Value at Risk (VaR)

A statistical measure estimating the maximum expected loss over a specific time period at a given confidence level. 'Our 1-day 95% VaR is $10M' means there's a 95% probability that losses won't exceed $10M in a single day. VaR became the standard risk metric in finance because it's intuitive and comparable across portfolios. However, VaR is dangerously misleading: it says nothing about how bad losses can be in the remaining 5% of cases. A portfolio with $10M VaR could lose $11M or $11B in its tail—VaR can't distinguish between them.

When to use it

When evaluating financial risk metrics and need to understand their limitations; when designing risk management systems that need to account for both normal and extreme scenarios; when someone presents VaR as comprehensive risk assessment; when deciding between risk metrics for different purposes.

How it can help

Use VaR as a starting point for risk communication, not as the final word. When someone quotes VaR, always ask: 'What happens in the tail?' VaR is useful for day-to-day risk management but dangerous for crisis preparation because it explicitly ignores the extreme events that actually cause ruin. Complement VaR with stress testing (what happens in specific extreme scenarios?), CVaR/Expected Shortfall (what's the average loss in the tail?), and scenario analysis (what could go catastrophically wrong?).