Fuzzy Reporting - Press Helps to Distort Housing Data
I want to talk about the renewed tendency of the press to reiterate the pronouncements of the National Association of Realtors (NAR) apparantly without consideration for how badly they were bamboozled by the NAR from 2003 to 2008.
Check out this opening paragraph announcing the latest “great news” on existing home sales (emphasis mine):
March 23 (Bloomberg) -- U.S. Sales of previously owned homes unexpectedly climbed in February as record foreclosures brought bargain hunters into the market to take advantage of lower prices.
I am not clear on this “unexpectedly” part. Home sales always rise in February compared to January. That’s just part of the seasonal dynamic of home sales.
In good years and bad, February is a better month for home sales than January. That’s why the data series is “seasonally adjusted."
Look at the purple lines connecting January to February in the chart below from Calculated Risk. This covers both bubble years and lean years.
As you can see, each of the last four years has presented a successively lower home sales figure than the year prior, but February was always higher than January, despite the yearly declines. So I am unclear what the intent was in stating that home sales "unexpectedly" rose in February.
In my book, when something happens each and every year, it is "expected."
The real story is that home sales are still declining, but you won’t find that emphasized in Bloomberg, Reuters, the NYT or elsewhere. For the most part, that critical element of the story is “hidden in plain sight."
The real fun will begin from August through December, when the series nearly always falls. That’s when you will begin to hear the NAR, and their friends in the media, display a renewed preference for the “seasonally adjusted” numbers, provided those can be spun to provide a bullish, upbeat view of the housing market.
Why do we care?
Because being less than forthright with ourselves is largely to blame for the mess we’re in.
Because we owe it to ourselves to take an unflinching look at the data and see what it says, without needing to couch it in the same soothing but patronizing tones that one might use on a 3 year old with a lightly skinned knee.
But even the numbers above, as misrepresented as they are, do not even remotely begin to display the true severity of the story.
For that we’ll have to use our ability to connect information separated by as many as a dozen paragraphs of material.
Here’s the second paragraph of the Bloomberg article.
Purchases increased 5.1 percent to an annual rate of 4.72 million from 4.49 million in January, the National Association of Realtors said today in Washington.
See what they did there? Announcing a 5.1% increase in home sales sounds good right? While we had to go and dig for a chart showing that a 5.1% increase is pretty much par for the course going from January to February, most folks would have been left with the impression that there was some sort of an unusual bounce there.
We now know that February sales are down each of the past four years and that this Jan-to-Feb bounce was nothing special at all. It was expected and ordinary. But I digress....
There’s another, even bigger, element to this story that’s just sitting there practically gnawing on the press’ lower legs, just 14 paragraphs down:
Home foreclosures were up 30 percent in February from a year earlier, according to RealtyTrac Inc., an Irvine, California-based seller of default data. A total of 290,631 properties got a default or auction notice or were seized by banks.
Lessee here...290k seized properties (many of which count as "sales" in the NAR methodology) times 12 months gives us 3.48 million homes in default, slated for auction, or seized.
When we divide these 3.48 million (annualized) defaulted or seized properties by the 4.72 million in total reported sales (also annualized), we immediately discover that a full 73% of recent housing activity was distressed.
By mixing foreclosure data with organic house sales, the NAR and the press have badly misrepresented the actual condition of the housing market.
This would be like counting auto repossessions as “sales” and then reporting a big increase in auto buying after a wave of repossessions had swept through a city.
It’s just plain wrong to count things this way, and everybody knows it. But still we do it.
Such Fuzzy Reporting was one of the prime causes of our self-induced housing excess and I am not at all clear on why continuing that practice makes sense at this time.
I think in the future I’ll refer to this as “Jiminy Cricket reporting.”