For the past six weeks combined, actual jobless claims filed nationwide have been virtually the same as the were during the six comparable weeks in early 2012.
You wouldn't know that from Christopher Rugaber's coverage at the Associated Press of the Department of Labor's unemployment claims report released yesterday. Rugaber, who described last month's jobs report showing the unemployment rate rising to 7.9 percent with a mediocre 157,000 jobs added (both figures are seasonally adjusted) as "mostly encouraging," wrote Thursday that the movement in jobless claims "suggests slow but steady improvement in the job market." If so, that "suggestion" is at best a whisper.
By ignoring the raw data, Rugaber and the rest of the press are painting a rosier picture of the claims situation than really exists. As seen in the following table for the first six truly comparable weeks in 2013 and 2012, year-over-year seasonally adjusted claims data would lead one to believe that this year's level of layoffs is somewhat lower than last year's, when in reality they're barely lower (early January's numbers were affected by the day of the week on which New Year's Day fell; the second and third weeks listed were differently affected in 2013 and 2012 because of the timing of the Martin Luther King holiday):
Until very recently, weekly claims had dropped by anywhere from 6 percent to 14 percent from the same week in the prior year.
Also note that the current week's numbers will probably be revised up a bit, as has been the case well over 90% of the time during the past two years.
With this in mind, let's look at a few paragraphs from Rugaber's writeup (bold is mine):
The number of Americans seeking unemployment benefits jumped 20,000 last week to a seasonally adjusted 362,000, though it remains at a level that suggests slow but steady improvement in the job market.
The Labor Department said Thursday that the four-week average, a less volatile measure, rose 8,000 to 360,750, the highest in six weeks. A department spokesman said heavy snowstorms in the Northeast didn't affect the total.
Applications for unemployment benefits are a proxy for layoffs. Even with last week's jump, they have trended downward recently. The four-week average has declined 7.5 percent since mid-November and fell to a five-year low three weeks ago.
As seen above, the only reason the seasonally adjusted four-week average fell so far three weeks ago is that it included two weeks (January 12 and 19) during which raw claims were actually greater than the were in 2012 -- but somehow, after seasonal adjustment, they were far lower.
As I've written several times before in comparable situations, I'm not saying that the DOL didn't perform the seasonal adjustment calculations correctly. I am saying that relying entirely on solely on seasonally adjusted figures, especially in a protracted period of economic malaise such as the one we've been living through for almost five years, can lead one to make wrong conclusions. This is one such instance. The roughly four-point difference between raw and seasonally adjusted claims in the table above probably washes away most of the improvement Rugaber, looking only at seasonally adjusted figures, claimed has taken place since mid-November.
One potential offsetting point is that job growth during late 2011 and early 2012 was pretty strong. But January 2013 wasn't very strong; based on info available so far, February also won't be particularly impressive. That certainly wasn't conveyed in Rugaber's "slow but steady improvement" analysis -- which is yet another reason why he and other AP reporters shouldn't be engaging in it.
Cross-posted at BizzyBlog.com.