What Is a CFD? The Trade Where You Never Own Anything

Two people wake up convinced the same company is going up. The first buys 100 shares through a normal brokerage account: the money leaves, the shares arrive, her name goes on the register. The second opens a CFD on the same company, same size, same morning. By the close the stock is up 3% and both of them are up almost exactly the same amount of money. They high-five. They have had, by every measure that shows up on a screen, an identical day.

Only one of them owns anything. The first has 100 shares that will sit there whatever happens next — through a broker collapse, a dividend, a takeover vote, a decade. The second has a contract with a company, and the value of that contract depends entirely on the company honouring it. That difference is invisible on a good day and it is the whole subject on a bad one.

What a CFD actually is

Strip the acronym and it is startlingly literal: a contract, for the difference. You and the broker agree on an opening price. Later you agree on a closing price. Whoever is on the wrong side of the gap pays the other the difference, multiplied by the size of the position. That's the product. There is no share, no barrel of oil, no gold, and no delivery of anything to anyone — the asset exists in the arrangement purely as a reference price, in the same way a bet on a horse doesn't involve you taking the horse home.

This is also why going short is no harder than going long. Short-selling a real share is a genuine faff — the shares have to be borrowed from someone who owns them, with a fee and a recall risk attached. A CFD has no such problem, because nothing needs to be borrowed: you're simply agreeing to settle a difference in the other direction. That symmetry is a real advantage, and it's the honest half of the CFD sales pitch.

The instrument came out of London's equity-derivatives desks in the mid-1990s, essentially an equity swap traded on margin, and it wasn't built for retail traders at all — it was a tool for taking positions in takeover battles without showing up on the share register. The earliest hard record is a UK Takeover Panel ruling that cleared their use in a bid for a regional electricity company in November 1994. Ordinary traders got access at the tail end of that decade, when a London broker started offering them to private clients.

You'll find a tidier origin story than that on most broker websites — a named pair of bankers, a named bank, a named deal — and it's worth knowing that the tidy version doesn't survive contact with the dates. The bank it usually credits didn't exist under that name until 1998, four years after the Panel was already ruling on these things; a competing version names an entirely different firm; and the motive it cites appears to have been retrofitted decades later. It's a small lesson worth carrying into everything else you read about this product: a great deal of it is repetition rather than record.

Who is actually on the other side

This is the part that gets skipped, and it's the part that explains everything else. When you buy a share, your broker is a messenger: it routes your order to an exchange, the exchange matches you with another trader who wanted to sell, and a clearing house stands between you to guarantee that both sides deliver. The broker's job is to carry the order and take a fee. It has no stake in whether you were right.

A CFD does not work like that. There is no exchange to route to, because CFDs aren't listed anywhere — they are over-the-counter contracts, written between you and the firm. Your order doesn't travel through the broker to a market. For your trade, the broker is the market. It is your counterparty, and the position exists only as a line in its book.

Two consequences follow, and it's worth being precise about both, because the sloppy version of this argument is everywhere and it's wrong.

The first is counterparty risk, and it is unambiguous. Your profit is a claim on the broker. Client money rules and segregated accounts exist to protect the cash you deposited, and in many jurisdictions a compensation scheme covers some of it — but a share you own is your property in a way a contract with a firm simply is not. If the firm fails, a shareholder still holds shares. A CFD client joins a queue.

The second is the conflict of interest, which is real but frequently overstated. When a broker takes the other side of your position and leaves it there — the market-maker model — your loss is mechanically its gain, and that is an uncomfortable arrangement no matter how well-regulated the firm is. But many brokers hedge their net exposure in the underlying market instead, earning from spread and financing rather than from your losses, and for that flow the conflict genuinely isn't there. The honest version isn't "your broker wants you to lose." It's that you usually cannot tell from the outside which model you're trading against, and the arrangement doesn't require you to be able to.

When you buy a share, your broker carries your order to a market. When you open a CFD, your broker is the market — and the difference only becomes visible on the day you'd least like to discover it.

The number they have to print on their own advertising

Here is an unusual thing about this corner of finance. In the UK and across the EU, a firm advertising CFDs to retail customers must display, on that advertising, the percentage of its own retail client accounts that lose money trading these products. Not an industry average, not a footnote in a PDF — its own figure, on its own marketing. Regulators imposed it after concluding that consumers were systematically misjudging the odds, and the resulting disclosures have been remarkably consistent ever since: for most major brokers, somewhere in the region of seven in ten retail accounts finish behind.

It's worth sitting with how strange that is. This is an industry legally obliged to publish its own batting average next to the advert, and the number has not moved much in years. Whatever else a CFD is, it is not a product with a hidden success rate — the success rate is printed on the tin, and people buy it anyway.

What you give up by not owning it

Ownership turns out to carry a lot of quiet cargo, all of which you leave behind. You get no vote, which matters more than it sounds when a takeover or a board fight arrives. You get no share certificate and no place on the register. And dividends arrive in a peculiar half-form: rather than being paid one, a long CFD holder typically receives a cash adjustment that mimics it — the sensation of a dividend, without the dividend. Hold a short position through the same date and the adjustment is debited from you instead, which surprises people exactly once.

What it costs to hold

CFDs are usually sold as cheap, and for a day trade that's fair. The costs are three: the spread between the buy and sell price, paid the moment you open; a commission on some markets, share CFDs especially; and overnight financing, charged every night the position stays open. The first two are small and obvious. The third is the one that decides outcomes, and it is the one nobody reads about before opening an account.

Financing is charged because you're holding a position larger than your deposit — the broker is funding the rest, and it charges for that, typically at a benchmark overnight interest rate plus a markup of two or three percent. The crucial detail is the base it's applied to. The charge is levied on the full position, not on the margin you put up. Trade at 5× and the nightly fee is calculated on five times your own money; the percentage looks modest against the position and lands five times harder against your account. This is why CFDs behave so differently over a day and over a quarter. Drag the clock below and watch a genuinely good trade change hands.

Note what the widget never touches: the trade itself. The market moves your way by the same 4% in every scenario. Your analysis was right the entire time — the holding period does all the damage. That is the single most useful thing to understand about this instrument: a CFD is a short-dated tool wearing a long-term tool's clothes, and the costs are structured so that patience, which is a virtue nearly everywhere else in investing, is here a slow leak.

Leverage, on top of all this

CFDs are leveraged by default: you post a margin deposit and control a much larger position, so both gains and losses are multiplied, and a move against you of 100 ÷ your leverage percent takes the deposit to zero. That machinery isn't unique to CFDs and we've given it its own article rather than compressing it here — if the words margin call and liquidation don't yet make you flinch, read that one first. What CFDs add to it is the structure above: an OTC contract, a broker counterparty, and a nightly charge on the full notional. Leverage decides how fast; the structure decides who you're facing while it happens.

Where you can trade them, and where you can't

CFDs are legal and widely traded across the UK, Europe, Australia and much of Asia and Latin America — and unavailable to ordinary traders in the United States. That split confuses people who assume a product shut out of the world's largest financial market must be exotic. It isn't, and the popular explanation is wrong in a way worth correcting: no American law bans CFDs by name, and it isn't really the SEC's doing either.

What US law actually says is that a leveraged derivative like this must be traded on a registered exchange when the customer is an ordinary retail client — a rule that lives mostly with the commodities regulator rather than the securities one. A CFD is over-the-counter by construction, so it fails that test and can't be offered. The tell that it's a venue rule rather than a safety judgement is who's exempt: an individual with more than $10 million in investments counts as an "eligible contract participant" and may trade them perfectly legally. The product isn't deemed too dangerous. You're deemed too small.

Everywhere they are permitted, they've been steadily fenced in. The UK, the EU's national regulators and Australia have landed on a similar package: leverage capped by asset class, a margin close-out rule, negative balance protection so a retail account can't be pushed below zero, and the risk-warning disclosure. The caps scale to volatility, which is a neat piece of design — the steadier the market, the more leverage you're allowed. It is not, however, one harmonised rulebook, and the details diverge more than the summaries suggest.

The Australian experience is the most instructive, because the regulator published the before and after. When the caps came in, aggregate quarterly retail losses across the sector fell by around 90%, margin close-outs dropped sharply, and instances of accounts going negative all but disappeared. The rules worked, in other words — and even so, the regulator's most recent review found around two-thirds of retail CFD investors still lost money over the year. That is the state of the art in consumer protection for this product: a dramatic improvement, landing at a majority of customers losing.

So who are they actually for?

Not for building wealth, and the structure says so plainly. Every feature of a CFD — the financing clock, the leverage, the absence of the asset — is optimised for short holding periods and hostile to long ones. If your plan is to own good businesses for a decade, the CFD is the wrong shape of tool in every dimension, and the version of that plan that uses one is just the same plan with a meter running.

Where they earn their place is narrower and less glamorous: short-dated directional trades, easy access to markets that are otherwise awkward to reach, and hedging an existing portfolio without selling it. Those are real uses, and the people doing them well share a trait — they know their holding period before they open the position, and they size it so being wrong is survivable. The instrument doesn't supply either of those. You do.

If you want to understand how one of these actually behaves, the sane order is: learn the mechanics, then feel them somewhere the mistakes are free, then decide whether you want the thing at all. Plenty of people work through that sequence and conclude they'd rather just own the shares. That is a perfectly good outcome, and considerably cheaper than the alternative route to the same conclusion.