An FHLB asks: Can alternative payment data improve fairness?

Alternative credit reporting data could help government related agencies make mortgage underwriting and homeownership more equitable among racial groups, but the goal may not be easy to achieve, according to a new study.

Considering data like residual income, rent or utility payments in underwriting can have a net benefit when it comes to narrowing the 30 percentage-point homeownership gap between white and Black households — if it’s done right, according to the report, the first in a series funded by the Federal Home Loan Bank of San Francisco.

Statistics show Black adults are, for example, disproportionately renters and under- or unbanked. So using rent histories to address barriers to building credit holds promise when it comes to helping them achieve homeownership, according to Federal Reserve System data cited in the study done by the Urban Institute’s Housing Finance Policy Center. 

More than one-fourth or 27% of Black adults have a thin credit file, and 13% are unbanked altogether, according to the Federal Reserve data. In comparison, only 9% of white adults are underbanked and 3% are unbanked. A host of private vendors and government-related entities have been rallying around this issue, spurred in part by the Biden administration’s goal to expand homeownership.

While all that is promising, broader systemic barriers to building wealth could limit whatever progress is made in using alternative data to build credit.

For example, some government-related entities in the housing market have been careful to only approve the use of positive vendor-reported rent records in underwriting. They’ve also tried to ensure that vendors don’t charge fees that prevent people from getting into a better financial position for homeownership.

Even when those precautions are taken, narrowing the homeownership gap using alternative data isn’t easy. While a large share of Black adults require this kind of assistance to get a mortgage, they also may be more prone to miss payments because of their lack of wealth, so their ability to establish a positive rent history may lag other groups, the report said. The share of Black adults who make on-time rent payments, is 55.3%, below the average of 67.8% for all renters. 

While complications like this do slow progress toward the goal, on a net basis, using positive alternative data as a way to expand homeownership is a productive exercise, the research finds.

“Overall, we find that even with the trade offs, this does have a potential of reducing the racial disparities and access to credit,” said Jung Hyun Choi, a senior research associate at the institute.

The research was designed to lay the groundwork for pilot programs that will be staged on a test-and-learn basis, according to Teresa Bazemore, president and CEO at the Federal Home Loan Bank of San Francisco. 

“You definitely want to be able to use this information to help people and not to create unintended consequences that are negative,” Bazemore said.

The FHLB plans to fund more research that will have an impact on these programs, including studies evaluating the efficacy of innovations designed to address student debt as a barrier to mortgage lending and another looking at whether artificial intelligence could be used to make more equitable loan decisions.

Credit scoring that accounts for alternative data in automated underwriting

used by government sponsored enterprises would help to further equity goals, researchers from the institute noted.

Fannie Mae and Freddie Mac, which back a significant number of loans in the United States, have long been reliant on a “classic” score from FICO. While the GSEs are reportedly close to accepting a Vantage Score model that would consider rent and utility payments, it hadn’t been confirmed at the time of this writing. The GSEs have, however, allowed alternative data to be considered apart from the credit score in underwriting.

While this opens up the credit box to some degree, it doesn’t address the barrier posed by a score limit, which is a prerequisite for loan eligibility and can affect secondary market pricing for mortgages, noted Michael Stegman, a nonresident fellow at the Urban Institute, visiting professor at Duke University, and a former senior policy advisor for housing at the National Economic Council during the Obama administration.

“You ought to be lowering the gateway or not using the gateway to block people from access to GSE loans if a full complement of data really shows good credit,” Stegman said.

The GSEs have said they need to move deliberately and carefully when it comes to this change.

Although the automation of alternative data in underwriting is opening the credit box up now, at one point, technology adoption did lead to a reduction in its use. Therefore, how it’s formulated and deployed must be carefully considered, noted Janneke Ratcliffe, a vice president at the institute who also was formerly assistant director at the Consumer Financial Protection Bureau.

“Rental history was something that was frequently used in the past, but with automated underwriting, it became too difficult. There were virtually no underwriters anymore who were going to go through boxes of checks and that kind of thing,” noted Ratcliffe, who also previously worked for mortgage company GE Capital and several nonprofit advocacy groups. 

“These new innovations I think have some promise to make [alternative data reporting] easier to opt into, to begin to automate more and more of these things, which haven’t been automated in this current iteration, but which, in the past, when we all did things manually, we used to look at,” she said.