How Credit Score Migration Analysis Can Identify Opportunities, Mitigate Losses

by Michael Cochrum
How Credit Score Migration Analysis Can Identify Opportunities, Mitigate Losses

Most credit unions use a credit score to decision and price loan applications.  In fact, in many cases the credit score is used as the primary factor in credit and pricing decisions.  Therefore, since the score is such an important factor, it stands to reason that we should have a good understanding of what it means and how a credit union should use credit scores for application processing and ongoing risk management.

The credit score, no matter what model a credit union uses, is intended to predict the probability of borrowers in a  particular score band defaulting on a loan.  The most common, generic models predict the probability today of an applicant defaulting – going 90 days or more delinquent – in the next 24 months.  There are other models that predict different probability, so it is important that a lender understand the model being used.  Often, lenders view credit scores similar to the certification that one may see on a piece of meat purchased at a market, as in “Grade A” for example.  “Grade A” in meat processing signifies that the meat being processed meets a standard of quality.  This is not true for credit scores, and this is where many lenders go astray.  “Grade A” meat will always be “Grade A,” but a borrower will not maintain the same probability of default over time.  In fact, not only does the borrower’s score change over time, so does the probability associated with a certain credit score.

Today, a credit score of 680 could be associated with a 2% probability of default.  However, two years from now, a borrower with the exact same score may carry an 8% probability of default.  This is because the models do not change, but borrower behavior changes within those score intervals.  In cases of a severe recession, as our country recently experienced, even the most reliable borrowers can be challenged to make timely payments.  Factor into this that the borrower may move from one score interval to another, then one can see that there can be a great deal of variation in credit risk over time.

Yes, it is important to know what a borrower’s score is at origination, and it is important to analyze how borrowers in a particular origination score interval perform over time to inform future lending decisions.  It’s also important to measure a portfolio’s current risk by obtaining new scores for each borrower at certain intervals during the life of the loan.  Often, lenders will manage their portfolios based on the risk of the borrower at origination.  However, by conducting regular credit score migration analysis, the lender can take action to better mitigate future losses and discover new opportunities for growth.

The question then becomes, “How often should new scores be obtained?”  Optimally, as often as it is cost effective.  This could be monthly if it proved valuable, but it certainly should be no less than annually.  If one considers the credit score’s purpose, a quarterly assessment seems reasonable as sudden changes in a borrower’s behavior, or the probability of default, can be identified in enough time to prevent losses and capitalize on opportunities.

About the Author

Michael Cochrum
Michael Cochrum is the Executive Lending Advisor for CU Direct.