The Frustrating Case of the Junk Statistic

Will someone help me to stop tearing the now diminishing quantity of hair out of my head?

Why are such vast numbers of journalists, agents, brokers (and even the occasional government official) head-over-heels in love with changes in Median Sales Price?  And ALWAYS as a means of talking about the direction of home values within a market?   These earnest people seem desperately to WANT a change in Median Sales Price to actually BE a measure of Home Appreciation yet it just isn’t.

I grabbed an example in 5 seconds.  Here is the highest Google page rank for “change in median sales price for real estate in Miami”:

The median home value in Miami is $301,600. Miami home values have gone up 5.2% over the past year and Zillow predicts they will rise 0.2% within the next year.
— Zillow

Let’s be clear, this is the way that Median Sales Price is typically is used.  Zillow just happened to have the #1 page rank and so they get called-out first.  Nevertheless, for a content creator claiming authority to use changes in Median Sales Price the way Zillow uses it here is a little like advertising all-beef hot dogs that are actually made with 90% pork - it obscures the point in a pretty harsh way!


It’s not that Median Sales Price is uninteresting, it’s just misapplied.  I suspect that these people WANT to tell their readers (or their agents, or their clients) the amount that homes in the subject market have appreciated or depreciated, but what they ACTUALLY tell them is a pretty deeply obscured story about the changes in THE MIX OF PROPERTIES THAT WERE BOUGHT/SOLD.  I am reminded of a bit of humor from the world of finance.  Jonathan Clements famously wrote in The Wall Street Journal that for mutual funds “Past performance may be be a guide to future results. But it’s a mighty tough guide to read.”

Here is a wonderful and head scratching - yet typical - example from USA Today dated 8/10/2010. It’s a profile of how the Topeka, Kansas property market was reacting to the 2007-2008 financial meltdown…

But home sales are still improving. In June [2010], Topeka-area sales were 11% above a year earlier. In June [2010], Median Sales Price - $105,500 - was up 1% from a year earlier. That was not a dramatic increase. But it was better than the first five months of the year, when prices dipped modestly.  The [Federal] tax credit drew in more first-time home buyers, which boosted the low- priced end of the market.

The author talks about prices dipping 1% in the midst of a first-time home buyer spree.  If demand was stronger due to a widely publicized government tax incentive then how did home prices dip?  I submit that you would be hard-pressed to find a pair of comparable homes that sold (one in mid Q1 and one in late Q2) that show a decrease in market value. Why?  Because it simply didn't happen.  Home values did not drop but the MIX of what sold did.  MANY more houses under the median sold and that simply changed the math.

The article continues…

New homes typically fetch higher prices than older ones, but with few being built, prices have been in the doldrums.

There is so much red meat here.  Where to start?   Remember we have a market that has expanded a sizeable 11% in a calendar year.  Next, we have frothy demand in the low-priced market (due to government stimulus) and at least moderate upper-price demand coming up against tight new construction inventory (due to cautious commercial lending).  So where do we come up with prices in the doldrums???   Hint: it’s in the 5th grade math.

I like this one so much that I keep the clipping from the original print edition. This guy isn’t just dishing-up junk news, the “doldrum” commentary shows that he’s buying into it!

Maybe the misuse of Median Sales Price is so over-the-top prevalent because so many people simply want it to be true?  Perhaps the home owning public is constantly on the lookout for cues about the equity in their homes.  Perhaps agents, brokers and Broker owners love a Silver Bullet that makes their day to day conversations more effective.  Perhaps journalists’ and broadcasters’ professional lives would be easier if that is what it truly meant.

How this will play out...

If the real estate content creation world really wants to make me happy, although I am pretty sure this isn’t part of their brief, things would play out in one of two ways:

  1. Can we shift to a measure that uses the change in sales price for an agreed upon “Basket of Homes” for each area?  Like the way you can tell if food prices are up or down by tracking a standard basket with a gallon of milk, a dozen eggs, 5 lbs of flour, etc.  

  2. Or can we just get people to write or say “overall, consumers bought more expensive homes” or “overall, consumers bought less expensive homes” when they cite changes in Median Sales Price?

The first would require a significant intellectual shift. The second one is awfully easy being that it is just a change in wording BUT being a realist, I am betting that neither one comes to pass!