Friday, April 3, 2009

Employment Revisions Egregious…

Frequent participant on this site, PointPark, penned a sad but true commentary on employment data revisions. I agree that they are egregious, there is simply no excuse for lack of accuracy in this data.
Employment Revisions...

By PointPark

Much like on ESPN's "SportsCenter," let's go inside the numbers, shall we?

The revisions to the unemployment data over the past six months has been, to say the least, breathtaking. Consider the following:

September 2008 job losses, originally reported as being -159k, were actually -403k after the final revision. That's a total difference of 153% to the downside.

October 2008 was almost as egregious. Originally reported as -240k, the final revision came in at -423k, for a total difference of 76%.

November (-533k to -597k) and December (-524k to -681k) revisions weren't as sharp, but still represented a respective 12% and 30% total change.

January 2009 saw a total revision of 24%, but it's the actual numbers that are the story. Originally, the losses were reported as -598k, but when the dust settled two months later, here in early April, we find out that 741k jobs were shed in January, making the first month of this year the first to lose over 700k jobs since the start of this recession-cum-depression.

If you apply even a conservative percentage of revision, say 30%, to February's (which were suspiciously unchanged in its first revision) and March's numbers, then you would have actual respective losses of 846k and 865k.

Just since September, using the revised and raw numbers we have as of today, the economy has lost a mind-numbing 4.16 million jobs that are never coming back. No magic wand waved or spell cast will change that. They're gone, baby, gone.

Now, adding the 30% revisions to Feb. and March, since September job losses would actually stand at a whopping 4.45 million.

Non-seasonally adjusted U3, BTW, is now 9%, while unadjusted U6 is already 16.2%, which is up from 9.3% a year ago and 10.6% in September.

Coming up, after the break, A-Rod's resurrection is days away. It will be a feat worthy of Jesus.

Hey, Jesus didn’t have steroids! Did he?

Seriously, the data from the government is just not trustworthy. John Williams at has devoted himself to tracking this data and providing numbers that are consistent with past calculations. What’s the point in tracking the data if you change it so that it’s not comparable to other timeframes and people can adjust it and tweak it at will?

Mish has also done a great job of reporting on their “birth/death” model and seasonal adjustments. Here’s his latest take on today’s report: Jobs Contract 15th Straight Month; Unemployment Rate Soars to 8.5%

And the unemployment data is just one of the problem areas. I’d say the largest is in the area of calculating inflation data with hedonistic adjustments and so forth that warp and skew not just inflation data, but also the growth data and the paychecks on which many people live. This is one of the root causes of the bubble and bust! Bad data leads to malinvestment as the distortions grow larger over time, one adding onto the other.

David Bowie - Changes: