Announcing the 2024 Tech Trendsetters winners.

Read Now
Inventory
info icon
Single family homes on the market. Updated weekly.Powered by Altos Research
721,576-14142
30-yr Fixed Rate30-yr Fixed
info icon
30-Yr. Fixed Conforming. Updated hourly during market hours.
6.95%0.01
Appraisals and ValuationsRegulatory

Federal agencies seek input on proposed AVM credibility, integrity rule

The agencies include FHFA, the CFPB, the FDIC and the Federal Reserve

Six federal agencies have requested comment from the public on a newly-proposed rule that is designed to “ensure the credibility and integrity of models used in real estate valuations.”

The proposed rule would also implement quality control standards that govern the use of automated valuation models (AVMs) used by mortgage originators and secondary market issuers in valuing real estate collateral securing mortgage loans.

The agencies involved include the Federal Housing Finance Agency; the Consumer Financial Protection Bureau; the National Credit Union Administration; the Federal Deposit Insurance Corporation; the U.S. Department of the Treasury; and the Federal Reserve System.

“Under the proposal, the agencies would require institutions that engage in covered transactions to adopt policies, practices, procedures, and control systems to ensure that AVMs adhere to quality control standards designed to ensure the credibility and integrity of valuations,” the announcement states.

The intent is to create a set of standards that increase confidence in the use of AVMs, according to the announcement.

“The proposed standards are designed to ensure a high level of confidence in the estimates produced by AVMs; help protect against the manipulation of data; seek to avoid conflicts of interest; require random sample testing and reviews; and promote compliance with applicable nondiscrimination laws,” the agencies said.

In a blog post released with the joint announcement, CFPB Director Rohit Chopra said AVMs have the potential to cause serious harm if the algorithms are inaccurate or biased.

“While machines crunching numbers might seem capable of taking human bias out of the equation, they can’t,” Chopra said. “Based on the data they are fed and the algorithms they use, automated models can embed the very human bias they are meant to correct. And the design and development of the models and algorithms can reflect the biases and blind spots of the developers.”

This could make bias more difficult to filter out, since algorithms can “cloak the biased inputs and design in a false mantle of objectivity,” he said.

Ensuring that standards are in place allows these tools to be used with greater confidence, and ultimately reduces costs associated with appraisals, according to the announcement.

“While advances in AVM technology and data availability have the potential to contribute to lower costs and reduce loan cycle times, it is important that institutions using AVMs take appropriate steps to ensure the credibility and integrity of their valuations,” the agencies said. “It is also important that the AVMs institutions use adhere to quality control standards designed to comply with applicable nondiscrimination laws.”

According to the proposed rule, these quality control standards would only apply to the use of AVMs in determining the value of collateral.

“The proposed rule would implement the statute by applying the quality control standards when an AVM is being used to make a determination of collateral value, as opposed to other uses such as monitoring value over time or validating an already completed valuation,” the proposed rule states.

Other AVM uses, including portfolio monitoring, are not designed to determine collateral value and would not be subject to the rules.

Comments must be submitted within 60 days of the proposed rule’s publication in the Federal Register.

Leave a Reply

Your email address will not be published. Required fields are marked *

Most Popular Articles

3d rendering of a row of luxury townhouses along a street

Log In

Forgot Password?

Don't have an account? Please