FICO scores cover a much wider demographic than proprietary models used by ABS issuers to predict collateral performance, according to FICO Senior Director Adem Yilmaz.
Yilmaz was responding to a report from Standard & Poor's released yesterday. The release stated that GM Financial's proprietary credit scoring was a better predictor of the performance of their subprime auto ABS than FICO scores. For ASR's related story, please click here.
S&P further stated that the difference in accuracy was most pronounced for those deals issued between 2006 and mid-2009. S&P suggested that these results can be extrapolated to cover other originators in the sector.
Meanwhile, Yilmaz said that all of the top 20 auto lenders make use of FICO scores, whether for originations, account management or targeting prospects.
"The FICO score facilitates communication between the dealer and the lender by providing a common language about the level of risk that an applicant represents," Yilmaz said.
He added that the score helps lenders and dealers to find prospects that cannot be scored using their proprietary methods.
Proprietary scores are built from a sub-segment of the population and measure their behavior under limited circumstances, while FICO scores cover much wider demographics, Yilmaz explained.
"As lenders expand their portfolio they might find out that the pool of loans they used to develop proprietary scores might not be representative of their expanded demographics," Yilmaz stated.
In yesterday's report, S&P Senior Director Amy Martin said that many subprime auto companies argue that their scoring methods do a better job than FICO scores in forecasting default probability. "Unlike FICO, proprietary scores can incorporate additional variables, such as borrower income, the amount of down payment, and loan-to-value ratio," Martin said.