no control depends on is a number that no control tests, and that is about to stop being true.

Europe gives the number weight

Article 34(5) of the Anti-Money Laundering Regulation (AMLR), Regulation (EU) 2024/1624, prescribes specific enhanced measures for the wealthiest customers. They apply where a higher-risk relationship involves at least EUR 5 million handled through personalised services, for a customer holding total assets of at least EUR 50 million in financial, investable or real estate assets, excluding the private residence. Thresholds in this field have historically attached to transactions: EUR 10 000 for occasional transactions, EUR 3 000 for cash, EUR 2 000 for gambling. This one attaches to the size of the customer.

Recital 100 anticipates the objection, telling credit institutions, financial institutions and trust or company service providers that they "need not carry out or request a precise calculation" of total wealth, but should take measures to establish whether the customer sits above the line. The Authority for Anti-Money Laundering and Countering the Financing of Terrorism (AMLA) is to issue guidelines on making that determination by 10 July 2027, the day the AMLR applies. The Council and Parliament, announcing their provisional agreement in January 2024, said that failure to apply these measures would count as an aggravating factor in the sanctioning regime. Firms will build the assessment before they are told how.

The rule has a serious case

None of this is drafting carelessness, and the provision deserves a better hearing than it usually gets. A discussion paper on high-risk high-net-worth customers, circulated during the Council's negotiations, set out the reasoning plainly. Successive disclosures had shown institutions setting aside their own due diligence for very wealthy customers whose wealth should have raised questions, while customers of modest means carrying the same risk indicators were losing access to banking altogether. Judgement, left unstructured, had produced a regime that was strictest with those least able to argue back.

A threshold answers that, and more besides. It is supervisable, comparable across twenty-seven member states and capable of being reported, which a narrative judgement about plausibility is not. The UK's own numbers make the case for structure rather than against it. Yet the cure has a property its designers appear not to have priced: it hands the institution the job of deciding, in its own file, whether the customer is rich enough to warrant the harder questions.

Estimates travel in one direction

Consider what a firm gains from each answer. Source of wealth enquiry is owed in any higher-risk relationship under Article 34(4), so the threshold governs an additional layer, not the question itself. Yet the same incentive runs one level up, to the risk rating that opens the sequence, and the AMLR attaches a published number to it. Placing a client above the line brings further measures, senior attention, a heavier file and a slower onboarding; placing the same client below it does not, and the recital has confirmed that no precise calculation was required. A relationship assessed as below the threshold leaves nothing behind: no additional file, no senior approval, no recorded reasoning, and nothing for a supervisor to sample later.

The measurement problem points the same way. A fortune in listed securities, deposits and registered property can be counted; the same fortune held through a discretionary trust, a private fund interest, an unlisted company or self-custodied digital assets cannot. The FCA's findings of July 2026 show where those assets collect. Thirty-two per cent of firms active in private markets reported politically exposed persons among their customers, against nine per cent of firms that are not; eighty-five per cent of firms outside private markets reported no customers at all using complex ownership structures. Under a threshold keyed to visible assets, opacity reads as modest means.

The client can read the number too. Fifty million is published, and the tranche any one institution manages is a matter the client controls. The incentive is not to lie about a fortune, but to avoid having it counted in one place.

Documents flatter the number

What Europe will accept as evidence compounds the difficulty. AMLA consulted until 8 May 2026 on the technical standards under Article 28(1), and was required to submit them to the Commission by 10 July 2026. Article 27 of that draft lists what evidence of source of funds and wealth may consist of: tax declarations and pay slips, audited accounts, loan agreements, land registry extracts, inheritance and sale documentation, asset registers and reputable media. Once adopted, the standard applies alongside the AMLR from 10 July 2027.

The list turns back on itself. These are the documents from which an institution would build its estimate of a client's wealth, and they are also the documents the estimate determines whether it must obtain. A file assembled from valuations produces a valuation, because almost every item on the list evidences a holding, a transfer or a price rather than the activity behind it. The question was never what a customer holds, but how they came to hold it.

Singapore governed the estimate

One regulator has already worked through this, and did so after being embarrassed rather than before. After the money laundering case of August 2023, the Monetary Authority of Singapore (MAS) issued Circular AMLD 08/2024 on establishing customers' sources of wealth, setting out three principles: materiality, prudence and relevance. It assumes firms will estimate. The industry paper implementing it, published in May 2025 by the AML/CFT Industry Partnership whose secretariat includes MAS, treats those principles as a cascade. Understand the entire body of wealth, corroborate what is material or higher risk, and where evidence is absent use benchmarks that are reasonable, documented and reviewed. "Benchmarks alone should not be mistaken for corroboration", it adds.

Two instructions in that paper answer what the AMLR leaves open. Where corroboration is triggered, the customer's total wealth is to be considered, not only the portion of assets brought to that bank; and obtaining documents is not the end of the exercise, because documents must be assessed for plausibility and for red flags. Enforcement has run on the same logic.

MAS imposed composition penalties totalling S$27.45 million on nine financial institutions on 4 July 2025. Its source of wealth finding was not that documents were missing: all nine failed to detect or adequately follow up significant discrepancies and red flags in the information they already held, and in some cases significant aspects of wealth were never corroborated. The difference is not severity, but sequence. Singapore governed the estimate; Europe has made it decisive and left the method to guidance that arrives as the duty begins.

Firms should audit the negatives

The UK has moved the other way, which sharpens the problem rather than removing it. The Money Laundering and Terrorist Financing (Amendment) Regulations 2026, in force from 30 June, narrow mandatory enhanced due diligence to unusually complex transactions and to jurisdictions subject to a Financial Action Task Force (FATF) call to action, on HM Treasury's reasoning that the earlier drafting encouraged risk-averse behaviour. In both regimes, more now rests on what the firm decides for itself and on whether that decision leaves a trace.

Three things follow. Treat the net worth figure as a control rather than a field, with a named owner, a recorded basis, a refresh trigger and documented assumptions; a client's declaration is an input to that assessment, not the conclusion of it. Sample the negatives, because relationships assessed as falling below the threshold are the ones nobody will look at again, and they are where structured wealth collects. Record the portion of a client's source of wealth that could not be corroborated and what followed, which both the Wolfsberg guidance and the MAS circular anticipate. None of this is expensive. But it has to be done before the file is challenged, because afterwards it is not evidence; it is reconstruction.

Comfort is not a control

Every incentive in the new arrangement points at the same answer, and it is the quiet one. A firm that estimates low buys a shorter file, a faster onboarding and a defensible position, since the rules never asked for precision; a firm that estimates high buys work alone. Supervisors will eventually test the population above the line, because it is countable and reportable, and the interesting population will be the one nobody was ever asked to justify. Yet incuriosity has never survived contact with a case that goes wrong, and the file that records a comfortable number without recording why it was believed will read, at that point, exactly like a decision not to ask. The threshold will be met. The question is whether anyone tested the number that decided it did not apply.

Does your firm know how many of its clients sit just below the £50m mark, and could it show what that judgement was based on?

If those files rest on a declaration made at onboarding and never revisited, we can help you test the basis and evidence the reasoning while the deadline is still ahead; contact us.

At OpusDatum, we test the assessments firms are least likely to revisit, including the relationships judged to fall below a threshold and the wealth figures nobody has challenged since onboarding. Our focus is control effectiveness rather than control existence, and the evidence a supervisor will want when it asks how a client was placed on the lighter side of a line.