Value at Risk (VaR)

VaR summarises the risk of a portfolio in a single number. It is defined by three elements: an amount of money, a time period and a confidence level.

For example, if a portfolio worth 10 million has a one-day VaR of 200,000 at 95% confidence, this means that on 95 out of 100 days, the portfolio is not expected to lose more than 200,000. On the remaining 5 days, losses are expected to exceed that amount.

VaR can be calculated in several ways, including using historical price movements, statistical assumptions about the distribution of returns, or computer simulations of many possible scenarios.

VaR is widely used by banks, investment firms and regulators to measure and limit market risk. However, it has important limitations. It says nothing about how large losses could be on the worst days beyond the confidence level, which is where the most damaging losses occur. It often relies on past data and assumptions that may underestimate the likelihood of extreme events. These weaknesses became clear during the 2008 financial crisis, and risk managers now also use stress tests and measures such as expected shortfall, which looks at the average loss beyond the VaR level.