If someone fools you once, shame on them. If they fool you with the same trick a second time, shame on you. If they "fool" you a third time — well, you must be in on it.
That's my take on Bloomberg News's virtually euphoric reaction to yesterday's new-home sales release from the Census Bureau. The wire service's Shobhana Chandra celebrated how seasonally adjusted December sales were at "the highest level in more than six years." The problem is that the bureau reported the same development two other times in 2014, only to see each improvement disappear in subsequent revisions. Excerpts follow the jump (bolds are mine):
Sales of New U.S. Homes Rose More Than Forecast to End 2014
Purchases of new homes in the U.S. jumped in December to the highest level in more than six years, a sign the industry is poised to keep expanding in 2015.
Sales increased 11.6 percent to a 481,000 annualized pace, exceeding all estimates in a Bloomberg survey of economists and the most since June 2008, figures from the Commerce Department showed Tuesday in Washington. Demand increased in three of four regions.
Gains in employment, borrowing costs close to historically low levels and easing credit may prompt more prospective buyers to enter the housing market. The improving outlook is driving up confidence among builders, and companies such as D.R. Horton Inc. and Lennar Corp. project the recovery will be sustained.
Chandra failed to note, as seen below, that we originaly saw "most since" figures compared to 2008 two other times in 2014, only to see the originally reported result fall steeply in subsequent revisions (HT, with one minor adjustment, to Zero Hedge):
As seen above, in both May and August, original results showing 504,000 units sold (seasonally adjusted and annualized) were sharply written down in subsequent months to the point of being the best in 11 months (May) and the best in three months (August, compared to the adjusted May).
The Census Bureau's figure this month only has to fall by 5,000 units for Chandra's "best since June 2008" statement to become untrue (it will then the best since July 2008), and by just 20,000 units to be demoted to being the best in just the past two months. Given that the average nosedive in revisions from May through November was 22,000, with the worst result coming in at a ridiculous -56,000, those changes are hardly outside the realm of possibility.
So Shobhana Chandra should really keep the chamapagne on ice, especially considering the fact that a downward revision of only 6,000 to December, with other months left alone, will cause full-year new-home sales for all of 2014 to come in lower than the 430,000 tallied in 2013.
In a bit of a role reversal, the Associated Press's Josh Boak, outside of the headline employed, was at least a bit restrained in his coverage:
US NEW HOME SALES JUMP 11.6 PERCENT IN DECEMBER
Sales of new U.S. homes accelerated strongly in December, a sign that home-buying may improve this year after a lackluster 2014.
The Commerce Department said Tuesday that new home sales climbed 11.6 percent last month to a seasonally adjusted annual rate of 481,000. The gains were not enough to offset essentially flat home-buying over the course of 2014. Just 435,000 new homes were bought last year, a modest 1.2 percent improvement from 2013.
The growth in December pointed to rising sales in 2015, buoyed by the combination of strong hiring in recent months and drastically lower mortgage rates.
Both reports gave too much weight to employment gains, and not enough to the fact that household incomes have yet to recover to pre-recession levels by a wide margin.
Cross-posted at BizzyBlog.com.
