The Federal Housing Finance Agency (FHFA) addresses appraisal bias reflected in the recently released Uniform Appraisal Dataset (UAD) Aggregate Statistics Data File and Dashboard.  The UAD information is derived from more than 47 million appraisals conducted between 2013 and June 30, 2022.  As the name suggests, the UAD standardizes various data elements regarding an appraisal. Lenders selling loans to Fannie Mae and Freddie Mac must submit appraisal information in the UAD format. 

FHFA focuses on appraisals that report a valuation lower than the contract sales price of the home, which the FHFA refers to as an “undervaluation.”  The FHFA assesses the relative rates of undervaluation in census tracts in which Whites comprise at least 50% of the residents (White tracts), minorities comprise between 50.1% and 80% of the residents (minority tracts), and minorities comprise over 80% of the residents (high minority tracts). The FHFA notes that while “controlling for observable characteristics may explain some of the gap in undervaluation between White and . . . minority areas, it is not likely to explain all of the difference.”

The FHFA cites 2021 appraisal statistics that reflect a rate of undervaluation of 13.4% in White tracts, 19.2% in minority tracts and 23.3% in high minority tracts. The FHFA notes that, based on these numbers, the proportion of properties that are undervalued in high minority tracts is 74% higher than in White tracts, and the proportion of properties that are undervalued in minority tracts is 43% higher than in White tracts. The FHFA also notes that in 2021 on a national basis, 15.2% of appraisals were below the contract sales price, 26.7% of appraisals were equal to the contract sales price, and 58.1% of appraisals were above the contract sales price. 

The FHFA advises that:

“From a practical perspective, compliance departments of lenders and appraisal management companies could use the UAD Aggregate Statistics Dashboards to narrow the scope of an exam or compliance review related to appraisal bias.  For example, in 2021 in the Charlotte, North Carolina Metropolitan Statistical Area, reviewers may want to focus on minority tracts …which have a higher proportion of undervaluation than high minority and White tracts.”

The FHFA concludes by stating:

“This research note highlights a few ways that the UAD Aggregate Statistics Dashboards can be used to explore potential appraisal bias—using the ‘tract percent minority population’ property characteristic and the ‘percent of appraisals below contract price’ property statistic.  The gap in undervaluation is notable, and the new datasets may be helpful in better understanding the disparities.”