FRM Part II · FRM Exam Part II · Parametric Approaches (II): Extreme Value
With the fitted POT model (n = 1,000, N_u = 50, u = 3%, ξ = 0.5, β = 1), a risk committee asks for the 99.9% VaR rather than the 99% VaR of 5.47%. Which value is correct?
The 99.9% VaR is about 15.14%. The tail ratio is 20 × 0.001 = 0.02, its −0.5 power is 7.071, subtracting 1 and multiplying by 2 gives 12.14, and adding the 3% threshold yields 15.14%. This is much larger than normal scaling would suggest.
- A7.27%
- B12.14%
- C17.14%
- D15.14%Correct
Explanation
n/N_u × (1 − p) = 20 × 0.001 = 0.02. 0.02^(−0.5) = 7.0711, minus 1 = 6.0711, times β/ξ = 2 gives 12.142. Adding u = 3 gives 15.14%. Omitting u gives 12.14%, forgetting the −1 gives 17.14%, and scaling the normal quantile ratio gives 7.27%, which ignores the heavy tail.
Did you get it right without looking?
One question tells you little. A timed set on Parametric Approaches (II): Extreme Value shows your real accuracy, how long you take and where you lose marks.
More Parametric Approaches (II): Extreme Value questions
- A bank's risk team compares the GEV block maxima approach with the peaks-over-threshold (POT) approach for estimating extreme losses. Which …
- An analyst fitting a GPD to excess losses over a threshold estimates a shape parameter ξ of 0.25 and a scale parameter β of 2 (in USD millio…
- A GPD is fitted to exceedances of operational losses over a threshold, and the estimated shape parameter is ξ = 0.6. Which statement about t…
- An analyst has 1,000 daily loss observations (in % of portfolio value). Fifty losses exceed the threshold u = 2.0%. A GPD fitted to the exce…
- Using a POT model with threshold u = 2.0%, shape ξ = 0.2 and scale β = 0.5, a risk manager has computed a 99% VaR of 2.95%. Using ES = (VaR …
- A risk manager models the maximum loss over blocks of n = 10 days with a Gumbel GEV (ξ = 0) with location m = 2.0% and scale β = 0.5%. The G…