Skip to content

FRM Part II · FRM Exam Part II · Credit Scoring and Rating

A card issuer's scorecard was developed on accepted applicants only. Management notes that applicants rejected historically have no observed repayment outcomes. A risk analyst proposes reject inference to address the resulting problem. Which statement best describes the issue and its effect if ignored?

The problem is sample selection bias. A model built only on accepted applicants does not represent the full applicant population, so it may misestimate risk for applicants similar to past rejects. Reject inference tries to correct this by inferring likely outcomes for rejected applications.

  1. AMulticollinearity, which inflates coefficient standard errors but leaves predictions unbiased
  2. BSample selection bias, which can make the model misestimate risk for the full through-the-door populationCorrect
  3. COverfitting, which is eliminated by adding more accepted applicants
  4. DLook-ahead bias, which arises because variables are measured after the outcome window

Explanation

Training only on accepted accounts creates sample selection bias because the sample is not representative of all applicants. The model may misrank or misprice applicants resembling past rejects. Reject inference attempts to assign outcomes to rejects; more accepted data does not fix the non-random selection.

Did you get it right without looking?

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

Practise Credit Scoring and RatingFree diagnostic mock

More Credit Scoring and Rating questions