What Is Value at Risk? VaR Explained for CFD Traders

Value at Risk (VaR) is a statistical measure used to estimate a portfolio's potential loss threshold over a defined period and at a chosen confidence level. It helps traders understand how much a portfolio could lose under the assumptions built into the risk model.

For traders using leveraged products such as Contracts for Difference (CFDs), looking only at individual stop-loss distances may not provide a complete picture of portfolio risk. Market gaps, volatility and slippage can affect several positions at the same time. VaR can help estimate combined portfolio risk, although it should not be treated as a worst-case loss measure.

• The value at risk meaning, in short: it estimates a loss threshold over a specified timeframe and confidence level, based on the assumptions and data used by the model.

• A VaR figure needs three elements to be meaningful: a time horizon, a confidence level and a loss amount or percentage.

• VaR does not show the maximum possible loss. Losses beyond the VaR threshold can be substantially larger.

• VaR is more useful when combined with other risk tools, such as stress testing and Expected Shortfall.

What Does Value at Risk Mean in Trading?

So, what is Value at Risk? In simple terms, it measures how much a portfolio could lose over a defined period at a specified confidence level.

Instead of asking how much a portfolio might lose in the most extreme possible scenario, VaR answers a narrower question: What loss threshold is the portfolio expected not to exceed at a given confidence level, based on the model being used?

A VaR figure has three main components:

• Time horizon: The period over which risk is measured, such as one day, 10 days or one month. Short-term traders may focus on a one-day VaR.
• Confidence level: The probability, according to the model, that the portfolio loss will remain within the VaR threshold. Common levels include 95% and 99%.
• Loss threshold: The estimated monetary amount or percentage that the portfolio is not expected to lose more than within that confidence level.

For example, suppose a trading account has a one-day 95% VaR of $500. Under the model's assumptions, this means there is a 95% probability that the portfolio will not lose more than $500 over one trading day.

It also implies a 5% modelled probability that the loss could exceed that level. This is sometimes described as approximately one day in 20, but it should not be interpreted as a prediction that a breach will occur exactly once every 20 trading days.

In institutional risk management, trading desks commonly operate within formal risk limits and reporting frameworks. Depending on the firm's policies, a breach of an internal risk limit may lead to escalation, hedging, reduced exposure or other risk-management action. Basel standards also require banks using internal models to maintain defined trading limits and produce regular desk-level risk reports.

The 3 Main Methods to Calculate Value at Risk

There are three widely used approaches to estimating VaR: historical simulation, the variance-covariance method and Monte Carlo simulation. Each uses a different way of estimating the distribution of potential portfolio returns.

These three approaches are widely recognised methods of estimating VaR.

1. Historical Simulation

Historical simulation applies past market movements to the current portfolio and ranks the resulting gains and losses from worst to best.

Suppose a trader uses 500 daily observations to estimate a one-day VaR at a 95% confidence level. The worst 5% of outcomes represent approximately 25 observations.

The VaR threshold would fall around the boundary between this worst-performing 5% and the remaining 95% of observations. The exact observation used can vary slightly depending on the statistical convention used to calculate the percentile.

One advantage of historical simulation is that it does not require returns to follow a normal distribution. However, its usefulness depends heavily on whether the historical sample provides a reasonable representation of future market conditions.

2. Variance-Covariance (Parametric) Method

The variance-covariance method estimates VaR using statistical measures such as expected portfolio return, standard deviation and correlations between assets.

In its common delta-normal form, the method assumes that portfolio returns are approximately normally distributed.

A simplified one-day calculation, where the expected daily return is assumed to be close to zero, is:

1-Day Parametric VaR ≈ Portfolio Value × Z-score × Standard Deviation

If expected return is included explicitly, the calculation may also adjust for the portfolio's estimated mean return.

For a one-tailed VaR calculation, the standard normal Z-score is approximately:

• 1.645 at a 95% confidence level
• 2.326 at a 99% confidence level

These values are standard normal quantiles used for one-tailed probabilities.

The distinction matters because these are VaR confidence levels, not conventional two-sided statistical confidence intervals. For example, a two-sided 95% confidence interval commonly uses a Z-score of approximately 1.96 instead.

3. Monte Carlo Simulation

Monte Carlo VaR generates a large number of possible future market scenarios using a statistical model.

Each simulated scenario produces a hypothetical portfolio gain or loss. These outcomes are then ranked to identify the VaR threshold at the required confidence level.

Monte Carlo methods are flexible because the model can include different volatility assumptions, correlations and non-linear price relationships. They can also incorporate non-normal or fat-tailed distributions if these features are explicitly built into the simulation.

However, the quality of the result still depends on the assumptions used. A sophisticated simulation does not automatically produce an accurate estimate if its underlying inputs are unrealistic.

Why Value at Risk Matters for Leveraged CFD Traders

VaR is widely used as a market-risk measure in institutional finance, although its regulatory role has changed. Under the current Basel market-risk framework, Expected Shortfall plays a central role in internal-model capital calculations, while VaR remains important for areas such as backtesting and internal risk monitoring.

Retail CFD traders can adapt the underlying VaR concept to assess how several positions may affect overall account risk.

Leverage lets you control a larger market exposure with a smaller amount of capital, but it can increase both your potential gains and losses.

For UK retail clients, rules from the Financial Conduct Authority (FCA) restrict CFD leverage to between 30:1 and 2:1, depending on the underlying asset. The rules also include account-level margin close-out protection and negative balance protection.

UK traders can also check whether a financial services firm is authorised using the FCA Financial Services Register.

VaR can support CFD risk management in several ways:

• Multi-asset risk assessment: Separate stop-loss distances do not show how several positions may behave together. Portfolio VaR can incorporate correlations or historical co-movements between assets, depending on the calculation method.
• Position sizing: A rising VaR can indicate that portfolio risk has increased because of higher volatility, larger positions or changing correlations. Traders may use this information alongside a structured position sizing process.
• Margin awareness: VaR can help estimate the effect of adverse market movements before an account approaches its margin threshold. For UK retail CFDs, FCA rules require margin close-out protection when account funds fall to 50% of the margin required to maintain open CFD positions.

Limitations of VaR and the Danger of Tail Risk

VaR is useful for summarising portfolio risk, but it has important limitations. Traders should understand what the figure does not show before relying on it.

1. VaR Is Not Your Maximum Possible Loss

A 99% VaR should never be interpreted as a maximum-loss figure.

A one-day VaR at the 99% confidence level means the model assigns approximately a 1% probability to a loss exceeding the VaR threshold.

Across roughly 250 trading days, that would correspond to around two or three breaches per year as a long-run statistical expectation if the model were perfectly calibrated and market conditions remained consistent. It does not mean breaches will occur on a predictable schedule.

More importantly, VaR itself does not show how severe losses could become once the threshold has been exceeded.

2. Normal Distribution and Fat-Tail Risk

Parametric VaR models commonly assume normally distributed returns. Real financial returns, however, can display more extreme observations than a simple normal distribution would predict.

These fat tails mean unusually large market movements may occur more often than a basic normal-distribution model suggests.

During periods of severe volatility, poor liquidity or sudden market gaps, losses can therefore be considerably larger than a parametric VaR estimate. Research and industry guidance have long identified this as an important limitation of normal-distribution VaR models.

3. Expected Shortfall (Conditional VaR)

Expected Shortfall (ES), sometimes called Conditional VaR (CVaR), helps address one of VaR's main weaknesses.

VaR tells you where the tail-loss threshold begins. Expected Shortfall estimates the average loss once that threshold has been exceeded.

For example:

This makes Expected Shortfall particularly useful when the size of extreme losses matters, rather than only the probability of crossing a particular threshold.

Regulatory research also highlights the high risks associated with retail CFD trading. The European Securities and Markets Authority (ESMA) reported in 2018 that 74–89% of retail CFD accounts typically lost money across the EU jurisdictions it analysed. The FCA stated in December 2022 that approximately 80% of customers lose money when investing in CFDs. These are historical regulatory findings rather than a fixed percentage that applies to every broker or every period.

For this reason, VaR should not be used on its own. Position sizing, margin management, stress testing and clear limits on leverage remain important when managing CFD risk.

Conclusion

With Value at Risk explained this way, it's clear the concept provides a structured way to estimate portfolio loss exposure over a defined period and confidence level. It can be particularly useful when several positions are open at the same time because it provides a portfolio-level view of risk rather than assessing each trade in isolation.

However, VaR is a statistical estimate rather than a maximum-loss guarantee. Extreme losses can exceed the calculated threshold, especially when market behaviour differs from the assumptions used by the model.

For CFD traders, VaR is best treated as one part of a broader risk-management process alongside position sizing, margin awareness, stress testing and controlled use of leverage.

To learn more about protecting trading capital and managing overall market exposure, see our guide to risk management in trading.

This article is for educational purposes only and does not constitute financial advice. CFD trading involves risk, and losses can occur quickly when leverage is used.