Fair Isaac Corp. says it is changing the way it calculates FICO credit scores to better predict the ability of a borrower to repay loans.
The new procedures for calculating “Classic FICO” scores, to be implemented in September, should increase their predictive power by 5 percent to 15 percent, Fair Isaac said in a press release.
The largest increases in predictive power will be seen among originations and new accounts, subprime borrowers, and borrowers with “thin” credit files, Fair Isaac said.
Heavy reliance by mortgage lenders on FICO scores has been cited as one possible reason for an increase in delinquencies and defaults in subprime mortgage loans. Critics say the scores are easily manipulated, and that their usefulness in predicting a borrower’s ability to repay a loan has diminished.
Fitch Ratings, which analyzes pools of mortgage loans that are sold to Wall Street investors, has reported that the gap in FICO scores between loans that default in their first year and those that are still current dropped from 30 points in 2003 to 10 points last year.
Fair Isaac has said the scores themselves are not to blame for rising delinquencies and foreclosures in subprime loans, but that lenders misuse or ignore them. The company said U.S. businesses have used more than 100 billion FICO credit scores “to make smart decisions about their customers and prospects.”
The new Classic FICO scoring model won’t change features such as the score range of 300 to 850, score reason codes, minimum scoring criteria, inquiry treatment, and related model parameters, Fair Isaac said, and lenders should be able to use the score with minimal changes to their operating systems.