Skip to content

FRM Part II · FRM Exam Part II · Credit Value at Risk

A bank validates its credit VaR model by comparing predicted portfolio loss quantiles with realized annual losses. Over 20 years, the 99% credit VaR was exceeded in 4 years. Which conclusion is most appropriate?

The model probably understates tail risk. At 99% confidence only about 0.2 exceedances are expected over 20 years, yet 4 occurred. The validator should examine assumptions such as default correlation, PDs and recoveries rather than treat the clustering as acceptable.

  1. AThe model is well calibrated because credit losses are naturally clustered
  2. BThe model likely understates tail risk, since about 0.2 exceedances would be expected at 99% over 20 years, and the validator should investigate correlation and parameter assumptionsCorrect
  3. CThe model overstates risk because exceedances are too frequent
  4. DNo conclusion is possible, since credit VaR cannot be backtested in any form

Explanation

At 99% confidence, expected exceedances over 20 years are 0.01 x 20 = 0.2. Observing 4 is far above that, indicating underestimation of tail loss, possibly from understated default correlation or PDs. Clustering of losses is itself a reason for concern rather than reassurance. Backtesting is hard with few observations but not meaningless.

Did you get it right without looking?

One question tells you little. A timed set on Credit Value at Risk shows your real accuracy, how long you take and where you lose marks.

Practise Credit Value at RiskFree diagnostic mock

More Credit Value at Risk questions