Special offer

Paired Sales/Matched Pairs - A flawed technique...

By
Real Estate Appraiser with Appraiser, Ken Rossman Cert General RZ3504

One or a few matched pairs is hardly conclusive proof of a market trend.  

Matched pairs are flawed as at technique and essentially meaningless because they can achieve ANY result you want them to by simply searching for a pair that supports your desired conclusion.

It is always a treat in litigation consultation/support to get an appraiser on the other side using paired sales to prove an idiotic (inflated or deflated) adjustment and then "pair"  his or her other sales for the same component or other components.

Years ago, when I did my narratives to earn my former SRA and IFA designations, I used homes in nearby Levittown (NY) for my subject properties. I had to go thru MANY MANY dozens of sales in each case to come up with two that could "prove" any given adjustment. This was in a vast sea of originally identical homes, probably as ideal a comp pool as could be desired.

While paired sales can be misused/manipulated, my primary objection to them is that they simply represent two random indications from a highly imperfect market that more often than not are meaningless.  

I did not make up my demo report analogy.   In both instances, in an extremely homogonous market, I had to test in excess of 50 nearly identical (in all but one or two significant differences) pairs to find one that supposedly "proved" any adjustment result that wasn't absurd...  

Unfortunately, paired  sale analysis assumes a much more efficient marketplace than that which exists in the real world, and therein lies the fundamental flaw in this method.   Stated differently it assumes that all market participants view all components (to value) equally with similar expectations and motivations that also rarely exist.  People buy what they buy for many different reasons (motivations).  Many times they buy in spite of a feature that has no particular value or use to them. Almost All home purchases involve some degree of settling for a multitude of factors that often can't be quantified.  

I personally try to find sales that have as high degree of similarity to the subject as possible and keep highly subjective adjustments to a minimum.  I spent many years as Realtor (before and during the early years of my appraisal practice) and nothing has been more useful to me than observing market participants reaction and their desire to pay for or not pay for different items...  I spend quite a bit of time on every assignment talking to the homeowners or purchasers and probing them about their likes, dislikes and desires.  I always inquire about the cost of recent improvements and significant amenities.  I ask if different features influenced their decision to buy.  I have many Realtor friends whom I frequently use as a sounding board when I come across something atypical.   

I wish the industry would divorce itself from paired sales analysis. It is an outdated and flawed method that has no statistical validity and bears no cognizant relationship with the market.

my 2 cts...

Comments (9)

Jesse Skolkin
Independent New York State Certified Real Estate Appraiser - Fresh Meadows, NY

Kenneth:

I wonder how many appraisers who claim to perform matched paired analysis actually perform such analysis.  I try to keep to the method you proscribe - utilize the most similar, most recent, most proximate sales (and listings) available in the subject market and keep the adjustments to a minimum.  However, the alternative methods available are also problematic, for example, the cost method, wherein a dollar of cost does not necessarily equate to a dollar of value (based upon the law of increasing and diminishing returns).

Regarding the Levittown market - it's not as homogeneous as it used to be, with the recent trend to expand or "dormer" the cape cod style homes which used to abound in that market.

Jun 08, 2009 06:07 AM
Kenneth M Rossman
Appraiser, Ken Rossman - Boynton Beach, FL
FL Certified General Real Estate Appraiser #RZ3504

In the late 80's / early 90's when I did the research for my demo's, there were 100's of sales of basic Levitt capes with no dormers or extensions. I tried to set up matched pairs for a one car garage and a deck/porch. The results were all over the lot depending upon which pairs were tested.

Jun 08, 2009 06:53 AM
David Mescon
DAVID B. MESCON REAL ESTATE APPRAISER AND CONSULTANT - Kailua-Kona, HI

Hi Kenneth,

It's so refreshing to hear someone else who isn't afraid to "question authority."  I won't go as far as to claim that paired sale anayisis is essentially meaningless, but my opinion is very close to yours.  Even in those rare situations where one element of comparison is isolated and there are no other elements to adjust for, I can usually find one or more other paired sales with divergent results.

Jun 10, 2009 08:31 AM
Tom Horn
Thomas Horn, Real Estate Appraiser - Alabaster, AL
Appraising The American Dream

Kenneth,

I agree with you.  There are way to many other factors involved in the real world.  It's my opinion that matched pairs only works within a market where all of the homes are extremely similar.  Thanks for making a great point.

Jun 15, 2009 12:19 AM
Michelle Tucci
LIRA Corp - East Islip, NY

Very true Ken.

Paired sales analysis is no longer realistic.  When a reviewer asks me for a paired sales analysis to "support" adjustments on the market grid, I cringe.  They just dont exist!!  And I have a nice lengthy statement to support that!

Jun 15, 2009 11:08 AM
Tim Bradley
Contour Investment Properties - Jackson Hole, WY
Commercial Real Estate Expert in Jackson Hole, WY

My master's thesis in 1987 was on the use of matched pairs analysis in appraisal. Thousands of data points, lots of statistical analysis, and at the end of the day my conclusion was : GIGO. Garbage In Garbage Out. Matched pairs is a beautiful theoretical concept, and virtually worthless in practice.

 

Just my $0.02, of course. :)

Jun 17, 2009 09:35 AM
Anonymous
mike mcgill

Does anyone have an example of a matched pair analysis adjustment grid?  I can't seem to find one and I know I have seem them in appraisal textbooks in the past. 

If you have one, please e-mail a copy.

Thank You,

Mike McGill, Appraiser in Boise, Id.

Oct 30, 2009 11:44 AM
#7
Anonymous
MrLOFT.ca
I am building a website that gives a property Estimate of Value using Paired Sales of the 3 Closest homes. It gives a pretty accurate time value adjustment. For this purposes i think paired sales is great!
Apr 22, 2014 01:15 PM
#8
Anonymous
Isaac

I utilize the Matched Pair Analysis all the time. It works perfect here in Southern California. I do agree that it has its flaws but it is not meaningless either. It is useful to an extend, but we are not all mathematicians, although I did take statistics in College I implement the standard deviation to support Square foot adjustments.

Oct 06, 2014 05:52 PM
#9