How Do I Make Adjustments In Square Footage

[Patrick Eubanks]

Thanks Carole!!

Your suggestion helps to shape my thinking as to how I should approach this issue. I don't want to be a "skippy" and just put any kind of adjustment on the form without having "rhyme or reason" how I arrived at the dollar value. What I'm finding out during this discussion is that what works in one situation (area) might not be the normal method somewhere else. I just need to adopt a standard that is acceptable so that when I get called to testify as to how I arrived at "this $ adjustment", I will be able to explain it in court .

 

[Austin]

Can anyone define what a properly adjusted sale looks like? For example, how do you know when adjustments are needed and when no further adjustments are needed? How do you know when size matters and when size does not matter. If size doesn't matter some times but some times it does matter as someone stated, then if you use parired sales how do you know if size matters or if it doesn't? Have you ever seen a cost manual where smaller houses cost the same as larger houses built to the same quality of construction rating? Have you ever seen a cost manual that based cost on room count or a meter to indicate when an improvement becomes an over or under improvement?

 

[jeff samolinski]

And thus the reason for using comparables that are bracketed by their size.

 

[Jungle Boy]

Around here 10 / sf adkustments are applied only to piece of crap properties.

 

[Fred]

Patrick,
The problem with your mentor’s “method” is that it is completely arbitrary and fixed. The size-price relationship is different within each data set.

Outside the realm of homogenous products and perfect markets
1. no rule of thumb can substitute for a knowledge of mathematics and use of sensible analysis of a sufficient sample of data (pairing is of limited use, graphing and regression are better).
2. size-price relationships are rarely linear over an extended range (so, in my view, adjusting all sales at the same sf/rate builds-in error).
3. three sales are not going to be enough

Economic theory suggests that a buyer will pay more for more size, but up to a point. Each extra dollop of more size contributes less to overall value than the preceding increment, until a point is reached where more size adds nothing and begins to detract. What that means is that 100 sf difference cannot be adjusted at the same price/sf as a 1,000 sf difference within a given data set.

Let’s apply that to a simple sales set.
Sale 1 is 2,000 sf for $100k.
Sale 2 is 2,200 sf for $110k
Sale 3 is 2,500 sf for $120k

Looking at sale 2 v 1, the addition of 200 sf added $10,000 or $100/sf (by pairing).
Looking at sale 3 v 1, the addition of 500 sf added $20,000 or $40/sf (by pairing).
Looking at sale 3 v 2, the addition of 300 sf added $10,000 or $33/sf (by paring).

This shows why the single “fixed” size adjustment cannot fit all sale pairs. If your subject is just like sale 1 and 2,000 sf, you would need to adjust sale 1 at $0/sf, sale 2 at $100/sf, and sale 3 $40/sf. Then all your sales would indicate $100k. If you try averaging those paired differences ($58/sf), then the results will be self-skewing. Sale 2 will adjust to $98k and sale 3 will adjust to $91k.

Austin,
There is no point in discussing size adjustments without you.

Can anyone define what a properly adjusted sale looks like? For example, how do you know when adjustments are needed and when no further adjustments are needed?

When the residuals are zero!

 

[Austin]

One comment:

I agree with Steven but would add that what Steven is said for one quality level of house. The problem is that larger houses have higher quality with more extras of higher quality thus the curve trend line over a wide range of properties. Then too, if you can't bracket all property features you can't measure them without using even more complicated methods, but it can be done-for a fee.

 

[Restrain]

If you're concerned about adjustments for square footage, try this. Adjust for everything else, then look at the differentials that are left and pair them for size differential. You'll find significant differences across your market so there's no hard and fast rule.

Roger

 

[tommcsherry]

Paired sales analysis is great in theory, but the times I've tried to use it have convinced me that it's basically useless...I've seldom found two homes that were equivalent enough to use...there ALWAYS seems to be some difference (besides size) that potentially factors in. So...for a while now, I've been wanting to look at regression and other types of statistical analysis....problem being, I don't know statistics...duh...

So, I was wondering...does anyone know of any regression anaylsis software out there that has been developed specifically for appraisal use? Hopefully one that wont cost an arm and a leg....

Thanks!

 

[DB]

There are many, but one of the best is contained within Excel, provided you are using Windows .... From what I have been told, one of the best folks to ask that question is Austin .... I keep hoping that he will help me learn to do RA ... Of all the folks here, He is supposed to be one of the most knowledgable on the subject of RA ...

 

[Austin]

Roger:

Good point-sometimes. Believe it or not, many times you can make one adjustment that will adjust for every thing at once. This is because value factors are correlated to each other. What I do is to make the size adjustment first, which is not really size only because it reflects covariance of value factors, then see what the graph looks like. If the slope of the trend line after doing that is some out of the range number such as having the subject and comps at around $75 per square foot including land, and the trend line slope being something like $150 or -$30 per square foot, then you know something needs an adjustment. If the slope of the trend line is $78 per square foot, then one adjustment at that amount will generally average out all other value influencing factors. It is all related to comparability in general. I have my own little program to guide me through the sequence of adjustments. Every time I make an adjustments, the trend line on the graph moves to tell me where I am going. When the variance about the trend line is reasonable I then make a size adjustment that makes the trend line have a slope of zero which means the set is perfectly adjusted.
When you finish your marketing grid graph the results with GLA as the X-axis and price as the Y-axis. Calculate the trend line and if the line is perfectly flat and all points are on the trend line, that defines a perfectly adjusted data set. If the trend line is not flat, the set is not adjusted. About 90% of appraisals I see are no more adjusted after the appraiser finishes than it was before he or she started adjusting. All they are doing is playing shuffle board.