Research Scan: Journal Enters DebateOn Fair-Lending Issue

The Journal of Financial Services Research has waded hip-deep into one of the most wrenching debates in banking research: Do lenders discriminate against minority borrowers?

The academic periodical devotes its most recent issue to nine fair- lending studies, which address issues varying from using the Home Mortgage Disclosure Act to detect bias, to overages, to the economics of low-income lending.

All the studies have been refereed, which means other economists have reviewed and approved the techniques used by the authors. The entire journal is available by calling 617-871-6600.

Highlights from the journal include:

Home Mortgage Disclosure Act data may serve as an effective tool for identifying potential lending discrimination, according to Federal Reserve economists.

Examiners currently use their own discretion when selecting files for review, producing inconsistent results and making it difficult for them to justify their findings statistically, write Robert B. Avery and Paul S. Calem of the Federal Reserve Board and Patricia E. Beeson of the Federal Reserve Bank of Cleveland.

But the Fed has developed a statistical program that scans HMDA data to identify bank products in specific markets that deserve more scrutiny. Then examiners use HMDA data to identify matched pairs and see whether white and black consumers with similar incomes, seeking similar loans, in similar communities are treated the same.

The process is more defensible than the existing method because it is based on statistical sampling techniques, they write. It also is more efficient than having examiners select files at random.

For more information on "Using HMDA as a Regulatory Screen for Fair Lending Compliance," call Mr. Calem 202-452-3116.

Statistical models used in past studies to detect lending bias were fundamentally flawed, according to Federal Deposit Insurance Corp. economist David K. Horne.

Models, including one used in the Federal Reserve Bank of Boston's historic 1992 study, fail to capture all the factors that go into underwriting decisions. For instance, the study did not account for monetary gifts from relatives for a down payment.

Another problem is that researchers are unable to verify the underlying data. This means there is no way to correct for errors that could bias the results. Mr. Horne tried to correct these errors by constructing new models. When he ran the Boston Fed's data through other models, he found no evidence of bias.

For more information about "Mortgage Lending, Race, and Model Specification," contact Mr. Horne at 202-898-3981.

Lenders reject minority group members more often than whites because more unqualified blacks and Hispanics apply for loans, according to an economist.

Eric Rosenblatt, director of credit policy at Fannie Mae, reviewed 12,000 conventional and government-backed mortgage applications filed with a major lender.

The different education levels, language skills, and cultural backgrounds of minorities make it more difficult for them to judge whether they qualify for credit, he found. As a result, more unqualified minorities apply for loans, which means lenders reject more minority borrowers.

For more information about "A Reconsideration of Discrimination in Mortgage Underwriting with Data from a National Mortgage Bank," call Mr. Rosenblatt at 202-752-7000.

For reprint and licensing requests for this article, click here.
MORE FROM AMERICAN BANKER