Stressed VaR

Standard value-at-risk is usually calculated from recent history, the last year or two of market moves. That's sensible for day-to-day risk management, but it has an awkward side effect: during a long calm period, recent volatility is low, so VaR quietly shrinks, and so does the capital a bank is required to hold, right before the calm ends. Stressed VaR was introduced specifically to stop capital from draining away during the quiet years before a crisis.

Same method, different window

The mechanics don't change: a bank still computes VaR the same way it always does, often via historical simulation over a fixed confidence level and holding period. What changes is the data window. Instead of the trailing year, the bank identifies a continuous historical period (commonly a year long) during which its current portfolio would have suffered the largest losses, typically a period covering 2008, or another severe stress episode, and calculates VaR using only that window's price moves.

Worked example

A bank's regular VaR, based on the last 12 months of low volatility, comes out to $20 million at 99% confidence. Regulators also require the bank to compute stressed VaR using a 12-month window from the 2008 financial crisis applied to the bank's current portfolio; that calculation produces $65 million. The bank's total market-risk capital charge is built from both numbers, so even though recent markets have been quiet, the capital held against this book is anchored by the $65 million stressed figure, not the $20 million calm one.

What this means in practice

Stressed VaR was added to the Basel framework after 2008 specifically because pre-crisis VaR-based capital had shrunk as volatility fell through the mid-2000s, leaving banks under-capitalized right as the crisis hit. By forcing a crisis-period floor into the capital calculation, regulators aimed to keep capital roughly stable across the cycle rather than pro-cyclically thin exactly when a shock is most likely.