Leave it to the Associated Press, with the assistance of the "magic" of seasonal adjustments, to make the November housing market appear as if it was a bit better than the two months that preceded it. It wasn't.
Thursday, the wire service grabbed the single crumb that was available, namely the Census Bureau's report earlier that day that annualized, seasonally adjusted housing starts had increased by about 4% and turned it into a decidedly positive headline: "Home construction up after 2 months of declines."
AP Economics Writer Jeannine Aversa watered down the headline in her very first sentence, describing the "up" part of the headline as a "nudge."
That's nowhere near enough. The available evidence indicates that November may have been the worst month the homebuilding industry has had in 4-5 decades of related recordkeeping.
One number that rarely gets press attention is the Bureau's "New Privately Owned Housing Units Completed" report (PDF here). Its November total of 44,600 is the lowest ever for any month in any year since the Bureau began reporting the figure in 1968. That result, combined with the lowest-ever November for both housing starts and building permits, would indicate that new home "construction" -- correctly defined as what happens between when a job starts and gets finished -- was at an all-time low, and certainly not "up" as the AP headline claims.
In another egregious error that has developed into an AP habit of late, Aversa followed previous AP writers in setting a benchmark for a "healthy housing market" that is far lower than truly healthy markets have historically been.
Here are the first five paragraphs of Aversa's report, plus a later one on building permits:
Home construction nudged up in November after two months of declines.
Builders broke ground last month on a seasonally adjusted 555,000 units, a 3.9 percent rise from October, the Commerce Department reported Thursday.
Even with the gain, housing starts are just 16 percent above the 477,000 unit pace from April 2009 - the lowest point on records dating back to 1959.
And they are down 76 percent from their peak in January 2006, and 45 percent below the 1 million annual rate that analysts say is consistent with a healthy housing market.
All the activity last month came from building single-family homes. They increased to a pace of 465,000 units, a 6.9 percent rise from October. Apartment construction fell 9.1 percent to a unit pace of 90,000.
... Housing permits, a barometer of future demand, fell 4 percent to an annualized rate of 530,000, reflecting weakness in apartment construction. It marked the lowest level in permits since April 2009.
Beyond what was noted earlier, the de facto AP communication asserting a bit of improvement doesn't hold up upon review of the data.
Let's begin by looking at the last two years of NSA housing starts info (Census Bureau PDF here):
- The 41,200 total is the lowest for any November since such recordkeeping began in 1959.
- (Red boxes) After some signs of life in August, where starts were about 6.5% ahead of August 2009, the bottom started dropping out. November's starts are 27% below August 2010; the previous year's August-November decline was 20%.
- (Green box) I believe Aversa meant to write that "All of the increase in activity came from building single-family homes." The 33,700 November figure for single-family starts is lower than last November, and 35% lower than April. I can see how November's relatively small decline from October can turn into a seasonally adjusted positive, but it's certainly not a cause for being impressed, and certainly not a reason to cause readers to believe that things are legitimately getting better, even minimally.
Now let's look at building permits (PDF here):
- (Blue boxes) The NSA also came in with the lowest November on record.
- (Orange box) The October-November decline in single-family permits compared to the same months in 2009 of about 13% is way larger than the 7% decline in starts in the same comparative periods, signaling that an already slow market is heading even further downward.
Combine these two reports with the "Completions" report noted earlier, and you have the picture of a new home market that barely has a pulse, and is heading towards a flatline.
Finally, Aversa's claim that 1,000,000 units a year represents a "healthy market" is completely false in historic context. 1.2 million is more like it.
Aversa and the AP have again clearly failed to present housing market reality.
Cross-posted at BizzyBlog.com.