Almost every statistic reported about the housing market is misleading in some way.  While the housing market is getting worse by the day, you wouldn't know it from information reported by the government, Wall Street, banks, realtors, media, and industry associations.  Worse, none of them have tried to educate the public about the problems with their data. 

Here you will find the most complete description of what is wrong with the housing market data.  Most of it is my own work.  Some is copied from other sources as noted.   If you have some "data lies" reports to add, please contact me at schberkland@sbcglobal.net.

Case-Shiller index is fatally flawed 

Instead of measuring housing values, as we are led to believe, the index measures the change in price from the last time that group of home sold.  As long as many homeowners are selling for more than they paid for their homes, the index will be upward biased.  It ignores incentives, new home sales, effect of all those expensive remodels, and all those homes that are languishing and not selling. In short, it is not a home price index at all.  Rather, it's a "sold price index". 

Quirky MBA categories understates subprime delinquencies 

5/20/07:  The Mortgage Bankers Association counts a subprime loan as delinquent only if it was made by a subprime lender.  In their category, Countrywide is a prime lender, so none of Countrywide's delinquent subprime loans are counted. 

Construction job losses not counted in unemployment data

Job losses due to the busting housing market are much higher than shown in the unemployment reports, because construction workers, realtors, and mortgage brokers are usually independent contractors and don't get unemployment pay.  In the past year, 600K illegal construction workers lost their jobs - none are in the data. 

 $/sq ft does not measure price changes of homes, but price changes of the mix of homes sold

Small homes cost $400/sq ft, while larger homes cost $300/sq ft.  When fewer small homes sell, the $/sq ft goes down.   Thus, $/sq ft is a distribution number.  There is not a number that measures the change in home prices! 

Median price, average price, and $/sq ft are not price indicators

These are distribution numbers, meaning they measure the change in the mix of homes sold.  In a cooling housing market, the first time buyer is squeezed out first, so the smaller homes make up a smaller percentage of sold homes every month.  This trend toward more large homes selling makes the median and average price go up, and the price/sq ft go down (since larger homes cost less per square foot). 

5/11/07, per a comment by Patrick B. on patrick.net, I want to reiterate that the reduction in subprime lending causes even fewer low end homes to be sold, so we should expect the median to rise sharply in the next few months.  Subprime lending was really hit in February, so closing from April 1 onward should show the effect of more large homes sold, skewing up the median and average, and lowering the $/sq ft.  I anticipate that next week's release of the April data will show a bigger shift than we have seen to date.

From the Credit Suisse Report:  The Data Lies

The data lies on housing starts, the OFHEO index, NAR home prices,  new and existing home sales.  Analyst Ivy Zelman looks behind the headlines, and explains why these indicators are plain wrong.

MBA Purchase Index rises, but lending is down

The MBA Purchase Index, which measures applications, rises as borrowers make multiple applications, and bankrupt lenders increase volume to those left in business.   People have misinterpreted the recent rise in the Index to signal more purchases of homes.

The MBA Refinance Index should be subject to that same error.