The Consumer Financial Protection Bureau (CFPB) announced Wednesday that it may propose a rule to regulate the use of automated valuation models (AVMs) by lenders and appraisers.
The government watchdog is concerned that automated valuation models may reflect bias in design and function. Specifically, the bureau raised the concern that mathematical models may rely on biased data, resulting in inaccurate valuations.
The bureau added that without proper safeguards these models could digitally redline neighborhoods and perpetuate historical disparities.
To address those potential risks, the CFPB and a handful of other regulators, including the Office of the Comptroller of the Currency, the Federal Deposit Insurance Corporation, the National Credit Union Administration and the Federal Housing Finance Agency are considering proposing an interagency requirement that institutions establish policies, practices, procedures and control systems to ensure that that their AVMs comply with applicable nondiscrimination laws, the bureau said.
The Dodd-Frank Act gives the CFPB and other regulators the authority to implement rules and establish quality control standards for automated valuation models. In a Wednesday press release, the bureau listed several “options” for how it might do so.
The options include requiring random sample testing and reviews, ensuring a high level of confidence in the estimates provided by the AVMs and protecting against data manipulation. How the CFPB will address the two latter points remains to be seen.
Firms that develop automated valuation models have argued they might help to eradicate appraisal bias. CFPB Director Rohit Chopra disagrees.
“It is tempting to think that machines crunching numbers can take bias out of the equation, but they can’t,” said Chopra.
But prior to introducing a proposed rule, the CFPB, in accordance with federal law, must convene a small business review panel to get feedback from small businesses that could be affected by the proposal.
The bureau notes that small entities such as real estate credit companies, secondary market financing companies and mortgage loan brokers may be directly impacted.
The timeline for when the panel will take place is unclear. The CFPB said in an outline published on their website that within 60 days of convening, the panel must issue a report summarizing the input received from small businesses before the CFPB can propose a rule.
The bureau added that they are considering a 12-month implementation period after issuance of an eventual interagency final rule.
Fair lending enforcement and addressing appraisal bias have been top priorities for the Biden-Harris administration.
In July 2021, the Department of Housing and Urban Development announced that it would spearhead an inter-agency task force to make recommendations on how to address appraisal bias. The report is expected in early 2022.
The post CFPB weighs rulemaking on algorithmic bias in valuation models appeared first on HousingWire.