I am not sure if you have ever taken a class in "Statistical Analysis", but I have taken several and always get a kick out people trying to defend their values or estimates with little true basis!
What I mean by that is that with bad statistical analysis I can prove almost anything if you let me select the sample. If I wanted my analysis to show the number of children per average home should I sample the one bedroom condos or the five bedroom homes to illustrate the answer I wanted?
I have yet to see a real estate class taught that covers statistical analysis (other than what can be done in Excel or Access type programs). They should require sampling strategies, bell curve analysis, standard deviations from the norm, percent confidence levels, etc. In a "Bell Curve" your average (one standard deviation from the norm) only covers 68% of the entire sample. If you go out two, you at least cover nearly 96% of the population. In the chart below which illustrates the Bell Curve for IQ distribution, you will see that if your IQ is over 145, (4 standard deviations from the norm) you are truly smarter than 99.8% of the people and will be extremely frustrated with a world based on "normal" being at 100!

Too many real estate valuation models are based on easier methods like Price Per Sq. Ft. Some extreme examples of other industries in which that doesn't make sense include:
- Automobile - do you buy your car by the pound? Cost per pound of a Harley Davidson vs a Hummer H2 vs a Honda?
- Restaurants - Does Caviar cost more or less per pound than regular eggs? If it is all just "beef" why does prime rib cost more than hamburger?
- Art - why does a gold necklace cost more than the value of the gold alone?
- Clothing - Does Victoria's Secret charge extra or less for their skimpy little outfits?

There is a huge attempt to treat real estate like a commodity. Companies like Zillow admits their numbers are off (varies by market) at first it doesn't sound bad if the number is give or take 10% but that is 10% on either end so their range is really 20%. For example, a $200,000 home could be $180K-$220K! If the true value was $220K and the price they sold for themselves without a Realtor is the low end (since they thought they saved money) is $180K, they left $40K on the table! That equates to selling at an 18% discount yet they "saved" about 6%. They could have hired five Realtors for that with non exclusive contracts since only one would end up receiving the offer (5x3%=15% +3% =18%). Obviously, that is a weird example, but I wanted to prove my point on "bad math"!
Just for fun, the next time you do a CMA, compare the following:
- Average price per sq. ft.;
- Average sales price;
- Median Sales price;
- Median Sales Price;
- Mode Sales Price;
Figures don't lie, but liars figure, while others are just plain ignorant and speak the loudest!

Comments (34)Subscribe to CommentsComment