Report shows that credit agencies behind tech curve

A new study concludes that federal credit institutions, such as the Federal Housing Administration, the Department of Veterans Affairs' loan program and the Rural Housing Service, face a tenuous future because they are not using new technologies that have been embraced by private-sector lenders.

New technologies that enable lenders to quickly distinguish good credit risks from bad risks are letting private-sector credit institutions provide loans more cheaply than ever, the study said. If federal credit agencies do not more widely use those technologies, the agencies are in danger of losing their business to the private sector, the study argued.

The PricewaterhouseCoopers study, "Credit Scoring and Loan Scoring: Tools for Improved Management of Federal Credit Programs," urges the Office of Management and Budget to encourage federal credit agencies to "make multiyear commitments of resources to adopt scoring-based systems."

OMB officials did not return phone calls for comment.

Computer systems that perform credit scoring and loan scoring make it easier for lenders to weed out bad risks, the study said. Armed with data from the credit scoring systems, lenders can quickly grant loans to the most deserving loan applicants, speeding the entire loan process.

The systems also help lenders save money by steering them away from bad credit risks and by allowing them to spend less time processing good credit risks. The savings eventually are passed along to the consumer in the form of less expensive loans, the study says.

If federal lenders fail to adopt the new technologies, good credit risks will increasingly turn to the private sector for loans. As a result, the government could become stuck with an increasing number of higher-risk borrowers, which likely will mean more loan defaults.

In addition, the processing of poor credit risks is more expensive than the work that goes into lending money to good credit risks.

The private sector, said Thomas Stanton, the author of the study and a fellow at the Center for the Study of American Government at Johns Hopkins University, already has started to siphon off the "better borrowers, and the government will be left with the poorer risks. They should be dealing with that proactively."

Key to staving off the fiscal decline of federal credit agencies is "top-level commitments and multiyear commitments of resources. You can't just be on an annual budget cycle," Stanton said.

The conventional mortgage market already has made "steady incursions" into the market share of federal mortgage programs, the study said. In 1970, for example, the FHA provided mortgage insurance for 24.6 percent of single-family mortgages, and the VA provided 10.8 percent of single-family mortgages. By the end of 1997, however, the FHA served only 8.6 percent of the market, and the VA served 3.3 percent of the market.

The study also shows that in 1986, FHA loans defaulted 1.9 times more than conventional loans, and VA loans defaulted 1.8 times more than conventional loans. But in 1998, FHA loans defaulted 4.7 times more often than conventional loans, and VA loans defaulted 4.2 times more often.

Stanton identified the government's failure to embrace new technologies as a major contributor to the widening gap between conventional and federal loans. This gap will continue to widen if federal credit agencies do not embrace new technologies, he said, adding that "at some point, the government simply can't tolerate a very high level of losses. The program ceases to be cost-effective."

Keith Pedigo, the director of the loan guaranty service for the VA, said although he agrees with the study that embracing new technologies is important, he takes issue with the study's assertion that federal agencies aren't using technology enough. "Speaking for the VA, we are pursuing this very aggressively, and we intend to continue to do so," Pedigo said.

In the past several years, he said, the agency has adopted several credit-scoring technologies. Today, about 20 percent of all VA loans are being underwritten through one system alone. The agency recently approved the use of another system, and he said the agency likely will approve using yet another system within the next month.

"Once that approval has been completed," Pedigo said, "then the two major underwriting systems that exist in the industry will be available for making VA loans."

Automated systems draw from more pools of loan information than underwriters would reasonably be able to examine, and they run the data through complicated models with dozens of variables that can affect the ranking of a loan application, he said. The systems do not replace the need for underwriters but instead serve as "powerful tools," he said.

The systems "speed up the process," Pedigo added. "Most of these systems can generate a decision within four minutes. It results in a more prudent decision, a more solid business decision, which of course reduces the risk."

Stanton said Pedigo is a model manager, but disagreed that the VA is on the right track. "I don't think that VA has all of the resources that they need," he said.