Grid adjustment question

[Terrel L. Shields]

Regression has been around a long time and is still used heavily in the Aerospace Industry

Regression (specifically multilinear regression) should not be the issue. The issue is how can you spend a lot of time developing a data set that is large enough to do the job, or can we limit the variables and use a much smaller data set with a resulting analysis that is still good enough to reliably SUPPORT our adjustments. We cannot seek a universal truth with any real expectations because RE markets are imperfect. If the clients insist upon low fees, then the effort in using regression means the appraiser must work a select market area where they can add to, manipulate and analyze a large data set on a continuing basis, or rely upon a down and dirty shortcut method.

The promise of AVMs and the Zillow's of the world is that we can quantify housing in the same way we quantify the sale of gold on the gold market, or oil on the oil market or a stock. These are wrong. Gold is an instantaneous and liquid market. Houses side by side can sell for much different prices because the seller and buyer's motivations and the element of time, differences in appointment, the buyer's financial situation, and the skill of the selling agent. And if a couple, one important difference is the decision between them as to what they agree upon.

I think a lot of appraisers get hung up on making far too many adjustments. To me, 4 basic adjustments are justified. Size, effective age (which is condition and age combined), land value (as if vacant) and site or outbuildings. Adjusting small amounts for site view, location, etc. often leads to a central value that is little different from using less adjustments - in fact, can cause the range of value to spread and certainly increases the total gross adjustments and those gross adjustments are a measure of the reliability. The larger the gross adjustments, the less accurate the value is almost always going to be.

 

[J Grant]

Regression (specifically multilinear regression) should not be the issue. The issue is how can you spend a lot of time developing a data set that is large enough to do the job, or can we limit the variables and use a much smaller data set with a resulting analysis that is still good enough to reliably SUPPORT our adjustments. We cannot seek a universal truth with any real expectations because RE markets are imperfect. If the clients insist upon low fees, then the effort in using regression means the appraiser must work a select market area where they can add to, manipulate and analyze a large data set on a continuing basis, or rely upon a down and dirty shortcut method.

The promise of AVMs and the Zillow's of the world is that we can quantify housing in the same way we quantify the sale of gold on the gold market, or oil on the oil market or a stock. These are wrong. Gold is an instantaneous and liquid market. Houses side by side can sell for much different prices because the seller and buyer's motivations and the element of time, differences in appointment, the buyer's financial situation, and the skill of the selling agent. And if a couple, one important difference is the decision between them as to what they agree upon.

I think a lot of appraisers get hung up on making far too many adjustments. To me, 4 basic adjustments are justified. Size, effective age (which is condition and age combined), land value (as if vacant) and site or outbuildings. Adjusting small amounts for site view, location, etc. often leads to a central value that is little different from using less adjustments - in fact, can cause the range of value to spread and certainly increases the total gross adjustments and those gross adjustments are a measure of the reliability. The larger the gross adjustments, the less accurate the value is almost always going to be.

View or site adjsutments can be large - if that is what the market shows buyers re paying.

We can't decide to only adjust for four things to suit our own purpose and ignore what the market is valuing - yes, a range of value can be wider if we follow the market data; however, properly applied adjustments usually prevent a wide range though are times when it is unavoidable.

 

[David S]

Regression (specifically multilinear regression) should not be the issue. The issue is how can you spend a lot of time developing a data set that is large enough to do the job, or can we limit the variables and use a much smaller data set with a resulting analysis that is still good enough to reliably SUPPORT our adjustments. We cannot seek a universal truth with any real expectations because RE markets are imperfect. If the clients insist upon low fees, then the effort in using regression means the appraiser must work a select market area where they can add to, manipulate and analyze a large data set on a continuing basis, or rely upon a down and dirty shortcut method.

The promise of AVMs and the Zillow's of the world is that we can quantify housing in the same way we quantify the sale of gold on the gold market, or oil on the oil market or a stock. These are wrong. Gold is an instantaneous and liquid market. Houses side by side can sell for much different prices because the seller and buyer's motivations and the element of time, differences in appointment, the buyer's financial situation, and the skill of the selling agent. And if a couple, one important difference is the decision between them as to what they agree upon.

I think a lot of appraisers get hung up on making far too many adjustments. To me, 4 basic adjustments are justified. Size, effective age (which is condition and age combined), land value (as if vacant) and site or outbuildings. Adjusting small amounts for site view, location, etc. often leads to a central value that is little different from using less adjustments - in fact, can cause the range of value to spread and certainly increases the total gross adjustments and those gross adjustments are a measure of the reliability. The larger the gross adjustments, the less accurate the value is almost always going to be.

Agreed. Right now I only do the adjustment for GLA, lot site, condition ( the reviewer said don't do age adjustment alone since you have no way to set $500 or $1,000 per year. Just put this into condition consideration) , location (next to Frwy, busy road, power / cell tower, gas station, shopping plaza, etc. ) and other major component, like pool (even w or w/o pool, there is no price / value difference in my local market), owned solar pane, outbuilding (not ADU Yet).Set ZERO adjustment for deck / porch / balcony , fireplace and air condition and heating (of cause the house must have heating capability to meet local law. Otherwise will adjust as function obsolescence) since those few $K thing has no major impact to the multi-million house price / value here.

 

[Terrel L. Shields]

View or site adjsutments can be large

"view" is not the same as the "site" adjustment. And the site is no different vacant or with a building. So, it is baked into the land value cake.

 

[David S]

View or site adjsutments can be large - if that is what the market shows buyers re paying.

View adjustment can be a little hard since you can't enter into the comps house to feel the view. Can't totally depends on MLS pictures. Also view adjustment is $100K ~ $1.0M thing. Normally I didn't climb up mountain or drive to beach for the appraisal.

 

[J Grant]

"view" is not the same as the "site" adjustment. And the site is no different vacant or with a building. So, it is baked into the land value cake.

You could say that ajd it is wrong It is time for res appraisal.

Buyers pay a lot for view differences, and they do it when lot sizes are the same, so there is no site adjustment, and they pay for a view in a condo when there is no lot component.

 

[Tom D]

big urban. here, part of the view is the lot size and how close are you to that stinking neighbor who plays loud music. the view adjustment of a park or playground means less neighbors and more street parking spaces, but there is no line for st parking on the grid.

 

[J Grant]

big urban. here, part of the view is the lot size and how close are you to that stinking neighbor who plays loud music. the view adjustment of a park or playground means less neighbors and more street parking spaces, but there is no line for st parking on the grid.

you can create a line item category in one of the blank spaces on the bottom of the tried and type in st parking on it

 

[J Grant]

View adjustment can be a little hard since you can't enter into the comps house to feel the view. Can't totally depends on MLS pictures. Also view adjustment is $100K ~ $1.0M thing. Normally I didn't climb up mountain or drive to beach for the appraisal.

View adjustments - get th best inro we can, pore over MLS photos, call agents and ask, they might exaggerate ina listing but tend to be more honest if asked directly, also overhead plat maps show distance and other buildings in the way - despite the fact we can not see the view for ourselves, the market data reaction and premium paid for it is the main basis for the adjustment -

 

[Tom D]

you can create a line item category in one of the blank spaces on the bottom of the tried and type in st parking on it

yes you can, but i have been told from reviewers that CU doesn't recognize those lines. anyway, city parking has more to do with row homes with only 1 side of the street parking or 2 sides. and semi-det where their are less homes and less cars fighting for spaces. but this is beyond the thinking of any underwriter or CU, so i lump it somewhere else. just normal city life, now that a lot of families have more than one car, parking in front of your house cannot always happen. my car in front of my row home is a wonderful view.