United States non-farm jobs fell by 95,000 jobs in September driven by a 159,000 loss of government jobs that was offset by a 64,000 gain in private sector payroll jobs. The government job losses resulted from Census cutbacks and continued local government sector declines. Non-farm job growth has been negative for four months now, i.e. technically a double-dip has already occurred, however for three of those months the job losses were small.

The unemployment rate held steady at 9.6 percent due to labor force gains of 48,000 being offset by household survey employment gains of 141,000. This is the second month that the household survey and the establishment survey employment numbers do not match well. This might indicate that folks are deciding to work for themselves, which happens when establishment job growth is weak.

Today’s Employment Situation press release also contained revisions and estimated revisions. July nonfarm job change was revised down 12,000 to -66,000, and August nonfarm job change was revised down 3,000 to -57,000. The BLS also provides advance estimates of their upcoming annual benchmark revision in February to March 2010 establishment payrolls, currently estimated to be a reduction of 366 thousand jobs.

The private sector payroll gains were mostly in Professional and Business Services, Education and Healthcare, and Leisure and Hospitality sectors. While Education and Healthcare jobs have grown for most months in this recession, Leisure and Hospitality jobs have declined in most of the months since mid-2008. It is unlikely that the Leisure and Hospitality job growth is sustainable in a relatively weak recovery such as this one. The Professional and Business Services job growth, while weak, is welcome since these are decently-salaried jobs.

Last month I forecasted a gain of 30 thousand September jobs based on about 70,000 private sector gains offset by about 40 thousand government job-losses. I refrained from forecasting labor force and household survey employment growth, just saying that the unemployment rate would not move much.

My October non-farm job forecast is based on these factors: ongoing state and local government job-losses and continued private-sector job growth at a low level. If private gains were 75,000 and government losses where 60,000 this would yield a 15,000 non-farm job gain. Again, I will refrain from projecting the components of the October unemployment rate, only to say that it is not likely to change much from September.