WSJ: Home Sales Data Phonied Up

We live in an era where data fetishists worship numbers, stats and data - but increasingly, we are coming to see that the data is phony.

Whether it's Bloomberg's crime stats that downgrade felonies to misdemeanors and misdemeanors to lost property or the BloomKlein test score miracle that wasn't or the unemployment rate which consistently undercounts the employment problem, so much of the data used by the media and the political establishment is just bullshit.

Today, the Wall Street Journal reveals that the National Association of Realtors has been phonying up the home sales data (surprise, surprise!):

The housing crash may have been more severe than initial estimates have shown.

The National Association of Realtors, which produces a widely watched monthly estimate of sales of previously owned homes, is examining the possibility that it over-counted U.S. home sales dating back as far as 2007.

The group reported that there were 4.9 million sales of previously owned homes in 2010, down 5.7% from 5.2 million in 2009. But CoreLogic, a real-estate analytics firm based in Santa Ana, Calif., counted just 3.3 million homes sales last year, a drop of 10.8% from 3.7 million in 2009. CoreLogic says NAR could have overstated home sales by as much as 20%.

While revisions wouldn't affect reported home-price numbers, they could show that the housing market faces a bigger overhang in inventory, given the weaker demand.

In December, NAR said that it would take 8.1 months to sell some 3.6 million homes listed for sale at the current pace, but the number of months it would take could be even higher if sales are revised down. Any revisions wouldn't have an impact on homeowners, but it could have consequences for the real-estate industry. Downward revisions would show that "this horrific downturn in the housing market has been even more pronounced than what people thought, and people already thought it was pretty bad," said Thomas Lawler, an independent housing economist.

NAR said the data, which are used by economists, investors and the real-estate industry to gauge the health of the housing market, could be revised downward this summer. Lawrence Yun, chief economist at NAR, wasn't specific about whether and by how much the revisions could reduce reported sales, and he raised the possibility that the CoreLogic estimates have understated the number of home sales. "This is a very important issue, and we are looking at it carefully right now," Mr. Yun said.

Barry Ritholtz notes that the overcounting of home sales goes back to 2000, not 2007, and takes a closer look at the problem here.

Accidental miscounting problem or fraud?

Hmmm....