October 09, 2003

Employment News Not as Bad as Expected

This morning's labor market news is not as bad as expected. However, alas, it still suggests that there is upward pressure on the unemployment rate.

Jobless claims drop to 382,000 - Oct. 9, 2003: The Labor Department report said 382,000 people filed for benefits in the week ended Oct. 4 -- the lowest level since the week of Feb. 8 -- compared with a revised reading of 405,000 in the prior week. Economists, on average, expected 394,000 new claims, according to Briefing.com...

Posted by DeLong at October 9, 2003 07:32 AM | TrackBack


The labor market broke a losing streak, but one analyst questions its recovery.
October 7, 2003: 2:55 PM EDT
By Les Christie, CNN/Money Contributing Writer

New York (CNN/Money) - Though many cheered last week's news of jobs creation in September, John Challenger, CEO of Challenger, Gray & Christmas, a job placement company, notes a number of disturbing trends in the Labor Department report.

The report showed payrolls expanding by 57,000 workers, the first increase in eight months. But Challenger gave several reasons to not get too excited:

Less pay "While many celebrated, the numbers indicate that things have not improved for most workers; in some cases, workers are earning 43 percent less," said Challenger.

More part-time work One stat that concerns Challenger is that nearly 5 million people now work part-time jobs, the highest level since November 1993. "Benefits are much worse for part-time workers. The number of people without health care insurance has gone over 40 million."

Long-term joblessness Another troubling trend Challenger identifies is that long-term joblessness, those out of work for 27 weeks or longer, has grown to 2.1 million, the highest level since November 1983.

Bleak outlook from small business Challenger also points out that a survey of small businesses (fewer than 500 employees) found that only 14 percent of them plan on adding workers in the near future. "That is troubling," says Challenger, because, "Small businesses represent more than 99 percent of all employers and historically have accounted for about 60 to 80 percent of all new jobs."


Posted by: Kosh on October 9, 2003 08:10 AM

I think you're expecting too much from this indicator. Once the economy has already turned (in either direction) claims tend to reverse and continue moving in that direction.

I don't know what the reason is, but looking at a a chart of claims vs. unemployment going back 30 years, it is the change in claims which predict future unemployment, rather than a comparison with 357,000. Although in a narrower sense the comparison with 357,000 does correlate with the direction of unemployment, the fact that there is usually so much momentum in the claims number makes me much more optimistic than you are today.

For an example of this consider unemployment claims in 2000 as they started moving up and reversed the downward trend, but remained below 350,000. Unemployment was still edging down at the time, but anyone who was optimistic about that got their predictions wrong. Due to momentum behind claims numbers (again, I don't know the economic reason for the momentum), claims numbers continued in nearly a straight line up to 500,000.

Sorry but I cannot find a link for the chart I'm looking at now.

Posted by: snsterling on October 9, 2003 08:27 AM

Found it. It's the third chart down. Jobless Claims 4 week average compared to unemployment rate.


Posted by: snsterling on October 9, 2003 08:37 AM


That is a very odd little piece, from an editorial point of view. British English text, with Dutch captions (except for some English mixed in), commenting on US labor market data. Very cool.

I understand John Challenger's concern about the wide variety of weak employment stats he cites. They have mostly to do with the impact on workers (or former workers). That's important. However, to the extent these data represent the current impact of earlier conditions in the labor market, they don't tell us much about where we are going. A look at new vs continuing claims, for instance, shows that layoffs are slowing, but that there is little change in the number of people on jobless benefits since July. It takes getting people off the jobless roles to bring down the duration of unemployment. Slowing the pace of layoffs is actually likely to lengthen the average time of unemployment, since there will be fewer recently unemployed in the mix. The employment diffusion index is at its highest level in over 2 years. Layoff announcemens and new claims are down. Weekly hours are still very low, which is not encouraging, but overall the mix of labor market indicators is as healthy as it has been in some time. Not out of the woods yet, by any stretch, but better.

Posted by: K Harris on October 9, 2003 09:07 AM

Don't worry. Next week, this number will be revised upward, to about 395,000-399,000. And a few weeks from now, it will be revised upward again.

Just as last week's number was just revised upward, over the "magic" 400,000 number.

What you smell is books being cooked.

Posted by: Chuck Nolan on October 9, 2003 09:12 AM

every single time this number gets reported, it is obviously too low....and the headlines read something like "Claims drop"...and then pundits and stock brokers are POSITIVE that this time is the Big One (tm)....and then without fail the numbers are ALWAYS revised up....

why does anyone still take this obviously flawed/cooked number seriously? Why not only report it after it has been revised?

didnt they just decide to "revise" the baseline numbers from march too?

Posted by: sampo on October 9, 2003 09:53 AM

A step in the right direction. I like that positive September employment reports came out in Germany and Australia today. Perhaps new jobs will be created in the developed world. And I am not being sarcastic, sometimes I really fear that they won't be again for many years. Hopefully a sign we are all stepping out of the muck.

Posted by: Joe Blog on October 9, 2003 10:10 AM

Careful, the employment report for September and the beginning of October are a bit better than the reports since the recovery from recession began in November 2001. But, we are not generating enough jobs to keep from adding to the jobless problem. Job growth is too slow to account simply for the population growth aspect of added labor supply. Month after month we are building a deficit of job creation.

The labor department has just told us that an added 110,000 jobs were lost from April 2002 to March 2003. We have been shedding jobs as never before during a recovery, and now are adding too few jobs to even account for new entrants to the labor force.

Better by a bit, but still quite worrisome.

Posted by: Anne on October 9, 2003 10:54 AM

So here's the thing. The claims data do get revised, often upward, and still the trend is toward lower claims. "They" didn't "decide" to revise the payroll figures from March on back. Benchmark revisions are a regularly scheduled event. More recent revisions have shifted the trend in payroll jobs from down to up. In fact, the timing of the improving trend in new claims and the less ugly trend in payroll jobs is very consistent.

I keep hearing the idea that the data are cooked, but I have yet to see a shred of evidence offered. Anybody?

Posted by: K Harris on October 9, 2003 12:14 PM

The revised employment figures are certainly not cooked, simply increasingly more precise figures. The revision for March 2002 to April 2003 showed 110,000 more jobs lost, while the revision for August 2003 showed fewer jobs were lost in that month. The monthly figures are improving but still not enough to keep the labor market from worsening.

Posted by: anne on October 9, 2003 12:30 PM

K Harris:

I asked that same question here last week. Revised figures generally seem to have been better than the original numbers since I have been following them. Were revised figures generally worse before the recovery started when unemployement claims were rising?

So what is the story, do revised figures reflect the general trend, are the early numbers high because they are designed to be conservative, or is it just random chance?

That aside, I am so relieved by the skeptical posts here. When I saw the good economic news today I was afraid I would not be able to get my daily dose of doom and gloom on Brad's weblog. I needn't have worried.

Delong's Daily Dose of Doom and Gloom. Catchy, isn't it?

Posted by: Jason on October 9, 2003 12:40 PM


Yes, there really really is an employment problem and it is severe and deepening, though there are finally signs of a turn round.

"Payrolls are down 2.7 million since the recession began in March 2001 and one million over the recovery.

"BLS's preliminary estimate of its annual adjustment to the payroll survey is -145,000. If this value is not substantially revised, the BLS will adjust the payroll survey to show that 145,000 fewer jobs existed in March 2003"

Posted by: anne on October 9, 2003 01:09 PM

Yes, the figures get revised about 1.5 to 2% upward within a week of listing. That occurs pretty consistantly but it wouldn't have much effect. If you're suspicious you could just follow the rolling average, which is published as well. Remember that these figures are seasonally adjusted--the real figures are lot lower, but seasonal adjustments include the relative hiring that's supposed to take place in this week of the year.

Anne supplied a link which is far more helpful in understanding the actual trends. Nothing here I would find to rejoice and be glad in. Again thats

Posted by: James R MacLean on October 9, 2003 11:16 PM

My problem with the repeated suggestions that the data are cooked is that there are so many people involved, not all of whom would share the political motives of whatever president might decide to make up happy numbers. Some things, like covering up the identity of criminals in the administration, involve small numbers of people who, if they don't share exactly the same motive, at least can be brought to understand that they should agree on silence. With states reporting claims data, with the next administration in charge of benchmark revisions, with hundreds of bureaucrats involved in counting, tallying and adjusting, I have never heard a peep about jiggering the BLS data - and I've been listening.

Posted by: K Harris on October 10, 2003 04:08 AM

Thanks for the link, Anne. No, it's not worth rejoicing in... but it IS enlightening.

Posted by: Theodore Salizar on October 11, 2003 01:31 PM

Thanks for the link, Anne. No, it's not worth rejoicing in... but it IS enlightening.

Posted by: Theodore Salizar on October 11, 2003 01:34 PM

Thanks for the link, Anne. No, it's not worth rejoicing in... but it IS enlightening.

Posted by: Theodore Salizar on October 11, 2003 01:34 PM

Sorry for the triple posting (computer kept timing out) :-(

Posted by: Theodore Salizar on October 11, 2003 01:38 PM
Post a comment