The Premier Online Community for Real Estate Appraisers!
-
Welcome to AppraisersForum.com, the premier online
community for the discussion of real estate appraisal. Register a free account to be able to post and unlock additional forums and features.
Regression for GLA Adjustment
- Thread starter NJAPRAZE
- Start date
-
- Tags
- adjustments regression
- Status
RCA
Elite Member
- Joined
- Jun 27, 2017
- Professional Status
- Certified General Appraiser
- State
- California
Depending on the neighborhood ... many variables can correlate to GLA, e.g. bathroom & bedroom count, garage car bays, lot size, .... The only way to extract the value contribution for GLA is through an advanced regression program like MARS. MARS does essentially a kind of sensitivity analysis - modifying parameters and then observing how the sum of squares of errors is impacted - only many thousands of times in a very logical and strict fashion. --- But it has a learning curve.
J Grant
Elite Member
- Joined
- Dec 9, 2003
- Professional Status
- Certified Residential Appraiser
- State
- Florida
The purpose of applying adjustments is to make the comp more equivalent to the subject and applied adjustments will narrow the adjusted value range of the comps.
Whether it is done with a sensitivity analysis right there on the comp grid by extraction or by a chart or graph or by regression, there is no way to arrive at a perfect one-perfect benchmark of an "accurate" adjustment.
The adjustments are supposed to represent an approximate market reaction to a feature or defect. If the adjustment fails to do that then it is a fail, regardless of the method or what fancy charts are in a report. Buyers adjust with their wallets - they pay more for a positive feature and expect a discount for a negative. Several methods that complement each other can provide a strong support.
For example, does the $ per sf adjustment have a reasonable relation to $ per sf depreciated cost ? Is it reasonable relative to the quality and size of the dwelling? Is it contributory value, or raw cost per sf ( contributory value should be the adjustment ) . Many appraisers don't adjust for minor differences per sf. A buyer does not really care if a house is 2000 sf or 2040 sf, They will care if it is 1800 sf vs 2000 sf.
Whether it is done with a sensitivity analysis right there on the comp grid by extraction or by a chart or graph or by regression, there is no way to arrive at a perfect one-perfect benchmark of an "accurate" adjustment.
The adjustments are supposed to represent an approximate market reaction to a feature or defect. If the adjustment fails to do that then it is a fail, regardless of the method or what fancy charts are in a report. Buyers adjust with their wallets - they pay more for a positive feature and expect a discount for a negative. Several methods that complement each other can provide a strong support.
For example, does the $ per sf adjustment have a reasonable relation to $ per sf depreciated cost ? Is it reasonable relative to the quality and size of the dwelling? Is it contributory value, or raw cost per sf ( contributory value should be the adjustment ) . Many appraisers don't adjust for minor differences per sf. A buyer does not really care if a house is 2000 sf or 2040 sf, They will care if it is 1800 sf vs 2000 sf.
Tom D
Elite Member
- Joined
- May 22, 2015
- Professional Status
- Certified Residential Appraiser
- State
- Pennsylvania
i really feel bad saying this, but with my big city work & lots of close comps, sometimes the only adjustment i have to make is GLA. and it's usually because i have to bracket the GLA
& even then not a big $ amount..
certainly the data set you have does affect the amount of adjustments. easy to get a good grade on your report. in my young days i did do suburban & rural suburban, so i did pay my dues to have my current privileged appraiser life.
& even then not a big $ amount..
certainly the data set you have does affect the amount of adjustments. easy to get a good grade on your report. in my young days i did do suburban & rural suburban, so i did pay my dues to have my current privileged appraiser life.
Terrel L. Shields
Elite Member
- Joined
- May 2, 2002
- Professional Status
- Certified General Appraiser
- State
- Arkansas
The problem is that sometimes we have a problem with double dipping adjustments. If you adjust for bedrooms, can you justify adjusting for SF at the same rate as you would otherwise? Other than using regression, I don't see how one extracts both a bedroom and a SF adjustment with confidence.
The same for a "quality" adjustment and then a separate adjustment for "granite countertops", "fireplace", "brick siding", and "crown molding". And I have seen people do that. I will never adjust for those individually. I might consider the aggregate as a quality issue. And you get into a real statistical conundrum when you apply too many adjustments. Each adjustment compounds the uncertainty. I've seen as many as 14 or more adjustments made between 3 comps, in about 7 categories. That's likely random adjustments made without real support just to target a certain value.
Jerry Bone Jr
Senior Member
- Joined
- Feb 23, 2004
- Professional Status
- Licensed Appraiser
- State
- Oregon
chad hampton
Senior Member
- Joined
- Nov 10, 2006
- Professional Status
- Certified Residential Appraiser
- State
- North Carolina
What do you do when all of your statistics come up with a result that you know is incorrect based on your expertise in the local market? experience is everything.
J Grant
Elite Member
- Joined
- Dec 9, 2003
- Professional Status
- Certified Residential Appraiser
- State
- Florida
My issue with statistics is it needs a volume of properties to work, and a number of those properties will not be comps for the subject and then the question is did that skew results....
Using the comps a buyer really would consider is about quality vs quantity. Plus buyers do not make a chart or graph, they have a rough idea of what they want to spend- it is not that hard to study the patterns and figure it out from the activity itself, verify with RE agents and other support.
Using the comps a buyer really would consider is about quality vs quantity. Plus buyers do not make a chart or graph, they have a rough idea of what they want to spend- it is not that hard to study the patterns and figure it out from the activity itself, verify with RE agents and other support.
TerryRohrer
Elite Member
- Joined
- Aug 13, 2005
- Professional Status
- Certified General Appraiser
- State
- Montana
What is the difference between a paired sale and a simple "regression"/best fit line with the same data? Graphically, they will be identical and the regression line would suggest an R2 of 1.00, perfect. So, by relying on two or 4 or six pairs will in both methods suggest nearly identical results. The regression analysis includes some measures of how sound the results are, whereas those using the same limited data for paired sales or sensitivity analysis just assume greatness without question.
Results improve with multivariate regression. In my experience, those most vocal and adamant against "statistics" and "regression" are the least likely to have learned how to apply it and gauge the results (to the point, in my opinion, of not even having tried). There is a lot of ground between all and none, but current trends suggest that appraisal work is showing heavily toward the all stats approach. Of course, "they" probably don't have a clue, either.
Results improve with multivariate regression. In my experience, those most vocal and adamant against "statistics" and "regression" are the least likely to have learned how to apply it and gauge the results (to the point, in my opinion, of not even having tried). There is a lot of ground between all and none, but current trends suggest that appraisal work is showing heavily toward the all stats approach. Of course, "they" probably don't have a clue, either.
J Grant
Elite Member
- Joined
- Dec 9, 2003
- Professional Status
- Certified Residential Appraiser
- State
- Florida
Can you create meaningful regression results using only 4-6 properties? Or using 10 properties?
My reports typically have 4-6 comps ( these days 5-6) but in addition I analyzed and review more sales and listings in addition, how many depend on the assignment. So although the 4-6 comps are what show up on the grid, 8-16 additional sales /listings/ pending contributed to the analysis . (typical number, could be more or less after elimination of others )
The typical buyer does nto run regression, so staying closer to their thought process and analyiss more closely can replicate/understand their motivations in the market -USPAP also has a section (paraphraisng it here ) where the appraiser , if they use statistics or AVM or other computer program methods should be able to replicate results if asked - how many can do that ?
If a person just feeds the data in and lets the machine do the work, where is the understanding -(not saying that represents any one person here, more of a general question )
My reports typically have 4-6 comps ( these days 5-6) but in addition I analyzed and review more sales and listings in addition, how many depend on the assignment. So although the 4-6 comps are what show up on the grid, 8-16 additional sales /listings/ pending contributed to the analysis . (typical number, could be more or less after elimination of others )
The typical buyer does nto run regression, so staying closer to their thought process and analyiss more closely can replicate/understand their motivations in the market -USPAP also has a section (paraphraisng it here ) where the appraiser , if they use statistics or AVM or other computer program methods should be able to replicate results if asked - how many can do that ?
If a person just feeds the data in and lets the machine do the work, where is the understanding -(not saying that represents any one person here, more of a general question )
- Status