Credit
scoring is the set of decision models and their underlying
techniques that aid lenders in the granting of consumer credit. These
techniques describe who should get credit, how much credit they should
receive, and which operational strategies will enhance the profitability
of the borrowers to the lenders (Thomas, Edelman, and Crook 2002).
Credit Scoring, as
defined by SAS, includes the following:
-
applying a statistical model to
assign a risk score to a credit application or an existing credit
account
-
building the statistical model
-
monitoring the accuracy of one
or more statistical models
-
monitoring the effect that score-based
decisions have on key business performance indicators
Although credit scoring
is not as glamorous as pricing exotic financial derivatives, it is
one of the most successful applications of statistical and operations
research techniques in finance and banking. Without an accurate and
automated risk assessment tool, the phenomenal growth of consumer
credit would not have been possible over the past 40 years (Thomas,
Edelman, and Crook 2002).
In its simplest form,
a scorecard is built from a number of characteristics (that is, input
or predictor variables). Each characteristic includes a number of
attributes. For example, age is a characteristic, and “25-33”
is an attribute. Each attribute is associated with a number of scorecard
points. These scorecard points are statistically assigned to differentiate
risk, based on the predictive power of the characteristic variables,
correlation between the variables, and business considerations.
For example, using the
Example Scorecard in Figure 1.1, an applicant who is 35, makes $38,000,
and is a homeowner would be accepted for credit by this financial
institution’s scorecard. The total score of an applicant is
the sum of the scores for each attribute that is present in the scorecard.
Lower scores imply a higher risk of default, and higher scores imply
a lower risk of default.
Credit Scoring for SAS
Enterprise Miner contains the following nodes, which are added to
your SAS Enterprise Miner toolbar to support scorecard development:
-
Interactive Grouping —
groups input variables into bins before the credit scorecard is built.
An initial, automatic grouping can provide optimal splits, but this
node enables you to regroup the variables through an interactive interface.
It also has the capability to screen or select variables.
-
Scorecard —
uses the grouped variables as inputs in a logistic regression model
and usually follows the Interactive Grouping node. In addition, it
scales the regression parameters to compute score points and the resulting
scorecard. Finally, the Scorecard node performs score and characteristic
(variable) analysis that helps in understanding the scorecard, and
aids in crafting a score-based strategy.
-
Reject Inference —
offers three standard, industry-accepted methods for inferring the
performance of the rejected applicant data by the use of a model that
is built on the accepted applicants.
-
Credit Exchange —
enables the use of scorecards in SAS Credit Risk for Banking. Because
it plays no part in the development of the scorecard, coverage of
this node is beyond the scope of this tutorial.