Following the governmemt's Employment Situation Summary yesterday, two words were noticeably absent at the Associated Press (here, here, and here), Bloomberg, Reuters, CNBC, and the New York Times: "seasonally adjusted."
While they told their readers of the number of jobs supposedly added in total (157,000) and in other sectors, the fact remains that in the real world, before seasonal adjustment, the government told us, as is the case every January, that employment declined steeply. In January 2013, the government estimates that 2.84 million jobs were lost.
This failure to refer to seasonal adjustments is odd, because while these same outlets typically use these two words when describing the weekly unemployment claims results released each Thursday and several other government reports which appear throughout the month, they typically fail to do so in describing the monthly number of jobs added or lost -- and for that matter, the unemployment rate, which was really 8.5 percent in January, only 0.3 points below where is was in January 2012 (December's, November's, and October's differentials were 0.7, 0.8 and 1.0 points, respectively).
By not using the term "seasonally adjusted" anywhere in their reports, these establishment press outlets have led readers to believe that their wonderfully positive number represent what actually happened. They don't. Thus, as a public service, yours truly will present the real numbers beside several of those reported at the aforementioned outlets, with a bit of analysis as to whether trends in those sectors appear to be favorable (SA means "seasonally adjusted; NSA means "before seasonal adjustment, i.e., the actuals):
Overall: 157K SA, -2.84 million NSA; January's actual job loss was worse than the 2.635 million seen in January 2012 which generated a 311K seasonally adjusted result. This looks like deterioration to me, not improvement. It also appears, looking at the three preceding years, that January's 157K SA figure could easily and justifiably been as low as 100K.
Private sector: +166K SA, -2.345 million NSA; again much worse than the -2.155 million seen in January 2012, and again with the seasonally adjusted number arguably overstated by 50K or more.
Health care: +23K SA, -57K NSA; the actual decline in January 2012 was far less (-36K), perhaps indicating that health care employment averaged over 12 months might be close to hitting a plateau.
Retail: +33K SA, -592K NSA; in this 2013's actual decline is worse than each of the past three years by 19K-32K. If retail stops adding people over time, look out below.
Construction: +28K SA, -272K NSA; this one's a real head-scratcher as to why we should be impressed. Contstruction employment is just over 100,000 greater than it was a year ago, and (seasonally adjusted) is still over (23 percent) below its January 2008 value of peak of 7.476 million. At that rate, it will take about 15 years for construction employment to get back to where it was.
Manufacturing: +4K SA, -90K NSA; the NSA number is worse than both 2012 (-65K) and 2011 (-80K). In the past 12 months, seasonally adjusted manufacturing employment is up by 109K, or just under 1 percent.
Besides the obvious negligence seen in their failure to appropriately label seasonally adjusted figures as, well, seasonally adjusted, it's also annoying to observe that the press either believes that news consumers are too easily confused if they are also told what really happened, or worse (and this is not a long-shot possibility), many if not most business reporters don't even understand the difference between raw and seasonally adjusted figures. That of course leads one to wonder why they are employed as business reporters in the first place.
Cross-posted at BizzyBlog.com.
