December 08, 2006
Employment remains solid
This week's employment data do not show an economy in recession.
The Bureau of Labor Statistics reported today that nonfarm payrolls showed a seasonally adjusted net increase of 132,000 workers in November. That's probably not enough to keep up with the growing population, and the unemployment rate indeed did edge up slightly from 4.4% to 4.5%. But a 4.5% unemployment rate is still well below the U.S. historical average over the last half century of 5.6%.
The BLS also modestly revised down its estimate of October employment growth from the initial figure of 92,000 jobs to a new estimate of 79,000. The September figure, which had initially been reported at 51,000, had tripled in last month's revision to 148,000 and has now quadrupled to 203,000. Other than October, all five months between May and September have been revised upward, on average by 60,000 jobs, and we have not yet heard the final word on revisions for October. Barry Ritholtz wisecracks
The initial BLS data does retain some merit, however, in that it is now a form of entertainment.
Dave Altig found cheer in the strong improvements in three other separate sources of data: the Monster Employment Index, the Automatic Data Processing Employment Report, and new filings for state unemployment benefits. ADP estimated U.S. total November employment growth of 158,000 based on the quarter-million establishments whose data they manage, while a fourth, yet again independent series (the BLS household survey) estimated a net increase of 277,000 people who were employed in November.
| BLS payroll |
| BLS payroll |
| BLS payroll |
| ADP |
| Household |
| Weighted |
I've argued for combining the BLS and ADP payroll with the BLS household survey data, giving the BLS payroll figure 8 times the weight of either ADP or the household. That weighted average suggests an initial estimate of November employment growth of around 150,000, or the sort of level you'd expect if economic growth was chugging along solidly.
Obviously you can't read too much into one month's numbers, even when the data are more reliable than these. And Calculated Risk has some concerns about construction and retail employment. But the employment indicators we have, and there are quite a few, all give a clear signal that no downturn in total employment has begun at this time.
Posted by James Hamilton at December 8, 2006 09:52 AMdigg this | reddit
Do you see employment as a lagging indicator? If so approximately how much of a lag?
Posted by: Dick at December 8, 2006 12:49 PM
No, but it shows a economy that is slowing down. That was not a impressive report compared to past Novembers. The seasonal trade made up alot of the "job growth".
December and January will grow more bearish.
Posted by: Dick at December 8, 2006 02:18 PM
But James...that entertainment from The Sweater...twas about keeping a straight face till 2 or 3 months have passed, yes?
And so this B hastey:
But the employment indicators we have, and there are quite a few, all give a clear signal that no downturn in total employment has begun at this time.
as those very signals have not yet settled.
Would it help with these revisions if the accuracy was also understood to improve? So that the first release would never dare to suggest accuracy within 1000 jobs as per 51,000 say. No, just round it to the nearest 100,000 so sweaters don't get all excited. And then after a month, to the nearest 25,000 (50,000? I'm easy) and then maybe we can get it down to the nearest 1,000 (I can be generous and professional).
Given all the extra accountants working over Fannie's position, (soon to close, forcing many sharp pencils to the street) surely there is an opportunity here to extend the series from 'final' to 'ossified' by pencilling in something confirming, with perhaps small adjustments, those 'final' numbers.
Or we could just wait the 2 or 3 months, and demonstrate professional calm.
Posted by: calmo at December 8, 2006 07:05 PM
So which survey are real estate agents still counted as "employed" under?
Posted by: Aaron Krowne at December 8, 2006 08:21 PM
Unemployment tends to drop significantly just before every recession. It's only after a recession starts that unemployment increases. As shown by the chart at http://research.stlouisfed.org/fred2/series/UNRATE, the unemployment rate is a lagging indicator. In fact, if anything you could say that periods of very low unenmployment actually precede recessions since they always coincide with the top of the economic cycle (see graph)
Posted by: olairet at December 9, 2006 05:00 AM
Dick and Olairet, initial unemployment claims are included in the Conference Board's index of leading indicators, nonfarm payroll employment in their coincident index, and duration of unemployment in their index of lagging indicators. We have plenty of discussion here at Econbrowser about leading indicators, but it's nuts to look only at leading indicators, and I stand by my statement: these data do not show an economy in recession. See also the as always excellent discussion at Macroblog.
Posted by: JDH at December 9, 2006 07:18 AM
and neither did the November 2000 numbers, nope no recession close(grimaces). James, it is the trend, not the present case figures that force the future's hand.
Posted by: Dick at December 9, 2006 12:53 PM
Unemployment may be at its low point at the peak of the cycle, but reaching a new low does not by itself tell you that you have reached the peak. No more than reaching a higher elevation says you have reached the peak of the mountain.
After the cycle peak has been passed it will be a natural effect of the slowing economy to see rising unemployment. So saying that unemployment is lowest at the peak is not much more useful than saying the mountain is highest at the peak.
Posted by: Zephyr at December 9, 2006 04:07 PM
Although 132,000 may not be enough to keep up with population growth, it may be enough to keep up with growth in the labor force due to declining participation rates.
There is a recent research paper by an economist at the Atlanta Fed that discusses this.
Posted by: GWG at December 11, 2006 07:57 AM
Good work JH,
If job growth is "solid"...
Would you like a healthy McSalad with your Big Mac?
Posted by: The Nattering Naybob at December 11, 2006 11:16 AM
There seems to be another Dick here so I am adding an "F"
JDH, Thanks for answering my question on employment as a lagging indicator. I was not making a judgement on whether it should or should not be analyzed. I was only interested in how you weighed it in your analysis.
Posted by: DickF at December 11, 2006 01:36 PM