Value at risk (VaR) is one of the most widely used measures of market risk. It answers a simple question: over a given period, what is the loss that I should not exceed, except in the worst few percent of cases? This calculator uses the parametric method, based on the normal distribution, to give the VaR of a portfolio from its value, its volatility, its expected return, the horizon and the confidence level. It also gives the expected shortfall.
How the calculation works
- Z-score from the confidence level: 1.282 (90%), 1.645 (95%), 2.326 (99%)
- Horizon volatility = annual volatility × √(days ÷ 252)
- Horizon return = annual return × days ÷ 252
- VaR = portfolio value × (z × horizon volatility − horizon return)
- Expected shortfall = value × (volatility × φ(z) ÷ (1 − confidence) − horizon return)
Example
A portfolio of $100,000 has an annual volatility of 18% and an expected return of 7%. Over 10 trading days the volatility is 18% × √(10 ÷ 252) = 3.59% and the expected return is 0.28%. At 95% confidence, z is 1.645, so the VaR is 100,000 × (1.645 × 3.59% − 0.28%) = $5,620, or 5.62%. The expected shortfall is $7,118.
Why it matters
Banks, funds and risk managers use VaR to set limits and to compare the risk of very different portfolios in a single unit, money. Individual investors can use it to see whether the possible loss over a short period fits their tolerance for risk, before taking a large position.
Good to know
VaR is a threshold, not a worst case. A 95% VaR is exceeded in one period out of twenty, by an amount the measure does not tell. The normal model understates extreme moves, so stress tests and historical simulation are used alongside it. Volatility changes over time, and the square-root-of-time rule is an approximation. This tool is for education and is not investment advice.
Related tools
The portfolio risk calculator computes the volatility of a mix of assets, the Sharpe ratio calculator compares return to risk, and the standard deviation and normal distribution calculators help with the statistics behind VaR.