Credit Scorecard

From Open Risk Manual
Revision as of 20:00, 11 September 2020 by Wiki admin (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Definition

A Credit Scorecard is a type of Risk Model used in the classification (scoring) of Credit Risk for individuals, corporations or other legal entities. The scorecard output (Credit Score) is an assessment of the relative likelihood of a certain credit event occurring, given a number of observable inputs.

A credit scorecard typically (but not necessarily) implements some type of quantitative credit scoring model. The main output of a scorecard is a Credit Score. A Credit Score is a quantitative (numerical) indicator of the Credit Quality (credit worthiness) of a borrower. A credit score is typically produced after the review (potentially using algorithms) of an individual's Credit History, financial condition and other characteristics.

Credit Scorecard Taxonomy

There are many credit scorecard types. The Credit Scorecard Taxonomy break thems down to the main classes by:

  • usage
  • type of input
  • type of output
  • manner of development and
  • algorithm used (if any)

Role of Credit Scorecards within a Credit Rating System

A credit scorecard is typically an integral component of a Credit Rating System, a framework that consistently classifies and quantifies all credit risk aspects of a portfolio. Depending on the type of scorecard, outputs may be used directly in such a system or they may need additional processes, e.g. mapping scores to ratings. For a discussion of the relation see Credit Rating versus Credit Score

Credit Scorecard Glossary

Term Description
Credit Scorecard Accuracy (or Gini) The determination of the degree to which a scorecard can correctly predict / classify the Credit Event it aims to capture. The Accuracy Ratio is one of the metrics used (See Also Gini Index)
Credit Scorecard Binning The practice of segmenting a continuous numerical score (e.g. from 0 to 1000) in a small sets of bins. See Binned Variable
Credit Scorecard Development Development of scorecards is highly dependent on context. An overview is given in the How to Build a Credit Scorecard entry.
Credit Scorecard Dataset A credit scorecard typically requires historical and other Credit Data for its development
Credit Scorecard Evaluation (or Validation) A credit scorecard is an example of a Risk Model and thus typically requires some form of Model Validation
Credit Scorecard Formula There are very many types of scorecards (See Credit Scorecard Taxonomy) so there is no single formula that is applicable. The closest to a formula would be the very popular Logistic Regression Credit Scorecard
Credit Scorecard Model For a catalog of credit scoring models see Credit Scoring Models
Credit Scorecard Quality See Credit Scorecard Accuracy
Credit Scorecard Reporting This is the production of a regular report to assess whether a credit scorecard still performs according to requirements

See Also