This couldn't have happened at a worse time. After last Friday's BLS truther conspiracy idiocy was catapulted by Jack Welch, the DoL spokesperson who handed out this morning's initial jobless claims to the press single-handedly handed the tin foil hat brigade some ammunition.
Instead of letting the data speak for itself, or issuing a public statement that everyone could digest equally, s/he told the members of the press *privately* that unique factors explained much, perhaps most, of this week's precipitous decline of over 30,000 claims. As a result, Bloomberg, CNBC, and the Wall Street Journal have all reported separately on what their reporters have said they overheard.
One state was responsible for either (a) much or (b) most of the decline. That was either because the state (a) did not report or (b) reported fewer claims when higher claims were expected on a (a) non-seasonally adjusted or (b) seasonally adjusted basis.
The answers to these three questions impact a great deal on how much we should read into the decline this week. Whichever reporter heard and reported most accurately and completely has the equivalent of very large and market-moving insider information. The public at large has no way to know what the truth is. With less than a month before the presidential election, and the BLS having already been accused of cooking the books, the truth Matters. Already there are claims that the state must have been Obama's home state of Illinois.
If *most* of the decline was due to an decline vs. a rise in NSA claims, then the culprit is almost certainly California. California's claims surged from 56.1 to 70.0 thousand claims in the reference week last year, a 13,900 gain. The state with the next largest number of claims, New York, only saw claims jump from 20.1 to 28.7 thousand claims, a gain of 8,600. The seasonal adjustment this week is very small. So even with an undercount, it is very unlikely New York could have swung the seasonally adjusted number by 15,000 or more. But even if California is the culprit, it is likely that this week's claims did drop by about 10,000 -- still a significant decline, and possibly setting up a new low in the 4 week average next week.
If on the other hand, if the state did not report at all, then about another half dozen states such as Texas, Florida, Illinois, Pennsylvania, and North Carolina might be the culprit. Those states could have NSA claims up to 20,000 for the reference week, and thus the seasonally adjusted 4 week average might not decline at all after next week's number is known.
In other words, next week we might know that unemployment claims really haven't improved at all. Or we might know that, despite the issue, a new post-recession 4 week average low was made.
But the reporter who most completely and accurately remembers the spokesperson's statement - and any investor s/he shared the information with - likely knows that information now.
Sharing this kind of information privately with reporters is a Big Mistake. Since the BLS can't put the genie back in the bottle, they owe it to the public to clarify which state was involved, whether it was a non-report or simply an unexpected report, and what the change in jobless claims would have been comparing the other 49 states last week vs. this week.
UPDATE: Henry Blodget of Business Insider spoke with a source in the DoL who has confirmed the California version of event I independently deduced above, including the fact that there almost certainly would have been a very substantial drop in initial claims this week anyway. The DoL needs to issue a press release explaining this so that all investors and members of the public are on an even playing field. It's not a matter of courtesy, we are entitled to the information.
I'm on Linked In and Twitter (@captivelawyer). Silver Oz's Linked In name is @silver_oz. NDD is a fossil and may be reached by etching a picture in stone on the wall of a cave.
The Bonddad Economic History Project
At the beginning of 2012, I decided to start looking at the actual, statistical history of the US economy starting in 1950. The reason is simple: to find out what really happened. So, when you see title of a post that begins with a year such as 1957, followed by "employment" or "Fed policy: you know what it's for. You can also access the information by typing in BE for Bonddad econ and a year to find information on a particular year.
Here is a link to pages that contain links to all the posts on the years listed.