May 5, 2012

Three Wires All Avoid Mentioning Seasonality of Jobs Data in Friday’s Reports; Strangely Enough, the Raw Numbers Stink

It is more than a little odd that each of the three wire services identified in today’s earlier post (at NewsBusters; at BizzyBlog), in reporting on yesterday’s OMG-awful jobs report, somehow failed to mention something about the data presented. Specifically, at Bloomberg, Reuters, and the Associated Press (here and here), five reporters in four stories somehow avoided using two truly required words in describing the data contained in many if not most government economic data releases: “seasonally adjusted.”

One is in an odd omission. A pair of such reports is a strange coincidence. The presence of four from three separate sources makes you wonder, especially since all three wire services found room for the two magic words (Bloomberg, though cryptically; Reuters; AP) in dispatches about Uncle Sam’s report on initial unemployment claims the previous day. A look at how dismal the not seasonally adjusted numbers were in April follows the jump, and shows how, bad as they turned out to be, the Obama administration caught a lucky break in the seasonal adjustment calculations. It may also explain why the wire services avoided mentioning it.

Here is how the raw (i.e., not seasonally adjusted) and seasonally adjusted data compare during the past decade:


While I’m not contending that the Bureau of Labor Statistics (BLS) performed their seasonal calculations incorrectly (but someone should consider digging deeper), the data seen above make a strong argument that the seasonally adjusted numbers as presented yesterday are higher than the underlying reality. Paraphrasing what I wrote yesterday at my home blog, concerning the overall nonfarm (red boxes) and private-sector (green boxes) payroll changes:

  • Not seasonally adjusted (NSA) additions of 896,000 in April 2012 trailed those seen in April 2010 by 233,000. Yet the seasonally adjusted (SA) result is only 124,000 lower.
  • This year’s NSA number trails April 2011 by even more (283,000), but the SA result is only 136,000 lower.
  • As to , the NSA figure of 896,000 is in this case 154,000 and 257,000 lower than April 2011 and 2010, yet the SA number is only 63,000 and 134,000 lower.

Assuming that BLS isn’t cooking the books — and this may not be as safe an assumption as it once was — the Obama administration caught a huge break, because it seems that both the overall nonfarm and private-sector numbers after seasonal adjustment could easily have been about zero.

Further, this year’s NSA numbers include 206,000 jobs from the “Birth/Death Model,” which is supposed to measure jobs being created or lost at new businesses (less those going out of business) which BLS believes is out there but cannot find. Last year’s comparable number was 172,000. Why this year’s number is higher is a mystery.

Of course, one can’t prove some kind of collusion in how all three wires somehow forgot to mention the seasonally adjusted nature of the jobs data yesterday. But at a minimum, in that it prevented relatively uninformed readers from wondering what the raw numbers were, it was very convenient. Those who are more conspiratorially oriented should note that all three news organizations are among the less than twenty outfits which have pre-embargo access privileges (i.e., they get to see the data 30-60 minutes and to interact with representatives at Hilda Solis’s Department of Labor before everyone else does), which in my opinion they should not be entitled.

Cross-posted at


1 Comment

  1. I track wires all day and this is something I hadn’t noticed. Hm.

    Comment by Melvin — May 6, 2012 @ 8:40 am

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.