On June 1, 2021, President Biden directed HUD to lead a "first-of-its-kind interagency initiative to address inequity in home appraisals."²

The interagency task force was established on Property Appraisal and Valuation Equity (PAVE).³ The PAVE website states its purpose as "an equitable path toward addressing the persistent misevaluation and undervaluation of properties experienced by families and communities of color."

The U.S. Appraisal Standards Board (USPAP) announced Ethics Rule and Advisory Opinion 16 on March 1, 2021,⁴ which the USPAP stated had been extensively rewritten regarding the non-use by appraisers of any factors based on any protected group's characteristics "or any unsupported conclusion that homogeneity of such characteristics is necessary to maximize value." (Emphasis added.) The Ethics Rule states "In summary, USPAP requires appraisers to know the law and obey it." "Bias is never permitted."

The Appraisal Foundation (TAF) states:

Unsupported, unexplained, or generalized terms [may indicate bias or illegal discrimination]. Therefore, appraisers should take great care in the selection of words and phrases in their narratives. Examples of neutral language that avoids use of subjective non-relevant, non-specific, generalizations might be to write "synthetic" rather than "man-made," "adults older than 65", rather than "elderly", or "one-unit housing", rather than "single-family."

On February 4, 2022, all of the federal regulators sent a letter⁵ to the TAF demanding a change in wording. These regulators want the Ethics Rule to definitively state that "an appraiser's use of or reliance on conclusions based on protected characteristics, regardless of whether the appraiser believes the conclusions are supportable, constitute illegal discrimination." (Emphasis added.)

Thus, prudential regulators are on record that no subjectivity is allowed for appraisers to rely upon protected characteristics in reaching valuation opinions — and they uniformly contend that existing law already prohibits such utilization.

This report was released by Freddie Mac on September 20, 2021:⁶

•The focus is on appraisal values versus the contract prices for home sales. It does not look at the valuation appeals process or how this might improve appraisal accuracy and the absence of bias.

•It concludes that "our preliminary modeling results suggest that a property is more likely to receive an appraisal lower than the contract price if it is in a minority tract". Yet, the study says that "our analysis has not yet determined the full root cause of the gap." The conclusion says that "we are testing whether alternatives to traditional appraisals offer a more objective analysis of property value" but provides no finishing analysis.

•It contends that UCDP data from 2015-20 show that Latino buyers received appraisals below contract prices in 15.4 % of cases and Blacks in 12.5% of cases — versus Whites in 7.4% of cases. It does not explain the gaps of 8% for Latinos or 5.2% for Blacks. It does not ask why this gap metric is 35% higher for Latino buyers than for Blacks.

•It claims the robustness checks on its data show that the gaps "mostly persist ... and seem pervasive", but again no details.

•Regarding the differences in the comparable houses used by appraisers, it concludes that "there is more variation in the sales price of comps used in appraisals for properties in Black tracts than White tracts, but there is less variation for Latino tracts [when compared to White tracts]." (Emphasis added). Indeed, the Black variance is 2.4%, but the Latino variance is at minus 1.6%. No explanation is provided for these data anomalies. No suggested reasons are proffered.

•The conclusions in this study have been directly challenged by the American Enterprise Institute (AEI) Housing Center Report issued in November 2021 and revised in January 2022 arguing that the "gaps" can be explained by variations in socio-economic status, first time home-buyer inexperience, and a concentration of FHA lending in certain census tracts.⁷

On January 26, 2022, the National Fair Housing Alliance ("NFHA") published its report "Identifying Bias and Barriers, Promoting Equity: An Analysis of USPAP Standards and Appraiser Qualification Criteria".⁸

This analysis was commissioned by the Appraisal subcommittee of the Federal Financial Institutions Examination Council ("FFIEC"). This 83-page review focused on the appraisal process with a goal of identifying any systemic racial bias and disparate impacts on "people of color."

Four major areas were identified for reform:

(1) re-examination of the governance structure of the appraisal industry and exploration of limits on the authority of the private Appraisal Foundation and the states;

(2) clarifications in fair lending requirements and training protocols absolutely prohibiting any discriminatory bias including phraseology in the free-form text section of appraisals;

(3) removing barriers to entry for minorities into the appraisal profession and lowering requirements for immediate injection of appraisers of color; and

(4) release by the GSEs of appropriate elements of their proprietary appraisal data sets for public input, the extension of appraiser's duty of care for "intended users" to include borrowers for any professional negligence; and finally, amending the value appeals process to be more fair, transparent and uniform.

The FFIEC regulators appear to be hard at work looking for ways to implement these reforms.

On February 26, 2022, Fannie Mae published the first of what will be a series of reports on appraisal accuracy: "A Closer Look at Divergent Appraisal Values for Black and White Borrowers Refinancing Their Homes".⁹ The study selected 1.8 million refinance transaction appraisals in 2019-2020 because borrowers would typically have contact with the appraisers and biases could arise. Fannie Mae used its two separate proprietary AVM models to compare against the appraisals and deemed a 10% upward disparity an overvaluation and a 10% downward disparity an undervaluation. Fannie utilized census tract majority population indicators from the American Community Survey 2018 and defined Black as including borrowers who solely identified as such — plus combinations of Black and any other race or ethnicity (e.g., Black/Hispanic). White borrowers, however, included only those who self-identified as solely white and did not include any combination of whites with other ethnicities.

These are the major points.

•The statistical differences between appraisal values and AVM values "are indicative of possible bias but not conclusive." Homes owned by White borrowers tended to have appraisals slightly higher than AVMs regardless of the demographics of the neighborhoods — Black borrowers tended to have slightly lower appraisals than the AVM calculation.

•By utilizing its proprietary risk assessment model — Collateral Underwriter — Fannie Mae identified comparable location variance as the top reason for overvaluation by appraisers and comparable selection (dissimilar properties) as the top cause of undervaluation. Other important factors identified included market adjustments and total space GLA adjustments for overvaluations. Market adjustments and condition quality adjustments were listed causes for undervaluations.

•The report ends with five recommendations: (1) increase the use of alternative-scope valuations like desktop appraisals and hybrid desktop/site visit appraisals (to remove bias); (2) continue to build/revise risk assessment software like Collateral Underwriter; (3) continue to modernize the appraisal process; (4) foster diversity in the appraiser workforce; and (5) enhance the tools appraisers use to complete and test their valuations.

On February 23, 2022, The Bureau announced this topic of new rulemaking for the use and control of AVMs.¹⁰ The authority the CFPB relies upon is section 1125 of FIRREA (1989) [12 U.S.C. § 3354] which delineates that AVMs must provide a high level of confidence, must not be subject to manipulation, must avoid conflicts, must pass random sampling/review and must satisfy anything else deemed "appropriate." The Bureau is considering either a flexible principles-based rule, which can vary from entity to entity, and it is also considering a prescriptive approach mandating specific protections for everyone. The timetable for implementation would be one year after a final rule is published.

At the heart of this rulemaking is the determination of which entities in the mortgage lending sphere would be covered by an AVM rule and in what situations.

While this passage is still being debated, the Bureau has indicated its initial preferences as outlined below:

•Covered AVMs are those which officially determine collateral worth and does not include use of AVMs for marketing or portfolio monitoring purposes.AVMs used by an already certified or licensed appraiser in the appraisal process would be excluded so long as appraisers are not employed by or in a third-party relationship with an originator or secondary market issuer.

•Post origination use of AVMs for loan modifications or other changes to existing loans would be excluded unless the refinancing transaction resulted in a new mortgage.

•Servicers would ordinarily be exempted from AVM rules unless they are engaged in decisions regarding what would constitute a mortgage origination.

•AVMs used for Home Equity Line of Credit (HELOC) credit line reductions or suspensions would be excluded, but use for increases in lines of credit would be covered.

•AVMs used to determine whether an appraisal waiver is or will be granted for the loan (waiver to be granted either by a GSE, a guarantor, insurer or underwriter), would be excluded, but the GSEs use of the models would be covered (as the substitute).

•AVMs used by "mortgage originators" depends on the Bureau's adopted definition — which might include only loan originators or may more broadly pick up all creditors [see Reg Z and the Truth in Lending Act (TILA)].

•AVMs used by "loan originators" would include mortgage servicers when they perform loan origination activities or refinancings changing the debt obligor.

•The definition of covered "secondary market issuers" could be broad and potentially include not just RMBS issuers, but also guarantors, insurers, underwriters and appraisal management companies.

•AVM rules applicable to mortgages might be broadened to include not just loans secured by a dwelling, but also consensual security interest loans and installment land contracts.

•Unlike TILA (which only applies to use of proceeds to primarily for personal, family, or households purposes); the AVM rule would apply no matter the use of proceeds.

•AVM rules would apply to dwellings whether or not secured by real property- this includes mobile homes and chattel/ which may be personality under state laws.

•AVM rules for new construction loans and homes under construction would apply if occupancy is to be within one year.

•AVM rules would "create an independent requirement for institutions to establish policies and procedures" in their usage.

As the Bureau recognizes, the technology underlying AVM models is evolving so quickly, that a do's and don'ts rule approach may not be practical for the dynamic environment. This militates in favor of flexible rules which allows different policies, practices, procedures, and control systems, for various institutions with dissimilar business model and risk profiles.

There is uncertainty about what is coming.

•Are there any critical data points necessary for accurate property valuation on a track to be disregarded because they are labeled as potentially carrying bias — and if so, what are they?

•Are the appraisal bias studies performed to date accurate, directionally clear, and ripe for the process changes being discussed?

•Will the Appraisal Institute and state licensing boards cede their longstanding oversight roles to federal regulators without a fight over Constitutional authority?

•Will the increased regulatory burdens on smaller lenders, fintechs and non-depository players deter innovation, increase costs ultimately passed to borrowers and further complicate an already confusing housing finance industry?

These are complex issues. But, like it or not, a new course is being charted. Diligence is imperative.

1 Tex. Dep't of Hous. & Cmty. Affs. v. Inclusive Communities Project, Inc., 576 U.S. 519 (2015).

2 https://bit.ly/35wOFn1

3 https://bit.ly/3JdtILL

4 https://bit.ly/3wVqCJw

5 https://bit.ly/3uNyOcb

6 https://bit.ly/3iYHVBw

7 https://bit.ly/3u1yMOY

8 https://bit.ly/3u0ejKa

9 https://bit.ly/3K5d6ae

10 https://bit.ly/3NDJOBz