Appraisal education, practice, and standards have not kept up with changes in data, computation, and communication.
Scope of work is supposed to include the extent of:
- Property identification
- Property inspection
- Data researched
- Analysis used
There is a missing piece.
This piece is ignored, avoided, deflected, and denied. The substitute is “my experience, superior knowledge, and well-established education is best for picking comps”.
The USPAP Standard 1 (appraisal development) gets it right. It says that the appraiser must use all relevant, available information.
“Information” is data that has been filtered for relevance. “Available” is available, or accessible.
Let’s go back to what we learned in school: “More data is always better – except when it becomes garbage data”. This means that somewhere in the middle is the ideal. There is even an official data science rule for this. (See below.)
Well-established, proven, and “accepted” appraisal practice has worked well for decades. Even longer. Narrative reports were common, electronic spreadsheets were just coming into use. Mostly it was a table, typed by the typist.
On the residential side, it was a typed form perfectly suited to perfectly fit in a typewriter on 8 ½ X 11 paper. It was ideal. It fit four columns. One for the subject, three for the three comps. Perfect.
Everybody knew that one comp might be biased (or just have bad information). A second comp helped. The third comp settled things. If one of the three was wrong, the third voted with the first, and settled things. Credibility! Achieved!
This third comp – and the reputation of the appraiser — was all that was needed. And it was reasonable, given the challenge of tracking down comps from agents, or “sold” books, which were always four to six months behind the times. Nobody did time adjustments, even though the lack was rebuked and scolded (just like today), but somehow accepted. However, it was nice to cite the local newspaper’s financial writer’s financial writings.
Four or five comps was all that was credibly reasonable to expect. It was hard. 80% of the appraiser’s job was to gather comps. Phone calls, trips to the county offices, asking friends at SREA or AIREA monthly meetings – worked well.
“Data researched” was obvious. See them? All three. All five. Right there.
Today things have changed. Dramatically. But appraiser education, practice, and standards have not kept up. There is a missing piece. A piece which, unless acknowledged, ensures the continued decline of the appraisal profession to just a vocation — to an obstacle to make a loan or make a deal.
The missing piece is selection or documentation of “all available” information. And the analysis of “all information” necessary for credible results.
In the Data Science, market analysis, intelligence valuation approach – this is called the “Bias-variance” tradeoff. It is the Goldilocks and Baby Bear point between too little data and too much data. One creates analytical bias, the other creates lost sureness (precision).
This concept of bias-variance tradeoff and its analysis is a key element of EBV© (Evidence Based Valuation) as found in the SGDS1 class.