WHEN YOU hear that mortgage rates just fell to new quarter-century lows, do you know who qualifies for them?
Those premium rates generally are reserved for homebuyers and refinancers whose FICO credit scores are at the top of the heap - 720 and up. People with much lower scores often pay a lot more.
Last week, for example, consumers with FICO scores in the 750 to 800 range were quoted an average of 5.96 percent for fixed-rate loans, according to a survey of approximately 5,000 creditors nationwide conducted by Informa Research Services Inc.
By contrast, borrowers with lower FICO scores, in the mid-500s, were quoted an average 8.84 percent. On a $200,000 mortgage, high scorers will pay about $400 less per month in principal and interest than low scorers. Spread out over 30 years, that differential would amount to more than $100,000.
But that's all right, you might say: People who present greater risks of default should pay higher rates. But what if you are judged to be a higher risk than you really are? What if your score is low because the software that generated it is not the most consumer-friendly model available? How fair is that?
It's not. But with virtually no public attention on the issue, it's happening to American mortgage applicants every day of the week. That's because most lenders and brokers still are not using the most accurate behavior-prediction model of the FICO scoring system developed by Fair, Isaac & Co. to replace older, less predictive models.
The new model, Next Generation FICO, was introduced more than two years ago to improve on earlier models' inabilities to "see through" commonplace credit behavior patterns associated with late payments, new credit accounts, multiple "inquiries," revolving bank-card credit lines and "subprime" credit histories.
The Next Generation model, available to lenders and brokers who request it from the three national credit repositories, more accurately separates serious credit risks from less serious risks than the earlier models, according to Fair, Isaac.
In a recent test, 57 percent of all consumers scored higher on the latest model than on the older model. One out of six consumers scored 50 points or more higher, and another one out of six scored 10 to 39 points higher.
To grasp the significance of this, you have to understand that these scores were generated from the identical credit files or histories. All the credit facts were the same, but the scores on 57 percent came in higher on the new, more analytical software model.
