Analysis of the “Appraisal Gap” White Paper
- USPAP Third Exposure Draft - August 4, 2022
- Appraisers Should Voluntarily Follow ANSI for Desktops? - July 25, 2022
- Desktops Are Being Done WRONG! - July 12, 2022
My critical analysis of Freddie Mac’s “appraisal gap” white paper
Appraisers, this is not an easy essay to write. It is my critical analysis of an academic study put out by a GSE, which I have read.
In September 2021, Freddie Mac’s Modeling, Econometrics, Data Science & Analytics; Single Family Risk Management; and Economic & Housing Research groups released a white paper (Economic & Housing Research Note) named “Racial and Ethnic Valuation Gaps in Home Purchase Appraisals.”
The paper uses the term ‘appraisal gaps.’ The paper’s definition of the ‘gap’ is: the percentage difference between minority and White groups in the share of properties or applicants receiving “appraisal value lower than contract price.”
The people who produced this paper concentrated on two primary aspects relating to ‘race’ and value opinions in appraisals of new home purchases, oriented toward Black vs White racial differences. It spanned years 1993 through 2020, and used the 2010 census for the racial profiling done.
After this paper was released to the media and stakeholder groups, some reporting about it used the term ‘groundbreaking’ to describe it. I don’t share that opinion. Breaking is mostly about breaking the morale of appraisers, who for the majority of reports, do their work appropriately per the USPAP mandates and GSE (and other agency) guidelines and lender assignment overlays we deal with, based on how appraisers understand those criteria.
The first consideration in the paper was appraised value vs the Contract Price. The groups utilized a third party study done in 2017 by Calem, Lambie-Hanson, and Nakamura to basically imply that appraisers are deficient in their work if the appraised value is LESS than the Contract Price in minority Census Tracts. More about ‘tracts’ below.
Yet in a footnote to the paper, the groups say
“We acknowledge that the sale price is not always equal to market value, and we expect that in all areas some appraisals will report values lower than the contract price. However, research data indicate that a high percentage of appraisals are at or above the purchase price.”
So let’s see here, some appraisals have values lower than the contract price, while a ‘high percentage’ are at or above that metric. It also implies that some appraisals are AT the contract price. But the paper goes on at great length about appraisals in minority tracts which report an opinion less than the contract price. Seems to me that the contradictory statement in their own paper refutes their conclusion that appraisers are wrongly reporting values in predominately minority occupied areas.
There, of course, was no discussion of any type of report reviews or calls with the appraiser to analyze the comparables used against the subject properties in order to properly evaluate the report conclusions.
The second consideration used for their analytics is Census Tracts vs the population racial component of area residents – and only 4 races were considered – in those Tracts. The US Census office keeps detailed records about the racial components across communities.
It should be noted that the Census Tract number has been inserted onto the GSE report forms since 1986, when GSE’s took over appraisal report form production. The tract number relates to the subject’s street address.
The tract number is used for compliance review tracking by the lender and regulatory agencies who can cross check purchases in communities based on locations in a Census Tract.
This tract number is not something that most appraisers pay any attention to. I have asked many on a particular forum I frequent. Their responses were as I stated. Appraisers don’t check Census maps to determine where tract boundaries are relative to where the subject and comps are located. And appraisers certainly don’t consciously evaluate the racial mix of residents in tracts to determine which homes they might or might not use for comps in a particular assignment. (If they do, then they need to be removed from this profession.)
This paper also says that a Census Tract (which is a count of people) is the same as what we call a Neighborhood in reports. Really? The Dictionary of Real Estate Appraisal, 4th Edition, defines Neighborhood as “A group of complimentary land uses; a congruous grouping of inhabitants, buildings, or business enterprises.” (In my Webster’s Dictionary, ‘congruous’ means “Harmoniously related or combined; Appropriate, consistent.”)
USPAP, Federal Laws, and the GSE guidelines prohibit the use of ‘race’ by appraisers in determining which properties will be in a report.
Yet this Freddie Mac research paper over many pages of written explanations, exhibits, dummy variables, models and high level math statistics concentrates on racial components in Census Tracts to try to make the case that appraisers are deficient in how appraisals are done, particularly toward Black people.
The folks who have said this paper is ‘groundbreaking’ probably have not thoroughly digested it, read the footnotes, and may not even have a true understanding of the current appraisal process. But it’s been thrown up on the ‘you’re the bad guys wall’ hoping it will stick along with the already debunked Dr. Perry Brookings Institution diatribe.
At the end, this paper actually admits that the group really doesn’t have a complete understanding of the ‘gap’ that they so critically discuss. It says
“…the potential appraisal gap is worthy of considerable research.”
“…our analysis has not yet determined the full root cause of the gap…”
And it’s conclusion makes this statement:
“We are testing whether alternatives to traditional appraisals offer a more objective analysis of property value.”
I wonder how many of the “alternatives” will allow as much ‘racial profiling’ as this paper does.