Amid the housing market’s high prices and rates—with little relief in sight—there has been a push for creative solutions to ensure homeownership doesn’t become even less attainable. This is especially critical for Black borrowers and other borrowers of color, who still experience the effects of systemic racism in mortgage lending. Today, there is a 30 percentage-point homeownership gap between Black households and white households.
Our recent report explores how homeownership could be more attainable if mortgage underwriting included additional factors that better predict mortgage performance. We found that incorporating cash-flow data and rental payment information into underwriting could expand access to loans for borrowers with missing or thin credit profiles, lower mortgage costs for borrowers with low credit scores, and disproportionately benefit Black and Latino borrowers—ultimately helping to close the racial homeownership gap.
The benefits of using cash-flow data in mortgage underwriting
With more than 96 percent of American households having bank or prepaid accounts, cash-flow information is a promising option for improving data access and automated underwriting systems. These data—which include accurate predictions of consumer credit risk and ability to pay through precise measures of income, rental payments, utility payments, and other relevant variables found in consumer deposit and card accounts—offer a timely and comprehensive picture of consumer finances outside of traditional credit reports.
This cash-flow information could help generate dependable credit risk assessments of the 45 to 60 million US consumers who lack sufficient credit history to generate reliable credit scores. These consumers are disproportionately people of color, who have long lacked equal access to credit. The information could also supplement and improve risk assessment within the current credit system through a combination of new and existing models that lenders are already familiar with.
Though research is limited, compelling evidence found these separate cash-flow-information-based measures and scores were highly predictive of credit risk across a diverse set of populations for which loan-level performance data were available when tested.
The benefits of using rental payment data in mortgage underwriting
Our report investigates the effects of including rental payment data in mortgage underwriting. Analysis of denial rates by race on Home Mortgage Disclosure Act purchase loans in 2021 shows a clear gap in both application submissions and application approvals for Black and Latino borrowers.
Among the 4.2 million borrowers who applied for conventional purchase mortgages, white applicants accounted for 72.4 percent of all applicants (despite accounting for only 66.7 percent of all households), compared with just 6.1 percent of Black applicants and 12.9 percent of Latino applicants (despite these households representing 12.9 and 13.6 percent of all households, respectively).
Though they represent a significantly lower share of applications, Black applicants make up 14.2 percent of those who were denied; Latino borrowers make up 20.4 percent of denied applicants, and white borrowers make up only 58.5 percent.
These denial outcomes confirm the struggles Black and Latino households face in becoming homeowners. Using rental payment data from the Understanding America Study, we found that a disproportionately high share of Black and Latino households could be reapproved if positive rental payment history were included in mortgage underwriting.
Black and Latino applicants accounted for 15.0 and 20.9 percent of applicants among those who could be reapproved. This means that for every potential additional Black homebuyer, there are only four potential additional white homebuyers. This ratio is substantially more equal than the current ratio of more than nine white homeowning households for each Black homeowning household. But the actual share of reapproved borrowers is likely lower because lenders would not be able to obtain rental payment data for all borrowers who were denied because of their credit, and even those who do provide the data may still be denied.
Efforts to include cash-flow and rental payment data in underwriting
Both government-sponsored enterprises have recently started a pilot and announced plans to include cash-flow data in mortgage underwriting in their Equitable Housing Finance Plans to improve racial equity in mortgage underwriting.
Last August, Fannie Mae added a new feature in its Desktop Underwriter automated underwriting system that incorporates consumers’ rental payments in the mortgage credit evaluation process for first-time homebuyers. If the initial application evaluated by Desktop Underwriter is not approved for sale to Fannie Mae, the system will now check whether a 12-month history of on-time rental payments would change that outcome. If it would, Fannie Mae will inform the lender, who can then ask the borrower for Fannie Mae’s permission to access their bank statements through an approved vendor to identify recurring rental payments and ensure the confidentiality of the consumer’s highly sensitive information.
Before launching the initiative, Fannie Mae estimated that about 17 percent of first-time homebuyers who were initially ineligible would have been approved with a consistent 12-month rental payment history. Between September 2021 and May 2022, Fannie Mae assisted underwriting about 2,000 borrowers through this approach, and about half of those were borrowers of color. Though promising, the current number of borrowers who were actually able to take advantage of the program was substantially limited by the small number of vendors with third-party systems that meet Fannie Mae’s security criteria.
In July 2022, Freddie Mac announced a similar initiative to consider on-time rental payments in its loan purchase decisions. Through designated third-party service providers, lenders and brokers can submit 12 months of borrower-permissioned rental payment data identified in bank accounts to Freddie Mac’s automated underwriting system when assessing the borrower’s purchase eligibility.
Freddie Mac also recently announced the inclusion of bank account cash-flow data when reviewing borrower eligibility for purchase loans beginning in November 2022. Meanwhile, several financial technology companies are advancing the use of cash-flow data in mortgage underwriting. Guild Mortgage has launched initiatives in this space, including a pilot program that uses inflows and outflows of income and expenses data and measures of residual income derived from consumers’ bank accounts provided by FormFree. Guild plans to analyze these data to provide additional insight into borrowers’ ability to repay their mortgage to price Federal Housing Administration (FHA) mortgages for applicants lacking FICO credit scores while using the FHA’s manual underwriting guidelines.
Research shows that using alternative data such as rental payment history and cash-flow data from bank accounts in the mortgage underwriting process would allow for a better assessment of a borrower’s credit risk. It would disproportionately benefit those without credit scores—namely, Black and Latino borrowers—therefore helping to narrow the racial homeownership gap.
Although there’s important progress being made in the market, it will take time for these new data to be fully integrated into the system. In the meantime, more research on incorporating cash-flow information will be needed along with a focus on data collection, data standardization, and regulation for consumer protection and access to credit.