April 06, 2003

More Bad Unemployment News

John S. Irons reports more bad employment news:

The employment situation remains poor. Data released by the Bureau of Labor Statistics shows that the unemployment rate for March remains unchanged at 5.8%, and, in addition, total nonfarm payroll employment declined by 108,000 after seasonal adjustment.

Overall, these numbers again point to the idea that the economy is in somewhat of a holding pattern. Overall growth appears to be weak, leading to the decline in employment, yet the unemployment rate has been holding steady at just under 6% for the past year.

Given how lousy the employment numbers have been over the past year, it is remarkable that the unemployment rate has not risen by more...

Posted by DeLong at April 6, 2003 08:43 PM | TrackBack


The labor market participation rate peaked in 2000, and has since come back down to the level last seen around 1994. That suggests the jobless rate is understating the level of unemployment somewhat. However, given the participation rate in 2000, near 65% vs just over 62% now, it is very likely the jobless rate was overstated in 2000, when it cracked below 4% for a while. Now, how do we accurately gauge the loss of output when we can't figure out how much the jobless rate has "really" risen?

Posted by: K Harris on April 7, 2003 05:17 AM

> Given how lousy the employment
> numbers have been over the past year,
> it is remarkable that the
> unemployment rate has not risen by more...

Careful....you've given Mickey Kaus an opening....

Posted by: P O'Neill on April 7, 2003 06:52 AM

You beat me to it again K Harris! This is a very important point.Over the past year, the overall participation rate -the ratio of the labor force, or the number of people employed plus those looking for a job relative to the civilian population - declined by 40 basis points. Had the participation rate stayed at February 2002 levels, the unemployment rate could now be as high as 6.4%. This decline in participation is unwinding much of the per capita GDP 'good news' juice squeezed out in the 90's. When you say 'output loss' I guess you mean the notorious 'output gap': again a good point.

Perhaps instead of focusing on the unemployment statistic - dependent as it is on the 'discouragement' factor, or, as used to be the case in Holland, on how many people are classified as long-term disabled - it would be useful to develop a new metric based on the particpation rate and the dependency ratio (possibly making a composite employment/creativity index by combining this with the average age of the population: to give an advantage to those societies where the dependent population was more weighted towards the young). New metrics for a new century with new problems.

Posted by: Edward Hugh on April 7, 2003 07:34 AM

Could combined declining payroll statistics and unchanged unemployment rate be partially the result of the mobilization of the Reserves and National Guard?

Posted by: wetzel on April 7, 2003 09:51 AM

We have lost 240,000 jobs in little more than 2 years. This a severe problem and will remain so for a long time. Middle class America is being slowly harmed. I am quite worried.

Posted by: jd on April 7, 2003 10:16 AM


Could be, but you wouldn't ever be able to know. The monthly employment report frequently shows sizable changes in the level of employment without an accompanying change in the jobless rate, so finding a particular reason for the lack of response in the jobless rate in the March data is kind of tricky. Part of the reason the two series don't track well is that the job tally that everyone watches is collected from establishments (employers) while the jobless rate is calculated from household data. Those two surveys sometimes just don't agree on what is happening in the labor market. Even the household report of the number of unemployed doesn't always track well with the jobless rate, because the denominator in calculating the jobless rate, the civilian labor force, tends to wiggle around a great deal month to month.

Posted by: K Harris on April 8, 2003 04:28 AM

Edward Hugh,

My e-mail server doesn't like your e-mail server (at bonobo land). Would you mind sending me your address?


Posted by: K Harris on April 8, 2003 07:48 AM

The longer I look at them, the more I'm convinced that current employment statistics are basically meaningless. As mentioned above, there are problems with them on the level of laborforce participation, with the often nebulous difference between full- and part-time workers, and with issues created by "under-employment"--people who have to be counted as employed, but in lines of work that do not match their interests or training. Getting all of the relevant information out of one or two numbers just doesn't seem very likely to work.

Not that I have a ready answer, mind you...

Posted by: Jonathan King on April 8, 2003 12:12 PM

Jonathan King,

If one or two numbers don't suffice, would you settle for hundreds? It is a pain to do, but if you are willing to wade through endless data series, you can get a pretty good impression of what is really going in the labor market, within the limits of the accuracy of the data. The labor market participation rate itself (which led us in earlier scribbles to doubt the unemployment rate) actually comes in several different flavors every month. You can know the jobless rate breakdown between races, sexes, educational levels and brief or boxer (OK, not by brief or boxer). Want to know the average length of unemployment? Pretty soon, you can get a good idea of the state of the labor market -- just ignore that gross jobless rate figure.

Posted by: K Harris on April 8, 2003 01:02 PM
Post a comment