If you have ever opened the homepage of a UK or EU CFD broker, you will have seen, in small text near the bottom, a sentence that reads something like this: "CFDs are complex instruments and come with a high risk of losing money rapidly due to leverage. 76% of retail investor accounts lose money when trading CFDs with this provider. You should consider whether you understand how CFDs work and whether you can afford to take the high risk of losing your money."
The exact percentage varies by provider. Across the industry, the published numbers cluster between roughly 65% and 85%, with most major providers landing in the mid-70s. The number is not a marketing flourish. It is a regulatory requirement, the methodology behind it is precisely defined, and what it tells you about CFD trading is, on close inspection, even less encouraging than it appears at first glance.
This piece walks through where the number comes from, what it includes and excludes, and what an honest look at the underlying data implies for anyone considering whether a CFD account is the right product for them. We are not making the argument that no one should ever trade a CFD. We are making the argument that the customer who reads the warning, understands what it actually says, and decides to proceed has made a more informed decision than the customer who skims past it.
Where the number comes from
In May 2018 the European Securities and Markets Authority adopted a temporary product intervention measure under Article 40 of MiFIR, restricting the sale of CFDs to retail clients across the European Union. The measure, set out in ESMA Decision (EU) 2018/796 and successors, imposed five specific requirements:
- Leverage limits ranging from 30 to 1 (for major currency pairs) down to 2 to 1 (for cryptocurrencies).
- A margin close-out rule requiring the provider to close out all open positions when account equity falls to 50% of the initial required margin.
- Negative balance protection on a per-account basis, so that a retail customer cannot lose more than the funds in the account.
- A prohibition on monetary and certain non-monetary benefits offered to retail customers as inducements to trade.
- A standardised risk warning, including the loss-percentage number, to be displayed on all marketing communications.
ESMA's temporary measures were converted into permanent rules at the national level by EU member states and (post-Brexit) by the UK. The UK's permanent rules are in FCA Policy Statement PS19/18, published in June 2019, with the relevant Handbook provisions sitting in COBS 22.5. The substantive rules are essentially unchanged from the original ESMA measures.
The methodology for the loss-percentage number is set out in detail in the FCA Handbook and the ESMA decisions. It is the percentage of retail clients who held an open CFD position with the provider during the most recent twelve months and whose total realised and unrealised profit and loss across all CFD positions during that period was negative. The calculation is performed by the provider, applies only to its own retail clients, and must be updated quarterly.
What the number includes and excludes
The methodology has several features that are worth understanding precisely.
First, it counts accounts, not customers. A customer with two accounts at the same provider (for example, one in their own name and one in a corporate vehicle) counts as two separate accounts. This is a relatively minor wrinkle.
Second, it counts retail accounts only. Customers who have been categorised as professional clients, either by virtue of their portfolio size and trading experience or by self-selection, are excluded from the calculation. Professional clients also receive substantially higher leverage limits and weaker protection. Some providers actively encourage qualifying customers to opt up to professional status, partly because this removes them from the loss-percentage calculation and partly because it removes them from the leverage caps.
Third, it counts realised and unrealised P&L combined. An account that has open profitable positions on the snapshot date but closed losing positions during the year is treated according to the net of the two. An account that has open losing positions on the snapshot date is counted as a loser, even if those positions might recover.
Fourth, it counts any negative net P&L, not just large losses. An account that lost £50 over the twelve-month period is counted exactly the same as an account that lost £50,000. The 76% figure tells you nothing about the distribution of losses among the losing accounts.
What the underlying distribution looks like
Several regulators and academic studies have looked at the distribution of P&L among retail CFD accounts in detail. The picture they paint is consistent and worth describing.
A 2021 study by the Belgian Financial Services and Markets Authority (FSMA), drawing on detailed account-level data from licensed Belgian CFD providers, found that the typical losing account lost a meaningful fraction of the funds initially deposited, while the typical winning account won a much smaller fraction. The distribution was strongly negatively skewed: a small number of accounts lost very large amounts, while the winning accounts mostly recorded modest gains. The mean P&L per account was substantially negative, even though the median P&L was less negative than the mean, because the worst-performing accounts dragged the average down disproportionately.
The FCA's PS19/18 cost-benefit analysis presented similar conclusions for the UK market: the average UK retail CFD customer lost approximately £2,200 per year before the intervention, and the regulator estimated that the new rules would reduce average annual losses by approximately £750 per customer. The estimate of average losses was net across both winners and losers, and the regulator was careful to note that the losses were concentrated among a relatively small subset of customers who lost very large sums.
Two implications follow from this. First, the 76% number understates the asymmetry between winners and losers. The 24% who win do not, on average, win nearly as much as the 76% who lose, lose. Second, even within the losing group, the distribution is wide: most losing customers lose a manageable amount, but a small minority lose amounts that are life-changing. The regulator's rules (margin close-out at 50%, negative balance protection) are explicitly designed to truncate the worst tail, and the evidence is that they have done so.
What the number does and does not tell a prospective customer
The number is a population-level statistic, derived from the combined experience of tens of thousands of retail accounts at the specific provider over the most recent year. It is not a prediction for any individual customer. A customer who is sceptical that the population statistic applies to them ("I am better educated, more disciplined, less impulsive than the average user") is making an empirical claim. The population data does not contradict the claim, but it also does not support it. Among the customers who lost money in the past twelve months, a substantial fraction would, if asked beforehand, have predicted they would be in the winning 24%.
What the number tells a prospective customer reliably is: at this provider, with these products, under the current regulatory framework, three times more retail accounts lose money over a one-year period than win. That ratio has been remarkably stable across providers, jurisdictions and time periods since the data has been collected. Retail CFD trading is a negative-expected-value activity for the typical retail user, and the regulatory disclosure is designed to make that fact unavoidable in the customer's decision-making process.
The honest summary
CFDs are a legitimate financial product with legitimate use cases (short-term hedging by sophisticated users, leveraged exposure where the user fully understands the leverage, exposure to instruments not easily accessible through other formats). The regulatory framework that has applied since 2018 has materially reduced the worst customer outcomes. The leverage caps work, the margin close-out rule works, and the negative balance protection rule has eliminated the historical horror stories of customers owing hundreds of thousands of pounds to their broker after a single bad day.
What the framework cannot do is convert CFD trading from a negative-expected-value activity for retail users into a positive-expected-value one. The published loss percentages, taken seriously, are the regulator's way of conveying this in plain language. A prospective customer who has read this far, has thought about the underlying distribution rather than just the headline percentage, and still wants to open an account is welcome to do so. They have made the decision with the same data the regulator has, which is the most that any regulatory regime can reasonably ask of an adult.