Expanded HMDA data shows racial disparities persist in mortgage lending

An FDIC economist found that racial disparities in mortgage lending persist even after weighting costs and denial rates by the widest range of data in recent HMDA data.
Stephen J. Popick, Senior Financial Economist at the FDIC, reviewed the 2020 Home Mortgage Disclosure Act (HMDA) figures, which contain a richer array of credit risk and cost data than pre-2018 data. .
His paper published June 9 by the FDIC showed how racial and ethnic differences are shrinking. He wrote that the remaining differences could be biased, or that they could “reflect the importance that lenders place on non-HMDA reportable data fields, such as residency and job stability requirements, bankruptcies and previous entries”.
“Therefore, differences in loan underwriting or pricing decisions between population groups in this document should not be construed as evidence of discrimination in fair lending,” he wrote. .
Banks and credit unions have criticized studies using HMDA data from before 2018 that claimed to show intentional or unintentional racial bias on the part of lenders, saying the data lacked the detail needed to determine it. When data collection was expanded to include more of these relevant details, they complained about the regulatory burden and pushed to exempt smaller institutions from having to provide data.
As of 2018, HMDA data includes data used to make lending decisions, including applicant’s credit score, debt-to-income ratio, and loan-to-worth ratio. New data allowing for better price analysis included interest rates, discount points, lender credits, points and fees.
Prior to the expansion of the dataset in 2018, HMDA data did not include credit factors commonly used by lenders to make lending decisions. This meant that studies from before 2018 could only show raw, unadjusted differences between blacks, Hispanics, and non-Hispanic whites.
“While the researchers used HMDA data before the expansion to understand mortgage lending patterns, they could not easily account for differences in credit risk affected by these patterns,” Popick wrote.
Popick compared raw denial rates with those controlled for credit factors, lender characteristics (such as total origination volume), loan timing, and loan location.
He found that checks explained about 70% of the difference in rejection rates for conventional loans.
“The impact of these checks on the differences in denial rates in FHA loans is significantly smaller, although this likely reflects the fact that minority and white borrowers are more similar in FHA purchase loans than in loans. conventional and that the underwriting requirements are quite different between these two types of loans,” he wrote.
After controls, blacks and Hispanics had rejection rates that were still 2 to 3 percentage points higher than non-Hispanic white applicants.
Minority borrowers also paid more than whites for loans after checks. On a $200,000 30-year purchase loan, blacks paid $1,583 more than similar white borrowers for conventional loans and $542 more for FHA loans in the form of interest, points of discount, lender credits and fees. Costs were accrued over 30 years with payments adjusted to present value.
Popick’s study also found:
- Black borrowers paid $1,211 more for a conventional refinance with no withdrawal on a 200,000 30-year loan and $1,909 more for a conventional refinance with withdrawal.
- Hispanic borrowers paid $1,725 more for a conventional purchase loan of $200,000 over 30 years. However, cost disparities with whites were significantly lower on other types of loans than disparities between blacks and whites.
- For conventional purchase loans among prime borrowers (those with credit scores above 660), interest rate differences between minority and white borrowers in the same credit score category narrow as credit ratings increase.