The CFD structure: synthetic exposure, broker as counterparty
A Contract for Difference is a bilateral contract between trader and broker. The contract pays the difference between entry and exit price multiplied by position size, on whatever underlying the contract references — an index, a share, a commodity, a currency pair.
There is no underlying ownership transfer. A long CFD on Apple shares does not produce share certificates, voting rights, or the ability to claim cash dividends in the conventional sense. The broker applies a dividend adjustment that mirrors the underlying when the underlying pays a dividend; this is a price adjustment to the CFD, not a cash credit.
The broker is the counterparty on every CFD position. When the trader profits, the broker loses, unless the broker has hedged the exposure in the underlying market. Most regulated brokers run hedging programs for aggregate client exposure; the trader's individual claim is contractual against the broker's balance sheet.
CFD margin and {{leverage}}
CFDs are leveraged. Position size scales beyond what cash on deposit could fund directly, with the broker holding Margin as collateral. ESMA caps the leverage on retail CFD positions: 1:30 for major FX, 1:20 for non-major FX and gold, 1:10 for non-gold commodities and major equity indices, 1:5 for individual equities and minor indices, 1:2 for cryptocurrencies.
Worked example on a major-index CFD. A long CFD on the US 500 index at a price of 5,200 with a contract size of 100 (broker-defined; some quote per-point):
Notional position value = 5,200 × 100 = $520,000. ESMA leverage cap = 1:10. Margin held = $520,000 ÷ 10 = $52,000.
Retail accounts trading indices typically use fractional contract sizes — 0.10 contracts on this index would hold $5,200 in margin and produce $10 per point of movement. Fractional sizing is the practical mechanism for opening index CFD positions on accounts that cannot hold full-contract margin.
CFD costs: spread, financing, and commission
CFDs carry the same Spread mechanic as forex: bid-ask gap paid at entry, again at exit. A typical major-index CFD quotes at 0.4–1.5 points spread; a major-share CFD at 0.05–0.20% of the price; a major commodity CFD at fractions of a cent per unit.
Overnight financing is a swap-equivalent calculated against the broker's reference rate (typically the relevant interbank rate plus a markup), differing by direction and instrument. A long CFD on a major index typically pays daily financing — long an index is being long the asset and short cash, which costs financing in a positive-rate environment. A short CFD on the same index typically receives a smaller positive amount.
Some brokers also charge a per-trade commission on CFDs, particularly on share CFDs. The commission is published in the contract specifications. The all-in cost per trade is spread + commission + average overnight financing for the holding period.
Counterparty risk and the limits of regulatory protection
The trader's claim on a CFD position is contractual against the broker, not against the underlying market. If the broker becomes insolvent, the trader's positions and unrealised P&L become claims against the broker's estate, not directly recoverable through the underlying exchange.
Regulated brokers in the EU, UK, and Australia operate under client money segregation rules — client funds are held separately from the broker's operational capital. In the event of insolvency, segregated client money is returned to clients before unsecured creditors are paid. The protection is significant but not unlimited: investor compensation schemes (the FSCS in the UK, equivalent schemes elsewhere) cover residual claims up to a per-client cap (typically £85,000 in the UK, similar floors elsewhere).
Counterparty risk does not exist when buying the underlying directly. A spot share purchase on a regulated exchange is held with a custodian; a futures position clears through a central counterparty. CFDs trade off a different risk model: ease of access, leverage, and short access at the cost of counterparty exposure.
What CFDs do, and what they do not
CFDs do offer: leveraged exposure to a wide range of underlyings without the operational complexity of futures or stock loan; fractional position sizes; the ability to short most underlyings without arranging stock borrow; access to instruments that retail traders may otherwise reach indirectly (commodities, indices, individual non-domestic shares).
CFDs do not offer: ownership of the underlying; access to corporate actions in the cash sense (dividend cash payment, share buyback participation, AGM voting); tax treatment equivalent to direct ownership in most jurisdictions; protection from broker counterparty risk beyond regulated client-money rules.
The choice between trading CFDs and trading the underlying directly is not universal. Day traders and short-term position traders typically prefer CFDs for leverage, ease of shorting, and unified-account access across asset classes. Long-term holders typically prefer direct ownership for the regulatory and tax structure. Neither is universally superior; they serve different timeframes and account structures.
The retail loss statistic and what it means
Every regulated CFD broker in the EU, UK, and Australia is legally required to publish the percentage of retail clients who lose money on CFDs. The figures cluster between 70% and 85% across brokers.
The statistic is a regulatory requirement, not a commercial claim. It reflects net P&L over a trailing period (typically the previous quarter or year). A loss-making client is one whose closed positions in the period netted negative; open positions are not included in the calculation.
The statistic does not say that 70–85% of trades are loss-making (the per-trade hit rate is typically closer to 40–55%). It says that the typical retail client's net P&L over a quarter is negative — usually because of cost layer (spread, financing, slippage) and occasional outsized losses from undersized stops or unprotected positions, not because the client picks the wrong direction more often than the right one. Reading the statistic as "7–8 in 10 trades fail" misreads what is being measured.