Will there be a need for traditional “appraisal” in 5 years? We think not.
Some 25 years ago, I was education chair for my Appraisal Institute Chapter. Our ‘custom’ seminar considered the topic of technology replacing appraisers. The conclusion was, especially for residential appraisal, that AVMs would take a great market share. Worse yet, because they were cheap, they would drive down fees.
One well-qualified speaker made it clear that AVMs could do some things better than a human, and they would get better as data and computer power grew. Some of this happened. AVMs do take a large market share today. Yet human appraisal seems to persist! Why? And how does this trend shape up?
Are we nearing the peak of AVM accuracy? And what about the valuation ‘substitutes’, including ‘data collectors’ and ‘hybrids’ and ‘evaluations’? And how about Artificial Intelligence?
No need for appraisers? Yes, I’m convinced — no need for appraisals. [“Please stay to the end.”]
Yet there will be a need for valuations . . . and the risk for loans, investments and equity … and forecasting. Yes, all those appraisal substitutes are there created because of perceived deficiencies in the appraisal product. And many perceived deficiencies really are because of real, actual inadequacies in PVO products (Point Value Opinion).
But what won’t be needed?
What will not be needed is subjective opinion. What will not be needed is a point-value with no hint of its dependability. An appraisal is defined as an opinion! “Worthy of belief!”
What won’t be needed is “trust me” comp selection and “supporting” the user’s expectation (as required by USPAP)! It has to feel safe. What won’t be needed is claims of “adjustment support” via magical methods, false formulas, cheat sheets.
What won’t be needed is justification (reconciliation) of why the different “approaches” don’t match up.
What is needed?
What is needed is an objective result. A result which can be replicated.
What is needed is a reliability score, based on the appropriateness of the purpose to the intended use of the analysis. What is needed is user-friendly understandability of the data considered, and the specific MS© (Competitive Market Segment). And the resulting trend, its velocity, and a forecast. What is needed is interactive dashboard delivery. A dashboard allowing a reviewer, underwriter, investor, or equity enforcer the ability to enhance decision-making.
All this is possible today. We have the computer power. We have the ability to model and interact with algorithms via human-understandable visualizations (graphs and maps). We can clarify the underlying economic and human behavior elements from an uncertainty-recognized viewpoint.
The good news is the learning path is now available. With EBV (Evidence Based Valuation)©, we set the foundation for all of this. The Valuemetrics curriculum focuses on the building blocks for the future. Our CAA (Community of Asset Analysts)© is one mutual-help resource. Some others, (such as the 10,000 Minority Appraisers Foundation), are building their education standards, and best practices around the future of data analysis methods.
Yes, there is a need for competent asset analysts, quickly.
Start now.
Paul Rayburn
May 1, 2024 @ 5:27 am
Touching on hot topics, mainly biases, will require all valuation entities to ensure their members understand what an appraisal should look like.
The “art” of appraising is not about an appraiser’s “artistic license”, it’s about the art of painting a “complete picture”.
The pressure for valuations to be “restorative” will place many appraisers in high pressure situations. There has probably never been a more important time to remove opinion from the equation and replace that with evidence, not just some hand picked evidence, all of the evidence.
James Scholl
May 1, 2024 @ 6:08 am
No issue with what you wrote. The issue is the use of fake standards like ANSI. If you or anyone can, please explain how you compare a subject with a second floor ceiling height of 6’11” with the comparable sales when you do not know the ceiling heights of the comparable sales? How do you choose comparable sales without this information? What are appraisers actually doing? How does this affect the quality of the Fannie / Freddie portfolio? There is no source for the second floor ceiling heights. No one is willing to take on ANSI.
Tom Stowe
May 1, 2024 @ 7:28 am
I was at an Appraisal Institute dinner meeting in the early 90s where the guest speaker was some senior vice president for Washington Mutual (at the beginning of their rise to power). He said that within five years they wouldn’t need any residential appraisals and soon after any commercial appraisals. I am still here, they are not.
Vince Slupski
May 6, 2024 @ 11:12 am
The aging and declining majority of appraisers simply doesn’t have the mathematical, statistical, and programming skills to create appraisal data models, and today’s clients won’t buy them from appraisers anyway. Younger people with those skills will work for the GSEs and big data AVM providers, creating cheap and fast AVMs that will fall under appraisal waivers. If the clients want it, AVMs will include graphs and reliability measures. This kind of thing can be programmed for orders of tens of thousands of valuations, but won’t be economic for the one-off products of individual appraisals. The 100-year-old model of appraisal has plenty of inertia, but as GSE requirements diminish in favor of hybrids and waivers, the industry will go the way of travel agents – some specialty niches will remain, but the US will have perhaps a third of the appraisers it had at its peak.
William Jones
October 10, 2024 @ 10:25 am
I wonder who will ensure that the data on which the AVM’s are relying are accurate, and who will be responsible if the result is based on inaccurate data obtained from an unreliable source. Other AVM models may be able to replicate the results of the first AVM, but just because they can be replicated won’t mean they are reliable or credible. A fun exercise is to take existing AVM’s and use them to estimate the value of your own home. You will be surprised at the variance.