Historical VaR: Estimating Loss From Past Returns

Key Takeaways

• Historical VaR reads a loss percentile straight from past returns, with no distribution assumption.
• A 1-day 95 percent VaR of 2 percent means losses should exceed 2 percent on about 1 day in 20.
• The biggest error is using too short a window, which omits crashes the future can still deliver.
• It anchors position sizing and risk limits to what the portfolio has actually lived through, not theory.

What It Is

Historical VaR, sometimes called historical simulation, answers a specific question: based on how this portfolio's holdings actually moved in the past, what loss sits at a chosen probability cutoff? You collect a window of historical returns, sort them from worst to best, and pick the return at your confidence level.

A 1-day 95 percent historical VaR of 2 percent means that, across the historical window, the portfolio lost more than 2 percent on roughly 5 percent of days. VaR is a threshold, not a worst case. It tells you the loss you should breach only on the bad tail of days, not the maximum you could ever lose.

The Intuition

Parametric methods assume returns follow a clean curve, usually the normal distribution. Real markets do not cooperate. Returns have fat tails, meaning extreme moves happen more often than a bell curve predicts, and they are often skewed toward losses.

Historical VaR sidesteps the assumption entirely. It treats the past return series as the best available picture of what can happen and lets the data speak. If the window contains a crash, the crash shows up in the tail. If it does not, the method has no way to know the crash exists. That dependence on the chosen window is the central trade-off.

How It Works

The procedure has four steps.

1. Choose a lookback window (for example, 500 trading days).
2. Revalue today's portfolio using each day's historical return.
3. Sort the simulated profit-and-loss results from worst to best.
4. Read the loss at the chosen percentile.

For a 95 percent VaR on 500 observations, you locate the 25th worst outcome, since 5 percent of 500 is 25. For 99 percent VaR, you take the 5th worst. The result is a currency loss or a percentage of portfolio value over the horizon.

Two design choices matter most. The window length sets how much history feeds the estimate. The weighting scheme decides whether every day counts equally or recent days count more. The RiskMetrics framework popularized exponentially weighting recent observations so the estimate reacts faster to changing volatility.

Worked Example

Suppose you hold a 1 million dollar equity portfolio and use the last 500 daily returns. You apply each historical daily return to the current portfolio value, producing 500 hypothetical 1-day profit-and-loss figures.

You sort those 500 results from worst to best. The 25th worst figure is a loss of 19,400 dollars, which is 1.94 percent of the portfolio.

1-day 95% Historical VaR = 19,400 dollars (1.94%)

The reading says that on a normal day you should not lose more than about 19,400 dollars, and you would expect to breach that level on roughly 1 trading day in 20. If you needed 99 percent VaR instead, you would read the 5th worst figure, which would be a larger loss.

Common Mistakes

1. Using a window too short to contain stress. A 250-day window covering only a calm year will report a comforting VaR that collapses the moment volatility returns. Many risk teams keep at least one full crisis in the window for this reason.
2. Assuming VaR is the worst case. VaR is the threshold of the tail, not its depth. The losses beyond VaR can be far larger, which is why expected shortfall exists.
3. Ignoring stale or illiquid positions. Historical simulation needs accurate past returns for every holding. Thinly traded assets with gappy price history distort the tail.
4. Treating non-overlapping horizons casually. Scaling a 1-day VaR to 10 days by multiplying by the square root of 10 assumes independent returns, which historical data often violates.
5. Forgetting that the portfolio changed. Historical VaR revalues today's positions with old returns. If your exposures shifted, last year's correlations may no longer apply.

Frequently Asked Questions

What is historical VaR in simple terms? Historical VaR estimates how much a portfolio could lose by sorting its past returns and reading the loss at a chosen probability. A 95 percent reading marks the loss you would expect to exceed on about 1 day in 20.

How does historical VaR affect investment decisions? Risk teams use it to set position limits and capital buffers tied to losses the portfolio has actually experienced. Because it uses real history, a manager can size exposure to what the strategy survived, not to a theoretical curve.

What is a real-world example of historical VaR? A trading desk holding 1 million dollars and using 500 days of returns finds the 25th worst day was a 1.94 percent loss. That 19,400 dollar figure becomes its 1-day 95 percent VaR.

How can investors avoid misusing historical VaR? Keep a long window that includes at least one stress period, and pair the number with expected shortfall to see the depth of the tail beyond the threshold.

How is historical VaR different from parametric VaR? Historical VaR reads losses from actual past returns and assumes no distribution shape. Parametric VaR assumes returns follow a curve, usually normal, and computes the loss from volatility and a z-score.

Sources

1. J.P. Morgan/Reuters. RiskMetrics Technical Document, 4th ed. (1996). https://www.msci.com/documents/10199/5915b101-4206-4ba0-aee2-3449d5c7e95a
2. CFA Institute. Measuring and Managing Market Risk. https://www.cfainstitute.org/insights/professional-learning/refresher-readings/2026/measuring-managing-market-risk
3. Basel Committee on Banking Supervision. Minimum Capital Requirements for Market Risk. https://www.bis.org/bcbs/publ/d457.htm
4. Investopedia. Value at Risk (VaR). https://www.investopedia.com/terms/v/var.asp

Disclaimer

This article is educational content only and is not financial advice. Nothing here is a recommendation to buy, sell, or hold any security. Consult a licensed advisor before making investment decisions.