Skip to content

FRM Part II · FRM Exam Part II · Risk Identification

A bank's internal loss database only records events with gross loss above USD 50,000. A risk manager fitting a severity distribution to this data without adjustment is most likely to encounter which bias?

A reporting threshold truncates the data from below, so observed frequency is understated and the average severity is overstated because small losses are absent. Analysts should fit truncated distributions or otherwise adjust for the threshold rather than treating the sample as complete.

  1. AUnderstated frequency and overstated average severity because small losses are missing from the sampleCorrect
  2. BOverstated frequency and understated average severity
  3. CNo bias, since large losses dominate capital
  4. DBias only in the correlation between business lines

Explanation

A collection threshold truncates the data from below. Observed counts understate true frequency of all events, and the sample mean severity is biased upward since small losses are excluded. Models must account for truncation, for example using conditional (shifted or truncated) distributions.

Did you get it right without looking?

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

More Risk Identification questions