FICO Seeks to Take Scores to New Level
Fair Isaac Corp., the creator of the ubiquitous FICO credit scores, wants the next generation of scores to not only reflect a record of how well or poorly consumers have paid their debts in the past but also to predict how they are likely to react to future economic events.
To help accomplish this, Fair Isaac has rolled out something it has called the FICO Economic Impact Service, a product that company executives have explained will allow them to "nuance" or "shade" credit scores and allow lenders to make more targeted underwriting decisions.
"Banks are experiencing tremendous pressure to balance demands for growth and profitability with the imperative to manage risk," said Andrew Jennings, chief research officer at FICO. "Changes in unemployment rates, interest rates and other economic indicators often anticipate important marketplace movements. It can be extraordinarily difficult for even large lenders to objectively assess such external risks and adapt their credit decision systems quickly enough to be effective. That's why we've developed this new way to enhance traditional risk scores and significantly strengthen lenders' abilities to manage risk and boost profitability in dynamic economic environments."
To do this, FICO said it examines up to 150 different economic indicators, for example, unemployment rate, interest rates and gross domestic product, then projects changes in risk using predictive analysis that the company pioneered. Lenders using the scores can choose how frequently they wish to examine economic indicators and correspondingly adjust their risk-management strategies.
"As a result, lenders can keep their strategies tuned to evolving market conditions and determine appropriate levels of loan-loss reserves and capital requirements for a more sustainable, healthier performing business," the company said.
Company executives pointed to the recent wide economic swings and resulting changes in consumer behavior and loan repayment patterns as part of the spur for the new product. For example, for the first time ever, there began to be a greater risk of consumer mortgage default than consumer credit card default. This even though, traditionally, a mortgage loan secured by real estate has been seen as more secure and less risky than credit card loans, which are unsecured.
FICO said the primary indicator of this change was the way the gap between default rates on credit cards versus mortgages has narrowed.
In 2008-2009, bank card accounts were 1.6 times more likely to become 90 days delinquent than were mortgage loans, according to the company's analysis. But in 2005 bank card accounts were more than three times more likely to become 90 days delinquent. And for borrowers with higher credit scores, the level of repayment risk actually has become greater for real estate loans than for bank cards. In 2009, 0.3% of consumers with scores between 760 and 789 defaulted on real estate loans, compared to 0.1 percent% who defaulted on bank cards, the company said.
The new product could help card issuers guard against these sorts of trends by providing more accurate risk prediction than those restricted to looking only at a consumers previous actions on loans, the company said. FICO said it compared the new product to a U.S. card issuers' actual delinquency rate, the rate predicted by cardholder credit scores and a rate predicted by the new product. Over a period stretching from June 2005 to June 2008, the company found that rate predicted by the Economic Impact Score tracked the actual default rate far more closely than the default rate predicted by the traditional score alone.
FICO does not see the new scores as supplanting its traditional scores but instead as an additional tool to use with the traditional score. The new tool could help a credit union predict the impact a job loss or slow down to some members-whether a member who loses a job would be more or less likely to default on the credit card, for example, or get a an idea about which members would respond better to different collections approaches.