Skip to content

FRM Part II · FRM Exam Part II · Non-parametric Approaches

A risk manager bootstraps a 99% historical simulation VaR using 1,000 resamples. The resulting VaR estimates are sorted. What is the most appropriate way to obtain the bootstrapped point estimate of VaR and its dispersion?

The best bootstrapped VaR estimate is the average of the VaRs from all resamples, and their standard deviation or percentiles give the estimate's precision. Picking the largest or smallest value, or ignoring the resamples, wastes the information that the bootstrap generates.

  1. ATake the average of the 1,000 VaR estimates as the best estimate and use their spread to assess precisionCorrect
  2. BTake the largest of the 1,000 VaR estimates as the best estimate to remain conservative
  3. CTake the VaR from the original sample only and ignore the resampled values
  4. DTake the smallest of the 1,000 VaR estimates since it is least affected by outliers

Explanation

The bootstrapped estimate is typically the mean of the resample VaRs, which is generally more precise than the single-sample estimate. The standard deviation or percentiles of the resample VaRs give the dispersion. Using extremes or ignoring the resamples discards the information the bootstrap provides.

Did you get it right without looking?

One question tells you little. A timed set on Non-parametric Approaches shows your real accuracy, how long you take and where you lose marks.

Practise Non-parametric ApproachesFree diagnostic mock

More Non-parametric Approaches questions