|
Decision
system construction, validation and regulatory review. |
|
Use of empirical methods
to classify loans as prime or subprime has
significant benefits from a regulatory or fair lending perspective. Decision criteria are documented and
consistent. Borrowers are
priced according to risk and prime borrowers never receive subprime rates. Behavioral models use
inputs such as payment history, updated credit scores, current delinquency
status, and change in property values to
re-estimate default probability and rank order risk of loss. Comparisons between pools and forecast loss
estimates are typically made using Behavioral Models. Behavioral models integrate into the Basel
II “Advance Practice” economic capital calculations. |
|
Management Consulting Services |
|
Mortgage Decision Technologies, Inc. |
|
To contact us: |
|
Mortgage credit models can
be subdivided into Prime and Subprime and
further broken into Origination models or Behavioral models. Prime mortgages make up
80-90% of all mortgage applications.
Prime mortgage origination models can
reliably rank order the risk of
loss at the date of funding such
that the highest risk loans will default more than thirty times as often as
the lowest risk loans. Models vary considerably
between lenders. Reasons for variance typically include
availability of data, operational considerations, or fair lending concerns. Subprime
models operate on a relatively small portion of the entire credit risk
spectrum. Some variables that are
important in Prime origination models are not statistically predictive
when only Subprime applications are examined. . |