Banks are looking for more data to help them make lending decisions, beyond the simple matter of whether prospective borrowers pay their bills on time.
The focus on "alternative" data to credit scores in underwriting is not new but has intensified after the lessons banks learned from the credit bubble.
"In the marketplace right now everyone is talking about rethinking underwriting," said Peter Carroll, a partner in the retail and business banking practice of the management consulting firm Oliver Wyman.
"Everyone realizes that credit scores, as clever as they are, have in some respects left out of the credit-assessment equation certain aspects of the borrower," Carroll said.
Though more data about people is available today than ever before, sifting through it is arduous.
"Basically people are saying we can either go back to human underwriting, which is cost-prohibitive and not that good anyway, or we can find data sources that potentially shed light on these other dimensions of the borrower," Carroll said.
Banks are especially interested in ways to improve their identity-verification processes, said Michelle Reinhard, the senior vice president of quality risk management for Huntington National Bank, the banking subsidiary of Huntington Bancshares Inc. in Columbus, Ohio.
"Having a larger tool for risk management is really where I think the whole industry is going, where you can clearly identify your customer," Reinhard said.
For example, a customer may have applied for credit, listing his home phone number on the application, and later applied again listing a mobile phone number. Reinhard said the discrepancy could trigger a red flag. "The institution may have had to work harder to verify that the consumer information was accurate," she said, but better identity-verification capabilities could have made this extra work unnecessary.
She said Huntington is trying to fine-tune its decisioning techniques.
Zoot Enterprises Inc. is expanding the data its credit decisioning software can use to help its bank customers evaluate applications. "We find that right now in the market there is a hunger for other data and additional intelligence," Tom Johnson, Zoot's vice president of product development, said in an interview.
In February the Bozeman, Mont., technology company struck a partnership with LexisNexis Risk Solutions, which it has worked with previously, to incorporate information from the New York data aggregator's public records vault into its prescreen applications; it expects to begin marketing the capability as part of its Prescreen 3.0 product this quarter. LexisNexis is a subsidiary of the London information company Reed Elsevier Inc. Zoot has used LexisNexis data in the past with other products but is expanding its relationship with the company through its new agreement.
Johnson said the details in the Lexis data can augment credit bureau files and other information sources that banks commonly use when deciding to extend credit or send prescreened offers to consumers. The potential outcome for banks is better-performing credit portfolios and identification of new customer prospects who previously were denied products, he said.
Carroll said the LexisNexis data included "fundamentally different" details that are "additive" to, "not competitive" with, traditional credit bureau reports.
For example, credit card issuers have long marketed their cards with direct mail. "If you look at everyone in the bureau file and pick the subset that you want to mail to, there's a whole bunch of people" with no file, he said. "If you've got another source of data like LexisNexis … you can use this to pick out good people who would previously have been no hits and thin files."
That also creates an opportunity to potentially market to underbanked customers, Johnson said.
Zoot also sees opportunities for banks to lower their screening costs by ruling out a credit applicant sooner.
"Every data source that you pull has costs associated with it," Johnson said. "The sooner in the process you can make a decision … the less expensive the transaction is going to be."
LexisNexis uses public records and other sources to track bankruptcies and liens — which will usually included in a standard credit report — and plenty of other things that often are not, according to Grayson Clarke, the senior director of credit risk decisioning with LexisNexis Risk Solutions.
This could include details about property values and ownership, an applicant's educational background, professional licenses, phone service history, subprime credit information such as a payday loan and ownership of other assets, such as boats and airplanes.
"Traditional bureau data captures whether or not someone is repaying an obligation," Clarke said, while the LexisNexis' data includes other indicators of a person's credit risk.
For example, Clarke said people who hold certain professional licenses have proven to be more stable credit risks than those without.
And while those people already may have been deemed good credit risks based on traditional data sources, the ability to tap into such additional details to reinforce that can make a bank's underwriting stronger and help with segmenting customers for marketing purposes, he said.
All this provides "an additional layer of data to the credit bureau files that improve the acceptance rates and improve" accuracy, said Zoot's president, Dennis Dixon.
Alan Riegler, a partner and senior consultant with the financial services consulting firm CCG Catalyst in Los Angeles, agrees that giving banks new types of data to evaluate can help mitigate risk, though he said the current economic climate might make it hard for some financial companies to buy new analytics software.
"You can determine from a credit score if [a person] has had a bankruptcy," Riegler said. "What is not as likely to be determined … is have you committed fraud."
