How home appraisal reforms aim to reduce bias

October 31, 2022 1:37 PM

Enjoy complimentary access to top ideas and insights — selected by our editors.

Fairness in lending continues to be top of mind for regulators, with a number of authorities creating new transparencies with the goal of eliminating discriminatory practices and appraisal bias, including the Federal Housing Finance Agency's recent announcement that it will make public appraisal data from Fannie Mae and Freddie Mac.

Meanwhile, stakeholders in the industry are taking a proactive approach and implementing changes to systems and processes, such as Black Knight's upgrade of its automated valuation model and the FHFA's new framework to review and recommend changes to automated models to advance fair lending.

Read our roundup to learn about these developments and other efforts to end appraisal bias and discrimination in the industry.

The seal of the Federal Housing Finance Agency (FHFA) is displayed outside the organization's headquarters in Washington, D.C., U.S., on Wednesday, March 20, 2019.

Andrew Harrer/Bloomberg

FHFA to make Fannie, Freddie appraisal data public

Looking to provide more transparency into home valuations, the Federal Housing Finance Agency is making available to the public the Uniform Appraisal Dataset records compiled by the government-sponsored enterprises.

The data is drawn from 47.3 million appraisal records collected from 2013 through the second quarter of 2022 on single-family properties in a manner that protects borrower privacy. In addition, FHFA is offering UAD Aggregate Statistics Dashboards on its website to provide user-friendly visualizations of the newly available data.

At the Mortgage Bankers Annual convention in Nashville, FHFA Director Sandra Thompson announced these new features as part of the regulator's efforts to reduce appraisal bias.

Read more:

FHFA to make Fannie, Freddie appraisal data public

Test-survey-study-online-computer-work-strategy

Increased study on how automated valuation models impact appraisal bias

The question of whether automated valuation models are a useful aid in reducing appraisal bias and promoting equitable lending remains unclear after two research studies returned contrasting results.

A Veros Real Estate Solutions study testing automated valuations in Chicago found no signs of racial bias, with median absolute AVM errors "statistically insignificant" at one basis point or less. In contrast, an Urban Institute study in Atlanta and Memphis revealed an error rate of as much as 5 percentage points.

The widely different results leave mortgage companies in a quandary as they await ratification of proposed CFPB rulemaking on the use of AVM technology in property appraisals that would require strict compliance with nondiscrimination laws.

Read more:

Two reports present very different pictures of appraisal bias

Upgraded automated valuation system aims to cut bias in comps

Black Knight is upgrading its automated valuation system with an optional add-on to its CA Risk Profiler, which determines the overall valuation risk of an appraisal, in response to the federal government's efforts to eliminate racial bias against protected classes in home-buying and refinancings.

Appraisers typically use comparable market prices to support their estimate of a property's value. However, some borrowers believe the process can discriminate against minorities by utilizing comps that are based on a person's race, prompting government action. Black Knight's new CA Risk Profiler Plus aims to fix that.

"U.S. policymakers are actively working to address valuation bias and its impact on America's racial wealth gap," said Mike Sklarz of Black Knight Collateral Analytics. With "greater process scrutiny and regulatory action in the near future" a very likely prospect, Black Knight has acted quickly with the revamp of its automated model.

Read more:

Black Knight's new offering examines appraisal comps for bias

Appraisal bias reform proposals gain bipartisan support

White House proposals seeking to tackle appraisal bias against Black and Latino homebuyers, which was highlighted in Fannie Mae and Freddie Mac studies on fair lending, have gained bipartisan support, but the Senate Committee on Banking, Housing and Urban Affairs want to see more data.

"These reforms would increase the cost of the appraisal process and that means homeowners and homebuyers would have to pay that if we're going to do that," said Sen. Pat Toomey, the top Republican on the committee. "I think that should be based on reliable data."

Toomey pointed out that the two reports did not explicitly blame appraisal discrepancies on racial bias, but only suggested it was a key driver. Committee Democrats had less concern with the findings, but agreed that more data was necessary before proceeding with the reforms.

Read more:

Senators call for more data backing appraisal bias proposals

New National Fair Housing Alliance framework seeks to advance fairness in lending

The drive for fairness in automated decision making has been boosted by a new framework developed by the National Fair Housing Alliance, which can identify the underlying assumptions and limitations of algorithmic tools used in automated underwriting and risk-based pricing.

Designed for regulators, businesses, researchers and other stakeholders, the alliance's Purpose, Process and Monitoring system examines automated models and produces recommendations to advance fairness and reduce privacy risks.

"NFHA's PPM framework provides an equity-centered auditing solution at a time when policymakers and civil rights organizations are calling for fairness, accountability, transparency, explainability and interpretability," said chief tech equity officer Michael Akinwumi.

Read more:

A.I. decisioning audit structure launched by housing group

Human loan officers responsible for bias fintechs say

CFPB concerns that loan decisions based on artificial intelligence may breach laws against discrimination in lending could soon lead to a raft of new legislation. The fintechs and banks on the cutting-edge of the use of AI in lending broadly speaking agree, but are quick to point out the upside of technology.

"AI is not perfect," said Jason Altieri, general counsel and chief compliance officer at Happy Money and former general counsel at LendingClub. "If you have humans designing something, bias will creep in." This tendency for bias, however, makes AI less biased than human loan officers, Altieri believes.

"For years you'd walk in, ask for a loan, and they would look at you and with absolute bias say, Do I know you, do you have an account here and do you look like me?" says Altieri. With AI, "you're just a series of data points," making loan decisions based on AI more fair than those of traditional underwriting.

Read more:

Unfair lending with AI? Don't point just at us, fintechs and online lenders say