When Freddie Mac published the study, “Discovery of Racial and Ethnic Valuation Gaps In Home Purchase Appraisals” in late 2021, its conclusion, “that Black and Latino applicants receive lower appraisal values than the contract price more often than White applicants,” sent shockwaves through the mortgage industry. Then in early 2022, Fannie Mae turned to real estate valuation technology to conduct its own study on appraisal bias. The GSE used automated valuation models (AVMs) to look at divergent appraisal values for black and white borrowers refinancing their homes.
Fannie Mae’s study, while finding that homes owned by white borrowers were more often overvalued than homes owned by Black borrowers, furthered the discussion around Appraisal Modernization, including the use of AVMs in property valuations.
Upcoming changes to AVM compliance and regulation
In addition to the Fannie Mae and Freddie Mac focus on appraisal bias, federal banking regulators and the Consumer Financial Protection Bureau (CFPB), all of whom are on the Interagency Task Force on Property Appraisal and Valuation Equity (PAVE), continued their work to examine requirements around AVM development and use. AVMs faced intense scrutiny. Questions were posed such as, “are they, in fact, objective,” or “is AVM technology just as susceptible to valuation bias?”
Since the Dodd-Frank Act and some light updates, there have not been any major policy changes to valuation guidelines since 2010. While the government agency concluded that AVMs were a good solution to some of the mortgage industry’s challenges, it was suggested to explore if we are replacing one problem with another.
Recently, the CFPB published its outline of proposals and alternatives under consideration as part of its small business advisory review panel for AVM rulemaking. These proposals are designed to strengthen oversight of AVMs to ensure that computer models used to determine home valuations are accurate and fair. The options outlined by the CFPB are being considered by federal banking partners as well as lenders, and AVM service providers. The CFPB released its small business review report in mid-May.
Highlights of potential changes
While federal banking regulators are speculated to release regulations establishing quality control standards for AVMs this summer, here are a few key components they are expected to address:
Valuation Quality
- AVMs should be used to improve appraisal quality
- Guidelines on how AVMs are used in underwriting, post-origination and securitization
- AVMs should adhere to standards and rigor when predicting values
- AVM vendors should be transparent in their process
Valuation Equity
- AVMs should try to reduce bias, where possible
- AVMs should try and test for bias and disparate impact
- AVMs should be used in the lending process to prevent appraisal bias
New standards and guidelines will most likely be presented in a way that helps AVM providers ensure their model is built to satisfy these requirements. Even more importantly, the AVM provider will be empowered to communicate these guidelines to the user, so the consuming entities can also comply with these rules.
Clear Capital’s role in evolving AVM compliance
Between the timeline of proposed guideline changes and locking-in policy adjustments, Clear Capital is acting as a valuation technology subject matter expert to help policymakers better understand AVMs and machine learning. As guidelines change, so too will Clear Capital’s methodologies and documentation so that lenders and AVM users can easily meet industry standards.
In the meantime, with current compliance and continuous oversight, our automated valuation model, ClearAVM™, remains steadfast as a critical tool for property valuations. ClearAVM adheres to industry best practices in support of equality for all in the lending process.
How ClearAVM supports Fair Lending Laws
- Red-flag attributes such as ethnicity/race, national origin, religion, sex/gender, marital status, Public Assistance Income, Exercise of Right, disability, or familial status are not used in the ClearAVM model.
- The proprietary models and variables used in ClearAVM’s programming are routinely analyzed.There is continuous oversight of the model to ensure compliance.
- Clear Capital’s team of legal experts are always engaged in conversations with industry leaders to understand the nuances, economic factors, and landscape that contribute to AVM regulations.
In addition, ClearAVM pairs its machine learning with field-gathered property data on 139+ million properties nationwide to deliver accurate, objective valuations. You can try ClearAVM to see how it fits your workflow.
If you have questions about AVMs or want to see how ClearAVM performs, reach out! Clear Capital is a national real estate valuation technology company with a simple purpose: to build confidence in real estate decisions to strengthen communities and improve lives. And we are always looking to connect with like-minded people in the industry.
