Guest Author: David A. Braun, MAI, SRA has been appraising 35 years. He is a certified USPAP instructor and an Appraisal Institute qualified instructor. David has been developing automating tools for appraisers for over a decade.

Myth #1. You have to learn regression or go out of business. No one knows when or to what degree your clients will embrace regression analysis as a viable valuation technique. However, it is a recognized valuation tool that is available for use today. The appraiser should decide when and if he/she wants to begin learning and training in this technique.

Myth #2. You can become competent in regression analysis in one to two weeks. It will take months to become competent enough to really apply regression analysis to real world assignments. When taught on a cumulative basis it is relatively easy to learn, but it is still a foreign subject to most appraisers. If the training method is for you to suit-up, study some statistical formulas, and then start the engine and do some low altitude loop-to-loops you are in the wrong classroom. It is best to start with some simple tasks and then progress in difficulty.

Myth #3. It is easy to extract the line-item adjustments with regression analysis but difficult to form the final opinion. In reality it is just the opposite. It is usually much easier to form an opinion of the final value than extracting line-item adjustments. Some adjustments are indeed easy to extract, but many require advanced knowledge and training to extract. Some cannot be extracted at all depending on the data-set used. When extracting adjustments the culprit is usually multicollinearity, but it tends to aid in the process of finding the final value by decreasing the number of property characteristics that need be considered in the analysis.

Myth #4. The coefficient is always the line-item adjustment. When using regression to only form a final value opinion the coefficients will only occasionally be appropriate to use as the line-item adjustments (again this is because of multicollinearity).

Myth #5. Regression is just another form of the comparative sales grid method. Maybe a little, but in general this is not true. Regression, like the comparative method, is a form of the “direct comparison approach”. So, both methods could be used independently of one another in the same appraisal. When only a few recent sales are available and the subject is very similar to the sales (requiring only a few small adjustments) then the comparative approach is going to be more reliable. When there are many sales (say 30 or more over the past two years) and the subject is not very similar to the sales (requiring many and sometimes large adjustments) then the regression will have the edge. I foresee appraisers using both methods in the same appraisal with the comparative approach being more qualitative much like the 2055 form is now. In this case the regression would be the more quantitative of the two approaches

Conclusion: I encourage appraisers to begin studying regression analysis over a long-term basis. This spreads out the study, practice time, and the monetary costs as well. Do it to become a better appraiser, not because “you fear you have to or go out of business”. The sum of the knowledge gained by studying both methods, comparative (sales grid) and regression analysis is definitely greater than the sum of the parts. This is because everything you know about one method increases what you know about the other.

I am providing an addendum from one of AVT’s training Workbooks as an example of how more can be learned by studying both methods. It discusses proper comparable selection for the comparative and the regression methods of the direct comparison approach. These analyses dictate what a “good” comp is. Understanding why the comp is good for one or the other analysis method explains a lot about how to choose your comparables in general. I believe that today many appraisers are second guessing their comp selection. While the practice of selecting comps can be daunting, all appraisers should understand and be able to explain the theory behind their choices.

  Download Comp Selection - AVT Workfook Chapter XIII - David Braun

Go to www.AVTtools.com for more information about incorporating regression analysis into your appraisal practice.

Author: David A. Braun, MAI, SRA has been appraising 35 years. He is a certified USPAP instructor and an Appraisal Institute qualified instructor. David has been developing automating tools for appraisers for over a decade. Contact him with questions or comments at the above website.