What is a CFD — contract for difference in forex and equities

A contract for difference, commonly abbreviated as CFD, is a derivative agreement that underpins roughly ninety per cent of European retail forex trading. When a trader says "I bought one lot of EUR/USD" through a regulated EU broker, they did not actually take ownership of one hundred thousand euro — they opened a contract with the broker, where the two parties settle only the difference between the opening and closing price of the underlying instrument. CFDs sit under ESMA oversight, carry hard leverage caps, include a built-in negative balance protection, and trigger specific tax treatment in each EU member state. Below I explain how the contract actually works.

What a CFD is in legal and economic terms

A contract for difference is a bilateral agreement between a retail client and a broker, in which both parties commit to exchanging the difference between the underlying instrument price at the moment the position opens and the moment it closes. If a client opens a long position on EUR/USD at 1.0850 and closes it at 1.0900, the broker pays the client fifty pips multiplied by the contract size — or pulls those funds from the deposit in the opposite scenario. At no point does an actual delivery of currency take place. That distinguishes a CFD from the interbank spot market, where two banks settle a physical exchange of currency at T+2.

Economically the effect for a retail client is almost indistinguishable from spot trading — the CFD price tracks the market quote, spreads and slippage behave the same way, and the profit-and-loss profile is identical. The difference shows up in three places: in the settlement mechanics (cash, not currency), in the legal status of the product (derivative, not the underlying), and in the scope of regulatory oversight (a separate framework for retail derivatives in the European Union, with a full ban in the United States under CFTC rules).

Long and short position mechanics

A trader opening a long position on one standard lot of EUR/USD at 1.0850 does not buy one hundred thousand euro for dollars. The broker simply registers a contract in the client book with an exposure of one hundred thousand euro, blocking around three thousand three hundred dollars as required margin under the one-to-thirty leverage cap. The rest of the contract value is funded synthetically on the broker side. If the price moves to 1.0900, each pip on a full lot is worth ten dollars, so closing the position generates five hundred dollars of realised profit credited to the account in the account's settlement currency.

A short CFD position is "natural" in a way short selling on a regulated stock market is not — the trader can sell EUR/USD without first borrowing one hundred thousand euro from a counterparty. The same mechanic matters even more on equity CFDs: shorting a single stock on a regulated exchange requires a stock loan, a borrow fee, and the entire prime brokerage infrastructure. A CFD on the same stock is just a contract on the price difference, so the broker can match both sides without involving stock loans at all — and that is the fundamental difference from an ETF, where the investor actually owns units of a fund; the full structural comparison lives in the piece on ETFs versus CFDs: owning the asset versus a contract on the difference. That is one of the main reasons CFD platforms became so popular with short-term traders — immediate access to the short side on every instrument, from an equity index to a commodity.

ESMA leverage caps and retail limits

Since August 2018 the European Securities and Markets Authority has enforced hard leverage caps for retail CFD clients. The maximum leverage on major currency pairs — EUR/USD, USD/JPY, GBP/USD and their close peers — is one to thirty. On minor pairs and major equity indices the cap drops to one to twenty. Commodities and less liquid indices sit at one to ten. Single-name equities are capped at one to five. Cryptocurrencies, where retail CFDs on them are even allowed in the relevant jurisdiction, run at one to two. Poland goes further than the EU baseline — the KNF banned the offering of cryptocurrency CFDs to retail clients altogether, regardless of the ESMA cap.

"Restrictions on CFDs are necessary because these products are complex and risky, and a majority of retail clients lose money on them. We are introducing leverage limits, negative balance protection and a standardised risk warning." — Steven Maijoor, Chair of the European Securities and Markets Authority, 2018

The standardised risk warning that every European CFD broker has to publish on its website is based on statistical data ESMA gathered across 2017 and 2018, when it pulled retail loss figures from more than one hundred brokers. The results were tightly clustered — depending on the broker, between seventy-four and eighty-nine per cent of retail accounts closed the reporting period at a net loss. That range is now the mandatory warning text in promotional materials, and the disclaimer cannot be removed or reduced in size.

Trading costs — spread, commission and swap

A CFD client pays in three places. The first is the spread, the gap between the bid and the ask. On EUR/USD an ECN account typically shows a spread between 0.1 and 0.17 pip, while a standard account with the broker mark-up runs between 0.8 and 1.2 pip. The second cost is commission — present mostly on ECN accounts, where the industry standard is three to seven dollars per one million in turnover per side. The third, the least visible to a new trader, is the swap point charged for holding the position overnight.

The swap point reflects the interest rate differential between the two currencies in the pair, adjusted by the broker margin. If euro rates sit below dollar rates, a long EUR/USD position generates a negative swap — the client pays a financing cost. In the opposite direction the client receives swap points. On equity CFDs the mechanic is analogous: a long position generates a financing cost calculated as the reference rate plus the broker mark-up, while a short position can produce a positive or negative carry depending on the stock borrow rate. A CFD does not give the client a legal right to a dividend from the underlying stock, but the broker still posts a so-called dividend adjustment — the long CFD receives a credit equivalent to the net dividend, while the short CFD receives a matching debit on the ex-dividend date.

Negative balance protection and the stop-out mechanism

The second pillar of the 2018 ESMA reform was a mandatory negative balance protection on retail CFD accounts. The direct trigger had landed three years earlier — on 15 January 2015 the Swiss National Bank scrapped the 1.20 floor on EUR/CHF and the franc appreciated by more than twenty per cent in a few minutes. Clients of many brokers ended up far below zero on their accounts, because at one-to-two-hundred leverage or higher the stop-loss queue could not clear. Some brokers went into liquidation — Alpari UK was wound up — and others tried to collect the residual debt from retail clients. Since 2018 such outcomes have been legally excluded within the European Union.

A modern European CFD account now has two layers of built-in protection. A margin call triggers when the equity drops to one hundred per cent of required margin — the broker pushes an alert to the client. A stop-out triggers at fifty per cent — the system automatically liquidates losing positions, starting with the largest one, until equity climbs back above the threshold. If equity still falls below zero despite the auto-liquidation, the broker writes off the deficit at its own cost. From the client's perspective this means a hard guarantee of never leaving a CFD account with a debt to the broker, even in the most extreme market moves. The broader regulatory layer is covered in the MiFID II regulation article.

Tax treatment of CFDs across the EU

Tax outcomes for CFDs vary by EU member state, but a few rules are common. Most jurisdictions classify CFD profit as capital gains rather than business or employment income, with a flat rate that does not blend with salary on the main tax return. The taxable base is the realised profit converted to the local currency, so trades closed on an account denominated in a foreign currency need conversion at the central-bank reference rate from the day before the trade. Open positions at year-end do not enter the calculation — only closed transactions count.

A locally licensed broker — XTB in Poland, IG in the UK, Saxo Bank in Denmark — usually issues a yearly tax statement with figures ready for the local return. A broker based in another EU country leaves the calculation to the client. The client has to pull the trade history, convert it to the local currency and complete the return manually. Trading costs — broker commissions and, depending on local interpretation, the spread — are deductible expenses. Losses can typically be carried forward against future trading gains, often for five years, but only within the same income category. The detailed walk-through for the UK case sits in the taxes and records section on forexmechanics.com.

Most common mistakes traders make on their first CFD account

1. Assuming that an equity CFD grants the right to a dividend in the classical sense — in reality the broker only books a net dividend adjustment, with no actual transfer of share ownership and no voting rights at the company's annual meeting.
2. Ignoring the swap cost on position-trading strategies — a trader holding a long EUR/USD contract for several months can hand back more in swap points than the move on the rate generated, if the interest rate differential moves against the position; I develop this point in the analysis of whether forex is suitable as a long-term investment.
3. Maxing out the one-to-thirty leverage cap on a single trade — formal compliance with the ESMA limit does not solve the underlying risk-of-ruin arithmetic, which I covered in the article on the one-to-thirty ESMA leverage cap.
4. Opening an account with an offshore broker to escape the leverage cap — that move pulls the client outside ESMA, but it also pulls them outside negative balance protection, segregated client funds and the complaints procedures of national regulators.
5. Skipping the annual tax return when trading with a non-domestic broker — the absence of a local tax statement does not remove the filing obligation, and tax authorities across the EU have increasingly good access to Common Reporting Standard data on cross-border accounts.

Your next step if you are opening your first CFD account

1. Verify the broker licence in the relevant EU regulator register. Only a broker authorised by a national EU regulator — KNF for Poland, CySEC for Cyprus, BaFin for Germany, the AMF for France, the CNMV for Spain — operates within ESMA leverage caps and negative balance protection. The broader walk-through on broker selection sits in the choosing a broker section on forexmechanics.com.
2. Choose the account type consciously — ECN or standard. An ECN account quotes a tighter spread but adds a commission per turnover; a standard account widens the spread and zeroes the commission. Short-term strategies often come out cheaper on ECN, longer-horizon ones simpler to track on the standard model. The full comparison sits in the article on spread versus commission.
3. Run a demo account for two to four weeks before going live. Demo lacks realistic slippage and the emotional pressure of a live account, but it lets you verify the order-entry mechanics, the behaviour of stop-loss orders and the way the broker calculates swap points. That window is now standard practice, and I recommend it to every client opening a first account.
4. Fund the account with capital you can realistically afford to lose. The ESMA statistic of seventy-four to eighty-nine per cent of accounts in net loss is not a rhetorical device — it is an empirically confirmed industry norm. Treat the first six to twelve months on a CFD account as educational capital rather than investment capital. After that period, the balance, the trading journal and the quality of the decision process decide whether scaling further makes sense, not the running balance on its own.