matched paired analysis

[joe caruso]

having trouble with matched paired analysis!


when i find the right two or three comps to compare how do i find the adjustment for sf, brick exterior, or age i can't seem to find the formula.

joe c

 

[Vegan702]

Appraising Residential Properties 4th ed has a good explanation on pages 339-342. You should pick that up if you are doing residential properties, or The Appraisal of Real Estate 13th ed is out now which is even better and goes more in depth in all aspects of appraising, not just residential.

 

[timd354]

having trouble with matched paired analysis!


when i find the right two or three comps to compare how do i find the adjustment for sf, brick exterior, or age i can't seem to find the formula.

joe c

In match pair analysis, you try to isolate one variable at a time and extract the market reaction for that variable from the market. Then you move on to the next variable and do the same thing and so on and so on. There is not a formula.

 

[Lee Lansford]

having trouble with matched paired analysis!


when i find the right two or three comps to compare how do i find the adjustment for sf, brick exterior, or age i can't seem to find the formula.

joe c

Great in theory, difficult in practice...matched-pairs, that is.

How often do you find 2 sales but with a SINGLE difference between the two? Even for, say, VERY SIMILAR houses, once the improvements have "aged", other factors (updating etc.) come into play.

When you find "the formula", let me know.

 

[Lee Lansford]

Appraising Residential Properties 4th ed has a good explanation on pages 339-342. You should pick that up if you are doing residential properties, or The Appraisal of Real Estate 13th ed is out now which is even better and goes more in depth in all aspects of appraising, not just residential.

I have the 3rd edition; are you referencing the section titled "Limitations of Paired Data Analyses"?

 

[Kali the Boston Terrier]

One thing that makes paired sales much easier, CSV or copy/pasting into excel. Sorting the fields allows you to look at large chunks of data and find paired analysis easier.

Another option is linear regression for large R2 type statistics like GLA and market conditions. Seeing that these items tend to account for 60% (GLA) of a property's value solving for these via a linear can take a lot of the guess work out of paired analysis.

Awhile ago on the forum Austin and Steve Santora played around with a concept called Multiple regression analysis, using a theory called stepwise. That is essentially borrowing the benefits from paired sale analysis, and applying it to multiple regression. Doing this usually tightens up the R2...just some food for thought.

BTW do not be discouraged when you do a paired sale analysis with different data sets, they rarely ever point to a definitive answer for an adjustment...welll at least in residential.

 

[fifteen]

He didn't say paired sales analysis. Use a simple approach, almost never less than 1/3 of your cost approach factor. It's tough enough without having to weed through what value does that tree have. I can figure it out, but who cares.
Paired sales analysis would be useful for market analysis of trends. Interpolate a sf adjustment factor? I have friends that narrow it down to 22.725 or 31.98. They usually have to explain it. This is not an exacting science. Ever done a review where a different adjustment factor was used for each comp? Now that's hilarious

having trouble with matched paired analysis!


when i find the right two or three comps to compare how do i find the adjustment for sf, brick exterior, or age i can't seem to find the formula.

joe c

 

[Randy Pippin]

Go online to M&S and build yourself a test house 5, 10 or 15 times. Each time only make 1 change and have everything else the same. Work your way through different bathroom counts, different size garages, bedrooms, fireplaces, porches, etc. It’s not perfect but you will have a true paired sales/cost of the same house. The M&S online cost of $10 is good for a week so there's no limit to the changes you can make.

 

[Michigan CG]

Go online to M&S and build yourself a test house 5, 10 or 15 times. Each time only make 1 change and have everything else the same. Work your way through different bathroom counts, different size garages, bedrooms, fireplaces, porches, etc. It’s not perfect but you will have a true paired sales/cost of the same house. The M&S online cost of $10 is good for a week so there's no limit to the changes you can make.

That is a really good idea!

 

[timd354]

Go online to M&S and build yourself a test house 5, 10 or 15 times. Each time only make 1 change and have everything else the same. Work your way through different bathroom counts, different size garages, bedrooms, fireplaces, porches, etc. It’s not perfect but you will have a true paired sales/cost of the same house. The M&S online cost of $10 is good for a week so there's no limit to the changes you can make.

That's a fine idea except for the fact that cost does not necessarily reflect market value or the market's reaction to the addition or subtraction of certain features. This is most easily demonstrated by the fact that the appropriate adjustment for GLA differences may be significantly different than the per sf cost to build a house.