http://twitter.com/#!/WJGBalderama/status/256405323366146049

Earlier today, the Bureau of Labor Statistics reported that initial claims for state unemployment benefits unexpectedly dropped 30,000 to 339,000 from 369,000 a week ago.

Economists were forecasting a reading of 370,000, according to Bloomberg.

The plunge in initial jobless claims is seen in some quarters as confirmation of last Friday’s unemployment report that the U.S. economy is improving.

Twitter users immediately began attacking former General Electric CEO Jack Welch, who suggested that last week’s unemployment report was manipulated to make the Obama administration look good ahead of next month’s election:

In a dark lonely room @jack_welch is pounding head on his desk. “Jobless claims fall to 339,000 lowest in 4 1/2 years.” #traitor

— Neo Blaque (@neoblaque) October 11, 2012

Jobless claims are now lower than when Obama took office! Take that Jack Welch and GOP obstructionist!

— SportsGirl101 (@Arianna8927) October 11, 2012

Lowest weekly unemployment claim since Feb 2008. read.bi/Qff1J7 I’m sure @jack_welch will have something crazy to say about this.

— Derek Young (@DerekMYoung) October 11, 2012

But Bloomberg reported something odd that raised doubts about today’s jobless claims report.

One state accounted for most of the plunge in claims, a Labor Department spokesman said as the data were issued to the press.

Forbes added fuel to the fire with this:

[I]t seems this week’s data was skewed by a large state failing to report quarterly figures. “Some are saying it was California, though we cannot confirm that,” says FTN Financial economist Chris Low. “As a result, the reported plunge in claims is suspect. We expect claims will continue trending lower through the quarter, as there is strong seasonal bias downward anyway…But the big drop this week will likely be revised or be followed by a pop upward next week.”

Marketwatch quotes another analyst who thinks the missing state is California:

Added Stephen Stanley of Pierpont Securities: “The formula for the size of a claimant’s benefit check is derived based on an average of their last few quarters of pay (the more you were earning before being laid off, the bigger your unemployment check would be). Thus, in many cases, it pays for a laid-off worker to game the formula by waiting until the beginning of the next calendar quarter to file (if they can wait that long), as they may have been getting paid more in the quarter when they were laid off than in the quarter that rolls out of the equation if they wait.”

“As a result, there is an accumulation of claims that are likely submitted over a period of several weeks but not processed until the turn of the quarter. Apparently, the state in question (and it pretty much has to be California to account for anything close to 30,000) forgot to include that stockpile of unprocessed claims in their tally for this week (which is the first week of a new calendar quarter). Since the seasonal factors expected an unadjusted surge of almost 20% in the period to account for the quarterly filing pattern, failure to adhere to that pattern in the raw data (unadjusted claims were only up 8.6%) creates a big drop seasonally adjusted.”

CNBC’s Kelly Evans is reporting in a confusing series of tweets that “one state did not process & report its typical seasonal workload” and that it’s likely we’ll see headline jobless claim numbers rebound next week.

Huh:”Labor Dept econ said 1 large state didn’t report add’l quarterly figures as expected, accounting for substantial part of the decrease.”

— Kelly Evans (@Kelly_Evans) October 11, 2012

Let the guessing games begin: which state didn’t turn in its jobless claims tally?

— Kelly Evans (@Kelly_Evans) October 11, 2012

Or rather, looks like one state reported a drop instead of an increase? bloomberg.com/news/2012-10-1…

— Kelly Evans (@Kelly_Evans) October 11, 2012

BBG: “The year’s increase was smaller than projected, because one large [state] showed a drop rather than an increase, the spokesman said.”

— Kelly Evans (@Kelly_Evans) October 11, 2012

Apols for the confusion here. Looks like all states DID report; one showed a drop instead of expected increase in quarterly refilings.

— Kelly Evans (@Kelly_Evans) October 11, 2012

It IS still the case that one state didn’t process as expected. Also likely we’ll see headline rebound next wk absent typical seas decline.

— Kelly Evans (@Kelly_Evans) October 11, 2012

So, (a) 1 state appears not to have processed normal seasonal workload; (b) seasonal spike didn’t occur; (c) means rebound next wk likely

— Kelly Evans (@Kelly_Evans) October 11, 2012

Just to be super, super clear: it *does* appear that one state did not process & report its typical seasonal workload. (see a/b/c tweet)

— Kelly Evans (@Kelly_Evans) October 11, 2012

Labor Dept did not estimate what state would have reported (understandably — as it couldn’t know).

— Kelly Evans (@Kelly_Evans) October 11, 2012

Evans’ colleague Becky Quick sums up Evans’ reporting as follows:

Great work by @kelly_evans. Jobless claims # is legit. There was NOT a state that didn’t report. State didn’t report a BUMP as expected.

— Becky Quick (@beckyquickcnbc) October 11, 2012

We’re not sure Evans said what Quick thinks she said.

If you’re feeling confused, join the club.

If anyone can cut through all the spin and confusion, it’s bond traders. This morning the bond market is reacting to the jobless claims numbers with a yawn, which means the prospect for employment growth in the U.S. remains bleak.

Read more: http://twitchy.com/2012/10/11/plunge-in-jobless-claims-but-are-the-numbers-legit/