Pick comps, make adjustments are the “recognized methods and techniques” which must be “correctly” used by appraisers.
Originally, the practice was just “find comps and reconcile.” Some hundred years ago, the main task of an appraiser (usually also the local broker) was to track down sales. This involved phone calls to agents, people they knew, and a drive to the county courthouse to scribble notes. (Copy machine – nope).
Who you knew was as important as what you knew. Knowing your area mattered. Judgment of “better or worse” required personal credibility. “Credible = worthy of belief” was the underlying principle of lender expectations and, later, the writing of appraisal standards by the earlier appraisal organizations: AIREA and SREA (American Institute of Real Estate Appraisers, and the Society of Real Estate Appraisers).
Credibility, “belief worthiness,” was all we could hope for. “Data” was not a word heard. “Computer” was a fictional future figure.
Where did “adjustments” come from? Many appraisals went to lenders. Some had appraiser reviewers. Some believed some comps were “not truly comparable.” And needed to be “adjusted.”
“Don’t you have any better comps?” they asked. “Here’s one more,” the appraiser replied. “But this one is also off, but the best I can do.” So the appraiser discussed the main difference and said, “If I adjust for that feature, it’s a good comp!” “OK,” said the reviewer, “but in the future, can you just do the adjustment up front, so I don’t have to ask?” “OK,” said the appraiser. And adjusted adjustments.
Time passed.
A new, smarter review appraiser said: “You adjusted?” “Why is your adjustment so big? It’s not worthy! How did you calculate it? You have to ‘support’ your opinion! Find a way!”
And so it was. And so it is.
Then things happened. We got comp books. We got internet. We got instant electronic data. Yay! Now we could just consider all (or nearly all) the sales which seemed to matter.
What used to take days, took just minutes from books, and — then just seconds on the computer!
Pull up sales, control your search parameters (based on your experience, training, and familiarity with that market). Hopefully you get 10 or 20, pick 6 or 8, and use 4 or 5 for your report. Throw away the rest! Done!
Oh yeah. Make some adjustments. Not too big, not too odd. Make them “just right.”
Technology came.
Then we got computer power, and easy, free analytic software (data science and AI). And we got telling visuals, graphs, dashboards, and simple statistics. Power. The ideal blend of the human expert and repeatable computer algorithms. Power. Art with science!
Best of all: we discovered that the importance of (dubious and unprovable) adjustments became less and less relevant! Or just plain un-necessary!
We do explainably adjust for “cash equivalency” and for time. However . . . other adjustments cannot be mathematically, explainably calculated (given the usual “population” of similar sales).
For those of us who use data science methods, the reliance on the “right data” in quality and amount – minimizes or eliminates other adjustments. The analytic result becomes near-obvious, easily explained, reliable, and reproducible.
August 28, 2024 @ 3:20 pm
The first residential article written for the SREA Journal instructed appraisers to watch the newspapers for reports of sale and put the information in a file to use them later.
A topic I have been thinking on is “Do Buyers or Sellers make adjustments to sales?”.
If we are to be mirroring the actions of buyers and sellers, and they do not adjust sales, then why do we?
I do many reports using appropriately selection selection criterion for the location, and physical attributes, examine a dozen or two of similar properties that sold, are pending or listed for sale on a $/SF basis. Next I consider how the subject relates to the data as far as terms, time, location and physical attributes. Then I conclude to a value per Square Foot for the subject.
If the is note requiring concessions because the market is strong and going up, I conclude above the mean property.
If the property is strong I conclude above the mean.
If the property is just average, I conclude at or near the mean.
If the market is weak or declining, I conclude below the mean.
If the location is weak I conclude below the mean.
If the property is weak, I conclude below the mean.
Assuming my selection criterion is appropriate I usually a fairly narrow standard deviation above and below the mean.
I think in terms of 1/4 steps to 1/2 steps, to one full step above or below the mean.
The beginning of my doing this started with the SOW Rule. The first time I did it for a long term client, I got a call and he said “This is the best report I have ever read, it is so clear and concise.”.