Grid adjustment question

[David S]

Seems a simple question, we do grid adjustments for every appraisal all the time. For GLA and lot site, if we use regression method, every sq ft should be same. Like 1000 sq ft house sold for $1.0 m, which translate $1,000 / sq ft. Then, for a 1,100 sq ft comps, not often to apply $100K adjustment in the grid, right? Most I saw $30K~$40K adjustment? The same for the lot size. How regression method works here?

 

[sputnam]

Only... what we adjust for is not SF. We adjust for market reaction to differences in SF. Given two dwellings that are identical in every way, except for 1 SF difference in size, a typical buyer wouldn't notice. Where that break point is could vary... might be 50 square feet... might be 100. For a very large dwelling, it might be more.

 

[J Grant]

Seems a simple question, we do grid adjustments for every appraisal all the time. For GLA and lot site, if we use regression method, every sq ft should be same. Like 1000 sq ft house sold for $1.0 m, which translate $1,000 / sq ft. Then, for a 1,100 sq ft comps, not often to apply $100K adjustment in the grid, right? Most I saw $30K~$40K adjustment? The same for the lot size. How regression method works here?

Appraisers usually adjust for the residual, extracted contributory value of $ per sf of a differential, rather than raw cost per sf to build, or dividing the sale price by sf which is what RE agents do

I don;t use regression to derive adjustments, like anything else it is how it is applied, mechanically as a formula vs market derived reaction -

 

[Jerry Bone Jr]

Think of regression as a matched pair analysis.
y=mx+b
m=rise/run
m=y2-y1/x2-x1
y = price or ASP
x = SF
m = slope = the incremental $ per SF adjustment rate
Not simply sale price per square foot.

 

[Raven59]

When developing a model to estimate the value of a property, it's essential to select the features that directly impact the price. Square footage is one of those characteristics that significantly affects the property's value. To build a regression model, choose the dependent variable (sale price) and independent variables (significant features such as square footage) and estimate the coefficients for each independent variable. These coefficients represent the value adjustment per unit of change in the variable. For example, if the square footage has a coefficient of $1000, then each additional square foot adds $1000 to the property's value. You can validate the model by checking the goodness-of-fit measures, such as R-squared.

However, choosing the variables to include in the model can be tricky. It's essential to be guided by both statistical analysis and an understanding of the local real estate market to avoid omitting essential variables or including irrelevant ones that may lead to biased or inaccurate results. Overfitting the model to the data is another risk, especially if too many variables are included or the data set needs to be more significant. Overfitting occurs when the model is too complex, capturing noise rather than the underlying relationship, which can result in poor predictive performance on new data.

Furthermore, multiple linear regression assumes a linear relationship between the dependent variable (property value) and the independent variable (property characteristics). However, in reality, some relationships may be nonlinear, or there may be interactions between variables that a simple linear model cannot capture. Therefore, it's important to be aware of your model's assumptions and limitations and adjust accordingly.

Adjustments based solely on differences in floor space (measured in sqft) often fail to account for the impact of volume space, which can significantly affect a property's value. As mentioned earlier, it is essential to identify the threshold at which the market becomes sensitive to such differences. Based on your example, it appears that you are referring to the overall price of a property (including both the land and any improvements made) rather than just its sqft variable (property characteristics). Sorry if I'm missing the point of your question.

 

[RCA]

Seems a simple question, we do grid adjustments for every appraisal all the time. For GLA and lot site, if we use regression method, every sq ft should be same. Like 1000 sq ft house sold for $1.0 m, which translate $1,000 / sq ft. Then, for a 1,100 sq ft comps, not often to apply $100K adjustment in the grid, right? Most I saw $30K~$40K adjustment? The same for the lot size. How regression method works here?

I learned 20+ years ago when I started appraisal that linear regression simply does not work for homes in urban areas of Northern California.

This is what a true regression model will look like for GLA:

1. A base amount that represents the starting point for GLA, assuming your smallest home is a certain size such as 800sf.
2. A segmented linear regression is invariably the best representation for price increases or decreases. This has linear segments between so-called "knots" or change points. For example, assuming the minimum GLA is 800sf, $300/sf from 800-1500sf, $250/sf from 1500-2500sf, $200/sf from 2500-5000sf and $0/sf for over 5000sf. In this case the knots are 1500sf, 2500sf and 5000sf.
3. The model in #2 would look like this Value(GLA) = $240,000 + max(0,GLA-800) x $300 - max(0,GLA-1500) x $50 - max(0, GLA-2500) x $50 - max(0,GLA-5000) x $200
4. By the way the function max(x,y) means that you take the value of x if it is larger than y, otherwise y. So, max(0, -$200) is just 0 or in other words max(0,y) only take the value of y if it is positive.
5. For the given model, we add $300 for each square foot of GLA above 800sf - out to infinity. Then we subtract $50/sf for each sf above 1500sf out to infinity, then subtract another $50/sf for each sf over 2500sf out to infinity then subtract $200/sf for each sf over 5000sf out to infinity.

You can do this in Excel. If we put the GLA in column 1 the function for the value would be:

=240000+MAX(0,A2-800)*300-MAX(0,A2-1500)*50-MAX(0,A2-2500)*50-MAX(0,A2-5000)*200

GLA	GLA Value Contribution
800​	$240,000​
1,000​	$300,000​
1,500​	$450,000​
1,600​	$475,000​
2,400​	$675,000​
2,500​	$700,000​
2,600​	$720,000​
3,200​	$840,000​
4,000​	$1,000,000​
4,900​	$1,180,000​
5,000​	$1,200,000​
5,100​	$1,200,000​
5,200​	$1,200,000​

And in the Sales Grid we would have (in this kind of sales grid, there is an extra column for variable "Value Contribution" and the adjustments are simply Subject_Value_Contribution - Comparable_Value_Contribution:

Subj GLA	Value Contrib.	Comp GLA	Value Contrib	Adjustment
4000​	$1,000,000​	3200​	$840,000​	$160,000​

By the way, if there is really no demand for homes over a certain size in a market area, the value contribution for the extra GLA may actually decrease above a certain size. I have seen this for example, where heating costs are high --- necessitating closing off parts of the house. Then throw in additional maintenance costs - making the extra GLA of such homes a liability for many. And some women have a very negative view of having to clean those superfluous areas of the home.

 

[David S]

Thanks all!

Another thing, in my local area, most single home are around $2.0M ~$3.0M. Seems hard to do the pairs analysis for Air Condition and fireplace adjustment since even two similar comps, one with AC or fireplace, another without. Maybe there is about $10K ~$20K sale price difference, it is hard to say this $10K ~$20K sale price difference is because AC or fireplace or because GLA or lot size difference since no two comps have exactly GLA and/or lot site. So just don't do the adjustment or $0 adjustment? Some peer told me use cost method to apply $5K adjustment.

 

[David S]

Right now I set 50 sq ft for GLA adjustment threshold (for under 2000 sq ft house) and 200 sq ft for site adjustment threshold ( for under 10,000 sq ft lot).

 

[J Grant]

Thanks all!

Another thing, in my local area, most single home are around $2.0M ~$3.0M. Seems hard to do the pairs analysis for Air Condition and fireplace adjustment since even two similar comps, one with AC or fireplace, another without. Maybe there is about $10K ~$20K sale price difference, it is hard to say this $10K ~$20K sale price difference is because AC or fireplace or because GLA or lot size difference since no two comps have exactly GLA and/or lot site. So just don't do the adjustment or $0 adjustment? Some peer told me use cost method to apply $5K adjustment.

If I can't expect a smaller adjustment, especially on a house of that price range, I might reconcile it slightly higher or lower to account for the feature ( or lack of it ) and ex[explain why.

 

[RCA]

Thanks all!

Another thing, in my local area, most single home are around $2.0M ~$3.0M. Seems hard to do the pairs analysis for Air Condition and fireplace adjustment since even two similar comps, one with AC or fireplace, another without. Maybe there is about $10K ~$20K sale price difference, it is hard to say this $10K ~$20K sale price difference is because AC or fireplace or because GLA or lot size difference since no two comps have exactly GLA and/or lot site. So just don't do the adjustment or $0 adjustment? Some peer told me use cost method to apply $5K adjustment.

Well, the problem of course is that "pairs analysis,' as an appraisal method, is a retarded approach in most urban areas. Oh, for a number of reasons:

1. It means you find specific pairs to match in all respects except one and then calculate the impact of that difference on price. These are most likely not representative but subject to statistical anomalies. They are, in other words, not reliable by themselves. You need to look at larger sets of data and extract relationships that take in to consideration many different variable combinations.

2. If you have 6 important variables that impact price, it is unlikely you will even be able to find six matched pairs to find adjustments. Outside of newer subdivisions, and especially in older areas where there has been numerous updates and modifications, this is not likely, - in my experience.

3. The data is non-parametric and calls for non-linear and non-parametric (e.g. ranking) methods. The "organization" teaches only parametric and linear methods.

It is extremely surprising that the leading appraisal organizations turn a blind like to these obvious defects in their courses for indoctrinating residential appraisers. They are not all idiots of course. So, you might think they are lazy. Some are absolutely lazy of course, but some work day and night to become bona-fide "leaders" in the likes of TAF and the Appraisal Institute. The Appraisal Institute, I must say, does a lot of good on the one hand, but they are "missing a card" on the other hand. The net impact of the Appraisal Institute on the success of residential appraisal, is I would seriously have to say, very negative due to the lack of creative researcch they have attempted (none), poor editiorial policy on the part of "The Appraisal Journal", and the absense of any significant progress in advancing "the state of the art."