About that weekly initial unemployment claims report

One might see blazing headlines that initial weekly unemployment claims for March 25, 2010 dropped 14,000 to 442,000. But note the below, contained in the press release:

This week's release reflects the annual revision to the weekly unemployment claims seasonal adjustment factors. The historical factors from 2005 forward have been revised.

From 2005 they revised the data? What kind of yearly seasonal adjustment is that?

Then from MarketWatch we have (looks like someone got on the phone?)

Latest figures reflect annual revisions to the data that put claims 10,000 lower than they would have been under the old methodology, a Labor official said.

Without the annual revision, claims would have totaled about 453,000

(Sorry MarketWatch, that's 11,000 lower via revisions than the old methodology, from your raw reported numbers).

Ok Folks, I have hunted high and low looking for information on annual seasonal adjustments which magically go back 5 years and so far have found no information. The DOL press release has circular link claiming to explain these revisions, which goes to basically a repeat of the original press release.

I don't know about you all and I have never thought too highly of this particular data metric from DOL, but this takes the cake. Revisions back 5 years with no explanation and magically we have an initial weekly unemployment claim report that sounds real good as a result?

Subject Meta: 

Forum Categories: 

Great post and catch

Outstanding post, of course, and the norm: I've run into endless circular links when attempting to ascertain source data, or the source of the "data."

The last verbal circular link was when I was attempting to find the originating source to that "30,000 predicted to die from flu pandemic" some months back. The CDC claimed those numbers originated out of the White House (their science council), upon contacting the appropriate party in the White House I was told they got that figure from the CDC (an examination of the CDC's site put the figures of deaths from flu, at the highest, to be around 1,000, and there was no differentiation whether those in the mortality group -- babies, young children, the elderly, those with immune system problems -- had actually been vaccinated or not).

Why all government stats are suspect? Hmmmmm....could it possibly be chronic experience with phony, unexplainable and unverifiable stats?

And again I say, the level of unemployment which correlates with all those food stamp increases, homelessness increases, and poverty increase, has to put it from 28% to 34%.

gov stats

Across the board need to be revamped in my view. It's 2010 and they should be able to get data, real time, then they need to add significant levels of detail.

Take unemployment, they need immigration status in those stats. Sorry it's a globalization/labor arbitrage age and somehow I don't think guest workers should be counted as employed, which artificially represses the real unemployment rate for Americans, esp. in some occupational categories.

That said, some stats seem to be better than others and the DOL to me really needs to be revised, #1! They could easily, in this day and age, get work stats on not just those showing up to the State unemployment offices, but temp labor, small business quarterly earnings and employees, the self-employed, those being rejected for UI, those not qualifying. Every month we have a horde of Journalists and bloggers digging around in the data, hunting high and low, having to extrapolate, guesstimate, subtract, add, do a log and on and on to try to get more accuracy and a picture of what's really going on.

EP and Bonddad

Great observation. Before I read it I read Bonddad. He and New Deal either didn't see what you did or are ignoring it. They predict 'job growth' this month. After I read your post I literally laughed. They are really a mixed bag. Sometimes they are dead on and other times I think they smoke 'green shoots'

Good catch

I don't understand the reason for this revision at all. We aren't talking about unemployment numbers. We are talking about unemployment claims. This should be straight-forward.
To go back and revise the numbers for the last 5 years requires some serious explanation.

Speaking of the unemployed

Remember when I pointed out that 5 million unemployed people will lose their benefits this summer?
Well, it's started.

As many as 130,000 Californians are expected to exhaust their unemployment benefits within the next three weeks, based on estimates from the state Employment Development Department. About 3,300 already have fallen off the unemployment rolls.
Currently, the unemployed in California are eligible for up to 99 weeks of benefits — 26 weeks of regular unemployment plus four extensions and so-called FedEd relief.
...
Congress is currently considering legislation that would allow more-recently unemployed to be eligible for some of the extended benefits as well as a 65% government subsidy for COBRA health insurance premiums. However, none of the bills would add any additional weeks of benefits for those who have received the maximum.

That's 130,000 by mid-April in just California, and Congress has no plans for addressing the problem. By mid-summer we are going to be looking at millions of very desperate people.

I thought they extended to 12-10

I saw Congress pass an extension to 12-10 for unemployment, nation wide.

I haven't kept track of all of the extensions but I'm positive this recently passed.