Skip to content

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

A retail bank builds a logistic regression scorecard to predict default on unsecured personal loans. An analyst proposes converting each continuous predictor, such as age and income, into a small number of grouped bins and replacing each bin with its weight of evidence before fitting the model. Which is the principal benefit of this approach?

Grouping variables into bins and using weight of evidence captures non-linear effects, dampens outliers, lets missing values be treated as a separate category, and keeps the scorecard transparent. It guarantees neither better performance nor stability, and validation on holdout data is still required.

  1. AIt guarantees that the model will have a higher out-of-sample Gini coefficient than any alternative specification
  2. BIt removes the need to validate the scorecard on a holdout sample
  3. CIt captures non-linear relationships with default and handles outliers and missing values while keeping the scorecard transparentCorrect
  4. DIt ensures that the population stability index will always remain below 0.10

Explanation

Binning with weight of evidence lets a linear-in-logit model reflect non-linear risk patterns, limits the influence of outliers, allows missing values to form their own bin, and yields an interpretable points-based card. It does not guarantee superior performance, replace validation, or control population drift.

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