October 23, 2003

Unemployment Insurance Claims

Those of you who remember that last week's New Unemployment Insurance Claims number was 384,000 may be surprised to learn that this week's number--386,000--is described as a "decrease."

Wampum, however, is here to explain all:

Wampum: New claims: Rinse, spin, repeat...: In case anyone missed it, here's last week's ruminations on newly filed unemployment claims:

Claims are up...no, wait, they're down...or up?

From the DOL press release this morning:

In the week ending Oct. 11, the advance figure for seasonally adjusted initial claims was 384,000, a decrease of 4,000 from the previous week's revised figure of 388,000. The 4-week moving average was 390,750, a decrease of 4,250 from the previous week's revised average of 395,000.

Of course, what the release doesn't tell us is that the 388K number was "revised" upwards by 6,000, from 382K last week. So while we're told that claims have again "dropped" this week, they actually increased by 2,000 from last week's "preliminary" number.

Confused yet?

For this week, change the revised figure to 390K, once again up 6,000 from last week's rosy preliminary figure of 384K. Change the current, also preliminary, number to 386K. Everything else stays the same...

Posted by DeLong at October 23, 2003 11:18 AM | TrackBack

Comments

As billmon said in a post last time this happened, the chocolate ration has been raised from 22g to 20g.

Posted by: Keith on October 23, 2003 11:35 AM

As billmon said in a post last time this happened, the chocolate ration has been raised from 22g to 20g.

Posted by: Keith on October 23, 2003 11:39 AM

I think we should form a pool on what this week's number will be revised to. I'll put $1 on 391,000.

Posted by: joe on October 23, 2003 11:54 AM

Darn, I simply do not see any significant gains in job creation. My guess is 6% GDP growth last quarter, back to 3 to 4% this quarter, then less. Employment is the need, the numbers are poor.

Posted by: anne on October 23, 2003 11:59 AM

Wow, the people at Labor have discovered a "Perpetual Falling Jobless Claims Machine". Revise, claim decrease, repeat.

On a more serious note, doesn't it seems like Labor's statisticians should make take this phenomenon into account? Over the last 4 weeks, the revisions (in thousands) have been +5, +6, +6, +6 unemployment claims. The pattern doesn't seem that hard to predict.

Posted by: Kash on October 23, 2003 01:17 PM

It's not the methodologie that's the problem,
it's the way it's presented that's dishonest.

I wrote a nasty letter to my Local Paper for the headline (Unemployment falls to lowest since last month!) it put on the AP writeup.

Posted by: Patrick (G) on October 23, 2003 02:16 PM

anne - Let me see if I get this straight. No significant employment growth, then no significant demand growth. No significant demand growth, then no sustainable GDP growth. No sustainable GDP growth, then no significant employment growth. Repeat.

Is there any hope?

Posted by: joe on October 23, 2003 02:23 PM

All recessions have self reinforcing aspects (as do expansions). Things will definitively turn around at some point and there is no time like the present. I can easily see people looking back at this point and observe:

1. Steadily rising stock market over the past year
2. Several quarters of rising business profits
3. The most productive work force in the world
4. Massive fiscal and monetary stimulus

And so forth and say that it should have been obvious that the economy was on the upswing. Time will tell.

Posted by: Joe Blog on October 23, 2003 04:42 PM

Joe Blog:

Yours was a kindly and gently form of the spin I read at NRO at how incredibly this recovery is. First the author talks about GDP growth - which by my calculations would leave us only $470 billion short of full employment. Then he spins the employment growth predictions in a way that Brad DeLong would have a lot of fun with (since he has been doing so quite rightfully of late). He also talks about higher profits and higher investment relative to last quarter but still far below 2000 levels. And the really funny thing wa the recovery in manufacturing, in industrial goods, in durable goods. This is sort of like the witch doctor who gives a well man some poison which makes him so ill that his temperature shoots up to 108 degrees and then gloats when the temp falls to 107 degrees.

Posted by: Harold McClure on October 23, 2003 08:45 PM

The fact that the 4-week average of new claims has fallen from the recent high of 459k in early May to 392k most recently seems to be overshadowed in some folks thinking by the steady 5-6k weekly revisions we've been seeing. Is it really more important that there is an apparent small downward bias to the first report of weekly claims than that there has been a fairly substantial decline in recent months? I detect a serious forest-and-trees problem here.

OK, there are serious problems yet to get through in the US labor market and the wider economy. However, looking at small technical problems (that 6k revision amounts to 1.6% of the total) and pretending they amount to something big is not going to help you point out those problems to those who don't already agree with you.

The thing to focus on, if you really want to understand what is going on, is the level of claims and the trend. Not as exciting as looking for criminal statistical behavior, I know. The trend is clearly improving, but the level remains quite high. The average level of new claims was below 350 from 1997 through 2000. In 2000, it was around 300. The 4-week average of 392 is not a sign of labor market health. The 92K gap between the recent average and the average though 2000 is a lot more important that the recent weekly obsession with revisions.

Posted by: K Harris on October 24, 2003 04:20 AM

What we all knew, and the new job loss - job gain
data shows that their are two aspects to employment. The claims data only deal with the job loss side of the equation. But the new job loss - job gains data implies that the problem is much more on the jobs being created side of the equation rather than the job loss side. It is a shame that this new series is quarterly and has such a reporting lagg. But the data shows the jobs loss(as a % ) are below level through most of the 1990s. But the job gains data is just as
low (as a % ) as the job loss data and is far, far, far below 1990s levels. In this context it is hard to have much confidence that the modestly lower claims data means employment is improving.
The employment diffusion data is not bullish either.

Posted by: Spencer on October 24, 2003 05:44 AM

Harold,

I am not trying to say the economy and job market is in good shape, it clearly is not. It's just that the sentiment here tends so far towards doom and gloom, that I find it hard not to point out that there are numerous positive leading indicators.

Posted by: Joe Blog on October 24, 2003 06:22 AM

The unemployment figures HAVE to be up! I just filed!

Posted by: Wayne on October 24, 2003 08:26 AM

The important story in this week's report wasn't initial claims. It was continuing claims. Continuing claims in the week of October 4 were revsied down 50,000 and then continuing claims in the week of October 11 plunged another 80,000 to the lowest level in six months. This figure picks up the impact of firings and hirings and was thus a very good sign for October employment.

Posted by: Bill on October 24, 2003 08:42 AM

Bill is on the beam, and also provides Spencer with the more timely data he is looking for. Note also that the Labor Department reports mass firings data a few weeks after the payroll data for each month. Mass firings are down sharply. That doesn't mean smaller firings (less than 50 at a pop) can't make up the difference, but there is no particular reason to think smaller firms would be following a trend just opposite to that of larger firms.

Posted by: K Harris on October 24, 2003 09:22 AM

"anne - Let me see if I get this straight. No significant employment growth, then no significant demand growth. No significant demand growth, then no sustainable GDP growth. No sustainable GDP growth, then no significant employment growth. Repeat.

Is there any hope?"

Joe - Sorry, I was away. This summary is precisely my worry. The problem is subtle. There will continue to be GDP growth, but growth that is likely to slow to account for productivity gains and natural labor force gains. We need GDP growth fast enough to add more than 125,000 jobs a month, month after month, to cut unemployment to decent levels. With more potent fiscal stimulus we could have had that, now I am worried.

Posted by: anne on October 24, 2003 12:15 PM

Joe Blog -
"It's just that the sentiment here tends so far towards doom and gloom"

That's what I come here for - Delong's Daily Dose.

Posted by: nordic on October 24, 2003 01:52 PM

Bill, what do we know about continuing claims? Can we tell how many people fell off because their eligibility terminated?

Posted by: Masaccio on October 24, 2003 02:31 PM
Post a comment