AP Lowers the Housing Recovery Bar By At Least 17%

October 20th, 2010 5:54 PM

In a report so riddled with errors, inconsistencies, incompleteness and sloppiness that it's really hard to know where to begin, Associated Press real estate writer Alan Zibel couldn't even keep his housing recovery benchmark remotely consistent with what it was only a month ago.

The Census Bureau's September release of information about August housing starts and building permits informed the country that those items came in at seasonally adjusted annual rates of 598,000 and 569,000, respectively (they were revised slightly upward in yesterday's reports covering September).

On September 21, after disclosing the housing starts number, but not the one for permits, Zibel quoted Paul Dales, U.S. economist with Capital Economics, who said:

"Homebuilding activity remains at an astoundingly weak level," Dales said, adding that construction has to be more than double current levels for the market to be considered healthy.

My math says that means that annualized starts have to reach more than 1.2 million before health returns.

Now look at what Zibel told us in his report yesterday:

Housing starts are up 28 percent from their bottom in April 2009. Still, they are down 73 percent from their peak in January 2006 and 40 percent below the 1 million annual rate that analysts say is consistent with healthy housing markets.

Really, Alan? How did the benchmark for healthy housing markets drop by 17% or more, from the "more than 1.2 million" cited in September to 1 million yesterday?

My answer: It didn't. A quick eyeball review of the nearly 52 years of data in the Census Bureau's housing starts report (the not seasonally adjusted one, because we're looking at full years of data) shows that every single year since 1959 -- good and bad -- has had over 1 million total starts. A convenient table from the National Association of Home Builders showing annual starts from 1978 to 2009 shows us that total starts were over 1 million every single year, even during the slump years of 1981, 1982 and 1991.

Perhaps the AP writer intended to refer to total starts including multi-family units in September and only to single-family units in October. If he tries to go there, I'm not buying it. Dale's quote in September refers to "homebuilding," which would ordinarily seem to refer only to single-family dwellings. If Dale is setting the single-family annual bar at 1.2 million, the benchmark for all housing starts should really be much higher than 1.2 million (I told you the reporting was pervasively sloppy).

So why did Zibel lower the bar so significantly two weeks out from the mid-term elections? And why was he so sloppy about it? For many readers here, the questions answer themselves.

Cross-posted at BizzyBlog.com.