US temporary jobs


6 Aug 2014

The YOY change, which typically leads changes in non-farm payrolls by about 5 months, does not suggest any improvement in job growth.

The current series from the Bureau of Labor Statistics began in 1990.


6 Aug 2014. Data through Jul 2014.

The data are from BLS (see Sources for details). It is visually apparent that the YOY change in temporary jobs tends to lead the YOY change in non-farm payrolls by a few months. (Click on graph for a larger image.)


Clippings below covered through 3 Sep 2010.

  • Definition (4 Jun 2010) A common misperception is that “temporary help services” refers to all temporary jobs. In fact, the classification refers to employees of temporary help agencies, including both the permanent staff who run these agencies, and those employed by the agencies and supplied temporarily to other firms. Nevertheless, the changes in temporary help services do indicate something important about temporary hiring overall.
  • Leading indicator (3 Sep 2010) See 5 Nov 2010 clipping for a scatter plot. It would appear that an increase in temporary jobs is still a positive sign for non-farm payrolls. However, although the qualitative relationship seems to be intact, the quantitative relationship is different this time.


  • BLS. On the home page, select tab “Databases and tables”; under the main heading “Employment”, subheading “Employment, Hours, and Earnings - National”, click “Top picks”; tick “Temporary help services Employment” (5th from the bottom of the page) and click “Retrieve data”; click “More formatting options”; tick “All years”, “Original Data Value” and “12 month percent change”, then click “Retrieve data”

See also

Selected commentary:

Clippings below were used in the construction of this page

Temporary help services definition

Nov 1986. Monthly Labor Review.

“Temporary help workers: who they are, what jobs they hold. WAYNE J.HOWE”

“It should be noted that the temporary help supply employment industry level derived from the cps is not strictly com-parable to that obtained from the Current Employment Statistics (CES) program. The monthly CES survey derives its estimates from the payroll records of a sample of establishments. In the case of temporary help workers, the CES estimates relate to the total number of jobs in the industry, including those of the personnel who operate the temporary help agencies. By contrast, the new cps data relate only to those workers who view their jobs as temporary. Thus, these data exclude not only the permanent employees who staff the many agencies, but may also exclude many of the workers who do not view their jobs as temporary, as they have a fairly continuous attachment to the temporary help agency. This explains at least part of the reason the number of temporary help workers identified in the cps in May in the temporary help industry as measured through the CEs-689,000. The cps data, while not representative of all the employees in the industry, are representative of the preponderance of them, particularly of those whose jobs are truly temporary. And, the cps supplies the only information available on the characteristics of these workers.”

[Note, the current series started in 1990, but this was just a change in the numerical classification system from SIC to NAIC. See undated clipping that follows this one.]

The place of temporary help services in the current classification

Undated. BLS page on NAICS cods for the CES.

“CES Series for Professional and Business Services Under NAICS”

561  	Administrative and support services  	60561000  	D  	1990
5611 	Office administrative services 	        60561100 	D 	1990
5612 	Facilities support services 	        60561200 	D 	1990
5613 	Employment services 	                60561300 	D 	1985
56131 	Employment placement agencies 	        60561310 	D 	1990
56132 	Temporary help services 	        60561320 	D 	1990

[The last column is “Beginning date of AE series]

Temporary help services definition, more recent

June 1998. BLS Working Paper

“The Implications of Flexible Staffing Arrangements for Job Security. Susan N. Houseman and Anne E. Polivka”

“Our temporary help agency category includes all of those who state they are paid by a temporary help agency. Thus, it includes the permanent staff of these agencies, though they are expected to represent a relatively small percentage of this category.

Table 1 reports the distribution of employment by employment arrangement according to data from the February 1995 CPS. It is interesting to note that agency temporaries account for only about one percent of total employment in these data. In the Current Employment Survey (CES), the Bureau of Labor Statistics’ establishment survey, the percentage of employment accounted for by agency temporaries is about double that number. …

The CES provides information on employment in the help supply services industry, SIC 7363, which is comprised primarily of temporary help agencies. From 1982 to 1996 the share of non-farm payroll employment in help supply services rose rapidly from 0.5 percent to 2.0 percent.”

Temporary help a leading indicator, with varying lag

Jun 2009. American Staffing Association report.

“Looking ahead for growth. Steven P Berchem”

“ASA analysis shows that temporary help employment has been a strong two-quarter leading indicator of nonfarm employment during periods of normal economic growth and a modest one-quarter leading indicator when the economy is emerging from a recession.

Most immediately, this means that as the economy comes out of the current recession, overall nonfarm job growth will likely resume about three months after temporary help employment begins to grow.

Two or three years from now, after the economy has fully recovered from the current recession and resumed normal growth, changes in temporary help em- ployment would likely be reflected in overall nonfarm employment in about six months.”

Source cited by Calculated Risk for caution on temps

6 Dec 2009. SF Chronicle.

“In economic woes, firms count on temp workers. Tom Abate”

“BLS economist Amar Mann said an analysis by the San Francisco office suggests that employers are getting more sophisticated about using temp hiring as a clutch to downshift into recessions and upshift into recoveries.

Mann said temp jobs started down a month after overall employment dropped during the 1990-91 recession. But by the 2001 downturn, employers started cutting temps about five months before they started issuing pink slips to the general workforce.

In the current recession, he said, companies began shedding temps 12 months before they started cutting permanent payrolls.

A similar pattern prevailed in the two prior recoveries, Mann said. Temp jobs came back at the same time as overall employment after the 1991 recovery. Temporary employment rebounded five months before the general job market turned positive following the 2001 dip.

If that pattern holds, it could be next summer before general payrolls start to grow.

Mann refused to speculate about the timing, but said temps are playing an increasing role in the job cycle.

“Employers are getting more savvy about using just-in-time labor on the way down and on the way up,” he said.”

Rosenberg says temporary no longer leads NFP

8 Feb 2010. Gluskin Sheff.

“Breakfast with Dave. David A Rosenberg”

“Of course, there are economists out there who see that temp agency jobs going up 52k and the rise in the workweek as “leading indicators” of future job creation because these measures worked in the past. But this is a different cycle than its other post-WWII predecessors — a credit contraction and wealth destruction cycle of massive proportions. Businesses have adjusted their order books, production schedules and staff requirements in line with a level of overall private sector credit that is $1 trillion smaller now than it was a year ago. And, as we saw in December with the $2.7 billion decline in consumer credit (11th contraction in a row, the longest stretch since the records began in 1943) this deleveraging cycle is ongoing. It may well be the case that temp agency employment and the workweek adjustments have actually become new cost- efficient strategies to manage the labour force in this new paradigm of credit contraction rather than traditional “leading barometers” of job growth. This is the time to think out-of-the-box and not to rely exclusively on relationships that worked in past cycles that basically have little in common with today’s post- bubble economic and financial realities.”

Current buzz says temporary services no longer leads NFP

15 Feb 2010. AP on Yahoo.

“Use of temps may no longer signal permanent hiring. By Jeannine Aversa”

“When employers hire temporary staff after a recession, it's long been seen as a sign they'll soon hire permanent workers.

Not these days.

Companies have hired more temps for four straight months. Yet they remain reluctant to make permanent hires because of doubts about the recovery's durability.”

Census hiring was probably mostly direct-hire

Leading indicator status weakening but intact

3 Dec 2010. My calculations.

For offsets of 0 to 10 months, the correlation was measured between the YOY change in temporary jobs and the YOY change in non-farm payrolls that number of months later.

Here are the results from Feb 2010:

Months  Pearson correlation coefficient

0   0.8736977
1   0.9044646
2   0.9239108
3   0.9342608
4   0.9373071
5   0.9370345
6   0.9313207
7   0.9207776
8   0.906417
9   0.8869937
10  0.863637

And here is the calculation for data as of Aug 2010:

[1] 0.000000 0.811123
[1] 1.0000000 0.8570393
[1] 2.000000 0.891816
[1] 3.0000000 0.9168012
[1] 4.0000000 0.9296742
[1] 5.00000 0.93329
[1] 6.0000000 0.9254738
[1] 7.000000 0.908802
[1] 8.0000000 0.8856485
[1] 9.0000000 0.8572294
[1] 10.0000000  0.8258115

Note that all along the optimum was rather ambiguous. In February a lag of 4 won by a hair. Later it was 5.

The following is a scatter plot of temporary help services, YOY percent, against non-farm payrolls, YOY percent, with the temporary help services 5 months earlier than the NFP. Data are through Nov 2010.