U.S. Employment Report Shows Markedly Slower Job Decline and Lower Unemployment Rate

The much-better-than-expected November employment report suggests that many employers have reached the point where they cannot extract more from a shrinking workforce. To keep output growing, they now have to add hours and jobs. The report shows sharp improvements in key "leading indicators," with both the workweek and temporary jobs higher. And although job losses continued in manufacturing and construction, the private services sector added jobs for the first time since December 2007.

The report cannot be dismissed as the result of seasonal quirks. It is mostly genuine good news. The argument goes that the seasonal adjustment process would treat part of the plunge in employment during November 2008 as a seasonal dip rather than a cyclical decline, and would distort the seasonal adjustment for November 2009, propping up employment. Looking at the historical pattern of November seasonals, one can make the case that a more "normal" seasonal factor would have put the job loss at around 60,000, not 11,000. But that is still massively better than the 190,000 loss originally announced for October. In addition, two-thirds of the upward revisions to September and October came in the raw, unadjusted data, not in the seasonal factor.

In the details, manufacturing jobs fell by 41,000 in November, fewer than in October, but showing less improvement than construction or services. Most major manufacturing segments still showed declines. But the payroll count is probably giving too gloomy a signal of the manufacturing labor market. Manufacturers are adding hours, as the workweek jumped from 40.1 hours in October to 40.4 hours last month, its longest since October 2008. Overtime rose from 3.3 to 3.4 hours. And it is likely that some of the increase in temporary help employment (which is classified in the service sector) reflected workers being deployed in manufacturing. Overall manufacturing hours worked rose 0.4%, signaling that production increases can no longer be achieved by productivity gains alone.

Construction payrolls fell by 27,000, much less than the 56,000 drop in October. The contrast between the residential and nonresidential sectors became even starker. The losses in the residential sector were only 3,000 this month, but the nonresidential sector lost 29,000 jobs. This is consistent with the story that residential construction activity is bottoming out, but that nonresidential has much further to fall. There were 5,000 jobs gained in heavy and civil engineering construction, perhaps helped by fiscal stimulus.

In the private services sector, jobs were added (51,000) for the first time since December 2007. Two major segments added jobs—professional and business services added 86,000 (up from 38,000 in October), while education and health services added 40,000 (the same as in October). The strongest job gains in the professional and business category were in employment services (61,000), of which temporary help contributed 52,000. Other service areas generally continued to shed jobs, but less rapidly than in October.

The government added 7,000 jobs, with the federal government up 1,000 and state and local governments up 6,000. October's originally announced decline of 16,000 state and local government jobs was revised to a 29,000 increase, somewhat surprising given state and local government budget problems.

The unemployment rate fell from 10.2% to 10.0%, reversing half of October's increase. The household survey showed employment up 227,000, and the labor force down 98,000. It is much better to focus on the payroll employment count than the household count, since the household survey is smaller than the payroll survey, so its employment numbers are more volatile from month to month. The household survey had shown huge job declines during the previous two months, so this month's increase should be viewed as a correction. The main beneficiaries of the lower unemployment rate were the short-term unemployed (those out of work for less than 15 weeks). Those most recently laid off seem to be the first to come back. The number of long-term unemployed continued to rise, and they will be the last to benefit from the turn in the labor market.

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—fell from a 17.5% to a 17.2% rate.

The increase in the manufacturing workweek helped move the overall workweek up from 32.0 hours to 32.2 hours. Total hours worked jumped 0.6% in November, more than reversing October's decline. Average hourly earnings growth was weak at 0.1%, but with hours higher there should still be a sharp increase in wage and salary income for November.

Overall, today's report says that we are nearer a bottom in the labor market that we thought. But we are not yet convinced that employment has troughed, nor that the unemployment rate has peaked. The November report looks better than one would have expected based on other evidence, such as the ADP employment report and the ISM non-manufacturing survey. So, it probably overstates how much the underlying picture has improved. But we can now be more confident that even if the bottom is not quite here, it will be no later than the first quarter of 2010.