Friday, July 8, 2011

FICO Scores In Residential Lending

The FICO (Fair Isaac Company) score is the accepted method of evaluating the credit-worthiness of an individual consumer. Without discussing the origins and mechanics of FICO scores, which will go beyond the scope of this post, quantifying their meaning is essential to valuing residential mortgage notes. One of the accepted fallacies of residential lending is that the 3 companies that report FICO scores--Equifax, Experian and TransUnion, do so in a way that makes their scoring interchangeable. In support of this viewpoint, most lenders either accept the median score or average the FICO scores, despite the fact that each of these companies follows a different philosophy in issuing a FICO score and frequently works with different information than the other 2.

Placing aside the issues that arise from treating the scoring of these 3 companies as equal, below is a general set of guidelines that links FICO score ranges with the likelihood of default by a consumer possessing a FICO score in that range on a loan or line of credit:

800+ = 1%

750-799 = 2%
700-749 = 5%
650-699 = 15%
600-649 = 31%
550-599 = 51%
500-549 = 71%
0-499 = 87%

This scale is posted at and comes from a book called CreditBooster: Ultimate Guide to a Better Credit Score Credit, Debt, Credit Scores, Credit Reports, Free Credit Reports, written in 2007 by the In Charge Education Foundation.

Looking at the posted delinquency rates, one can clearly see why FICO scores are so strongly tied to interest rates, which are merely an expression of the compensation that owner will accept for the risk of lending.

In your work sheet, using a logic function, like the IF() function, you can convert the FICO scores reported in your model into likelihoods of default. These numbers can be used to explore the way that a note is pricing buyer risk, once you are aware of all other considered factors and hold them constant. One can also use these percentages as a method of discounting note cash flows for consumer risk.

If there are more in-depth FICO default scales or if you are aware of another way to use FICO default scores, please feel free to post a response to this post.