Credit Scoring Models and Scoring Dynamics in the U.S.

Credit scores are used extensively in household debt underwriting and pricing as an indicator of borrowers’ credit risks. In addition, they feature prominently in consumers’ financial lives in a variety of ways beyond household borrowing, including rental, insurance, and employment decisions. Currently, multiple competing models are available for scoring consumers’ credit risks, and lawmakers recently approved a provision that would require mortgage-finance companies Fannie Mae and Freddie Mac to consider multiple credit scores in determining a mortgage applicant’s credit risk. These models typically do not produce identical credit scores for the same consumer and therefore can provide different indicators of a particular consumer’s credit risk. Using a novel dataset that includes credit scores of the same individual derived from three different scoring models we study similarities and differences between existing credit scoring models in the U.S. Our results indicate that, in general, various credit scoring models evaluate consumers’ relative credit risks at a point in time in a consistent fashion, but the models appear to interpret new credit information differently.