Want a Loan? Act Responsibly

As bankers claw their way out of the credit crunch, they're likely to get a lot more curious about our ability to repay loans. To do that, they'll no doubt search for statistical correlations between financial risk and our behavior in different realms—as shoppers, students, even drivers.

Indeed, quantitative analysts in major banks and researchers at credit risk companies are hard at work looking for ways to understand borrowers better. The logical first step is to pore over more of our data. Clark Abrahams, chief risk officer at SAS, a large software company that creates analytic programs for the banking industry, suggests lenders may one day take into account lots of nontraditional metrics, such as whether the borrower has a good reputation on eBay (EBAY) or pays cell-phone bills on time before deciding whether to extend credit.

At first blush, it may seem odd that banks need more data on borrowers. After all, mortgage bankers and credit-card companies feasted on financial data during the lending spree that helped inflate the housing bubble. Lenders studied individuals' borrowing and payment patterns and stuffed mailboxes with microtargeted pitches for new loans and credit cards. But they focused their analysis on the borrowers' appetite for credit—not on people's ability to afford it or the risk that they would default. New risk models, analysts say, are sure to call for more financial data, including revenue projections for each borrower. In other words, they'll want to know more about how much we make and how we spend it.

Ranking You on Responsibility

Already, marketers, advertisers, and political consultants are harvesting mountains of data about people and building sophisticated mathematical models to predict their behavior. "If they're smart, [bankers] will be using these techniques to figure out each customer's risk, and to give them customized offers," says Dave Morgan, founder of Tacoda, a behavioral targeting advertising company bought last year by Time Warner's (TWX) AOL.

A hot spot of this type of data study is at the San Rafael (Calif.) research labs of Fair Isaac (FIC), creator of the widely used FICO credit risk scores. In the short term, Fair Isaac is sifting through financial data to calculate not only each borrower's risk, but also how much debt each one can take on.

Looking further ahead, Fair Isaac predicts that based on analysis of our data, we'll each have scores that can predict far more than our financial behavior. Fair Isaac research fellow Larry Rosenberger speculates that one day, each of us will be scored for broad values such as "responsibility." Such a score, still years away, could be used to appraise a person's worthiness for a whole range of benefits, from housing loans to employment in a nursery school to rates on car insurance. Colleges might even find it useful in admissions decisions.

Looking for a Broader Read

And how would data-crunching companies come up with such scores? That's where new sources of data come in. According to Fair Isaac CEO Mark Greene, research indicates that "bad people are bad people are bad people." In other words, their behavior in one domain predicts what they might do in another. People who get in traffic accidents and don't pay their taxes on time, Greene says, "are often bad credit risks."

This means that more of our lives—our school records, for example, or claims made on insurance policies—could provide the data for broader responsibility scores.

Already, a number of industries have used Fair Isaac's FICO credit risk score for a broader read on a person's responsibility. The FICO score is based on limited data regarding credit and payment history. But it turned out to be a predictor for auto and home insurance claims. And recently Rosenberger was stunned to see a study pointing to a new correlation: People who pay their bills on time seem more likely to stick to exercise regimens at the health club. Could losing weight boost a person's responsibility score?

What's Your eBay Status?

This type of scoring may sound menacing. What's to stop banks from using nontraditional statistics to unearth measures that divide society ethnically and regionally, leading to new forms of discrimination, so-called red-lining. Let's say that analysts find that people who get new treads on car tires default on loans more often than those who buy new tires. Chances are, most of those economical drivers make less money and live in low-income neighborhoods. If so, the behavior may point to a demographic grouping. "We have to ask ourselves as a society [whether] we want to be making those calls," Abrahams says.

For now, reaching beyond standard financial statistics remains a research project. Use of "responsibility" scores, for one, will depend on privacy rules and regulations that societies develop, Fair Isaac says.

What's more, some of these new scores and metrics could prove helpful to customers. In a climate of tight credit, banks may be reluctant to lend to those who lack traditional credit histories. Incorporating new data, from school grades to one's status on eBay, could open doors for first-time borrowers.

But chances are, banks will also use these new sources of data to figure out the rest of us, too