The Trouble With Case Shiller, Again
Courtesy of Lee Adler of the Wall Street Examiner
Here are the key excerpts from that piece which tell you why Case Shiller should be ignored.
The mainstream media wasted hours reporting on and analyzing the Case Shiller Housing Index today. Did even one pundit mention what's wrong with the Case Shiller data? If Dow Jones used the methodology to report the Dow Industrials that Standard and Poors uses to construct the Case Shiller Index (CSI) the Dow would be reported as being at 10,650 when, in reality, it's at roughly 10,800. 10,650 is where the Dow's 3 month moving average was at the end of May.
"What does that have to do with now?" you ask. I say, "Absolutely nothing!" The fact that the numbers aren't too far apart is coincidental. The same is true of the housing market. The CSI doesn't remotely represent the current state of the housing market, which is, in fact, much lower than where the CSI shows it to be, and it has been trending drastically lower when the CSI shows it being relatively flat recently.
Why is the Case Shiller data both inaccurate and misleading relative to current market conditions? Let's see. The data is collected from a month ended two months ago, in today's case, July. That data is from public sources for CLOSED sales, NOT then current contracts. So the currently reported data represents sales that happened mostly in two months before that, which would be May. Then that data is aggregated with the data collected in the two previous months, for sales contracts from the 2 months before that (March and April), and reported as an average price for the 3 months. The theoretical midpoint of the data reported now is from 3½ months ago, representing the average contract price from roughly 5½ months ago. In other words, today's CSI represents average selling prices as of the MIDDLE OF APRIL when the homebuyers' tax credit was still skewing the market upward.
As soon as that distortion was removed from the market, prices collapsed. CSI isn't showing that. Other data, such as the Commerce Department's new home sales price data is. Still more current data than that is also showing a sharp drop which is continuing right up to today.
The media and its captive economist pundits treat the CSI as somehow being a good indication of current market conditions. It's anything but. It's wrong, and dangerously misleading.
The National Association of Realtors knows the real story. They track current contract prices in their MLS databases. But they choose not to share this information with the public. It is not in their industry's interest to do so. If you want to know the real story, you need a few friends in the business around the US. Or you could use some of the tracking services whose data sources are current listings, and which have proven to be highly reliable trend indicators when compared with the closed sales data released months later.
The closed sales information doesn't reach the public until after the sales are reported in the public records of each county. That's usually a month after the sale closes, hence the 3 month lag between the time the property goes under contract and the time the public has first news of that sale. It takes S&P another month to collect and assemble the data into the CSI indexes. That's why they are so late. Case Shiller's methodology of then reporting only a 3 month moving average as if it were the most recent prices further compounds the misimpression. The goal of smoothing the data only serves to slow it down and render it even more inaccurate.
A good real time market indicator is the data from Housingtracker.net. It reports current listing prices from 55 large US metros. I have compared that data with subsequently released actual closed sales data and the listing prices have proven to be a good indicator of the market's direction in real time. Naturally, listings prices are higher than sales prices, but a measure of listings versus sales prices published by Zillow.com shows a remarkably consistent spread that hovers around 10%. The Housingtracker data thus shows both the trend, and with the subtraction of 10%, a reasonable indication of current price level.
Housingtracker indicates a current median asking price for the 55 US metros of $218,886 as of January 30. That's a month to month increase of 0.5% and a year to year increase of 3.8%. That's the real state of prices. It's a far cry from Case Shiller's ancient history showing prices down 1.3% month to month, and 3.7% year over year. As far as representing the current market, those numbers just are not true.
I ignore Case Shiller in my work, but here's a chart of several indicators which I do track which will give you some idea of where housing prices really have been and where they are now.
