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 it 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 certain scores.
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. While the models do not change, borrower behavior can change within those score intervals. In cases of a severe recession, as recently experienced in the United States, 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, and 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. But 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.
Let’s take a look at several examples:
John Doe had a credit score of 720 when his line of credit was originated two years ago. He has had an open line of credit of $20,000 since origination that he has never used. Borrowers in this score range typically have a probability of default equal to 2% and carry a 10% average utilization. This would mean that you might forecast a loss of $40 for this account. When a new score is obtained for Mr. Doe today, he has a score of 640 and he has increased his utilization of this line of credit to double the average. If a drop in credit score of this magnitude indicates a probability of default five times greater than when the borrower had a 720 score, this would mean that the actual loss probably would be 10% of the now $4,000 balance, or $400 — ten times greater than what is being allowed. A reduction in credit line or an increase in the allowance for loss are indicated here.
Two years ago, Jane Smith got a car loan for $10,000, even though she had a very low credit score of 600. Her rate on that loan is 12%. She’s done a good job of making her payments on time and has also paid off some collection items that appeared on her credit report and were affecting her score. Today, her credit score is 680. She thinks she may be able to refinance her car loan at a much lower rate with another lender, so she applies and gets the loan, plus her first credit card with a $1,000 line of credit. Her current lender could have made the same offer based on her current score and the lender’s current lending criteria. Since the lender didn’t get new scores for Jane, they never had the opportunity to keep the loan.
These are just two scenarios reinforcing why lenders should track borrowers’ credit score throughout the life of their relationships. But the question then becomes, “How often should new scores be obtained?” Optimally, as often as is cost effective. This could be monthly, if it proved valuable, but 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 early enough to prevent losses and capitalize on opportunities.