Working Paper 24-06: Home Purchase Appraisals in Minority Neighborhoods

[TerryRohrer]

Seems to me that there is a problem inherent in defining a low valuation in terms of the contract sales price.

And yet that has become the standard measure of "appraisal accuracy" over the past two years. Give the biggest crooks the loudest bullhorn and have them repeat something ad nauseum to create a new standard.

 

[shrubberyvaluation]

Regardless of property type or location, I think the prospect of less well informed/advised buyers is due to the brokers acting differently with the data than appraisers act. The effects of that will be different when the quantity/similarity of the data is lower than when that quantity/similarity is higher. Complexity becomes an effect-multiplier.

If an appraiser is working to specs we should expect that they are "seeing" a lot more of the data than 95% of the brokers out there, and is routinely looking at more sales data in a week and specifically for value than the average broker is looking at/analyzing in a year. It simply isn't common for a broker to provide relevant comparables to an appraiser that the appraiser wasn't already aware of during their normal course of business. That is, if/when that appraiser is working to specs.

Harder for appraisers WILL also be harder x 2 for the brokers and buyers/sellers just because they're not trained to look at everything nor are they performing their own different-valuation processes all day, every day. They are trained to look for what it takes to make their deal - because THAT is their job - and because many brokers don't really understand how appraisers pick their comps. The flip side of the "data availability" coin is that easier for appraisers will also be easier for the brokers and buyers/sellers. More agreement despite their different search criteria and different modes of research and analyses.

Just from a logic and consistency and critical thinking perspective, "The effect dissipates when the data and analysis are easier" directly contradicts the appraiser bias allegation. Based on their own conclusions, the appraisers AREN'T coming in low when the data is easier, which would not be the case if appraisers were acting with that kind of bias. You basically cannot have that kind of "dissipates when its easy" in the same room as "appraisers as a group are racists bias".

Screwing up on the selective basis regardless of how easy the analysis is would accrue to a violation of the ETHICS RULE. By comparison, screwing up more on everything as the analysis gets harder accrues to the COMPETENCY RULE.

A good point about agents acting differently. You also have more inexperienced agents as the commissions are typically lower in lower income neighborhoods.

 

[shrubberyvaluation]

And yet that has become the standard measure of "appraisal accuracy" over the past two years. Give the biggest crooks the loudest bullhorn and have them repeat something ad nauseum to create a new standard.

The two metrics have been

A) contract price
B) AVMS

These are the gold standard in which appraiser's are measured for accuracy.

In the study that was done in Baltimore where they did a secret shopper type thing with fake homeowners both black and white they turned in the black appraiser to the state for being biased against blacks. The white appraiser's actually valued the homes where a fake black homeowner a little higher than with a fake white homeowner.

 

[George Hatch]

One advantage the AVMs have over the appraisers in the low-comparability datasets is they can throw more data at the problem from different competing neighborhoods. If there are 10 "more heterogenous composition" neighborhoods in town which are generally comparable to the subject's neighborhood in terms of composition and previous pricing trends the AVMs can add that data to their analysis whereas an appraiser basically can't; not without having already identified all the neighborhoods which fit into that tranch.

 

[George Hatch]

A good point about agents acting differently. You also have more inexperienced agents as the commissions are typically lower in lower income neighborhoods.

I don't think the "rookie" factor is even the dominant variable. Brokers at all experience levels have been trained on the intro to appraising module but almost none of them EVER get to using that training to the level of detail in their instruction.

How many times have you asked a broker how they got to their list price only to be met with an explanation that starts with "we feel that....". Because that's nowhere near how an appraiser would answer that question. Even when the brokers or their buyers do know what they're talking about it is usually the result of them doing all their work in their head, not on paper and actively trying to work toward a more refined range of value indicators.

 

[Tim Hicks (Texas)]

Seems to me that there is a problem inherent in defining a low valuation in terms of the contract sales price.

I believe that there are some appraisers on this forum that believe if there is a contract price that is proof the market acceptance of that price and the market value. They always forget the part about the "informed buyer," and the fiduciary responsibility of the appraiser to the lender and the public. Oh, and the appraiser's responsibility to follow USPAP in the appraisal process.

 

[Elliott]

I looked over the 'working paper' and what struck me is the amount of regurgitation of numbers, over slicing and dicing. But a couple of fundamental problems with the analysis, again its based on 'census tracts' and appraisers don't define neighborhoods by census tracts and every report in the last twenty years should name the streets. So it would be beneficial for the authors to do the work right. Second, there seems to be a conclusion of 'correlation' but I couldn't see a stated R2 (so flawed statistical analysis) or hiding the validity of the analysis).

And as Judges are wont to say, "Appraisal isn't science, its as much art as science."

 

[Tim Hicks (Texas)]

At what point did lenders decide that the appraisal opinion of value is theirs to decide if acceptable. Shouldn't it be if you don't like our opinion of value, then you are free to get another opinion of value instead of crucifying the appraiser whether justifiable or not. That is our job to render an opinion of value. You can trust it or not, but don't make it a crucifying offense to render a value you don't expect. Now, completing a fraudulent appraisal or misleading appraisal is a different story. But, that typically is not the case when the opinion of value is lower than the contract price. Some lenders understand and trust their appraisers and some simply do not.

 

[George Hatch]

I believe that there are some appraisers on this forum that believe if there is a contract price that is proof the market acceptance of that price and the market value. They always forget the part about the "informed buyer," and the fiduciary responsibility of the appraiser to the lender and the public. Oh, and the appraiser's responsibility to follow USPAP in the appraisal process.

Seems to me that there is a problem inherent in defining a low valuation in terms of the contract sales price.

1 Buyer + 1 Seller is how the brokers define MV. Not the appraisers. So this whole controversy starts off on the wrong foot by comparing appraiser performance to the brokers' definition of value instead of the appraiser's definition of MV. *Sometimes* the results of the two different definitions can be the same, especially in a very competitive market. But other times the results will be very different, like when there's a lot variability in the quantity and comparability of the data.

If we were tasked to use the same definition of value the brokers use then we wouldn't have this problem. Nor would there be any demand for appraisals.

I don't think anyone here believes that the contract is proof of anything other than what the one buyer and the one seller came to terms on. However, the gap between that contract vs the appraisal is what all the critics are studying so that's really the only basis upon which a response can be meaningful to the question that's being framed.

 

[George Hatch]

I'll take the quantity/comparability issue one step further and ask the question if its even possible for the R2 factors to be similar between the increased quantity/comparability datasets vs the decreased quantity/comparability datasets.

Can any appraiser get to the same confidence level when the data is hard that they can get to when the data is easy? Can any broker or buyer do any better than an appraiser in that low quantity/comparability market area?

I will go so far as to say it can't be done. Not on any consistent basis and when using the same amount of time/effort to do the research and analysis.