September U.S. Employment Report Shows Modest Private Gains, Heavy Government Cutbacks

The big story was the state and local cutbacks, of which 50,000 were in local education. That means fewer school teachers. The federal government's extra assistance to education probably came too late to help—but may reverse some of the job cuts in coming months. Education took a big hit everywhere, losing 73,000 jobs in the private and public sectors combined.

Combining September's state and local cutback with the revisions to July and August leaves state and local employment down 162,000 so far in the new fiscal year (which began in July for most states and localities). That is the biggest three-month decline on record, in data going back to 1955, as states and localities continue to struggle to balance their budgets.

In the private-sector details, manufacturing lost 6,000 jobs. That is less than the 28,000 drop in August, but that decline had been exaggerated by seasonal quirks in the motor vehicles sector. Excluding motor vehicles, September was down 7,000, in line with August's 6,000 loss.

The manufacturing workweek edged down to 40.1 hours, from 40.2 hours, and production-worker hours fell 0.1%. Manufacturing is still growing, since productivity gains mean that output is still rising, but it has lost momentum as the impetus from the turn in the inventory cycle has faded.

Construction lost 21,000 jobs, reversing most of a 31,000 increase in August (of which 10,000 had been strike-related). There were 3,000 jobs lost in residential construction and 18,000 in nonresidential construction. Residential construction activity is already extremely low; the nonresidential sector has greater downside.

In the private-services sector, 86,000 jobs were added in September. Job creation in services has been very stable in recent months (83,000 in August, 80,000 in July). The health sector added 24,000 jobs, in line with recent months, but there was a surprise 34,000 jump in food services and drinking places (surprising since the previous five months combined had only added 13,000 jobs). This jump looks odd and it is likely that there will be a reversal next month. On the downside, private education services (private colleges and schools) lost 15,000 jobs, also cutting back at the start of the school year.

Business services added 14,000 jobs, more than accounted for by a 17,000 increase in temporary help, very similar to August's 18,000 increase. Temporary help is usually regarded as a leading indicator of permanent hiring, but it has not worked that way so far in this recovery, instead just signaling a sustained increase in reliance on temps.

The government overall shed 159,000 jobs, driven down by the 83,000 loss in state and local governments and by a loss of 77,000 temporary Census workers. Since only 6,000 temporary Census workers remained in place as of mid-September, Census changes will have a negligible influence on future employment reports.

The private workweek was steady, at 34.2 hours, and with private employment barely higher and the workweek flat, that left total hours-worked in the private sector unchanged. Hours worked rose 1.9% overall at an annualized rate in the third quarter, slower than the 3.3% advance in the second. September average hourly earnings were flat, so when combined with no increase in hours, that left total private payrolls up only 0.1% in total, giving little extra spending power for those in work.

The unemployment rate was steady at 9.6% as household employment rose by 141,000 and the labor force rose by 48,000. As a result, unemployment fell, just not enough to make the rate tick down.

But the most comprehensive measure of underemployment (U-6)—which includes workers who would like a job but are not currently looking, plus those working part time who would rather work full time—rose from 16.7% to 17.1%. That reflected a sharp 612,000 increase in those who could not find full-time work. That indicates plenty of slack still in the workweek before employers need to re-hire.

The announcement that the "benchmark" revision to employment levels in March 2010 is likely to be a reduction of 366,000 (0.3% of the employment level) does not come as a surprise. The benchmark figure comes from a complete employment count using data from unemployment insurance contributions. Without the benchmark data, statisticians must make assumptions about births and deaths of new firms, and in hard times their assumptions about births tend to be too optimistic. The revisions will mean that monthly employment changes between March 2009 and March 2010 will be revised down by just over 30,000 per month in the data release due on February 4, 2011.

The payrolls report suggests that the economy continues to crawl forward at a sub-2% growth pace, enough to generate a handful of private-sector jobs but not enough to prevent the unemployment rate from rising further in coming months. We believe that the figures leave the Fed on a path to a new program of quantitative easing in November, but they do not settle the debate. State and local employment will not be such a drag in coming months. If the undershoot had been in the private sector, that probably would have sealed the case for QE II, but the private sector continued to crawl forward.