Every regulated broker must publish how many of its customers lose money. Read fourteen of those disclosures beside the same firms’ filed accounts, as one analysis did in April 2026, and the arithmetic becomes uncomfortable.
There is a figure in the British trading industry that nobody advertises and everybody is required to publish. Since 2019, every regulated provider of leveraged products has had to display the percentage of its retail customers who lose money. In the April 2026 sample discussed below, every one of the fourteen firms reported a majority of its retail accounts losing money, and the worst of them reported more than four in five.
Set that against the same companies’ filed accounts and you can work out, roughly but reliably, where the money goes.
What do the numbers say?
An April 2026 analysis by The Investors Centre, worked from the published loss disclosures and the filed accounts of fourteen FCA-authorised UK brokers, puts annual UK retail CFD losses at roughly one billion pounds. That is a calculation rather than a total the regulator collects: no official body publishes an industry-wide loss figure, which is precisely why somebody had to build one. Beside it sits a firm-level number of a completely different kind. CMC Markets reported net operating income of £4,685 per active UK retail client in its FY25 accounts, which is one company in one financial year and an average of nothing wider.
The second of those is audited. The first is arithmetic performed on audited inputs by somebody outside the industry, and the distance between those two things is worth a section of its own.
Four ways the billion-pound estimate could be wrong
Fourteen brokers is not the market. FCA figures record 74 firms holding CFD retail permission as at December 2025, so a fourteen-firm sample covers the large end and models the rest. The fourteen do not share a financial year, so any annual total welds together periods that do not line up.
Active client is a term each firm defines for its own reporting, and the definitions have moved. And a loss disclosure counts accounts that finished a window down, which is a different quantity from money lost.
Any of those could push the figure in either direction, which is the honest position to take on it. What an estimate of this kind is good for is scale. It establishes that the sums leaving British retail accounts run to a national-level number rather than a hobbyist one. It will not settle an argument to the nearest hundred million, and anyone quoting it to two decimal places has not read how it was assembled.
If commissions are free, where does the revenue come from?
Four places, none of which feature prominently in a television advert: the spread, the daily financing charge on anything leveraged and held open, the currency conversion applied when you buy something priced in dollars, and platform or data subscriptions. What the list conceals is that those four behave in completely different ways depending on what the market happens to be doing while you are in it.
| Cost line |
On a quiet session |
When the market is moving |
| Spread |
Tightest, and the version that gets advertised |
Widens, sometimes sharply, exactly when the volume arrives |
| Overnight financing |
Accrues nightly at the same rate |
Accrues nightly at the same rate, on a position you now care about |
| Currency conversion |
Applied inside the exchange rate |
Applied inside a rate that is itself moving |
| Slippage between click and fill |
Small enough to ignore |
The gap that decides whether the quoted spread was the price you paid |
| Platform and data fees |
Fixed |
Fixed, and the only line here you can budget for in advance |
Five revenue lines and how each behaves either side of a busy market. Direction is the whole of what this shows: none of these movements is quoted in advance, and any figure in the cells would have to be lifted off a real statement.
Not untrue, but incomplete in a way that flatters. Removing the commission genuinely removed a cost. It also moved the revenue to lines that very few customers ever total up. The word free is doing a great deal of work in that sentence. The regulator has taken an increasing interest in exactly this, which is why the risk warnings and loss disclosures exist in the form they do.
What should you do differently if you trade anyway?
Two things, both dull and both effective. Read the loss disclosure for the specific firm before funding an account, not to rank one provider against another, but to confirm that at no authorised firm do most clients win. Then price your own trading pattern rather than the advertised example: somebody buying US shares monthly should care far more about the currency conversion charge than about the headline commission rate.
Why did the industry drop commissions at all?
Competition, mostly, plus a shift in what customers noticed. Commission was the most visible cost in trading – a clear line item, easy to compare across providers, and so it became the number everybody advertised against. Once one large platform removed it, the rest followed or lost accounts.
What did not change was the need to earn revenue from the same customers. What changed was that the replacement lines cannot be laid side by side. Two spreads quoted at different moments are not comparable numbers. A financing rate expressed as a benchmark plus a percentage is not comparable with one expressed in pounds per lot. Competition carried on perfectly happily. It simply moved onto a measure nobody can rank.
Consumer pricing does this everywhere, and calling it fraud would be lazy. It is a design decision about which of a customer’s costs are easy to hold up against a rival’s.
Who is on the other side of these trades?
For contracts for difference, frequently the broker itself. Many CFD providers act as counterparty rather than passing trades to an external market, which means a customer’s loss can be the firm’s gain directly. Providers hedge and manage that exposure in various ways, and the arrangement is disclosed, but a great many retail traders do not realise it is the arrangement. Worth understanding before opening a leveraged account, mainly because it explains why the loss disclosure figures look the way they do.
What would actually reduce the billion-pound figure?
Nothing the industry can do unilaterally, and probably nothing dramatic. Leverage produces those outcomes by design: it amplifies both directions, and the cost of holding leveraged positions accrues daily regardless of which way the market goes. The regulator has already capped retail leverage and mandated the risk warnings, which reduced the harm without eliminating it.
At an individual level the levers are duller and more effective. Trade less often. Hold for shorter periods when leveraged. And choose a venue where the cost per trade does not quietly consume the difference between a decent decision and a bad outcome.
Does the platform choice genuinely change the outcome?
On the available evidence, yes, and by more than most strategy adjustments. Compare CFD venues the way the team at The Investors Centre does, on the costs a funded account actually paid rather than on published pricing, and the gap shows up immediately: the distance between the cheapest and the dearest venue for an identical pattern of trading is wide enough to swallow a mediocre strategy’s entire edge.
None of which is a reason to trade. The loss disclosures remain what they are and no amount of platform optimisation alters the arithmetic of leverage. The narrower point is that if the decision to trade has already been taken, the boring administrative choice deserves at least as much attention as the exciting one.
…
