Suitably Flip

Jobless Thursday

You know, in the private sector, you go to jail for what used they used to winkingly call "managing" earnings - using accounting gimmickry to nudge revenues and expenses from period to period in order to meet analysts' expectations or smooth volatility.

But in the realm of government data reporting, the art of data massage has become impressively robust.

For the 59th week of the last 60, the previous initial jobless claim report was revised upward, from 386,000 to 389,000.  And once again, this enables the Labor Department to report a week-over-week decline in new jobless claims, from the adjusted 389,000 to an unadjusted 388,000.  Upon next week's revision, this week will almost certainly have shown another increase.

If that sounds familiar, it may be because last week, the government reported a decline of 2,000 (but only after upwardly revising the previous week by 8,000).

Looking back over the last five weeks, the cumulative reported weekly changes (from previous weeks' adjusted data to the new unadjusted numbers) showed a net decline of 1,000, despite an actual cumulative net increase of 24,000.  And that's without the 5th revision factored in, at which point the cumulative increase will be closer to 30,000.

In addition to serving as fodder for another round of "Jobless Claims Fall" headlines, this week's underestimate has the additional side effect of avoiding the probably true headline "Jobless Claims Reach New 2012 High" from being written (at least for another week).  They started at 390,000 in early January and, assuming next week brings an upward revision of more than 2,000 (revisions have ranged from +3,000 to +10,000 over the last month), then we're already sitting at year-to-date highs.

The lie of unadjusted unemployment claims at least used to be a predictable one.  For the last year or so, the upward revision was almost invariably 3,000 or 4,000.  While last week's oops was only 3,000, the two preceeding weeks were truly wild pitches that needed revisions of 8,000 and 10,000.  So we no longer have the luxury of appplying a known truth adjustment factor to reveal the real data.  Alas, the only thing we know for sure is that initial unemployment claims are some amount higher than 388,000.

Update:  Ed at Hot Air has more, including some of the early media reports, dutifully noting that jobless claims "eased" last week.

Comments

Jobless Thursday

You know, in the private sector, you go to jail for what used they used to winkingly call "managing" earnings - using accounting gimmickry to nudge revenues and expenses from period to period in order to meet analysts' expectations or smooth volatility.

But in the realm of government data reporting, the art of data massage has become impressively robust.

For the 59th week of the last 60, the previous initial jobless claim report was revised upward, from 386,000 to 389,000.  And once again, this enables the Labor Department to report a week-over-week decline in new jobless claims, from the adjusted 389,000 to an unadjusted 388,000.  Upon next week's revision, this week will almost certainly have shown another increase.

If that sounds familiar, it may be because last week, the government reported a decline of 2,000 (but only after upwardly revising the previous week by 8,000).

Looking back over the last five weeks, the cumulative reported weekly changes (from previous weeks' adjusted data to the new unadjusted numbers) showed a net decline of 1,000, despite an actual cumulative net increase of 24,000.  And that's without the 5th revision factored in, at which point the cumulative increase will be closer to 30,000.

In addition to serving as fodder for another round of "Jobless Claims Fall" headlines, this week's underestimate has the additional side effect of avoiding the probably true headline "Jobless Claims Reach New 2012 High" from being written (at least for another week).  They started at 390,000 in early January and, assuming next week brings an upward revision of more than 2,000 (revisions have ranged from +3,000 to +10,000 over the last month), then we're already sitting at year-to-date highs.

The lie of unadjusted unemployment claims at least used to be a predictable one.  For the last year or so, the upward revision was almost invariably 3,000 or 4,000.  While last week's oops was only 3,000, the two preceeding weeks were truly wild pitches that needed revisions of 8,000 and 10,000.  So we no longer have the luxury of appplying a known truth adjustment factor to reveal the real data.  Alas, the only thing we know for sure is that initial unemployment claims are some amount higher than 388,000.

Update:  Ed at Hot Air has more, including some of the early media reports, dutifully noting that jobless claims "eased" last week.